*3,479*
*1,185*
*43MB*

*English*
*Pages 655
[656]*
*Year 2021*

Table of contents :

Chapter 1

Chapter 2

Chapter 3

Chapter 4

Chapter 5

Chapter 6

Chapter 7

Chapter 8

Chapter 9

Chapter 10

Chapter 11

Chapter 12

Chapter 13

Chapter 14

Chapter 15

Chapter 16

Chapter 1

Chapter 2

Chapter 3

Chapter 4

Chapter 5

Chapter 6

Chapter 7

Chapter 8

Chapter 9

Chapter 10

Chapter 11

Chapter 12

Chapter 13

Chapter 14

Chapter 15

Chapter 16

- Author / Uploaded
- Adel S. Sedra
- Kenneth C. Smith
- Tonny Chan Carusone
- Vincent Gaudet

EIGHTH INTERNATIONAL EDITION

Instructor’s solution manual for

Microelectronic Circuits Adel S. Sedra University of Waterloo

Kenneth C. Smith University of Toronto

Tonny Chan Carusone University of Toronto

Vincent Gaudet University of Waterloo New York Oxford OXFORD UNIVERSITY PRESS

Sedra_FM_BM.indd 3

9/30/2014 9:36:34 PM

Exercise 1–1

Rs

Chapter 1

Solutions to Exercises within the Chapter

vs (t)

Ex: 1.1 When output terminals are open-circuited, as in Fig. 1.1a:

RL

For circuit a. v oc = v s (t)

Given v s (t) = 10 mV and Rs = 1 k.

For circuit b. v oc = is (t) × Rs When output terminals are short-circuited, as in Fig. 1.1b: For circuit a. isc =

vo

v s (t) Rs

If RL = 100 k v o = 10 mV ×

100 = 9.9 mV 100 + 1

If RL = 10 k

For circuit b. isc = is (t)

v o = 10 mV ×

For equivalency

If RL = 1 k

Rs is (t) = v s (t)

v o = 10 mV ×

Rs

a

10 9.1 mV 10 + 1 1 = 5 mV 1+1

If RL = 100 v o = 10 mV ×

vs (t)

100 0.91 mV 100 + 1 K

For v o = 0.8v s , RL = 0.8 RL + Rs

b

Since Rs = 1 k,

Figure 1.1a

RL = 4 k

a

Ex: 1.4 Using current divider:

is (t)

io

Rs is 10 A

Rs

RL

b Figure 1.1b io = i s × Ex: 1.2

Rs Rs + RL

Given is = 10 µA, Rs = 100 k. For

Rs voc vs

isc

RL = 1 k, io = 10 µA ×

100 = 9.9 µA 100 + 1

For RL = 10 k, io = 10 µA × For

v oc = 10 mV

RL = 100 k, io = 10 µA ×

isc = 10 µA Rs =

Sedra, Smith, Carusone, Gaudet

100 K 100 K + 1 M

0.9 µA

Ex: 1.3 Using voltage divider: RL Rs + RL

100 = 5 µA 100 + 100

For RL = 1 M, io = 10 µA ×

v oc 10 mV = 1 k = isc 10 µA

v o (t) = v s (t) ×

100 9.1 µA 100 + 10

For io = 0.8is ,

100 = 0.8 100 + RL

⇒ RL = 25 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 1–2

Ex: 1.5 f =

1 1 = −3 = 1000 Hz T 10

ω = 2π f = 2π × 103 rad/s 1 1 = s = 16.7 ms f 60

Ex: 1.6 (a) T =

1 1 (b) T = = −3 = 1000 s f 10

T

v A = 0.25 V ⇒ D = 0000 v A = 3.75 V ⇒ D = 0000 (b) (i) 1 level spacing: 20 × +0.25 = +0.25 V

Ex: 1.7 If 6 MHz is allocated for each channel, then 470 MHz to 608 MHz will accommodate 806 − 470 = 23 channels 6 Since the broadcast band starts with channel 14, it will go from channel 14 to channel 36. 1 T

v A = 0 V ⇒ D = 0000 v A = 1 V ⇒ D = 0000

1 1 (c) T = = 6 s = 1 µs f 10

Ex: 1.8 P =

Ex: 1.9 (a) D can represent 15 equally-spaced values between 0 and 3.75 V. Thus, the values are spaced 0.25 V apart.

(ii) 2 level spacings: 21 × +0.25 = +0.5 V (iii) 4 level spacings: 22 × +0.25 = +1.0 V (iv) 8 level spacings: 23 × +0.25 = +2.0 V (c) The closest discrete value represented by D is +1.25 V; thus D = 0101. The error is -0.05 V, or −0.05/1.3 × 100 = −4%. Ex: 1.10 Voltage gain = 20 log 100 = 40 dB

v2 dt R

Current gain = 20 log 1000 = 60 dB

0

Power gain = 10 log Ap = 10 log (Av Ai )

V2 V2 1 × ×T = T R R Alternatively,

=

= 10 log 105 = 50 dB

P = P1 + P3 + P5 + · · · 2 4V 2 1 1 4V = √ + √ R R 2π 3 2π 2 1 4V + ··· + √ R 5 2π 1 V2 8 1 1 = × 2 × 1+ + + + ··· R π 9 25 49 It can be shown by direct calculation that the infinite series in the parentheses has a sum that approaches π 2 /8; thus P becomes V 2/R as found from direct calculation. Fraction of energy in fundamental = 8/π 2 = 0.81 Fraction of energy in first five harmonics 1 1 8 = 0.93 = 2 1+ + π 9 25 Fraction of energy in first seven harmonics 8 1 1 1 = 2 1+ + + = 0.95 π 9 25 49 Fraction of energy in first nine harmonics 8 1 1 1 1 = 2 1+ + + + = 0.96 π 9 25 49 81 Note that 90% of the energy of the square wave is in the first three harmonics, that is, in the fundamental and the third harmonic.

Sedra, Smith, Carusone, Gaudet

Ex: 1.11 Pdc = 15 × 8 = 120 mW √ (6/ 2)2 = 18 mW PL = 1 Pdissipated = 120 − 18 = 102 mW 18 PL × 100 = 15% × 100 = Pdc 120 10 Ex: 1.12 v o = 1 × 6 10−5 V = 10 µV 10 + 10 η=

(10 × 10−6 )2 = 10−11 W 10 With the buffer amplifier:

PL = v 2o /RL =

vo = 1 × =1× PL =

Ri RL × Av o × Ri + Rs RL + Ro

10 1 ×1× = 0.25 V 1+1 10 + 10

v 2o 0.252 = 6.25 mW = RL 10

Voltage gain =

vo 0.25 V = 0.25 V/V = vs 1V

= −12 dB Power gain (Ap ) ≡

PL Pi

where PL = 6.25 mW and Pi = v i i1 , v i = 0.5 V and ii =

1V = 0.5 µA 1 M + 1 M

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 1–3

Thus,

Ex: 1.16 Refer the solution to Example 1.3 in the text.

Pi = 0.5 × 0.5 = 0.25 µW

v i1 = 0.909 V/V vs

and Ap =

6.25 × 10−3 = 25 × 103 0.25 × 10−6

v i1 = 0.909 v s = 0.909 × 1 = 0.909 mV

10 log Ap = 44 dB

v i2 v i1 v i2 = × = 9.9 × 0.909 = 9 V/V vs v i1 vs

Ex: 1.13 Open-circuit (no load) output voltage = Av o v i

v i2 = 9 × v S = 9 × 1 = 9 mV v i3 v i2 v i1 v i3 = × × = 90.9 × 9.9 × 0.909 vs v i2 v i1 vs

Output voltage with load connected = Av o v i 0.8 =

RL RL + Ro

= 818 V/V

1 ⇒ Ro = 0.25 k = 250 Ro + 1

v i3 = 818 v s = 818 × 1 = 818 mV vL v i3 v i2 v i1 vL = × × × vs v i3 v i2 v i1 vs

Ex: 1.14 Av o = 40 dB = 100 V/V 2 RL v2 RL P L = o = Av o v i RL RL + Ro 2 1 2 1000 = 2.5 v 2i = v i × 100 × 1+1 Pi =

= 0.909 × 90.9 × 9.9 × 0.909 744 V/V v L = 744 × 1 mV = 744 mV Ex: 1.17 Using voltage amplifier model, the three-stage amplifier can be represented as

v 2i v 2i = Ri 10,000

Ro

PL 2.5v 2 = −4 i 2 = 2.5 × 104 W/W Ap ≡ Pi 10 v i

10 log Ap = 44 dB

vi

Ri

Ex: 1.15 Without stage 3 (see figure) vL = v s 1 M 100 k (10) 100 k + 1 M 100 k + 1 k 100 ×(100) 100 + 1 k vL = (0.909)(10)(0.9901)(100)(0.0909) vs

Avovi

Ri = 1 M Ro = 10 Av o = Av 1 ×Av 2 ×Av 3 = 9.9×90.9×1 = 900 V/V The overall voltage gain Ri RL vo = × Av o × vs Ri + Rs RL + Ro

= 81.8 V/V This figure belongs to Exercise 1.15. Stage 1

Stage 2

100 k

vs

1 k

1 k

vi1

Sedra, Smith, Carusone, Gaudet

1 M

10vi1

100vi2

vi2

100 k

Microelectronic Circuits, 8th International Edition

vL

100

© Oxford University Press 2021

Exercise 1–4

For RL = 10 :

Ex: 1.20 Using the transresistance circuit model, the circuit will be

Overall voltage gain =

10 1M × 900 × = 409 V/V 1 M + 100 K 10 + 10

ii

Ro

For RL = 1000 :

Overall voltage gain

is

Rs

Ri

1000 1M × 900 × = 810 V/V = 1 M + 100 K 1000 + 10

Rmii

RL

vo

∴ Range of voltage gain is from 409 V/V to 810 V/V. ii Rs = is Ri + Rs

Ex: 1.18

ii

v o = R m ii ×

io

RL RL + Ro

vo RL = Rm ii RL + Ro

is

Rs

Ri

Ro

RL

Ais ii

ii = i s

Rs Rs + Ri

io = Ais ii

Now

RL vo vo ii Rs = × = Rm × is ii is RL + Ro Ri + Rs

= Rm

Rs RL × Rs + Ri RL + Ro

Ex: 1.21

Ro Rs Ro = Ais is Ro + RL Rs + Ri Ro + RL

Thus, io Rs Ro = Ais is Rs + Ri Ro + RL Ex: 1.19

Ri vs

vi = vs

vi

Ri

Ri Ri + Rs

v o = Gm v i (Ro RL ) Ri = Gm v s (Ro RL ) Ri + Rs Thus, vo Ri = Gm (Ro RL ) vs Ri + Rs

Sedra, Smith, Carusone, Gaudet

Ro Gmvi

RL

vo

v b = ib rπ + (β + 1)ib Re = ib [rπ + (β + 1)Re ] But v b = v x and ib = ix , thus vx vb Rin ≡ = = rπ + (β + 1)Re ix ib Ex: 1.22 f 10 Hz 10 kHz 100 kHz 1 MHz

Gain 60 dB 40 dB 20 dB 0 dB

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 1–5

V2 = Vs

Gain (dB) 60 40

which is an HP STC function. 1 f3dB = ≤ 100 Hz 2πC(Rs + Ri )

20 dB/decade

20 1

10

10

0 10

10

10

10

10

s Ri Ri = 1 1 Rs + Ri Rs + +Ri s+ sC C(Rs + Ri )

f (Hz)

C≥

3 dB frequency

1 = 0.16 µF 2π(1 + 9)103 × 100

Ex: 1.25 T = 50 K ni = BT 3/2 e−Eg/(2kT)

Ex: 1.23

−5 ×50)

= 7.3 × 1015 (50)3/2 e−1.12/(2×8.62×10

Vi

Ri GmVi

Ro RL

Vo

9.6 × 10−39 /cm3

CL

T = 350 K ni = BT 3/2 e−Eg/(2kT) −5 ×350)

= 7.3 × 1015 (350)3/2 e−1.12/(2×8.62×10 = 4.15 × 1011 /cm3

Vo = Gm Vi [Ro RL CL ] =

Gm Vi 1 1 + + sCL Ro RL

Ex: 1.26 ND = 1017 /cm3 From Exercise 1.26, ni at

Vo Gm 1 Thus, = × sCL 1 1 Vi 1+ + 1 1 Ro RL + Ro RL Vo Gm (RL Ro ) = Vi 1 + sCL (RL Ro )

=

ni2 pn ∼ = ND (4.15 × 1011 )2 1017

= 1.72 × 106 /cm3

1 CL (RL Ro )

1 4.5 × 10−9 (103 Ro )

Ex: 1.27 At 300 K, ni = 1.5 × 1010 /cm3 pp = NA

For ω0 to be at least wπ × 40 × 103 , the highest value allowed for Ro is Ro =

nn = ND = 1017 /cm3

=

which is of the STC LP type. ω0 =

T = 350 K = 4.15 × 1011 /cm3

3

10 2π × 40 × 103 × 103 × 4.5 × 10−9 − 1

103 = 7.64 k = 1.131 − 1

Want electron concentration = np =

∴ NA = pp = =

The dc gain is

1.5 × 1010 = 1.5 × 104 /cm3 106 ni2 np

(1.5 × 1010 )2 1.5 × 104

= 1.5 × 1016 /cm3

Gm (RL Ro ) To ensure a dc gain of at least 40 dB (i.e., 100), the minimum value of Gm is ⇒ RL ≥ 100/(10 7.64 × 10 ) = 113.1 mA/V 3

Ex: 1.24 Refer to Fig. E1.24

Sedra, Smith, Carusone, Gaudet

3

Ex: 1.28 (a) νn−drift = −μn E Here negative sign indicates that electrons move in a direction opposite to E. We use

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 1–6

νn-drift = 1350 ×

1 ∵ 1 µm = 10−4 cm 2 × 10−4

Ex: 1.30 Using Eq. (1.44),

= 6.75 × 106 cm/s = 6.75 × 104 m/s

Dn Dp = = VT μn μp

(b) Time taken to cross 2-µm

Dn = μn VT = 1350 × 25.9 × 10−3

length =

−6

2 × 10 30 ps 6.75 × 104

∼ = 35 cm2 /s

(c) In n-type silicon, drift current density Jn is Jn = qnμn E = 1.6 × 10−19 × 1016 × 1350 ×

1V 2 × 10−4

= 1.08 × 104 A/cm2 (d) Drift current In = AJn = 0.25 × 10−8 × 1.08 × 104 = 27 µA The resistance of the bar is R=ρ×

L A L A

= qnμn ×

2 × 10−4 0.25 × 10−8

= 1.6 × 10−19 × 1016 × 1350 × = 37.0 k

Dp = μp VT = 480 × 25.9 × 10−3 ∼ = 12.4 cm2 /s

Ex: 1.31 Equation (1.49) 2 s 1 1 V0 + W= q NA ND 2 s NA + ND V0 = q NA ND 2 s NA + ND 2 V0 W = q NA ND NA ND 1 q W2 V0 = 2 s NA + ND

Ex: 1.32 In a p+ n diode NA ND 2 s 1 1 V0 Equation (1.49) W = + q NA ND

Alternatively, we may simply use the preceding result for current and write

We can neglect the term

R = V/In = 1 V/27 µA = 37.0 k

thus

Note that 0.25 µm = 0.25 × 10 2

Ex: 1.29 Jn = qDn

−8

2

cm .

dn(x) dx

n0 = 10 /cm = 10 /(µm) 5

3

2 s · V0 qND

Equation (1.50) xn = W

From Fig. E1.29, 17

W

W

3

Dn = 35 cm2 /s = 35 × (104 )2 (µm)2 /s

1 1 as compared to , NA ND

NA NA + ND

NA NA

=W ND NA + ND

= 35 × 108 (µm)2 /s

Equation (1.51), xp = W

dn 105 − 0 = = 2 × 105 µm−4 dx 0.5 dn(x) Jn = qDn dx

since NA ND ND NA W =W NA ND

−19

= 1.6 × 10

−6

= 112 × 10

× 35 × 10 × 2 × 10 8

5

A/µm2

Aq

= 112 µA/µm2 103 µA 1 mA = 9 µm2 ⇒A= Jn 112 µA/(µm)2

Sedra, Smith, Carusone, Gaudet

NA ND W NA + ND

NA ND W NA

= AqND W

For In = 1 mA = Jn × A

Equation (1.52), QJ = Aq

Equation (1.53), QJ = A 2 s q

Microelectronic Circuits, 8th International Edition

NA ND V0 NA + ND

© Oxford University Press 2021

Exercise 1–7

NA ND A 2 s q V0 since NA ND NA = A 2 s qND V0

=

ND = 1016 /cm3 In the n-region of this pn junction nn = ND = 1016 /cm3

since

Dp Dn + Lp N D Ln N A

can be neglected as compared to

Dp Lp N D

Dn Ln N A

Dp ∴ IS ∼ = Aqn2i Lp N D Ex: 1.35 IS =

Aqn2i

Dp Dn + Lp N D Ln N A

10

+

1016 5 × 10−4 × 2

= 1.46 × 10

−14

⎞

⎟ 18 ⎟ 10 × 10−4 × 1018 ⎠

IS eV/V T = 1.45 × 10−14 e0.605/(25.9×10

−3 )

Ex: 1.36 W = =

2 × 1.04 × 10−12

1.6 × 10−19

1 1 + NA ND 1

1018

+

1

(V0 − VF ) (0.814 − 0.605)

1016

= 1.66 × 10−5 cm = 0.166 µm Ex: 1.37 W =

2 s q

Reverse current I = IS = Aqn2i

Dp Dn + Lp N D Ln N A

= 10−14 × 1.6 × 10−19 × (1.5 × 1010 )2 10 18 × + 5 × 10−4 × 1016 10 × 10−4 × 1018

Ex: 1.38 Equation (1.69), q N N 1 s A D Cj0 = A 2 NA + ND V0 1.04 × 10−12 × 1.6 × 10−19 = 10−4 2 1018 × 1016 1 0.814 1018 + 1016

Cj0 Cj = VR 1+ V0

= 1.72 pF Ex: 1.39 Cd =

= 0.2 mA

Using Eq. (1.52), NA ND W QJ = Aq NA + ND 18 10 × 1016 × 6.08 × = 10−4 × 1.6 × 10−19 1018 + 1016

3.2 × 10−12 = 2 1+ 0.814

I = IS (eV/V T − 1)

2 s q

(0.814 + 2)

Equation (1.70),

A

1016

= 3.2 pF

= 10−4 × 1.6 × 10−19 × (1.5 × 1010 )2 ⎛ ⎜ ×⎜ ⎝

1

= 7.3 × 10−15 A

Dn Dp and have approximately Lp Ln

similar values, if NA ND , then the term

1018

+

= 9.63 pC

As one can see from above equation, to increase minority-carrier concentration (pn ) by a factor of 2, one must lower ND (= nn ) by a factor of 2.

1

10−5 cm

n2 (1.5 × 1010 )2 pn = i = = 2.25 × 104 /cm3 nn 1016

Equation (1.64) IS = Aqn2i

1.6 × 10−19

= 6.08 × 10−5 cm = 0.608 µm

Ex: 1.33 In Example 1.10, NA = 1018 /cm3 and

Ex: 1.34

2 × 1.04 × 10−12

1 1 + NA ND

Sedra, Smith, Carusone, Gaudet

(V0 + VR )

dQ d = (τ T I) dV dV

d [τ T × IS (eV/V T − 1)] dV d (eV/V T − 1) = τ T IS dV 1 V/V T e = τ T IS VT τT = × IS eV/V T VT τT ∼ I = VT

=

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 1–8

Ex: 1.40 Equation (1.73), τp =

In Example 1.6, NA = 1018 /cm3 ,

Lp2

ND = 1016 /cm3

Dp

Assuming NA ND ,

(5 × 10−4 )2 = 10 = 25 ns Equation (3.57), τT Cd = I VT

Sedra, Smith, Carusone, Gaudet

τ T τ p = 25 ns 25 × 10−9 ∴ Cd = 0.1 × 10−3 25.9 × 10−3 = 96.5 pF

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–1

Solutions to End-of-Chapter Problems

(f) P = V 2 /R = 112 /(1 × 103 ) = 0.121 W

1.1 (a) V = IR = 5 mA × 1 k = 5 V

Thus, a rating of

P = I 2 R = (5 mA)2 × 1 k = 25 mW

suffice, though

(b) R = V/I = 5 V/1 mA = 5 k

1 W should theoretically 8

1 W would be prudent to allow 4 for inevitable tolerances and measurement errors.

P = VI = 5 V × 1 mA = 5 mW (c) I = P/V = 100 mW/10 V = 10 mA R = V/I = 10 V/10 mA = 1 k

1.4 See figure on next page, which shows how to realize the required resistance values.

(d) V = P/I = 1 mW/0.1 mA = 10 V R = V/I = 10 V/0.1 mA = 100 k (e) P = I 2 R ⇒ I = P/R I = 1000 mW/1 k = 31.6 mA V = IR = 31.6 mA × 1 k = 31.6 V Note: V, mA, k, and mW constitute a consistent set of units. 5V V = = 5 mA 1.2 (a) I = R 1 k V 5V (b) R = = = 5 k I 1 mA (c) V = IR = 0.1 mA × 10 k = 1 V

1.5 Shunting the 10 k by a resistor of value of R results in the combination having a resistance Req , Req =

10R R + 10

Thus, for a 1% reduction, R = 0.99 ⇒ R = 990 k R + 10 For a 5% reduction, R = 0.95 ⇒ R = 190 k R + 10 For a 10% reduction, R = 0.90 ⇒ R = 90 k R + 10

1V V = = 0.01 A = 10 mA R 100 Note: Volts, milliamps, and kilohms constitute a consistent set of units.

For a 50% reduction,

1.3 (a) P = I 2 R = (20 × 10−3 )2 × 1 × 103

(a) 1 M results in

= 0.4 W

Req =

(d) I =

Thus, R should have a

1 -W rating. 2

R = 0.50 ⇒ R = 10 k R + 10 Shunting the 10 k by 10 × 1000 10 = = 9.9 k 1000 + 10 1.01

a 1% reduction;

(b) P = I 2 R = (40 × 10−3 )2 × 1 × 103

(b) 100 k results in

= 1.6 W

Req =

Thus, the resistor should have a 2-W rating.

a 9.1% reduction;

(c) P = I 2 R = (1 × 10−3 )2 × 100 × 103

(c) 10 k results in

= 0.1 W 1 Thus, the resistor should have a -W rating. 8

Req =

10 10 × 100 = = 9.09 k 100 + 10 1.1

10 = 5 k 10 + 10

a 50% reduction.

(d) P = I 2 R = (4 × 10−3 )2 × 10 × 103 = 0.16 W Thus, the resistor should have a

1 -W rating. 4

(e) P = V 2 /R = 202 /(1 × 103 ) = 0.4 W 1 Thus, the resistor should have a -W rating. 2

Sedra, Smith, Carusone, Gaudet

1.6 Use voltage divider to find VO VO = 5

2 =2V 2+3

Equivalent output resistance RO is RO = (2 k 3 k) = 1.2 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–2

This figure belongs to Problem 1.4. All resistors are 5 k 20 kΩ

1.67 kΩ

12.5 kΩ

23.75 kΩ

The extreme values of VO for ±5% tolerance resistor are 2(1 − 0.05) VOmin = 5 2(1 − 0.05) + 3(1 + 0.05) = 1.88 V VOmax = 5

2(1 + 0.05) 2(1 + 0.05) + 3(1 − 0.05)

1.8 9 V

10 k 6 V R 20 // 10 k 6.67 k

= 2.12 V 10 k

5 V

3 V

3 k

R 10 k // 20 k 6.67 k

10 k

VO (a)

2 k

9 V

RO

The extreme values of RO for ±5% tolerance resistors are 1.2 × 1.05 = 1.26 k and 1.2 × 0.95 = 1.14 k. 1.7 VO = VDD

R2 R1 + R2

10 k 4.5 V 10 k R 10 // 10 5 k

To find RO , we short-circuit VDD and look back into node X, RO = R2 R1 =

R1 R2 R1 + R2

Sedra, Smith, Carusone, Gaudet

(b)

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–3

9 V

15 V

10 k

R

10 k

3 V R 10 // 10 // 10 3.33 k 10 k

5.00 V 4.7 k

10 k

(c)

Thus 1 1 1 + = 10 R 9.4 ⇒ R = 156.7 ≈ 157 k

9 V

10 k

10 k

Now,

6 V

RO = 10 k R 4.7 k 9.4 = 3.133 k 3 To make RO = 3.33, we add a series resistance of approximately 200 , as shown below,

= 9.4 4.7 =

10 k R 10 // 10 // 10 3.33 k (d)

15 V

Voltage generated: 157 k

+3V [two ways: (a) and (c) with (c) having lower output resistance]

10 k 200

+4.5 V (b)

+6V [two ways: (a) and (d) with (d) having a lower output resistance]

4.7 k

VO

1.9

RO 15 V

1.10 I1

10 k I

VO

R1

I2

V

R2

4.7 k

V = I(R1 R2 )

VO = 15

=I

4.7 = 4.80 V 10 + 4.7

To increase VO to 10.00 V, we shunt the 10-k resistor by a resistor R whose value is such that 10 R = 2 × 4.7.

Sedra, Smith, Carusone, Gaudet

R 1 R2 R1 + R2

I1 =

V R2 =I R1 R1 + R2

I2 =

V R1 =I R2 R1 + R2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–4

1.11 Connect a resistor R in parallel with RL . To make IL = I/4 (and thus the current through R, 3I/4), R should be such that 6I/4 = 3IR/4 ⇒ R = 2 k 3 I 4

I 4 RL 6 k IL

I

1.1R2 = R2 +

1.1R 4

11R = 2.75 R 4 1.1R 11R Rin = R 4 4 R R =R = 4 5 ⇒ R2 =

R

0.2I I

1.12 The parallel combination of the resistors is R where

}

N

The voltage across them is I

i=1

Rin

11R 4

R 4 (b) Connect a resistor in series with the load resistor R so as to raise the resistance of the load branch by 10%, thereby restoring the current division ratio to its desired value. The added series resistance must be 10% of R (i.e., 0.1R).

1 = 1/Ri R i=1

V = I × R = N

1.1R 4

R

1/Ri

Thus, the current in resistor Rk is

0.8 I

I/Rk Ik = V/Rk = N i=1 1/Ri

0.2 I 0.1R

1.13

1.1R 4

I R 0.2 I

I

0.8 I R

Rin

R1

Rin = 1.1R

Rin

To make the current through R equal to 0.2I, we shunt R by a resistance R1 having a value such that the current through it will be 0.8I; thus R 4 The input resistance of the divider, Rin , is 0.2IR = 0.8IR1 ⇒ R1 =

1 R = R 4 5 Now if R1 is 10% too high, that is, if

Rin = R R1 = R

R 4 the problem can be solved in two ways:

R1 = 1.1

1.1R 4

1.1R 5 that is, 10% higher than in case (a).

=

1.14 For RL = 10 k , when signal source generates 0−0.5 mA, a voltage of 0−2 V may appear across the source To limit v s ≤ 1 V, the net resistance has to be ≤ 2 k. To achieve this we have to shunt RL with a resistor R so that (R RL ) ≤ 2 k. R RL ≤ 2 k. RRL ≤ 2 k R + RL For RL = 10 k

(a) Connect a resistor R2 across R1 of value such that R2 R1 = R/4, thus

R ≤ 2.5 k

R2 (1.1R/4) R = R2 + (1.1R/4) 4

The resulting circuit needs only one additional resistance of 2 k in parallel with RL so that

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–5

v s ≤ 1 V. The circuit is a current divider, and the current through RL is now 0–0.1 mA.

1 1.5 V 3

1.16

is

RL

vs

00.5 mA

R

1.15 (a) Between terminals 1 and 2:

1 1 k

1.5 V

VTh 2

1 k RTh

RTh

1 k

1 k 1 0.5 k 0.75 V

2 (b) Same procedure is used for (b) to obtain

0.5 k 2

12.31 k

0.75 V 3 (c) Between terminals 1 and 3, the open-circuit voltage is 1.5 V. When we short circuit the voltage source, we see that the Thévenin resistance will be zero. The equivalent circuit is then

Sedra, Smith, Carusone, Gaudet

0.77 V

I

3 k

Now, when a resistance of 3 k is connected between node 4 and ground,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–6

I=

0.77 12.31 + 3

= 0.75 + 0.5 = 1.25 mA

= 0.05 mA

V = I 3 R3 = 1.25 × 2 = 2.5 V

1.17

To summarize:

10 V 5 V R1 10 k

R2 5 k I2

I2 = 0.5 mA

I3 = 1.25 mA

V = 2.5 V

(b) A node equation at the common node can be written in terms of V as 5−V V 10 − V + = R1 R2 R3

I1

V

I1 = 0.75 mA

Thus,

I3

5−V V 10 − V + = 10 5 2 ⇒ 0.8V = 2

R3 2 k

⇒ V = 2.5 V (a) Node equation at the common mode yields

Now, I1 , I2 , and I3 can be easily found as

I3 = I1 + I2

10 − 2.5 10 − V = 10 10 = 0.75 mA 5 − 2.5 5−V = I2 = 5 5 = 0.5 mA V 2.5 = 1.25 mA = I3 = R3 2 I1 =

Using the fact that the sum of the voltage drops across R1 and R3 equals 10 V, we write 10 = I1 R1 + I3 R3 = 10I1 + (I1 + I2 ) × 2 = 12I1 + 2I2 That is, 12I1 + 2I2 = 10

(1)

Similarly, the voltage drops across R2 and R3 add up to 5 V, thus

Method (b) is much preferred, being faster, more insightful, and less prone to errors. In general, one attempts to identify the lowest possible number of variables and write the corresponding minimum number of equations.

5 = I2 R2 + I3 R3 = 5I2 + (I1 + I2 ) × 2

1.18 Find the Thévenin equivalent of the circuit to the left of node 1.

which yields 2I1 + 7I2 = 5

(2)

Equations (1) and (2) can be solved together by multiplying Eq. (2) by 6: 12I1 + 42I2 = 30

(3)

Now, subtracting Eq. (1) from Eq. (3) yields 40I2 = 20 ⇒ I2 = 0.5 mA Substituting in Eq. (2) gives 2I1 = 5 − 7 × 0.5 mA ⇒ I1 = 0.75 mA

Between node 1 and ground,

I3 = I1 + I2

RTh = (1 k 1.2 k) = 0.545 k

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–7

VTh = 10 ×

vπ rπ = vs r π + Rs

1.2 = 5.45 V 1 + 1.2

Find the Thévenin equivalent of the circuit to the right of node 2.

(1)

At the output side, we find v o by multiplying the current gm v π by the parallel equivalent of ro and RL , v o = −gm v π (ro RL )

(2)

Finally, v o /v s can be obtained by combining Eqs. (1) and (2) as rπ vo =− gm (ro RL ) vs r π + Rs 1.21 (a) T = 10−4 ms = 10−7 s f =

1 = 107 Hz T

Between node 2 and ground,

ω = 2π f = 6.28 × 107 rad/s

RTh = 9.1 k 11 k = 4.98 k

(b) f = 1 GHz = 109 Hz

VTh = 10 ×

11 = 5.47 V 11 + 9.1

T=

The resulting simplified circuit is 0.545 k 1

I5

ω = 2π f = 6.28 × 109 rad/s

2 4.98 k

R5 2 k 5.45 V

I5 =

V5

1 = 10−9 s f

5.47 V

5.47 − 5.45 4.98 + 2 + 0.545

(c) ω = 6.28 × 102 rad/s ω = 102 Hz f = 2π 1 T = = 10−2 s f (d) T = 10 s 1 = 10−1 Hz T

= 2.66 µA

f =

V5 = 2.66 µA × 2 k

ω = 2π f = 6.28 × 10−1 rad/s

= 5.32 mV

(e) f = 60 Hz

1.19 We first find the Thévenin equivalent of the source to the right of v O .

T=

V =4×1=4V Then, we may redraw the circuit in Fig. P1.19 as shown below 3k 1k vo

5 V

4V

ω = 2π f = 3.77 × 102 rad/s (f) ω = 1 krad/s = 103 rad/s ω = 1.59 × 102 Hz f = 2π 1 T = = 6.28 × 10−3 s f (g) f = 1900 MHz = 1.9 × 109 Hz T=

Then, the voltage at v O is found from a simple voltage division. v O = 4 + (5 − 4) ×

1 = 4.25 V 3+1

1.20 Refer to Fig. P1.20. Using the voltage divider rule at the input side, we obtain

Sedra, Smith, Carusone, Gaudet

1 = 1.67 × 10−2 s f

1 = 5.26 × 10−10 s f

ω = 2π f = 1.194 × 1010 rad/s

1.22 (a) Z = R + = 103 +

1 jωC

1 j2π × 10 × 103 × 10 × 10−9

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–8

= (1 − j1.59) k (b) Y = =

1 + jωC R

1 + j2π × 10 × 103 × 0.01 × 10−6 104

=

104 1 = Y 1 + j6.28

f = 100 kHz,

Z = j0.628

f = 1 GHz,

Z = j6.28 k

=

104 (1 − j6.28) 1 + 6.282

1 (c) Y = + jωC R 1 + j2π × 10 × 103 × 100 × 10−12 = 100 × 103 −5

= 10 (1 + j0.628)

1 + jωC jωL

1 − ω2 LC jωL

⇒Z=

= (247.3 − j1553)

Z=

Z = j0.377 m

1.24 Y =

= 10−4 (1 + j6.28) Z=

f = 60 Hz,

jωL 1 = Y 1 − ω2 LC

The frequency at which |Z| = ∞ is found letting the denominator equal zero: 1 − ω2 LC = 0 1 ⇒ω= √ LC At frequencies just below this, ∠Z = +90◦ .

105 1 + j0.628

At frequencies just above this, ∠Z = −90◦ .

= (71.72 − j45.04) k

Since the impedance is infinite at this frequency, the current drawn from an ideal voltage source is zero.

(d) Z = R + jωL = 100 + j2π × 10 × 103 × 10 × 10−3 = 100 + j6.28 × 100

1.25

= (100 + j628),

Rs vs

Rs

is

1.23 (a) Z = 1 k at all frequencies (b) Z = 1 /jωC = − j

Thévenin equivalent

1 2π f × 10 × 10−9

At f = 60 Hz,

Z = − j265 k

At f = 100 kHz,

Z = − j159

At f = 1 GHz,

Z = − j0.016

v oc = v s isc = is v s = is R s

1 (c) Z = 1 /jωC = − j 2π f × 10 × 10−12

Thus,

At f = 60 Hz,

Z = − j0.265 G

Rs =

At f = 100 kHz,

Z = − j0.16 M

(a) v s = v oc = 1 V

At f = 1 GHz,

Z = − j15.9

(d) Z = jωL = j2π f L = j2π f × 10 × 10 At f = 60 Hz,

Z = j3.77

At f = 100 kHz,

Z = j6.28 k

At f = 1 GHz,

Z = j62.8 M

(e) Z = jωL = j2π f L = j2π f (1 × 10−6 )

Sedra, Smith, Carusone, Gaudet

Norton equivalent

v oc isc

is = isc = 0.1 mA −3

Rs =

v oc 1V = 10 k = isc 0.1 mA

(b) v s = v oc = 0.1 V is = isc = 1 µA Rs =

v oc 0.1 V = = 0.1 M = 100 k isc 1 µA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–9

After a 10% increase in RL , the new values will be

1.26

Rs

VL =

1.1 × 200 = 0.978 V 5 + 1.1 × 200

IL =

1 = 4.44µA 5 + 1.1 × 200

vs

RL vo

These values represent a 0.2% and 9% change, respectively. Since the load voltage remains relatively more constant than the load current, a Thévenin source is more appropriate here.

vo RL = vs RL + Rs Rs 1+ vo = vs RL Thus, vs 1+

Rs 100

and vs 1+

Rs 10

(b) The nominal values are

= 40

(1)

VL =

50 × 1 = 0.909 V 5 + 50

IL =

1 = 18.18 mA 5 + 50

After a 10% increase in RL , the new values will be = 10

(2)

Dividing Eq. (1) by Eq. (2) gives 1 + (Rs /10) =4 1 + (Rs /100)

VL =

1.1 × 50 = 0.917 V 5 + 1.1 × 50

IL =

1 = 16.67 mA 5 + 1.1 × 50

These values represent a 1% and 8% change, respectively. Since the load voltage remains relatively more constant than the load current, a Thévenin source is more appropriate here.

⇒ RS = 50 k Substituting in Eq. (2) gives

(c) The nominal values are

v s = 60 mV The Norton current is can be found as is =

vs 60 mV = 1.2 µA = Rs 50 k

VL =

0.1 × 1 = 47.6 mV 2 + 0.1

IL =

1 = 0.476 mA 2 + 0.1

1.27 The nominal values of VL and IL are given by

After a 10% increase in RL , the new values will be

RL VS VL = RS + RL

VL =

1.1 × 0.1 = 52.1 mV 2 + 1.1 × 0.1

IL =

1 = 0.474 mA 2 + 1.1 × 0.1

IL =

VS RS + RL

After a 10% increase in RL , the new values will be VL =

1.1RL VS RS + 1.1RL

IL =

VS RS + 1.1RL

(a) The nominal values are VL =

200 × 1 = 0.976 V 5 + 200

1 IL = = 4.88µA 5 + 200

Sedra, Smith, Carusone, Gaudet

These values represent a 9% and 0.4% change, respectively. Since the load current remains relatively more constant than the load voltage, a Norton source is more appropriate here. The Norton equivalent current source is IS =

VS 1 = = 0.5 mA RS 2

(d) The nominal values are VL =

16 × 1 = 96.4 mV 150 + 16

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–10

IL =

1 = 6.02 mA 150 + 16

1.30

Rs

After a 10% increase in RL , the new values will be VL =

1.1 × 16 = 105 mV 150 + 1.1 × 16

IL =

1 = 5.97 mA 150 + 1.1 × 16

io vo

vs

io Rs vo

is

These values represent a 9% and 1% change, respectively. Since the load current remains relatively more constant than the load voltage, a Norton source is more appropriate here. The Norton equivalent current source is VS 1 = 6.67 mA = IS = RS 150

v o = v s − io R s

vo Open-circuit (io 0) vs voltage Slope Rs

1.28

PL = v 2O × =

v 2S

= v 2S

1 RL

vs is Rs

0

R2L 1 × (RL + RS )2 RL RL (RL + RS )2

Short-circuit (vo 0) current

Rs vs

RL vO

1.31

Since we are told that the power delivered to a 16 speaker load is 75% of the power delivered to a 32 speaker load, PL (RL = 16) = 0.75 × PL (RL = 32) 16 32 = 0.75 × (RS + 32)2 (RS + 32)2 √ √ 16 24 = RS + 32 RS + 32 √ √ √ √ ⇒ ( 24 = 16)RS = 16 × 32 − 24 × 16 0.9RS = 49.6

⫹ v ⫺ s

1.29 The observed output voltage is 1 mV/ ◦ C, which is one half the voltage specified by the sensor, presumably under open-circuit conditions: that is, without a load connected. It follows that that sensor internal resistance must be equal to RL , that is, 5 k.

io

⫹

⫹

RL is

vo

Rs vo

⫺

RL

⫺

RL represents the input resistance of the processor For v o = 0.95v s 0.95 =

RL ⇒ RL = 19Rs RL + Rs

For io = 0.95is 0.95 =

Rs ⇒ RL = RS /19 Rs + RL

1.32 Case

RS = 55.2

io

Rs

Sedra, Smith, Carusone, Gaudet

io

ω (rad/s)

f (Hz)

T(s)

5 × 10

0.2 × 10−9

2 × 109

3.18 × 108

3.14 × 10−9

c

6.28 × 1010

1 × 1010

1 × 10−10

d

3.77 × 102

60

1.67 × 10−2

e

6.28 × 104

1 × 104

1 × 10−4

f

6.28 × 105

1 × 105

1 × 10−5

a

3.14 × 10

b

10

Microelectronic Circuits, 8th International Edition

9

© Oxford University Press 2021

Chapter 1–11

1.33 (a) v = 10 sin(2π × 103 t), V √ (b) v = 120 2 sin(2π × 60), V (c) v = 0.1 sin(2000t), V (d) v = 0.1 sin(2π × 103 t), V 1.34 Comparing the given waveform to that described by Eq. (1.2), we observe that the given waveform has an amplitude of 0.5 V (1 V peak-to-peak) and its level is shifted up by 0.5 V (the first term in the equation). Thus the waveform looks as follows:

v

1V

T

which is 2% lower than the rms value of the square wave. 1.38 If the amplitude of the square wave is Vsq , then the power delivered by the square wave to a 2 /R . If this power is to be resistance R will be Vsq equal to that delivered by a sine wave of peak ˆ then amplitude V,

...

0

Vsq t

1 = 1 kHz I

1.35 (a) Vpeak = 117 × 2 = 165 V √ (b) Vrms = 33.9/ 2 = 24 V √ (c) Vpeak = 220 × 2 = 311 V √ (d) Vpeak = 220 × 2 = 311 kV 1.36 The two harmonics have the ratio 126/98 = 9/7. Thus, these are the 7th and 9th harmonics. From Eq. (1.2), we note that the amplitudes of these two harmonics will have the ratio 7 to 9, which is confirmed by the measurement reported. Thus the fundamental will have a frequency of 98/7, or 14 kHz, and peak amplitude of 63 × 7 = 441 mV. √ The rms value of the fundamental will be 441/ 2 = 312 mV. To find the peak-to-peak amplitude of the square wave, we note that 4V/π = 441 mV. Thus, Peak-to-peak amplitude π = 693 mV 2

1 1 Period T = = = 71.4 µs f 14 × 103

Sedra, Smith, Carusone, Gaudet

√ ˆ 2)2 (V/ = R R √ ˆ 2 . This result is independent of Thus, Vsq = V/ frequency. 1.39

√

= 2V = 441 ×

T

2 Vsq

2π 1 = = 10−3 s f0 ω0

Frequency f =

0 ⫺Vsq

Average value = 0.5 V Peak-to-peak value = 1 V Lowest value = 0 V Highest value = 1 V Period T =

1.37 The rms value of a symmetrical square ˆ Taking wave with peak amplitude Vˆ is simply V. the root-mean-square of the first 5 sinusoidal terms in Eq. (1.2) gives an rms value of, 2 2 2 2 1 1 1 1 4Vˆ + + + √ 12 + 3 5 7 9 π 2 ˆ = 0.980V

Decimal Binary 0

0

6

110

11

1011

28

11100

59

111011

1.40 (a) For N bits there will be 2N possible levels, from 0 to VFS . Thus there will be (2N − 1) discrete steps from 0 to VFS with the step size given by Step size =

VFS 2N − 1

This is the analog change corresponding to a change in the LSB. It is the value of the resolution of the ADC. (b) The maximum error in conversion occurs when the analog signal value is at the middle of a step. Thus the maximum error is

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–12

Note that there are two possible representations of zero: 0000 and 1000. For a 0.5-V step size, analog signals in the range ±3.5 V can be represented.

Step

Input

1 VFS 1 × step size = 2 2 2N − 1 This is known as the quantization error. (c)

5V ≤ 2 mV 2N − 1

2 − 1 ≥ 2500 2N ≥ 2501 ⇒ N = 12, For N = 12, 5 = 1.2 mV 212 − 1

Quantization error =

+2.5 V

+5

0101

−3.0 V

−6

1110

+2.7

+5

0101

−2.8

−6

1110

1.42 (a) When bi = 1, the ith switch is in position 1 and a current (Vref /2i R) flows to the output. Thus iO will be the sum of all the currents corresponding to “1” bits, that is, Vref b1 b2 bN iO = + 2 + ··· + N R 21 2 2

N

Resolution =

Steps Code

(b) bN is the LSB b1 is the MSB

1.2 = 0.6 mV 2

(c) iOmax =

10 V 10 k

1.41

1 1 1 1 + 2 + 3 + 4 21 2 2 2 +

b3

b2

b1

b0

Value Represented

0

0

0

0

+0

0

0

0

1

+1

0

0

1

0

+2

0

0

1

1

+3

0

1

0

0

+4

0

1

0

1

+5

0

1

1

0

+6

0

1

1

1

+7

1

0

0

0

–0

1

0

0

1

–1

1

0

1

0

–2

1

0

1

1

–3

1

1

0

0

–4

1

1

0

1

–5

1

1

1

0

–6

1

1

1

1

–7

Sedra, Smith, Carusone, Gaudet

1 1 1 1 + 6 + 7 + 8 25 2 2 2

= 0.99609375 mA Corresponding to the LSB changing from 0 to 1 the output changes by (10/10) × 1/28 = 3.91 µA. 1.43 There will be 44,100 samples per second with each sample represented by 16 bits. Thus the throughput or speed will be 44, 100 × 16 = 7.056 × 105 bits per second. 1.44 Each pixel requires 8 + 8 + 8 = 24 bits to represent it. We will approximate a megapixel as 106 pixels, and a Gbit as 109 bits. Thus, each image requires 24 × 10 × 106 = 2.4 × 108 bits. The number of such images that fit in 16 Gbits of memory is

2.4 × 108 = 66.7 = 66 16 × 109

1.45 (a) Av =

vO 10 V = 100 V/V = vI 100 mV

or 20 log 100 = 40 dB Ai =

iO v O /RL 10 V/100 0.1 A = = = iI iI 100 µA 100 µA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–13

= 1000 A/A

9.9 Vrms . The input needed is 9.9 V/200 = 49.5 mVrms .

or 20 log 1000 = 60 dB Ap =

vO

v O iO vO iO = × = 100 × 1000 v I iI vI iI

VDD 1.0 200 V/V

= 105 W/W or 10 log 105 = 50 dB (b) Av =

vI

vO 1V = 1 × 105 V/V = vI 10 µV

VDD 1.0

or 20 log 1 × 10 = 100 dB 5

Ai = =

iO v O /RL 1 V/10 k = = iI iI 100 nA

1.47

0.1 × 10−3 0.1 mA = = 1000 A/A 100 nA 100 × 10−9

+3 V 1 mA 2.2 V

or 20 log Ai = 60 dB Ap =

t

20 mA (average)

ii

v O iO vO iO = × v I iI vI iI

vo vi

= 1 × 105 × 1000

0.2 V

20 mA (average)

RL 100

= 1 × 108 W/W 3 V

or 10 log AP = 80 dB (c) Av =

vO 5V = = 5 V/V vi 1V

or 20 log 5 = 14 dB iO v O /RL 5 V/10 Ai = = = iI iI 1 mA =

v O iO vO iO = × v I iI vI iI

io 2.2 V/100 = ii 1 mA

22 mA = 22 A/A 1 mA

or 20 log Ai = 26.8 dB √ po (2.2/ 2)2 /100 = Ap = pi 10−3 0.2 √ × √ 2 2

= 5 × 500 = 2500 W/W or 10 log Ap = 34 dB

= 242 W/W

1.46 For ±5 V supplies:

or 10 log AP = 23.8 dB

The largest undistorted sine-wave output is of √ 4-V peak amplitude or 4/ 2 = 2.8 Vrms . Input needed is 14 mVrms . For ±10-V supplies, the largest undistorted sine-wave output is of 9-V peak amplitude or 6.4 Vrms . Input needed is 32 mVrms . For ±15-V supplies, the largest undistorted sine-wave output is of 14-V peak amplitude or

Sedra, Smith, Carusone, Gaudet

or 20 log 11 = 20.8 dB

=

or 20 log 500 = 54 dB

vo 2.2 = vi 0.2

= 11 V/V

Ai =

0.5 A = 500 A/A 1 mA

Ap =

Av =

Supply power = 2 × 3 V ×20 mA = 120 mW √ (2.2/ 2)2 v2 = 24.2 mW Output power = orms = RL 100 Input power =

24.2 = 0.1 mW (negligible) 242

Amplifier dissipation Supply power − Output power

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–14

= 120 − 24.2 = 95.8 mW Amplifier efficiency = =

=

Output power × 100 supply power

= 5.56 V/V Much of the amplifier’s 100 V/V gain is lost in the source resistance and amplifier’s output resistance. If the source were connected directly to the load, the gain would be

24.2 × 100 = 20.2% 120

RL RL + Ro RL Ri = Av o v s Ri + Rs RL + Ro

1.48 v o = Av o v i

0.1 vo = 0.0196 V/V = vs 5 + 0.1 This is a factor of 284× smaller than the gain with the amplifier in place!

Rs

Ro vi

vs

Ri

1 100 × 100 × 5+1 200 + 100

(b) vi

Av ovi RL

The equivalent current amplifier has a dependent current source with a value of

100 V/V 100 V/V × ii = × 1000 × v i 200 200

Thus,

= 500 × ii

vo Ri RL = Av o vs Ri + Rs RL + Ro

Thus,

(a) Av o = 100, Ri = 10Rs , RL = 10Ro :

=

vo 10Rs 10Ro = 100 × × vs 10Rs + Rs 10Ro + Ro

io ii io = × is is ii

5 200 × 500 × 5+1 200 + 100

= 277.8 A/A

= 82.6 V/V or 20 log 82.6 = 38.3 dB

Using the voltage amplifier model, the current gain can be found as follows,

(b) Av o = 100, Ri = Rs , RL = Ro : 1 1 vo = 100 × × = 25 V/V or 20 log 25 = 28 dB vs 2 2

ii vi io io = × × is is ii vi

(c) Av o = 100 V/V, Ri = Rs /10, RL = Ro /10:

=

vo Ro /10 Rs /10 = 100 vs (Rs /10) + Rs (Ro /10) + Ro

= 277.8 A/A

= 0.826 V/V or 20 log 0.826 = −1.7 dB

5 100 V/V × 1000 × 5+1 200 + 100

1.50 In Example 1.3, when the first and the second stages are interchanged, the circuit looks like the figure above, and

1.49 (a) vi vo vo = × vs vs vi

100 k v i1 = 0.5 V/V = vs 100 k + 100 k

This figure belongs to Problem 1.49. 5 k

io

200 k vi

vs

1 k

100 vi

vo

100

Figure 1

ii

is

5 k

io

500 ii

1 k

200

100

Figure 2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–15

This figure belongs to Problem 1.50.

Av 1 =

v i2 1 M = 100 × v i1 1 M + 1 k

2=

= 99.9 V/V Av 2 =

Ais ii = 1 ×

v i3 10 k = 10 × v i2 10 k + 1 k

vL 100 = 0.909 V/V =1× v i3 100 + 10 vL Total gain = Av = = Av 1 × Av 2 × Av 3 v i1 Av 3 =

= 99.9 × 9.09 × 0.909 = 825.5 V/V The voltage gain from source to load is

io Rm = ii Ro + RL 5000 10 + 1000 = 4.95 A/A = 13.9 dB =

vo ii io vo = × × vs vs ii io

= 825.5 × 0.5

=

= 412.7 V/V The overall voltage has reduced appreciably. This is because the input resistance of the first stage, Rin , is comparable to the source resistance Rs . In Example 1.3 the input resistance of the first stage is much larger than the source resistance. 1.51 The equivalent circuit at the output side of a current amplifier loaded with a resistance RL is shown. Since Ro Ro + RL

1 io × × RL Rs + Ri ii

1 × 4.95 × 1000 1000 + 100 = 4.90 V/V = 13.8 dB =

The power gain is v o io = 4.95 × 4.90 v s ii = 24.3 W/W = 27.7 dB 1.53 Gm = 60 mA/V

we can write Ro 1 = (Ais ii ) Ro + 1

(1)

Ro = 20 k RL = 1 k

and Ro Ro + 12

(2)

io

Ro

vi = vs = vs

Dividing Eq. (1) by Eq. (2), we have

Aisii

The current gain is

The voltage gain is

vL vL v i1 v i1 = × = Av · vs v i1 vS vS

0.5 = (Ais ii )

10 + 1 = 1.1 mA 10

1.52

= 9.09 V/V

io = (Ais ii )

Ro + 12 ⇒ Ro = 10 k Ro + 1

Ri Rs + Ri

2 2 = vs 1+2 3

v o = Gm v i (RL Ro )

RL

= 60

20 × 1 vi 20 + 1

= 60

20 2 × vs 21 3

Overall voltage gain ≡

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

vo = 38.1 V/V vs

© Oxford University Press 2021

Chapter 1–16

This figure belongs to Problem 1.52. ii Rs

vi

1 k

vs

Ri 100

Ro 10

io vo

Rm ii

Ro

Rs 10 k

vo

ii vs 5-mV peak

Ri

vi

Av ovi RL

1 k

1.54

100 vi

1 M

100vi 500

vo

20 log Av o = 40 dB ⇒ Av o = 100 V/V

500 500 + 100

or 20 log 83.3 = 38.4 dB v 2o /500 = A2v × 104 = 1.39 × 107 W/W v 2i /1 M

or 10 log (1.39 × 107 ) = 71.4 dB. For a peak output sine-wave current of 20 mA, the peak output voltage will be 20 mA × 500 = 10 V. Correspondingly v i will be a sine wave with a peak value of 10 V/A √ v = 10/83.3, or an rms value of 10/(83.3 × 2) = 0.085 √ V. Corresponding output power = (10/ 2)2 /500 = 0.1 W 1.55 20

200 k 1V

1 M

vi

vo = 1 V × ×1×

1 M 1 M + 200 k 100 100 + 20

Sedra, Smith, Carusone, Gaudet

Current gain =

v o /100 = 0.69 × 1.2 × 104 v s /1.2 M

= 8280 A/A

or

Power gain =

v 2o /100 v 2s /1.2 M

78.4 dB = 5713 W/W

or 10 log 5713 = 37.6 dB (This takes into account the power dissipated in the internal resistance of the source.)

vo v ib v ia vo = × × = vs v ib v ia vs 100 10 10 × × 100 × × 100 +1000 10 + 10 100 100 + 100

= 83.3 V/V

Ap =

1 100 × = 0.69 V 1.2 120 vo Voltage gain = = 0.69 V/V or −3.2 dB vs =

1.56 (a) Case S-A-B-L (see figure on next page):

vo Av = vi = 100 ×

RL 1 k

100 1vi

vo

vo = 22.7 V/V and gain in dB 20 log 22.7 = vs 27.1 dB (b) Case S-B-A-L (see figure on next page): vo v o v ia v ib = · · vs v ia v ib v s 100 × = 100 × 100 + 10 K 100 K 10 × × 100 K + 1 K 10 K 10 K + 100 K vo = 0.89 V/V and gain in dB is 20 log 0.89 = vs −1 dB. Obviously, case a is preferred because it provides higher voltage gain. 1.57 Each of stages #1, 2, ..., (n − 1) can be represented by the equivalent circuit: v i1 v i2 v i3 v in vo vo = × × × ··· × × vs vs v i1 v i2 v i(n−1) v in where

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–17

This figure belongs to 1.56, part (a).

This figure belongs to 1.56, part (b).

v i1 10 k = = 0.5 V/V vs 10 k + 10 k 200 vo = 10 × = 1.67 V/V v in 1 k + 200 10 k v i2 v i3 v in = = ··· = = 10× v i1 v i2 v i(n−1) 10 k + 10 k

1.58 Deliver 0.5 W to a 100- load. Source is 30 mV rms with 0.5-M source resistance. Choose from these three amplifier types:

A

= 9.09 V/V

Thus, vo = 0.5 × (9.09)n−1 × 1.67 = 0.833 × (9.09)n−1 vs vo For v s = 5 mV and v o = 3 V, the gain must vs be ≥ 600, thus 0.833 × (9.09)

n−1

B

C

Ri 1 M

Ri 10 k

Ri 10 k

Av 10 V/ V

Av 100 V/ V

Av 1 V/ V

Ro 10 k

Ro 1 k

Ro 20

Choose order to eliminate loading on input and output: A, first, to minimize loading on 0.5-M source B, second, to boost gain C, third, to minimize loading at 100- output.

≥ 600

⇒ n=4 Thus four amplifier stages are needed, resulting in vo = 0.833 × (9.09)3 = 625.7 V/V vs

We first attempt a cascade of the three stages in the order A, B, C (see figure above), and obtain

and correspondingly

1 1 M v i1 = = vs 1 M + 0.5 M 1.5

v o = 625.7 × 5 mV = 3.13 V This figure belongs to 1.57. Rs 10 k vs

5 mV

#1

#2

vi1

vi2

vin

#n

vo RL 200

Ri1 10 k

Ron 1 k

1 k

#m

vim

vi(m 1)

vim

Ri(m 1)

Sedra, Smith, Carusone, Gaudet

10 k

10vim

10 k vi(m 1)

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–18

⇒ v i1 = 30 ×

1 = 20 mV 1.5

10 k v i2 =5 = 10 × v i1 10 k + 10 k ⇒ v i2 = 20 × 5 = 100 mV 10 k v i3 = 100 × = 90.9 v i2 10 k + 1 k

= 833 mV

⇒ v i3 = 100 mV × 90.9 = 9.09 V

(b)

100 vo = 0.833 =1× v i3 100 + 20

vo 833 mV = 83.3 V/V = vs 10 mV

(c)

⇒ v o = 9.09 × 0.833 = 7.6 V

vo 100 = 1000 × = 500 V/V vi 100 + 100

(d)

v2 7.62 = 0.57 W Po = orms = RL 100

Rs

which exceeds the required 0.5 W. Also, the signal throughout the amplifier chain never drops below 20 mV (which is greater than the required minimum of 10 mV).

Rp

Ri

...

Connect a resistance RP in parallel with the input and select its value from

1.59

(Rp Ri ) 1 Ri = (Rp Ri ) + Rs 2 Ri + Rs ⇒1+ (a) Current gain = = Ais

io ii

⇒

Ro Ro + RL

= 100

(b) Voltage gain =

io R L vo = vs ii (Rs + Ri )

io R L = ii R s + R i

v o = 10v 1 + 20v 2 we use two transconductance amplifiers and sum their output currents. Each transconductance amplifier has the following equivalent circuit:

1 10 + 0.1

vi

= 9 V/V =19.1 dB

(c) Power gain = Ap =

Ri 10 k

Gmvi

Ro 10 k

Gm 20 mA/V

v o io v s ii

Consider first the path for the signal requiring higher gain, namely v 2 . See figure at top of next page.

= 9 × 90.9 = 818 W/W = 29.1 dB

The parallel connection of the two amplifiers at the output and the connection of RL means that the total resistance at the output is

1.60 (a) v o = 10 mV ×

1 = 16.7 k 0.11 − 0.05

1.61 To obtain the weighted sum of v 1 and v 2

= 90.9 A/A = 39.2 dB

= 90.9 ×

1 11 1 + = Rp Ri 100

Rp =

10 11

100 Rs RS = = 12 ⇒ Rp Ri = Rp Ri 11 11

100 20 × 1000 × 20 + 100 100 + 100

Sedra, Smith, Carusone, Gaudet

10 k 10 k 10 k =

Microelectronic Circuits, 8th International Edition

10 k. 3

© Oxford University Press 2021

Chapter 1–19

This figure belongs to Problem 1.61.

Rs1 10 k

Thus the component of v o due to v 2 will be v o2 = v 2

10 10 × Gm2 × 10 + 10 3

v1

10 = 33.3v 2 = v 2 × 0.5 × 20 × 3 To reduce the gain seen by v 2 from 33.3 to 20, we connect a resistance Rp in parallel with RL , 10 Rp = 2 k ⇒ Rp = 5 k 3 We next consider the path for v 1 . Since v 1 must see a gain factor of only 10, which is half that seen by v 2 , we have to reduce the fraction of v 1 that appears at the input of its transconductance amplifier to half that that appears at the input of the v 2 transconductance amplifier. We just saw that 0.5 v 2 appears at the input of the v 2 transconductance amplifier. Thus, for the v 1 transconductance amplifier, we want 0.25v 1 to appear at the input. This can be achieved by shunting the input of the v 1 transconductance amplifier by a resistance Rp1 as in the figure in the next column.

Sedra, Smith, Carusone, Gaudet

Rp1

Ri1

vi1

10 k

The value of Rp1 can be found from (Rp1 Ri1 ) = 0.25 (Rp1 Ri1 ) + Rs1 Thus, 1+

Rs1 =4 (Rp1 Ri1 )

⇒ Rp1 Ri1 = Rp1 10 =

10 Rs1 = 3 3

10 3

⇒ Rp1 = 5 k The final circuit will be as follows:

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–20

When driven by a source with source resistance Rin as shown in the figure above,

1.62

gm

gmv1

v1

R 5 k

Thus, vo vo = 0.5 vs vi

vo R

Rin Rin × vs = × v s = 0.5 × v s Rs + Rin Rin + Rin

vi =

100 mA/V

io

1.64 Voltage amplifier:

v2

gmv2

Rs 1 to 10 k

Ro

vs

Ri

vi

io = gm v 1 − gm v 2

Avo vi

vo

io RL 1 k to 10 k

v o = io RL = gm R(v 1 − v 2 ) v1 = v2 = 1 V v 1 = 1.01 V v 2 = 0.99 V

∴ vo = 0 V For Rs varying in the range 1 k to 10 k and v o limited to 10%, select Ri to be sufficiently large:

∴ v o = 100 × 5 × 0.02 = 10 V

Ri ≥ 10 Rsmax

1.63 (a)

Ri = 10 × 10 k = 100 k = 1 × 105

i2

ix

vx

For RL varying in the range 1 k to 10 k, the load voltage variation limited to 10%, select Ro sufficiently low:

i1

vi

Ri

gmvi

⎫ i = v x /Ri + gm v x ix = i 1 + i 2 ⎪ x ⎪ ⎪ ⎪ 1 ⎪ ⎪ ix = v x + gm i1 = v i /Ri ⎬ Ri v 1 ⎪ x i2 = gm v i ⎪ ⎪ ⎪ = ⎪ ⎪ ix 1/Ri + gm ⎭ vi = vx Ri = = Rin 1 + gm Ri

Ro ≤

RLmin 10

Ro =

1 k = 100 = 1 × 102 10

Now find Av o : v omin = 10 mV ×

RLmin Ri × Av o Ri + Rsmax Ro + RLmin

1 = 10 × 10−3 ×

100 k 100 k + 10 k

× Av o ×

(b)

1 k 100 + 1 k

⇒ Av o = 121 V/V

Rs Rin

vs

Rin

vi

vs 2

Sedra, Smith, Carusone, Gaudet

Ri

gmvi

vo

Ro vi

Ri

Microelectronic Circuits, 8th International Edition

Avovi

vo

© Oxford University Press 2021

Chapter 1–21

Values for the voltage amplifier equivalent circuit are Ri = 1 × 105 , Av o = 121 V/V, and Ro = 1 × 102 1.65 Transresistance amplifier: To limit v o to 10% corresponding to Rs varying in the range 1 k to 10 k, we select Ri sufficiently low; Ri ≤

v c = −βib RL

Rsmin 10

The voltage v b can be related to ib by writing for the input loop:

Thus, Ri = 100 = 1 × 10 2

v b = ib rπ + (β + 1)ib RE

To limit v o to 10% while RL varies over the range 1 k to 10 k, we select Ro sufficiently low; Ro ≤

(1)

Thus, v b = [rπ + (β + 1)RE ]ib

RLmin 10

(2)

Dividing Eq. (1) by Eq. (2) yields

Thus, Ro = 100 = 1 × 10 2

vc βRL =− vb rπ + (β + 1)RE

Now, for is = 10 µA,

Q.E.D

The voltage v e is related to ib by R

v e = (β + 1)ib RE

i i 10 µA

R 1 to 10 k

R

R i

R 1 to 10 k

v

That is, v e = [(β + 1)RE ]ib

(3)

Dividing Eq. (3) by Eq. (2) yields

v omin = 10

−5

Rsmin RLmin Rm Rsmin + Ri RLmin + Ro

1000 1000 Rm 1 = 10−5 1000 + 100 1000 + 100 ⇒ Rm = 1.21 × 105

ve (β + 1)RE = vb (β + 1)RE + rπ Dividing the numerator and denominator by (β + 1) gives RE ve = vb RE + [rπ /(β + 1)]

Q.E.D

= 121 k 1.67 Ro = = 2 k v o = 10 ×

10 V Open-circuit output voltage = Short-circuit output current 5 mA 2 =5V 2+2

1.66 The node equation at E yields the current through RE as (βib + ib ) = (β + 1)ib . The voltage v c can be found in terms of ib as

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–22

Av =

vo 10(2/4) = vi 1 × 10−6 × (200 5) × 103

the input side does not know what happens at the output side). If we neglect this internal feedback, that is, assume g12 = 0, we can compare the two models and thus obtain:

1025 V/V or 60.2 dB Ai = =

io v o /RL = ii v i /Ri

Ri = 1/g11 Av o = g21

v o Ri 5 k = 1025 × v i RL 2 k

Ro = g22

= 2562.5 A/A or 62.8 dB 1.69

The overall current gain can be found as io 5 V/2 k v o /RL = = is 1 µA 1 µA

Rs

2.5 mA = 2500 A/A = 1 µA

Vs

Ci

Ri

Vi

or 68 dB. Ap =

v 2o /RL 52 /(2 × 103 ) = 2 2 ii R i 200 10−6 × 5 × 103 200 + 5

= 2.63 × 106 W/W or 64.2 dB

1.68

g22

I1

g12 I2 g21V1

V1 1/g11

I2 V2

Ro

I1 V1

1 sCi 1 Ri + Vi sCi ⎛ = 1 Vs ⎜ Ri sCi Rs + ⎜ ⎝ 1 Ri + sCi Ri

I2

=

Ri 1 + sCi Ri ⎞ = Ri Rs + ⎟ 1 + sCi Ri ⎟ ⎠

Ri Rs + sCi Ri Rs + Ri

Vi Ri = = Vs (Rs + Ri ) + sCi Ri Rs

Ri

AvoV1

V2

which is a low-pass STC function with

K=

The correspondences between the current and voltage variables are indicated by comparing the two equivalent-circuit models above. At the outset we observe that at the input side of the g-parameter model, we have the controlled current source g12 I2 . This has no correspondence in the equivalent-circuit model of Fig. 1.16(a). It represents internal feedback, internal to the amplifier circuit. In developing the model of Fig. 1.16(a), we assumed that the amplifier is unilateral (i.e., has no internal feedback, or that

Sedra, Smith, Carusone, Gaudet

Ri (Rs + Ri ) C i Ri Rs 1+s Rs + Ri

Ri and ω0 = 1/[Ci (Ri Rs )]. Rs + Ri

For Rs = 10 k, Ri = 40 k, and Ci = 5 pF, ω0 = f0 =

1 = 25 Mrad/s 5 × 10−12 × (40 10) × 103 25 = 4 MHz 2π

The dc gain is K=

40 = 0.8V/V 10 + 40

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–23

f (Hz) |T|(dB) ∠T(◦ )

1.70 Using the voltage-divider rule.

R1

Vi

0

40

0

100

40

0

1000

40

0

104

37

−45◦

105

20

−90◦

106

0

−90◦

C R2

Vo

Vo = Vi

T(s) =

T(s) =

R2 R2 + R1 +

R2 R1 + R2

1 sC

⎛ ⎜ ⎜ ⎝

⎞ s s+

1 C(R1 + R2 )

⎟ ⎟ ⎠

which from Table 1.2 is of the high-pass type with K=

R2 R1 + R2

ω0 =

1 C(R1 + R2 )

As a further verification that this is a high-pass network and T(s) is a high-pass transfer function, see that as s ⇒ 0, T(s) ⇒ 0; and as s → ∞, T(s) = R2 /(R1 + R2 ). Also, from the circuit, observe as s → ∞, (1/sC) → 0 and Vo /Vi = R2 /(R1 + R2 ). Now, for R1 = 10 k, R2 = 40 k and C = 1 µF,

1.72 Rs = 100 k, since the 3-dB frequency is reduced by a very high factor (from 5 MHz to 100 kHz) C2 must be much larger than C1 . Thus, neglecting C1 we find C2 from 100 kHz

1 2π C2 Rs

Thevenin equivalent at node A

Rs Node A

Shunt capacitor

1 ω0 = f0 = −6 2π 2π × 1 × 10 (10 + 40) × 103 = 3.18 Hz 1 40 K |T(jω0 )| = √ = √ = 0.57 V/V 10 + 40 2 2 1.71 The given measured data indicate that this amplifier has a low-pass STC frequency response with a low-frequency gain of 40 dB, and a 3-dB frequency of 104 Hz. From our knowledge of the Bode plots for low-pass STC networks [Fig. 1.23(a)], we can complete the table entries and sketch the amplifier frequency response.

=

C1

Initial capacitor

1 2π C2 × 105

⇒ C2 = 15.9 pF If the original 3-dB frequency (5 MHz) is attributable to C1 , then 5 MHz = ⇒ C1 =

1 2π C1 Rs

1 2π × 5 × 106 × 105

= 0.32 pF

T , dB 40 37 30

20 dB/decade

3 dB

1.73 For the input circuit, the corner frequency f01 is found from 1 2π C1 (Rs + Ri )

20

f01 =

10

For f01 ≤ 100 Hz,

0 10

C2

102

103

104

Sedra, Smith, Carusone, Gaudet

105

106 f (Hz)

1 ≤ 100 2π C1 (10 + 100) × 103

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–24

R 1 k

R 10 k C V

R 100 k

R 100V 1 k

V

which is of the form shown for the low-pass function in Table 1.2 with K = 1 and ω0 = 1/RC.

C

⎛

⎞

V

⎜ For (b) Vo = Vi ⎝

⎟ 1 ⎠ R+ sC

1 ⇒ C1 ≥ = 1.4 × 10−8 F 2π × 110 × 103 × 102 −7

Thus we select C1 = 1 × 10 F = 0.1 µF. The actual corner frequency resulting from C1 will be f01 =

R

Vo sRC = Vi 1 + sCR Vo = Vi

s s+

1 RC

which is of the form shown in Table 1.2 for the high-pass function, with K = 1 and ω0 = 1/RC.

1 = 14.5 Hz −7 2π × 10 × 110 × 103

For the output circuit, 1.75 Using the voltage divider rule, 1 f02 = 2π C2 (Ro + RL ) For f02 ≤ 100 Hz,

Vs

1 ≤ 100 2π C2 (1 + 1) × 103 ⇒ C2 ≥

1 = 0.8 × 10−6 2π × 2 × 103 × 102

Vl = Vs

Select C2 = 1 × 10−6 = 1 µF.

RL

=

1 f02 = = 80 Hz 2π × 10−6 × 2 × 103 s 1+ 2π f01

s 1+

s 2π f02

RL + Rs +

RL RL + Rs

1.74 Circuits of Fig. 1.22:

Vi

Vo Vi

C R

(b)

1/sC 1/sC + R

s+

1 C(RL + Rs )

RL RL + Rs

ω0 =

1 ≤ 100 2π C(RL + Rs )

Vo

⇒C≥

(a)

For (a) Vo = Vi

s

1 C(RL + Rs )

For f0 ≤ 100 Hz

C

1 sC

which is of the high-pass STC type (see Table 1.2) with K=

R

Vl

RL

This will place the corner frequency at

T(s) = 100

C

Rs

Vo 1 = Vi 1 + sCR

Sedra, Smith, Carusone, Gaudet

1 2π × 100(20 + 5) × 103

Thus, the smallest value of C that will do the job is C = 0.064 µF or 64 nF.

1.76 From our knowledge of the Bode plots of STC low-pass and high-pass networks, we see that this amplifier has a midband gain of 40 dB, a low-frequency response of the high-pass STC type with f3dB = 102 Hz, and a high-frequency

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–25

response of the low-pass STC type with f3dB = 106 Hz. We thus can sketch the amplifier frequency response and complete the table entries as follows. T , dB

40

20 dB/decade

20 dB/decade

1.78 Since when C is connected to node A the 3-dB frequency is reduced by a large factor, the value of C must be much larger than whatever parasitic capacitance originally existed at node A (i.e., between A and ground). Furthermore, it must be that C is now the dominant determinant of the amplifier 3-dB frequency (i.e., it is dominating over whatever may be happening at node B or anywhere else in the amplifier). Thus, we can write

30 3 dB

20

3 dB

200 kHz =

40 dB

10 0

1

10

10

10

10

10

10

3-dB Bandwidth

f

10

10 f (Hz)

1 2π C(Ro1 Ri2 )

⇒ (Ro1 Ri2 ) = = 0.8 k

f

f (Hz) 1 10 102 103 104 105 106 107 108 |T|(dB) 0 20 37 40 40 40 37 20

1 2π × 200 × 103 × 1 × 10−9

0

A

#1

T(s) =

1+

#3

Ro1

vo1

C

1

B

1 nF

C 1.77 Since the overall transfer function is that of three identical STC LP circuits in cascade (but with no loading effects, since the buffer amplifiers have infinite input and zero output resistances) the overall gain will drop by 3 dB below the value at dc at the frequency for which the gain of each STC circuit is 1 dB down. This frequency is found as follows: The transfer function of each STC circuit is

#2

Ri2

s ω0

where ω0 = 1/CR

Now Ri2 = 100 k.

Thus,

Thus Ro1 0.8 k 1

|T(jω)| = 1+ 20 log

2

1 = −1 ω1 dB 2 1+ ω0

⇒1+

ω ω0

ω1 dB ω0

2 = 100,1

ω1dB = 0.51ω0 ω1dB = 0.51/CR

Sedra, Smith, Carusone, Gaudet

Similarly, for node B, 20 kHz =

1 2π C(Ro2 Ri3 )

⇒ Ro2 Ri3 =

1 2π × 20 × 103 × 1 × 10−9

= 7.96 k Ro2 = 8.65 k The designer should connect a capacitor of value Cp to node B where Cp can be found from

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–26

10 kHz = ⇒ Cp =

1 2π Cp (Ro2 Ri3 )

C2

R2

1 2π × 10 × 103 × 7.96 × 103

GmR2Vi

Note that if she chooses to use node A, she would need to connect a capacitor 10 times larger!

1.79 The LP factor 1/(1 + jf /105 ) results in a Bode plot like that in Fig. 1.23(a) with the 3-dB frequency f0 = 105 Hz. The high-pass factor 1/(1 + 102 /jf ) results in a Bode plot like that in Fig. 1.24(a) with the 3-dB frequency

To (s) =

R3 Vo = −Gm R2 Vi R2 + R3 + 1/sC2

= −Gm (R2 R3 )

s

The Bode plot for the overall transfer function can be obtained by summing the dB values of the two individual plots and then shifting the resulting plot vertically by 60 dB (corresponding to the factor 1000 in the numerator). The result is as follows:

1 C2 (R2 + R3 )

s+

which is an HP, with 3-dB frequency =

f0 = 102 Hz.

=

1 2π C2 (R2 + R3 )

1 = 14.5 Hz 2π 100 × 10−9 × 110 × 103

∴ T(s) = Ti (s)To (s) =

1 1+

Av (dB)

× − 909.1 ×

s 2π × 159 × 103

60

s s + (2π × 14.5)

20 dB/decade

20 dB/decade

40

R3

= 2 nF

50

Vo

59.2 dB 20 dB/decade

20 dB/decade

14.5 Hz

30

159 kHz

Bandwidth = 159 kHz – 14.5 Hz 159 kHz

20 10

1.81 Vi = Vs

0

1

f = 10 Av ~ – 40

10 102 103 104 105 106 107 108

60

60

60

57

40

20

0 (dB)

57

2

1/sC1 Vi (s) 1 = = 1.80 Ti (s) = Vs (s) 1/sC1 + R1 sC1 R1 + 1 LP with a 3-dB frequency f0i =

1 1 = = 159 kHz 2π C1 R1 2π 10−11 105

For To (s), the following equivalent circuit can be used:

Sedra, Smith, Carusone, Gaudet

(a) To satisfy constraint (1), namely, x Vs Vi ≥ 1 − 100 we substitute in Eq. (1) to obtain

Bandwidth = 10 − 10 = 99,900 Hz 5

(1)

f (Hz)

102 103 104 105 106 107 108 (Hz) 60

Ri Rs + Ri

x Ri ≥1− Rs + Ri 100

(2)

Thus 1 Rs + R i ≤ x Ri 1− 100 x 1 Rs 100 ≤ x −1= x Ri 1− 1− 100 100 which can be expressed as

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–27 x 1− Ri 100 ≥ x Rs 100

Substituting Rs = 10 k and x = 10% in (3) results in 100 − 1 = 90 k Ri ≥ 10 100

resulting in

Substituting f3dB = 2 MHz, CL = 20 pF, and RL = 10 k in Eq. (4) results in

Rs

V s

Vi Ri

Ro Gm Vi

RL

Vo CL

Ro ≤

1 2π × 2 × 10 × 20 × 10−12 − 6

1 104

= 6.61 k Substituting A0 = 100, x = 10%, RL = 10 k, and Ro = 6.61 k, Eq. (5) results in

Ri ≥ Rs

100 −1 x

(3)

(b) The 3-dB frequency is determined by the parallel RC circuit at the output f0 =

Gm ≥

100 = 27.9 mA/V 10 (10 6.61) × 103 1− 100

1.82 Using the voltage divider rule, we obtain Vo Z2 = Vi Z1 + Z 2

1 1 1 ω0 = 2π 2π CL (RL Ro )

where Thus, 1 f0 = 2π CL

1 1 + RL R0

Z1 = R 1

To obtain a value for f0 greater than a specified value f3dB we select Ro so that 1 1 1 ≥ f3dB + 2π CL RL Ro 1 1 + ≥ 2π CL f3dB RL Ro

1 2π f3dB CL −

(4)

1 RL

(c) To satisfy constraint (c), we first determine the dc gain as dc gain =

Ri Gm (Ro RL ) Rs + Ri

1 + sC1 Vo R1 = 1 1 Vi + + s(C1 + C2 ) R1 R2 C1 = C1 + C2

1 C 1 R1 1 1 1 s+ + (C1 + C2 ) R1 R2 s+

This in turn will happen if

Ri Gm (Ro RL ) ≥ A0 Rs + Ri

1 1 = C 1 R1 C1 + C2

The first factor on the left-hand side is (from constraint (2)) greater or equal to (1 − x/100). Thus

Sedra, Smith, Carusone, Gaudet

Y1 Vo = Vi Y1 + Y 2

This transfer function will be independent of frequency (s) if the second factor reduces to unity.

For the dc gain to be greater than a specified value A0 ,

A0 Gm ≥ x (Ro RL ) 1− 100

It is obviously more convenient to work in terms of admittances. Therefore we express Vo /Vi in the alternate form

and substitute Y1 = (1/R1 ) + sC1 and Y2 = (1/R2 ) + sC2 to obtain

1 1 ≥ 2π CL f3dB − Ro RL Ro ≤

1 1 and Z2 = R2 sC1 sC2

1 1 + R1 R2

which can be simplified as follows: (5)

C1 + C2 = R1 C1

1 1 + R1 R2

Microelectronic Circuits, 8th International Edition

(1)

© Oxford University Press 2021

Chapter 1–28

1+

R1 C2 =1+ C1 R2

The gain drops by 3 dB at the corner frequencies of the two STC networks, that is, at f = 10 Hz and f = 10 kHz.

or C 1 R1 = C 2 R2

1.84 Use the expression in Eq. (1.26), with

When this condition applies, the attenuator is said to be compensated, and its transfer function is given by C1 Vo = Vi C1 + C2

k = 8.62 × 10−5 eV/K; and Eg = 1.12 V we have

which, using Eq. (1), can be expressed in the alternate form Vo = Vi

B = 7.3 × 1015 cm−3 K−3/2 ;

1

R2 = R1 R1 + R2 1+ R2

T = −55◦ C = 218 K: ni = 2.68 × 106 cm−3 ;

N = 1.9 × 1016 ni

That is, one out of every 1.9 × 1016 silicon atoms is ionized at this temperature.

Thus when the attenuator is compensated (C1 R1 = C2 R2 ), its transmission can be determined either by its two resistors R1 , R2 or by its two capacitors. C1 , C2 , and the transmission is not a function of frequency.

T = 0◦ C = 273 K: ni = 1.52 × 109 cm−3 ;

N = 3.3 × 1013 ni

T = 20◦ C = 293 K: ni = 8.60 × 109 cm−3 ; 1.83 The HP STC circuit whose response determines the frequency response of the amplifier in the low-frequency range has a phase angle of 5.7◦ at f = 100 Hz. Using the equation for ∠T(jω) from Table 1.2, we obtain tan−1

f0 = 5.7◦ ⇒ f0 = 10 Hz 100

The LP STC circuit whose response determines the amplifier response at the high-frequency end has a phase angle of −5.7◦ at f = 1 kHz. Using the relationship for ∠T(jω) given in Table 1.2, we obtain for the LP STC circuit. −tan−1

N = 5.8 × 1012 ni

T = 75◦ C = 348 K: ni = 3.70 × 1011 cm−3 ;

N = 1.4 × 1011 ni

T = 125◦ C = 398 K: ni = 4.72 × 1012 cm−3 ;

N = 1.1 × 1010 ni

1.85 Use Eq. (1.26) to find ni , ni = BT 3/2 e−Eg /2kT

3

10 = −5.7◦ ⇒ f0 10 kHz f0

At f = 100 Hz, the drop in gain is due to the HP STC network, and thus its value is 20 log

1

1+

10 100

2

= −0.04 dB

Similarly, at the drop in gain f = 1 kHz is caused by the LP STC network. The drop in gain is 20 log 1+

1

1000 10, 000

2

= −0.04 dB

Sedra, Smith, Carusone, Gaudet

Substituting the values given in the problem, −5 ×300)

ni = 3.56 × 1014 (300)3/2 e−1.42/(2×8.62×10 = 2.2 × 106 carriers/cm3

1.86 The concentration of free carriers (both electrons and holes) in intrinsic silicon is found in Example 3.1 to be 1.5 × 1010 carriers/cm3 at room temperature. Multiplying this by the volume of the wafer gives 1.5 × 1010 ×

π × 152 × 0.3 = 4

7.95 × 1010 free electrons

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–29

1.87 Since NA ni , we can write

At 398 K, hole concentration is

pp ≈ NA = 5 × 1018 cm−3

pn = 2.23 × 108 /cm3

Using Eq. (1.27), we have np =

n2i = 45 cm−3 pp

1.90 (a) The resistivity of silicon is given by Eq. (1.41). For intrinsic silicon,

1.88 Hole concentration in intrinsic Si = ni ni = BT

3/2

p = n = ni = 1.5 × 1010 cm−3

−Eg /2kT

e

= 7.3 × 10 (300) 15

3/2 −1.12/(2×8.62×10−5 ×300)

e

Using μn = 1350 cm2 /V · s and μp = 480 cm2 /V · s, and q = 1.6 × 10−19 C we have

= 1.5 × 1010 holes/cm3 In phosphorus-doped Si, hole concentration drops below the intrinsic level by a factor of 108 .

ρ = 2.28 × 105 -cm. Using R = ρ ·

∴ Hole concentration in P-doped Si is

A = 3 × 10−8 cm2 , we have

1.5 × 1010 = 1.5 × 102 cm−3 pn = 108

R = 7.6 × 109 .

Now, nn ND and pn nn = n2i

(b) nn ≈ ND = 5 × 1016 cm−3 ;

2

nn = n2i /pn =

L with L = 0.001 cm and A

(1.5 × 1010 ) 1.5 × 102

pn =

n2i = 4.5 × 103 cm−3 nn

= 1.5 × 1018 cm−3

Using μn = 1200 cm2 /V · s and

ND = nn = 1.5 × 1018 atoms/cm3

μp = 400 cm2 /V · s, we have ρ = 0.10 -cm; R = 3.33 k.

1.89 T = 27◦ C = 273 + 27 = 300 K

(c) nn ≈ ND = 5 × 1018 cm−3 ;

At 300 K, ni = 1.5 × 1010 /cm3 pn =

Phosphorus-doped Si:

n2i = 45 cm−3 nn

Using μn = 1200 cm2 /V · s and μp = 400 cm2 /V · s, we have

nn ND = 1017 /cm3 2

pn =

n2i (1.5 × 1010 ) = = 2.25 × 103 /cm3 ND 1017

Hole concentration = pn = 2.25 × 103 /cm3

ρ = 1.0 × 10−3 -cm; R = 33.3 . As expected, since ND is increased by 100, the resistivity decreases by the same factor.

T = 125◦ C = 273 + 125 = 398 K At 398 K, ni = BT

(d) pp ≈ NA = 5 × 1016 cm−3 ; np =

3/2 −Eg /2kT

e

= 7.3 × 1015 × (398)3/2 e−1.12/(2×8.62×10

= 4.5 × 103 cm−3

= 4.72 × 1012 /cm3

ρ = 0.31 -cm; R = 10.42 k

−5 ×398)

pn =

n2i = 2.23 × 108 /cm3 ND

Sedra, Smith, Carusone, Gaudet

n2i nn

(e) Since ρ is given to be 2.8 × 10−6 -cm, we directly calculate R = 9.33 × 10−2 .

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–30

1.91 Cross-sectional area of Si bar

= 4.05 × 106 cm/s

= 5 × 4 = 20 µm2

νn 4.05 × 106 = = 2.8125 νp 1.44 × 106

−4

Since 1 µm = 10

cm, we get

or

νn = 2.8125 νp

= 20 × 10−8 cm2

Or, alternatively, it can be shown as

Current I = Aq(pμp + nμn )E

μn νn μn E 1350 = = = νp μp E μp 480

= 20 × 10−8 × 1.6 × 10−19 (1016 × 500 + 104 × 1200) ×

= 2.8125

1V 10 × 10−4

= 160 µA

1.94 Jdrift = q(nμn + pμp )E

1.92 Use Eq. (1.45):

Here n = ND , and since it is n-type silicon, one can assume p n and ignore the term pμp . Also,

Dp Dn = = VT μn μp

E=

Dn = μn VT and Dp = μp VT where VT = 25.9 mV.

1V 1V = = 103 V/cm 10 µm 10 × 10−4 cm

Need Jdrift = 2 mA/µm2 = qND μn E Doping Concentration μn μp Dn Dp (carriers/cm3 ) cm2 /V · s cm2 /V · s cm2 /s cm2 /s

2 × 10−3 A = 1.6 × 10−19 N D × 1350 × 103 10−8 cm2

Intrinsic

1350

480

35

12.4

⇒ ND = 9.26 × 1017 /cm3

1016

1200

400

31

10.4

750

260

19.4

6.7

17

10

1018

380

160

9.8

4.1

1.95 pn0 =

2

n2i (1.5 × 1010 ) = = 2.25 × 104 /cm3 ND 1016

From Fig. P1.95, 1.93 Electric field: 3V 3V = E= 10 µm 10 × 10−6 m 3V = 10 × 10−4 cm

dp 108 pn0 − pn0 108 pn0 =− − dx W 50 × 10−7 since 1 nm = 10−7 cm 108 × 2.25 × 104 dp =− dx 50 × 10−7

= 3000 V/cm

= −4.5 × 1017 Hence Jp = −qDp

10 µm

dp dx

= −1.6 × 10−19 × 12 × (−4.5 × 1017 )

3 V

= 0.864 A/cm2

νp-drift = μp E = 480 × 3000

1.96 From Table 1.3,

= 1.44 × 106 cm/s

VT at 300 K = 25.9 mV

νn-drift = μn E = 1350 × 3000

Using Eq. (1.46), built-in voltage V0 is obtained:

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–31 NA ND = 25.9 × 10−3 × V0 = VT ln n2i

10 × 10 ln 2 1.5 × 1010 17

16

1.97 Equation (1.49):

Holes

W

xp

xn

Electrons

V0 =

1 1 V0 ← Eq. (1.50) + NA ND

= 5 × 1016 cm−3 and ni = 1.5 × 1010 cm−3 , we have V0 = 778 mV. Using Eq. (1.50) and s = 11.7 × 8.854 × 10−14 F/cm, we have W = 2 × 10−5 cm = 0.2 µm. The extension of the depletion width into the n and p regions is given in Eqs. (1.51) and (1.52), respectively:

= 0.298 µm ND 1016 = 0.328 × 17 NA + ND 10 + 1016

= 0.03 µm Use Eq. (1.53) to calculate charge stored on either side: NA ND W, where junction area QJ = Aq NA + ND = 100 µm2 = 100 × 10−8 cm2 QJ = 100 × 10−8 × 1.6 × 10−19

Sedra, Smith, Carusone, Gaudet

2 × 1.04 × 10−12

1.98 Using Eq. (1.46) and NA = ND

NA 1017 = 0.328 × 17 NA + ND 10 + 1016

Hence, QJ = 4.8 × 10−14 C

2 1.6 × 10−19 × 1016 × 0.2 × 10−4

since NA ND , we have QJ = 3.2 fC.

Use Eqs. (1.51) and (1.52) to find xn and xp :

× 0.328 × 10−4

qND 2 ·W 2 s

= 0.31 V NA ND W∼ QJ = Aq = AqND W NA + ND

= 0.328 × 10−4 cm = 0.328 µm

xp = W

V0 ,

2 s 1 V0 q ND

So V0 =

W= 1 2 × 1.04 × 10−12 1 × 0.754 + 1.6 × 10−19 1017 1016

xn = W

Here W = 0.2 µm = 0.2 × 10−4 cm

Depletion width W=

1 1 + NA ND

W

2 s q

Since NA ND , we have

= 0.754 V

2 s q

W=

1017 · 1016 1017 + 1016

xn = W·

NA = 0.1 µm NA + ND

xp = W·

ND = 0.1 µm NA + ND

Since both regions are doped equally, the depletion region is symmetric. Using Eq. (1.53) and A = 20 µm2 = 20 × 10−8 cm2 , the charge magnitude on each side of the junction is QJ = 1.6 × 10−14 C.

1.99 Using Eq. (1.47) or (1.48), we have charge stored: QJ = qAxn ND . Here xn = 0.1 µm = 0.1 × 10−4 cm

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–32

A = 10 µm × 10 µm = 10 × 10−4 cm × 10 × 10−4 cm = 100 × 10−8 cm2

Charge density

A 4× increase in ND therefore results in a 2× decrease in the width of the depletion region on the n-type side of the junction, as illustrated below.

qNA

So, QJ = 1.6×10−19 ×100×10−8 ×0.1×10−4 ×1018

p-type

n-type

= 1.6 pC

1 X 2 n

NA ND 1.100 V0 = VT ln n2i

X

If NA or ND is increased by a factor of 10, then new value of V0 will be 10 N A ND V0 = VT ln n2i

4qND 1.102 The area under the triangle is equal to the built-in voltage.

The change in the value of V0 is VT ln10 = 59.6 mV.

V0 =

1 Emax W 2

Using Eq. (1.49): 1.101 This is a one-sided junction, with the depletion layer extending almost entirely into the more lightly doped (n-type) material. A thin space-charge region in the p-type region stores the same total charge with higher much higher charge density.

p-type

⇒ Emax

Emax =

n-type

X

qND

Increasing ND by 4× will increase the charge density on the n − type side of the junction by 4×. We see from Eq. (1.45) that since NA ND ni (i.e. at least a couple of orders of magnitude) a 4× increase in ND will have comparatively little effect on V0 . Thus, we can assume V0 is unchanged in Eq. (1.49), and the junction width (residing almost entirely in the n-type material) is 2 s 1 1 V0 ∝ √ q NA ND

Sedra, Smith, Carusone, Gaudet

NA ND 2kTNA ND ln s (NA + ND ) n2i

Finally, we can substitute for ni using Eq. (1.26)

Xn

W≈

2 s 1 1 × V0 + q NA ND 2qNA ND V0 = s (NA + ND )

Substituting Eq. (1.45) for V0 ,

Charge density

qNA

Emax V0 = 2

E max = 2kTNA ND Eg ln(NA ND ) − 2 ln B − 3 ln T + s (NA + ND ) kT 1.103 Using Eq. (1.46) with NA = 1017 cm−3 , ND = 1016 cm−3 , and ni = 1.5 × 1010 , we have V0 = 754 mV Using Eq. (1.55) with VR = 5 V, we have W = 0.907 µm. Using Eq. (1.56) with A = 1 × 10−6 cm2 , we have QJ = 13.2 × 10−14 C. 1.104 Equation (1.65): IS = Aqn2i

Dp Dn + Lp N D Ln N A

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–33

A = 100 µm2 = 100 × 10−8 cm2

Now I = Ip + In = 100 I n + In ≡ 1 mA

2 IS = 100 × 10−8 × 1.6 × 10−19 × 1.5 × 1010

In =

10

+

5 × 10−4 × 1016

18 10 × 10−4 × 1017

= 7.85 × 10−17 A

At 300 K,

= 7.85 × 10

−17

×e

750/25.9

ni = 7.3 × 1015 × (300)

∼ = 0.3 mA

=

2 s q

1 1 + NA ND

2 s 1 1 + q NA ND

(V0 + VR )

V0 1 +

VR V0

Qj = A 2 s q

= A 2 s q

NA N D NA + ND

NA N D NA + ND

· (V0 + VR )

V0 ·

VR 1+ V0

= QJ0 1 +

VR V0

Dp Dn + Lp N D Ln N A

Here Ip = Aqn2i

ni = 7.3 × 1015 × (305)

× e−1.12/(2×8.62×10

)

−5 ×305

n2i (at 305 K) = 2.152 n2i (at 300 K)

Thus IS approximately doubles for every 5◦ C rise in temperature.

1.108 Equation (1.63) I = Aqn2i

In = Aqn2i V/V e T −1

Dp V/VT −1 e Lp N D

Dn V/VT e −1 Ln N A

Ip Dp Ln NA = · · In Dn Lp ND

Dp Dn + Lp N D Ln N A

10 10 10 × × 16 20 5 10

Ip = 100 In

Sedra, Smith, Carusone, Gaudet

eV/VT − 1

Dp V/VT e −1 Lp N D

Dn V/VT e −1 Ln N A

For p+ −n junction NA ND , thus Ip In and I Ip = Aqn2i

Dp V/VT −1 e Lp N D

For this case using Eq. (1.65): IS Aqn2i

Dp = 104 × 10−8 × 1.6 × 10−19 Lp N D 2 × 1.5 × 1010

18

=

3/2

= 2.152 × 1010

So Ip = Aqn2i

1.106 Equation (1.62):

At 305 K,

so

Equation (1.55):

In = Aqn2i

)

−5 ×300

n2i (at 305 K) = 4.631 × 1020

VR 1+ V0

I = Aqn2i

×e−1.12/(2×8.62×10

n2i (at 300 K) = 2.232 × 1020

= W0

3/2

= 1.4939 × 1010 /cm2

1.105 Equation (1.54):

Ip = 1 − In = 0.9901 mA 1.107 ni = BT 3/2 e−Eg /2kT

I∼ = IS eV/VT

W=

1 mA = 0.0099 mA 101

10 10 × 10−4 × 1017

= 3.6 × 10−16 A I = IS eV/VT − 1 = 1.0 × 10−3

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–34 −3 3.6 × 10−16 eV/(25.9×10 ) − 1 = 1.0 × 10−3

q N N 1 s A D Cj0 = A 2 NA + ND V0

⇒ V = 0.742 V 1.109 VZ = 12 V Rated power dissipation of diode = 0.25 W. If continuous current “I” raises the power dissipation to half the rated value, then 12 V × I =

1.112 Equation (1.72):

1 × 0.25 W 2

NA ND V0 = VT ln n2i

= 25.9 × 10

−3

1017 × 1016 × ln 2 1.5 × 1010

= 0.754 V

I = 10.42 mA

Cj0 = 100 × 10−8

Since breakdown occurs for only half the time, the breakdown current I can be determined from

1.04 × 10−12 × 1.6 × 10−19 1 1017 × 1016 2 1017 + 1016 0.754

I × 12 ×

1 = 0.25 W 2

= 31.6 fF

⇒ I = 41.7 mA

Cj =

1.110 Equation (1.73), Cj =

Cj0 VR m 1+ V0

Cj0

31.6 fF

=

VR 1+ V0

1+

3 0.754

= 14.16 fF

0.4 pF For VR = 1 V, Cj = 1 1/3 1+ 0.75 = 0.3 pF

1.113 Equation (1.66):

For VR = 10 V, Cj =

0.4 pF 10 1+ 0.75

1/3

α = A 2 s q

NA ND NA + ND

= 0.16 pF

Equation (1.68):

1.111 Equation (1.81):

α Cj = √ 2 V0 + VR

Cd =

τT I VT

5 pF =

Substitute for α from Eq. (1.66):

τT 25.9 × 10−3

× 1 × 10−3

τ T = 5 × 10−12 × 25.9

NA ND √ A 2 s q s N + ND ×√ Cj = √ A s 2 V0 + VR = A s ×

= 129.5 ps For I = 0.1 mA: τT Cd = ×I VT =

129.5 × 10−12 25.9 × 10−3

= s A

1

2 s NA + ND q NA ND

1

2 s 1 1 + q NA ND × 0.1 × 10−3 = 0.5 pF

Sedra, Smith, Carusone, Gaudet

=

(V0 + VR )

(V0 + VR )

s A W

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 1–35

p n, np

n region

p region

pn(xn) pn(x)

np (xp) np (x)

pn0 np0

Wp

xp

Wn

xn

0 Depletion region

Widths of p and n regions

1.114 (a) See figure at top of page.

Qp = Aq ×

(b) The current I = Ip + In .

# 1" pn (xn ) − pn0 (Wn − xn ) 2

=

# 1 " Aq pn0 eV/V T − pn0 (Wn − xn ) 2

=

1 Aqpn0 eV/V T − 1 (Wn − xn ) 2

dp dx

=

1 ni 2 Aq (Wn − xn ) eV/V T − 1 2 ND

pn (xn ) − pn0 dp pn0 eV/V T − pn0 = = dx Wn − xn Wn − xn V/V e T −1 = pn0 Wn − xn V/V n2i e T − 1 = ND (Wn − xn )

=

1 (Wn − xn )2 · Ip 2 Dp

1 Wn2 · Ip for Wn xn 2 DP

Find current component Ip : pn (xn ) = pn0 eV/V T and pn0 = Ip = AJp = AqDp

∴ Ip = AqDp = Aqn2i

n2i ND

(c) For Q Qp , I Ip , Q

dp dx

Dp (Wn − xn )ND

Thus, τT =

× eV/V T − 1

dI dQ = τT dV dV But I = IS eV/V T − 1

Dn × eV/V T − 1 Wp − xp NA

I = Ip + In =

Aqn2i

1 Wn2 , and 2 Dp

Cd =

Similarly, In = Aqn2i

1 Wn2 I 2 Dp

Dp Dn + (Wn − xn ) ND Wp − xp NA

!

× eV/V T − 1

I so Cd ∼ . = τT · VT (d) Cd =

The excess change, Qp , can be obtained by multiplying the area of the shaded triangle of the pn (x) distribution graph by Aq.

Sedra, Smith, Carusone, Gaudet

IS eV/V T I dI = dV VT VT

1 Wn2 1 × 10−3 = 8 × 10−12 F 2 10 25.9 × 10−3

Solve for Wn : Wn = 0.64 µm

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–1

Chapter 2

v d = (Gm v 2 − Gm v 1 )R = Gm R(v 2 − v 1 )

Solutions to Exercises within the Chapter

Therefore:

Ex: 2.1 The minimum number of terminals required by a single op amp is 5: two input terminals, one output terminal, one terminal for positive power supply, and one terminal for negative power supply. The minimum number of terminals required by a quad op amp is 14: each op amp requires two input terminals and one output terminal (accounting for 12 terminals for the four op amps). In addition, the four op amps can all share one terminal for positive power supply and one terminal for negative power supply.

v 3 = μGm R(v 2 − v 1 ) That is, the open-loop gain of the op amp is A = μGm R. For Gm = 20 mA/V and μ = 50, we have: A = 50 × 20 × 5 = 5000 V/V, or equivalently, 74 dB.

Ex: 2.4 The gain and input resistance of the inverting amplifier circuit shown in Fig. 2.5 are R2 and R1 , respectively. Therefore, we have: − R1 R1 = 100 k and

Ex: 2.2 Relevant equations are: v 3 = A(v 2 − v 1 ); v Id = v 2 − v 1 , 1 v Icm = (v 1 + v 2 ) 2 (a)

4 v3 = 0 − 3 = −0.004 V = −4 mV v1 = v2 − A 10 v Id = v 2 − v 1 = 0 − (−0.004) = +0.004 V

−

R2 = −10 ⇒ R2 = 10 R1 R1

Thus: R2 = 10 × 100 k = 1 M Ex: 2.5

= 4 mV 1 v Icm = (−4 mV + 0) = −2 mV 2

R 10 k vi

(b) −10 = 10 (2 − v 1 ) ⇒ v 1 = 2.01 V 3

v Icm =

v Id = v 2 − v 1 = 2 − 2.01 = −0.01 V = −10 mV

ii

vo

1 1 (v 1 + v 2 ) = (2.01 + 2) = 2.005 V 2 2

2V (c) v 3 = A(v 2 − v 1 ) = 103 (1.998 − 2.002) = −4 V v Id = v 2 − v 1 = 1.998 − 2.002 = −4 mV v Icm =

1 1 (v 1 + v 2 ) = (2.002 + 1.998) = 2 V 2 2

(d) −1.2 = 103 [v 2 − (−1.2)] = 103 (v 2 + 1.2)

From Table 1.1 we have: v o ; that is, output is open circuit Rm = i i

io = 0

The negative input terminal of the op amp (i.e., v i ) is a virtual ground, thus v i = 0: v o = v i − Rii = 0 − Rii = −Rii v o Rii Rm = =− = −R ⇒ Rm = −R ii i o = 0 ii

v Id = v 2 − v 1 = −1.2012 − (−1.2)

= −10 k vi Ri = and v i is a virtual ground (v i = 0), ii

= −0.0012 V = −1.2 mV

thus Ri =

⇒ v 2 = −1.2012 V

1 1 (v 1 + v 2 ) = [ − 1.2 + (−1.2012)] 2 2 −1.2 V

v Icm =

Ex: 2.3 From Fig. E2.3 we have: v 3 = μv d and

Sedra, Smith, Carusone, Gaudet

0 = 0 ⇒ Ri = 0 ii

Since we are assuming that the op amp in this transresistance amplifier is ideal, the op amp has zero output resistance and therefore the output resistance of this transresistance amplifier is also zero. That is Ro = 0 .

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–2

R 10 k

R1

10 k

0.5 mA

Rf

v1

vo

R2

vi

Ex: 2.7

v2

vO

Connecting the signal source shown in Fig. E2.5 to the input of this amplifier, we have: v i is a virtual ground that is v i = 0, thus the current flowing through the 10-k resistor connected between v i and ground is zero. Therefore, v o = v i − R × 0.5 mA = 0 − 10 k × 0.5 mA = −5 V.

Ex: 2.6

i2 R1 1 k v1

2V

we want to have: Rf =1 and R1

Rf =4 R2

Therefore 4V 4V ⇒ Rf ≥ 4 k ≤ 1 mA ⇒ Rf ≥ Rf 1 mA

iO iL

Since it is required that v O = −(v 1 + 4v 2 ),

It is also desired that for a maximum output voltage of 4 V, the current in the feedback resistor not exceed 1 mA.

R2 5 k

i1

For the circuit shown above we have: Rf Rf v1 + v2 vO = − R1 R2

vO RL 1 k

Let us choose Rf to be 4 k, then Rf R1 = Rf = 4 k and R2 = = 1 k 4 Ex: 2.8

v 1 is a virtual ground, thus v 1 = 0 V i1 =

2 V − v1 2−0 = 2 mA = R1 1 k

Assuming an ideal op amp, the current flowing into the negative input terminal of the op amp is zero. Therefore, i2 = i1 ⇒ i2 = 2 mA v O = v 1 − i2 R2 = 0 − 2 mA × 5 k = −10 V iL =

vo −10 V = −10 mA = RL 1 k

iO = iL − i2 = −10 mA − 2 mA = −12 mA Voltage gain = 14 dB Current gain = 14 dB

−10 V vO = = −5 V/V or 2V 1V iL −10 mA = −5 A/A or = i1 2 mA

−10(−10 mA) PL = 25 W/W or 14 dB = Pi 2 V × 2 mA PL Note that power gain in dB is 10 log10 . Pi

=

Sedra, Smith, Carusone, Gaudet

We want to design the circuit such that v O = 2v 1 + v 2 − 4v 3 Thus we need to have

Power gain

Rc Rc Ra Ra v1 + v2 R1 Rb R2 Rb Rc v3 − R3

vO =

Ra R1

Rc Rb

= 2,

Ra R2

Rc Rb

= 1, and

Rc =4 R3

From the above three equations, we have to find six unknown resistors; therefore, we can arbitrarily choose three of these resistors. Let us choose Ra = Rb = Rc = 10 k.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–3

Then we have 10 Rc = = 2.5 k R3 = 4 4 Ra Rc 10 10 =2 = 2, ⇒ × R1 Rb R1 10

Using the superposition principle to find the contribution of v 1 to v O , we set v 2 = v 3 = 0. Then we have (refer to the solution of Exercise 2.9): v O = 6v 1 To find the contribution of v 2 to v O , we set v 1 = v 3 = 0, then: v O = 2v 2

⇒ R1 = 5 k Ra Rc 10 10 × =1⇒ =1 R2 Rb R2 10

To find the contribution of v 3 to v O we set v 1 = v 2 = 0, then 7 k v 3 = −7v 3 1 k Combining the contributions of v 1 , v 2 , and v 3 to v O we have: v O = 6v 1 + 4v 2 − 7v 3 . vo = −

⇒ R2 = 10 k

Ex: 2.9 Using the superposition principle to find the contribution of v 1 to the output voltage v O , we set v 2 = 0

Ex: 2.11

7 k

R2

i

1 k

R1

vO

vO

v1

1 k

vi v2 3 k

v + (the voltage at the positive input of the op amp 3 is: v + = v 1 = 0.75v 1 1+3 7 k v + = 8 × 0.75v 1 = 6v 1 Thus v O = 1 + 1 k To find the contribution of v 2 to the output voltage v O we set v 1 = 0. Then v + =

vO R2 R2 =1+ =2⇒ = 1 ⇒ R1 = R2 vi R1 R1 If v O = 10 V, then it is desired that i = 10 µA. Thus, i=

10 V 10 V = 10 µA ⇒ R1 + R2 = R1 + R2 10 µA

R1 + R2 = 1 M and R1 = R2 ⇒ R1 = R2 = 0.5 M

1 v 2 = 0.25v 2 1+3

Ex: 2.12

Hence 7 k v + = 8 × 0.25v 2 = 2v 2 vO = 1 + 1 k

(a)

R2

Combining the contributions of v 1 and v 2

R1

v vI

to v O , we have v O = 6v 1 + 2v 2

vO

Ex: 2.10 7 k

v I − v − = v O /A ⇒ v − = v I − v O /A

1 k

1 k v1

v3

(1)

But from the voltage divider across v O , vO

v− = vO

R1 R1 + R2

(2)

Equating Eq. (1) and Eq. (2) gives v2 3 k

Sedra, Smith, Carusone, Gaudet

vO

R1 vO = vI − R1 + R2 A

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–4

which can be manipulated to the form vO = vI

1 + (R2 /R1 ) 1 + (R2 /R1 ) 1+ A (b) For R1 = 1 k and R2 = 9 k the ideal value 9 for the closed-loop gain is 1 + , that is, 10. The 1 10 actual closed-loop gain is G = . 1 + 10/A

vO 10 V = = 10 V/V or 20 dB vi 1V 10 mA iL =∞ = iI 0 PL v O × iL 10 × 10 = = =∞ PI v I × iI 1×0 Ex: 2.14 (a) Load voltage =

If A = 103 , then G = 9.901 and G − 10 × 100 = −0.99% −1% 10 For v I = 1 V, v O = G × v I = 9.901 V and

(b) Load voltage = 1 V

=

v O = A(v + − v − ) ⇒ v + − v − =

1 k × 1 V 1 mV 1 k + 1 M

Ex: 2.15

9.901 vO = A 1000

9.9 mV If A = 104 , then G = 9.99 and = −0.1%. For v I = 1 V, v O = G × v I = 9.99 V, therefore, v+ − v− =

9.99 vO = = 0.999 mV 1 mV A 104

(a) R1 = R3 = 2 k, R2 = R4 = 200 k

If A = 105 , then G = 9.999 and = −0.01%

Since R4 /R3 = R2 /R1 we have:

For v I = 1 V, v O = G × v I = 9.999 thus,

Ad =

9.999 vO = = 0.09999 mV v+ − v− = A 105 0.1 mV

(b) Rid = 2R1 = 2 × 2 k = 4 k Since we are assuming the op amp is ideal, Ro = 0

Ex: 2.13

(c)

iI = 0 A, v 1 = v I = 1 V

Acm ≡

1V v1 = = 1 mA 1 k 1 k i2 = i1 = 1 mA

i1 =

v1 vI 1 V

iO vO

1 k

9 k

i2 i1

vO R2 200 = 100 V/V = = v I2 − v I1 R1 2

iI

iL

1 k

vO R4 = v Icm R3 + R4

R2 R3 1− R1 R4

The worst-case common-mode gain (i.e., the largest Acm ) occurs when the resistor tolerances are such that the quantity in parentheses is maximum. This in turn occurs when R2 and R3 are at their highest possible values (each one percent above nominal) and R1 and R4 are at their lowest possible values (each one percent below nominal), resulting in Acm

R4 = R3 + R4

| Acm |

1.01 × 1.01 1− 0.99 × 0.99

R4 200 ×0.04 ×0.04 0.04 V/V R3 + R4 202

The corresponding CMRR is v O = v 1 + i2 × 9 k = 1 + 1 × 9 = 10 V vO 10 V = = 10 mA 1 k 1 k iO = iL + i2 = 11 mA

iL =

Sedra, Smith, Carusone, Gaudet

CMRR =

100 | Ad | = = 2500 | Acm | 0.04

or 68 dB.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–5

Ex: 2.16 We choose R3 = R1 and R4 = R2 . Then for the circuit to behave as a difference amplifier with a gain of 10 and an input resistance of 20 k, we require R2 = 10 and R1

R

RId = 2R1 = 20 k ⇒ R1 = 10 k and

C

i

Ad =

Ex: 2.18

vi(t) i

R2 = Ad R1 = 10 × 10 k = 100 k

vO (t)

1 CR

t

冮0vI (t) dt

Therefore, R1 = R3 = 10 k and R2 = R4 = 100 k.

The signal waveforms will be as shown.

vI (t)

Ex: 2.17 Given v Icm = +5 V

2.5 V

v Id = 10 sin ωt mV 2R1 = 1 k, R2 = 0.5 M

2 s

t

0

R3 = R4 = 10 k 1 1 v I1 = v Icm − v Id = 5 − × 0.01 sin ωt 2 2 = 5 − 0.005 sin ωt V

2.5 V

1 v I2 = v Icm + v Id 2 = 5 + 0.005 sin ωt V

2.5 V

vO (t)

t

v − (op amp A1 ) = v I1 = 5 − 0.005 sin ωt V v − (op amp A2 ) = v I2 = 5 + 0.005 sin ωt V v Id = v I2 − v I1 = 0.01 sin ωt v Id v O1 = v I1 − R2 × 2R1 = 5 − 0.005 sin ωt − 500 k ×

0.01 sin ωt 1 k

= (5 − 5.005 sin ωt) V v Id v O2 = v I2 + R2 × 2R1

When v I = +2.5 V, the current through the capacitor will be in the direction indicated, i = 2.5 V/R, and the output voltage will decrease linearly from +2.5 V to −2.5 V. Thus in (T/2) seconds, the capacitor voltage changes by 5 V. The charge equilibrium equation can be expressed as i(T/2) = C × 5 V

= (5 + 5.005 sin ωt) V v + (op amp A3 ) = v O2 ×

2.5 V

10 R4 = v O2 R3 + R4 10 + 10

1 1 v O2 = (5 + 5.005 sin ωt) 2 2 = (2.5 + 2.5025 sin ωt)V

2.5T 1 2.5 T = 5C ⇒ CR = = × 2 × 10−6 R 2 10 4 = 0.5 µs

=

Ex: 2.19

v − (op amp A3 ) = v + (op amp A3 )

= 10.01 sin ωt V

Sedra, Smith, Carusone, Gaudet

= 1(1 + 1000) × 0.01 sin ωt

R Vi

= (2.5 + 2.5025 sin ωt) V R4 R2 1+ v Id vO = R3 R1 0.5 M 10 k 1+ × 0.01 sin ωt 10 k 0.5 k

C

Vo

The input resistance of this inverting integrator is R; therefore, R = 10 k.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–6

10−3 s = 0.1 µF 10 k From Eq. (2.27) the transfer function of this integrator is: C=

1 Vo ( jω) =− Vi ( jω) jωCR For ω = 10 rad/s, the integrator transfer function has magnitude Vo 1 = V 1 × 10−3 = 100 V/V i ◦

and phase φ = 90 . For ω = 1 rad/s, the integrator transfer function has magnitude Vo 1 = V 1 × 10−3 = 1000 V/V i and phase φ = 90◦ . The frequency at which the integrator gain magnitude is unity is 1 1 = −3 = 1000 rad/s ωint = CR 10

and phase φ = −90◦ . If we add a resistor in series with the capacitor to limit the high-frequency gain of the differentiator to 100, the circuit would be:

R C

R1

Vi

is 10−3 s, we have: CR = 10−3 s ⇒

For ω = 103 rad/s, the differentiator transfer function has magnitude Vo = 103 × 10−2 = 10 V/V V i

Since the desired integration time constant

Vo

At high frequencies the capacitor C acts like a short circuit. Therefore, the high-frequency gain R . To limit the magnitude of of this circuit is: R1 this high-frequency gain to 100, we should have: R 1 M R = 100 ⇒ R1 = = = 10 k R1 100 100

Ex: 2.20

Ex: 2.21 Refer to the model in Fig. 2.29 and observe that

R v + − v − = VOS + v 2 − v 1 = VOS + v Id

Vi

C

Vo

and since v O = v 3 = A(v + − v − ), then

v O = A(v Id + VOS )

C = 0.01 µF is the input capacitance of this differentiator. We want CR = 10−2 s (the time constant of the differentiator); thus, R=

10−2 = 1 M 0.01 µF

From Eq. (2.33), the transfer function of the differentiator is Vo (jω) = −jωCR Vi (jω) Thus, for ω = 10 rad/s the differentiator transfer function has magnitude Vo = 10 × 10−2 = 0.1 V/V V i and phase φ = −90◦ .

Sedra, Smith, Carusone, Gaudet

(1)

where A = 104 V/V and VOS = 5 mV. From Eq. (1) we see that v id = 0 results in v O = 50 V, which is impossible; thus the op amp saturates and v O = +10 V. This situation pertains for v Id ≥ −4 mV. If v Id decreases below −4 mV, the op-amp output decreases correspondingly. For instance, v Id = −4.5 mV results in v O = +5 V; v Id = −5 mV results in v O = 0 V; v Id = −5.5 mV results in v O = −5 V; and v Id = −6 mV results in v O = −10 V, at which point the op amp saturates at the negative level of −10 V. Further decreases in v Id have no effect on the output voltage. The result is the transfer characteristic sketched in Fig. E2.21. Observe that the linear range of the characteristic is now centered around v Id = −5 mV rather than the ideal situation of v Id = 0; this shift is obviously a result of the input offset voltage VOS .

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–7

◦

(b) If at room temperature (25 C), VOS is trimmed to zero and (i) the circuit is operated at a constant temperature, the peak of the input sine wave can be increased to 10 mV. (ii) However, if the circuit is to operate in the temperature range of 0◦ C to 75◦ C (i.e., at a temperature that deviates from room temperature by a maximum of 50◦ C), the input offset voltage will drift from by a maximum of 10 µV/◦ C × 50◦ C = 500 µV or 0.5 mV. This will reduce the allowed peak amplitude of the input sinusoid to 9.5 mV.

R3 = 9.9 k 10 k With this value of R3 , the new value of the output dc voltage [using Eq. (2.38)] is: VO = IOS R2 = 10 nA × 10 k = 0.01 V

Ex: 2.25 Using Eq. (2.39) we have: VOS 2 mV t ⇒ 12 = 2 mV + t CR 1 ms 12 V − 2 mV × 1 ms 6 s ⇒t= 2 mV

v O = VOS +

RF

C R

Ex: 2.22 (a) The inverting amplifier of −1000 V/V gain will exhibit an output dc offset voltage of ±VOS (1 + R2 /R1 ) = ±3 mV × (1 + 1000) = ±3.03 V. Now, since the op-amp saturation levels are ±10 V, the room left for output signal swing is approximately ±7 V. Thus to avoid op-amp saturation and the attendant clipping of the peak of the output sinusoid, we must limit the peak amplitude of the input sine wave to approximately 7 V/1000 = 7 mV.

Vi

Ex: 2.23 (a) If the amplifier is capacitively coupled in the manner of Fig. 2.32(a), then the input offset voltage VOS will see a unity-gain amplifier [Fig. 2.32(b)] and the dc offset voltage at the output will be equal to VOS , that is, 3 mV. Thus, almost the entire output range of ±10 V will be available for signal swing, allowing a sine-wave input of approximately 10-mV peak without the risk of output clipping. Obviously, in this case there is no need for output trimming. (b) We need to select a value of the coupling capacitor C that will place the 3-dB frequency of the resulting high-pass STC circuit at 1000 Hz, thus 1000 =

1 2π CR1

⇒C=

1 = 0.16 µF 2π × 1000 × 1 × 103

Ex: 2.24 From Eq. (2.35) we have: VO = IB1 R2 IB R2

From Eq. (2.37) the value of resistor R3 (placed in series with positive input to minimize the output offset voltage) is R1 R2 10 k × 1 M = R3 = R1 R2 = R1 + R2 10 k + 1 M

Sedra, Smith, Carusone, Gaudet

With the feedback resistor RF , to have at least ±10 V of output signal swing available, we have to make sure that the output voltage due to VOS has a magnitude of at most 2 V. From Eq. (2.34), we know that the output dc voltage due to VOS is RF RF ⇒ 2 V = 2 mV 1 + VO = VOS 1 + R 10 k RF = 1000 ⇒ RF 10 M 10 k The corner frequency of the resulting STC 1 network is ω0 = CRF

1+

We know RC = 1 ms and R = 10 k ⇒ C = 0.1 µF Thus ω0 = f0 =

= 100 nA × 1 M = 0.1 V

= 9.9 k

VOS

Vo

1 = 1 rad/s 0.1 µF × 10 M

1 ω0 = = 0.16 Hz 2π 2π

Ex: 2.26 20 log A0 = 106 dB ⇒ A0 = 200,000 V/V ft = 3 MHz fb = ft /A0 =

3 MHz = 15 Hz 200,000

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 2–8

At fb , the open-loop gain drops by 3 dB below its value at dc; thus it becomes 103 dB. For f fb , | A| ft /f ; thus 3 MHz = 104 V/V 300 Hz

At f = 300 Hz, | A| =

Since the nominal dc gain is much lower than A0 , R2 1+ f3dB ft R1 =

20 MHz = 200 kHz 100

or 80 dB 3 MHz = 103 V/V 3 kHz

or 60 dB

Ex: 2.29 For the input voltage step of magnitude V the output waveform will still be given by the exponential waveform of Eq. (2.56) if

At f = 12 kHz, which is two octaves higher than 3 kHz, the gain will be 2 × 6 = 12 dB below its value at 3 kHz; that is, 60 − 12 = 48 dB.

SR SR ⇒V≤ resulting in that is, V ≤ ωt 2πft

At f = 3 kHz, | A| =

At f = 60 kHz, | A| =

ωt V ≤ SR

3 MHz = 50 V/V 60 kHz

V ≤ 0.16 V From Appendix F we know that the 10% to 90% rise time of the output waveform of the form of 2.2 . Eq. (2.56) is tr 2.2 × time constant = ωt

or 34 dB

Ex: 2.27

Thus, tr 0.35 µs

A0 = 106 V/V or 120 dB The gain falls off at the rate of 20 dB/decade. Thus, it reaches 40 dB at a frequency four decades higher than fb , 104 fb = 100 kHz ⇒ fb = 10 Hz

If an input step of amplitude 1.6 V (10 times as large compared to the previous case) is applied, the output will be slew-rate limited and thus linearly rising with a slope equal to the slew rate, as shown in the following figure.

vo

The unity-gain frequency ft will be two decades higher than 100 kHz, that is, ft = 100 × 100 kHz = 10 MHz

Slope SR Alternatively, we could have found ft from the gain-bandwidth product

1.6 V

ft = A0 fb = 106 × 10 Hz = 10 MHz

0 tr

At a frequency f fb , | A| = ft /f For f = 10 kHz, | A| =

tr = 10 MHz = 103 V/V or 60 dB 10 kHz

t

0.9 × 1.6 − 0.1 × 1.6 1 V/µs

⇒ tr = 1.28 µs Ex: 2.30 From Eq. (2.57) we have: SR = 318 kHz 2π VO max

Ex: 2.28

fM =

20 log A0 = 106 dB ⇒ A0 = 200,000 V/V

Using Eq. (2.58), for an input sinusoid with frequency f = 5 fM , the maximum possible amplitude that can be accommodated at the output without incurring SR distortion is: fM 1 = 5 × = 1 V (peak) VO = VO max 5fM 5

ft = 20 MHz For a noninverting amplifier with a nominal dc gain of 100, R2 = 100 1+ R1

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–1

Solutions to End-of-Chapter Problems

2.4 The minimum number of pins required by a dual op amp is 8. Each op amp has 2 input terminals (4 pins) and one output terminal (2 pins). Another 2 pins are required for power supplies.

2.1 Refer to Fig. P2.1. v+ = vI ×

1 1 k = vI 1 k + 1 M 1001

v O = Av + = Av I A = 1001

Similarly, the minimum number of pins required by a quad op amp is 14:

1 1001

vO vI

4 × 2 + 4 × 1 + 2 = 14 2.5 v O = −2.000 V v − = −1.000 V

4 1 A = 4004 V/V

= 1001 ×

v

2.2

v0

v

#

v1

v2

vd = v2 − v1

vO

vO /vd

1

0.00

0.00

0.00

0.00

–

v + = v − = −1.000 V

2

1.00

1.00

0.00

0.00

–

3

(a)

1.00

(b)

1.00

Measured voltage at positive input = −1.005 V vO Amplifier gain A = v + −v −

4

1.00

1.10

0.10

10.1

101

5

2.01

2.00

–0.01

–0.99

99

6

1.99

2.00

0.01

1.00

100

7

5.10

(c)

(d)

–5.10

For ideal op amp,

=

−2.000 −1.005 − (−1.000)

= 400 V/V 2.6 i = Gm (v 2 − v 1 ) v O = Rm i

Experiments 4, 5, and 6 show that the gain is approximately 100 V/V. The missing entry for experiment #3 can be predicted as follows: vO 1.00 = = 0.01 V. (a) v d = A 100

= Rm Gm (v 2 − v 1 ) vO = Rm Gm Gain A = v2 − v1 = 1 × 106 × 40 × 10−3 = 40,000 V/V

(b) v 1 = v 2 − v d = 1.00 − 0.01 = 0.99 V

v1

The missing entries for experiment #7:

Gmv1

−5.10 = −0.051 V (c) v d = 100

i

Rmi vO

v2

(d) v 2 = v 1 + v d = 5.10 − 0.051 = 5.049 V

Gmv2

All the results seem to be reasonable. 2.3 v1 = vIcm − 0.90 0.75 –1.20 –0.15 0.10

vId 2

v2 = vIcm + 0.90 1.25 –1.00 0.05 –0.10

Sedra, Smith, Carusone, Gaudet

vId 2

vId = v2 − v1 0 0.50 0.20 0.20 –0.20

1 (v +v2 ) 2 1 0.90 1.00 –1.10 –0.05 0.00

vIcm =

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–2

1 (v 1 + v 2 ) 2 v d = 0.005 sin(2π 1000)t = v 2 − v 1 2.7 v cm = 2 sin(2π 60)t =

(b) G = −2 V/V =

−R2 ⇒ R1 = 10 k, R2 = 20 k R1

v 1 = v cm − v d /2 = sin(120π )t − 0.0025 sin(2000π t)

(c) G = −5 V/V

v 2 = v cm + v d /2 = sin(120π t) + 0.0025 sin(2000π t)

=

2.8 Refer to Fig. E2.3.

=

−R2 ⇒ R1 = 10 k, R2 = 50 k R1

(d) G = −100 V/V −R2 ⇒ R1 = 10 k, R2 = 1 M R1

v d = R(Gm2 v 2 − Gm1 v 1 ) v O = v 3 = μv d = μR (Gm2 v 2 − Gm1 v 1 )

2.12 Gain is 46 dB

vO =

R2

1 1 μR(Gm v 2 + Gm v 2 − Gm v 1 + Gm v 1 ) 2 2 1 v O = μRGm (v 2 − v 1 ) + μRGm (v 1 + v 2 ) 2 v Id 2v Icm we have v O = Ad v Id + Acm v Icm

R1

vI

⇒ Ad = μRGm , Acm = μRGm Ad = 20 log Gm CMRR = 20 log Acm Gm For a CMRR ≥ 60 dB, Gm Gm ≥ 1000 ⇒ ≤ 0.1% Gm Gm

vO

46 dB = 20 log |G| |G| = 200 ∴

R2 vO = −200 V/V = − vI R1

⇒ R2 = 200R1 ≤ 1 M

2.9 Circuit a b c d

vo /vi (V/V)

Rin (k)

−100 = −5 20 –5 –5 –5

R2 = 1 M

20

1 M = 5 k 200 Rin = R1 = 5 k

20 20 20

R1 =

Note that in circuit (b) the 20-k load resistance has no effect on the closed-loop gain because of the zero output resistance of the ideal op amp. In circuit (c), no current flows in the 20-k resistor connected between the negative input terminal and ground (because of the virtual ground at the inverting input terminal). In circuit (d), zero current flows in the 20-k resistor connected in series with the positive input terminal. 2.10 (a) G = −2 V/V (c) G = −0.5 V/V (e) G = −5 V/V

For the largest possible input resistance, choose

(b) G = −10 V/V (d) G = −50 V/V

2.13

vI

0V 4 V

Average = −2 V Highest = 0 V Lowest = −4 V The current through R1 is VI /R1 .

0 mA 2.11 (a) G = −1 V/V −R2 ⇒ R1 = R2 = 10 k = R1

Sedra, Smith, Carusone, Gaudet

2 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–3

ii

2.14 The circuit will be as follows: ii

R R

vi

5 k

15 k

Rin = If R1 and R2 have ±1% tolerance, then VO will exhibit ±2% variability and thus will be in the range of 5 × 1.02 = 5.1 V to 5 × 0.98 = 4.9 V. Variation of the −15-V supply by ±1% results in a ±1% variation in the output voltage. Thus the total variation in the output voltage can be ±3%, resulting in VO in the range 4.85 V to 5.15 V. 2.15 (a) Choose R1 = 1 k and R2 = 200 k. R2 /R1 (b) G = − 1 + R2 /R1 1+ A 200 = −192.3 V/V =− 1 + 200 1+ 5000 To restore the gain to −200 V/V, we need to change the effective value of R1 to R1 , −200 = −

200/R1 1 + 200/R1 1+ 5000

A

200 200 200 1 + = 200 + R1 5000 R1 200 = 200.04 R1

(b) For A finite: vo v i = − , v o = v i − ii R f A −v o vo Rf − ii R f ⇒ R m = ⇒ vo = =− 1 A ii 1+ A vi Rf Ri = = ii 1+A 2.17 From Example 2.2, vO R2 R4 R4 =− + 1+ vI R1 R2 R3 Here R1 = R2 = R4 = 1 M 1M 1 v ∴ O =− 1+1+ =− 2+ vI R3 R3 1 M vO +2=− vI R3

This effective value can be realized by shunting R1 (1 k) with Ra , 1 1 1 ⇒ Ra = 24 k = + 0.960 1 Ra (c) From Appendix J, we find the closest available 1% resistor as either 23.7 k or 24.3 k.

1 M vO +2 vI

vO = −100 V/V vI 1 M = 10.2 k R3 = − (−100 + 2) (a)

(b)

⇒ R1 = 0.960 k

vO = −10 V/V vI

R3 = − (c)

1 M = 125 k (−10 + 2)

vO 1 M =∞ = −2 V/V, R3 = − vI (−2 + 2)

1V = 0.1 mA 10 k I2 = I1 = 0.1 mA 2.18 (a) I1 =

2.16

−I 2 × 10 k = −I3 × 100 ≡ Vx

(a) For A = ∞: v i = 0

∴ I3 = I2 ×

v o = −ii Rf vo Rm = = −Rf ii

Sedra, Smith, Carusone, Gaudet

vo

vi =0 ii

R3 = −

Thus,

0.96 ×

V 5 V

15 V

Rf

10 k 100

= 0.1 × 100 = 10 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–4

Now IL = I2 + I3 = 10.1 mA and Vx = −I2 × 10 k = −0.1 × 10 k = −1 V (b) Vx = RL IL + VO = 10.1 RL + VO RL =

If iI is in a direction opposite to that shown in the figure, then v x = 10iI v O = v x + iL RL = 10iI + 11iI = 21iI

Vx − V O −1 − (−13) = = 1.19 k 10.1 mA 10.1

The maximum value of iI will result in v O = +12 V. Thus

(c) 100 ≤ RL ≤ 1 k

VO = Vx − RL IL = −1 − RL × 10.1

12 = 0.57 mA 21 Thus, the allowable range of iI is

RL = 100 , VO = −2.01 V; RL = 1 k,

−0.57 mA ≤ iI ≤ +0.57 mA

VO = −11.1 V

(d) Since Rin = 0, the value of the source resistance will have no effect on the resulting iL ,

12 = 21iImax ⇒ iImax =

IL stays fixed at 10.1 mA.

∴ −11.1 V ≤ VO ≤ −2.01 V

iL = 0.2 × 11 = 2.2 mA

2.19 2.20 v O = −(2v 1 + 4v 2 + 8v 3 ) iI 10 k

iL

vx i

R1 , R2 , R3 ≥ 1 k

RL

Rf Rf Rf = 2, = 4, =8 R1 R2 R3

R

R3 = 1 k ⇒ Rf = 8 k R2 = 2 k

iI

vO

R1 = 4 k 2.21 (a) v O = −(v 1 + 2v 2 + 3v 3 ) Rf = 1 ⇒ R1 = 100 k, R1

(a) v x = −iI × 10

Rf = 2 ⇒ R2 = 50 k R2

i = −v x /R = iI (10/R) 10 iL = iI + i = i I 1 + R

100 Rf k = 1 ⇒ R3 = R3 3 100 k

Thus,

v

iL 10 =1+ iI R

R 100 k

100 k v

iL = 11 ⇒ R = 1 k. For iI

100 k 100 k

(b) Rin = 0 (because of the virtual ground at the input). Ro = ∞ (because iL is independent of the value of RL ).

100 k v 100 k

(c) If iI is in the direction shown in the figure above, the maximum allowable value of iI will be determined by v O reaching −12 V, at which point

Ri1 = 100 k

v x = −iImax × 10

Ri2 = 50 k

and

Ri3 = 33.3 k

v O = −12 = v x − iLmax × 1 = −10 iImax − 11 iImax

(b) v O = −(v 1 + v 2 + 2v 3 + 2v 4 )

⇒ iImax =

12 = 0.57 mA 21

Sedra, Smith, Carusone, Gaudet

v

Rf = 1 ⇒ R1 = 100 k R1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–5

100 = 1.67 k 6 Suggested configurations:

100 k

Ri1 =

100 k

100 k

100 k

v v

100 k

v O = −(2v 1 + 2v 2 + 2v 3 )

v

v

100 k

100 k 100 k

v1

Rf = 1 ⇒ R2 = 100 k R2

v2

v

100 k

100 k 100 k

100 k

vO

100 k

100 Rf = 2 ⇒ R3 = k R3 2

v3

100 k

100 Rf = 2 ⇒ R4 = k R4 2 v O = −(3v 1 + 3v 2 )

Ri1 = 100 k, Ri2 = 100 k, Ri3 = 50 k, Ri4 = 50 k

100 k

(c) v O = −(v 1 + 5v 2 )

100 k

100 k

100 k v

100 k

vO

100 k

100 k

v 100 k

100 k

v1

100 k

v

v2

100 k 100 k

100 k 100 k

In order to have coefficient = 0.5, connect one of vO = 0.5. the input resistors to v O . v1

R1 = 100 k 100 k 5 Ri1 = 100 k

R2 =

100 k

v1

Ri2 = 20 k

100 k

100 k 100 k

(d) v O = −6v 1

100 k

100 k R1 = 6

vO

100 k 100 k

100 k 100 k 100 k

100 k 100 k 100 k

Sedra, Smith, Carusone, Gaudet

v1

100 k

2.22 Using the circuit of Fig. 2.11:

vO

To obtain v O = v 1 + 2v 2 − 3v 3 − 5v 4 , we can arbitrarily select Rc = Rb , then Ra =1 R1

and

Ra =2 R2

If we select R2 = 10 k, then Ra = 20 k and R1 = 20 k.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–6

10 k

10 k vI

10 k

vO

10 k

For the second summer, to obtain Rc =3 R3

10 k vI

10 k

Rc =5 R4

requires

vO

and vO = −5 v4

then

vO = −3 v3

We can select Rc = Rb = 60 k, resulting in R3 = 20 k and R4 = 12 k. vO = −0.5 V/V vI

2.23

Rin = 10 k (smallest gain magnitude)

R2 10 k

10 k

R1 10 k

vO

vI

10 k 10 k vI

vO vO = −0.5 V/V vI

Rin = 20 k (smallest gain magnitude)

Closed-loop gain is R2 10 k vO =− =− vI R1 10 k

10 k

= −1 V/V

vI

10 k

10 k

For v I = +1.00 V,

vO

v O = −1 × 1.00 = − 1.00 V The two resistors are 1% resistors vO = 10(1 − 0.01) = 0.98 V/V v 10(1 + 0.01) I min vO = 10(1 + 0.01) = 1.02 V/V v 10(1 − 0.01) I max Thus the measured output voltage will range from −0.98 V to −1.02 V.

vO = −2 V/V vI

Rin = 5 k (largest gain magnitude)

2.25 vO R2 = −10 V/V = − vI R1 ⇒ R2 = 10R1

2.24 There are four possibilities: vO = −2 V/V Rin = 10 k (largest gain vI magnitude)

Sedra, Smith, Carusone, Gaudet

Total resistance used is 110 k ∴ R1 + R2 = 110 k R1 + 10R1 = 110 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–7

R2

Thus, the tolerance of the closed-loop gain is ±2x%. For Gnominal = −100 and x = 1, the closed-loop gain will be in the range −98 V/V to −102 V/V.

R1

vI

vO

R2 /R1 1 + R2 /R1 1+ A For A = 500, G = −50, and R2 = 100 k,

2.29 G = −

100/R1 1 + 100/R1 1+ 500 100 100 = 50 + 0.1 1 + Thus, R1 R1

−50 = − R1 = 10 k and R2 = 10R1 = 100 k 2.26 The circled numbers indicate the order of the analysis steps. The additional current supplied by the op amp comes from the power supplies (not shown).

1 V

1 mA 2

6 6 mA 0V 1

100 = 50.1 R1

⇒ R1 = 1.796 k 1.8 k The input 2-k resistor can be shunted by a resistance Ra to obtain an equivalent R1 of 1.796 k,

3 1 mA 10 k

1 k

0.9 ×

10 V 4

5 mA

5

2 k

1 1 1 = + 1.796 Ra 2 ⇒ Ra = 17.61 k 18 k 2.30 Output voltage ranges from –10 V to +10 V and open-loop gain is 5000 V/V

2.27 The inverting amplifier has a gain of −10 V/V resulting in an output voltage of +10 V. Thus, there is 2 mA current in RL . With R2 = 10 k, as shown in Fig. P2.17, there is 1 mA current in R2 for a total of 3 mA coming from the ideal op-amp output. If the output current must be limited to 2.5 mA with the same gain, the output voltage must remain +10 V and the current in R2 must be limited to 2.5 − 2 = 0.5 mA. This requires,

∴ Voltage at the inverting input terminal will 10 −10 to (i.e., –2 mV to +2 mV). vary from 5000 5000 Thus the virtual ground will depart by a maximum of ±2 mV. 2.31

10 = 20 k 0.5 To keep the same gain, −R2 /R1 = −10. Hence, R2 =

R1 =

R2 = 2 k 10

2.28 R1 = R1nominal (1 ± x%) R2 = R2nominal (1 ± x%) |G| =

v O = −Av − = v − − ii R2

R2 R2nominal 1 ± x% = R1 R1nominal 1 ± x%

ii R2 = v − (1 + A)

Thus, |G|max

R2nominal × (1 + 2x%) R1nominal

|G|min

R2nominal × (1 − 2x%) R1nominal

Sedra, Smith, Carusone, Gaudet

v− =

ii R 2 1+A

Again v I = ii R1 + v − = ii R 1 + ii

R2 1+A

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–8

So Rin =

vI R2 = R1 + ii A+1

1 =1+

Q.E.D.

2.32 Refer to Fig. P2.32. Vo = Vi

−R2 /R1 1 + R2 /R1 1+ A

(1)

A 1+

R2 R1

Solving for A, we get 1 R2 A= 1+ −1 R1 1 = (1 − Gnominal ) −1

where (b) Gnominal = −100 = −

R1 = R1 Rc Thus,

Rin = 1 k = R1

1 1 1 = + R1 R1 Rc

R1 = 1 k R2 = 100R1 = 100 k

Substituting in Eq. (1), R1 R2 Vo Rc =− R2 R2 Vi R1 1+ + R1 Rc 1+ A R2 Vo = − , we have to make To make Vi R1 1+

R1 = Rc

1+

R2 R2 + R1 Rc A

That is, A

Q.E.D

R2 R1

1+

A 1+

R2 R1

Sedra, Smith, Carusone, Gaudet

(2)

1 + R2 /R1 1+

|G| A 1− = 1 + R2 /R1 |G| 1+ A − A Thus

Gain error G − Gnominal G = − 1 = Gnominal Gnominal − 1 + R2 /R1 1 A − 1 = = /R ) (1 + R /R 1 + R 2 1 2 1 1 + 1 + A A 1

R2 /R1 1 + R2 /R1 1+ A − A

(1)

Dividing Eq. (2) by Eq. (1), we have

−R2 /R1 R2 A 1+ 1+ R1

and Gnominal = −

=

A = 909 V/V

|G| − |G| =

2.33 (a) Equation (2.5): vO = vI

= 101 × 9

R2 /R1 1 + R2 /R1 1+ A If A is reduced by A, |G| will be correspondingly reduced by |G|,

which yields

G=

Again Gnominal = −100 and max = 10% 1 ∴ A = (1 − Gnominal ) −1 1 −1 = [1 − (−100)] 0.1

2.34 (a) |G| =

R1 R1 =1+G+G Rc Rc

Rc A−G = R1 1+G

R2 R1

1 1 R2 − 1+

|G| R1 A − A A = 1 + R2 /R1 |G| 1+ A − A ⎛ ⎞ 1 1 ⎜ R2 ⎟ 1+ − 1⎠ ⎝

A R1 A 1− A = R2 /R1 1+ A − A fOR

A R2 A, 1 and 1 + A R1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–9

|G| |G|

R2 R1 A A A

1+

So

vx R2 =− vI R1

Thus,

R2

|G| / |G| 1 + R2 /R1

A/ |A| A (b)

R4

vx

Q.E.D

|G|

A R2 = 0.1, = 100, ≤ 0.001 |G| R1 A

A ≥ (1 + 100) × 0.1/0.001

iI vI

That is, Amin 104 V/V.

R3

iI

R1

0V

vO

2.35 iI R1

vI

iI

R2

vO A

vO

A

Looking at the figure above, the input current may be written as iI =

v I − v O /A R1

The same current flows through R2 resulting in vO − v O = iI × R 2 A vO v I − v O /A − vO = × R2 A R1 R2 vO R2 1+ − vO = vI × A R1 R1 ⇒

vO = vI

1 A

R2 /R1 1 + RR21 − 1

The denominator of the closed-loop gain expression becomes zero, and the closed-loop gain infinite, if A = 1 + (R2 /R1 ). 2.36 R2 /R1 = 500, R2 = 100 k ⇒ R1 = 200 (a) Rin = R1 = 200 −R2 R4 vO R4 = + (b) 1+ = −500 vI R1 R2 R3 If R2 = R1 = R4 = 100 k ⇒ R3 = 200 Rin = R1 = 100 k 2.37 iI =

vI vI , v x = −iI R2 = − R2 R1 R1

Sedra, Smith, Carusone, Gaudet

100 k 500 − 2

Now, because of the virtual ground at the negative input terminal of the op amp, v O appears across the series combination of R4 and (R2 R3 ); thus (R2 R3 ) vx = vO (R2 R3 ) + R4 = vO

R2 R3 R2 R3 + R2 R4 + R3 R4

vO R2 R3 + R2 R4 + R3 R4 = vx R2 R3 R4 R4 + R3 R2 vO vO vx = vI vx vI R4 R2 R4 1+ =− + R1 R3 R2 = 1+

Q.E.D.

2.38 (a) R1 = R R R R = + =R 2 2 2 R R R3 = (R2 R) + = (R R) + = R 2 2 R R R4 = (R3 R) + = (R R) + = R 2 2 See Fig. 1 on next page. R2 = (R R) +

(b) See Fig. 2 on next page. We utilize the results of (a) above as follows: At node 1 we have a resistance R to ground and, looking left, a resistance R. These two resistances must carry equal currents ⇒ I1 = I. A node equation at node 1 results in the current through (R/2) as 2I. At node 2, we have a resistance R to ground and an equal resistance looking to the left. These two resistances must carry equal currents ⇒ I2 = 2I.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–10

R

R/2 2

1

R/2 3

R/2 4

R

I1

R

I2

I3

I4

R

I

1 R

2

R I1 I

4I R/2

3

R I2 2I

8I R/2

4

I3 4I

R1 = 30 R2

Choosing R2 = 500 ⇒ R1 = 15 k. The circuit will work properly as long as the op amp does not saturate (i.e., as long as VO ≤ −10 V). But V O = Vx − IL RL = −0.1 × 15 − 3.1RL The maximum allowed value for RL can now be found by substituting VO = −10 V, thus

Figure 2 A node equation at 2 results in the current leaving node 2 as 4I. We continue in the same fashion to find I3 = 4I and the current from 3 to 4 as 8I.

RL = 2.74 k 2.40 To obtain an input resistance of 100 k, we select R1 = 100 k. From Example 2.2 we have (1 − x)R4 vO R2 (1 − x)R4 1+ =− + vI R1 R2 R3 + xR4

(c) V1 = −I1 R = −IR V2 = −I2 R = −2IR V3 = −I3 R = −4IR V4 = V3 −

I = 0.1 mA 1.5V = 15 k R= 0.1 mA R1 3.1 = 0.1 1 + R2

= −IR1 − IL RL

0V

2I R/2

IL = 3.1 mA

⇒

Figure 1 I

R1 R1 I = 1+ R2 R2

which is independent of the value of RL . Now, for our specific design:

I R

IL = I + I

R × 8I = −8IR 2

The minimum gain magnitude is obtained when x = 1, vO R2 =− = −1 vI R1 Thus, R2 = 100 k.

2.39 1.5 V I1

R1

Vx

IL RL

R I2

I

R2

The circuit shown above provides a constant current through RL that is independent of the value of RL . The current through RL is determined by the input current and the ratio of the two resistors R1 and R2 . Specifically, I1 = I, Vx = −I1 R1 , I2 = −Vx /R2 = I1 (R1 /R2 ) = I(R1 /R2 ), thus

Sedra, Smith, Carusone, Gaudet

The maximum gain magnitude is obtained when x = 0, R4 R2 R4 vO 1+ = −100 =− + vI R1 R2 R3 100 100 + ⇒1+ = 100 100 R3 100 = 1.02 k ⇒ R3 = 98 When the potentiometer is set exactly in the middle, x = 0.5 and vO R2 0.5R4 0.5R4 =− + 1+ vI R1 R2 R3 + 0.5R4 0.5 × 100 100 0.5 × 100 1+ =− + 100 100 1.02 + 0.5 × 100 = −2.48 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–11

2.41

v1

R2

10 k

v1 10 k v2

v1

Rf 50 k

10 k

Rf

R1

v2

vO

vO v O = −3 sin ωt − 3

Rf Rf Rf v1 + v1 + v2 R1 R2 R3 50 50 50 =− v1 + v1 + v2 10 10 10

vO

vO = −

= −(10v 1 + 5v 2 )

t

0

Now v 1 = 1 V and v 2 = −1 V v O = −(10 × 1 − 5) = −5 V

3 2.42

Rf i1

R1

v1 R2 v2

vO For the summer circuit, we should have

i2

Rf Rf = 1 and =2 R1 R2

The output of the weighted summer circuit is Rf Rf vO = − v1 + v2 R1 R2 v2 v O = − 2v 1 + 2 v1 v2 i1 = and i2 = R1 R2

Select Rf = 2R2 = 20 k. Thus R2 = 10 k, and R1 = 20 k.

Since i1 , i2 ≤ 50 µA for 1-V input signals

2.44 This is a weighted summer circuit: Rf Rf Rf Rf vO = − vO + v1 + v2 + v3 R0 R1 R2 R3

∴ R1 , R2 ≥ 20 k

We may write:

Here

Rf = 2, if R1 = 20 k, Rf = 40 k R1

1 Rf = ⇒ R2 = 2Rf = 80 k R2 2 2.43 For v 1 and v 2 , we assume v 1 = 3 sin ωt v 2 = 1.5 V The output signal required is

Sedra, Smith, Carusone, Gaudet

v 0 = 5V × a0 , v 2 = 5V × a2 , v 1 = 5V × a1 , v 3 = 5V × a3 , 5 5 5 5 a0 + a1 + a2 + a3 v O = −Rf 80 40 20 10 a a1 a2 a3 0 + + + v O = −Rf 16 8 4 2 Rf 0 2 a0 + 21 a1 + 22 a2 + 23 a3 vO = − 16 For −12 V ≤ v O ≤ 0, Rf 0 2 × 1 + 2 × 1 + 22 × 1 + 23 × 1 16

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–12

15Rf = 12 16 ⇒ Rf = 12.8 k

2.48

=

1 M

2.45

10 V

1 k v

R3 10 k R2 10 k

R1 10 k

vI

vO

(a) Source is connected directly. 1 10 mV 1001 10 mV vO = = 10 µA iL = 1 k 1 k Current supplied by the source is 10 µA. v O = 10 ×

Short-circuit R2 : vO =2 vI

(b) Inserting a buffer.

1 M 10 k

R3

R1 v I

vO 1 k

10 V

10 k 10 k

R2

Short-circuit R3 : vO =1 vI

vO v O = 10 V

R4 R3 + R4 R4 1 + R2 /R1 vO R2 = = 1+ vI R1 R3 + R4 1 + R3 /R4

2.46 v + = v I

2.47

10 V = 10 mA 1 k Current supplied by the source is 0.

iL =

The load current iL comes from the power supply of the op amp. 2.49

10 k

1 k

10 k

vO

vO = vI −

v O = 10v I

vO = vI

Sedra, Smith, Carusone, Gaudet

vO

vI

10 10 1+ vO = vI 1 + 10 1 Rin = 11 k

A

vI

vO/A

1 k

vO A

1 1+

1 A

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–13

Error of gain magnitude vO − 1 v 1 I =− 1 A+1 A (V/V)

1000

100

10

vO (V/V) vO

0.999

0.990

0.909

Gain error –0.1%

–1%

–9.1%

R2 R1

(%)

a b c d e f g

–1 1 –1 10 –10 –10 +1

10 10 100 10 100 1000 2

–0.83 0.91 –0.98 5 –9 –9.89 0.67

17 9 2 50 10 1.1 33

v O = −15 +

30 × 25 = +10 V 25 + 100 + 25

⇒ −10 V ≤ v O ≤ + 10 V Pot has 20 turns, and for each turn:

R2 R1

v O =

Case

Gain (V/V)

Rin

R1

R2

a

–10

10 k

10 k

100 k

b

–1

100 k

100 k

100 k

c

–2

100 k

100 k

200 k

d

+1

∞

∞

0

e

+2

∞

100 k

100 k

f

+11

∞

10 k

100 k

g

–0.5

20 k

20 k

10 k

2 × 10 =1V 20

2.54

R2 R1

G=1+

G (V/V)

and to the top:

for a noninverting amplifier: Rin = ∞

A (V/V)

G=−

G0 (V/V)

2.53 Refer to Fig. P2.53. When potentiometer is set to the bottom: 30 × 25 = −10 V v O = v + = −15 + 25 + 100 + 25

2.50 For an inverting amplifier Rin = R1 ,

Case

vO

vI

2.51 For a gain of 1 + (R2 /R1 ) = 20 V/V and a gain error of 1%, Eq. (2.12) provides the minimum required op-amp gain A. 20 × 100 1= A + 20

vO R2 =1+ vI R1 vO R2 =1=1+ vI R1

A + 20 = 2000

(a)

⇒ A = 1980

Set R2 = 0 and eliminate R1 (b)

2.52 For a non inverting amplifier [Eq. (2.11)]: G0 G0 − G × 100 = G0 G0 1+ A for an inverting amplifier (Eq. 2.5):

G=

G=

G0 1 − G0 1+ A

=

Sedra, Smith, Carusone, Gaudet

|G0 | − |G| × 100 |G0 |

vO R2 =2=1+ vI R1

R2 = 1; set R1 = R2 = 10 k R1 (c)

vO R2 = 21 = 1 + vI R1

R2 = 20; set R1 = 10 k, R2 = 200 k R1 (d)

vO R2 = 100 = 1 + vI R1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–14

R2 = 99; set R1 = 10 k, R2 = 990 k R1

when all inputs are present:

2.55 V− = V+ = V; thus the current in the V moving-coil meter will be I = , independent of R the resistance of the meter. To obtain I = 100 µA when V = 10 V, we select 10 = 100 k 0.1 mA The meter resistance does not affect the voltmeter calibration. R=

2.56 Refer to the circuit in Fig. P2.56: (a) Using superposition, we first set v P1 = v P2 , . . . , = 0. The output voltage that results in response to v N1 , v N2 , . . . , v Nn is Rf Rf Rf v ON = − v N1 + v N2 + · · · + v Nn RN1 RN2 RNn Then we set v N1 = v N2 = · · · = 0, then:

v O = v ON + v OP Rf Rf =− v N1 + v N2 + · · · + RN1 RN2 RP RP Rf v N1 + v N2 + · · · 1+ RN RP1 RP2 (b) v O = −4v N1 + v P1 + 3v P2 Rf = 4, RN1 = 10 k ⇒ Rf = 40 k RN1 Rf RP RP 1+ =1⇒5 =1 RN RP1 RP1 Rf RP RP 1+ =3⇒5 =3 RN RP2 RP2

(1) (2)

1 1 1 1 1 , = + + in RP RP RP0 RP1 RP2 Eqs. (1) and (2) and selecting (arbitrarily) RP0 = 10 k results in RP1 = 10 k and RP2 = 3.33 k. The result is the following circuit: Substituting for

RN = RN1 RN2 RN3 · · · RNn 40 k

The circuit simplifies to that shown below.

v OP = 1 +

+v P2

Rf RN

⎛

⎜ ×⎜ ⎝v P1

1/RP1 1 1 1 + + ··· + RP0 RP1 RPn

vN ⎞

vp

⎟ 1/RP2 1/RPn ⎟ · · · + v Pn 1 1 ⎠ 1 1 + ··· + + ··· + RP0 RPn RP0 RPn

vp

10 k

10 k

v

3.33 k

10 k

Rf RN Rp1

vOP

vP1

2.57

Rp2 vP2

Rf vPn Rpn

vI3

RP0

RN

RP1

vI1 v OP

RP2

RP RF v P1 = 1+ RN RP1 + v P2

vO

vI2 RP4

RP RP + ··· + v Pn RP2 RPn

vI4 RP0

where RP = RP0 RP1 · · · RPn

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–15

Adapting the expression given in Problem 2.56 to the circuit above yields RP Rf Rf vO = − v I3 + 1 + v I1 RN RN RP1 RP RP + v I2 + v I4 RP2 RP4

2.58 Refer to Fig. P2.58. Setting v 2 = 0, we obtain the output component due to v 1 as v O1 = −10v 1

where RP = RP0 RP1 RP2 RP3 .

Setting v 1 = 0, we obtain the output component due to v 2 as 10R 10R v O2 = v 2 1 + = 10v 2 R 10R + R

We require

The total output voltage is

v O = −9v I3 + v I1 + 2v I2 + 4v I4

v O = v O1 + v O2 = 10(v 2 − v 1 )

Equating the coefficients of v I3 , we have

For v 1 = 10 sin(2π × 60t) − 0.1 sin(2π × 1000 t)

Rf =9 RN

v 2 = 10 sin(2π × 60t) + 0.1 sin(2π × 1000 t) v O = 2 sin(2π × 1000t)

Selecting RN = 10 k ⇒ Rf = 90 k. Equating the coefficients of v I1 provides RP Rf 1+ =1 RN RP1

2.59

Thus, 10

RP RP =1⇒ = 0.1 RP1 RP1

x1 1x

(1)

Similarly, equating the coefficients of v I2 gives 10

RP RP =2⇒ = 0.2 RP2 RP2

vI

vO

vI

vO

(2)

and equating the coefficients of v I4 gives 10

RP RP =4⇒ = 0.4 RP4 RP4

(3)

R

Now, summing Eqs. (1), (2), and (3) provides 1 1 1 = 0.7 (4) + + RP RP1 RP2 RP4 But, 1 1 1 1 1 = + + + RP RP0 RP1 RP2 RP4

vO R2 1−x 1 =1+ =1+ =1+ −1 vI R1 x x

Thus,

∴

1 1 1 1 1 − = + + RP RP0 RP1 RP2 RP4

(5)

Equations (4) and (5) can be combined to obtain RP RP = 0.7 ⇒ RP0 = 1− RP0 0.3 Selecting RP0 = 10 k Equation (6) ⇒ RP = 3 k Equation (1) ⇒ RP1 = 30 k Equation (2) ⇒ RP2 = 15 k Equation (3) ⇒ RP4 = 7.5 k

Sedra, Smith, Carusone, Gaudet

(6)

vO 1 = vI x

0≤x≤1⇒∞≥

vO ≥1 vI

Add a resistor as shown: vO (1 − x) × 10 k =1+ vI x × 10 k + R vO = 11 For v I max x = 0,

10 k vO = 11 = 1 + vI R

10 k R R = 1 k

10 =

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–16

2.60 An ideal op amp has infinite gain and, thus zero differential input. The op-amp inputs are tied directly to the sinusoidal signal. Thus, the common-mode input is VICM = sin(ωt).

=

11.11 = 10.52 V/V 1.056

2.62

In an inverting configuration, the ideal op amp again ensures zero differential input. Since the positive op-amp input terminal is grounded, the common-mode input is zero.

G=

Voltage followers must tolerate a common-mode input whereas the inverting configuration does not.

or

If R1 = 10 k ⇒ R2 = 90 k

x F

According to Eq. (2.11): R2 R1 1 + R2 /R1 1+ A 10 1 + 90/10 = 9.09 V/V = G= 1 + 90/10 1.1 1+ 100 In order to compensate the gain drop, we can shunt R1 with a resistor Rsh . The value of Rsh can be obtained from 90 90 + 1+ 10 Rsh ⇒ 10 = 90 90 + 1+ 10 Rsh 1+ 100 90 90 10 + 0.1 10 + = 10 + Rsh Rsh 1+

vO = G= vi

1+

9 90 = Rsh Rsh

4

10

0.1 3

10

1

10 2

10

10

Thus for: 10 102 x = 0.01: G0 (V/V) 1 A (V/V) 104 105 106

103 104 7 8 10 10 too high to be practical

10 102 103 104 x = 0.1: G0 (V/V) 1 3 A (V/V) 10 104 105 106 107 x = 1: G0 (V/V) 1 10 102 103 104 2 A (V/V) 10 103 104 105 106 x = 10: G0 (V/V) 1 10 102 103 104 A (V/V) 10 102 103 104 105

Eq. (2.20): Rid = 2R1 ⇒ R1 =

90 k

Rsh

vI

10 k

(a) Ad =

vO

90 10 10 = = = 9.52 V/V 1 + 90/10 1.05 1+ 200

90 90 + 10 81 = 90 90 1+ + 10 81 1+ 200 1+

Sedra, Smith, Carusone, Gaudet

R2 = 1 V/V ⇒ R2 = 10 k R1

R1 = R2 = R3 = R4 = 10 k (b) Ad =

1+

20 = 10 k 2

(Refer to Fig. 2.16.)

If A = 200, then:

Gcompensated

0.01

100 100 −1 x x

2.63 If we assume R3 = R1 , R4 = R2 , then

⇒ Rsh = 81 k

Guncompensated

G0 A ≥ 100 ⇒ A ≥ 100 − 1 G0 x G0 x A F

1+

⇒ A ≥ G0 F, where F =

R2 = 10 V/V 1+ R1

2.61 A = 100 V/V

G0 − G G0 G0 /A × 100 , ≤x × 100 = G0 G0 G0 1+ 1+ A A

R2 = 5 V/V ⇒ R2 = 50 k = R4 R1

R1 = R3 = 10 k (c) Ad =

R2 = 100 V/V ⇒ R2 = 1 M = R4 R1

R1 = R3 = 10 k (d) Ad =

R2 = 0.5 V/V ⇒ R2 = 5 k = R4 R1

R1 = R3 = 10 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–17

This figure belongs to Problem 2.64.

2.64 Refer to the figure above:

(b) Refer to the circuit in Fig. 2.20(b):

v B = v A = 3 + 0.05 sin ωt, V v C = v D = 3 − 0.05 sin ωt, V

1 v I1 = v Icm − v Id 2 1 v I2 = v Icm + v Id 2

Current through R2 , 2R1 , and R2 is i = (v B − v C )/2R1 = 0.05 sin ωt, mA v E = v B + iR2 = 3 + 2.55 sin ωt, V v F = v C − iR2 = 3 − 2.55 sin ωt, V 1 v F = 1.5 − 1.275 sin ωt, V 2 v O = (v F − v E ) × 1 = −5.1 sin ωt, V

vG = vH =

2.65 (a) Refer to the circuit in Fig. 2.20(a). 1 v I1 = v Icm − v Id 2 1 v I2 = v Icm + v Id 2 R2 v I1 v O1 = 1 + R1 R2 R2 1 1+ v Icm − v Id = 1+ R1 2 R1 R2 v I2 v O2 = 1 + R1 R2 R2 1 1+ v Icm + v Id = 1+ R1 2 R1 R2 v Id v Od = v O2 − v O1 = 1 + R1 1 R2 v Icm v Ocm = (v O1 + v O2 ) = 1 + 2 R1 Ad1 = 1 +

R2 R1

Acm1 = 1 +

R2 R1

Ad1 CMRR = 20 log A

cm1

= 0 dB

Sedra, Smith, Carusone, Gaudet

1 v − (A1 ) = v I1 = v Icm − v Id 2 1 v − (A2 ) = v I2 = v Icm + v Id 2 Current through R1 in the upward direction is i= v O1 v O2 v Od

v Id v − (A2 ) − v − (A1 ) = 2R1 2R1

1 1+ = v − (A1 ) − iR2 = v Icm − 2 1 1+ = v − (A2 ) + iR2 = v Icm + 2 R2 = v O2 − v O1 = 1 + v Id R1

R2 v Id R1 R2 v Id R1

1 (v O1 + v O2 ) = v Icm 2 R2 =1+ R1

v Ocm = Ad1

Acm1 = 1

= 20 log 1 + R2 R1 cm1

Ad1 CMRR = 20 log A

Comment: In circuit (a), the first stage amplifies the differential signal and the common-mode signal equally. On the other hand, in circuit (b), the amplifies the differential signal by first stage R2 and the common-mode signal by 1+ R1 unity, thus providing a substantial CMRR. Circuit (a) is useless as a differential amplifier! 2.66 Adding a 100-k potentiometer (whose resistance can be set to any desired value Riv in the range of 0 to 100 k) in series with a fixed resistance Rif makes

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–18

This figure belongs to Problem 2.67

2R1 = R1f + R1v The resulting differential gain will be R4 R2 Ad = 1+ R3 R1f + R1v The minimum gain of 2 is obtained with R1v , at its maximum value of 100 k. Thus R2 R4 (1) 1+ 2= R3 R1f + 100 The maximum gain is obtained with R1v = 0. Thus R2 R4 1+ (2) 1000 = R3 R1f Selecting R3 = R4 at a convenient value of, say, 10 k, then Eqs. (1) and (2) can be solved to obtain R1f = 100.2 and R2 = 100.1 k. Practical values of 100 and 100 k can be used, and the gain range required will be covered.

2.67 (a)

10 vB = 2 V/V =1+ vA 10

1.0 V

1.0 V 1.0 V

1.0 V

vC 20 = −2 V/V =− vA 10

2V

vO (b) v O = v B − v C = 4v A ⇒ = 4 V/V vA (c) v B and v C can be ±4 V, or 8 VP−P . −8 V ≤ v O ≤ 8 V, or 16 VP−P 8 v Orms = √ = 5.66 V 2 This corresponds to 4 VP−P or 1.414 V rms at VI 2.68

R2

R3

vO

vI2

R1

vI1

2V

R4

From Eq. (2.20) Rid = 2R1 = 2 × 5 k = 10 k The two resistance ratios

R1 = R3 = 5 k, R2 = R4 = 100 k Equation (2.15):

R2 R4 = = 20 R3 R1

From Eq. (2.16), vO =

R2 v id R1

Ad =

vO R2 = = 20 V/V v id R1

Sedra, Smith, Carusone, Gaudet

1%. R R ∴ 2 = 0.99 4 R1 R3

R2 R4 and differ by R1 R3

Now for this case, Acm can be found from Eq. (2.19) R2 R4 R3 1− × Acm = R4 + R3 R1 R4 R4 100 R3 × 1 − 0.99 × 100 + 5 R3 R4

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–19

R

Acm = 0.0095 Neglecting the effect of resistance variation on Ad ,

vI/2R R vI/2

I

R2 100 = 20 V V = Ad = R1 5 Ad CMRR = 20 log Acm 20 = 20 log 0.0095

vI

vI 2

vO

vI/2R R

R

vI vI vI + = 2R 2R R Ri = R I=

= 66.4 dB 2.69 Refer to Fig. P2.69:

2.70

Considering that v − = v + : v2 vO − v1 = ⇒ vO = v2 − v1 2 2 v1 =R v 1 only: Ri = I

v1 +

R2 i1

R1

R I v1

vO

vicm

R

vO

0V

R v 2 only: Ri =

Ricm i2

R3

R4

When R2 /R1 = R4 /R3 , the output voltage v O will be zero. Thus, v icm i1 = R1 + R2

R

v2 = 2R I

R

and i2 =

R

vO

v icm R3 + R4

Thus, ii = v icm

I v2

1 1 + R1 + R2 R3 + R4

and

R

R

Ricm = (R1 + R2 ) (R3 + R4 )

v between 2 terminals: v Ri = = 2R I v I connected to both input terminals

2.71 From Eq. (2.19), R2 R 3 R4 1− Acm = R4 + R3 R1 R4

R R 0V

v

vO

I R

Sedra, Smith, Carusone, Gaudet

R

Q.E.D.

The second factor in this expression in effect determines Acm . The largest, |Acm |, will occur when R2 and R3 are at their lowest (or highest) values and R1 and R4 are at their highest (or lowest) values, as thiswill provide the maximum R2 R3 from unity. Thus, deviation of R1 R4 R2 = R2nominal (1 − ) R3 = R3nominal (1 − )

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–20

100 k

R1 = R1nominal (1 + ) R4 = R4nominal (1 + )

100 k

vcm

where R2nominal R4nominal = =K R1nominal R3nominal

100 k

vcm

A B

Substituting in the expression for Acm , R4nominal (1 + ) Acm = × R4nominal (1 + ) + R3nominal (1 − ) (1 − )2 1− (1 + )2

vO

100 k

(c) The circuit becomes as shown below:

For 1, Acm

100 kΩ

K × 4 K+1

vI1

Since

100 kΩ

A 10 kΩ

Ad = K

Ad = K+1 CMRR = A 4

vI2

vO

100 kΩ

B

cm

which in dB becomes K+1 CMRR = 20 log 4

10 kΩ

Q.E.D.

For Ad = 100 V/V and = 0.01, 101 = 68 dB CMRR = 20 log 0.04 To obtain CMRR = 80 dB, 101 = 104 4 101 = 0.25 × 10−2 4 × 104

2.72 (a) Refer to Fig. P2.72 and Eq. (2.19): R2 R3 R4 1− Acm = R3 + R4 R1 R4 100.100 100 1− = 100 + 100 100.100

(1011) v (1011) v

R2 R4 = R1 R3

R2 =1 R1

(b) v A = v B v A = v Icm

100 100 + 100

v Icm v Icm and v B = 2 2 ⇒ −5 V ≤ v Icm ≤ 5 V

vA =

Sedra, Smith, Carusone, Gaudet

(100 // 10) k

I1

I2

(100 // 10) k

vO

100 k

vO =

Refer to Eq. (2.17): ⇒ Ad =

Applying Thévenin’s theorem to v I1 together with the associated (100-k, 10-k) voltage divider, and similarly to v I2 and the associated (100-k, 10-k) voltage divider, we obtain the following circuit: 100 k

That is, the resistor tolerance should be a maximum of 0.25%.

Acm = 0

100 kΩ

100 k (100 10) k

10 10 v I2 − v I1 11 11

= 10(v I2 − v I1 ) For v I1 = v I2 = v Icm , v O = 0, thus Acm = 0 For v I2 − v I1 = v Id , v O = 10 v Id Thus, Ad = 10. To obtain the input common-mode range, we note that for v I1 = v I2 = v Icm ,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–21

v+ = v− =

6.8 = 0.9 ⇒ R5 = 756 6.8 + R5

10 100 v Icm × 11 100 + (100 10)

= 0.833v Icm For v + and v − in the range

2.74 See partial analysis on circuit diagram below.

−2.5 V ≤ v + , v − ≤ +2.5 V

From input loop:

the range of v Icm will be

v Id = 2i1 R1

−3 V ≤ v Icm ≤ +3 V

(1)

From the loop containing R2 , +, −, R2 : v G = i1 R2 + 0 + i1 R2 = 2i1 R2 Thus, we can find the current through RG as v G /RG = 2i1 (R2 /RG ). Finally, from the loop containing v O , R2 , RG , and R2 : R2 R2 − 2i1 R2 v O = −2i1 1 + 2 RG R2 = −4i1 R2 1 + RG

2.73 Refer to Fig. P2.69. Using superposition: v O = v O1 + v O2 Calculate v O1 : v + = v1 −

βv O1 = v− 2

βv O1 βυ O1 − vO v1 2 = 2 ⇒ v O1 = R R β −1

Substituting for 2i1 from Eq. (1), we obtain the gain as R2 R2 vO 1+ Q.E.D. = −2 v Id R1 RG

Calculate v O2 : v O2 v O2 v O2 = v+ ⇒ v2 − = − βv O2 v− = 2 2 2 v2 ⇒ v O2 = 1−β v2 v1 + v O = v O1 + v O2 = β −1 1−β =

2.75 We solve this problem by replacing the network in the dashed box above with its Thévenin equivalent. The resistance seen looking into the network is dominated by the 100 resistor to ground, which is much smaller than the other resistances in parallel with it. Thus,

1 (v 2 − v 1 ) 1−β

Ad =

vO 1 = v2 − v1 1−β

Q.E.D

RTh = 100

R6 Ad = 10 V/V ⇒ β = 0.9 = R5 + R6

The open-circuit voltage at the output of the network can be shown to be (3 − 6x) VTh = 0.01 (1 + x − x2 )

RId = 2R = 2 M ⇒ R = 1 M R ⇒ R5 + R6 ≤ 10 k 100 Selecting R6 = 6.8 k

R5 + R6 ≤

The resulting circuit is the superposition of a difference amplifier with respect to v Id having a

This figure belongs to Problem 2.74. i1 R1 0

R2 0

vId

i1

vG

i1

2i1

R2 RG

R2 R RG 2

RG

2i1 R2

R1

Sedra, Smith, Carusone, Gaudet

i1 2i1

i1

2i1

R2 RG

vO R2

R2 i1 2i1

R2 RG

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–22

100 k

(b) (i)

vO = −1 V/V vi

50 k

v I1

v Id

v I2

50 k

25 k

A vI

vO

25 k C

3 V

O

25 k B

100 k 10 k

vO

25 k D

x 10 k

100

(1 x)

(ii) 3 V

vO = +1 V/V vi

Two possibilities: 25 k A

100 k

50 k v Id

v I2

C

D

vI

A vO

B

v I1

25 k

C

25 k

vi

O vO 25 k D

B

vO 50 k

The circuit on the left ideally has infinite input resistance vO = +2 V/V (iii) vi

100 k RTh

25 k

VTh

C

A vI

gain of 2 (neglecting RTh 100 k), and a non-inverting configuration with respect to VTh having a gain, (iv)

100 50 (1 + )≈1 50 + 100 + 0.1 50

25 k

B

(3 − 6x) (1 + x − x2 )

R2 = 1. R1 Connect C and O together, and D to ground. 2.76 (a) Refer to Eq. (2.17): Ad =

O vO

25 k D

25 k

The change in v O as X goes from 0 to 1 is −60 mV. This range can be increased by increasing the value of the 100 resistor.

vO 1 = + V/V vi 2

Thus, the output is v O = 2v Id + VTh = 2v Id + 0.01

25 k

25 k

A

vI

25 k

C O

25 k

vO

D

B

2.77 Referring to Fig. 2.20(a), the feedback resistor around amplifier A2 is increased to R2 + R2 . With a common-mode input v Icm , the output of A1 is R2 v Icm 1 + R1 whereas the output of A2 is R2 + R2 v Icm 1 + R1

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–23

Thus, the first stage produces a differential output v Icm R2 /R1 . This is amplified by the final difference-amplifier stage having a gain R4 /R3 resulting in an output v O = v Icm ⇒

R2 R4 R1 R3

vO R2 R4 = v Icm R1 R3

The overall common-mode gain can be made zero by modifying the corresponding resistor R1 so that the two noninverting stages at the input have the same gain. That is, 1+

R2 + R2 R2 =1+ R1 R1 + R1

⇒ R1 =

R1 R2 R2

If all resistors have ±1% tolerance, the differential gain will be slightly affected; Eq. (1) indicates that in the worst case Ad can deviate by approximately ±4% of the nominal value. The common-mode gain, however, undergoes dramatic change because of the significant effect of resistor tolerances on the operation of the difference amplifier in the second stage. Equation (2.19) can be employed to evaluate the worst-case common-mode gain. For our case, Acm2 = 0.5[1 − (1 ± 0.04] = ± 0.02 The common-mode gain of the first stage will remain approximately unity. Thus the ±1% resistor tolerances will mainly affect the common-mode gain of the instrumentation amplifier, increasing it in the worst case to |Acm | = 0.02 Correspondingly, the CMRR will be reduced to

2.78 (a) Refer to Fig. 2.20(a). R2 = 101. If The gain of the first stage is 1 + R1 the op amps of the first stage saturate at ±12 V then −12 V ≤ 101v Icm ≤ +12 V

CMRR = 20 log 20 log

As explained in the text, the disadvantage of circuit in Fig.2.20(a) isthat v Icm is amplified by a R2 in the first stage and gain equal to 1 + R1 therefore a very small v Icm range is acceptable to avoid saturation. (b) In Fig. 2.20(b), when v Icm is applied, v − for both A1 and A2 is the same and therefore no current flows through 2R1 . This means the voltage at the output of A1 and A2 is the same as v Icm . ⇒ −12 V ≤ v Icm ≤ 12 V

2.80 See analysis on the circuit diagrams below. Note that circuit (a) has the advantage of infinite input resistance. However, it has the limitation that the load impedance must be floating. This constraint is removed in circuit (b), but the input resistance is finite (2R1 ).

2.81

(1)

Acm = 0 CMRR = ∞ For R2 = R3 = R4 = 100 k, and 2R1 = 10 k. 100 = 21 V/V Ad = 1 1 + 5 CMRR = ∞

Sedra, Smith, Carusone, Gaudet

= 60.4 dB

If 2R1 is reduced to 1 k, Ad increases to 201 V/V while Acm remains unchanged at 0.02 V/V. Thus, CMRR increases to about 80 dB. We conclude that increasing the gain of the first stage increases CMRR.

This circuit allows for a much larger range of v Icm .

Acm = 0

as opposed to the ideal infinite value!

⇒ −0.12 V ≤ v Icm ≤ 0.12 V

2.79 Ideally, R4 R2 Ad = 1+ R3 R1

21 0.02

| Ad | | Acm |

1 Vo (s) =− Vi (s) sCR

1 Vo (jω) =− Vi (jω) jωCR Vo = 1 ∠φ = +90◦ V ωCR i For C = 1 nF and R = 10 k, CR = 1 × 10−9 × 10 × 104 = 10−4 s 1 (a) |Vo /Vi | = 1 at ω = = 104 rad/s. CR 104 Correspondingly, f = = 1.59 kHz. 2π (b) At f = 1.59 kHz, the output sine wave leads the input sine wave by 90◦ .

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–24

This figure belongs to Problem 2.80.

t = 20 s

(c) If the frequency is lowered by a factor of 10, the gain increases by a factor of 10 and, correspondingly, the output voltage increases by a factor of 10.

See figure below.

(d) The phase relation between input and output remains unchanged. 2.82 CR = 1 s and Rin = 100 k. ⇒ R = 100 k 1 = 10 µF C= 100 × 103 When a dc voltage of −1 V is applied, a dc current of 1 V/100 k = 0.01 mA will flow as shown in the figure below. C 10 F

1V

100 k

0.01 mA

0.01 mA

vO

2.83 The current drawn from the input of an inverting integrator with ideal op amp is iI =

The capacitor voltage v O will rise linearly from its initial value of −10 V. Thus, vO = =

In this case 0.01 =

It − 10 C

10−5 t − 10 = t − 10, V 10 × 10−6

Thus v O will reach 0 V at

Sedra, Smith, Carusone, Gaudet

0.1 0.1 ⇒R= = 10 k R 0.01

The change in output voltage over 1 ms is 1 CR

−1 V = −

t = 10 s and will reach 10 V at

vI R

C=

1 104

10−3

0.1dt 0

10−3

0.1dt = 0

Microelectronic Circuits, 8th International Edition

1 × 10−4 = 10 nF 104

© Oxford University Press 2021

Chapter 2–25

Vo (dB) Vi

v O (t) = − =−

It C

0.1 × 10−3 t 159 × 10−12

= −0.63 × 106 t, V At the end of the pulse, t = 10 µs, resulting in v O (10 µs) = −0.63 × 106 × 10 × 10−6

6 dB/octave

= −6.3 V 0

f (log scale)

See Fig. 2.

10 kHz

The integrator frequency is ωint =

vO (V)

1 = 10 kHz CR

1 ⇒ CR = 1.59 µs CR For Rin = 10 k, R = 10 k, and 2.84 2π × 100 × 103 =

1.59 × 10−6 C= = 159 pF 104 To limit the dc gain to 40 dB (i.e., 100 V/V), we connect a resistance RF across C (as in Fig. 2.25) with RF = 100 R = 1 M. The resulting low-pass filter will have a 3-dB frequency of f3dB

1 1 = = 2π CRF 2π × 159 × 10−12 × 106

= 1 kHz When a 10-µs, 1-V pulse (see Fig. 1) is applied at the input, a current of 1 V/10 k = 0.1 mA flows into the integrator. Now we consider two cases: with and without RF .

10 t (s)

6.28 Figure 2 (b) For an integrator with RF , the 0.1-mA current flows through the parallel combination of C and RF . The result is

Figure 3 v O (t) = v Ofinal − (v Ofinal − v Oinitial )e−t/τ where v Ofinal = −IRF = −0.1 × 10−3 × 106 = −100 V

Figure 1

v Oinitial = 0 (a) For an integrator without RF , the 0.1-mA current flows through C and the output voltage decreases linearly from 0 V as

Sedra, Smith, Carusone, Gaudet

τ = CRF = 159 × 10−12 × 106 = 159 µs Thus,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–26

v O (t) = −100(1 − e−t/159 ), V where t is in µs. At the end of the pulse, t = 10 µs, v O (10 µs) = −100(1 − e−10/159 ) = −6.1 V Beyond t = 10 µs, the capacitor discharges through RF . Thus, including RF results in the nonideal integrator response shown in Fig. 3. 2.85 For a time constant of CR = 10−5 s with C = 10−9 F, 10−5 = 10 k 10−9 The magnitude response of the ideal differentiator is shown in Fig. 2.28, and the phase response is constant at +90◦ . The 0-dB frequency is R=

1 1 = = 15.9 kHz 2π × CR 2π × 10−5

provide a high-frequency gain magnitude of 100 V/V, 4 R = 100 ⇒ Ri = R = 10 = 100 R 100 100 i The magnitude and phase response of this modified circuit is shown in the dashed line in the plot. A corner frequency appears in the magnitude response at 1/(2π CRi ) = 1.59 MHz. The phase response decreases to 0◦ around 1.59 MHz, with a value of 45◦ at 1.59 MHz. 1 . If |T| = 100 V/V at ωRC f = 10 kHz, then for |T| = 1 V/V, f has to be 10 kHz ×100 = 1 MHz. 2.86 |T| =

Also, RC =

1 1 = 0.159 µs = ωint 2π × 1 MHz

The magnitude and phase responses are plotted with the solid line.

2.87 For 0 ≤ t ≤ 0.5 ms:

With a series resistor Ri connected to C, the frequency response becomes

v O (t) = 0 −

−

1 RC

t v I dt 0

−sCR R =− Ri + 1/sC 1 + sCRi

t t =− RC 1 ms v O (0.5 ms) = −0.5 V

v O (t) = 0 −

Vo Vi

for 0.5 ≤ t ≤ 1 ms: 1 v O (t) = v O (0.5 ms) − RC

t −1 dt 0.5

v O (t) = −0.5 +

1 (t − 0.5) RC

f (log scale) 15.9 kHz

1.59 MHz

Vo Vi

90°

f (log scale) 1.59 MHz

At high frequencies, the capacitor acts as a short-circuit, and the differentiator becomes an R inverting configuration with a gain − . Thus, to Ri

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–27

vO (V)

2.88 Each pulse provides a constant current of 1V through the capacitor and thus deposits a R 1V charge of × 10 µs on the capacitor, resulting R in a change of the output voltage of 0.5

t (ms)

1

10−5 1 × 10 × 10−6 = − −3 = −0.01 V RC 10 Therefore a total of 100 pulses are required to cause a change of −1 V in v O (t).

−

0.5

vi (t) 0.5 =0V 1 Another way of thinking about this circuit is as follows: 1V flows for 0 ≤ t ≤ 0.5 ms, a current I = R through R and C in the direction indicated on the diagram. At time t we write: v O (1 ms) = −0.5 +

−1 −I t= t I · t = −Cv O (t) ⇒ v O (t) = C RC which indicates that the output voltage is linearly decreasing, reaching –0.5 V at t = 0.5 ms.

I

vO

R

vo (t), mV t (ms) 10 20 30

2.89 The input resistance of Fig. 2.23 is R1 = 10 k. Using the results from Example 2.4 for dc gain K and 3-dB frequency ω0 ,

C

vI

t (ms)

K=−

R2 R1

R2 10 ⇒ R2 = 100 k −10 = −

Then for 0.5 ≤ t ≤ 1 ms, the current flows in the opposite direction, v O rises linearly, reaching 0 at t = 1 ms.

ω0 =

For v I = ±2 V:

2π × 2 × 105 =

We obtain the following waveform (assuming time constant is the same).

⇒ C2 =

If CR is also doubled, then the waveform becomes the same as the first case. vO (V )

1 C 2 R2 1 C2 × 105

1 = 15.9 pF 105 × 2π × 105

Vo (dB) Vi

20 dB

0

1

t (ms)

1

Sedra, Smith, Carusone, Gaudet

2 MHz

Microelectronic Circuits, 8th International Edition

f (log scale)

© Oxford University Press 2021

Chapter 2–28

2.90 Equation (2.5) can be generalized as follows: Vo Z2 /Z1 =− 1 + Z2 /Z1 Vi 1+ A For Z1 = R, Z2 = 1/sC, and A = A0 , Vo =− Vi =−

=−

v O (T) = 1000(1 − e−0.001 ) = −0.9995 −1 V 2.91

vI

1/sCR 1 1 1+ + A0 sA0 CR

1V

1 1 1 CR(1 + ) s + A0 (A0 + 1)CR 1

0

A0 /[(A0 + 1)CR] 1 s+ (A0 + 1)CR

t

0.2 ms

vO (V )

which is low-pass STC function. The pole (or 3-dB) frequency is ωP =

1 (A0 + 1)CR

0.2 ms

t

The ideal integrator has ωP = 0. Observe that as A0 → ∞, ωP → 0. The dc gain is −A0 , which is the dc gain of the op amp. If an ideal Miller integrator is fed with a −1-V pulse signal of width T = CR, the output voltage can be found as follows: The −1-V pulse will cause a current I = 1 V/R to be drawn through R and C. The capacitor voltage, which is v O , will rise linearly according to 1 1 It = t C CR Thus, at t = T (the end of the pulse) the output voltage reaches 1 V and then stays constant at this value. vO =

If the integrator is made with an op amp having a finite A0 = 1000, the response to the −1-V step will be that of an STC low-pass circuit. Thus, v O = v Ofinal − (v Ofinal − v Oinitial )e−t/τ

5

v O = −CR

dv I dt

Therefore: For 0 ≤ t ≤ 0.2 ms: 1V = −5 V 0.2 ms and v O = 0 otherwise.

v O = −1 ms ×

2.92

where

R2

v Ofinal = −1 V × dc gain = −1 V × −1000

R1

= 1000 V

C

Vi

v Oinitial = 0 V τ=

Vo

1 = (A0 + 1)CR = 1001CR ωP

Thus, v O = 1000(1 − e−t/1000CR ) At t = T, which is equal to CR,

Sedra, Smith, Carusone, Gaudet

Vo Z2 =− =− Vi Z1

R2 R1 +

Microelectronic Circuits, 8th International Edition

1 sC

© Oxford University Press 2021

Chapter 2–29

where

Thus, Vo (R2 /R1 )s =− 1 Vi s+ CR1

ω1 =

1 C 1 R1

and

ω2 =

1 C 2 R2

Q.E.D.

Assuming ω2 ω1 , then

which is that of an STC high-pass type.

(a) For ω ω1 ,

R2 High-frequency gain (s → ∞) = − R1

R2 /R1 Vo − Vi 1 + (ω1 /jω)

3-dB frequency (ω3dB ) =

1 CR1

(b) For ω1 ω ω2 Vo −(R2 /R1 ) Vi

For a high-frequency input resistance of 1 k, we select R1 = 1 k. For a high-frequency gain of 40 dB,

(c) For ω ω2

R2 = 100 ⇒ R2 = 100 k R1

Vo R2 /R1 − Vi 1 + j(ω/ω2 )

For f3dB = 2 kHz,

The resulting Bode plot will be as shown:

1 = 2 × 103 2π CR1 ⇒ C = 79 nF The magnitude of the transfer function reduces from 40 dB to unity (0 dB) in two decades. Thus f (unity gain) =

2000 f3dB = = 20 Hz 100 100 Design: Gain of 40 dB ⇒ f1 = 200 Hz ⇒

R2 = 100 R1

1 = 200 2π C1 R1

f2 = 200 kHz ⇒

1 = 200 × 103 2π C2 R2

Input resistance (at ω ω1 ) = 2 k ⇒ R1 = 2 k Thus, R1 = 2 k, R2 = 200 k, C1 0.4 µF, and C2 4 pF. 2.94 Inverting configuration:

2.93 Refer to the circuit in Fig. P2.93. Z2 1 Vo =− =− Vi Z1 Z1 Y2 = −

= −

1 R1 + sC1

1

1 + sC2 R2

R2 /R1 1 1+ (1 + sC2 R2 ) sC1 R1

Vo R2 /R1 (jω) = − Vi [1 + (ω1 /jω)][1 + j(ω/ω2 )]

Sedra, Smith, Carusone, Gaudet

R2 VO = VOS 1 + R1 R2 −0.2 = VOS 1 + R1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–30 100 = VOS 1 + 2

Thus,

⇒ VOS 4 mV 2.95 Input offset voltage = 3 mV Output dc offset voltage = 3 mV × closed loop gain = 3 mV × 1000

2.99

=3V The maximum amplitude of an input sinusoid that results in an output peak amplitude of 12 – 3 = 9 V is given by 9 = 9 mV 1000 If amplifier is capacitively coupled then

vi =

v i max =

12 = 12 mV 1000 (a) IB = (IB1 + IB2 )/2

2.96

1 M

Open input: v O = v + + R2 IB1 = VOS + R2 IB1

(1000/99) k

VO

5.3 = VOS + 10,000IB1

(1)

(a) VO = 200 × 10−9 × 1 × 106 = 0.2 V

Input connected to ground: VOS v O = v + + R2 IB1 + R1 R2 + R2 IB1 = VOS 1 + R1

(b) Largest output offset is

5 = VOS × 101 + 10,000IB1

VO = 1 mV × 100 + 0.2 V = 0.3 V

Equations (1), (2)

100 = 1 +

R2 ⇒ R1 = 10.1 k R1

(c) For bias current compensation, we connect a resistor R3 in series with the positive input terminal of the op amp, with R3 = R1 R2 ,

⇒ IB1 = 530 nA

R3 = 10 k 1 M 10 k

IB IB1 = 530 nA

200 = 20 nA 10 The offset current alone will result in an output offset voltage of

and both flow into the op-amp input terminals.

IOS =

(2)

⇒ 100VOS = −0.3 ⇒ VOS = −3 mV

(b) VOS = −3 mV

IOS × R2 = 20 × 10−9 × 1 × 106 = 20 mV (d) VO = 100 mV + 20 mV = 120 mV = 0.12 V 2.97 VOS = ±2 mV VO = 0.01 sin ωt × 100 + VOS × 100 = 1 sin ωt ± 0.2 V 2.98 VOS =

−1.8 V −1.8 mV 1001

Sedra, Smith, Carusone, Gaudet

(c) In this case, Since R is very large, we may ignore VOS compared to the voltage drop across R.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–31

VOS RIB , Also Eq. (2.46) holds:

5-k resistance in series with R3 . The resulting dc offset voltage at the output will be

R3 = R1 R2

VO = IOS R2 = 0.2 × 100 = 20 mV

therefore from Eq. (2.40): VO = IOS × R2 ⇒ IOS =

Since IOS can be of either polarity,

−0.6 10 M

VO = ±20 mV

IOS = −60 nA

The same result could have been found by replacing R3 in Eq. (1) by (R3 + R4 ) where R4 = 5 k.

2.100

If the signal source resistance is 15 k, then the resistances can be equalized by adding a 5-k resistor in series with the negative input lead of the op amp.

R2 = 100 k 100 k 9 R3 = 5 k

R1 =

2.101

IB1 = 2 ± 0.1, µA, VOS = 0

R2

IB2 = 2 ± 0.1, µA

C1

From Eq. (2.36):

R3 VO = −IB2 R3 + R2 IB1 − IB2 R1

Thus,

R2 VO = IB1 R2 − IB2 R3 1 + R1

R1

C2

R3

(1)

The maximum value of VO is obtained when IB1 = 2.1 µA and IB2 = 1.9 µA, 100 VOmax = 2.1 × 100 − 1.9 × 5 1 + 100/9 = 210 − 95 = 115 mV The minimum value of VO is obtained when IB1 = 1.9 µA and IB2 = 2.1 µA,

R2 = R3 = 100 k 1+

R2 = 100 R1

R1 =

100 k = 1.01 k 99

VOmin = 1.9 × 100 − 2.1 × 5 × 10

1 1 = 2π × 100 ⇒ C1 = R1 C 1 1.01 × 2π × 105

= 190 − 105 = 85 mV

= 1.58 µF

Thus the dc offset at the output will be in the range of 85 mV to 115 mV. The bulk of the dc offset at the output, that due to IB , can be reduced to zero by making the dc resistances seen by the two input terminals equal. Currently, the positive input terminal sees a resistance R3 = 5 k and the negative input terminal sees a resistance equal 100 100 = 10 k. Thus the two to R1 R2 = 9 resistances can be made equal by connecting a

1 1 = 2π × 10 ⇒ C2 = R3 C 2 2π × 106

Sedra, Smith, Carusone, Gaudet

vO

= 0.16 µF 2.102 The component of VO due to VOS is 1 M VO1 = VOS 1 + 10 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–32

1 M

VA

1 M 1 k

I0

= 5(1 + 100) = 505 mV

VO

VOS

Since the two op-amp input terminals see equal resistances of (10 k 1 M), the input bias current IB will not result in a dc voltage at the output; however, the input offset current IOS will give rise to a dc output offset of

v + = v − = VOS

VO2 = IOS × 1 M

VO = VA + 1 M ×

= 0.2 µA × 1 M = 0.2 V

= 1001VOS

Thus,

= 1001 × 3 mV 3V

VO = VO1 + VO2 = 0.505 + 0.2

A large capacitor placed in series with the 1-k resistor results in

= 0.705 V

I = 0 VA = VOS VOS = VOS + 1000VOS 1 k

2.103

1 M VA 1 M I

1 k

1 M VOS

I

VO

VOS

v + = v − = VOS No dc current flows through R1 , C branch ∴ VO = VA + VOS

v − = v + = VOS VA = 2VOS = 6 mV

= 2 VOS + VOS

6 mV VOS = = 6 nA 1 M 1 M VA VO = VA + 1 M × I + 1 k VOS 2 VOS = 2 VOS + 1 M × + 1 M 1 k

= 3 VOS

I=

= 2003 VOS = 2003 × 3 mV ∼ =6V For capacitively coupled input,

Sedra, Smith, Carusone, Gaudet

= 3 × 3 mV = 9 mV 2.104 At 0◦ C, we expect ±20 × 25 × 1000 µV = ±500 mV = ±0.5 V At 75◦ C, we expect ±20 × 50 × 1000 µ = ±1 V We expect these quantities to have opposite polarities.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–33

2.105 R3 = R1 R2 = 10 k 1 M = 9.09 k Now, with the input grounded and assuming VOS = 0, the measured +0.3-V at the output is entirely due to IOS , that is,

At f = 100 kHz, the gain is 40 dB or 100 V/V. Thus

0.3 = IOS R2 = IOS × 1 M

Since ft = A0 fb ⇒ fb =

ft = 100 kHz × 100 = 10 MHz 10 MHz = 125 Hz. 80,000

Thus, IOS = 0.3 µA

2.108 f = 10 kHz |A| = 20 × 103

If VOS = ±1 mV, then it alone will result in an output voltageof R2 = VOS × 101 = ±101 mV or VOS 1 + R1 +0.1 V

f = 100 kHz |A| = 4 × 103

If VOS is positive, 0.1 V of the output 0.3-V offset will be due to VOS , leaving 0.2 V as the result of 0.2 V IOS ; thus in this case, IOS = = 0.2 µA. On 1 M the other hand, if VOS is negative, then −0.1 V of the output 0.3 V is due to VOV , with the result that IOS must be causing 0.4 V of output offset. In this 0.4 V = 0.4 µA. Thus, the possible case, IOS = 1 M range of IOS is 0.2 µA to 0.4 µA.

1 M

10 nF 10 k Vos

IB1

|A| =

A0 1 + (f /fb )2

Substituting the given data, we obtain A0 = 20 × 103 1 + (10/fb )2

(1)

A0 = 4 × 103 1 + (100/fb )2

(2)

Dividing Eq. (1) by Eq. (2), we have 1 + (100/fb )2 =5 1 + (10/fb )2 2500 102 1002 ⇒ 1 + 2 = 25 1 + 2 = 25 + 2 fb fb fb

2.106

R3

Thus, a change of a decade in f does not result in a factor of 10 reduction in gain; in fact, the gain reduces by only a factor of 5. It follows that the first frequency (10 kHz) is less than fb . Therefore, we must use the exact expression for |A|, that is,

Vo

IB2

10,000 2500 − 2 = 24 fb2 fb 7500 = 17.68 kHz fb = 24

(a) R3 = R RF = 10 k 1 M

Now, substituting in Eq. (1) yields 10 2 3 A0 = 20 × 10 1 + = 22,976 V/V 17.68

⇒ R3 = 9.9 k

and the unity-gain frequency is

(b) As discussed in Section 2.8.2, the dc output voltage of the integrator input is when the RF + IOS RF grounded is VO = VOS 1 + R 1 M + 20 nA × 1 M VO = 2 mV 1 + 10 k = 0.202 V + 0.02 V

ft = A0 fb = 22.976 × 103 × 17.68 × 103 = 406.2 MHz 2.109 Using Eq. (2.51), f3dB =

ft 2000 = = 400 kHz 1 + R2 /R1 5

From Eq. (2.53), the closed-loop gain is

VO = 0.222 V

|G| = 2.107 At very low frequencies, the gain is A0 , thus 20 log A0 = 98 dB ⇒ A0 80,000 V/V

Sedra, Smith, Carusone, Gaudet

1+

5

f f3dB

2

Solving for |G| = 5 × 0.99 yields f = 56.8 kHz. The amplifier behaves like a low-pass STC

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–34

circuit with a time constant 1 τ= = 398 ns. 2π × 400 × 103

G=−

tr = 2.2τ = 875 ns

50 1+j

|G| =

2.110 ft = A0 fb

f f3dB 50

1 + (f /f3dB )2

50 = 49.75 V/V For f = 0.1 f3dB , |G| = √ 1.01

A0

fb (Hz)

ft (Hz)

105

102

107

For f = 10 f3dB , |G| = √

106

1

106

which is a 20-dB reduction.

105

103

108

107

10−1

106

2 × 105

10

2 × 106

1+

R2 = 100 V/V R1

f3dB =

(a) A0 = 2 × 105 V/V 5 × 102 = 50 Hz 10

ft

ft = A0 fb = 2 × 10 × 50 = 10 Hz = 10 MHz 7

(b) A0 = 20 × 10 V/V 5

fb =

=

R2 1+ R1

G( jω) =

5

= 4.975 V/V

2.113 ft = 20 MHz and closed-loop gain

2.111 The gain drops by 20 dB at f 10fb . Thus

fb =

50 1 + 100

100 1+j

f f3db

⇒ φ = −tan−1

10 = 1 Hz 10

20 MHz = 200 kHz 100

f f3dB

For φ = −6◦

ft = A0 fb = 20 × 105 × 1 = 2 MHz

f = f3dB × tan 6◦ = 21 kHz

(c) A0 = 1800 V/V

φ = 84◦ ,

0.1 MHz = 10 kHz 10 ft = A0 fb = 1800 × 10 = 18 MHz

f = f3dB × tan 84◦ = 1.9 MHz

fb =

2.114 (a) G = −50 V/V ⇒

(d) A0 = 100 V/V f3dB =

0.1 GHz = 10 MHz 10 ft = A0 fb = 100 × 10 = 1 GHz fb =

ft R2 1+ R1

=

ft 51

For f3dB = 100 kHz ⇒ ft = 100 × 51 = 5.1 MHz

(e) A0 = 25 V/mV = 25 × 103 V/V 250 = 25 kHz fb = 10

(b) G = +50 V/V ⇒ 1 +

fb = A0 × fb = 25 × 103 × 25 × 103 = 625 MHz

f3dB =

2.112 Gnominal = −50 ⇒

R2 = 50 R1

108 = 1.96 MHz 51

Sedra, Smith, Carusone, Gaudet

R2 1+ R1

=

ft 50

(c) G = +2 V/V ⇒ 1 +

ft = 108 Hz

=

ft

R2 = 50 R1

For f3dB = 100 kHz, ft = 5 MHz

A0 = 104

f3dB of closed-loop amplifier =

R2 = 50 R1

f3dB =

ft 1+

R2 R1

ft R2 1+ R1

=

R2 =2 R1

ft 2

For f3dB = 5 MHz, ft = 10 MHz (d) G = −2 V/V ⇒

Microelectronic Circuits, 8th International Edition

R2 =2 R1

© Oxford University Press 2021

Chapter 2–35

f3dB =

ft R2 1+ R1

=

ft 3

For f3dB = 5 MHz, ft = 15 MHz (e) G = −1000 V/V ⇒ f3dB =

ft R2 1+ R1

=

R2 = 1000 R1

ft 1001

For f3dB = 10 kHz, ft = 10 × 1001 = 10.1 MHz R2 =0 (f) G = +1 V/V ⇒ R1 f3dB =

ft R2 R1

1+

R2 =1 R1

ft = = R2 2 1+ R1

For f3dB = 1 MHz, ft = 2 MHz 2.115 Gain = 1 +

R2 = 96 V/V R1

f3dB = 8 kHz ft = 96 × 8 = 768 kHz

G0

G0 ω = jf 1+j 1+ ω1 f1

G=

G0 2 f 1+ f1

G20 2 f 1+ f1 The gain will drop by 3 dB when f3dB 2 √ = 2 1+ f1 √ f3dB = f1 2−1 Q.E.D (b) 40 dB = 20 log G0 ⇒ G0 = 100 = 1 +

for f3dB = 32 kHz 768 = 24 V/V 32

f3dB =

R2 = 10 ⇒ R1 = 1 k and R2 = 9 k R1

When a 100-mV (i.e., 0.1-V) step is applied at the input, the output will be −t/τ

v O = 0.1 × 10(1 − e

), V

where τ=

2.117 (a) Assume two identical stages, each with a gain function:

overall gain of the cascade is

ft

2.116 1 +

Correspondingly, the op amp must have an ft of R2 = 36.6 MHz ft = f3dB 1 + R1

= ft

(g) G = −1 V/V ⇒

Gain =

f3dB

G=

For f3dB = 1 MHz, ft = 1 MHz

f3dB

1 or τ 1 1 = = 3.66 MHz = 2π τ 2π × 43.49 × 10−9

ω3dB =

1 ω3dB

ft R2 1+ R1

=

R2 R1

2 MHz = 20 kHz 100

(c) Each stage should have 20-dB gain or R2 1+ = 10 and therefore a 3-dB frequency of R1 2 × 106 = 2 × 105 Hz 10 √ 2−1 The overall f3dB = 2 × 105 f1 =

= 128.7 kHz,

v O reaches 1% of the 1-V final value at time t, 1 − e−t/τ = 0.99

which is 6.4 times greater than the bandwidth achieved using a single op amp, as in case (b) above.

e−t/τ = 0.01 2.118 ft = 100 × 5 = 500 MHz if a single op amp is used.

t = 4.6τ For t to be 200 ns, 200 = 43.49 ns 4.6 Thus we require a closed-loop 3-dB frequency

τ=

Sedra, Smith, Carusone, Gaudet

With an op amp that has only ft = 50 MHz, multiple stages in series are required. Note that the magnitude response of K non-inverting stages in series, each having a dc gain of G0 , will be

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–36 ⎤K

⎢ |G(f )|K = ⎢ ⎣

⎥ G0 ⎥ 2 ⎦ f 1 + f3dB

V1 V2

where for each stage, f3dB = ft /G0 = 50/G0 MHz. Thus ⎡ ⎤K G 0 K ⎦ |G(f )| = ⎣ 2 1 + G500 f To ensure the 3dB-frequency of the cascade is over 5 MHz, we require ⎡ ⎤K G 0 K ⎦ > √1 |G(5)| = ⎣ G0 2 2 1 + 10 while maintaining an overall dc gain of GK0 = 100. These requirements can be met with K = 3 (but not with K = 2). Specifically, with K = 3 each stage has a dc gain of G0 = 1001/3 = 4.64 V/V and 3-dB frequency 50 = 10.8 MHz f3dB = 4.64 The overall 3-dB frequency is the frequency at which the gain of each stage is 1 dB below its dc value. That is, ⎡ ⎤

1+

f f3dB

1

f f3dB

⎥ −1/20 ⎥ = 0.891 2 ⎦ = 10

V1 + VO /A VO /A VO + VO /A + + =0 R1 R2 Rf ! " A Rf 1+Rf /R1 R2 ⇒ VO = − V1 A R1 1 + 1+R /R f 1 R2 With V1 = 0, a similar analysis yields ! " A Rf 1+Rf /R1 R2 VO = − V2 A R2 1 + 1+R /R f 1 R2 Combining these with superposition yields,

Rf Rf + VO = − R1 R2

2.119 (a)

1+

A 1+Rf /R1 R2

1+

" (1)

A 1+Rf /R1 R2

Rf Rf Rf =1+ + =4 R1 R2 R1 R2

VO = −(V1 + 2V2 )

A/4 1 + A/4

2.121 The peak value of the largest possible sine wave that can be applied at the input without

0.2589f3dB = 5.5 MHz R2 = K, f3dB = R1

!

In this case, since the ideal op-amp circuit will result in VO = −(V1 + 2V2 ), we know Rf /R1 = 1 and Rf /R2 = 2. Thus,

= 0.2589 √

Vo

Substitution into (1) yields the final result,

2

⇒f =

R2

ft R2 1+ R1

=

ft 1+K

GBP = Gain × f3dB K ft = ft GBP = K 1+K K+1 R2 ft (b) 1 + = K, f3dB = R1 K

±14 V = 0.14 V = 140 mV. 100 140 Thus the rms value = √ 100 mV 2

output clipping is

2.122

i

100 kΩ

ft GBP = K = ft K For the same closed-loop gain, the noninverting configuration realizes a higher GBP, and it is independent of the closed-loop gain and equal to ft of the op amp.

1 kΩ vI

1

iO

vO

⎢ ⎢ ⎣

Rf

R1

⎡

iL

RL

2.120 With a finite op-amp gain A, the voltage at the negative op-amp input terminal is −VO /A. A nodal equation at the negative op-amp input terminal with V2 = 0 is

Sedra, Smith, Carusone, Gaudet

(a) RL = 1 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–37

for v Omax = 10 V: Vp =

10 100

0.8 × 10 (0.9 − 0.1) P = = 40 V/µs tr 0.2

SR =

Vp = 0.1 V When the output is at its peak,

vo

10 = 10 mA iL = 1 k −10 i= = −0.1 mA; therefore 100 k iO = 10 + 0.1 = 10.1 mA is just over iOmax = 10 mA.

P

Therefore, the input amplitude must be reduced slightly to ensure an undistorted output. With Vp = 99 mV, the output peak is v O = 9.9 V < v Omax and the output current is 9.9 = 10 mA = iOmax 100 At the negative peak of the output voltage, v O = 9.9 V, iL = −9.9 mA, i = 0.099 mA, and iO = −10 mA = iOmax , the maximum allowed. iO = 9.9 +

0.9 0.5 0.1

If output is at its peak: 10 V = 50 mA 0.2 which exceeds iOmax = 20 mA. Therefore v O cannot go as high as 10 V. Instead:

iL =

vO 10 vO + ⇒ vO = 2V 10 mA = 200 100 k 5.01 2 = 0.02 V = 20 mV Vp = 100 10 V 10 V (c) RL = ?, iOmax = 10 mA = + RLmin 100 k 10 ⇒ RLmin = 1.01 k RLmin

2.123 Op-amp slew rate = 1.2 V/µs.

t tr

tf

2.125 (a) This inverting amplifier has a 3-dB frequency given by Eq. (2.51) f3dB =

ft = 0.5 MHz 1+5

The frequency at which the amplifier gain drops by 1 dB is found from ⎤ ⎡ ⎢ ⎢ ⎣

(b) RL = 200

10 − 0.1 =

W

1+

1

f f3dB

⎥ −1/20 ⎥ = 0.891 2 ⎦ = 10

2

f f3dB

⇒f =

= 0.2589 √

0.2589f3dB = 254 kHz

(b) An output sinusoid with peak amplitude Vo at an angular frequency ω has a maximum rate-of-change of ωVo . Equating this with the op-amp slew rate at ω = 2π × 254 kHz gives the maximum amplitude of an undistorted sinewave at that frequency, This figure belongs to Problem 2.123. vi (V)

2

For the input pulse to rise 2 V, it will take 2 = 1.67 µs. 1.2 ∴ The minimum pulse width = W = 1.67 µs

t W

The output will be a triangular with 2-V peak and 1.2 V/µs slopes.

vO 2.124 W = 2 µs

(V)

2

tr + tf = 0.2 W = 0.4 µs tr = tf = 0.2 µs

Sedra, Smith, Carusone, Gaudet

t

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 2–38

2π × 2.54 × 105 Vomax = 106 ⇒ Vomax =

ωmax = 2 × 106 rad/s and fmax = 318.3 kHz

106 = 0.627 V 2π × 2.54 × 105

(c) At the peak of the output sinusoid, the feedback resistor (having a value of 5 × 20 = 100 k) will draw only 0.627/100 = 0.006 mA. This is negligible compared to the op-amp output current limit of ±1 mA. Thus, we focus on the current required to charge and discharge the load capacitance at the output zero-crossings, where the op amp is on the verge of slew rate limiting at 1 V/ s. The required current is, dvO = C × 106 dt Equating this with the op-amp output current limit gives the largest capacitive load that can be driven in this way,

C×

Cmax × 106 = 10−3 ⇒ Cmax = 10−9 = 1 nF

(b) The output will distort at the value of Vi that dv O results in = SR. dt max

ω × 10Vimax = SR Vimax =

10 × 106 = 0.795 V 2π × 200 × 103 × 10

(c) Vi = 50 mV

Vo = 500 mV = 0.5 V

Slew rate begins at the frequency for which ω × 0.5 = SR ⇒f =

10 × 106 = 3.18 MHz 2π × 0.5

However, the small-signal 3-dB frequency is f3dB =

ft 1+

R2 R1

=

20 × 106 = 2 MHz 10

Thus the useful frequency range is limited to 2 MHz.

dv O 2.126 v O = 10 sin ωt ⇒ dt dv O = 10ω cos ωt ⇒ dt max

(d) For f = 50 kHz, the slew-rate limitation occurs at the value of Vi given by

= 10ω

= 3.18 V

The highest frequency at which this output is possible is that for which dv O = SR ⇒ 10ωmax = 40 × 10+6 ⇒ ωmax dt max

Such an input voltage, however, would ideally result in an output of 31.8 V, which exceeds VOmax .

ωi × 10Vi = SR ⇒ Vi =

= 4 × 10 rad/s 6

Thus Vimax =

10 × 106 2π × 50 × 103 × 10

VOmax = 1 V peak. 10

⇒ fmax = 637 kHz 2.127 (a) Vi = 0.5, Vo = 10 × 0.5 = 5 V Output distortion will be due to slew-rate limitation and will occur at the frequency for dv O = SR which dt max

ωmax × 5 = 10 × 106

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–1

Chapter 3

(c)

Solutions to Exercises within the Chapter

V5V

Ex: 3.1 Refer to Fig. 3.3(a). For v I ≥ 0, the diode conducts and presents a zero voltage drop. Thus v O = v I . For v I < 0, the diode is cut off, zero current flows through R, and v O = 0. The result is the transfer characteristic in Fig. E3.1. Ex: 3.2 See Fig. 3.3(a) and 3.3(b) During the positive half of the sinusoid, the diode is forward biased, so it conducts resulting in v D = 0. During the negative half cycle of the input signal v I , the diode is reverse biased. The diode does not conduct, resulting in no current flowing in the circuit. So v O = 0 and v D = v I − v O = v I . This results in the waveform shown in Fig. E3.2. vˆI 10 V = = 2 mA Ex: 3.3 ˆiD = R 5 k 1 dc component of v O = vˆO π 1 10 = vˆI = π π = 3.18 V

I 0 mA 2.5 k

5 V (d)

V0V

05 2.5 2 mA

I

2.5 k

5 V (e)

I 3V 0 V3V

2V

Ex: 3.4

0 1V

(a)

5 V

I

2.5 k

I

1 k 50 2 mA 2.5

(f)

5 V

V0V

1 k 0

3 V (b)

2.5 k

1 V

V5V

51 1 4 mA

I

V1V

I

I0A

Sedra, Smith, Carusone, Gaudet

0

2 V

5 V

3 3 mA 1

Ex: 3.5 Vavg =

10 π

10 10 k 50 + R = π = 1 mA π ∴ R = 3.133 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–2

Assume VD = 0.7 V

Ex: 3.6 The maximum current arises when |vI | = 20 V. In this case,

ID =

20 − 5 iD = R

Use Eq. (3.5) and note that

To ensure this is 50 mA, R=

5 − 0.7 = 0.43 mA 10 k

V1 = 0.7 V, I1 = 1 mA

20 − 5 = 0.3 k = 300 50

R 10 k

Ex: 3.7 Equation (3.5) I2 V2 − V1 = 2.3 V T log I1

VCC 5 V

V2 − V1 = 2.3 × 25 × 10

VD

At room temperature VT = 25 mV −3

ID

10 × log 0.1

I2 I1 I2 V2 = V1 + 2.3 × VT log I1

V2 − V1 = 2.3 × VT log

= 120 mV Ex: 3.8 i = IS ev /VT

(1)

First iteration

1 (mA) = IS e0.7/VT

(2)

V2 = 0.7 + 2.3 × 25 × 10−3 log

Dividing (1) by (2), we obtain

= 0.679 V

0.43 1

i (mA) = e(v −0.7)/VT

where i is in mA. Thus,

Second iteration 5 − 0.679 = 0.432 mA I2 = 10 k

for i = 0.1 mA,

V2 = 0.7 + 2.3 × 25.3 × 10−3 log

⇒ v = 0.7 + 0.025 ln(i)

v = 0.7 + 0.025 ln(0.1) = 0.64 V and for i = 10 mA, v = 0.7 + 0.025 ln(10) = 0.76 V Ex: 3.9 iD = IS ev /VT ⇒ I = i e−v /VT = 0.25 × e−300/25 S

D

= 1.23 × 106 mA = 1.23 × 10−9 A Ex: 3.10 T = 125 − 25 = 100◦ C IS = 10−14 × 1.15T = 1.17 × 10−8 A

0.432 1

= 0.679 V 0.68 V we get almost the same voltage. ∴ The iteration yields ID = 0.43 mA, VD = 0.68 V b. Use constant voltage drop model: VD = 0.7 V ID =

constant voltage drop

5 − 0.7 = 0.43 mA 10 k

Ex: 3.13 10 V

1V = 1 µA Ex: 3.11 At 20 C I = 1 M Since the reverse leakage current doubles for every 10◦ C increase, at 40◦ C ◦

I

I = 4 × 1 µA = 4 µA

⇒ V = 4 µA × 1 M = 4.0 V @ 0◦ C ⇒V=

I=

R

1 µA 4

2.4 V

1 × 1 = 0.25 V 4

Ex: 3.12 a. Use iteration: Diode has 0.7 V drop at 1 mA current.

Sedra, Smith, Carusone, Gaudet

Diodes have 0.7 V drop at 1 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–3

∴ 1 mA = IS e0.7/VT

(1)

(d)

At a current I(mA), I = IS eVD /VT

V 0.7 V

(2)

Using (1) and (2), we obtain I = e(VD −0.7)/VT

2.5 k

For an output voltage of 2.4 V, the voltage drop 2.4 = 0.8 V across each diode = 3 Now I, the current through each diode, is

0.7 5 2.5 1.72 mA

I

5 V

(0.8−0.7)/0.025

I=e

(e)

= 54.6 mA 10 − 2.4 R= 54.6 × 10−3

I 3V 0 V 3 0.7 2.3 V

2V

= 139

0 1V

Ex: 3.14

2.3 1 2.3 mA

I

(a) 5 V

I

2.5 k

5 0.7 1.72 mA 2.5

(f)

5V 5 1.7 1 3.3 mA

I

V 0.7 V

3 V

2 V (b)

1 V

5 V

I 0 mA

1 k

1 k V 1 0.7

0

1.7 V

0 I

2.5 k V5V

Ex: 3.15 With a reverse voltage, the diode in Fig. 3.10 will not conduct. Thus, the voltage drop on R will be zero, and the reverse voltage on the diode is −VD = −VDD . To ensure we respect the peak inverse voltage, we require VDD = VD > −30 V. Hence, the minimum voltage on VDD is −30 V. Ex: 3.16 When conducting a reverse current of 20 mA, the reverse voltage is

(c) V5V

VZ = VZT + IZ rz = 3.5 V + (20 mA − 10 mA)10

= 3.6 V

I 0 mA

2.5 k

The maximum current, Imax , is the current at 200 mW power dissipation. 200 mW = Imax (3.5 V + 10(Imax − 10 mA))

5 V

Sedra, Smith, Carusone, Gaudet

2 − 0.1Imax 0.2 = 3.5Imax + 10Imax

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–4

15 V

2 ⇒ Imax + 0.34Imax − 0.02 = 0 1 −0.34 + 0.342 + 4 × 0.02 ⇒ Imax = 2 = 51 mA

R IL VO

Ex: 3.17 rd =

VT ID

ID = 0.1 mA

rd =

25 × 10−3 = 250 0.1 × 10−3

ID = 1 mA

rd =

25 × 10−3 = 25 1 × 10−3

ID = 10 mA

rd =

25 × 10−3 = 2.5 10 × 10−3

(b) For VO = 3 V, voltage drop across each 3 = 0.75 V 4

diode =

Ex: 3.18 For small signal model, iD = v D /rd where rd =

(1)

VT ID

iD = IS eV/VT IS =

iD eV/VT

=

5 × 10−3 = 4.7 × 10−16 A e0.75/0.025

For exponential model,

(c) If iD = 5 − i L = 5 − 1 = 4 mA.

iD = IS eV/V T

Across each diode the voltage drop is ID VD = VT ln IS 4 × 10−3 = 25 × 10−3 × ln 4.7 × 10−16

iD2 = e(V2 −V1 ) /V T = ev D /V T iD1 iD = iD2 − iD1 = iD1 ev D /V T − iD1 = iD1 ev D /V T − 1

(2)

In this problem, iD1 = ID = 1 mA.

= 0.7443 V

Using Eqs. (1) and (2) with VT = 25 mV, we obtain

Voltage drop across 4 diodes

v D (mV) iD (mA) iD (mA) small exponential signal model a b c d

− 10 −5 +5 + 10

− 0.4 − 0.2 + 0.2 + 0.4

− 0.33 − 0.18 + 0.22 + 0.49

= 4 × 0.7443 = 2.977 V so change in VO = 3 − 2.977 = 23 mV. Ex: 3.20 When the diode current is halved, the voltage changes by VZ = rz IZ = 80 ×

−5 mA = −200 mV 2

⇒ VZ = 6 − 0.2 = 5.8 V Ex: 3.19

When the diode current is doubled,

VO 20 mV = 20 . (a) In this problem, = iL 1 mA

VZ = rz IZ = 80 × 5 mA = 400 mV

∴ Total small-signal resistance of the four diodes = 20 20 ∴ For each diode, rd = = 5 . 4 VT 25 mV But rd = ⇒5= . ID ID

⇒ VZ = 6 + 0.4 = 6.4 V

∴ ID = 5 mA

= 5.6 V

Finally, the value of VZ0 is that obtained by using the model at zero current. VZ = VZ0 + rz IZ ⇒ VZ0 = VZ − rz IZ = 6 V − 80 × 5 mA

15 − 3 = 2.4 k. and R = 5 mA

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–5

Ex: 3.22

Ex: 3.21 vS

vS

Vs 12 2

input Vs

VD

0 u

2

t

u

VD 0

output ( )

t

VS (a) The diode starts conduction at v S = VD = 0.7 V

√ v S = Vs sin ωt, here Vs = 12 2

(a) As shown in the diagram, the output is zero between (π − θ) to (π + θ)

At ωt = θ,

= 2θ

v S = Vs sin θ = VD = 0.7 V √ 12 2 sin θ = 0.7 0.7 θ = sin−1 2.4◦ √ 12 2

Here θ is the angle at which the input signal reaches VD .

Conduction starts at θ and stops at 180 − θ. ∴ Total conduction angle = 180 − 2θ = 175.2◦ (b) v O,avg =

=

1 2π

(π −θ )

(Vs sin φ − VD ) dφ

(b) Average value of the output signal is given by ⎤ ⎡ (π −θ ) 1 ⎣ VO = 2× (Vs sin φ − VD ) dφ ⎦ 2π

θ

1 −θ [−Vs cos φ − VD φ]φ=π φ−θ 2π

1 [Vs cos θ − Vs cos (π − θ) − VD (π − 2θ )] = 2π But cos θ 1,

cos (π − θ ) − 1,

=

=

1 −θ [−Vs cos φ − VD φ]πφ=θ π

2

2Vs VD − 2π 2

Vs − VD , π

for θ small.

(c) Peak current occurs when φ =

VD Vs − π 2

π . 2

Peak current

√ For Vs = 12 2 and VD = 0.7 V √ 12 2 0.7 − = 5.05 V v O,avg = π 2 (c) The peak diode current occurs at the peak diode voltage. √ V − VD 12 2 − 0.7 ∴ ˆiD = s = R 100 = 163 mA

θ

and

π − 2θ π v O,avg =

∴ Vs sin θ = VD VD θ = sin−1 Vs VD 2θ = 2 sin−1 Vs

=

Vs − V D Vs sin (π /2) − VD = R R

If v S is 12 V(rms), √ √ then Vs = 2 × 12 = 12 2 √ 12 2 − 0.7 163 mA Peak current = 100 Nonzero output occurs for angle = 2 (π − 2θ)

√ PIV = +V S = 12 2

The fraction of the cycle for which v O > 0 is

17 V

=

Sedra, Smith, Carusone, Gaudet

2 (π − 2θ ) × 100 2π

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–6 0.7 2 π − 2 sin−1 √ 12 2 × 100 = 2π

Ex: 3.24 Full-wave peak rectifier:

D1

97.4 % Average output voltage VO is √ Vs 2 × 12 2 − 0.7 = 10.1 V VO = 2 − VD = π π Peak diode current ˆiD is √ ˆiD = Vs − VD = 12 2 − 0.7 R 100

R

C

D2 Vp

= 163 mA

t

assume ideal diodes

Vr{

PIV = Vs − VD + VS √ √ = 12 2 − 0.7 + 12 2

t T 2

= 33.2 V Ex: 3.23

The ripple voltage is the amount of voltage reduction during capacitor discharge that occurs when the diodes are not conducting. The output voltage is given by

Vs

output

input

v O = Vp e−t/RC

2VD t

0 Vs

vO

vS vS

2V sin–1 D VS

(a) VO,avg =

1 2π

(Vs sin φ − 2VD ) dφ

=

2 −θ [−Vs cos φ − 2VD φ]πφ=θ 2π

=

1 [Vs cos φ − Vs cos(π − θ) − 2VD (π − 2θ )] π

But cos θ ≈ 1,

π − 2θ ≈ π . Thus 2Vs − 2VD π

discharge is only half T the period. We also assumed t . 2 T/2 Vr = Vp 1 − e− RC T/2 , for CR T/2 RC T/2 Thus Vr Vp 1 − 1 + RC T/2

e− RC 1 −

Vr =

Vp 2fRC

(a)

Q.E.D.

QSUPPLIED = QLOST

√ 2 × 12 2 − 1.4 = 9.4 V = π (b) Peak diode current =

←

To find the average diode current, note that the charge supplied to C during conduction is equal to the charge lost during discharge.

cos (π − θ ) ≈ − 1

⇒ VO,avg

T/2

Vp − Vr = Vp e− RC

iCav t = CV r

Peak voltage R

√ 12 2 − 1.4 Vs − 2VD = = R 100 = 156 mA

Sedra, Smith, Carusone, Gaudet

Vp Vp = iD,av − IL t = C 2fRC 2fR =

Vp π ωR

iD,av =

√ PIV = Vs − VD = 12 2 − 0.7 = 16.3 V

SUB (a)

Vp π + IL ωtR

where ωt is the conduction angle.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–7

Note that the conduction angle has the same expression as for the half-wave rectifier and is given by Eq. (3.28), 2Vr ∼ ωt = (b) Vp

D4

ac line voltage

π Vp 2Vr ·R Vp

vO

vS

R

Substituting for ωt, we get ⇒ iD,av =

D1

D2

C

D3

+ IL

Since the output is approximately held at Vp , Vp ≈ IL · Thus R Vp + IL ⇒ iD,av ∼ = π IL 2Vr Vp = IL 1 + π Q.E.D. 2Vr

At the end of the discharge interval v O = Vp − 2VD − Vr The discharge occurs almost over half of the time period T/2. For time constant RC

e−t/RC 1 −

T 2

1 T × 2 RC

If t = 0 is at the peak, the maximum diode current occurs at the onset of conduction or at t = −ωt.

1 T ∴ VP − 2VD − Vr = Vp − 2VD 1 − × 2 RC

During conduction, the diode current is given by

T ⇒ Vr = Vp − 2VD × 2RC √ Here Vp = 12 2 and Vr = 1 V

iD = iC + iL iD,max = C

dv S + iL dt t=−ωt

assuming iL is const. iL

Vp = IL R

d Vp cos ωt + IL dt = −C sin ω t × ωVp + IL =C

VD = 0.8 V T=

1 1 = s f 60

√ 1 = (12 2 − 2 × 0.8) ×

1 2 × 60 × 100 × C

√ (12 2 − 1.6) = 1281 µF C= 2 × 60 × 100

= −C sin(−ωt) × ωVp + IL For a small conduction angle

Without considering the ripple voltage, the dc output voltage √ = 12 2 − 2 × 0.8 = 15.4 V

sin(−ωt) ≈ − ωt. Thus ⇒ iD,max = Cωt × ωVp + IL Sub (b) to get 2V r ωVp + IL iD,max = C Vp

If ripple voltage is included, the output voltage is √ Vr = 14.9 V = 12 2 − 2 × 0.8 − 2

Substituting ω = 2π f and using (a) together with Vp /R IL results in Vp iDmax = IL 1 + 2π Q.E.D. 2Vr

IL =

14.9 0.15 A 100

The conduction angle ωt can be obtained using √ Eq. (3.28) but substituting Vp = 12 2 − 2 × 0.8: 2Vr 2×1 ωt = = √ Vp 12 2 − 2 × 0.8

Ex: 3.25

= 0.36 rad = 20.7◦

The output voltage, v O , can be expressed as v O = Vp − 2VD e−t/RC

The average and peak diode currents can be calculated using Eqs. (3.31) and (3.32):

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–8

Vp 2Vr

iDav = IL 1 + π

14.9 V , 100

, where IL =

√ Vp = 12 2 − 2 × 0.8, and Vr = 1 V; thus

VA = 0.7 V + 1 k × 1 mA = 1.7 V For v I = −1 V, the diode is cut off. ∴ vO = 0 V

iDav = 1.45 A

iDmax = I 1 + 2π

Vp 2Vr

v A = −12 V Ex: 3.27

= 2.76 A

PIV of the diodes = VS − VDO

To provide a safety margin, select a diode capable of a peak current of 3.5 to 4A and having a PIV rating of 20 V.

Ex: 3.26

vO R

IL

v I > 0 ∼ diode is cut off, loop is open, and the opamp is saturated: vO = 0 V

i vI

vI

√ = 12 2 − 0.8 = 16.2 V

i

vD

vA

v I < 0 ∼ diode conducts and closes the negative feedback loop:

iD

vO = vI

vO

iR

1 k

Ex: 3.28 Reversing the diode results in the peak output voltage being clamped at 0 V:

vO t The diode has 0.7 V drop at 1 mA current.

10 V

iD = IS ev D /VT iD = e(v D −0.7)/V T 1 mA iD + 0.7 V ⇒ v D = VT ln 1 mA For v I = 10 mV,

v O = v I = 10 mV

It is an ideal op amp, so i+ = i− = 0. 10 mV = 10 µA 1 k 10 µA + 0.7 = 0.58 V v D = 25 × 10−3 ln 1 mA

∴ i D = iR =

v A = v D + 10 mV = 0.58 + 0.01 = 0.59 V For v I = 1 V vO = vI = 1 V vO 1 = = 1 mA 1 k 1 k v D = 0.7 V

iD =

Sedra, Smith, Carusone, Gaudet

Here the dc component of v O = VO = −5 V Ex: 3.29 The capacitor voltage accounts for the shift of the voltage waveforms from vI to vO . Thus, from Fig. 3.30, we see the capacitor voltage is Vp + VCC = 8 V. The diode’s peak inverse voltage arises at the peaks of vO , PIV = vOmax − VCC = VCC + 2Vp − VCC = 2Vp = 6 V Ex: 3.30 Cj0 = 100 fF, V0 = 3 V, m = 3. Using Equation (3.47), 100 fF at VR = 1 V: Cj = 3 = 42.2 fF, and 1 + 13 100 fF at VR = 3 V: Cj = 3 = 12.5 fF. 1 + 33 Ex: 3.31 The reverse current is

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 3–9

−iD = ID + iP = ID + R × P At an incident light power of P = 1 mW, −iD = 10−4 + 0.5 × 1 = 0.5 mA At an incident light power of P = 1µW, −iD = 10−4 + 0.5 × 10−3 = 6 × 10−4 mA = 0.6 µA Ex: 3.32 Neglecting dark current, iP = R × P t

The capacitance is 10 pF per mm2 or, equivalently, 1 nF per cm2 . Thus, Cj = 1 × 3.33 = 3.33 nF Ex: 3.33 9 − 3 × 1.8 = 0.18 k = 180 R= 20 Ex: 3.34 9 − 3 × 2.2 R= = 0.12 k = 120 20

10−6 = 0.3 × 0.01 × A ⇒A=

10−6 = 3.33 × 10−4 m2 = 0.3 × 10−2

3.33 cm2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–1

Solutions to End-of-Chapter Problems

vO = 0 V Neither D1 nor D2 conducts, so there is no output.

3.1 Refer to Fig. P3.1.

(d)

(a) Diode is conducting, thus V = −3 V

vO

+3 − (−3) = 0.6 mA I= 10 k (b) Diode is reverse biased, thus

5V t

I=0 V = +3 V (c) Diode is conducting, thus

Vp+ = 5 V,

V = +3 V

Both D1 and D2 conduct when v I > 0

+3 − (−3) = 0.6 mA 10 k (d) Diode is reverse biased, thus

(e)

I=

Vp− = 0 V,

f = 1 kHz

vO

I=0 V = −3 V

5V 5 V

3.2

t

(a) Vp+ = 5 V,

vO

Vp− = −5 V,

f = 1 kHz

D1 conducts when v I > 0 and D2 conducts when v I < 0. Thus the output follows the input. (f)

5V 0

t

vO 5V

Vp+ = 5 V

t

Vp− = 0 V f = 1 kHz

(b)

vO

Vp+ = 5 V,

Vp− = 0 V,

f = 1 kHz

D1 is cut off when v I < 0 (g)

0

t 5 V

vO t

Vp+ = 0 V

5 V

Vp− = −5 V f = 1 kHz

(c)

vO Vp+ = 0 V,

0

Sedra, Smith, Carusone, Gaudet

t

Vp− = −5 V,

f = 1 kHz

D1 shorts to ground when v I > 0 and is cut off when v I < 0 whereby the output follows v I .

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–2

When v I > 0, D1 is cut off and D2 is conducting. The output becomes 1 V.

(h)

vO 0 V

When v I < 0, D1 is conducting and D2 is cut off. The output becomes:

t

vO = vI + 1 V v O = 0 V ∼ The output is always shorted to ground as D1 conducts when v I > 0 and D2 conducts when v I < 0.

3.3 3 V

(i)

vO

3 mA

1 k

5V

0V

t 2.5 V

Vp+ = 5 V,

D1 ON

I 2 mA

Vp− = −2.5 V,

D2 ON

f = 1 kHz

3 k

When v I > 0, D1 is cut off and v O follows v I . When v I < 0, D1 is conducting and the circuit becomes a voltage divider where the negative peak is 1 k × −5 V = −2.5 V 1 k + 1 k

vO 5V

Vp+ = 5 V,

Vp− = −2.5 V,

3 V (a)

(a) If we assume that both D1 and D2 are conducting, then V = 0 V and the current in D2 will be [0 − (−3)]/3 = 1 mA. The current in the 1 k will be (3 − 0)/1 = 3 mA. A node equation at the common anodes node yields a 2 mA current in D1 . Thus, we obtain the results shown in Fig. (a):

(j)

2.5 V

V0V

1.0 mA

t

I = 2 mA V=0V 3 V

f = 1 kHz

When v I > 0, the output follows the input as D1 is conducting.

3 k

When v I < 0, D1 is cut off and the circuit becomes a voltage divider.

–1.5 V

(k)

0

3 – (–3) 1.5 mA 31

D1 OFF

D2 ON

vO

V –1.5 V

5V 1.5 mA

1V

1 k

t 5 V

Vp+ = 1 V,

3 V (b)

4 V

Vp− = −4 V,

Sedra, Smith, Carusone, Gaudet

f = 1 kHz

(b) In (b), the two resistors are interchanged. With some reasoning, we can see that with D2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–3

1

conducting, the intermediate node V will be below 0 V due to the voltage division between the two resistors. Thus, D1 is off giving the results shown in Fig. (b):

I

1.5 V

VD

I = 0 mA

V = −1.5 V

(a)

√ 120 2 ≥ 4.2 k 3.4 R ≥ 40 The largest reverse voltage appearing across the diode is equal to the peak input voltage: √ 120 2 = 169.7 V

1 I

1.5 V

VD

3.5 The maximum reverse voltage arises when

(b)

vI = V+ − 10 = 3 − 10 = 7 V The maximum forward current arises when vI = Imax × R + V+ = 50 × 0.1 + 3 = 8 V 3.8

Thus, the safe operating range is −7 V ≤ vI ≤ 8 V.

(a)

3.6

D1

100

i1

200 V

D1 5V

V2 5V

1 V D2

V2V

2 V Conducting

Assuming ideal diodes with vI = +200 V, the diode D1 is conducting with zero voltage drop across it. The peak current is thus 200 − 5 = 1.95 A 100 and the peak voltage is

Cutoff

I

2 k

3 V

2 (3) 2 k

2.5 mA

i1 =

(b)

v2 = 5 − (−5) = 10 V

3 V

3.7 (a) The diode is reverse biased, thus I=0A

2 k Conducting

D1

1 V

V1V

VD = −1.5 V (b) The diode is forward biased, thus VD = 0 V I=

3 (1) 2 1 mA

I

2 V D2

Cutoff

1.5 V = 1.5 A 1

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–4

Thus the average value of iB becomes

3.9

iB |av =

10 V

60 × 112.5◦ = 18.75 mA 360◦

3.10 A B X Y 0 0 1 1

100 mA

0 1 0 1

0 0 0 1

0 1 1 1

X = AB, Y = A + B From Fig. P4.5 we see that when v I < VB ; that is, v I < 3 V, D1 will be conducting the current I and iB will be zero. When v I exceeds the battery voltage (3 V), D1 cuts off and D2 conducts, thus steering I into the battery. Thus, iB will have the waveform shown in the figure. Its peak value will be 60 mA. To obtain the average value, we first determine the conduction angle of D2 , (π − 2θ), where 3 = 30◦ θ = sin−1 6

X and Y are the same for A=B X and Y are opposite if A = B 3.11 The analysis is shown on the circuit diagrams below. 3.12 For v I > 0 V: D is on, v O = v I , iD = v I /R For v I < 0 V: D is off , v O = 0, iD = 0

Thus π − 2θ = 180◦ − 60 = 120◦ The average value of iB will be 60 × 120◦ = 20 mA 360◦ If the peak value of v I is reduced by 10%, i.e. from 6 V to 5.4 V, the peak value of iB does not change. The conduction angle of D2 , however, changes since θ now becomes 3 = 33.75◦ θ = sin−1 5.4 iB |av =

and thus π − 2θ = 112.5◦

These figures belong to Problem 3.11.

reverse biased

20 20 10 k

10 10 5 k

10 k V I

20 20 20 2.5 V

5

I

20 k

2.5 0.083 mA 10 20

Sedra, Smith, Carusone, Gaudet

I0

2.5 V

V 0.083 20 1.67 V (a)

V

1.5 V

V 1.5 2.5 1 V (b)

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–5

3.13

R

Using an ideal diode model,

D

vI

vO(peak) 5 iD(peak) = = = 25 mA R 0.2 iD(peak) iD(avg) = = 12.5 mA 2

12 V

vI A 12 V

2

0

conduction occurs

For a conduction angle (π − 2θ ) that is 25% of a cycle 1 π − 2θ = 2π 4 π θ= 4 A = 12/sin θ = 17 V ∴ Peak-to-peak sine wave voltage = 2A = 34 V Given the average diode current to be 1 2π

2π

A sin φ − 12 dφ = 100 mA R

0

1 2π

−17 cos φ − 12φ R

φ = 0.75π = 0.1 φ = 0.25π

R = 8.3 A − 12 = 0.6 A R Peak reverse voltage = A + 12 = 29 V Peak diode current =

3.14 The peak current is given by 24 − 12 = 50 R 24 − 12 ⇒R= = 0.24 k = 240 50 3.15 v I = A sin θ = 12 ∼ conduction through D occurs

Sedra, Smith, Carusone, Gaudet

For resistors specified to only one significant digit and peak-to-peak voltage to the nearest volt, choose A = 17 so the peak-to-peak sine wave voltage = 34 V and R = 8 . Conduction starts at v I = A sin θ = 12 17 sin θ = 12 π rad θ= 4 Conduction stops at π − θ. ∴ Fraction of cycle that current flows is

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–6 π − 2θ × 100 = 25% 2π Average diode current = 1 −17 cos φ − 12φ φ = 3π/4 = 103 mA 2π 8 φ = π/4

For VD = 0.71 V, (2)

I = IS e0.71/VT Combining (1) and (2) gives I = 10−3 e(0.71 − 0.7)/0.025 = 1.49 mA

Peak diode current 17 − 12 = 0.625 A = 8 Peak reverse voltage =

For VD = 0.8 V, (3)

I = IS e0.8/VT

A + 12 = 29 V

Combining (1) and (3) gives

3.16

= 54.6 mA

V 3V 0 –3 V

RED ON OFF OFF

I = 10−3 × e(0.8 − 0.7)/0.025

GREEN OFF OFF ON

Similarly, for VD = 0.69 V we obtain - D1 conducts - No current flows - D2 conducts

I = 10−3 × e(0.69 − 0.7)/0.025 = 0.67 mA and for VD = 0.6 V we have I = 10−3 e(0.6 − 0.7)/0.025

3.17

= 18.3 µA

500

To increase the current by a factor of 10, VD must be increased by VD ,

iD

10 = e VD /0.025

vI 5V

⇒ VD = 0.025 ln10 = 57.6 mV

iD

3.19 IS can be found by using IS = ID · e−VD /VT .

1 500

Let a decrease by a factor of 10 in ID result in a decrease of VD by V: ID = IS eVD /VT

vI 5V 500

iD

1 k vO

vI 5V

iD

Thus the result in each case is a decrease in the diode voltage by 60 mV.

10% of ID gives VD = 0.64 V (b) VD = 0.650 V, ID = 1 mA ⇒ IS = 5.11 × 10−15 A;

vI 7.5 V

10% of ID gives VD = 0.59 V (c) VD = 0.650 V, ID = 10 µA ⇒ IS = 5.11 × 10−17 A;

3.18 I = IS eVD /VT

Sedra, Smith, Carusone, Gaudet

10 = eV/V T ⇒ V 60 mV

(a) VD = 0.700 V, ID = 1 A ⇒ IS = 6.91 × 10−13 A;

1 500

10−3 = IS e0.7/VT

ID = IS e(VD −V)/VT = IS eVD /VT · e−V/V T 10 Taking the ratio of the above two equations, we have

(1)

10% of ID gives VD = 0.59 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–7

(d) VD = 0.700 V, ID = 100 mA ⇒ IS = 6.91 × 10−14 A;

If D2 has saturation current IS , then D1 , which is 20 times larger, has saturation current 20IS . Thus we can write

10% of ID gives VD = 0.64 V

ID2 = IS eVD2 /VT ID1 = 20I S eVD1 /VT

3.20

Taking the ratio of the two equations above, we have 1 (VD2 −VD1 )/VT 5 1 V/V T ID2 e e = = = ID1 3 20 20 100 = 87.7 mV ⇒ V = 0.025 ln 3

I

ID1

ID2

D1

D2

To instead achieve V = 60 mV, we need

VD

ID2 I1 − I2 1 0.06/0.025 e = = = 0.55 ID1 I2 20

Solving the above equation with I1 still at 8 mA, we find I2 = 5.16 mA.

ID1 = IS1 eVD /VT

(1)

ID2 = IS2 eVD /VT

(2)

Summing (1) and (2) gives

3.22 Given two different voltage/current measurements for a diode, we can write ID1 = IS eVD1 /VT ID2 = IS eVD2 /VT

ID1 + ID2 = (IS1 + IS2 )eVD /VT But ID1 + ID2 = I Thus I = (IS1 + IS2 ) eVD /VT

(3)

Taking the ratio of the above two equations, we have ID1 = IS e(VD1 −VD2 )/VT ⇒ VD1 − VD2 ID2 ID1 = VT ln ID2

From Eq. (3) we obtain I VD = VT ln IS1 + IS2

For ID = 1 mA, we have 1 × 10−3 A = −230 mV V = VT ln 10 A

Also, Eq. (3) can be written as IS2 I = IS1 eVD /VT 1 + IS1

⇒ VD = 570 mV (4)

Now using (1) and (4) gives ID1 =

IS1 I =I 1 + (IS2 /IS1 ) IS1 + IS2

We similarly obtain ID2 =

IS2 I =I 1 + (IS1 /IS2 ) IS1 + IS2

3.21 We can write a node equation at the anodes: ID2 = I1 − I2 = 5 mA ID1 = I2 = 3 mA

For ID = 3 mA, we have 3 × 10−3 A = −202 mV V = VT ln 10 A ⇒ VD = 598 mV Assuming VD changes by –2 mV per 1◦ C increase in temperature, we have, for ±20◦ C changes: For ID = 1 mA, 530 mV ≤ VD ≤ 610 mV For ID = 3 mA, 558 mV ≤ VD ≤ 638 mV Thus the overall range of VD is between 530 mV and 638 mV. 3.23 VT =

kT q

We can write the following equation for the diode voltages:

where k = 1.38 × 10−23 J/K = 8.62 × 10−5 eV/K

V = VD2 − VD1

T = 273 + x◦ C

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–8

q = 1.60 × 10−19 C

ID × 10 gives V D = 0.86 V ID /10 gives VD = 0.74 V

Thus −5

◦

VT = 8.62 × 10 × (273 × x C), V x ◦ C VT [mV] −55 0 +55 +125

18.8 23.5 28.3 34.3

(e) VD = 0.6 V, ID = 10 µA ⇒ IS = 3.78 × 10−16 A ID × 10 gives V D = 0.66 V ID /10 gives VD = 0.54 V

3.24 i = I S ev /0.025 ∴ 10,000IS = IS ev /0.025

3.27 The voltage across three diodes in series is 2.0 V; thus the voltage across each diode must be 0.667 V. Using ID = IS eVD /VT , the required current I is found to be 3.9 mA.

v = 0.230 V At v = 0.7 V, i = IS e0.7/0.025 = 1.45 × 1012 IS

If 1 mA is drawn away from the circuit, ID will be 2.9 mA, which would give VD = 0.794 V, giving an output voltage of 1.98 V. The change in output voltage is −22 mV.

3.25 I1 = IS e0.7/VT = 10−3 i2 = IS e0.5/VT 0.5 − 0.7 0.025

i2 = 0.335 µA 3.26 IS can be found by using IS = ID · e−VD /VT . Let an increase by a factor of 10 in ID result in an increase of VD by V: ID = IS eVD /VT 10I D = IS e(VD +V)/VT = IS eVD /VT · eV/V T Taking the ratio of the above two equations, we have V/V T

10 = e

ID × 10 gives V D = 0.76 V ID /10 gives VD = 0.64 V

for VT = 25 mV at 17◦ C

i2 i2 = −3 = e i1 10

(d) VD = 0.70 V, ID = 1 mA ⇒ IS = 6.91 × 10−16 A;

⇒ V 60 mV

Thus the result is an increase in the diode voltage by 60 mV. Similarly, at ID /10, VD is reduced by 60 mV. (a) VD = 0.70 V, ID = 10 mA ⇒ IS = 6.91 × 10−15 A; ID × 10 gives V D = 0.76 V ID /10 gives VD = 0.64 V (b) VD = 0.70 V, ID = 1 mA ⇒ IS = 6.91 × 10−16 A;

3.28 Connecting an identical diode in parallel would reduce the current in each diode by a factor of 2. Writing expressions for the currents, we have ID = IS eVD /VT ID = IS e(VD −V)/VT = IS eVD /VT · e−V/V T 2 Taking the ratio of the above two equations, we have 2 = eV/V T ⇒ V = 17.3 mV Thus the result is a decrease in the diode voltage by 17.3 mV.

3.29 We can write the following node equation at the diode anodes: ID2 = 10 mA − V/R ID1 = V/R We can write the following equation for the diode voltages: V = VD2 − VD1 We can write the following diode equations:

ID × 10 gives V D = 0.76 V

ID2 = IS eVD2 /VT

ID /10 gives VD = 0.64 V

ID1 = IS eVD1 /VT

(c) VD = 0.80 V, ID = 10 A ⇒ IS = 1.27 × 10−13 A;

Taking the ratio of the two equations above, we have

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–9 ID2 10 mA − V/R = = e(VD2 −VD1 )/VT = eV/V T ID1 V/R

= 514.6 mV VR1 = 4 µA × 520 k = 2.08 V

To achieve V = 50 mV, we need

1 µA 4 V2 = 560 − 2.3 × 1 × 25 log 4

At 0◦ C, I =

ID2 10 mA − 0.05/R = e0.05/0.025 = 7.39 = ID1 0.05/R Solving the above equation, we have R = 42 3.30 For a diode conducting a constant current, the diode voltage decreases by approximately 2 mV per increase of 1◦ C. T = −20◦ C corresponds to a temperature decrease of 40◦ C, which results in an increase of the diode voltage by 80 mV. Thus VD = 770 mV. T = + 85◦ C corresponds to a temperature increase of 65◦ C, which results in a decrease of the diode voltage by 130 mV. Thus VD = 560 mV. 3.31

10 V

= 525.4 mV 1 VR1 = × 520 = 0.13 V 4 3.32 For a diode conducting a constant current, the diode voltage decreases by approximately 2 mV per increase of 1◦ C. A decrease in VD by 100 mV corresponds to a junction temperature increase of 50◦ C. The power dissipation is given by PD = (10 A) (0.6 V) = 6 W The thermal resistance is given by 50◦ C T = 8.33◦ C/W = PD 6W 3.33 Given two different voltage/current measurements for a diode, we have

I R1

D1

V1

D2

V2

ID1 = IS e(VD1 −VD2 )/VT ⇒ VD1 − VD2 ID2 ID1 = VT ln ID2 For the first diode, with ID = 0.1 mA and VD = 700 mV, we have ID = 1 mA:

1.0 = 57.6 mV V = VT ln 0.1 ⇒ VD = 757.6 mV

At 20◦ C:

ID = 3 mA:

3 = 85 mV ⇒ VD = 785 mV V = VT ln 0.1

VR1 = V2 = 520 mV R1 = 520 k 520 mV = 1 µA 520 k Since the reverse current doubles for every 10◦ C rise in temperature, at 40◦ C, I = 4 µA I=

ID2

40°C

4 μA

20°C

For the second diode, with ID = 1 A and VD = 700 mV, we have ID = 1.0 mA: 0.001 = −173 mV V = VT ln 1 ⇒ VD = 527 mV ID = 3 mA:

40 mV 1 μA 480

520 mV

V2 = 480 + 2.3 × 1 × 25 log 4

Sedra, Smith, Carusone, Gaudet

V2

0.003 = −145 mV V = VT ln 1 ⇒ VD = 555 mV For both ID = 1.0 mA and ID = 3 mA, the difference between the two diode voltages is approximately 230 mV. Since, for a fixed diode

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–10

current, the diode voltage changes with temperature at a constant rate (–2 mV per ◦ C temp. increase), this voltage difference will be independent of temperature!

r

V

V = Ir + 2.3VT log

2.0 Diode characteristic 1.0 0.8 0.6 0.4 0.2 0

3.34

I

i (mA)

Load line

0.4

I IS

= 0.025 × ln

Thus, a change in current, from I1 to I2 , results in a change in voltage, I2 V2 − V1 = r(I2 − I1 ) + 2.3VT log I1 In this case, r = 10, I1 = 1 mA and I2 = 2 mA. Thus, 2 V2 − V1 = 10 × (2 − 1) + 2.3VT log 1

3 10−12

0.6

0.8

1.0

v (V)

= 0.661 V For i = 0.4 mA, v = 0.668 V Now we can refine the diagram to obtain a better estimate 0.4

i (mA)

= 27.3 mV 0.35

3.35

Load line

R 1 kΩ VDD

i

1V

v

0.30 0.660

0.664

0.67

v (V)

From this graph we get the operating point i = 0.338 mA, v = 0.6635 V

IS = 10

−15

−12

A = 10

mA

Now we compare graphical results with the exponential model.

Calculate some points v = 0.6 V,

At i = 0.338 mA 0.338 i = 0.025 × ln v = VT ln IS 10−12

i = IS ev /V T

= 10−12 e0.6/0.025

0.03 mA

= 0.6637 V

v = 0.65 V,

i 0.2 mA

v = 0.7 V,

i 1.45 mA

The difference between the exponential model and graphical results is = 0.6637 − 0.6635

Make a sketch showing these points and load line and determine the operating point. The points for the load line are obtained using VDD − VD R From this sketch one can see that the operating point must lie between v = 0.65 V to v = 0.7 V i For i = 0.3 mA, v = VT ln IS

ID =

Sedra, Smith, Carusone, Gaudet

= 0.0002 V = 0.2 mV 3.36 IS = 10−15 A = 10−12 mA Use the iterative analysis procedure: 1. VD = 0.7 V, ID =

Microelectronic Circuits, 8th International Edition

1 − 0.7 = 0.3 mA 1K

© Oxford University Press 2021

Chapter 3–11

R 1 k 1V

ID

VD

0.3 ID = 0.025 ln 2. VD = VT ln IS 10−12

= 0.6882 V 1 − 0.6882 4 i= = 0.6235 mA 0.5 k Stop as we are getting the same result. 3.38

= 0.6607 V

||

1 − 0.6882 = 0.6235 mA 0.5 k 0.6235 4. v = 0.7 + 2.3 × 0.025 log 1 3 i=

(a)

1 − 0.6607 = 0.3393 mA 1K 0.3393 = 0.6638 V 3. VD = 0.025 ln 10−12 ID =

3 V

10 k

1 − 0.6638 = 0.3362 mA 1K 0.3362 = 0.6635 V 4. VD = 0.025 ln 10−12 ID =

V

1 − 0.6635 ID = = 0.3365 mA 1 k Stop here as we are getting almost same value of ID and VD

I 3 V V = −3 + 0.7 = −2.3 V 3 + 2.3 10 = 0.53 mA

I=

3.37

500

(b)

1V

3 V

i

v

10 k V

1 − 0.7 = 0.6 mA a) ID = 0.5 k b) Diode has 0.7 V drop at 1 mA current. Use Eq. (3.5): i2 v 2 = v 1 + 2.3VT log i1 1. v = 0.7 V 1 − 0.7 = 0.6 mA 1 i= 0.5 k

Cutoff

I

3 V I =0A V = 3 − I (10) = 3 V (c)

3 V

0.6 2. v = 0.7 + 2.3 × 0.025 log 1

I

= 0.6872 V 1 − 0.6872 = 0.6255 mA 2 i= 0.5 0.6255 3. v = 0.7 + 2.3 × 0.025 log 1

V 10 k

= 0.6882 V

3 V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–12

V = 3 − 0.7 = 2.3 V

assume D1 is also conducting, placing +0.7 V at its anode. Thus, as shown in Fig. (a),

2.3 + 3 I= = 0.53 mA 10 (d)

V =0V 0 − (−3) = 1 mA ID2 = 3 3 − 0.7 = 2.3 mA I + ID2 = 1 ⇒ I = 2.3 − 1 = 1.3 mA

3 V I

Cutoff

These results are consistent with our assumptions.

V

(b) Again assuming D2 is on, a voltage division between the two resistors will negatively bias D1 . Hence, we assume D1 is off. I=0

3 V

ID2 =

3 − 0.7 − (−3) = 1.325 mA 3+1

I =0A

V = −3 + 1.325 × 1 = −1.675 V

V = −3 V

Thus, the anode of D1 is at −1.675 + 0.7 = −0.975 V which confirms that it is reversebiased.

3.39

+3V 3.40

I ID2

1 k

0.7 V I

ON 0.7 V D2 V

D1 ID2

ON

vI

3 k

iD

R

v I,peak − 0.7 ≤ 40 mA R √ 120 2 − 0.7 = 4.23 k R≥ 40 √ Reverse voltage = 120 2 = 169.7 V. iD,peak =

(a)

3 V

3V

ID2

I0

The design is essentially the same since the supply voltage 0.7 V

3 k

D2

D1

OFF

ON V

ID2

3.41 Eq. (3.8), VZ = VZ0 + rz IZ , also applies to the test voltage and current. Thus, VZT = VZ0 + rz IZT

1 k 3 V

We can use this equation to complete the table as follows.

(b)

(a) Assuming a current flow from the positive to the negative supply voltages, D2 will be conducting. A voltage division between the two resistors would place forward bias on D1 , so let’s

Sedra, Smith, Carusone, Gaudet

(a) (b) (c) (d)

VZT 3V 5V 6V 9.1 V

IZT 2 mA 5 mA 2 mA 1 mA

rz 50 100 150 100

Microelectronic Circuits, 8th International Edition

VZ0 VZ @ IZ = 10 mA 2.9 V 3.4 V 4.5 V 5.5 V 5.7 V 7.2 V 9V 10 V

© Oxford University Press 2021

Chapter 3–13

Thus current through each branch is 0.2707 mA.

3.42

The 1 mA will split in =

R 3 kΩ

branches. Choose N = 4.

VDD

ID

There are 4 pairs of diodes in parallel.

3V

a) ID =

1 = 3.69 0.2707

∴ We need 8 diodes. Current through each pair of diodes 1 mA = 0.25 mA = 4 ∴ Voltage across each pair

0.25 = 2 0.7 + 0.025 ln 2

3 − 0.7 = 0.767 mA 3

b) ID = IS eVD /VT ⇒ IS = ID e−VD /VT IS = 0.767 × e−700/25 = 5.3 × 10−16 A

= 1.296 V

c) IS increases by a factor of 100

ID ID − ln ⇒ VD = VT ln 100IS IS = −VT ln 100 = −115 mV

SPECIAL NOTE: There is another possible design utilizing only 6 diodes and realizing a voltage of 1.313 V. It consists of the series connection of 4 parallel diodes and 2 parallel diodes.

3 − (0.7 − 0.115) 3 − 0.585 = 3 3 = 0.805 mA

⇒ ID =

We can iterate the anlysis to improve the accuracy. With ID = 0.805 mA, ID = VD = VT ln 100IS −3 0.805 × 10 25 × ln = 586 mV 5.3 × 10−14

3.44 Refer to Example 3.2. (a) 10 V 5 I 1.861 10 k 0.86 mA

This is almost identical to the value computed above. Thus, the current increases by 0.805 − 0.767 = 0.038 mA = 38 µA.

10 0 1 mA 10

4

V0V 3

1 0.7 V

3.43 Available diodes have 0.7 V drop at 2 mA current since 2VD = 1.4 V is close to 1.3 V, use N parallel pairs of diodes to split the 1 mA current evenly, as shown in the figure next.

5 k

2 0.7 10 5 1.86 mA

10 V

N (b)

⫹

10 V

V

1 mA

⫺ The voltage drop across each pair of diodes is 1.3 V. ∴ Voltage drop across each diode 1.3 = 0.65 V. Using = 2 I2 = I1 e(V2 −V1 )/VT = 2e(0.65−0.7)/0.025 = 0.2707 mA

Sedra, Smith, Carusone, Gaudet

5 k

ID2

I0

V 0.7 V

Cutoff 10 k

ID2

10 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–14 5 k

10 − (−10) − 0.7 = 1.29 mA 15 VD = −10 + 1.29 (10) + 0.7 = 3.6 V ID2 =

V

5 k

0 mA

1.5 V

2.5 V

3.45 (a)

(b) Cutoff (a) I =

1 V

V = 0.072 × 20 = 1.44 V

D1 2 V

2.5 − 0.7 = 0.072 mA 5 + 20

(b) The diode will be cut off, thus

V

D2 2 k

I=0 V = 1.5 − 2.5 = −1 V

I

3.47 VZ0 = 3 V − 5 mA × 100 = 2.5 V.

3 V IZ

V = 2 − 0.7

1 k

= 1.3 V 1.3 − (−3) I= 2 = 2.15 mA

100

VS VZ

(b)

2.5 V

The power dissipated in the zener is PZ = VZ IZ = (2.5 V + 100 × IZ ) × IZ

3 V

We use this to find the current at PZmax = 40 mW,

1 V

2 40 mW = 2.5 V × IZmax + 100 × IZmax

I

2 k

2 ⇒ 100IZmax + 2.5IZmax − 0.04 = 0 √ −2.5 + 2.52 + 4 × 100 × 0.04 ⇒ IZmax = 2 × 100

V

D1

= 0.0111 A = 11.1 mA

2 V

This current corresponds to

D2 Cutoff

VS = 2.5 + 1.1 × 11.1 = 14.71 V and VZ = 2.5 + 0.1 × 11.1 = 3.61 V

V = 1 + 0.7 = 1.7 V I=

3 − 1.7 = 0.65 mA 2

3.48 The voltage change per degree Celsius of the series-connected diodes is the sum of the two diode voltage changes. Thus, the temperature coefficient is +1.5 − 2 = −0.5 mV/◦ C.

3.46

10 10 5 k 0.7 V V 10 5 10 10 2.5 V

I 20 k

iD = IS ev /VT iD2 = e(V2 −V1 )/VT = ev /VT iD1

(a)

Sedra, Smith, Carusone, Gaudet

3.49 Use the exponential diode model to find the percentage change in the current.

For +5 mV change we obtain

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–15

iD2 = e5/25 = 1.221 iD1

3.51

% change =

1.221 − 1 iD2 − iD1 × 100 = × 100 iD1 1

I

= 22.1%

D1

For –5 mV change we obtain

D3

iD2 = e−5/25 = 0.818 iD1 % change =

vo

0.818 − 1 iD2 − iD1 × 100 × 100 = iD1 1

vi

R D2

= −18.1%

D4

Maximum allowable voltage signal change when the current change is limited to ±10% is found as follows: The current varies from 0.9 to 1.1 iD2 = eV/VT iD1 For 1.1, V = 25 ln (1.1) = +2.38 mV For ±10% current change the voltage signal change is from –2.63 mV to +2.38 mV

3.50 The dc current I flows through the diode VT and giving rise to the diode resistance rd = I the small-signal equivalent circuit is represented by

Rs rd

vs

I

(a) The current through each diode is

For 0.9, V = 25 ln (0.9) = −2.63 mV

vo

VT /I VT rd = vs = vs VT r d + RS VT + IRS + RS I 25 mV Now, v o = 10 mV × 25 mV + 103 I

0.05 VT 2VT = = I I I 2 From the equivalent circuit vo R R = = vi R + (2rd 2rd ) R + rd I

vo vi

rd

0 µA ∞ 0 1 µA 50 k 0.167 10 µA 5 k 0.667 100 µA 500 0.952 50 0.995 1 mA 5 0.9995 10 mA

rd

rd

vo vI

rd

rd R

10 k

vo

Equivalent Circuit

1 mA 0.24 mV 0.1 mA 2.0 mV 9.6 mV 1 µA For v o =

I : 2

rd =

vo = vs ×

I

10 k

(b) For signal current to be limited to ±10% of I (I is the biasing current), the change in diode voltage can be obtained from the equation

1 0.025 vs = vs × 2 0.025 + 103 I

⇒ I = 25 µA

Sedra, Smith, Carusone, Gaudet

iD = eVD /V T = 0.9 to 1.1 I v D = −2.63 mV to +2.32 mV

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–16

±2.5 mV so the signal voltage across each diode is limited to 2.5 mV when the diode current remains within 10% of the dc bias current.

From the results of (a) above, for I = 1 mA, v o /v i = 0.995; thus the maximum input signal will be vˆ i = vˆ o /0.995 = 1/0.995

∴ v o = 10 − 2.5 − 2.5 = 5 mV

= 1.005 V

5 mV = 0.5 µA and i = 10 k 2.5 2.5 mV mV

The same result can be obtained from the figure above where the signal across the two series diodes is 5 mV, thus vˆ i = vˆ o + 5 mV = 1 V + 5 mV = 1.005 V. i vo

10 mV

R 2.5 mV

3.52

10 k

0.1 A

2.5 mV

The current through each diode 0.5 µA = 0.25 µA 2 The signal current i is 0.5 µA, and this is 10% of the dc biasing current. =

∴ DC biasing current I = 0.5 × 10 = 5 µA (c) Now I = 1 mA. ∴ ID = 0.5 mA

Small signal resistance of each diode

Maximum current derivation 10%. 0.5 = 0.05 mA ∴ id = 10 and i = 2id = 0.1 mA. ∴ Maximum v o = i × 10 k = 0.1 × 10 =1V

Ten diode connected in parallel and fed with a total current of 0.1 A. So the current through each 0.1 = 0.01 A diode = 10

=

VT 25 mV = = 2.5 iD 0.01 A

Equivalent resistance, Req , of 10 diodes connected in parallel is given by Req =

2.5 = 0.25 10

If there is one diode conducting 0.1 A current, then the small signal resistance of this diode =

25 mV = 0.25 0.1 A

This value is the same as of 10 diodes connected in parallel. If 0.2 is the resistance for making connection, the resistance in each branch = rd + 0.2 = 2.5 + 0.2 = 2.7 For a parallel combination of 10 diodes, equivalent resistance, Req , is Req =

2.7 = 0.27 10

If there is a single diode conducting all the 0.1 A current, the connection resistance needed for the single diode will be = 0.27 − 0.25 = 0.02 .

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–17

3.53 As shown in Problem 3.50, vo VT 0.025 = = vi VT + RS I 0.025 + 104 I

(1)

Here RS = 10 k The current changes are limited ±10%. Using exponential model, we get

The variation is –2.63 mV to 2.38 mV for ±10% current variation. Thus the largest symmetrical output signal allowed is 2.38 mV in amplitude. To obtain the corresponding input signal, we divide this by (v o /v i ): (2)

I I I I

Now for the given values of v o /v i calculate I and vˆ S using Equations (1) and (2)

I I

I in mA vˆ s in mV

0.5 0.1 0.01 0.001

0.0025 0.0225 0.2475 2.4975

vo =0 vi

vo = 1 × 10−3 = 0.001 V/V vi vo = 10 × 10−3 = 0.01 V/V = 10 µA, vi vo = 100 × 10−3 = 0.1 V/V = 100 µA, vi vo = 500 × 10−3 = 0.5 V/V = 500 µA, vi vo = 600 × 10−3 = 0.6 V/V = 600 µA, vi vo = 900 × 10−3 = 0.9 V/V = 900 µA, vi vo = 990 × 10−3 = 0.99 V/V = 990 µA, vi

I = 1 µA,

For 1.1, v = 2.38 mV

vo vi

where I is in mA

Now I = 0 µA,

For 0.9, v = −2.63 mV

2.38 mV v o /v i

VT vo rd2 I 1 m −I = = = VT VT vi rd1 + rd2 1m + I 1m−I Thus vo = I, vi

iD2 = ev /VT = 0.9 to 1.1 iD1 iD2 and here v = 25 × 10−3 ln iD1

vˆs =

where we have replaced the large capacitors C1 and C2 by short circuits:

I = 1 mA

4.76 23.8 238 2380

= 1000 µA,

vo = 1000 × 10−3 = 1 V/V vi

3.55

I

3.54

D1

1 mA C 2 vo C1

D1

D2

vi

vI

I

i2 D2

i1

i3

v1

v3

v 2

v4

D3 iO

vO

i4 R 10 k D4

I When both D1 and D2 are conducting, the small-signal model is I = 1 mA

rd1 vi

vo rd2

Sedra, Smith, Carusone, Gaudet

Each diode exhibits 0.7 V drop at 1 mA current. Using diode exponential model we have i2 v 2 − v 1 = VT ln i1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–18

For I = 0.5 mA, the output will saturate at 0.5 mA ×10 k = 5 V.

and v 1 = 0.7 V, i1 = 1 mA i ⇒ v = 0.7 + VT ln 1

vo (V )

= 700 + 25 ln(i)

I 1 mA

10

Calculation for different values of v O : v O = 0, iO = 0, and the current I = 1 mA divide equally between the D3 , D4 side and the D1 , D2 side. I i1 = i2 = i3 = i4 = = 0.5 mA 2 v = 700 + 25 ln(0.5) 683 mV

I 0.5 mA

5

10.7 5.68

5.68

10.7

v1 (V)

5

v 1 = v 2 = v 3 = v 4 = 683 mV From the circuit, we have v I = −v 1 + v 3 + v O = −683 + 683 + 0 = 0 V 1 = 0.1 mA 10 k Because of symmetry of the circuit, we obtain

For v O = 1 V, iO =

3.56 Representing the diode by the small-signal resistances, the circuit is

iO I + = 0.5 + 0.05 = 0.55 mA and 2 2 i4 = i1 = 0.45 mA i2 = 685 mV v 3 = v 2 = 700 + 25 ln 1 i4 = 680 mV v 4 = v 1 = 700 + 25 ln 1 i3 = i 2 =

vO (V)

iO i3 = i2 i4 = i1 v 3 = v 2 v 4 = v 1 (mA) (mA) (mA) (mV) (mV)

v I = −v 1 + v 3 + v O (V)

0

0

0.5

0.5

683

683

0

+1

0.1

0.55

0.45

685

680

1.005

+2

0.2

0.6

0.4

∼ 687

677

2.010

+5

0.5

0.75

0.25

∼ 693

665

5.028

+9

0.9

0.95

0.05

∼ 699

∼ 625

9.074

0.995

0.005

∼ 700

568

10.09

9.99

0.999 0.9995 0.0005 ∼ 700

510

10.18

10

1

0

10.7

+ 9.9 0.99

1

0

700

v I = −v 1 + v 2 + v O = −0.680 +0.685 + 1 = 1.005 V Similarly, other values are calculated as shown in the table. The largest values of v O on positive and negative side are +10 V and −10 V, respectively. This restriction is imposed by the current I = 1 mA

rd

vi

C

vo

VT rd ID 1 sC

1 = 1 1 + sCrd rd + sC Vo 1 = Vi 1 + jωCrd ωCrd Phase shift = −tan−1 1 VT = −tan−1 ωC I Vo = Vi

For phase shift of −45◦ , we obtain −45 = −tan−1 2π × 100 × 103 × 10 × 10−9 ×

0.025 I

⇒ I = 157 µA

A similar table can be generated for the negative values. It is symmetrical.

Now I varies from

For v I > +10.7, v O will be saturated at +10 V and it is because I = 1 mA. Similarly, for v I < −10.7 V, v O will be saturated at −10 V.

Range of phase shift is −84.3◦ to −5.71◦

Sedra, Smith, Carusone, Gaudet

157 µA to 157 × 10 µA 10 Range is 15.7 µA to 1570 µA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–19

5 V

3.57 V

I

R

R

IL

Vo

VO

ID

RL

(a)

VO rd = = V + R + rd

VT /I VT R+ I

⇒ IS = 6.91 × 10−16 A

VT = IR + VT

Voltage drop across each diode =

V + − VO V + − 0.7 = . R R VT = VT + (V + − 0.7)

∴ ID = IS eV/V T = 6.91 × 10−16 × e0.75/0.025

For no load, I = ∴

VO V +

= 7.38 mA IL = 1.5/1.5 = 1 mA

R V

VO

rd

Small-signal model

(b) If m diodes are in series, we obtain mVT mrd VO = = V + mrd + R mVT + IR =

mVT mVT + (V + − 0.7m)

I = ID + IL = 7.39 mA + 1 mA = 8.39 mA ∴R=

5 − 1.5 = 417 8.39 mA

Use a small-signal model to find voltage VO when the value of the load resistor, RL , changes: rd =

VT 0.025 = 3.4 = ID 7.39

When load is disconnected, all the current I flows through the diode. Thus ID = 1 mA VO = ID × 2rd

(c) For m = 1 VO VT = 1.75 mV/V = V + VT + V + − 0.7 For m = 4 VO mVT = 8.13 mV/V = V + mVT + 15 − m × 0.7 3.58

= 1 × 2 × 3.4 = 6.8 mV With RL = 1 k, IL

1.5 V = 1.5 mA 1

IL = 0.5 mA

Diode has 0.7 V drop at 1 mA current VO = 1.5 V when RL = 1.5 k

ID = −0.5 mA

ID = IS eV/V T

VO = −0.5 × 2 × 3.4

1 × 10−3 = IS e0.7/0.025

= −3.4 mV

Sedra, Smith, Carusone, Gaudet

1.5 = 0.75 V. 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–20

11 V

With RL = 750 , 1.5 = 2 mA 0.75

IL

I

R

IL = 1 mA ID = −1 mA

VO VZ

VO = −1 × 2 × 3.4 rZ

= −6.8 mV

VZ0

With RL = 500 , 1.5 = 3 mA 0.5

IL

(b)

9 V

IL = 2.0 mA ID = −2.0 mA

I

R

VO = −2 × 2 × 3.4 VO VZ

= −13.6 mV

0.5 mA

3.59

IL

RL

10 V (c)

I

R and

VO VZ

IZ

IL

R=

10 − 7.5 10 − VO = = 167 I 15

When the supply undergoes a change VS , the change in the output voltage, VO , can be determined from

RL

(RL rz ) VO = VS (RL rz ) + R

(a)

=

To obtain VO = 7.5 V, we must arrange for IZ = 10 mA (the current at which the zener is specified). Now,

1.5 0.03 = 0.15 (1.5 0.03) + 0.167

For VS = +1 V (10% high), VO = +0.15 V and VO = 7.65 V. For VS = −1 V (10% low), VO = −0.15 V and VO = 7.35 V.

VO 7.5 = 5 mA = IL = RL 1.5

When the load is removed and VS = 11 V, we can use the zener model to determine VO . Refer to Fig. (b). To determine VZ0 , we use

Thus

VZ = VZ0 + IZT rz

I = IZ + IL = 10 + 5 = 15 mA

7.5 = VZ0 + 0.01 × 30

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–21 VO 1 = − (rd R) = − 1 1 IL + rd R

⇒ VZ0 = 7.2 V From Fig. (b) we have I=

11 − 7.2 = 19.3 mA 0.167 + 0.03

=−

1 ID ID + V + − 0.7 VT

=−

VT × ID

=−

V + − 0.7 VT × ID VT + V + − 0.7

Thus VO = VZ0 + Irz = 7.2 + 0.0193 × 30 = 7.78 V To determine the smallest allowable value of RL while VS = 9 V, refer to Fig. (c). Note that IZ = 0.5 mA, thus VZ = VZK VZ0 = 7.2 V 9 − 7.2 = 10.69 mA I= 0.167

VO mV ≤5 IL mA

V + − 0.7 VT 5 mV × ≤ ID VT + V + − 0.7 mA

mV 25 15 − 0.7 ≤5 × ID 0.025 + 15 − 0.7 mA i.e.,

IL = I − IZ = 10.69 − 0.5 = 10.19 mA RL =

For

1 VT +1 V + − 0.7

VO 7.2 = = 707 IL 10.19

ID ≥ 4.99 mA

VO = 7.2 V

ID 5 mA

3.60

R=

V

15 − 0.7 V + − 0.7 = ID 5 mA

R = 2.86 k

R

Diode should be a 5-mA unit; that is, it conducts 5 mA at VD = 0.7 V, thus 5 × 10−3 = IS e0.7/0.025 .

IL VO

(c) For m diodes connected in series we have

VO IL

⇒ IS = 3.46 × 10−15 A

ID =

V + − 0.7m R

and rd =

rd

R

So now

Small-signal model (a) From the small-signal model

(b) At no load, ID = rd =

VO 1 = −(R mrd ) = − 1 1 IL + R mrd

=−

1 ID ID + V + − 0.7m mVT

=−

mV T ID

VO = −I L (rd R) VO = − (rd R) IL

VT ID

V+ V + − 0.7 R

=−

mV T +1 − 0.7m

mVT V + − 0.7m ID V + − 0.7m + mVT

VT ID

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–22

3.62 (a) Three 6.8-V zeners provide 3 × 6.8 = 20.4 V with 3 ×10 = 30- resistance. Neglecting R, we have

3.61

Load regulation = −30 mV/mA.

I vO 1.5 V VO IL

(b) For 5.1-V zeners we use 4 diodes to provide 20.4 V with 4 ×30 = 120- resistance. Load regulation = −120 mV/mA

ID

3.63

82 IL varies from 2 to 7 mA

vS

8

To supply a load current of 2 to 7 mA, the current I must be greater than 7 mA. So I can be only 10 mA or 15 mA.

vO

Small-signal model

Now let us evaluate VO for both 10-mA and 15-mA sources. For the 10-mA source: Since IL varies from 2 to 7 mA, the current ID will varry from 8 to 3 mA.

From the small-signal model we obtain 8 v O 8 = = v S 8 + 82 90

Correspondingly, the voltage across each diode changes by VD where

Now v S = 1.0 V.

3 = e VD /VT 8

∴ v O =

⇒ VD = 25 ln

3 = −24.5 mV 8

and the output voltage changes by VO = 2 × VD = −49 mV

8 = e VD /VT 13

8 13

= −12.1 mV

and the output voltage changes by VO = 2 × VD = −24.2 mV which is less than half that obtained with the 10-mA supply. Thus, from the point of view of reducing the change in VO as IL changes, we choose the 15-mA supply. Note, however, that the price paid is increased power dissipation.

Sedra, Smith, Carusone, Gaudet

= 88.9 mV

3.64 (a) VZT = VZ0 + rz IZT 10 = VZ0 + 7 (0.025)

With I = 15 mA, the diodes current changes from 13 to 8 mA. Correspondingly, the voltage across each diode changes by VD where

⇒ VD = 25 ln

8 8 v S = × 1.0 90 90

⇒ VZ0 = 9.825 V (b) The minimum zener current of 5 mA occurs when IL = 20 mA and VS is at its minimum of 20(1 − 0.25) = 15 V. See the circuit below. R≤

15 − VZ0 20 + 5

where we have used the minimum value of VS , the maximum value of load current, and the minimum required value of zener diode current, and we assumed that at this current VZ VZ0 . Thus, R≤

15 − 9.825 + 7 25

≤ 207 .

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–23

VS 15 V

20 +

1 (20) = 25 V. 4

VZ = VZ0 + rZ IZ

R = 9.825 + 7 ×

20 mA VO IZmin 5 mA

25 − VZ 207

207VZ = 207 (9.825) + 7 (25) − 7VZ ⇒ VZ = 10.32 V

IZ RL

IZmax =

25 − 10.32 = 70.9 mA 0.207

PZ = 10.32 × 70.9 = 732 mW

∴ use R = 207

3.65

(c) Line regulation =

mV 7 = 33 207 + 7 V

±25% change in v S ≡ ± 5 V VO changes by ±5 × 33 = ±0.165 mV

vS

vO

R

±0.165 × 100 = ±1.65% 10

corresponding to

(d) Load regulation = − (rZ R) Using the constant voltage drop model:

= −(7 207) = −6.77 or –6.77 V/A

ideal

VD 0.7 V

VO = −6.77 × 20 mA = −135.4 mV 0.1354 × 100 = −1.35% corresponding to − 10 (e) The maximum zener current occurs at no load IL = 0 and the supply at its largest value of

25 V

R 1 k

vS

vO

(a) v O = v S + 0.7 V,

For v S ≤ − 0.7 V

v O = 0, for v S ≥ −0.7 V

207 vO VZ

0.7 V

IL 0

0

vS

slope 1

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–24 √ vˆO = 10 2 − 0.7 = 13.44 V

(b)

vS

10 V

vO t

0.7 V 10 V

Conduction begins at √ 10 2 sin θ = 0.7 0.7 θ = sin−1 = 2.84◦ √ 10 2 = 0.0495 rad

0.7 V

Conduction ends at π − θ. ∴ Conduction angle = π − 2θ = 3.04 rad

u

The diode conducts for 3.04 × 100 = 48.4% of the cycle 2π

(c) The diode conducts at an angle 0.7 θ = sin−1 = 4◦ and stops 10

v O,avg =

at π − θ = 176◦

1 2π

Thus the conduction angle is π − 2θ

= 4.15 V

= 172◦ or 3 rad.

iD,avg =

v O,avg =

=

−1 2π

π −θ

√ (10 2sinφ − 0.7) dφ

θ

v O,avg = 4.15 mA R

π −θ (10 sin φ − 0.7) dφ

3.67

θ

−1 [−10 cos φ − 0.7φ]π−θ θ 2π

0.5 k

= −2.85 V (d) Peak current in diode is 10 − 0.7 = 9.3 mA 1 (e) PIV occurs when v S is at its the peak and v O = 0. PIV= 10 V

3.66

ideal 0.7 V

12 : 1

vS

10 Vrms

120 Vrms 60 Hz

Sedra, Smith, Carusone, Gaudet

1 k vO

√ vˆO = 10 2 − VD = 13.44 V Conduction starts at θ = sin−1

0.7 √ = 10 2

2.84◦ = 0.05 rad and ends at π − θ. Conduction angle = π − 2θ = 3.04 rad in each half cycle. Thus the

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–25

fraction of a cycle for which one of the two 2(3.04) × 100 = 96.8% diodes conduct = 2π Note that during 96.8% of the cycle there will be conduction. However, each of the two diodes conducts for only half the time, i.e., for 48.4% of the cycle. v O,avg

1 = π

π −θ √ 1 −12 2 cos φ − 1.4φ θ π √ 2(12 2 cos θ) 1.4 (π − 2θ ) = − π π =

π −θ √ (10 2sinφ − 0.7)dφ θ

= 7.65 V iR,avg =

v O,avg 7.65 = = 7.65 mA R 1

3.69

= 8.3 V iL,avg =

8.3 = 16.6 mA 0.5 k

vS vS

120 Vrms

3.68

R

vO

12 : 1 D4

10 Vrms vs

120 Vrms

D1

Refer to Fig. 3.22.

R For VD Vs , conduction angle π , and

1 k D3

D2

VD 0.7 V

v O,avg =

2 2 Vs − VD = Vs − 0.7 π π

(a) For v O,avg = 10 V π × 10.7 = 16.8 V 2 √ 120 2 = 10.1 to 1 Turns ratio = 16.8

√ Peak voltage across R = 10 2 − 2VD √ = 10 2 − 1.4

Vs =

= 12.74 V

(b) For v O,avg = 100 V

vS

π × 100.7 = 158.2 V 2 √ 120 2 Turns ratio = = 1.072 to 1 158.2 Vs =

10 2 V 1.4 V t

3.70 Refer to Fig. 3.23. θ = sin−1

1.4 √ = 5.68◦ = 0.1 rad 10 2

For 2VD Vs 2 2 Vs − 2VD = Vs − 1.4 π π

Fraction of cycle that D1 & D2 conduct is

VO,avg =

π − 2θ × 100 = 46.8% 2π

(a) For VO,avg = 10 V

Note that D3 & D4 conduct in the other half cycle so that there is 2 (46.8) = 93.6% conduction interval. v O,avg

2 = 2π

π −θ

√ (10 2sinφ − 2VD ) dφ

θ

Sedra, Smith, Carusone, Gaudet

2 · Vs − 1.4 π π ∴ Vˆ s = 11.4 = 17.9 V 2 √ 120 2 Turns ratio = = 9.5 to 1 17.9 10 V =

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–26

ωt ∼ =

(b) For VO,avg = 100 V

2 · Vs − 1.4 π π = 159 V ⇒ Vs = 101.4 2 √ 120 2 Turns ratio = = 1.07 to 1 159 100 V =

2 × 0.1 Vp − 0.7 Vp − 0.7

= =

√

2Vr / Vp − VD

0.2

= 0.447 rad ∴ Fraction of cycle for

3.71 R

12 : 1

conduction =

120 Vrms

10 Vrms VS

R 1 k

C vO

0.447 × 100 2π

= 7.1%

Vp − 0.7 (ii) ωt 2 × 0.01 = 0.141 rad Vp − 0.7 Fraction of cycle =

Vp Vr

VpVD

0.141 × 100 = 2.24% 2π

(c) (i) Use Eq (3.31): ⎛

iD,avg T (i) Vr ∼ = Vp − V D CR

[Eq. (3.28)]

T 0.1 Vp − VD = Vp − VD CR 1 = 166.7 µF C= 0.1 × 60 × 103 (ii) For

Vr = 0.01 Vp − VD

Vp − VD T = CR

⎞ 2 Vp − VD ⎠ = IL ⎝1 + π Vr

2 Vp − VD v O,avg = 1+π R 0.1 Vp − VD 2 12.77 = 1+π 0.1 103 = 192 mA (ii) iD,avg =

C = 1667 µF 1 (a) (i) v O, avg = Vp − VD − VΓ 2 √ √ 1 10 2 − 0.7 0.1 = 10 2 − 0.7 − 2 √ 0.1 = 10 2 − 0.7 1 − 2 = 12.77 V

√ 13.37 1 + π 200 3 10

= 607 mA (d) Adapting ⎛ Eq. (3.32), we obtain ⎞ 2 Vp − VD ⎠ (i) iD,peak = IL ⎝1 + 2π Vr 2 12.77 1 + 2π = 0.1 103

√ 0.01 (ii) v O, avg = 10 2 − 0.7 1 − 2

= 371 mA

= 13.37 V

(ii) iD,peak

(b) (i) Using Eq (3.30), we have the conduction angle =

= 1201 mA 1.2 A

Sedra, Smith, Carusone, Gaudet

2 13.37 = 1 + 2π 0.01 103

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–27 Vp − VD 3.72 (i) Vr = 0.1 Vp − VD = 2fCR The factor of 2 accounts for discharge occurring 1 only during half of the period, T/2 = . 2f 1 1 C= = 83.3 µF = (2fR) 0.1 2 (60) 103 × 0.1 1 (ii) C = = 833 µF 2 (60) × 103 × 0.01

1 = 833 µF 2 (0.01) fR

1 (a) VO = Vp − 2VD − Vr 2 1 Vp − 2VD × 0.1 2

= Vp − 2VD × 0.95 √ = (10 2 − 2 × 0.7) × 0.95 = 12.1 V √ (ii) VO = (10 2 − 2 × 0.7) × 0.995 = 12.68 V

0.1 = (13.44) 1 − 2

(b) (i) Fraction of cycle =

= 12.77 V

0.01 = 13.37 V (ii) VO = (13.44) 1 − 2 (b) (i) Fraction of cycle =

=

(ii) C =

(i) VO = Vp − 2VD −

1 (a) (i) VO = Vp − VD − Vr 2 0.1 = Vp − V D 1 − 2

Vp − 2VD 3.73 (i) Vr = 0.1 Vp − VD × 2 = 2fCR Vp − 2VD 1 = 83.3 µF C= Vp − 2VD 2 (0.1) fR

2Vr / Vp − VD π

2ωt × 100 2π

× 100

1 2 (0.1) × 100 = 14.2% π √ 2 2 (0.01) × 100 (ii) Fraction of cycle = 2π

2ωt × 100 2π

√

2 (0.1) × 100 = 14.2% π √ 2 (0.01) × 100 = 4.5% (ii) Fraction of cycle = π 1 12.1 (c) (i) iD, avg = 1+π = 97 mA 1 0.2 =

√ 12.68 1 + π/ 0.02 = 249 mA 1 1 12.1 ˆ 1 + 2π = 182 mA (d) (i) iD = 1 0.2

=

(ii) iD, avg =

= 4.5%

1 12.68 1 + 2π = 576 mA (ii) ˆiD = 1 0.02

(c) Use Eq. (3.34):

(i) iD, avg = IL 1 + π

=

12.77 1+π 1

(ii) iD, avg = = 310 mA

Vp − VD 2Vr

1 2 (0.1)

= 102.5 mA

1 13.37 1+π√ 1 2 (0.01)

(d) Use Eq. (3.35): 1 = 192 mA (i) ˆiD = IL 1 + 2π √ 2 (0.1) 1 ˆ = 607 mA (ii) iD = IL 1 + 2π √ 0.02

Sedra, Smith, Carusone, Gaudet

3.74 Refer to Fig. P3.74. When v I is positive, v A goes positive, turning on the diode and closing the negative feedback loop around the op amp. The result is that v − = v I , v O = 2v − = 2v I , and v A = v O + 0.7. Thus (a) v I = +1 V, v − = +1 V, v O = +2 V, and v A = +2.7 V. (b) v I = +3 V, v − = +3 V, v O = +6 V, and v A = +6.7 V. When v I goes negative, v A follows, the diode turns off, and the feedback loop is opened. The op amp saturates with v A = −13 V, v − = 0 V and v O = 0 V. Thus

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–28

vO

3.76

10 V

0.7 V

5 V 0

vS

vI

R 1 k vO

t

5 V

1 ms vS (V) 3 2.3

(c) v I = −1 V, v − = 0 V, v O = 0 V, and v A = −13 V.

0.7 0

(d) v I = −3 V, v − = 0 V, v O = 0 V, and v A = −13 V. Finally, if v I is a symmetrical square wave of 1-kHz frequency, 5-V amplitude, and zero average, the output will be zero during the negative half cycles of the input and will equal twice the input during the positive half cycles. See figure above. Thus, v O is a square wave with 0-V and +10-V levels, i.e. 5-V average and, of course, the same frequency (1 kHz) as the input.

t1

T 4

t2 T 2

T

t

3

First find t1 and t2 3 0.7 = T t1 4 ⇒ t1 = 0.0583 T

3.75

t2 =

D

vD vS

R

vO

iD = IS ev D /V T iD = e[v D −v D (at 1 mA)]/V T iD (1 mA)

iD v D − v D (at 1 mA) = VT ln 1 mA

v O /R v D = v D (at 1 mA) + VT ln 1

=

T − t1 2

T − 0.0583 T 2

t2 = 0.442 T 1 × area of shaded triangle T 1 T = × (3 − 0.7) × − t1 T 4

v O (ave.) =

=

1 × 2.3 × T T

1 − 0.0583 4

= 0.441 V √ √ 3.77 120 2 ± 10%: 20 2 ± 10%

v O = v S − v D (at 1 mA) − VT ln R

⇒ Turns ratio = 6:1 √ 20 2 ± 10% vS = 2

where R is in k.

PIV= 2Vs − VD

vO = vS − vD

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–29 √ 20 2 =2× × 1.1 − 0.7 2

= 2VS − 0.7 = 39.2 V

= 30.4 V Using a factor of 1.5 for safety, we select a diode having a PIV rating of approximately 45 V.

If choosing a diode, allow a safety margin by moving a factor of 1.5, thus PIV 60 V

3.78 The circuit is a full-wave rectifier with center tapped secondary winding. The circuit can − be analyzed by looking at v + O and v O separately.

3.79

0.7 V vS vS

D1

D3

120 Vrms 60 Hz

vS

R 200

C vO

vO

vS vS

D4

R

VO = 12 V ± 1 V (ripple) RL = 200

D2

(a) VO = Vp − VD − 1 ⇒ Vp = 13 + 0.7 = 13.7 V 13.7 Vrms = √ = 9.7 V 2 (b) Vr = 2=

13.7 − 0.7 60 × C × 200

⇒C= v O, avg =

1 2π

Vp − VD fCR

13 = 542 µF 2 × 60 × 200

(VS sinφ − 0.7) dφ = 12

Vp

2Vs − 0.7 = 12 = π

VpVD

where we have assumed Vs 0.7 V and thus the conduction angle (in each half cycle) is almost π . 12 + 0.7 π = 19.95 V Vs = 2

Vp

t

PIV

= 2VS 40 V

(c) When the diode is cut off, the maximum reverse voltage across it will occur when v S = −Vp . At this time, v O = VO and the maximum reverse voltage will be

Looking at D4 ,

Maximum reverse voltage = VO + Vp

PIV= VS − VO−

= 12 + 13.7 = 25.7 V

= VS + (VS − 0.7)

Using a factor of safety of 1.5 we obtain

Thus voltage across secondary winding

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–30

PIV = 1.5 × 25.7

(b) Vr =

= 38.5 V

(d) iDav = IL 1 + π

2(Vp − VD ) Vr

2=

13.7 − 0.7 2 × 60 × 200 × C

⇒C=

VO 2(Vp − VD ) = 1+π RL Vr

Vp − V D 2fCR

12 = 271 µF 2 × 2 × 60 × 200

(c) Maximum reverse voltage across D1 occurs when v S = −Vp . At this point v O = VO . Thus maximum reverse voltage = VO + Vp = 12 + 13.7 = 25.7. The same applies to D2 .

12 2(13.7 − 0.7) = 1+π 0.2 2

In specifying the PIV for the diodes, one usually uses a factor of safety of about 1.5,

= 739 mA

(e) iDmax = IL 1 + 2π

2(Vp − VD ) Vr

PIV = 1.5 × 25.7 = 38.5 V Vp − VD (d) iDav = IL 1 + π 2 Vr

12 2(13.7 − 0.7) = 1 + 2π 0.2 2

12 13.7 − 0.7 = 1+π 0.2 2×2

= 1.42 A

= 399 mA

3.80

(e) iDmax = IL 1 + 2π

D1 Vs 0.7 V Vs 0.7 V

120 Vrms 60 Hz

C

R

vO

D2

=

Vp − VD 2 Vr

12 13.7 − 0.7 1 + 2π 0.2 2×2

= 739 mA

3.81

(a) VO = Vp − VD − 1

D4

D1

C R

⇒ Vp = VO + VD + 1 = 13 + 0.7 = 13.7 V

120 Vrms 60 Hz

13.7 Vrms = √ = 9.7 V 2

This voltage appears across each half of the transformer secondary. Across the entire secondary we have 2 × 9.7 = 19.4 V (rms).

(a) VO = Vp − 2VD − 1

vS

vO D2

D3

⇒ Vp = VO +2VD +1 = 12+2×0.7+1 = 14.4 V 14.4 Vrms = √ = 10.2 V 2 (b) Vr =

PIV

Sedra, Smith, Carusone, Gaudet

⇒C=

Vp − 2 V D 2fCR

14.4 − 1.4 = 271 µF 2 × 2 × 60 × 200

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–31

(c) The maximum reverse voltage across D1 occurs when Vs = −Vp = −14.4 V. At this time D3 is conducting, thus

vI

C 100 F

Maximum reverse voltage = −Vp + VD3 = −14.4 + 0.7 = −13.7 V The same applies to the other three diodes. In specifying the PIV rating for the diode we use a factor of safety of 1.5 to obtain

R 100

vO

(a) vI 12 V 11.3 V

Vr

vO

vI

PIV = 1.5 × 13.7 = 20.5 V Vp − 2 VD (d) iDav = IL 1 + π 2 Vr

t

12 14.4 − 1.4 = 1+π 0.2 2×2

T

1 ms

12 V

(b)

= 400 mA (e) iDmax = IL 1 + 2π

Vp − 2 VD 2 Vr

12 V 11.3 11.3 Vr 10.2 V

12 14.4 − 0.7 = 1 + 2π 0.2 2×2

} T

0

= 740 mA

Vr 1.13 V

4

t (c)

3.82

v O = 11.3 −

During the diode’s off interval (which is almost equal to T) the capacitor discharges and its voltage is given by v O (t) = 11.3 e−t/CR where C = 100 µF and R = 100 , thus

1.13 Vr = 11.3 − = 10.74 V 2 2

To obtain the interval during which the diode conducts, t, refer to Fig. (c). Vr 12 = T/4 t 1.13 × 1 Vr × (T/4) = 12 12 × 4

CR = 100 × 10−6 × 100 = 0.01 s

⇒ t =

At the end of the discharge interval, t T and

= 23.5 µs

v O = 11.3 e−T/CR

Now, using the fact that the charge gained by the capacitor when the diode is conducting is equal to the charge lost by the capacitor during its discharge interval, we can write

Since T = 0.001 s is much smaller than CR, T v O 11.3 1 − CR The ripple voltage Vr can be found as T Vr = 11.3 − 11.3 1 − CR =

11.3T 11.3 × 0.001 = = 1.13 V CR 0.01

The average dc output voltage is

Sedra, Smith, Carusone, Gaudet

iCav × t = C Vr ⇒ iCav =

100 × 10−6 × 1.13 C Vr = = 4.8 A t 23.5 × 10−6

iDav = iCav + iLav where iLav is the average current through R during the short interval t. This is approximately 11.3 11.3 = = 0.113 A. Thus R 100

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–32

iDav = 4.8 + 0.113 = 4.913 A

⇒ C = 868 µF

Finally, to obtain the peak diode current, we use

To specify the diodes, we determine iDav and iDmax , iDav = IL (1 + π (Vp − 0.7)/2 Vr )

iDmax = iCmax + iLmax =C

11.3 dv I + dt R

=C×

= 0.1(1 + π 12.5/2 )

12 11.3 + T/4 R

= 785 mA

12 × 4 11.3 + = 100 × 10−6 × 1 × 10−3 100 = 4.8 + 0.113 = 4.913 A

iDmax = IL (1 + 2π (Vp − 0.7)/2 Vr ) = 0.1(1 + 2π 12.5/2 ) = 1.671 A

which is equal to the average value. This is a result of the linear v I which gives rise to a constant capacitor current during the diode conduction interval. Thus iCmax = iCav = 4.8 A. Also, the maximum value of iL is approximately equal to its average value during the short interval t.

3.83 Refer to Fig. P3.78 and let a capacitor C be connected across each of the load resistors R. The − two supplies v + O and v O are identical. Each is a full-wave rectifier similar to that of the tapped-transformer circuit. For each supply, VO = 12 V

To determine the required PIV rating of each diode, we determine the maximum reverse voltage that appears across one of the diodes, say D1 . This occurs when v S is at its maximum negative value −Vp . Since the cathode of D1 will be at +12.5 V, the maximum reverse voltage across D1 will be 12.5 + 13.2 = 25.7 V. Using a factor of safety of 1.5, then each of the four diodes must have PIV = 1.5 × 25.7 = 38.6 V 3.84 v I > 0: D1 conducts and D2 cutoff v I < 0: D1 cutoff, D2 conducts ∼

Vr = 1 V (peak to peak)

vO = −1 vI vO

Thus v O = 12 ± 0.5 V

slope 1

It follows that the peak value of v S must be 12.5 + 0.7 = 13.2 V and the total rms voltage across the secondary will be 2 × 13.2 = 18.7 V (rms) = √ 2 Transformer turns ratio =

120 = 6.43:1 18.7

Vp − 0.7 2fCR

12.5 1= 2 × 60 × C × 120

Sedra, Smith, Carusone, Gaudet

vO = 0 V

Keeps D2 off so no current flows through R ⇒ v− = 0 V Virtual ground as feedback loop is closed through D1

12 = 120 0.1

Now, the value of C can be found from Vr =

(a) v I = +1 V

v A = −0.7 V

To deliver 100-mA dc current to each load, R=

vI

(b) v I = +3 V vO = 0 V v A = −0.7 V v− = 0 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–33

(c) v I = −1 V

3.86

v O = +1 V v A = 1.7 V v− = 0 V ∼ Virtual ground as negative feedback loop is closed through D2 and R. (d) v I = −3 V ⇒ v O = +3 V v A = +3.7 V

From the figure we see that √ v Oav = −5 2 = −7.07 V

v− = 0 V 3.85 The input is an AM signal with carrier frequency ω0 = 2π × 106 rad/s and time-varying amplitude Vp (t) = 1.5 + sin(2π × 103 t) Volts. With CR = 0.1 ms, the output time constant is much longer than the period of the carrier, 1/106 = 1s. Thus, the output drops very little between successive peaks of the carrier. However, the time constant is much smaller than the period of Vp (t) so that vO will follow changes in Vp (t). A sketch appears in Fig. 1.

3.87

3 V ID R VD

vO Vp(t)

2.5 V

(a) VD = 2.2 V, ID = 10 mA.

0.5 V t

1 ms Figure 1

⇒R=

3 − 2.2 = 80 10

(b) VD = 1.8 V, ID = 10 mA. With CR = 0.1 s, the output time constant is much smaller than the period of the carrier (1 s). Thus, the circuit behaves as a precision rectifier. A sketch appears in Fig. 2.

⇒R=

3 − 1.8 = 120 10

3.88

vO vO

Vp(t)

10 V V1

...

10 V

t

1 s 2 s Figure 2

Sedra, Smith, Carusone, Gaudet

0V

t

V2 5 s

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 3–34

V1 = 5 + 5e−5/10 = 8.03 V

V1 0

V2 = V1 − 10 = −1.97 V

V´1 V2 T1

3.89

t

V´2 T2

where α 1

(a) During the interval T2 , we have

10 V 10 V

|v O | = |V2 | e−t/(CR/2)

At the end of T2 , t = T, and v O = |V2 | where

(b)

|V2 | = |V2 | e−T/(CR/2)

20 V

|V2 | 1 −

0V

Now

T RC/2

= |V2 | (1 − 2α)

V1 + |V2 | = 20 ⇒ V1 + |V2 | − αV 1 = 20

(c)

(1)

0V

and V2 + V1 = 20 ⇒ V1 + |V2 | − 2α |V2 | = 20 (2)

20 V

From (1) and (2) we find that V1 = 2|V2 |

(d) Then using (1) and neglecting αV1 yields

0V 20 V

3 |V2 | = 20 ⇒ |V2 | = 6.67 V V1 = 13.33 V The result is

(e)

13.33 V 10 V 6.67 V 10 V (g)

18 V

(f) Here there are two different time constants involved. To calculate the output levels, we shall consider the discharge and charge wave forms. During T1 ,

2 V

v O = V1 e−t/RC

At t = T1 = T, v O = V1

(h) Using a method similar to that employed for case (f) above, we obtain

−T/RC

= V1 e

13.33 V

where for T CR

V1 V1 (1 − T/CR) = V1 (1 − α)

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

6.67 V

© Oxford University Press 2021

Chapter 3–35

3.90 Cj0 = 150 fF, V0 = 3 V, m = 3. Using Equation (1.71), 150 fF Cj = 3 1 + V3R ⇒ VR = 3

150 Cj

3.91 (a) Assuming an ideal op amp, there is a virtual short-circuit between input terminals so that the photodiode reverse voltage is, VR = v− = v+ = 3 V

1/3

Thus,

−3

at Cj = 80 fF: VR = 3

150 80

vO = 3 − iD × 250 = 3 + (ID + iP ) × 250 = 3 + (ID + R × P) × 250

1/3 − 3 = 0.70 V,

and

150 at Cj = 20 fF: VR = 3 20

vO = 3 + ID × 250 = 3 + 10−7 × 250 ∼ =3V

1/3 − 3 = 2.87 V.

Thus, the range of required bias voltages is 0.70 V < VR < 2.87 V.

Sedra, Smith, Carusone, Gaudet

(b) With no illumination, P = 0 and

(c) With P = 10 W, vO = 3 + (10−7 + 10−2 × 0.5) × 250 = 4.25 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 4–1

Chapter 4

IS 10−16 = = 10−18 A β 100 IC 1 mA = VT ln = 25 ln IS 10−16 A

ISB =

Solutions to Exercises within the Chapter Ex: 4.1 iC = IS ev BE /VT iC2 v BE2 − v BE1 = VT ln iC1 0.1 v BE2 = 700 + 25 ln 1

VBE

= 25 × 29.9336 = 748 mV

Ex: 4.6

= 642 mV = 0.64 V 10 v BE3 = 700 + 25 ln 1

VCC 5 V

= 758 mV = 0.76 V Ex: 4.2 ∴ α =

10 A

β β +1

C

B

150 50 VB , the transistor is operating in the active mode, as assumed.

VC RC 1 k

IC Ex: 4.14

10 V 5 V IE

5 k VE

IB

The transistor is operating at a constant emitter current. Thus, a change in temperature of +30◦ C results in a change in VEB by VEB = −2 mV × 30 = −60 mV

VB IB VC

100 k IC

5 k

Thus, VE = −60 mV Since the collector current remains unchanged at αIE , the collector voltage does not change:

10 V VB = 1.0 V

Ex: 4.16 Refer to Fig. 4.19(a): v CE iC = IS ev BE /VT + ro

Thus, VB = 0.01 mA IB = 100 k

ro =

Thus, 10 − 1.7 10 − VE = = 1.66 mA 5 k 5

and β +1=

IE 1.66 = 166 = IB 0.01

VA IS ev BE /VT

Substituting in Eq. (1), we have v CE v BE /VT 1+ iC = I S e VA which is Eq. (4.18).

⇒ β = 165 α=

165 β = = 0.994 β +1 165 + 1

Ex: 4.17 ro =

Q.E.D.

VA 100 = IC IC

At IC = 0.1 mA, ro = 1 M

Assuming active-mode operation,

At IC = 1 mA, ro = 100 k

IC = αIE = 0.994 × 1.66 = 1.65 mA

At IC = 10 mA, ro = 10 k

Sedra, Smith, Carusone, Gaudet

(1)

Now using Eqs. (4.21) and (4.22), we can express ro as

VE = +1.7 V

IE =

VC = 0 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 4–4

Ex: 4.18 IC =

VCE ro

Ex: 4.20 For VBB = 0 V, IB = 0 and the transistor is cut off. Thus,

where VA 100 = 100 k ro = = IC 1

IC = 0

11 − 1 = 0.1 mA 100 Thus, IC becomes 1.1 mA.

VC = VCC = +10 V

and

IC =

Ex: 4.21 Refer to the circuit in Fig. 4.22 and let VBB = 1.7 V. The current IB can be found from IB =

Ex: 4.19

VCC 10 V VBB IB

VBB − VB 1.7 − 0.7 = 0.1 mA = RB 10

Assuming operation in the active mode, IC = β IB = 50 × 0.1 = 5 mA

IC

RC 10 k

RB 10 k

VBE 0.7 V

VCE

(a) For operation in the active mode with VCE = 5 V,

Thus, VC = VCC − RC IC = 10 − 1 × 5 = 5 V which is greater than VB , verifying that the transistor is operating in the active mode, as assumed. (a) To obtain operation at the edge of saturation, RC must be increased to the value that results in VCE = 0.3 V:

IC =

VCC − VC 10 − 5 = 0.5 mA = RC 10

RC =

IB =

0.5 IC = = 0.01 mA β 50

=

VBB = VBE + IB RB

VCC − 0.3 IC

10 − 0.3 = 1.94 k 5

(b) For operation at the edge of saturation,

(b) Further increasing RC results in the transistor operating in saturation. To obtain saturation-mode operation with VCE = 0.2 V and βforced = 10, we use

VCE = 0.3 V

IC = βforced × IB

= 0.7 + 0.01 × 10 = 0.8 V

VCC − VCE 10 − 0.3 = = 0.97 mA IC = RC 10 IB =

The value of RC required can be found from

0.97 IC = = 0.0194 mA β 50

RC =

VBB = VB + IB RB = 0.7 + 0.0194 × 10 = 0.894 V (c) For operation deep in saturation with βforced = 10, we have VCE 0.2 V 10 − 0.2 = 0.98 mA 10 IC 0.98 = = 0.098 mA IB = βforced 10

IC =

VBB = VB + IB RB = 0.7 + 0.098 × 10 = 1.68 V

Sedra, Smith, Carusone, Gaudet

= 10 × 0.1 = 1 mA

=

VCC − VCE IC

10 − 0.2 = 9.8 k 1

Ex: 4.22 Refer to the circuit in Fig. 4.23(a) with the base voltage raised from 4 V to VB . If at this value of VB , the transistor is at the edge of saturation then, VC = VB − 0.4 V Since IC IE , we can write VE VB − 0.7 10 − VC = = RC RE RE

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 4–5 10 − (VB − 0.5) 4.7 VB − 0.7 IE = 6IB = 3.3 Dividing Eq. (1) by Eq. (2), we have

Thus,

IC = 5IB =

10 − (VB − 0.4) VB − 0.7 = 4.7 3.3 ⇒ VB = +4.7 V

(1) (2)

10.5 − VB 3.3 5 = × 6 VB − 0.7 4.7

Ex: 4.23

⇒ VB = +5.18 V

10 V 0.5 mA RC VC 6 V

Ex: 4.25 Refer to the circuit in Fig. 4.26(a). The largest value for RC while the BJT remains in the active mode corresponds to VC = +0.4 V

VB 4 V VE 3.3 V RE

0.5 mA

Since the emitter and collector currents remain unchanged, then from Fig. 4.26(b) we obtain IC = 4.65 mA Thus, RC =

To establish a reverse-bias voltage of 2 V across the CBJ, VC = +6 V

=

VC − (−10) IC

+0.4 + 10 = 2.24 k 4.65

Ex: 4.26

From the figure we see that

10 V

10 − 6 = 8 k RC = 0.5 and 3.3 = 6.6 k RE = 0.5 where we have assumed α 1.

1 mA

RE VE 0.7 V

VC 4 V 1 mA

Ex: 4.24

10 V

10 V

IC 5IB 4.7 k VB 0.5 VB

RC

For a 4-V reverse-biased voltage across the CBJ, VC = −4 V Refer to the figure.

IB VB 0.7 6IB

IC = 1 mA =

3.3 k

VC − (−10) RC

−4 + 10 = 6 k 1 10 − VE RE = IE

⇒ RC =

The figure shows the circuit with the base voltage at VB and the BJT operating in saturation with VCE = 0.2 V and βforced = 5.

Sedra, Smith, Carusone, Gaudet

Assuming α = 1, RE =

10 − 0.7 = 9.3 k 1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 4–6

Ex: 4.27 Refer to the circuit in Fig. 4.27: IB =

Ex: 4.30 15 V

5 − 0.7 = 0.043 mA 100

To ensure that the transistor remains in the active mode for β in the range 50 to 150, we need to select RC so that for the highest collector current possible, the BJT reaches the edge of saturation, that is, VCE = 0.3 V. Thus,

2.78 mA

2 k 9.44 V Q2

IC3 2.75 mA Q3

VC2

VCE = 0.3 = 10 − RC ICmax

IC3 /b

where

VE3 2.7 k

ICmax = βmax IB

470 IC3 a

= 150 × 0.043 = 6.45 mA Thus, RC =

10 − 0.3 = 1.5 k 6.45

From the figure we see that IC3 × 0.47 α

For the lowest β,

VE3 =

IC = βmin IB

IC3 × 0.47 + 0.7 (1) α A node equation at the collector of Q2 yields VC2 = VE3 + 0.7 =

= 50 × 0.043 = 2.15 mA and the corresponding VCE is

2.75 =

VCE = 10 − RC IC = 10 − (1.5 × 2.15) = 6.775 V Thus, VCE will range from 0.3 V to 6.8 V.

IC3 VC2 + 2.7 β

Substituting for VC2 from Eq. (1), we obtain 2.75 =

(0.47 IC3 /α) + 0.7 IC3 + 2.7 β

Substituting α = 0.99 and β = 100 and solving for IC3 results in Ex: 4.28 Refer to the solution of Example 4.10.

IC3 = 13.4 mA

VBB − VBE IE = RE + [RBB /(β + 1)]

Now, VE3 and VC2 can be determined: 13.4 IC3 × 0.47 = × 0.47 = +6.36 V α 0.99 = VE3 + 0.7 = +7.06 V

VE3 =

5 − 0.7 = 1.177 mA = 3 + (33.3/51)

VC2

IC = αIE = 0.98 × 1.177 = 1.15 mA Thus the current is reduced by

Ex: 4.31 5 V

IC = 1.28 − 1.15 = 0.13 mA which is a −10% change.

Q1 off

Ex: 4.29 Refer to the circuit in Fig. 4.30(b). The total current drawn from the power supply is I = 0.103 + 1.252 + 2.78 = 4.135 mA

10 k

0

5 V

IE IB

Q2 on IE

VE IE 1 1 k

Thus, the power dissipated in the circuit is P = 15 V × 4.135 mA = 62 mW

Sedra, Smith, Carusone, Gaudet

5 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 4–7

From the figure we see that Q1 will be off and Q2 will be on. Since the base of Q2 will be at a voltage higher than −5 V, transistor Q2 will be operating in the active mode. We can write a loop equation for the loop containing the 10-k resistor, the EBJ of Q2 and the 1-k resistor: −IE × 1 − 0.7 − IB × 10 = −5 Substituting IB = IE /(β + 1) = IE /101 and rearranging gives IE =

5 − 0.7 = 3.9 mA 10 +1 101

Q1 will be saturated. Assuming this to be the case, the analysis steps will be as follows: VCEsat |Q1 = 0.2 V VE = 5 V − VCEsat = +4.8 V 4.8 V = 4.8 mA 1 k = VE + VBE1 = 4.8 + 0.7 = +5.5 V

IE1 = VB1

10 − 5.5 = 0.45 mA 10 = IE1 − IB1 = 4.8 − 0.45 = 4.35 mA

IB1 = IC1

βforced =

Thus,

IC 4.35 = 9.7 = IB 0.45

VE = −3.9 V

which is lower than βmin , verifying that Q1 is indeed saturated.

VB2 = −4.6 V

Finally, since Q2 is off, IC2 = 0

IB = 0.039 mA Ex: 4.32 With the input at + 10 V, there is a strong possibility that the conducting transistor

Ex: 4.33 VO = +10 − BVBCO = 10 − 70 = −60 V

This figure belongs to Exercise 4.32.

5 V

4 5.5 V

Q1 on VCEsat 0.2 V

1

10 k

4.8 V

10 V 5 10 5.5 0.45 mA 10

3 Q2 off

2

4.8 mA 1 k

5 V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–1

Solutions to End-of-Chapter Problems 4.1 1. 2. 3. 4. 5. 6.

For iB = 20 µA, iC = iE − iB = 1000 − 20 = 980 µA

Active Saturation Active Saturation Active Cutoff

β=

980 iC = 49 = iB 20

α=

49 β = = 0.98 β +1 50

For iB = 50 µA,

4.2 The EB junctions have a 4:1 area ratio.

iC = iE − iB = 1000 − 50 = 950 µA

VBE /VT

IC = IS e

−3

0.5 × 10

= IS1 × e

0.75/0.025

⇒ IS1 = 4.7 × 10−17 A

β=

950 iC = 19 = iB 50

α=

β 19 = = 0.95 β +1 20

IS2 = 4IS1 = 1.87 × 10−16 A 4.7

4.3 Old technology: 10−3 = 2 × 10−15 eVBE /VT 10−3 = 0.673 V VBE = 0.025 ln 2 × 10−15

5 V 2 k

New technology:

aiE B

4.4 IB = 10 µA

DE iE

IC = 800 µA β=

VC

C

10−3 = 2 × 10−18 eVBE /VT 10−3 = 0.846 V VBE = 0.025 ln 2 × 10−18

VE

E

IC = 80 IB

2 mA

80 β = = 0.988 α= β +1 81 4.5 Transistor is operating in active region:

The figure shows the circuit, where

β = 50 → 300

α=

IB = 10 µA

β 100 = = 0.99 β +1 101

IE = (β + 1)IB = 0.51 mA → 3.01 mA

5 × 10−15 IS = = 5.05 × 100−15 A α 0.99 The voltage at the emitter VE is

Maximum power dissipated in transistor is

VE = −VDE

IB × 0.7 V + IC × VC = 0.01 × 0.7 + 3 × 10 30 mW

= −VT ln(IE /ISE ) 2 × 10−3 = −0.025 ln 5.05 × 10−15

4.6 For iB = 10 µA,

= −0.668 V

iC = iE − iB = 1000 − 10 = 990 µA

The voltage at the collector VC is found from

990 iC = β= = 99 iB 10

VC = 5 − IC × 2

ISE =

IC = βIB = 0.5 mA → 3 mA

α=

99 β = = 0.99 β +1 100

Sedra, Smith, Carusone, Gaudet

= 5 − αIE × 2 = 5 − 0.99 × 2 × 2 = 1.04 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–2

4.8. v CE (V) v BC (V) iBC (µA) iBE (µA) iB (µA) iC (mA)

iC /iB

0.4

0.3

0.016

14.46

14.48

1.446

100

0.3

0.4

0.89

14.46

15.35

1.445

94

0.2

0.5

14.46

62.96

1.398

22.2

48.5

The equations utilized are v BC = v BE − v CE = 0.7 − v CE

(1)

iBC = ISC ev BC /VT = 10−13 ev BC /0.025

(2)

v BE /VT

iBE = ISB e

−17 0.7/0.025

= 10

e

(3)

iB = iBC + iBE

(4)

iC = IS ev BE /VT − iBC = 10−15 e0.7/0.025 − iBC

(5)

IS 10−14 = = 2 × 10−16 A β 50 10 × 10−6 = 0.025 ln 2 × 10−16

ISB = VEB

= 0.616 V Thus, VB = −VEB = −0.616 V

Performing these calculations for v CE = 0.4 V, 0.3 V, and 0.2 V, we obtain the results shown in the table above.

The collector current can be found as

To obtain vCE that results in iC = 0, we first use Eq. (5),

= 50 × 10 = 500 µA = 0.5 mA

0 = 10−15 e0.7/0.025 − iBC ⇒ iBC = 1.446 mA Substituting in Eq. (2), we obtain 1.446 × 10−3 = 10−13 evBC /0.025 ⇒ vBC = 0.585 V Finally, substituting in Eq. (1) gives 0.0585 = 0.7 − vCE ⇒ vCE = 0.115 V

IC = βIB

The collector voltage can now be obtained from VC = −5 + IC × 8.2 = −5 + 0.5 × 8.2 = −0.9 V The emitter current can be found as IE = IB + IC = 10 + 500 = 510 µA = 0.51 mA

4.10

4.9

IB =

5 IE = = 0.238 A = 238 mA β +1 20 + 1

IC = IS eVEB /VT αIE = IS eVEB /VT where α=

20 = 0.95 21

IS = αIE e−VEB /VT = 0.95 × 5e−(0.8/0.025) = 6 × 10−14 A

The emitter–base voltage VEB is found as the voltage drop across the diode DB , whose scale current is ISB = IS /β, it is conducting a 10-µA current. Thus, 10 µA VEB = VT ln ISB where

Sedra, Smith, Carusone, Gaudet

A transistor that conducts IC = 1 mA with VEB = 0.70 V has a scale current IS = 1 × 10−3 e−0.70/0.025 = 6.9 × 10−16 A The emitter–base junction areas of these two transistors will have the same ratio as that of their scale currents, thus 6 × 10−14 EBJ area of first transistor = = 87 EBJ area of second transistor 6.9 × 10−16

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–3

4.11 IC = IS eVBE /VT

VBE = 0.76 + 0.025 ln(10−3 )

200 × 10−6 = IS e30

= 0.587 V

⇒ IS = 1.87 × 10

−17

A

For the transistor that is 32 times larger,

4.15

IS = 32 × 1.87 × 10−17

α

= 6 × 10−16 A At VBE = 30 VT , the larger transistor conducts a current of IC = 32 × 200 µA = 6.4 mA

0.5 0.8 0.9 0.95 0.98 0.99 0.995 0.999

α β= 1−α

4

1

2

9

19

49

99

199

β

10

20

50

100

200

4.17 β = IS1 AE1 200 × 200 = 250,000 = = IS2 AE2 0.4 × 0.4

α 1−α

(1)

β → β + β α + α 1 − α − α

IC2 = IS2 eVBE2 /VT

β + β =

For IC1 = IC2 we have

Subtracting Eq. (1) from Eq. (2) gives

IS1 = = 250,000 IS2

VBE2 − VBE1 = 0.025 ln(250,000) = 0.31 V

β =

α + α α − 1 − α − α 1−α

β =

α (1 − α − α)(1 − α)

(2)

(3)

Dividing Eq. (3) by Eq. (1) gives 1 β α = β α 1 − α − α

4.13 IC1 = 10−13 e700/25 = 0.145A = 145 mA IC2 = 10−18 e700/25 = 1.45 µA

For α 1, the second factor on the right-hand side is approximately equal to β. Thus α β β Q.E.D. β α

For the first transistor 1 to conduct a current of 1.45 µA, its VBE must be 1.45 × 10−6 VBE1 = 0.025 ln 10−13

For

= 0.412 V 4.14 5 × 10−3 = IS e0.76/0.025

(1)

IC = IS e0.70/0.025

(2)

Dividing Eq. (2) by Eq. (1) yields

β = −10% and β = 100, β

α −10% = −0.1% α 100

4.18 iC = IS ev BE /VT

−3 −0.06/0.025

IC = 5 × 10 e

= 5 × 10−15 e0.7/0.025 = 7.2 mA

= 0.45 mA For IC = 5 µA,

iB will be in the range

5 × 10−6 = IS eVBE /0.025

(3)

Dividing Eq. (3) by Eq. (1) yields

Sedra, Smith, Carusone, Gaudet

1000

α → α + α

IC1 = IS1 eVBE1 /VT

10−3 = e(VBE −0.76)/0.025

500

β α= 0.5 0.67 0.91 0.95 0.98 0.99 0.995 0.998 0.999 β +1

1 × 10−3 = 6 × 10−16 eVBE /VT 1 × 10−3 = 0.704 V VBE = VT ln 6 × 10−16

e(VBE2 −VBE1) /VT

999

4.16

At IC = 1 mA, the base–emitter voltage of the larger transistor can be found as

4.12

1

is, 144 µA to 36 µA.

7.2 7.2 mA to mA, that 50 200

iE will be in the range (7.2 + 0.144) mA to (7.2 + 0.036) mA, that is, 7.344 mA to 7.236 mA.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–4

4.19 Transistor

a

b

c

d

e

VBE (mV)

700

690

580

780

820

IC (mA)

1.000

1.000

0.230

10.10

73.95

IB (µA)

10

20

5

120

1050

IE (mA)

1.010

1.020

0.235

10.22

75

α

0.99

0.98

0.979

0.988

0.986

β

100

50

46

84

70

IS (A)

6.9 × 10−16

1.0 × 10−15 1.9 × 10−14 2.8 × 10−16 4.2 × 10−16

C

(b)

4.20 First we determine IS , β, and α: 1 × 10−3 = IS e700/25

iC

⇒ IS = 6.91 × 10−16 A

aiE

1 mA IC = 100 = β= IB 10 µA α=

iB B

100 β = = 0.99 β +1 101

v BE

Then we can determine ISE and ISB : ISE = ISB

DE ISE IS /a iE

IS = 6.98 × 10−16 A α

E a 0.99 ISE 6.98 1016 A

IS = 6.91 × 10−18 A = β

The figure below shows the four large-signal models, corresponding to Fig. 4.5(a) to (d), together with their parameter values.

(d) B

iB vBE

iC C DB IS ISB b

biB

iE E b 100 ISB 6.91 1018 A

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–5

4.21 Refer to the circuit in Fig. 4.6(b). ISB =

−15

IS 5 × 10 = β 50

= 10−16 A

IC 0.5 × 10−3 = = 10−5 A β 50 IB VB = VBE = VT ln ISB −5 10 = 0.025 ln 10−16 IB =

= 0.633 V We can determine RB from RB =

VCC − VB IB

5 − 0.633 = 437 k 10−5 To obtain VCE = 1 V, we select RC according to

=

VCC − VCE RC = IC =

Dividing the numerator and denominator of the right-hand side by ISC ev BC /VT and replacing v BE − v BC by VCEsat gives IS eVCEsat /VT − 1 ISC βforced = 1 IS eVCEsat /VT + 1 β ISC This equation can be used to obtain eVCEsat /VT and hence VCEsat as IS 1 + βforced eVCEsat /VT = ISC 1 − βforced /β ISC 1 + βforced Q.E.D. ⇒ VCEsat = VT ln IS 1 − βforced /β For β = 100 and ISC /IS = 100, we can use this equation to obtain VCEsat corresponding to the given values of βforced . The results are as follows: βforced

50

10

5

1

0

VCEsat (V) 0.231 0.178 0.161 0.133 0.115

5−1 = 8 k 0.5

4.24 At iC = 1 mA, v EB = 0.7 V At iC = 10 mA,

10 v EB = 0.7 + VT ln 1

4.22 IS = 10−15 A Thus, a forward-biased EBJ conducting a current of 1 mA will have a forward voltage drop VBE : I VBE = VT ln IS −3 10 = 0.691 V = 0.025 ln 10−15

= 0.7 + 0.025 ln(10) = 0.758 V At iC = 100 mA, v EB = 0.7 + 0.025 ln

100 1

= 0.815 V

ISC = 100IS = 10−13 A

Note that v EB increases by about 60 mV for every decade increase in iC .

Thus, a forward-biased CBJ conducting a 1-mA current will have a forward voltage drop VBC : 1 × 10−3 = 0.576 V VBC = VT ln 1 × 10−13

4.25

When forward-biased with 0.5 V, the emitter–base junction conducts I = IS e0.5/0.025 = 10−15 e0.5/0.025 = 0.49 µA and the CBJ conducts I = ISC e0.5/0.025 = 10−13 e0.5/0.025 = 48.5 µA 4.23 Dividing Eq. (4.14) by Eq. (4.15) and substituting iC /iB = βforced gives βforced =

v BE /VT

v BC /VT

− ISC e IS e (IS /β)ev BE /VT + ISC ev BC /VT

Sedra, Smith, Carusone, Gaudet

Referring to the figure, we see that IE = IB + IC =

IC + IC β

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–6

Thus, IC =

4.27 IE 1+

1 β

=

1 1+

1 10

= 0.909 mA

IB = 0.091 mA For direction of flow, refer to the figure. IB VEB = VT ln ISB where

4.28 (a) Refer to Fig. P4.28(a).

10−15 IS = = 10−16 A ISB = β 10 0.091 × 10−3 VEB = 0.025 ln 10−16

10.7 − 0.7 = 2 mA 5 k Assuming operation in the active mode,

I1 =

IC = αI1 I1 = 2 mA

= 0.688 V

V2 = −10.7 + IC × 5

Thus, VE = VB + VEB = 0 + 0.688 = 0.688 V If a transistor with β = 1000 is substituted, IC =

B

IE 1 1+ β

=

1 1+

1 1000

= 0.999 mA

Thus, IC changes by 0.999 − 0.909 = 0.09 mA, a 9.9% increase.

= −10.7 + 2 × 5 = −0.7 V Since V2 is lower than VB , which is 0 V, the transistor is operating in the active mode, as assumed. (b) Refer to Fig. P4.28(b). Since VC = −4 V is lower than VB = −2.7 V, the transistor is operating in the active mode. −4 − (−10) = 2.5 mA 2.4 k IC IC = 2.5 mA IE = α V3 = +12 − IE × 5.6 = 12 − 2.5 × 5.6 = −2 V

IC = 4.26 The two missing large-signal equivalent circuits for the pnp transistor are those corresponding to the npn equivalent circuits in Fig. 4.5(b) and 4.5(d). They are shown in the figure.

iE

v EB

DE (IS /a)

0 − (−10) = 0.5 mA 20 Assuming active-mode operation, and utilizing the fact that β is large, IB 0 and

IC =

E

iB

(c) Refer to Fig. P4.28(c) and use

V4 2 V Since VC < VB , the transistor is indeed operating in the active region.

B

aiE iC C

IC IC = 0.5 mA α (d) Refer to Fig. P4.28(d). Since the collector is connected to the base with a 10-k resistor and β is assumed to be very high, the voltage drop across the 10-k resistor will be close to zero and the base voltage will be equal to that of the collector: I5 = I E =

VB = V7 This also implies active-mode operation. Now, VE = VB − 0.7 Thus,

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–7

VE = V7 − 0.7 VE − (−10) I6 = 3 V7 − 0.7 + 10 V7 + 9.3 = = (1) 3 3 Since IB = 0, the collector current will be equal to the current through the 9.1-k resistor, +10 − V7 9.1 Since α1 1, IC = IE = I6 resulting in

IC =

Observe that with VC at 3 V and VB at 4.3 V, the transistor is operating in the active region. Refer to the analysis shown in the figure, which leads to β≡

3.952 IC = 82.3 = IB 0.048

(c)

(2)

V7 + 9.3 10 − V7 = 9.1 3 ⇒ V7 = −4.5 V and I6 =

−4.5 + 9.3 V7 + 9.3 = = 1.6 mA 3 3

4.29 (a)

Observe that the transistor is operating in the active region and note the analysis performed on the circuit diagram. Thus, IC = IE − IB = 3 − 0.04 = 2.96 mA and β≡ Since VC is lower than VB , the transistor is operating in the active region. From the figure above, we see that

2.96 IC = 74 = IB 0.04

4.30

IC = 1 mA

2.5 V

IB = 0.0215 mA

IE 0.5 mA RE

Thus,

VE

1 IC = 46.5 = β≡ IB 0.0215 (b)

VC 0.5 V IC

RC

2.5 V From the figure we see that VC = −0.5 V is lower than the base voltage (VB = 0 V); thus the transistor will be operating in the active mode. β 100 IE = × 0.5 IC = αIE = β +1 100 + 1

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–8

= 0.495 mA VC − (−2.5) RC = IC

4.33 ro = 1/slope = 1/(0.8 × 10−5 )

−0.5 + 2.5 = 4.04 k 4 k 0.495 The transistor VEB can be found from 0.5 mA VEB = 0.64 + VT ln 0.1 mA =

= 0.68 V

= 125 k VA ro = IC VA ⇒ VA = 125 V 1 mA At IC = 10 mA, 125 k =

ro =

Thus,

VA 125 V = 12.5 k = IC 10 mA

VE = +0.68 V 4.34

and 2.5 − 0.68 = 3.64 k 0.5 The maximum allowable value for RC while the transistor remains in the active mode corresponds to VC = +0.4 V. Thus,

iC (mA)

RE =

RCmax =

0.4 − (−2.5) = 5.86 k 0.495

4.31 Refer to the circuit in Fig. P4.31. Since VC = 0 V is equal to VB , the transistor will be operating in the active mode. The transistor VBE can be found from 0.5 mA VBE = 0.7 + 0.025 ln 1 mA = 0.68 V

1.1

VA

0 5 0.3

10 15

v CE (V)

Slope of iC –v CE line corresponding to v BE = 710 mV is Slope =

0.2 mA 1.3 − 1.1 = = 0.02 mA/V 15 − 5 10 V

Near saturation, VCE = 0.3 V, thus iC = 1.1 − 0.02 × (5 − 0.3)

Thus, VE = −0.68 V 101 β +1 IC = IC = 0.5 × IE = α β 100 = 0.505 mA The required value of RE can be found from RE =

1.3

VE − (−1.5) IE

= 1.006 1 mA iC will be 1.2 mA at, 1.2 − 1.1 = 10 V 0.02 The intercept of the iC –v CE straight line on the iC axis will be at v CE = 5 +

iC = 1.1 − 5 × 0.02 = 1 mA

−0.68 + 1.5 = 1.62 k 0.505 To establish VC = 0 V, we select RC according to RE =

1.5 − 0 = 3 k RC = 0.5

Thus, the Early voltage is obtained as Slope =

iC (at v CE = 0) VA

1 = 50 V 0.02 VA 50 V = ro = = 50 k IC 1 mA

⇒ VA =

4.32 Since the BJT is operating at a constant emitter current, its |VBE | decreases by 2 mV for every ◦ C rise in temperature. Thus,

which is the inverse of the slope of the iC –v CE line.

|VBE | at 0◦ C = 0.7 + 0.002 × 25 = 0.75 V

4.35 Refer to the circuit in Fig. P4.35.

◦

|VBE | at 100 C = 0.7 − 0.002 × 75 = 0.55 V

Sedra, Smith, Carusone, Gaudet

(a) For active-mode operation with VC = 2 V:

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–9

IC =

VCC − VC 10 − 2 = = 8 mA RC 1

IB =

8 IC = = 0.16 mA β 50

VC = VCC − VECsat = 3 − 0.2 = 2.8 V

VBB = IB RB + VBE

To operate at the edge of saturation,

= 0.16 × 10 + 0.7 = 2.3 V

VEC = 0.3 V

(b) For operation at the edge of saturation: VCE = 0.3 V VCC − VCE 10 − 0.3 = 9.7 mA = IC = RC 1 IB =

Since 12.2 is lower than the transistor β of 50, we have verified that the transistor is operating in saturation, as assumed.

IC 9.7 = = 0.194 mA β 50

Thus, 3 − 0.3 = 2.7 mA 1 2.7 IC IB = = = 0.054 mA β 50

IC =

RB =

VBB = IB RB + VBE

IC /IB = β = 50

and

2.3 2.3 = 42.6 k = IB 0.054

= 0.194 × 10 + 0.7 = 2.64 V (c) For operation deep in saturation with βforced = 10:

4.37

VCC

VCE = 0.2 V IC = IB =

VCC − VCE 10 − 0.2 = = 9.8 mA RC 1 IC βforced

IE RC

9.8 = 0.98 mA = 10

VC

VBB = IB RB + VBE

IB

= 0.98 × 10 + 0.7 = 10.5 V

IC

VB IB IE

4.36 3 V VB 2.3 V RB 10 k V

ECsat

0.2 V

VC 2.8 V

IB IC

Since the meter resistance is small, VC VB and the transistor is operating in the active region. To obtain IE = 1 mA, we arrange that VBE = 0.7 V. Since VC VB , VC must be set to 0.7 by selecting RC according to VC = 0.7 = VCC − IE RC

1 k

Thus, 0.7 = 9 − 1 × RC ⇒ RC = 8.3 k

Assume saturation-mode operation. From the figure we see that 2.8 VC = = 2.8 mA 1 k 1 VB 2.3 IB = = 0.23 mA = RB 10 IC =

Thus, βforced ≡

IC 2.8 = = 12.2 IB 0.23

Sedra, Smith, Carusone, Gaudet

Since the meter reads full scale when the current flowing through it (in this case, IB is 50 µA), a full-scale reading corresponds to β≡

1 mA IC = 20 IB 50 µA

If the meter reads 1/5 of full scale, then IB = 10 µA and β=

1 mA = 100 10 µA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–10

A meter reading of 1/10 full scale indicates that 1 mA = 200 β= 5 µA

circuit diagrams on the next page; the order of the steps is shown by the circled numbers. 4.41 ICBO approximately doubles for every 10◦ C rise in temperature. A change in temperature from 25◦ C to 125◦ C—that is, an increase of 100◦ C—results in 10 doublings or, equivalently, an increase by a factor of 210 = 1024. Thus ICBO becomes

4.38

2.5 V

ICBO = 10 nA × 1024 = 10.24 µA

IC

5 k VC

4.42

IB

VE 0.7 V IE

10 k

b 50 2.5 V

IE VE − (−2.5) 10 −0.7 + 2.5 = 0.18 mA = 10 Assuming the transistor is operating in the active mode, we obtain IE =

0.18 IE = = 3.5 µA β +1 50 + 1 β 50 IC = IE = × 0.18 = 0.176 mA β +1 51 IB =

VC = +2.5 − IC RC = 2.5 − 0.176 × 5 = 1.62 V Since VC > VB , active-mode operation is verified.

From the figure we can write IS v BE /VT e − ICBO iB = β iC = IS ev BE /VT + ICBO 1 v BE /VT e iE = I S 1 + β

(1) (2) (3)

When the base is left open-circuited, iB = 0 and Eq. (1) yields IS v BE /VT e ICBO = β or equivalently,

4.39 Refer to Fig. 4.15(a) with RC = 5.1 k and RE = 6.8 k. Assuming VBE 0.7 V, then VE = −0.7 V, and −0.7 − (−15) = 2.1 mA IE = 6.8 IC = αIE 2.1 mA

(4) v BE /VT

Substituting for IS e

in Eqs. (2) and (3) gives

iC = iE = (β + 1)ICBO 4.43 (a) If the junction temperature rises to 50◦ C, which is an increase of 30◦ C, the EB voltage decreases to

VC = 15 − 2.1 × 5.1 4.3 V

4.40 In all circuits shown in Fig. P4.40, we assume active-mode operation and verify that this is the case at the end of the solution. The solutions are indicated on the corresponding

Sedra, Smith, Carusone, Gaudet

IS ev BE /VT = βICBO

v EB = 692 − 2 × 30 = 632 mV (b) First, we evaluate VT at 20◦ C and at 50◦ C: VT =

kT q

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–11

This figure belongs to Problem 4.40.

1.5 V

(a) IC a 0.26 0.98 0.26 0.255 mA

4

2.7 k VC 1.5 0.255 2.7 0.81 V 5

6 I I I 0.005 mA B E C 0.8 V 1

VE 0.8 V 2

0.8 (1.5) 2.7 0.26 mA 1.5 V

2.7 k

3 IE

1.5 V

(b) 1.5 0.8 2 0.35 mA

3

6

IE

IB IE IC 0.007 mA

4

IB

1 0.8 V

2 k VE 0.8 V

2

VC 1.5 2 0.343 0.81 V 5 2 k

IC a 0.35 0.98 0.35 0.343

1.5 V

3 V

(c) 3 IE

3 1.8 0.12 mA 10

10 k VE 1.8 V 2

1 V

6

1 0.8 V IB IC/50 2.4 A 4 IC a 0.12 0.98 0.12 0.118 mA

VC 0.118 2 0.236 V 5 2 k

3 V

(d) 4

IC a 0.15 0.147 mA

8.2 k

VC 3 0.147 8.2 1.8 V 5

1.5 V 0.15 6 IB 50 3 A IB 3

V IE E 4.7

0.8 V 1

VE 1.5 0.8 0.7 2 4.7 k

0.7 0.15 mA 4.7

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–12

where k = 8.62 × 10−5 eV/K.

v BE (V) 0.65

0.70

0.72

0.73

0.74

1.45

3.22

4.80

7.16

Thus,

Intercept (mA)

At 20◦ C, T = 293 K and VT = 8.62 × 10−5 × 293 = 25.3 mV

Slope (mA/V) 0.002 0.015 0.032 0.048 0.072

0.2

At 50◦ C, T = 323 K and VT = 8.62 × 10−5 × 323 = 27.8 mV If the transistor is operated at v BE = 700 mV, then (i) At 20◦ C, iE becomes iE = 0.5e(700−692)/25.3 = 0.69 mA (ii) At 50◦ C, iE becomes iE = 0.5e(700−632)/27.8 = 5.77 mA

4.44 v BE = 0.7 V at iC = 10 mA For v BE = 0.5 V, iC = 10e(0.5−0.7)/0.025 = 3.35 µA At a current IC and a BE voltage VBE , the slope of the iC –v BE curve is IC /VT . Thus,

4.46

10 mA = 400 mA/V 25 mV 3.35 µA = of 500 mV = 25 mV

Slope at VBE of 700 mV = Slope at VBE 0.134 mA/V

R2 68 k I2 I1 R1 6.8 k

400 3000 Ratio of slopes = 0.134

IB VBE

IE

VE I 1.1 mA

4.45 Use Eq. (4.18): v CE iC = IS ev BE /VT 1 + VA with IS = 10−15 A and VA = 100 V, to get v CE iC = 10−15 ev BE /0.025 1 + 100

At 25◦ C, assume IE = 1 mA. Thus, VBE = 0.68 V VBE 0.68 V = = 0.1 mA R1 6.8 k

vBE 0.65 V 0.70 V 0.72 V 0.73 V 0.74 V

I1 =

v CE

IE = I − I1 = 1.1 − 0.1 = 1 mA

(V)

iC (mA)

iC (mA)

iC (mA)

iC (mA)

iC (mA)

0

0.196

1.45

3.21

4.81

7.16

15

0.225

1.67

3.70

5.52

8.24

To find the intercept of the straight-line characteristics on the iC axis, we substitute v CE = 0 and evaluate iC = 10−15 ev BE /VT A for the given value of v BE . The slope of each straight line is equal to this value divided by 100 V ( VA ). Thus we obtain

Sedra, Smith, Carusone, Gaudet

which is the value assumed. IE I2 = I1 + IB = I1 + β +1 1 = 0.11 mA 101 Note that the currents in R1 and R2 differ only by the small base current, 0.01 mA. Had I1 and I2 been equal, then we would have had = 0.1 +

I1 R1 = VBE I2 R2 I1 R2 = VBE

R2 R1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–13

VE = −(I1 R1 + I2 R2 ) R2 (1) = −VBE 1 + R1 6.8 = −11 VBE = −7.48 V = −VBE 1 + 0.68

At IC = 100 µA, ro =

50 V = 500 k 0.1 mA

4.48 The equivalent circuits shown in the figure correspond to the circuits in Fig. 4.19.

which gives this circuit the name “VBE multiplier.” A more accurate value of VE can be obtained by taking IB into account: VE = −(I1 R1 + I2 R2 ) R2 = − VBE + VBE + IB R2 R1 R2 VBE − IB R2 =− 1+ R1

(2)

= −7.48 − 0.01 × 68 = −8.16 V As temperature increases, an approximate estimate for the temperature coefficient of VE can be obtained by assuming that IE remains constant and ignoring the temperature variation of β. Thus, we would be neglecting the temperature change of the (IB R2 ) terms in Eq. (2). From Eq. (2) we can obtain the temperature coefficient of VE by utilizing the fact the VBE changes by – 2 mV/◦ C. Thus, Temperature coefficient of VE R2 × −2 =− 1+ R1

E

(b)

iE DB (IS/b)

vEB

B

biB

C

iB

iC

iC 1 mA = 100 = iB 10 µA iC 0.08 mA = 80 βac = = iB v CE constant 1.0 µA

4.49 β =

v CE ro

= −11 × −2 = +22 mV/◦ C

iC = iB × βac +

At 75◦ C, which is a temperature increase of 50◦ C,

where VA 100 = 100 k = ro = IC 1

VE = −8.16 + 0.022 × 50 = −7.06 V As a check on our assumption of constant IE , let us find the value of IE at 75◦ C: I1 (75◦ C) =

ro

Thus, 2 × 103 = 180 µA 100 = 0.18 mA and iC = 1.18 mA

VBE (75◦ C) R1

iC = 2 × 80 +

0.68 − 2 × 10−3 × 50 6.8 = 0.085 mA

=

4.50 Refer to the circuit in Fig. P4.35 (with VBB = VCC ) and to the BJT equivalent circuit of Fig. 4.21.

IE (75◦ C) = I − I1 (75◦ C) = 1.1 − 0.085 = 1.015 mA which is reasonably close to the assumed value of 1 mA.

IC =

VCC − 0.2 RC

IB =

VCC − 0.7 RB

βforced ≡ 4.47 ro =

VA 50 V = IC IC

Thus,

IC IB

VCC − 0.2 VCC − 0.7

Thus,

βforced =

50 V = 50 k At IC = 1 mA, ro = 1 mA

Pdissipated = VCC (IC + IB )

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

RB RC

(1)

© Oxford University Press 2021

Chapter 4–14 5 V

= VCC (βforced IB + IB ) = (βforced + 1)VCC IB

(2)

For VCC = 5 V and βforced = 10 and Pdissipated ≤ 20 mW, we can proceed as follows:

VC 5 0.1 10 4 V

RB = 8.96 RC

4

1.7 V

Using Eq. (1) we can determine (RB /RC ): RB 5 − 0.2 10 = 5 − 0.7 RC ⇒

IE 0.1 mA 3

IC

10 k

VE 1.0 V 1 1.0 IE 10 0.1 mA 2

10 k

(3)

(b) 5 V

Using Eq. (2), we can find IB :

0 3

(10 + 1) × 5 × IB ≤ 20 mW

10 k

⇒ IB ≤ 0.36 mA

VC 5 0 5 V 4

Thus, 0V

VCC − 0.7 ≤ 0.36 mA RB

VE 0 V

⇒ RB ≥ 11.9 k

1

10 k 0 2

From the table of 1% resistors in Appendix J we select

(c)

RB = 12.1 k RC = 1.35 k

The analysis and the results are given on the circuit diagrams of Figs. a through c. The circled numbers indicate the order of the analysis steps.

From the table of 1% resistors in Appendix J we select

4.52 Refer to the circuit in Fig. P4.52.

Substituting in Eq. (3), we have

VE = 1 V

RC = 1.37 k

2.5 − 1 = 0.15 mA 10 VB = VE − 0.7 = 0.3 V IE =

For these values: 5 − 0.2 = 3.5 mA IC = 1.37 5 − 0.7 = 0.36 mA IB = 12.1 Thus, βforced = Pdissipated

0.3 VB = = 0.003 mA 100 k 100 IC = IE − IB = 0.15 − 0.003 = 0.147 mA

IB =

VC = −2.5 + 10 × 0.147 = −1.03 V

3.5 = 9.7 0.36 = VCC (IC + IB )

Observe that VC < VB , confirming our implicit assumption that the transistor is operating in the active region.

= 5 × 3.86 = 19.3 mW

β=

0.147 IC = 49 = IB 0.003

α=

49 β = = 0.98 β +1 49 + 1

4.51 5 V 10 k

IC

IE 0.13 mA 3 VC 5 0.13 10 3.7 V 4

2 V

4.53 A loop equation for the EB loop yields

10 k

VE 2 0.7 1.3 V 1 1.3 IE 0.13 mA 2 10

(a)

Sedra, Smith, Carusone, Gaudet

3 = IB RB + VBE + IE RE ⇒ IE =

3 − 0.7 RB RE + β +1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–15 3 V IB RB

IE =

IC RC

VE VE = = VE RE 1

(3)

Substituting from Eqs. (1), (2), and (3) into

1 k VC

IE = IB + IC

VB

gives VE

IE

VE = 0.23 − 0.1 VE + 2.8 − VE

RE 1 k

⇒ VE = 1.44 V VC = 1.64 V VB = 2.14 V

2.3 IE = RB 1+ 101 (a) For RB = 100 k,

3 − 2.14 = 0.086 mA 10 3 − 1.64 IC = = 1.36 mA 1 Thus,

IB =

2.3 = 1.16 mA 100 1+ 101 VE = IE RE = 1.16 × 1 = 1.16 V IE =

1.36 IC = = 15.8 IB 0.086 which is lower than the value of β, verifying saturation-mode operation.

VB = VE + 0.7 = 1.86 V

(c) For RB = 1 k, we assume saturation-mode operation:

Assuming active-mode operation, we obtain IC = αIE = 0.99 × 1.16 = 1.15 mA

VB = VE + 0.7

VC = 3 − 1.15 × 1 = +1.85 V Since VC > VB − 0.4, the transistor is operating in the active region, as assumed.

VC = VE + 0.2

IE =

2.3 = 2.09 mA 10 1+ 101 VE = 2.09 × 1 = 2.09 V

3 − (VE + 0.7) = 2.3 − VE 1 3 − (VE + 0.2) = 2.8 − VE IC = 1 VE IE = = VE 1 These values can be substituted into

VB = 2.09 + 0.7 = 2.79 V

IE = IB + IC

Assuming active-mode operation, we obtain

to obtain

IC = αIE = 0.99 × 2.09 = 2.07 mA

VE = 2.3 − VE + 2.8 − VE

VC = 3 − 2.07 = +0.93 V

⇒ VE = 1.7 V

Since VC < VB − 0.4, the transistor is operating in saturation, contrary to our original assumption. Therefore, we need to redo the analysis assuming saturation-mode operation, as follows:

VB = 2.4 V

(b) For RB = 10 k,

VB = VE + 0.7 3 − VB RB

2.3 − VE 3 − VE − 0.7 = 10 10 3 − VC 3 − VE − 0.2 = IC = RC 1

=

= 2.8 − VE

Sedra, Smith, Carusone, Gaudet

VC = 1.9 V Now checking the currents, 3 − 2.4 = 0.6 mA 1 3 − 1.9 IC = = 1.1 mA 1 Thus, the transistor is operating at a forced β of

IB =

VC = VE + VCEsat = VE + 0.2 IB =

IB =

(1)

(2)

βforced =

IC 1.1 = 1.83 = IB 0.6

which is much lower than the value of β, confirming operation in saturation.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–16

3 V

(c)

4.54

3 V

(a) IC

0.5 mA

0.5 mA

IB 0

V2 3 0.5 3.6 1.2 V

0V

V7 3 0.5 3.6 1.2 V

43 k

3.6 k

43 k

3.6 k

IB

0

V6

V5 0.7 V

0

4.7 k IE 0.7 (–3) 4.7 0.5 mA 3 V

V1 0.7 V

3 V

(d)

0.5 mA

3 1.45 6.2 0.25 mA

IE

(b)

110 k

3.6 k

V9 3 0.25 10 0.5 V

0.75 V

V3 3 0.5 3.6 1.2 V

~0 IB

V8 0.75 0.7 1.45 V

0.75 V ~ 0 mA

3 V

~ 0.5 mA

6.2 k

10 k

0.25 mA

3 V

VE 0.7 V 4.7 k

3 V

I4 0.7 (3) 0.5 mA 4.7

3 V

(e) 6 I 480 0.0125 mA

V10 3 0.0125 300 0.75 V

180 k

3 1.45 IE 0.25 mA 6.2 6.2 k V11 0.75 0.7 1.45 V

0 300 k

V12 3 10 0.25 10 k 0.5 V 0.25 mA

3 V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–17

3 V

4.55 (a) See solution and answer on the figure, which corresponds to Fig. P4.54(a).

IC

3 V

3.6 k V7

V6

IC a 0.5 0.99 0.5 0.495 mA

43 k 3.6 k

VB 0.005 43 0.215 V

IB IE /(b 1) V5

V2 3 0.495 3.6 1.218 V

43 k

4.7 k

IE 0.5 IB 101

0.005 mA V1 0.215 0.7 0.915 V

3 V (c)

0.5 mA

3 V IE

3 V (a)

6.2 k

3 V

0.75 V

a 0.49 0.484 mA

V8 IB

3.6 k IB

0.5 101

V3 3 0.484 3.6 1.258 V

0.005 mA

110 k V9 VE 0.7

IC

4.7 k 0.7 (3) I4 4.7

10 k

0.49 mA

3 V

3 V (b)

(d)

This figure belongs to Problem 4.55, part (e). 3 V IE 180 k

3 V IE 6.2 k

6.2 k V11

VBB

V10

IC

V10

V12

V11

IB

IB

RB

V12

IC

10 k

10 k

300 k 3 V

Sedra, Smith, Carusone, Gaudet

(e)

3 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–18

(b) See solution and answer on the figure, which corresponds to Fig. P4.61(b). a (c)

4.56 Refer to Fig. P4.56. To obtain IE = 0.2 mA and VE = 2 V we select RE =

2V = 10 k 0.2 mA

Now, the base voltage has to be

Writing an equation for the loop containing the BEJ of the transistor leads to IE =

VB = VE + 0.7 = 2.7 V For IB2 = 0.1 mA, the value of RB2 must be

3 − 0.7 = 0.449 mA 43 4.7 + 101

RB2 =

VB 2.7 V = 27 k = IB2 0.1 mA

V5 = −3 + 0.449 × 4.7 = −0.9 V

The base current IB is

V6 = −0.9 + 0.7 = −0.2 V

IB =

IC = αIE = 0.99 × 0.449 = 0.444 mA

0.2 IE = = 0.002 mA β +1 100 + 1

and the current IB1 through RB1 is

V7 = 3 − 0.444 × 3.6 = +1.4 V

IB1 = IB2 +IB = 0.1+0.002 = 0.102 mA

(d)

The required value of RB1 can now be found as

An equation for the loop containing the EBJ of the transistor yields

RB1 =

IE =

3 − 0.75 − 0.7 = 0.213 mA 110 6.2 + 101

9 − 2.7 = 61.8 k 0.102

The collector current IC is IC = αIE = (

β ) IE = 0.99×0.2 = 0.198 mA β +1

V8 = +3 − 0.213 × 6.2 = +1.7 V

To obtain VC = +5 V, we select RC as

VB = 1 V

RC =

IC = αIE = 0.99 × 0.213 = 0.21 mA

Consulting the table of 5% resistors, in Appendix J, we find the nearest resistance values to be:

V9 = −3 + 0.21 × 10 = −0.9 V (e) See figure on the previous page. First, we use Thévenin’s theorem to replace the voltage divider feeding the base with VBB and RB : VBB

6 × 300 = +0.75 V = −3 + 480

RB = 180 300 = 112.5 k Next we write an equation for the loop containing the EBJ to obtain

9−5 = 20.2 k 0.198

RE = 10 k, RB1 = 62 k, RB2 = 27 k, and RC = 20 k We can use these values to determine 27 ) − 0.7 27 + 62 0.2 mA IE = 2762 10 + 101 VE = 0.2 × 10 = 2 V 9×(

3 − 0.75 − 0.7 = 0.212 mA IE = 112.5 6.2 + 101

VB = 2.7 V 2.7 V = 0.1 mA IB2 = 27 k IC = 0.198 mA

V11 = +3 − 0.212 × 6.2 = +1.7 V

VC = 9 − 0.198 × 20 = 5.04 V

V10 = 1.7 − 0.7 = +1 V

Which are equal or very close to the desired values!

IC = αIE = 0.99 × 0.212 = 0.21 mA V12 = −3 + 0.21 × 10 = −0.9 V Note that these results are nearly identical to those for case (d) above.

Sedra, Smith, Carusone, Gaudet

4.57 Refer to the circuit in Fig. P4.57. (a) For v I = 0, both transistors are cut off and all currents are zero. Thus VB = 0 V

and

Microelectronic Circuits, 8th International Edition

VE = 0 V

© Oxford University Press 2021

Chapter 4–19

(b) For vI = −2 V, Q1 will be off and Q2 will be on, and the circuit reduces to that in Fig. 1. Since VB > −2 V, VC will be lower than VB , and Q2 will be operating in the active region. Thus

2.5 V IC

10 k 5 V

Q1 IB

2 − 0.7 = 1.1 mA 10 1+ 51 VE = −IE × 1 = −1.1 V

VB

IE =

VE

IE

1 k

VB = VE − 0.7 = −1.8 V Figure 3 5 − 0.7 = 3.6 mA 10 1+ 51 VE = 3.6 × 1 = +3.6 V IE =

VB = +4.3

Figure 1

(c) For vI = +2.5 V, Q1 will be conducting and Q2 will be cut off, and the circuit reduces to that in Fig. 2. 2.5 V 10 k 2.5 V

Q1 IB

VB

VE IE 1 k

which is greater than VC . This is an impossible situation, indicating that Q1 must be saturated. Assuming saturation-mode operation, we obtain VE = VC − VCEsat = 2.5 − 0.2 = +2.3 V VE = 2.3 mA IE = 1 k VB = VE + 0.7 = 3.0 V 5−3 IB = = 0.2 mA 10 IC = IE − IB = 2.3 − 0.2 = 2.1 mA βforced =

IC 2.1 = 10.5 = IB 0.2

which is lower than β, verifying that Q1 is operating in saturation.

Figure 2 Since VB will be lower than +2.5 V, VC will be higher than VB , and the transistor will be operating in the active mode. Thus, 2.5 − 0.7 = 1.5 mA 10 1+ 51 VE = IE × 1 = +1.5 V

4.58 5 V

IE =

10 k IC

VB = VE + 0.7 = +2.2 V

5 (VB 0.4) 10

VCE 0.3 V

VB

(d) For vI = +5, Q1 will be on and Q2 will be cut off, and the circuit reduces to that in Fig. 3. Here we do not know whether Q1 is operating in the active mode or in saturation. Assuming active-mode operation, we obtain

Sedra, Smith, Carusone, Gaudet

VC VB 0.4

V 0.7 IE B 10

Microelectronic Circuits, 8th International Edition

VB 0.7

10 k

Figure 1

© Oxford University Press 2021

Chapter 4–20

Figure 1 shows the circuit with the value of VB that results in operation at the edge of saturation. Since β is very high,

Assuming operation in the active mode, we have IC = αIE IE = 2.3 mA VC = +3 − IC × 1 = 3 − 2.3 = +0.7 V

IC IE

Since VC > VB − 0.4, the BJT is operating in the active mode, as assumed.

VB − 0.7 5 − (VB − 0.4) = 10 10 ⇒ VB = 3.05 V

(c) For VB = +1 V, VE = 1 − 0.7 = +0.3 V

5 V IC

IB VB

VB 0.5 V

VCEsat 0.2 V

IC = αIE IE = 3.3 mA

VC = 3 − 3.3 × 1 = −0.3 V

VB 0.7 V

10 k

IE

0.3 − (−3) = 3.3 mA 1 Assuming operation in the active mode, we have

IE =

10 k

Now VC < VB − 0.4, indicating that the transistor is operating in saturation, and our original assumption is incorrect. It follows that

Figure 2

VC = VE + VCEsat

Figure 2 shows the circuit with the value of VB that results in the transistor operating in saturation, with VB − 0.7 = 0.1VB − 0.07 10 5 − (VB − 0.5) = 0.55 − 0.1VB IC = 10 IB = IE − IC = 0.2VB − 0.62 IE =

For βforced = 3,

= 0.3 + 0.2 = 0.5 V 3 − 0.5 3 − VC = = 2.5 mA 1 1 IB = IE − IC = 3.3 − 2.5 = 0.8 mA IC =

βforced =

IC 2.5 = 3.1 = IB 0.8

(d) When VB = 0 V, IE = 2.3 mA. The emitter current becomes 0.23 mA at VB = −3 + 0.23 × 1 + 0.7 = −2.07 V

IC =3 IB

(e) The transistor will be at the edge of conduction when IE 0 and VBE = 0.5 V, that is,

0.55 − 0.1VB =3 0.2VB − 0.62

VB = −3 + 0.5 = −2.5 V

⇒ VB = 3.44 V

In this case, VE = −3 V

4.59 Refer to the circuit in Fig. P4.59. (a) For VB = −1 V, VE = VB − VBE = −1 − 0.7 = −1.7 V −1.7 + 3 VE − (−3) = = 1.3 mA 1 1 Assuming active-mode operation, we have

IE =

VC = +3 V (f) The transistor reaches the edge of saturation when VCE = 0.3 V but IC = αIE IE : VE = VB − 0.7

IC = αIE IE = 1.3 mA

VB − 0.7 − (−3) = VB + 2.3 1 VC = VE + 0.3 = VB − 0.4

VC = +3 − IC × 1 = 3 − 1.3 = +1.7 V

IC =

Since VC > VB − 0.4, the transistor is operating in the active mode as assumed.

Since

(b) For VB = 0 V, VE = 0 − VBE = −0.7 V IE =

−0.7 − (−3) = 2.3 mA 1

Sedra, Smith, Carusone, Gaudet

IE =

3 − VB + 0.4 3 − VC = = 3.4 − VB 1 1

IC IE 3.4 − VB = VB + 2.3 VB = 0.55 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–21

5 V

For this value, VE = 0.55 − 0.7 = −0.15 V

4 mA

VC = −0.15 + 0.3 = +0.15 V

4 mA VC 4 V

(g) For the transistor to operate in saturation with βforced = 2,

0 1 k

VE = VB − 0.7 VB − 0.7 − (−3) = VB + 2.3 1 VC = VE + VCEsat = VB − 0.7 + 0.2 = VB − 0.5 IE =

VB 1.5 V

0

3 − (VB − 0.5) = 3.5 − VB 1 IB = IE − IC = 2 VB − 1.2

1 k

2 mA

IC =

5 V Figure 2

IC 3.5 − VB =2 = IB 2 VB − 1.2 ⇒ VB = +1.18 V

(c)

5 V

4.60 (a) VB = 0 V

4 mA

5 V 4 mA

4

VE 2 V

2 mA

4 1.3 2.7 mA 5 VC 2.7 1 2.7 V 6 1.3 mA

VC VB 0.4

IC

VB 0.4 1

1 k 0.3 V

VB IE

VE VB 0.7

1 k 3

2 mA

2 0.7 1.3 mA VE 0.7 V 1 0.7 mA 1 k 2

1 k 2 mA

VB 0.7 1 Figure 3

5 V Figure 1 The analysis is shown on the circuit diagram in Fig. 1. The circled numbers indicate the order of the analysis steps. (b) The transistor cuts off at the value of VB that causes the 2-mA current of the current source feeding the emitter to flow through the 1-k resistor connected between the emitter and ground. The circuit under these conditions is shown in Fig. 2. Observe that VE = −2 mA ×1 k = −2 V, IE = 0, and VB = VE + 0.5 = −1.5 V. Since IC = 0, all the 4 mA supplied by the current source feeding the collector flows through the collector 1-k resistor, resulting in VC = +4 V.

Sedra, Smith, Carusone, Gaudet

Figure 3 shows the transistor at the edge of saturation. Here VCE = 0.3 V and IC = αIE IE . A node equation at the emitter gives IE = 2 + VB − 0.7 = VB + 1.3 mA A node equation at the collector gives IC = 4 − (VB − 0.4) = 4.4 − VB mA Imposing the condition IC IE gives 4.4 − VB = VB + 1.3 ⇒ VB = +1.55 V Correspondingly, VE = +0.85 V VC = +1.15 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–22 1.5 − 0.7 = 0.079 mA 10.5 10 + 101 IC = αIE = 0.99 × 0.079 = 0.078 mA

4.61

⇒ IE =

VCC 5 V 0.1 mA RC 10 k

R1 0 VB 1.5 V

RE 10 k

R2

VC = +5 − 0.078 × 10 = +4.22 V Since VB = IE RE + VBE = 0.79 + 0.7 = 1.49 V, we see that VC > VB − 0.4, and thus the transistor is operating in the active region, as assumed. 4.62

3 V Figure 1

IE

RE

From Fig. 1 we see that R1 + R2 = VCC

VE 0.7 V

VCC 5 = = 50 k 0.1 mA 0.1

R2 = 1.5 R1 + R2

VBC 1 V

R2 = 1.5 5× 50 ⇒ R2 = 15 k R1 = 50 − 15 = 35 k

R2 15 = 1.5 V =5× R1 + R2 35 + 15

RB = R1 R2 = 15 35 = 10.5 k

5 V

IC VBB

RC

3 V

For β = 100, to obtain the collector current, we replace the voltage divider with its Thévenin equivalent, consisting of VBB = 5 ×

VC 1 V

IC

Refer to the figure. To obtain IE = 0.5 mA we select RE according to 3 − 0.7 = 4.6 k 0.5 To obtain VC = −1 V, we select RC according to RE =

−1 − (−3) = 4 k 0.5 where we have utilized the fact that α 1 and thus IC IE = 0.5 mA. From the table of 5% resistors in Appendix J we select RC =

RC 10 k VC

RB

RE = 4.7 k

RC = 3.9 k

and

For these values,

IB

3 − 0.7 = 0.49 mA 4.7 IC IE = 0.49 mA

IE =

IE

RE 10 k

VBC = 0 − VC = −(−3 + 0.49 × 3.9) = −1.1 V Figure 2

4.63

Refer to Fig. 2. Assuming active-mode operation, we can write a loop equation for the base–emitter loop: VBB = IB RB + VBE + IE RE 1.5 =

IE × 10.5 + 0.7 + IE × 10 β +1

Sedra, Smith, Carusone, Gaudet

Writing a loop equation for the EBJ loop, we have 3 = IE RE + VEB + IB RB = IE × 2.2 + 0.7 +

Microelectronic Circuits, 8th International Edition

(1)

IE × 20 β +1

© Oxford University Press 2021

Chapter 4–23 3 V

4.64

RE

2.2 k IE

IB

VE

Assume active-mode operation: IE =

VB RB 20 k

VC IC

RC 2.2 k 3 V

3 − VEB RB RE + β +1

3 − 0.7 = 0.887 mA 20 2.2 + 51 IE 0.887 IB = = = 0.017 mA β +1 51 IE =

3 V

3 − 0.7 = 0.86 mA 20 2.2 + 41 VE = 3 − 0.86 × 2.2 = +1.11 V ⇒ IE =

RE 2.2 k VE

IE

VB = VE − 0.7 = +0.41 V

VB RB 20 k

Assuming active-mode operation, we obtain 40 × 0.86 = 0.84 mA IC = αIE = 41 VC = −3 + 0.84 × 2.2 = −1.15 V Since VC < VB + 0.4, the transistor is operating in the active mode, as assumed. Now, if RB is increased to 100 k, the loop equation [Eq. (1)] yields 3 − 0.7 = 0.5 mA IE = 100 2.2 + 41 VE = 3 − 0.5 × 2.2 = +1.9 V

IB

VC RC 2.2 k

IC

3 V

IC = IE − IB = 0.887 − 0.017 = 0.870 mA VB = IB RB = 0.017 × 20 = 0.34 V VE = VB + VEB = 0.34 + 0.7 = 1.04 V VC = −3 + IC RC = −3 + 0.87 × 2.2 = −1.09 V

VB = VE − VEB = 1.9 − 0.7 = +1.2 V Assuming active-mode operation, we obtain 40 × 0.5 = 0.48 mA 41 VC = −3 + 0.48 × 2.2 = −1.9 V

IC = αIE =

Since VC < VB + 0.4, the transistor is operating in the active mode, as assumed. If with RB = 100 k, we need the voltages to remain at the values obtained with RB = 20 k, the transistor must have a β value determined as follows. For IE to remain unchanged,

Thus, VC < VB + 0.4, which means active-mode operation, as assumed. The maximum value of RC that still guarantees active-mode operation is that which causes VC to be 0.4 V above VB : that is, VC = 0.34 + 0.4 = 0.74 V. Correspondingly, RCmax =

0.74 − (−3) = 4.3 k 0.87

4.65 5 V RE

IE

3 − 0.7 3 − 0.7 = 20 100 2.2 + 2.2 + 41 β +1 ⇒

100 20 = 41 β +1

β +1=

RB IC

410 = 205 2

β = 204

Sedra, Smith, Carusone, Gaudet

VC RC

5 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–24

We require IE to be nominally 1 mA (i.e., at β = 100) and to remain within ±10% as β varies from 50 to 150. Writing an equation for the loop containing the EBJ results in IE =

5 − 0.7 RB RE + β +1

RC = 4 k Finally, for our design we need to determine the range obtained for collector current and collector voltage for β ranging from 50 to 150 with a nominal value of 100. We compute the nominal value of IE from 4.3 = 0.96 mA 50 4+ 101 We utilize Eqs. (2) and (3) to compute IEmin and IEmax , IEnominal =

Thus, 4.3 =1 (1) RB RE + 101 4.3 (2) = IEmin RB RE + 51 4.3 (3) = IEmax RB RE + 151 If we set IEmin = 0.9 mA and solve Eqs. (1) and (2) simultaneously, we obtain

4.3 = 0.86 mA 50 4+ 51 4.3 = 0.99 mA IEmax = 50 4+ 151 Thus, IEmin =

IEmax

=

0.99 = 1.03 0.96

RE = 3.81 k

IEnominal

RB = 49.2 k

IEmin 0.86 = 0.9 = IEnominal 0.96

Substituting theses values in Eqs. (2) and (3) gives IEmin = 0.9 mA

which meet our specifications. The collector currents are ICnominal = 0.99 × 0.96 = 0.95 mA

IEmax = 1.04 mA

ICmin = 0.99 × 0.86 = 0.85 mA

Obviously, this is an acceptable design. Alternatively, if we set IEmax in Eq. (3) to 1.1 mA and solve Eqs. (1) and (3) simultaneously, we obtain RE = 3.1 k

ICmax = 0.99 × 0.99 = 0.98 mA and the collector voltages are VCnominal = −5 + 0.95 × 4 = −1.2 V VCmin = −5 + 0.85 × 4 = −1.6 V

RB = 119.2 k Substituting these values in Eqs. (2) and (3) gives

VCmax = −5 + 0.98 × 4 = −1.1 V

IEmin = 0.8 mA IEmax = 1.1 mA

4.66

Obviously this is not an acceptable design (IEmin is 20% lower than nominal). Therefore, we shall use the first design. Specifying the resistor values to the nearest kilohm results in

3 V

RE = 4 k

2.3 V

RB = 50 k

IB

To obtain the value of RC , we note that at the nominal emitter current value of 1 mA, VC = −1 V,

100 k

VC

IC RC

IC = αIE = 0.99 mA −1 − (−5) = 4.04 k 0.99 Specified to the nearest kilohm, RC =

Sedra, Smith, Carusone, Gaudet

IB =

2.3 V = 0.023 mA 100 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–25

Since VC = 2 V is lower than VB , which is +2.3 V, the transistor will be operating in the active mode. Thus, IC = βIB = 50 × 0.023 = 1.15 mA To obtain VC = 2 V, we select RC according to RC =

VC 2V = 1.74 k = IC 1.15 mA

Now, if the transistor is replaced with another having β = 100, then IC = 100 × 0.023 = 2.3 mA which would imply VC = 2.3 × 1.74 = 4 V

Figure 2 (β = 100)

which is impossible because the base is at 2.3 V. Thus the transistor must be in the saturation mode and

Thus, IE1 = 0.0023 mA 101 V1 = 0.0023 × 100 IB1 =

VC = VE − VECsat = 3 − 0.2 = 2.8 V

= 0.23 V 4.67 (a) β = ∞

V2 = V1 + 0.7 = 0.93 V

The analysis and results are given in the circuit diagram in Fig. 1 below. The circled numbers indicate the order of the analysis steps.

Then we write a node equation at C1 :

(b) β = 100

IC1 = IB2 +

IC1 = αIE1 = 0.99 × 0.228 = 0.226 mA

By reference to Fig. 2, we can write an equation for the loop containing the EBJ of Q1 as follows: 3 = IE1 × 9.1 + 0.7 + IB1 × 100

−3 + 4.3IE2 + 0.7 + 3 IE2 + 101 9.1 4.3IE2 0.7 IE2 + + = 101 9.1 9.1 ⇒ IE2 = 0.31 mA

0.226 =

Substituting IB1 = IE1 /(β + 1) = IE1 /101 and rearranging, we obtain IE1 =

V3 − (−3) 9.1 Substituting for IC1 = 0.226 mA, IB2 = IE2 /101, and V3 = V4 + 0.7 = −3 + IE2 × 4.3 + 0.7 gives

3 − 0.7 = 0.228 mA 100 9.1 + 101

This figure belongs to Problem 4.67. 3 V 3 – 0.7 IE1 0.25 mA 9.1

4

2 V1 0 100 k

0

1

V3

9.1 k V2 0.7 V 3 Q1

0.25 mA 5 0

5.1 k 9 0.37 mA V5 3 0.37 5.1 1.1 V 10 Q2

9.1 k

6

7

4.3 k

V3 3 0.25 9.1 0.725 V 3 V

Figure 1

Sedra, Smith, Carusone, Gaudet

V4 V3 0.7 1.425 V IE2 –1.425 – (–3) 0.37 mA 4.3

8

(β = ∞)

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–26 5 − V3 5−0 = = 10 k 0.5 0.5 V4 = 0 + 0.7 = 0.7 V

IB2 = 0.0031

R2 =

IC1 − IB2 = 0.226 − 0.0031 = 0.223 mA V3 = −3 + 0.223 × 9.1 = −0.97 V −1 V

5 − V4 5 − 0.7 = = 8.6 k 0.5 0.5 V5 − (−5) −2 + 5 = = 6 k R4 = 0.5 0.5 V6 = V5 − 0.7 = −2 − 0.7 = −2.7 V

R3 =

V4 = V3 − 0.7 = −1.7 V IC2 = αIE2 = 0.99 × 0.31 = 0.3 mA V5 = +3 − 0.3 × 5.1 = +1.47 V

V6 − (−5) −2.7 + 5 = = 2.3 k 1 1 5−1 5 − V7 = = 4 k R5 = 1 1 Consulting the table of 5% resistors in Appendix J, we select the following resistor values:

R6 =

4.68 Figure 1 below shows the circuit with β = ∞; the required voltage values are indicated. The resistor values are obtained as follows: V2 = −0.7 V V2 − (−5) 0.5 mA ⇒ R1 = 8.6 k

R1 =

R1 = 8.2 k R2 = 10 k R3 = 8.2 k

This figure belongs to Problem 4.68. 5 V 0.5 mA

0.5 mA

1 mA

R3

R2

R5

V4 0

V7 1 V

Q2 V3 = 0

Q1

0

0.5 mA

V2

Q3

R1

V5 = –2 V

R4

0.5 mA

V6 R6

1 mA –5 V

Figure 1 This figure belongs to Problem 4.68. 5 V I1

IE2

10 k

8.2 k

3.9 k V4

IB2

V7

Q2

IC3

IC1 Q1 V2 IE1 8.2 k

IC2

V3

IB3 Q3

I2 V5 6.2 k

V6

IE3

2.4 k

–5 V

Figure 2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–27

R4 = 6.2 k R5 = 3.9 k

5 V

R6 = 2.4 k

The circuit with the selected resistor values is shown in Fig. 2. Analysis of the circuit proceeds as follows:

IC 10 k

IB

20 k

VE + 0.2

V2 = −0.7 V V2 − (−5) −0.7 + 5 = = 0.524 mA 8.2 8.2 = αIE1 = 0.99 × 0.524 = 0.52 mA

IE1 = IC1

VE + 0.7 V VE

IE

The current I1 through the 10-k resistor is given by

1 k

IE2 101 Noting that the voltage drop across the 10 k resistor is equal to (IE2 × 8.2 + 0.7), we can write

Now, imposing the constraint

I1 × 10 = 8.2IE2 + 0.7

IE = IC + IB

Thus, IE2 = 8.2IE2 + 0.7 10 0.52 − 101

results in

I1 = IC1 − IB2 = IC1 −

⇒ IE2 = 0.542 mA V4 = 5 − 0.542 × 8.2 = 0.56 V V3 = 0.56 − 0.7 = −0.14 V IC2 = αIE2 = 0.99 × 0.542 = 0.537 mA IE3 101 Since the voltage drop across the 6.2-k resistor is equal to (0.7 + IE3 × 2.4),

I2 = IC2 − IB3 = 0.537 −

I2 × 6.2 = 0.7 + 2.4IE3 IE3 6.2 0.537 − = 0.7 + 2.4IE3 101

(a)

VE = 0.5 − 0.1(VE + 0.2) + 0.25 − 0.05(VE + 0.7) ⇒ VE = 0.6 V VC = 0.8 V VB = 1.3 V 5 − 0.8 = 0.42 mA 10 5 − 1.3 = 0.185 mA IB = 20 0.42 βforced = = 2.3 0.185 which is less than the value of β1 verifying saturation-mode operation. IC =

(b)

⇒ IE3 = 1.07 mA

5 V

V6 = −5 + 1.07 × 2.4 = −2.43 V V5 = V6 + 0.7 = −1.73

IE

IC3 = α × IE3 = 0.99 × 1.07 = 1.06 mA

1 k VE

V7 = −3.9 × 1.06 = 0.87 V

VE – 0.7 4.69 (a) Assuming saturation-mode operation, the terminal voltages are interrelated as shown in the figure, which corresponds to Fig. P4.69(a). Thus we can write IE =

VE = VE 1

IC =

5 − (VE + 0.2) = 0.5 − 0.1(VE + 0.2) 10

IB =

5 − (VE + 0.7) = 0.25 − 0.05(VE + 0.7) 20

Sedra, Smith, Carusone, Gaudet

VE – 0.2

IB

10 k

IC

1 k

–5 V (b) Assuming saturation-mode operation, the terminal voltages are interrelated as shown in the figure, which corresponds to Fig. P4.69(b). We can obtain the currents as follows:

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–28 5 − VE = 5 − VE 1 VE − 0.2 = VE − 0.2 IC = 1 VE − 0.7 − (−5) = 0.1 VE + 0.43 IB = 10 Imposing the constraint

and denoted its voltage as V. We next draw a circle around the two transistors to define a “supernode.” A node equation for the supernode will be

IE =

IE3 + IC4 = IB3 + I + IE4

(1)

where 5 − (V + 0.2) = 0.48 − 0.1V (2) 10 5 − (V − 0.5) = 0.183 − 0.033V (3) IC4 = 30 (V − 0.5) = 0.1V − 0.05 (4) IB3 = 10 V = 0.05V (5) I= 20 V − 0.7 = 0.1V − 0.07 (6) IE4 = 10 Substituting from Eqs. (2)–(6) into Eq. (1) gives IE3 =

IE = IB + IC results in 5 − VE = VE − 0.2 + 0.1 VE + 0.43 ⇒ VE = +2.27 V VC = +2.07 V VB = 1.57 V 2.07 = 2.07 mA 1 1.57 − (−5) IB = = 0.657 mA 10 IC 2.07 = 3.2 βforced = = IB 0.657 IC =

0.48 − 0.1V + 0.183 − 0.033V = 0.1V − 0.05 + 0.05V + 0.1V − 0.07

which is lower than the value of β, verifying saturation-mode operation.

⇒ V = 2.044 V Thus

(c) We will assume that both Q3 and Q4 are operating in saturation. To begin the analysis shown in the figure, which corresponds to Fig. P4.69(c), we denote the voltage at the collector of Q3 as V and then obtain the voltages at all other nodes in terms of V, utilizing the fact that a saturated transistor has |VCE | = 0.2 V and of course |VBE | = 0.7 V. Note that the choice of the collector node to begin the analysis is arbitrary; we could have selected any other node

VC3 = V = 2.044 V VC4 = V − 0.5 = 1.54 V Next we determine all currents utilizing Eqs. (2)–(6): IE3 = 0.276 mA

IC4 = 0.116 mA

IB3 = 0.154 mA

I = 0.102 mA

IE4 = 0.134

This figure belongs to Problem 4.69, part (c).

5 V IE3

IC4

10 k

30 k Supernode

V + 0.2 V – 0.5

Q3 IB3

10 k

V – 0.5

V I

Q4 20 k

V – 0.7 10 k IE4

(c)

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 4–29

The base current of Q4 can be obtained from

βforced3 =

IC3 0.120 = = 0.8 IB3 0.154

βforced4 =

IC4 0.116 = = 6.4 IB4 0.018

IB4 = IE4 − IC4 = 0.134 − 0.116 = 0.018 mA Finally, the collector current of Q3 can be found as IC3 = I + IB4 = 0.102 + 0.018 = 0.120 The forced β values can now be found as

Sedra, Smith, Carusone, Gaudet

Both βforced values are well below the β value of 50, verifying that Q3 and Q4 are in deep saturation.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 5–1

W 1 = = 2 mA/V2 L 1 × 0.5

Chapter 5

∴ k n = k n

Solutions to Exercises within the Chapter

At v DS = 0.2 V, vDS < vOV , thus the transistor is operating in the triode region,

Ex: 5.1 ox 34.5 pF/m Cox = = = 8.625 fF/µm2 tox 4 nm

1 iD = kn (vOV vDS − v2DS ) 2 1 = 2(0.5 × 0.2 − × 0.22 ) = 0.16 mA 2 At vDS = 0.5 V, vDS = vOV , thus the transistor is operating in saturation,

μn = 450 cm2 /V · S k n = μn Cox = 388 µA/V2 vOV = v GS − Vt = 0.5 V W 1 W = k n vOV ⇒ = 5.15 1 k L L L = 0.18 µm, so W = 0.93 µm gDS =

Ex: 5.2 ox 34.5 pF/m Cox = = 24.6 fF/µm2 = tox 1.4 nm μn = 216 cm2 /V · s

1 2 1 kn v = × 2 × 0.52 = 0.25 mA 2 OV 2 At vDS = 1 V, v DS > vOV and the transistor is operating in saturation with iD = 0.25 mA. iD =

Ex: 5.6 VA = VA L = 5 × 0.8 = 4 V λ=

1 = 0.25 V−1 VA

vDS = 0.8 V > vOV = 0.2 V

k n = μn Cox = 531 µA/V2

⇒ Saturation: iD =

1 W iD = k n v2OV 2 L 1 50 = × 531 × 10 × v2OV 2 ∴ vOV = 0.14 V

1 W 2 k v (1 + λvDS ) 2 n L OV

16 1 × 400 × × 0.22 (1 + 0.25 × 0.8) 2 0.8 = 0.192 mA VA 4 = 25 k = ro = iD 0.16 iD =

vGS = Vt + vOV = 0.49 V vDS, min = vOV = 0.14 V 1 W 2 k v in saturation 2 n L OV Change in iD is: Ex: 5.3 iD =

(a) double L, 0.5

where iD is the value of iD without channel-length modulation taken into account. vDS 1V = 0.04 mA = ⇒ iD = ro = iD 25 k 40 µA Ex: 5.7

1.8 V

(b) double W, 2 (c) double vOV , 2 = 4 2

(d) double vDS , no change (ignoring length modulation)

vG iD

(e) changes (a)–(d), 4

vD

Case (c) would cause leaving saturation if vDS < 2vOV

Vtp = −0.5 V

Ex: 5.4 For saturation v DS ≥ vOV , so vDS must be changed to 2v OV iD =

1 W 2 k v , so iD increases by a factor of 4. 2 n L OV

k p = 100 µA/V2 W = 10 ⇒ k p = 1 mA/V2 L (a) Conduction occurs for VSG ≥ |Vtp | = 0.5 V ⇒ vG ≤ 1.8 − 0.5 = +1.3 V

Ex: 5.5 vOV = 0.5 V gDS =

W k n L

v OV

1 = 1 k

Sedra, Smith, Carusone, Gaudet

(b) Triode region occurs for vDG ≥ |Vtp | = 0.5 V ⇒ vD ≥ vG + 0.5

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 5–2

1.8 V

(c) Conversely, for saturation vDG ≤ |Vtp | = 0.5 V ⇒ vD ≤ vG + 0.5

R

(d) Given λ ∼ = 0, iD =

VD

1 W k |vOV |2 = 50 µA 2 pL

∴ |vOV | = 0.32 V = vSG − |Vtp | ⇒ vSG = |vOV | + |Vtp | = 0.82 V vG = 1.8 − vSG = 0.98 V vD ≤ vG + 0.5 = 1.48 V (e) For λ = −0.2 V−1 and |vOV | = 0.32 V,

Saturation mode (v GD = 0 < Vtn ) :

1 = 100 k iD = 50 µA and ro = |λ|iD

W 1 μn Cox (VD − Vtn )2 = 0.032 mA 2 L VD = 0.7 V = 1.8 − ID R

(f) At vD = +1 V, vSD = 0.8 V,

ID =

1.8 − 0.7 = 34.4 k 0.032 mA

iD =

1 W k |vOV |2 (1 + |λ||vSD |) 2 nL = 50 µA (1.16) = 58 µA

∴R=

At vD = 0 V, vSD = 1.8 V,

Ex: 5.10

iD = 50 µA (1.36) = 68 µA

1.8 V

vDS 1V = 100 k = ro = iD 10 µA

R2

R

which is the same value found in (c).

Ex: 5.8 VDD − vD 1 − 0.2 = 8 k RD = = ID 0.1 W 2 1 ID = μn Cox VOV ⇒ 2 L 1 5 2 100 = × 400 × V ⇒ 2 0.4 OV VOV = 0.2 V ⇒ VGS = VOV + Vt = 0.2 + 0.4 = 0.6 V VS = −0.6 V ⇒ RS =

VS − VSS ID

−0.6 − (−1) 0.1 RS = 4 k

Q1

Q2

Since Q2 is identical to Q1 and their VGS values are the same, ID2 = ID1 = 0.032 mA For Q2 to operate at the triode–saturation boundary, we must have VD2 = VOV = 0.7 − 0.5 = 0.2 V ∴ R2 =

1.8 V − 0.2 V = 50 k 0.032 mA

=

Ex: 5.9 Vtn = 0.5 V μn Cox = 0.4 mA/V2 0.72 µm W = = 4.0 L 0.18 µm λ=0

Sedra, Smith, Carusone, Gaudet

Ex: 5.11 RD = 6.55 × 2 = 13.1 k VGS = 2 V, assume triode region: ⎫ W 1 2 ⎪ ⎪ ID = k n (VGS − Vtn )VDS − VDS ⎬ L 2 ⇒ ⎪ ⎪ VDD − VDS ⎭ ID = R 2 − VDS V2 = 2 × (2 − 0.5)VDS − DS 13.1 2 2 ⇒ VDS − 3.076VDS + 0.15 = 0

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 5–3

⇒ VDS = 0.05 V < VOV ⇒ triode region 2 − 0.05 = 0.15 mA ID = 13.1

Since VDG = 0, the transistor is operating in saturation, and 1 2 k p VOV = 1 mA 2 1.8 − VS 1.8 − 1 ∴ R= = 0.8 k = ID 1 ID =

Ex: 5.12 As indicated in Example 5.6, VD ≥ VG − Vt for the transistor to be in the saturation region.

= 800

VDmin = VG − Vt = 5 − 1 = 4 V VDD − VDmin ID

ID = 0.5 mA ⇒ RDmax = =

10 − 4 = 12 k 0.5

Ex: 5.13 ID = 0.32 mA =

1 W 2 1 2 k V = × 1 × VOV 2 n L OV 2

Ex: 5.15 VI = 0: since the circuit is perfectly symmetrical, VO = 0 and therefore VGS = 0, which implies that the transistors are turned off and IDN = IDP = 0. VI = 2.5 V: if we assume that the NMOS is turned on, then VO would be less than 2.5 V, and this implies that PMOS is off (VSGP < 0) . IDN =

⇒ VOV = 0.8 V

1 W k (VGS − Vt )2 2 nL

VGS = 0.8 + 1 = 1.8 V VG = VS + VGS = 1.6 + 1.8 = 3.4 V RG2

2.5 V

3.4 VG = = 3.4 M = I 1 µA

RG1 =

5 − 3.4 = 1.6 M 1 µA

QN

VD = 3.4 V, then RD =

IDN

VI 2.5 V

VS = 5 k RS = 0.32

10 k

5 − 3.4 = 5 k 0.32

VO RL

Ex: 5.14

1.8 V IDN =

R

ID

1 × 1(2.5 − VO − 1)2 2

IDN = 0.5(1.5 − VO )2 Also: VO = RL IDN = 10IDN IDN = 0.5(1.5 − 10IDN )2 2 ⇒ 100IDN − 32IDN + 2.25 = 0 ⇒ IDN

= 0.104 mA

IDP = 0

VO = 10 × 0.104 = 1.04V Vtp = −0.4 V k p

= 0.1 mA/V

VO 2

IDP

W 10 µm = ⇒ k p = 5.56 mA/V2 L 0.18 µm VSG = |Vtp | + |VOV |

QP 10 k

–2.5 V

= 0.4 + 0.6 = 1 V VS = +1 V

Sedra, Smith, Carusone, Gaudet

–2.5 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 5–4

VI = −2.5 V: Again if we assume that QP is turned on, then VO > −2.5 V and VGSN < 0, which implies that the NMOS QN is turned off. IDN = 0 Because of the symmetry, IDP = 0.104 mA, VO = −IDP × 10 k = −1.04 V

Sedra, Smith, Carusone, Gaudet

√ √ Ex: 5.16 Vt = 0.8 + 0.4 0.7 + 3 − 0.7 = 1.23 V Ex: 5.17 v DSmin = v GS + |Vt | =1+2=3V ID =

1 × 2 [1 − (−2)]2 2

= 9 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–1

Solutions to End-of-Chapter Problems 5.1 k n = μn Cox =

m2 F F C/V C 1 = = = V · s m2 V·s V·s s V2

=

A V2

ox 34.5 pF/m = tox 4 nm pF F = 8.625 × 10−3 2 or m µm2 (a) Cox =

k n = μn Cox = 388 µA/V2 (b) For

Since k n = k n W/L and W/L is dimensionless, k n has the same dimensions as k n ; that is, A/V2 .

2.4 W = , k n = 5.17 mA/V2 L 0.18

∴ 0.1 mA = iD =

1 k n v2OV 2

vOV = 0.2 V 5.2 The transistor size will be minimized if W/L is minimized. To start with, we minimize L by using the smallest feature size,

vGS = 0.7 V

L = 0.18 µm

(c) gDS =

rDS = rDS =

k n

1 (W/L) (v GS − Vt )

k n

1 (W/L) v OV

v GS = 0.89 V.

Condition 1:

= 250 ⇒ (W/L) v OV,1 = 10 Condition 2: 1 400 × 10 (W/L) v OV,2 −6

= 1000 ⇒ (W/L) v OV,2 = 2.5 If condition 1 is met, condition 2 will be met since the over-drive voltage can always be reduced to satisfy this requirement. For condition 1, we want to decrease W/L as much as possible (so long as it is greater than or equal to 1), while still meeting all of the other constraints. This requires our using the largest possible v GS,1 voltage. v GS,1 = 1.8 V so v OV,1 = 1.8 − 0.5 = 1.3 V, and W/L =

10 v OV,1

10 = 7.69 = 1.3

Condition 2 now can be used to find v GS,2 v OV,2 =

2.5 2.5 = = 0.325 W/L 7.69

⇒ v GS,2 = 0.825 V ⇒ 0.825 V ≤ v GS ≤ 1.8 V

⇒ vG = −0.9 V (b) For the p-channel transistor to operate in saturation, the drain voltage must not exceed the gate voltage by more than |Vtp |. Thus v Dmax = −0.9 + 0.5 = −0.4 V Put differently, vSD must be at least equal to |vOV |, which in this case is 0.4 V. Thus v Dmax = −0.4 V. (c) In (b), the transistor is operating in saturation, thus 1 iD = k p |vOV |2 2 1 0.2 = × k p × 0.42 2 ⇒ k p = 2.5 mA/V2 For vD = −20 mV, the transistor will be operating in the triode region. Thus 1 2 iD = k p v SD |vOV | − v SD 2 1 2 = 2.5 0.02 × 0.4 − (0.02) 2

For vD = −2 V, the transistor will be operating in saturation, thus

tox = 4 nm m2 cm2 = 450 × 10−4 V·s V·s

Sedra, Smith, Carusone, Gaudet

= 0.5 + 0.4 = 0.9 V

= 0.02 mA

5.3 Lmin = 0.18 µm

μn = 450

5.4 Vtp = −0.5 V (a) |vSG | = |Vtp | + |vOV |

1 = 400 × 10−6 (W/L)v OV,1

rDS,2 =

1 = k n vOV 100

∴ vOV = 0.39 V.

Two conditions need to met for v OV and rDS

rDS,1

v DS ≥ 0.2 V

iD =

1 1 k p |vOV |2 = × 2.5 × 0.42 = 0.2 mA 2 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–2

5.5 tox = 6 nm, μn = 460 cm2 /V·s, Vt = 0.5 V, and W/L = 10.

5.7 Cox = 25 fF/µm2 , vOV = 0.15 V L = 0.065 µm, vDS = 0 V

W 3.45 × 10−11 = 460×10−4 × ×10 k n = μn Cox L 6 × 10−9

W = 0.65 µm

= 2.645 mA/V2

Q = Cox .W.L.vOV = 0.16 fC

(a) v GS = 2.5 V

v DS = 1 V

and

v OV = v GS − Vt = 2 V Thus v DS < v OV ⇒ triode region, 1 ID = k n v DS v OV − v 2DS 2 1 = 2.645 1 × 2 − × 1 = 4 mA 2 (b) v GS = 2 V

v DS = 1.5 V

and

5.8 With v DS small compared to VOV , Eq. (5.13a) applies: rDS =

1 W μn Cox (vOV ) L

(a) vOV is doubled → rDS is halved. Factor = 0.5 (b) W is doubled → rDS is halved. Factor = 0.5

v OV = v GS − Vt = 2 − 0.5 = 1.5 V

(c) W and L are doubled → rDS is unchanged. Factor = 1.0

Thus, v DS = v OV ⇒ saturation region,

(d) If oxide thickness tox is halved. Since

iD =

1 1 k n v 2OV = × 2.645 × 1.52 2 2

= 3 mA (c) v GS = 2.5 V

v DS = 0.2 V

and

Cox =

ox tox

then Cox is doubled. If W and L are also halved, rDS is halved. Factor = 0.5.

v OV = 2.5 − 0.5 = 2 V

5.9 k n = 10 mA/V2 , Vtn = 0.4 V,

Thus, v DS < v OV ⇒ triode region, 1 iD = k n v DS v OV − v 2DS 2

small v DS

1 = 2.645[0.2 × 2 − 0.22 ] = 1 mA 2 (d) v GS = v DS = 2.5 V

Thus, v DS > v OV ⇒ saturation region,

=

gDS =

1 = k n v OV rDS iD vGS 1.2 V

400 A

v OV = 2.5 − 0.5 = 2 V

iD =

iD = k n (v GS − Vt )v DS = k n v OV v DS

1 k n v 2OV 2

vGS 1.0 V

300 A

vGS 0.8 V

200 A

vGS 0.6 V

100 A

1 × 2.645 × 22 = 5.3 mA 2

vGS 0.4 V vDS 50 mV

0

5.6 tox = 2 ∼ 10 nm ox Cox = tox

This table belongs to Exercise 5.9.

ox = 34.5 pF/m

−1 = 58 ∼ 290 m2 /F Cox

µm pF

2

VGS VOV (V) (V)

gDS rDS (mA/V) ()

0.4

0

0

∞

0.6

0.2

2.0

500

0.8

0.4

4.0

250

so

1.0

0.6

6

167

d = 24 ∼ 54 µm

1.2

0.8

8

125

For 10 pF:

Area = 580 ∼ 2900 µm

2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–3

5.10 tox = 1.4 nm, Vt = 0.35 V

⇒ VDS sat = 0.5 V

Lmin = 65 nm, small v DS ,

In saturation: W 1 1 2 2 = k n VOV VOV iD = k n 2 L 2

k n = 540 µA/V2 , 0 < v GS < 1.0 V.

rDS = kn

W L

1

The lowest rDS is obtained at the largest vGS , i.e., vGS = 1 V, thus

100 = 540 × 10−6

1 W (1 − 0.35) L

W = 28.5 ⇒ L

∂iD ∂v DS v DS

For equal drain currents:

If VDS = 0.5VOV ⇒ rds =

1 k n (VOV − 0.5VOV )

=

vDS = iD

1 W k n (vGS − Vt ) L

1 400 × 10−6 × 20 × (1.8 − 0.5)

rDS = 96.2

vDS = rDS × iD = 19.2 mV

Wp = 80 L

5.15 For triode-region operation with v DS small,

Thus k n (VOV

1 − 0.8VOV )

5 k n VOV

rDS ≡ 1=

If VDS = VOV ,

v DS 1 = iD k n (v GS − Vt )

1 1 = k n (1.2 − 0.8) 0.4 k n

⇒ k n = 2.5 mA/V

1 ⇒∞ = 0

rDS =

1 2.5(VGS − 0.8)

5.12 VDS sat = VOV

0.2 =

1 2.5(VGS − 0.8)

VOV = VGS − Vt = 1 − 0.5 = 0.5 V

⇒ VGS = 2.8 V

rds

W (vGS − Vt )vDS , L

iD k n (v GS − Vt )v DS

2 k n VOV

If VDS = 0.8VOV ⇒ rds = = 1/k n (0.2VOV ) =

rDS =

= 20 × 4 ⇒

1.25 ⇒ rds = k n VOV

= 1/k n (0.5VOV ) =

Wn Wp = μp Cox L L

Wn μ Wp Wp = ×4 = n =4⇒ Wn μp L L

1 = 0 ⇒ rds = k n VOV

If VDS = 0.2VOV

k n = μn Cox

For the same performance of a p-channel device:

1 k n (VOV − VDS )

If VDS

1 W k |VOV |2 2 nL

5.14 For small v DS , iD k n

= VDS

−1 ∂ 1 2 k n VOV v DS − v DS = ∂v DS 2 ∂ ∂ 2 −1 v DS = kn (VOV v DS ) − 1/2k n ∂v DS ∂v DS −1 1 = k n VOV − · 2VDS 2 =

5.13 iD =

μ 1 Wp = 2.5 = n = Wn μp 0.4

= 1.85 µ m

iD = 0.5 mA

μn Cox

⇒ W = 28.5 × 0.065

5.11 rds = 1/

2 1 4 mA × × (0.5 V) 2 2 V

iD =

(vGS − Vt )

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

(k)

© Oxford University Press 2021

Chapter 5–4

For a device with twice the value of W, k n will be twice as large and the resistance values will be half as large: 500 and 100 , respectively. 5.16 Vtn = 0.5 V, ID = 0.05 =

k n = 1.6 mA/V2

1 2 × 1.6 × VOV 2

W = 20 L For v GS = 2 V and v DS = 0.1 V, 1 iD = k n (v GS − Vt )v DS − v 2DS 2 1 = 1 (2 − 0.5) × 0.1 − × 0.12 2

⇒ VOV = 0.25 V and VDS ≥ 0.25 V

= 0.145 mA = 145 µA

VGS = 0.5 + 0.25 = 0.75 V

For v GS = 2 V, pinch-off will occur for

ID = 0.2 =

1 2 × 1.6 × VOV 2

v DS = v GS − Vt = 2 − 0.5 = 1.5 V

⇒ VOV = 0.5 V and VDS ≥ 0.5 V

and the resulting drain current will be

VGS = 0.5 + 0.5 = 1 V

iD =

5.17 For VGS = VDS = 1 V, the MOSFET is operating in saturation, ID =

1 k n (VGS − Vt )2 2

0.4 =

1 k n (1 − Vt )2 2

=

1 k n (v GS − Vt )2 2

1 × 1 × (2 − 0.5)2 2

= 1.125 mA

(1)

5.19

VDD

1 0.1 = k n (0.8 − Vt )2 2

(2)

Dividing Eq. (1) by Eq. (2) and taking square roots gives 2=

1 − Vt 0.8 − Vt

iD

⇒ Vt = 0.6 V

vS

Substituting in Eq. (1), we have 0.4 =

1 k n × 0.42 2 v GD = 0 ⇒ saturation

⇒ k n = 5 mA/V2

iD =

5.18 iD = k n (v GS − Vt )v DS 25 = k n (1 − Vt ) × 0.05

(1)

50 = k n (1.5 − Vt ) × 0.05

(2)

Dividing Eq. (2) by Eq. (1), we have 2=

1.5 − Vt 1 − Vt

1 k n (v GS − Vt )2 2

v GS = VDD − v S ∴ iD =

1 k n [(VDD − Vt ) − v S ]2 2

0 ≤ v S ≤ (VDD − Vt ) iD = 0, v S ≥ (VDD − Vt )

⇒ Vt = 0.5 V Substituting in Eq. (1) yields 25 = k n × 0.5 × 0.05 ⇒ k n = 1000 µA/V2 For k n = 50 µA/V2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–5

5.20 VDS = VD – VS

W is kept at 10) L VA = ro iD = 20 k × 0.1 mA = 2 V (for the standard device)

VGS = VG − VS

W = 5 × 10 × 0.18 = 9 µm (so

VOV = VGS − Vt = VGS − 1.0 According to Table 5.1, three regions are possible.

Case

a

VS

VG

VD

+1.0 +1.0 +2.0

VA = 5 × 2 = 10 V (for the new device)

VGS VOV VDS Region of operation –1.0 +1.0

0

Cutoff

5.23

NMOS

1

2 −1

3 −1

4 −1

0.2 V−1

b

+1.0 +2.5 +2.0 +1.5 +0.5 +1.0

Sat.

λ

0.05 V

c

+1.0 +2.5 +1.5 +1.5 +0.5 +0.5

Sat.

VA

20 V

10 V

20 V

5V

d

+1.0 +1.5

Sat.∗

ID

0.5 mA

0.4 mA

0.2 mA

0.01 mA

Triode

ro

40 k

25 k

100 k

500 k

e

0

+0.5 –0.5 –1.0

+2.5 1.0 +2.5 +1.5 +1.0

0

f

+1.0 +1.0 +1.0

g

–1.0

0

h

–1.5

0

i

–1.0

0

j

+0.5 +2.0 +0.5 +1.5 +0.5

0

–1.0

0

Cutoff

0

+1.0

0

+1.0

Sat.

0

+1.5 +0.5 +1.5

Sat.

+1.0 +1.0

0

+2.0

Sat.

0

Triode

0.1 V

0.05 V

5.24 PMOS with Vtp = –1 V

Case

VS

VOV VSD

Region of

VG

VD

VSG

a

+2 +2

0

0

0

2

Cutoff

b

+2 +1

0

+1

0

2

Cutoff–Sat.

operation

∗ With the source and drain interchanged: V = 0, V = S G

+1.5, VD = +1, VGS = +1.5, VOV = 0.5, VDS = +1

5.21

iD

c

+2

0

0

+2

1

2

Sat.

d

+2

0

+1

+2

1

1

Sat–Triode

e

+2

0

+1.5

+2

1

0.5

Triode

f

+2

0

+2

+2

1

0

Triode

5.25

0.205 mA 0.200 mA

2 V

vDS 1V

ro =

1.5 V

0.8 V

v DS 0.5 = 100 k = iD v GS const. 0.005

VA ∼ = ID ro = 0.2 × 100 = 20 V λ=

vG

1 = 0.05 V−1 VA

5.22 VA = dependent.

v G = +2 V → 0 V

VA L,

where VA is completely process VA Also, ro = . Therefore, to achieve iD

desired ro (which is 5 times larger), we should increase L (L = 5 × 0.18 = 0.9 µm). To keep iD unchanged, the unchanged. Therefore:

Sedra, Smith, Carusone, Gaudet

Vtp = −0.4 V

W ratio must stay L

As v G reaches +1.6 V, the transistor begins to conduct and enters the saturation region, since v DG will be negative. The transistor continues to operate in the saturation region until v G reaches +0.4 V, at which point v DG will be 0.4 V, which is equal to |Vtp |, and the transistor enters the triode region. As v G goes below +0.4 V, the transistor continues to operate in the triode region.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–6

This figure belongs to 5.26, part (a).

Fig. 1 1 5.26 iD = k n v OV v DS − v 2DS 2 iD 1 = v OV v DS − v 2DS kn 2

iD /kn (V2) (1)

Figure 1 shows graphs for iD /k n versus v DS for various values of v OV . Since the right-hand side of Eq. (1) does not have any MOSFET parameters, these graphs apply for any n-channel MOSFET with the assumption that λ = 0. They also apply to p-channel devices with v DS replaced by v SD , k n by k p , and v OV with |v OV |. The slope of each graph at v DS = 0 is found by differentiating Eq. (1) relative to v DS with v OV = VOV and then substituting v DS = 0. The result is d(iD /k n ) = VOV dv DS v DS =0, v OV =VOV Figure 1 shows the tangent at v DS = 0 for the graph corresponding to v OV = VOV3 . Observe that 1 2 at it intersects the horizontal line iD /k n = VOV3 2 1 v DS = VOV3 . Finally, observe that the curve 2 representing the boundary between the triode region and the saturation region has the equation iD /k n =

1 2 v 2 DS

Figure 2 shows the graph for the relationship iD /k n =

1 2 v 2 OV

which describes the MOSFETs operation in the saturation region, that is, v DS ≥ v OV

Sedra, Smith, Carusone, Gaudet

Slope VOV

1 2 V 2 OV

1 iD /kn v 2OV 2

0 1 V VOV 2 OV Fig. 2

vOV (V)

Here also observe that this relationship (and graph) is universal and represents any MOSFET. The slope at v OV = VOV is d(iD /k n ) = VOV d v OV v OV =VOV Replacing k n by k p and v OV by |v OV | adapts this graph to PMOS transistors. 5.27 k n = 0.4 mA/V2

and

Vt = 0.5 V

For v GS = v DS = 1.8 V, the MOSFET is operating in saturation. Thus, to obtain iD = 2 mA, we write W 1 × (1.8 − 0.5)2 2 = × 0.4 × 2 L ⇒

W = 5.92 L

For L = 0.18 µm W = 1.07 µm

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–7

5.28 For the channel to remain continuous, v DS ≤ v GS − Vt Thus for v GS = 1.0 V to 1.8 V and Vt = 0.5, v DS ≤ 1 − 0.5 That is, v DSmax = 0.5 V. W 20 = = 20 k n = 100 µA/V2 L 1 W = 100 × 20 = 2000 µA/V2 k n = k n L

5.29

incremental resistance r at the bias point can be obtained by differentiating Eq. (1) relative to v and then substituting v = |Vt | + |VOV | as follows: W ∂i = k (v − |Vt |) ∂v L ∂i W VOV = k ∂v v =|Vt |+VOV L W ∂i =1 k VOV Q.E.D r=1 ∂v L 5.31

= 2 mA/V2 For operation as a linear resistance,

iD

iD = k n (v GS − Vt )v DS

vDS

and rDS =

v DS 1 ≡ = iD k n (v GS − Vt )

1 2(v GS − 0.8)

At v GS = 1.0 V, rDS =

v DS = v GS

1 = 2.5 k 2(1 − 0.8)

rDS =

1 k n (v DS − Vt )2 2

2iD ∴ v DS = + Vt kn iD =

At v GS = 4.8 V, 1 = 0.125 k 2(4.8 − 0.8)

Thus, rDS will vary in the range of 2.5 k to 125 . (a) If W is halved, k n will be halved and rDS will vary in the range of 5 k to 250 . (b) If L is halved, k n will be doubled and rDS will vary in the range of 1.25 k to 62.5 . (c) If both W and L are halved, k n will remain unchanged and rDS will vary in the original range of 2.5 k to 125 . 5.30 (a) Refer to Fig. P5.30. For saturation-mode operation of an NMOS transistor, v DG ≥ −Vtn ; thus v DG = 0 results in saturation-mode operation. Similarly, for a p-channel MOSFET, saturation-mode operation is obtained for v GD ≥ −|Vtp |, which includes v GD = 0. Thus, the diode-connected MOSFETs of Fig. P5.30 have the i−v relationship W 1 (v − |Vt |)2 (1) i = k 2 L where k represents k n in the NMOS case and k p in the PMOS case. (b) If either of the MOSFETs in Fig. P5.30 is biased to operate at v = |Vt | + |VOV |, then its

Sedra, Smith, Carusone, Gaudet

/kn

5.32 The cutoff–saturation boundary is determined by v GS = Vt , thus v GS = 0.4 V at the boundary (see figure on next page). The saturation–triode boundary is determined by v GD = Vt , and v DS = VDD = 1 V, and since v GS = v GD + v DS , one has v GS = 0.4 + 1.0 = 1.4 V at the boundary. 5.33 (a) Let Q1 have a ratio (W/L) and Q2 have a ratio 1.03 (W/L). Thus W 1 (1 − Vt )2 ID1 = k n 2 L

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–8

This figure belongs to Problem 5.32.

ID + ID = ID (1 + λ × 2) = 1.2 ID where ID is the drain current without channel-length modulation taken into account. Thus 100 ID = 1.1 and 1.2 × 100 = 109 µA 100 + ID = 1.2 ID = 1.1 ⇒ ID = 9 µA or 9%

ID2 =

1 W k 2 n L

× 1.03 × (1 − Vt )2

Thus, ID2 = 1.03 ID1 That is, a 3% mismatch in the W/L ratios results in a 3% mismatch in the drain currents.

To reduce ID by a factor of 2, we need to reduce λ by a factor of 2, which can be obtained by doubling the channel length to 2 µm. 5.36 VA = VA L = 5 × 0.54 = 2.7 V

(b) Let Q1 have a threshold voltage Vt = 0.6 V and Q2 have a threshold voltage Vt + Vt = 0.6 + 0.01 = 0.61 V.

1 1 = 0.37 V−1 = VA 2.7 W 1 v2OV (1 + λvDS ) iD = k n 2 L 5.4 1 × 0.252 (1 + 0.37 × 1) = × 0.4 × 2 0.54

Thus

= 0.17 mA

ID1

ro =

ID2

W 1 (1 − 0.6)2 = k n 2 L W 1 (1 − 0.61)2 = k n 2 L

VA 1 W 2 k vOV 2 n L 2.7 = 1 5.4 × 0.252 × 0.4 × 2 0.54

and (1 − 0.61)2 ID2 = = 0.95 ID1 (1 − 0.6)2 That is, a 10-mV mismatch in the threshold voltage results in a 5% mismatch in drain currents. VA 20 = , 0.1 mA ≤ iD ≤ 1 mA 5.34 ro = iD iD ⇒ 20 k ≤ ro ≤ 200 k v DS v DS 1 ⇒ iD = = iD ro ro iD At iD = 0.1 mA, iD = 5 µA, = 5% iD iD = 5% At iD = 1 mA, iD = 50 µA, iD ro =

5.35 L = 1.0 µm = 5× minimum. Thus 0.5 V−1 = 0.1 V−1 5 If v DS is increased from 1 V to 2 V, the drain current will change from

λ=

ID = 100 µA = ID (1 + λ × 1) = 1.1 ID to

Sedra, Smith, Carusone, Gaudet

λ=

= 21.6 k vDS 0.5 V = 0.023 mA = iD = ro 21.6 k 5.37 Quadrupling W and L keeps the current iD unchanged. However, the quadrupling of L increases VA by a factor of 4 and hence increases ro by a factor of 4. Halving VOV results in decreasing iD by a factor of 4. Thus, this alone increases ro by a factor of 4. The overall increase in ro is by a factor of 4 × 4 = 16.

5.38 Refer to the circuit in Fig. P5.33 and let VD1 = 2 V and VD2 = 2.5 V. If the two devices are matched, 1 2 2 ID1 = k n (1 − Vt ) 1 + 2 VA 1 2.5 2 ID2 = k n (1 − Vt ) 1 + 2 VA 0.5 1 ID = ID2 − ID1 = k n (1 − Vt )2 2 VA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–9 ID 0.5 0.01 = 1 VA 2 k n (1 − Vt ) 2 ⇒ VA = 50 V (or larger to limit the mismatch in ID to 1%). If VA = 100 V/µm, the minimum required channel length is 0.5 µm.

v GS = −1 V v SG = +1 V v DS = −1.6 V v SD = +1.6 V Vtp = −0.4 V VA = −4 V

5.39

k p = k p

W L

λ = −0.25 V−1

= 100 µA/V2

Vtp = −0.5 V λ = −0.1 V−1 vG = 0,

vS = +1 V ⇒ vSG = 1 V

|vOV | = vSG − |Vtp | = 1 − 0.5 = 0.5 • For v D = +0.8 V, v SD = 0.2 V < |vOV | ⇒ triode-region operation, 1 2 iD = k p v SD |vOV | − v SD 2 1 = 100 0.2 × 0.5 − × 0.22 = 8 µA 2 • For v D = +0.6 V, v SD = 0.4 V < |vOV | ⇒ triode-region operation, 1 2 iD = k p v SD |vOV | − v SD 2 1 = 100 0.4 × 0.5 − × 0.42 = 12 µA 2 • For v D = +0.5 V, v SD = 0.5 V = |VOV | ⇒ saturation-mode operation, iD = =

1 k p |vOV |2 (1 + |λ|v SD ) 2

1 × 100 × 0.25(1 + 0.1 × 0.5) = 13.13 µA 2

• For v D = 0 V, v SD = 1 V > |vOV | ⇒ saturation-mode operation, iD =

1 × 100 × 0.25(1 + 0.1 × 1) = 13.75 µA 2

• For v D = −1 V, v SD = 2 V > |vOV | ⇒ saturation-mode operation, iD =

1 × 100 × 0.25(1 + 0.1 × 2) = 15 µA 2

5.40 |Vtp | = 0.4 V, |VA | = 4 V |v GS | = 1 V, |v DS | = 1.6 V

2=

1 k p [−1 − (−0.4)]2 (1 − 0.25 × −1.6) 2

⇒ k p = 7.94 mA/V2 5.41 Case (a): Since VGS > 0 and VDS > 0, transistor 1 must be NMOS. Also, since in both situations VDG is positive, the NMOS transistor is operating in saturation and we can obtain Vt from (1 − Vt )2 100 = ⇒ Vt = 0.5 V (1.5 − Vt )2 400 To find kn we write 100 =

1 kn (1−0.5)2 ⇒ kn = 800 µA/V2 2

Case (b): Here VSG and VSD are both greater than zero, thus transistor 2 must be PMOS. Also, since in both situations VGD is positive, the PMOS transistor must be operating in saturation and we can find | Vt | from, (2− | Vt |)2 50 ⇒| Vt |= 1.5 V = (3− | Vt |)2 450 To find kp we write 50 =

1 kp (2−1.5)2 ⇒ kp = 400 µA/V2 2

Case (c): Since VSG and VSD are both positive, transistor 3 must be PMOS. At the second operating point, VGD is positive and thus the transistor is operating in saturation. However, at the first operating point, VGD is −1 V and we cannot tell whether the transistor is in saturation or in the triode region. Assuming saturation-mode operation in both cases, we can find | Vt | from (2− | Vt |)2 200 ⇒| Vt |= 1 V = (3− | Vt |)2 800

iD = 2 mA |vOV | = |v GS | − |Vtp | = 0.6 V |v DS | > |vOV | ⇒ saturation mode

Sedra, Smith, Carusone, Gaudet

1 k p (v GS − Vtp )2 (1 + λ v DS ) 2

iD =

We thus see that at the first operating point VDG =| Vt | confirming operation in saturation, as already assumed (actually, at the boundary

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–10

This table belongs to Problem 5.41. VS

VG

VD

ID

(V) (V)

(V)

(µA)

Case Transistor

a

1

b

2

c

3

d

4

0

1

2.5

100

0

1.5

2.5

400

5

3

–4.5

50

5

2

–0.5

450

5

3

4

200

5

2

0

800

–2

0

0

72

–4

0

–3

270

Type

NMOS

PMOS

PMOS

NMOS

72 (2 − Vt )2 ⇒ Vt = −0.136 = (4 − Vt )2 270 Since Vt must be positive, we conclude that our assumption of saturation-mode operation in the second case is not correct, and the transistor must be operating in triode. Thus (4 − Vt ) × 1 −

1 2

×

12

=

72 ⇒ Vt = 0.8 V or 2.7 V 270

The second solution does not make physical sense as it results in ID = 0 at the first operating point. Thus Vt = 0.8 V

0.5

400

–1.5

400

–1

100

+0.8

Sat.

Sat.

Sat. Triode

(c) V3 = VSD = VSG = 1 V (d) V4 = +1.25 − VSG = 1.25 − 1 = 0.25 V Now place a resistor R in series with the drain. For the circuits in (a) and (b) to remain in saturation, VD must not fall below VG by more than Vt . Thus, IR ≤ Vt Rmax =

0.5 Vt = = 5 k I 0.1

For the circuits in (c) and (d) to remain in saturation, VD must not exceed VG by more than |Vt |. Thus IR ≤ |Vt | which yields Rmax = 5 k. Now place a resistor RS in series with the MOSFET source. The voltage across the current source becomes (a) VCS = 2.5 − VDS − IRS

(1)

To keep VCS at least at 0.5 V, the maximum RS can be found from 0.5 = 2.5 − 1 − 0.1 × RSmax

and kn can be found from 72 =

800

Sat.

Sat.

Case (d): Since VGS > 0 and VDS > 0, transistor 4 must be NMOS. At the first operating point, VDG = 0 and the transistor must be in saturation. At the second operating point, VDG is negative by 3 V and it is not clear whether the transistor is in saturation or in triode region. Assuming saturation in both cases,

− Vt )2

(V)

Vt

Sat.

1 kp (2−1)2 ⇒ kp = 400 µA/V2 2

1 (2 2

W L (µ A/V2 ) μCox

Sat.

between saturation and triode). We can find kp by writing 200 =

Mode

1 kn (2−0.8)2 ⇒ kn = 100 µA/V2 2

The results are provided in the filled-out table above.

⇒ RSmax = 10 k V1 = 2.5 − 0.5 = 2 V (b) VCS = 1 − VDS − IRS − (−1.5) = 2.5 − VDS − IRS

5.42 Refer to the circuits in Fig. P5.42.

which is identical to Eq. (1). Thus

(a) V1 = VDS = VGS = 1 V

RSmax = 10 k

(b) V2 = +1 − VDS = 1 − 1 = 0 V

V2 = −1.5 + 0.5 = −1 V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–11

VDD 1.8 V

(c) VCS = 2.5 − IRS − VSD which yields

ID2

RSmax = 10 k

ID1

R2

V3 = 2.5 − 0.5 = 2 V

R

VD2

(d) VCS = 1.25 − IRS − VSD − (−1.25)

Q2

= 2.5 − VSD − IRS

VGS

which yields

Q1

VD1

RSmax = 10 k V4 = −1.25 + 0.5 = −0.75 V

= 0.5 + 0.25 = 0.75 V VD1 = VGS1 = 0.75 V

5.43

R=

1 V ID 0.1 mA RD

(b) Note that both transistors operate at the same VGS and VOV , and ID2 = 0.5 mA

VD 0.2 V

But ID2

0.1 mA

VS RS

W 2 1 ID = μn Cox VOV 2 L

and

VGS = Vt + VOV = 0.5 + 0.25 = 0.75 VS = 0 − VGS = −0.75 V

0.8 1 − VD 1 − 0.2 = = 8 k RD = = ID 0.1 0.1

1.44 2 1 × 0.4 × V 2 0.36 OV

R2 = =

VS − (−1) −0.75 + 1 = 2.5 k = ID 0.1

0.05 =

VDS2 = VOV

VD2 = 0.25 V

⇒ VOV = 0.25 V

(a) ID1 = 50 µA

W2 1 × 0.4 × × 0.252 2 0.36

Thus,

4 1 2 × 0.4 × × VOV 2 0.5

5.44

0.5 =

which is 10 times W1 , as needed to provide ID2 = 10ID1 . Since Q2 is to operate at the edge of saturation,

Since VDG > 0, the MOSFET is in saturation.

RS =

W2 1 2 VOV = kn 2 L2

⇒ W2 = 14.4 µm

1 V

0.1 =

1.8 − 0.75 VDD − VD1 = 21 k = ID1 0.05

VDD − VD2 ID2

1.8 − 0.25 = 3.1 k 0.5

5.45 Refer to Fig. P5.45. Both Q1 and Q2 are operating in saturation at ID = 0.1 mA. For Q1 , ID = 0.1 = ⇒

W1 2 1 μn Cox V 2 L1 OV1 W1 1 × 0.4 × (0.6 − 0.4)2 2 L1

W1 = 12.5 L1

⇒ VOV = 0.25 V

W1 = 12.5 × 0.18 = 2.25 µm

VGS1 = Vt + VOV

For Q2 , we have

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–12 W2 1 2 μn Cox VOV2 2 L2

ID = 0.1 = ⇒

Refer to Fig. P5.47(b): The transistor is operating in saturation, thus

1 W2 (1.4 − 0.6 − 0.4)2 × 0.4 × 2 L2

W2 = 3.125 L2

0.1 =

1 2 ⇒ VOV = 0.2 V × 5 × VOV 2

VGS = 0.7 V

W2 = 3.125 × 0.18 = 0.56 µm R=

1 2 k n VOV 2

ID =

1.8 − 1.4 = 4 k 0.1

⇒ V3 = 0.7 V Refer to Fig. P5.47(c): Assuming saturation-mode operation, we have

5.46

ID

1 k p |VOV |2 2

ID =

VDD

0.1 =

R

1 × 5 × |VOV |2 2

⇒ |VOV | = 0.2 V

VO vI

VSG = |Vt | + |VOV | = 0.5 + 0.2 = 0.7 V V4 = VSG = 0.7 V V5 = −1 + ID × 10 = −1 + 0.1 × 10 = 0 V

Assuming linear operation in the triode region, we can write

Refer to Fig. P5.47(d): Both transistors are operating in saturation at equal |VOV |. Thus

VO 50 mV = 1 mA = rDS 50 W (VGS − Vt )VDS ID = k n L ID =

1 = 0.5 × ⇒

0.1 =

W × (1.3 − 0.4) × 0.05 L

VSG = |Vt | + |VOV | = 0.7 V V6 = 1 − VSG = 1 − 0.7 = 0.3 V

(b) Circuit (a): The 0.1-mA current source can be replaced with a resistance R connected between the MOSFET source and the −1-V supply with

VDD − VO 1.3 − 0.05 = ID 1

= 1.25 k

R= 5.47 (a) Refer to Fig. P5.47(a): Assuming saturation-mode operation, we have ID = 0.1 =

1 × 5 × |VOV |2 ⇒ |VOV | = 0.2 V 2

V7 = +1 − 2 VSG = 1 − 2 × 0.7 = −0.4 V

W = 44.4 L

R=

Since VDG = 0, the MOSFET is indeed in saturation.

1 2 k n VOV 2

−0.7 + 1 V1 − (−1) = = 3 k 0.1 mA 0.1

We use the nearest 1% resistor which is 3.01 k. Circuit (b): The 0.1-mA current source can be replaced with a resistance R,

1 2 × 5 VOV 2

R=

1 − 0.7 1 − V3 = = 3 k 0.1 mA 0.1

⇒ VOV = 0.2 V

We use the nearest 1% resistor which is 3.01 k.

VGS = |Vt | + VOV = 0.5 + 0.2 = 0.7 V

Circuit (c): The 0.1-mA current source can be replaced with a resistance R,

V1 = 0 − VGS = −0.7 V

1 − 0.7 1 − V4 = = 3 k 0.1 mA 0.1

V2 = 1 − 0.1 × 10 = 0 V

R=

Since VDG = 0 V, the MOSFET is indeed in saturation.

We use the nearest 1% resistor which is 3.01 k.

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–13

Circuit (d): The 0.1-mA current source can be replaced with a resistance R, R=

0.6 V7 − (−1) = = 6 k 0.1 mA 0.1

We use the nearest 1% resistor, which is 6.04 k.

of the NMOS transistor. This compensates for the factor 4 in the process transconductance parameter, resulting in k p = k n , and the two transistors are matched. The solution will be identical to that for (a) above with V5 =

2 =1V 2

I6 = 360 µA 5.48 (a) Refer to the circuit in Fig. P5.48(a). Since the two NMOS transistors are identical and have the same ID , their VGS values will be equal. Thus VGS =

5.49 Refer to the circuit in Fig. P5.49. (a) Q1 and Q2 are matched. Thus, from symmetry, we see that the 200-µA current will split equally between Q1 and Q2 :

2 =1V 2

V2 = 1 V

ID1 = ID2 = 100 µA

VOV = VGS − Vt = 0.6 V W 1 2 VOV I1 = ID = μn Cox 2 L

V1 = V2 = 2.5 − 0.1 × 20 = 0.5 V To find V3 , we determine VGS of either Q1 and Q2 (which, of course, are equal), W 1 ID1 = μn Cox (VGS − Vt )2 2 L 1

1 1 = × 400 × × 0.62 2 0.2 = 360 µA (b) Refer to the circuit in Fig. P5.48(b). Here QN and QP have the same ID = I3 . Thus W 1 2 VOVN (1) I3 = μn Cox 2 L W 1 2 VOVP (2) I3 = μp Cox 2 L

100 =

1 × 125 × 20 × (VGS − 0.7)2 2

⇒ VGS = 0.983 V Thus, V3 = −0.983 V

Equating Eqs. (1) and (2) and using 2 2 = VOVP : μn Cox = 4μp Cox gives 4VOVN

(b) With VGS1 = VGS2 , but (W/L)1 = 1.5(W/L)2 , transistor Q1 will carry a current 1.5 times that in Q2 , that is,

|VOVP | = 2 VOVN

ID1 = 1.5ID2

Now,

But,

VGSN = VOVN + Vt = VOVN + 0.4

ID1 + ID2 = 200 µA

VSGP = |VOVP | + |Vt | = 2 VOVN + 0.4

Thus

But

ID1 = 120 µA

VSGP + VGSN = 2

ID2 = 80 µA

3VOVN + 0.8 = 2

V1 = 2.5 − 0.12 × 20 = 0.1 V

⇒ VOVN = 0.4 V

V2 = 2.5 − 0.08 × 20 = 0.9 V

|VOVP | = 0.8 V

To find V3 , we find VGS from the ID equation for either Q1 or Q2 , W 1 (VGS − Vt )2 ID1 = μn Cox 2 L

VGSN = 0.8 V VSGP = 1.2 V V4 = VGSN = 0.8 V I3 =

1

1 1 × 400 × × 0.42 = 160 µA 2 0.2

(c) Refer to Fig. P5.48(c). Here the width of the PMOS transistor is made 4 times larger than that

Sedra, Smith, Carusone, Gaudet

1 120 = × 125 × 20 × (VGS − 0.7)2 2 ⇒ VGS = 1.01 V V3 = −1.01 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–14

5.50

W 1 × 0.4 × (1.3 − 0.4)2 2 L W = 0.162 L =

1 V ID RD 0.2 V

W RD VD = 1.3 − ID RD = 1.3 − 0.162 L For the MOSFET to be at the edge of saturation, we must have

VGS 0.6 V

0.6 V

RS

VD = VOV = 1.3 − 0.4 = 0.9 Thus

W RD 0.9 = 1.3 − 0.162 L W RD 2.5 k ⇒ Q.E.D L

1 V

Since VDG > 0, the MOSFET is operating in saturation. Thus ID =

1 W k (VGS − Vt )2 2 nL

ID =

5.52

1 k n (VGS − Vt )2 2

1.3 V

1 = × 4 × (0.6 − 0.4)2 2

RD

= 0.08 mA RD =

0.8 1 − VD 1 − 0.2 = = 10 k = ID 0.08 0.08

RS =

−0.6 − (−1) −0.6 + 1 = 5 k = ID 0.08

For ID to remain unchanged from 0.08 mA, the MOSFET must remain in saturation. This in turn can be achieved by ensuring that VD does not fall below VG (which is zero) by more than Vt (0.4 V). Thus

VOV = VGS − Vt = 1.3 − 0.4 = 0.9

1 − ID RDmax = −0.4 RDmax =

0.1 mA VD

To operate at the edge of saturation, we must have

1.4 = 17.5 k 0.08

VD = VOV = 0.9 V Thus,

5.51

RD =

1.3 − 0.9 = 4 k 0.1

1.3 V 5.53

RD ID

ID = 180 µA R=

and

VD = 1 V

1 VD = 5.6 k = ID 0.18

Transistor is operating in saturation with |VOV | = 1.8 − VD − |Vt | = 1.8 − 1 − 0.5 = 0.3 V: ID =

Sedra, Smith, Carusone, Gaudet

1 W k |VOV |2 2 pL

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–15

VDD 1.8 V

ID = =

VD

ID R

1 W k (VGS − Vt )2 2 nL

1 × 1.5 × (5 − 6ID − 1.5)2 2

which results in the following quadratic equation in ID : 36ID2 − 43.33ID + 12.25 = 0 The physically meaningful root is ID = 0.454 mA

180 = ⇒

1 W × 100 × × 0.32 2 L

This should be compared to the value of 0.5 mA found in Example 5.6. The difference of about 10% is relatively small, given the large variations in k n and Vt (50% increase in each). The new value of VD is

W = 40 L

W = 40 × 0.18 = 7.2 µm

VD = VDD − RD ID = 10 − 6 × 0.454 = +7.28 V 5.54 Refer to the circuit in Fig. P5.54. All three transistors are operating in saturation with ID = 90 µA. For Q1 , ID =

W1 1 μn Cox (VGS1 − Vt )2 2 L1

1 W1 90 = × 90 × (0.8 − 0.5)2 2 L1

as compared to +7 V found in Example 5.6. We conclude that this circuit is quite tolerant to variations in device parameters.

5.56

VDD 10 V

W1 ⇒ = 22.2 L1

1 A

W1 = 22.2 × 0.5 = 11.1 µm

0.5 mA

RG1

RS

For Q2 ,

VS VG 6 V

W2 1 ID = μn Cox (VGS2 − Vt )2 2 L2

5 V

1 W2 90 = × 90 × (1.5 − 0.8 − 0.5)2 2 L2 ⇒

RD

RG2

0.5 mA

W2 = 50 L2

W2 = 50 × 0.5 = 25 µm Refer to the circuit in the figure above.

For Q3 , ID =

W3 1 μn Cox (VGS3 − Vt )2 2 L3

RG1 =

90 =

W3 1 × 90 × (2.5 − 1.5 − 0.5)2 2 L3

=

⇒

W3 =8 L3

W3 = 8 × 0.5 = 4 µm 5.55 Refer to the circuits in Fig. 5.24: VGS = 5 − 6ID

Sedra, Smith, Carusone, Gaudet

VDD − VG 1 µA

10 − 6 = 4 M 1

RG2 =

6 = 6 M 1 µA

RD =

5V = 10 k 0.5 mA

To determine VS , we use W 1 (VSG − |Vt |)2 ID = k p 2 L

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–16

1 × 4 × (VSG − 1.5)2 2

0.5 =

and, as a check, ID = µA.

⇒ VSG = 2 V

(g) Refer to Fig. P5.57(g). Our work is considerably simplified by observing that this circuit is similar to that in Fig. P5.57(e) with the 1-mA current source replaced with a 0.5-k resistor. Thus V7 = V5 = 1.5 V and, as a check, 2 − 1.5 = 1 mA. ID = 0.5

Thus, VS = VG + VSG = 6 + 2 = 8 V RS =

10 − 8 = 4 k 0.5

5.57 (a) Refer to Fig. P5.57(a): The MOSFET is operating in saturation. Thus ID =

1 2 k n VOV 2

10 =

1 2 × 2000 × VOV ⇒ VOV = 0.1 V 2

VGS = Vt + VOV = 0.5 + 0.1 = 0.6 V V1 = 0 − VGS = −0.6 V (b) Refer to Fig. P5.57(b): The MOSFET is operating in saturation. Thus 100 =

1 2 × 2000 × VOV ⇒ VOV = 0.316 V 2

(h) Refer to Fig. P5.57(h). Our work is considerably simplified by observing that this circuit is similar to that in Fig. P5.57(a) with the 10-µA current source replaced with a 140-k resistor. Thus V8 = V1 = −0.6 V and, as a check, −0.6 + 2 ID = = 0.01 mA = 10 µA. 140 5.58 (a) Refer to the circuit in Fig. P5.58(a). Transistor Q1 is operating in saturation. Assume that Q2 also is operating in saturation, VGS2 = 0 − V2 = −V2 and V2 = −1 + ID × 4

VGS = 0.5 + 0.316 = 0.816 V

⇒ ID = 0.25(V2 + 1)

V2 = −0.816 V

Now,

(c) Refer to Fig. P5.57(c). The MOSFET is operating in saturation. Thus

ID =

1=

2 − 0.6 = 0.01 mA = 10 140

1 2 × 2 × VOV ⇒ VOV = 1 V 2

1 k n (VGS2 − Vt )2 2

Substituting ID = 0.25(V2 + 1) and VGS2 = −V2 , 1 × 5(−V2 − 0.4)2 2

VGS = 0.5 + 1 = 1.5 V

0.25(V2 + 1) =

V3 = −1.5 V

0.1(V2 + 1) = V22 + 0.8 V2 + 0.16

(d) Refer to Fig. P5.57(d). The MOSFET is operating in saturation. Thus

V22 + 0.7 V2 + 0.06 = 0

1 2 ⇒ VOV = 0.1 V 10 = × 2000 × VOV 2

⇒ V2 = −0.6 V Thus,

VGS = 0.5 + 0.1 = 0.6 V

ID = 0.25(V2 + 1) = 0.1 mA

V4 = 0.6 V

and

(e) Refer to Fig. P5.57(e). The MOSFET is operating in saturation. Thus

VGS2 = 0.6 V

1 2 ⇒ VOV = 1 V 1 = × 2 × VOV 2 VGS = 0.5 + 1 = 1.5 V

VGS1 = 0.6 V ⇒ V1 = 0.5 − 0.6 = −0.1 V

V5 = VGS = 1.5 V (f) Refer to Fig. P5.57(f). To simplify our solution, we observe that this circuit is that in Fig. P5.57(d) with the 10-µA current source replaced with a 140-k resistor. Thus V6 = V4 = +0.6 V

Sedra, Smith, Carusone, Gaudet

Since Q1 is identical to Q2 and is conducting the same ID , then

which confirms that Q2 is operating in saturation, as assumed. (b) Refer to the circuit in Fig. P5.58(b). From symmetry, we see that

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–17

1 × 0.2 × |VOV |2 2

V4 = +1 V

0.1 =

Now, consider the part of the circuit consisting of Q2 and the 4-k resistor. We observe the similarity of this part with the circuit between the gate of Q2 and the −1 V supply in Fig. P5.58(a). It follows that for the circuit in Fig. P5.58(b), we can use the solution of part (a) above to write

⇒ |VOV | = 1 V

ID2 = 0.1 mA

and

VGS2 = 0.6 V

Thus, V5 = V4 − VGS2 = 1 − 0.6 = 0.4 V Since Q1 is conducting an equal ID and has the same VGS , ID1 = 0.1 mA

and

VGS1 = 0.6 V

⇒ V3 = V4 + VGS1 = 1 + 0.6 = 1.6 V We could, of course, have used the circuit’s symmetry, observed earlier, to write this final result.

VSG = |Vtp | + |VOV | = 1 + 1 = 2 V VG = 10 − 2 = 8 V V D = VG = 8 V VSD = 2 V (ii) R = 10 k IR = 0.1 × 10 = 1 V which just satisfies the condition for saturation-mode operation in (a) above. Obviously ID and |VOV | will be the same as in (i) above. VSG = 2 V VG = 8 V VD = VG + IR = 8 + 1 = 9 V VSD = 1 V (iii) R = 30 k

5.59

IR = 0.1 × 30 = 3 V

10 V

VSG

VSD

R

IR

which is greater than |Vtp |. Thus the condition in (a) above is not satisfied and the MOSFET is operating in the triode region.

10 V

VSG

VSD 3V

I 30 k Fig. 1

0.1 mA (a) From Fig. 1 we see that VDG = IR

Fig. 2

Since for the PMOS transistor to operate in saturation,

From Fig. 2, we see that

VDG ≤ |Vtp |

VSD = VSG − 3

It follows the

Now, for triode-mode operation, 1 2 ID = k p (VSG − |Vtp |)VSD − VSD 2 1 0.1 = 0.2 (VSG − 1)(VSG − 3) − (VSG − 3)2 2

IR ≤ |Vtp |

Q.E.D

(b) (i) R = 0, the condition above is satisfied and ID = I =

1 k p |VOV |2 2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–18

2 ⇒ VSG − 2VSG − 4 = 0

I2 = 50 µA

⇒ VSG = 3.24 V VSD = VSG − 3 = 0.24 V

This is the same current that flows through Q4 , which is operating in saturation and is matched to Q3 . Thus

(iv) R = 100 k

VGS4 = VGS3 = VGS1 = 1 V 10 V

VSG

100 k

Thus,

VSD 10 V

V2 = 2 − 1 = 1 V This is equal to the voltage at the gate of Q3 ; thus Q3 is indeed operating in saturation, as assumed.

0.1 mA

Fig. 3

I2 = ID3 = 10 ID1 = 10 × 50 µA

Here also (see Fig. 3) the MOSFET will be operating in the triode region, and VSD = VSG − 10 V Since we expect VSD to be very small, we can 2 term in the expression for ID and neglect the VSD write ID k p (VSG − |Vt |)VSD 0.1 = 0.2(VSG − 1)(VSG − 10)

5.61 See Table on next page. 5.62 Using Eq. (5.30), we can write

2φf + VSB − 2φf Vt = Vt0 + γ

γ = 0.5 V1/2

VSD = VSG − 10 = 0.055 V

2φf = 0.6 V

5.60 Refer to the circuit in Fig. P5.60. First consider Q1 and Q2 . Both are operating in saturation and since they are identical, they have equal VGS : 2 = =1V 2

and VSB = 0 to 4 V At VSB = 0, Vt = Vt0 = 1.0 V At

Thus,

=

Transistor Q4 will carry I2 but will retain the same VGS as before, thus V2 remains unchanged at 1 V.

Vt0 = 1.0 V

⇒ VSG = 10.055 V

ID2 = ID1 =

= 0.5 mA

where

2 − 11VSG + 9.5 = 0 ⇒ VSG

VGS1 = VGS2

If Q3 and Q4 have W = 5 µm, nothing changes for Q1 and Q2 . However, Q3 , which has the same VGS as Q1 but has 10 times the width, will have a drain current 10 times larger than Q1 . Thus

W 1 μn Cox (VGS1 − Vt )2 2 L

0.5 1 × 400 × (1 − 0.5)2 2 0.5

VSB = 4 V, Vt = 1 + 0.5

√

0.6 + 4 −

√

0.6

= 1.69 V

= 50 µA

ID3 = ID1 = 50 µA

If the gate oxide thickness is increased by a factor of 4, Cox will decrease by a factor of 4 and Eq. (5.31) indicates that γ will increase by a factor of 4, becoming 2. Thus at VSB = 4 V, √

√ Vt = 1 + 2 0.6 + 4 − 0.6

Thus,

= 3.74 V

Now, Q3 has the same VGS at Q1 and is matched to Q1 . Thus if we assume that Q3 is operating in saturation, we have

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–19

This table belongs to Problem 5.61. L (µm)

0.5

0.25

0.18

0.13

tox (nm)

10

5

3.6

2.6

3.45

6.90

9.58

13.3

173

345

479

665

1.73

3.45

4.79

6.65

34.5 ox 34.5(pF/µm) = (fF/µm2 ) = tox tox (nm) tox (nm) ox = 34.5 pF/m µA = μn (cm2 /V·s) × Cox (fF/µm2 ) × 10−1 k n V2 (μn = 500 cm2 /V·s) mA W = k kn n L V2 for L = Lmin and W/L = 10 Cox =

A(µm2 ) = W(µm) × L(µm) W for = 10 L

2.50 0.625 0.324 0.169

VDD (V) Vt (V)

5

2.5

1.8

1.3

0.7

0.5

0.4

0.4

16

6.90

4.69

2.69

80

17.3

8.44

3.50

32

27.7

26.1

20.7

n

4n

ID (mA) for VGS = VDS = VDD , P(mW) P mW A µm2

ID =

1 k n (VDD − Vt )2 2

P = VDD ID

Devices/Chip

5.63

7.72n 14.8n

5.64

(a) At 20◦ C, Vt = 0.5 V and kn = 1 mA/V2 , thus

vD

1 kn (VGS − Vt )2 2 1 0.5 = × 1(VGS − 0.5)2 2 ⇒ VGS = 1.5 V iD =

(b) At 50◦ C, T = 30◦ C, thus Vt = −2 × 30 = −60 mV ⇒ Vt = 0.5 − 0.06 = 0.44 V kn = −0.3% × 30 = −9% ⇒ kn = −9% ⇒ kn = 1 − 0.09 = 0.91 mA/V2 Thus, 1 × 0.91(VGS − 0.44)2 2 2 × 0.5 = 0.44 + = 1.488 V 0.91

0.5 = ⇒ VGS

Sedra, Smith, Carusone, Gaudet

The NMOS depletion-type MOSFET has the same i−v characteristics as the enhancement-type NMOS except that Vtn is negative, for the depletion device: 1 iD = k n (v GS − Vtn )v DS − v 2DS , 2 for v DS ≤ v GS − Vtn iD =

1 k n (v GS − Vtn )2 , 2

for v DS ≥ v GS − Vtn

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–20

For our case, v GS = 0, Vtn = −3 V, and k n = 2 mA/V2 . Thus 1 iD = 2 3v D − v 2D , for v D ≤ 3 V 2

and ∂iD /iD = −0.002/◦ C, VGS = 5 V ∂T and Vt = 1 V

1 × 2 × 9 = 9 mA, for v D ≥ 3 V 2

iD =

−0.002 =

For

∂k n /k n 2 × −0.002 − ∂T 5−1

∂k n /k n = −0.003/◦ C ∂T

1 v D = 0.1 V, iD = 2 3 × 0.1 − × 0.12 2 = 0.59 mA (triode)

⇒

For

1 5.68 iD = k n (v GS − Vtn )v DS − v 2DS , 2 for v DS ≤ v GS − Vtn

For

iD =

or − 0.3%/◦ C

1 v D = 1 V, iD = 2 3 × 1 − × 1 2 =5 mA (triode)

v D = 3 V,

1 k n (v GS − Vtn )2 (1 + λ v DS ), 2 for v DS ≥ v GS − Vtn

iD = 9 mA (saturation)

For our case,

For v D = 5 V,

Vtn = −2 V, k n = 0.2 mA/V2 , λ = 0.02 V−1

iD = 9 mA (saturation)

and v GS = 0. Thus 1 iD = 0.2 2 v DS − v 2DS , for v DS ≤ 2 V 2

5.65 Here the scaling factor is S=

5 µm = 156.25 32 nm

iD = 0.4(1 + 0.02 v DS ), for v DS ≥ 2 V

Therefore, Moore’s law predicts that the number of transistors on a chip of equal area, fabricated in the 32-nm process, would be 20, 000 × S2 = 20, 000 × 156.252

For v DS = 1 V, 1 = 0.3 mA iD = 0.2 2 − 2

4.88 × 108 or 488 million transistors.

For v DS = 2 V,

5.66 |Vt | = |Vt0 | + γ = 0.7 + 0.5

√

2φf + |VSB | −

0.75 + 3 −

√

0.75

2φf

iD = 0.4(1 + 0.02 × 2) = 0.416 mA For v DS = 3 V, iD = 0.4(1 + 0.02 × 3) = 0.424 mA

= 1.24 V

For v DS = 10 V,

Thus,

iD = 0.4(1 + 0.02 × 10) = 0.48 mA

Vt = −1.24 V

If the device width W is doubled, k n is doubled, and each of the currents above will be doubled. If both W and L are doubled, k n remains unchanged. However, λ is divided in half; thus for v DS = 2 V, iD becomes 0.408 mA; for v DS = 3 V, iD becomes 0.412 mA; and for v DS = 10 V, iD becomes 0.44 mA.

1 W kn (VGS − Vt )2 2 L 1 ∂k n W ∂iD = (VGS − Vt )2 ∂T 2 ∂T L W ∂Vt (VGS − Vt ) −k n L ∂T

5.67 (a) iD =

∂k /k 2 ∂iD /iD ∂Vt = n n − ∂T ∂T VGS − Vt ∂T (b) For ∂Vt = −2 mV/◦ C = −0.002 V/◦ C ∂T

Sedra, Smith, Carusone, Gaudet

5.69 The depletion-type MOSFET operates in the triode region when v DS ≤ v GS − Vt : that is, v DG ≤ −Vt , where Vt is negative. In the case shown in Fig. 1, v DG = 0. Thus the condition for triode-mode operation is satisfied, and

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 5–21

Here v GS = 0 and v DS = −v; thus v DS ≥ −Vt . Thus the device is operating in saturation, and

i

v

iD =

1 k n (0 − Vt )2 2

iD =

1 k n Vt2 2

But i = −iD ; thus Fig. 1

1 iD = k n (v GS − Vt )v DS − v 2DS 2

1 i = − k n Vt2 , 2

for v ≤ Vt

Figure 3 is a sketch of the i−v relationship for the case Vt = −2 V and k n = 2 mA/V2 .

which applies when the channel is not depleted, that is, when v GS ≥ Vt . For our case, 1 2 i = k n (v − Vt )v − v , for v ≥ Vt 2

Here

Thus,

i = −4 mA, for v ≤ −2 V

i=

1 k n (v 2 − 2Vt v), 2

i = v(v + 4), for v ≥ −2 V and

for v ≥ Vt

i (mA)

For v ≤ Vt , the source and the drain exchange roles, as indicated in Fig. 2.

6 4 2 –3

–2

–1

0

vDS

2

v (V)

–4

iD

–6

Fig. 3

Fig. 2

Sedra, Smith, Carusone, Gaudet

1 –2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–1

Av = −k n VOV RD

Chapter 6 Solutions to Exercises within the Chapter

−10 = −0.4 × 10 × VOV × 17.5 Thus,

Ex: 6.1 Refer to Fig. 6.2(a) and 6.2(b).

VOV = 0.14 V

Coordinates of point A: Vt and VDD ; thus 0.4 V and 1.8 V. To determine the coordinates of point B, we use Eqs. (6.7) and (6.8) as follows: √ 2k n RD VDD + 1 − 1 VOV B = k n RD √ 2 × 4 × 17.5 × 1.8 + 1 − 1 = 4 × 17.5 = 0.213 V

VGS = Vt + VOV = 0.4 + 0.14 = 0.54 V W 1 2 ID = k n VOV 2 L

Thus, VGS B = Vt + VOV B = 0.4 + 0.213 = 0.613 V

= 1.8 − 17.5 × 0.04 = 1.1 V

and

1 × 0.4 × 10 × 0.142 = 0.04 mA 2 RD = 17.5 k =

VDS = VDD − RD ID

Ex: 6.3 IC R C VT

VDS B = VOV B = 0.213 V

Av = −

Thus, coordinates of B are 0.613 V and 0.213 V. At point C, the MOSFET is operating in the triode region, thus 1 iD = k n (v GS C − Vt )v DS C − v 2DS C 2 If v DS C is very small, iD k n (v GS C − Vt )v DS C = 4(1.8 − 0.4)v DS C = 5.6v DS C , mA

−320 = −

But iD =

VDD − v DS C

VDD 1.8 = 0.1 mA = RD 17.5

1 × RC ⇒ RC = 8 k 0.025 VC = VCC − IC RC = 10 − 1 × 8 = 2 V Since the collector voltage is allowed to decrease to +0.3 V, the largest negative swing allowed at the output is 2 − 0.3 = 1.7 V. The corresponding input signal amplitude can be found by dividing 1.7 V by the gain magnitude (320 V/V), resulting in 5.3 mV.

Ex: 6.4

RD 0.1 = 0.018 V = 18 mV, which Thus, v DS C = 5.6 is indeed very small, as assumed.

VDD

iD

RD

Ex: 6.2 Refer to Example 6.1 and Fig. 6.4(a).

vDS

Design 1:

vgs

VOV = 0.2 V, VGS = 0.6 V

vGS

ID = 0.08 mA Now,

VGS

Av = −k n VOV RD Thus, −10 = −0.4 × 10 × 0.2 × RD

VDD = 5 V

⇒ RD = 12.5 k

VGS = 2 V

VDS = VDD − RD ID

Vt = 1 V

= 1.8 − 12.5 × 0.08 = 0.8 V

λ=0

Design 2:

k n = 20 µA/V2

RD = 17.5 k

RD = 10 k

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–2

W = 20 L (a) VGS = 2 V ⇒ VOV = 1 V

Ex: 6.6 ID = 0.1 mA, gm = 1 mA/V, k n = 500 µA/V2

ID =

gm =

2 ID 2 × 0.1 = 0.2 V ⇒ VOV = VOV 1

VDS

ID =

1 W 2 W 2 ID k V ⇒ = 2 2 n L OV L k n VOV

1 W 2 k V = 200 µA = 0.2 mA 2 n L OV = VDD − ID RD = +3 V

(b) gm = k n

W VOV = 400 µA/V = 0.4 mA/V L

v ds = −gm RD = −4 V/V (c) Av = v gs

=

2 × 0.1 = 10 500 × 0.22 1000

(d) v gs = 0.2 sinωt V Ex: 6.7

v ds = −0.8 sinωt V v DS = VDS + v ds ⇒ 2.2 V ≤ v DS ≤ 3.8 V (e) Using Eq. (6.28), we obtain 1 k n (VGS − Vt )2 2

iD =

W VOV L Same bias conditions, so same VOV and also same L and gm for both PMOS and NMOS.

gm = μn Cox

μn Cox Wn = μp Cox Wp ⇒

1 + k n (VGS − Vt )v gs + k n v 2gs 2 iD = 200 + 80 sinωt

⇒

μp Wn = 0.4 = μn Wp

Wp = 2.5 Wn

+ 8 sin2 ωt, µA = [200 + 80 sinωt + (4 − 4 cos2ωt)] = 204 + 80 sinωt − 4 cos2ωt, µA Thus, ID shifts by 4 µA and 2HD =

ˆi2ω 4 µA = 0.05 (5%) = ˆiω 80 µA

Ex: 6.8 1 W ID = k p (VSG − |Vt |)2 2 L 1 16 = × 60 × × (1.6 − 1)2 2 0.8 ID = 216 µA 2 × 216 2 ID = = 720 µA/V |VOV | 1.6 − 1

gm =

= 0.72 mA/V Ex: 6.5

|λ| = 0.04 ⇒ |VA | =

(a) VGS = 1.5V ⇒ VOV = 1.5 − 1 = 0.5 V

ro =

gm =

2 ID VOV

1 W 2 1 k n VOV = × 60 × 40 × 0.52 2 L 2 ID = 300 µA = 0.3 mA ID =

2 × 0.3 = 1.2 mA/V gm = 0.5 VA 15 = 50 k ro = = ID 0.3 (b) ID = 0.5 mA ⇒ gm = =

2 × 60 × 40 × 0.5 × 103

gm = 1.55 mA/V VA 15 ro = = = 30 k ID 0.5

Sedra, Smith, Carusone, Gaudet

|VA | × L 25 × 0.8 = 92.6 k = ID 0.216

Ex: 6.9 gm ro =

W ID L

2 ID VA 2VA × = VOV ID VOV

VA = VA × L = 6 × 3 × 0.18 = 3.24 V gm r o =

2 μn Cox

1 1 = = 25 V/µm |λ| 0.04

2 × 3.24 = 32.4 V/V 0.2

Ex: 6.10 Refer to the solution of Example 6.3. From vo Eq. (6.47), Av ≡ = −gm RD (note that RL vi is absent). Thus, gm RD = 25

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–3

5 V

Ex: 6.11

RD

RG

vo

it

Req

D

vt

0 i

vi

ro G 1 gm

i

S

Substituting for gm = k n VOV , we have k n VOV RD = 25

it =

where k n = 1 mA/V2 , thus VOV RD = 25

(1)

Next, consider the bias equation VGS = VDS = VDD − RD ID

vt vt +i= + gm v t ro ro

∴ Req =

vt 1 = ro it gm

Ex: 6.12 Given: gm =

Thus, Vt + VOV = VDD − RD ID Substituting Vt = 0.7 V, VDD = 5 V, and 1 1 1 2 2 2 k n VOV = × 1 × VOV = VOV 2 2 2 we obtain 1 2 RD (2) 0.7 + VOV = 5 − VOV 2 Equations (1) and (2) can be solved to obtain ID =

VOV = 0.319 V

∂iC ∂v BE iC = IC

where IC = IS ev BE /VT ∂iC IS ev BE /VT IC = = ∂v BE iC = IC VT VT Thus, gm =

IC VT

Ex: 6.13 IC 0.5 mA = 20 mA/V gm = = VT 25 mV

and Ex: 6.14

RD = 78.5 k The dc current ID can be now found as 1 2 = 50.9 µA k n VOV 2 To determine the required value of RG we use Eq. (6.48), again noting that RL is absent: ID =

RG Rin = 1 + gm RD RG 0.5 M = 1 + 25

For β = 50: gm =

IC 0.5 mA = 20 mA/V = VT 25 mV

IB =

0.5 IC = = 10 µA β 50

rπ =

β 50 = 2.5 k = gm 20

For β = 200,

⇒ RG = 13 M Finally, the maximum allowable input signal vˆ i can be found as follows: 0.7 V Vt = = 27 mV vˆ i = |Av | + 1 25 + 1

Sedra, Smith, Carusone, Gaudet

IC = 0.5 mA (constant)

gm =

IC = 20 mA/V VT

IB =

0.5 mA IC = = 2.5 µA β 200

rπ =

β 200 = = 10 k gm 20

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–4

ic = βib = β

Ex: 6.15

=

1 mA = 40 mA/V 25 mV

gm = re =

IC = 1 mA

β = 100

β v be = gm v be rπ

ie = ib + βib = (β + 1)ib = (β + 1)

VT αVT 25 mV = 25 = IE IC 1 mA

rπ =

v be rπ

=

β 100 = 2.5 k = gm 40

v be rπ

v be v be = rπ /(β + 1) re

Ex: 6.18 Ex: 6.16 IC 1 mA = 40 mA/V gm = = VT 25 mV v ce = −gm RC Av = v be

C

gmvbe

= −40 × 10

ib

= −400 V/V

B

VC = VCC − IC RC

re

= 15 − 1 × 10 = 5 V

vbe

v C (t) = VC + v c (t)

= VC + Av v be (t)

E

= 5 − 400 × 0.005 sinωt v be − gm v be re 1 = v be − gm re 1 β = v be − rπ/β+1 rπ β +1 v be β = − = v be rπ rπ rπ

= 5 − 2 sinωt

ib =

iB (t) = IB + ib (t) where 1 mA IC = = 10 µA IB = β 100 and ib (t) =

gm v be (t) β

40 × 0.005 sinωt 100 = 2 sinωt, µA

=

Ex: 6.19

10 V

Thus, iB (t) = 10 + 2 sinωt, µA

RE 10 k

Ex: 6.17

CC1

ib B

ic C

vbe

vi

CC2

bib

rp

vo

RC 7.5 k

E

Sedra, Smith, Carusone, Gaudet

10 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–5 10 − 0.7 = 0.93 mA 10 IC = αIE = 0.99 × 0.93

VE = −0.1 − 0.7 = −0.8 V

IE =

(b) gm =

= 0.92 mA VC = −10 + IC RC = −10 + 0.92 × 7.5 = −3.1 V vo αRC Av = = vi re

IC 0.99 40 mA/V = VT 0.025

rπ =

β 100 = 2.5 k gm 40

ro =

VA 100 = 101 100 k = IC 0.99

(c) See figure below. Rsig = 2 k RB = 10 k rπ = 2.5 k

25 mV where re = = 26.9 0.93 mA

gm = 40 mA/V

0.99 × 7.5 × 103 = 276.2 V/V Av = 26.9

RC = 8 k

For vˆ i = 10 mV, vˆ o = 276.2 × 10 = 2.76 V

Vy Vπ Vy = × Vsig Vsig Vπ

Ex: 6.20

=

RB rπ × −gm (RC RL ro ) (RB rπ ) + Rsig

=

10 2.5 × −40(8 8 100) (10 2.5) + 2

10V

RL = 8 k ro = 100 k

= −0.5 × 40 × 3.846 = −77 V/V

8 k

If ro is negelected, X

Y

10 k

Z

I 1 mA

Vy = −80, for an error Vsig

of 3.9%.

Ex: 6.21 2ID 2 × 0.25 = 2 mA/V gm = = VOV 0.25 Rin = ∞ Av o = −gm RD = −2 × 20 = −40 V/V Ro = RD = 20 k

IE = 1 mA 100 × 1 = 0.99 mA IC = 101 1 × 1 = 0.0099 mA IB = 101 (a) VC = 10 − 8 × 0.99 = 2.08 2.1 V VB = −10 × 0.0099 = −0.099 −0.1 V

Av = Av o

RL 20 = −40 × RL + Ro 20 + 20

= −20 V/V Gv = Av = −20 V/V vˆ i = 0.1 × 2VOV = 0.1 × 2 × 0.25 = 0.05 V vˆ o = 0.05 × 20 = 1 V

This figure belongs to Exercise 6.20c.

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–6

Ex: 6.22 IC = 0.5 mA gm =

IC 0.5 mA = 20 mA/V = VT 0.025 V

rπ =

β 100 = 5 k = gm 20

Rin = rπ = 5 k Av o = −gm RC = −20 × 10 = −200 V/V Ro = RC = 10 k Av = Av o

RL 5 = −200 × RL + Ro 5 + 10

= −66.7 V/V Gv =

Rin 5 Av = × −66.7 Rin + Rsig 5+5

= −33.3 V/V

For IC = 0.5 mA and β = 100,

vˆ π = 5 mV ⇒ vˆ sig = 2 × 5 = 10 mV

re =

vˆ o = 10 × 33.3 = 0.33 V

VT αVT 0.99 × 25 = = 50 IE IC 0.5

rπ = (β + 1)re 5 k

Although a larger fraction of the input signal reaches the amplifier input, linearity considerations cause the output signal to be in fact smaller than in the original design!

For vˆ sig = 100 mV, Rsig = 10 k and with vˆ π limited to 10 mV, the value of Re required can be found from 10 Re + 100 = 10 1 + 50 5

Ex: 6.23 Refer to the solution to Exercise 6.21. If vˆ sig = 0.2 V and we wish to keep

⇒ Re = 350

vˆ gs = 50 mV, then we need to connect a 3 resistance Rs = in the source lead. Thus, gm 3 = 1.5 k Rs = 2 mA/V

= 40.4 k Gv = −β = −100

RD RL Gv = Av = − 1 + Rs gm =−

Rin = (β + 1)(re + Re ) = 101 × (50 + 350)

2020 = −5 V/V 0.5 + 1.5

vˆ o = Gv vˆ sig = 5 × 0.2 = 1 V (unchanged)

1 = 10 mA/V 0.1 k

But

Ex: 6.24 From the following figure we see that

gm =

vˆ sig = ˆib Rsig + vˆ π + ˆie Re

Thus,

ˆie Rsig + vˆ π + ˆie Re = β +1

2ID VOV

2ID 0.2 ⇒ ID = 1 mA 10 =

=

Sedra, Smith, Carusone, Gaudet

10 = −19.8 V/V 10 + 101 × 0.4

Ex: 6.25 1 = Rsig = 100 gm ⇒ gm =

vˆ π vˆ π Rsig + vˆ π + Re (β + 1)re re Rsig Re vˆ sig = vˆ π 1 + + re rπ

RC RL Rsig + (β + 1)(re + Re )

Gv = Q.E.D.

Rin × gm RD Rin + Rsig

= 0.5 × 10 × 2 = 10 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–7

1 = 200 gm

Ex: 6.26 IC = 1 mA re =

VT VT 25 mV = 25 = IE IC 1 mA

Rin = re = 25

⇒ gm = 5 mA/V But

gm = k n

Av o = gm RC = 40 × 5 = 200 V/V Ro = RC = 5 k RL 5 = 100 V/V = 200 × Av = Av o RL + Ro 5+5

W VOV L

Thus, 5 = 0.4 ×

W × 0.25 L

Ex: 6.27

W = 50 L 1 W 2 ID = k n VOV 2 L 1 = × 0.4 × 50 × 0.252 2 = 0.625 mA

Rin = re = 50

RL = 1 k to 10 k

Rin Gv = × Av Rin + Rsig =

25 × 100 = 0.5 V/V 25 + 5000

⇒ IE =

VT 25 mV = 0.5 mA = re 50

IC IE = 0.5 mA Gv = 40 =

RC RL re + Rsig

⇒

Correspondingly, Gv =

RL RL = RL + Ro RL + 0.2

will range from Gv =

RC R L (50 + 50)

1 = 0.83 V/V 1 + 0.2

to

RC RL = 4 k

Gv =

Ex: 6.28 Refer to Fig. 6.42(c). Ro = 100

10 = 0.98 V/V 10 + 0.2

Ex: 6.30 IC = 5 mA

Thus, 1 = 100 ⇒ gm = 10 mA/V gm But

re =

VT VT 25 mV =5 = IE IC 5 mA

Rsig = 10 k RL = 1 k Rin = (β + 1) (re + RL )

2ID gm = VOV

= 101 × (0.005 + 1)

Thus,

= 101.5 k

10 × 0.25 = 1.25 mA 2 RL 1 vˆ o = vˆ i × = 0.91 V =1× RL + Ro 1 + 0.1

Gv o = 1 V/V

ID =

vˆ gs = vˆ i

1 gm 1 + RL gm

=1×

0.1 = 91 mV 0.1 + 1

Rout = re +

Rsig β +1

10,000 = 104 101 RL RL Gv = = Rsig RL + Rout RL + r e + β +1 =5+

1 = 0.91 V/V 1 + 0.104 re v π = v sig Rsig r e + RL + β +1 =

Ex: 6.29 Ro = 200

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–8 Rsig RL + vˆ sig = vˆ π 1 + re (β + 1) re 10,000 1000 + = 1.1 V/V vˆ sig = 5 1 + 5 101 × 5

0.5 =

1 × 1(VGS − 1)2 2

⇒ VGS = 2 V If Vt = 1.5 V, then

Correspondingly, ID =

vˆ o = Gv × 1.1 = 0.91 × 1.1 = 1 V

⇒ Ex: 6.31

1 × 1 × (2 − 1.5)2 = 0.125 mA 2

0.125 − 0.5 ID = = −0.75 = −75% ID 0.5

Ex: 6.33

Rsig

RD =

Q1

→ RD = 6.2 k

vsig

re1

Q2

ID =

vo Rin

RL (b2 1) (re2 RL)

1 W 2 1 2 k V ⇒ 0.5 = × 1 × VOV 2 n L OV 2

⇒ VOV = 1 V ⇒ VGS = VOV + Vt = 1 + 1 = 2 V

Rout

⇒ VS = −2 V RS =

From the figure we can write Rin = (β1 + 1) [re1 + (β2 + 1)(re2 + RL )] re1 + Rsig /(β1 + 1) Rout = RL re2 + β2 + 1 vo = vsig

VDD − VD 5−2 = = 6 k ID 0.5

RL re1 + Rsig /(β1 + 1) RL + re2 + β2 + 1

VS − VSS −2 − (−5) = 6 k = ID 0.5

→ RS = 6.2 k If we choose RD = RS = 6.2 k, then ID will change slightly: ID =

1 × 1 × (VGS − 1)2 . 2

Also,

For IE2 = 5 mA, β1 = β2 = 100, RL = 1 k, and Rsig = 100 k, we obtain

VGS = −VS = 5 − RS ID

25 mV =5 re2 = 5 mA 5 5 = 0.05 mA IE1 = β2 + 1 101

2 ID = (4 − 6.2 ID )2

25 mV = 500 0.05 mA Rin = 101 × (0.5 + 101 × 1.005) = 10.3 M 0.5 + (100/101) 20 Rout = 1 0.005 + 101 re1 =

vo = vsig

1 0.5 + (100/101) 1 + 0.005 + 101 = 0.98 V/V

Thus,

⇒ 38.44 ID2 − 51.6 ID2 + 16 = 0 ⇒ ID = 0.49 mA, 0.86 mA ID = 0.86 results in VS > 0 or VS > VG , which is not acceptable. Therefore ID = 0.49 mA and VS = −5 + 6.2 × 0.49 = −1.96 V VD = 5 − 6.2 × 0.49 = +1.96 V RG should be selected in the range of 1 M to 10 M. Ex: 6.34 1 W 2 k V 2 n L OV 0.5 × 2 = =1 1 = 1 V ⇒ VGS = 1 + 1 = 2 V

ID = 0.5 mA = Ex: 6.32 1 W ID = k n (VGS − Vt )2 2 L

Sedra, Smith, Carusone, Gaudet

2 ⇒ VOV

⇒ VOV

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–9 5−2 = 6 k 0.5 ⇒ RD = 6.2 k (standard value). For this RD we have to recalculate ID :

= VD ⇒ RD =

ID =

highly desirable result obtained at the expense of increased current and hence power dissipation. Ex: 6.36 Refer to Fig. 6.55. Since the circuit is to be used as a common-base amplifier, we can dispense with RB altogether and ground the base; thus RB = 0. The circuit takes the form shown in the figure below.

1 × 1 × (VGS − 1)2 2

1 (VDD − RD ID − 1)2 2 (VGS = VD = VDD − RD ID ) =

10 V

1 ID = (4 − 6.2 ID )2 ⇒ ID ∼ = 0.49 mA 2 VD = 5 − 6.2 × 0.49 = 1.96 V

RC vo

Ex: 6.35 Refer to Example 6.12. (a) For design 1, RE = 3 k, R1 = 80 k, and R2 = 40 k. Thus, VBB = 4 V. IE =

VBB − VBE R1 R2 RE + β +1

RE

For the nominal case, β = 100 and 4 − 0.7 = 1.01 1 mA IE = 4080 3+ 101 For β = 50, 4 − 0.7 = 0.94 mA 4080 3+ 51 For β = 150,

IE =

4 − 0.7 = 1.04 mA 4080 3+ 151 Thus, IE varies over a range approximately 10% of the nominal value of 1 mA. IE =

(b) For design 2, RE = 3.3 k, R1 = 8 k, and R2 = 4 k. Thus, VBB = 4 V. For the nominal case, β = 100 and 4 − 0.7 = 0.99 1 mA 48 3.3 + 101 For β = 50, IE =

4 − 0.7 = 0.984 mA 48 3.3 + 51 For β = 150,

4 − 0.7 = 0.995 mA 48 3.3 + 151 Thus, IE varies over a range of 1.1% of the nominal value of 1 mA. Note that lowering the resistances of the voltage divider considerably decreases the dependence on the value of β, a IE =

Sedra, Smith, Carusone, Gaudet

5V

To establish IE = 1mA, 5 − VBE IE = RE 5 − 0.7 1 mA = RE ⇒ RE = 4.3 k vo = gm RC , where The voltage gain vi IC gm = = 40 mA/V. VT To maximize the voltage gain, we select RC as large as possible, consistent with obtaining a ±2-V signal swing at the collector. To maintain active-mode operation at all times, the collector voltage should not be allowed to fall below the value that causes the CBJ to become forward biased, namely, −0.4 V. Thus, the lowest possible dc voltage at the collector is −0.4 V + 2V = +1.6 V. Correspondingly, RC =

IE =

vi

10 − 1.6 10 − 1.6 = 8.4 k IC 1 mA

Ex: 6.37 Refer to Fig. 6.56. For IE = 1 mA and VC = 2.3 V, VCC − VC IE = RC 1=

10 − 2.3 RC

⇒ RC = 7.7 k Now, using Eq. (6.147), we obtain

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–10 VCC − VBE RB RC + β +1

IE =

Ex: 6.39 Refer to Fig. 6.57(a) and (c) and to the values found in the solution to Exercise 6.38 above.

10 − 0.7 RB 7.7 + 101 ⇒ RB = 162 k

Rin = RG1 RG2 = 52.5 = 1.67 M

1=

Ro = RD ro = 18200 = 16.5 k Gv = −

Selecting standard 5% resistors (Appendix J), we use RB = 160 k

and

RC = 7.5 k

The resulting value of IE is found as 10 − 0.7 = 1.02 mA 160 7.5 + 101 and the collector voltage will be IE =

VC = VCC − IE RC = 2.3 V

VS = 3.5 V and ID = 0.5 mA; thus VS 3.5 = = 7 k ID 0.5

VDD = 15 V and VD = 6 V; thus VDD − VD 15 − 6 = 18 k = ID 0.5 mA

RD =

1.67 × 2 × (2001820) 1.67 + 0.1

= −17.1 V/V Ex: 6.40 To reduce v gs to half its value, the unbypassed Rs is given by Rs =

1 gm

From the solution to Exercise 6.38 above, gm = 2 mA/V. Thus

Ex: 6.38 Refer to Fig. 6.57(b).

RS =

=−

Rin gm (ro RD RL ) Rin + Rsig

To obtain VOV , we use

1 = 0.5 k 2 Neglecting ro , Gv is given by

Rs =

Gv = −

=−

Rin RD RL ×− 1 Rin + Rsig + Rs gm

1820 1.67 × 1.67 + 0.1 0.5 + 0.5

= −8.9 V/V

1 2 ID = k n VOV 2 1 2 0.5 = × 4VOV 2 ⇒ VOV = 0.5 V

Ex: 6.41 Refer to Fig. 6.58(a). For VB = 5 V and 50-µA current through RB2 , we have 5V = 100 k 0.05 mA The base current is IE 0.5 mA IB = = 5 µA β +1 100 RB2 =

Thus, VGS = Vt + VOV = 1 + 0.5 = 1.5 V We now can obtain the dc voltage required at the gate,

The current through RB1 is

VG = VS + VGS = 3.5 + 1.5 = 5 V

IRB1 = IB + IRB2 = 5 + 50 = 55 µA

Using a current of 2 µA in the voltage divider, we have 5V = 2.5 M RG2 = 2 µA

Since the voltage drop across RB1 is VCC − VB = 10 V, the value of RB1 can be found from 10 V = 182 k RB1 = 0.055 µA

The voltage drop across RG1 is 10 V, thus RG1 =

The value of RE can be found from

10 V = 5 M 2 µA

IE =

This completes the bias design. To obtain gm and ro , we use gm =

2ID 2 × 0.5 = 2 mA/V = VOV 0.5

ro =

VA 100 = = 200 k ID 0.5

Sedra, Smith, Carusone, Gaudet

VB − VBE RE

5 − 0.7 = 8.6 k 0.5 The value of RC can be found from ⇒ RE =

VC = VCC − IC RC

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 6–11

6 = 15 − 0.99 × 0.5 × RC

IC = αIE IE = 0.5 mA

RC 18 k

VC = VCC − RC IC

This completes the bias design. The values of gm , rπ , and ro can be found as follows:

For VC = 1 V and VCC = 5 V, we have

IC 0.5 mA = 20 mA/V gm = VT 0.025 V rπ =

β 100 = = 5 k gm 20

ro =

VA 100 = 200 k IC 0.5

1 = 5 − RC × 0.5 ⇒ RC = 8 k To obtain the required value of RE , we note that the voltage drop across it is (VEE − VBE ) = 4.3 V. Thus,

Ex: 6.42 Refer to Fig. 6.58(b) and to the solution of Exercise 6.41 above.

4.3 = 8.6 k 0.5 Rin Gv = gm (RC RL ) Rin + Rsig

Rin = RB1 RB2 rπ

=

RE =

= 182 100 5 = 4.64 k Ro = RC ro = 18 200 = 16.51 k Rin Gv = − gm (RC RL ro ) Rin + Rsig 4.64 × 20 × (18 20 200) Gv = − 4.64 + 10 = −57.3 V/V Ex: 6.43 Refer to the solutions of Exercises 6.41 and 6.42 above. With Re included (i.e., left unbypassed), the input resistance becomes [refer to Fig. 6.59(b)] Rin = RB1 RB2 [(β + 1)(re + Re )] Thus, 10 = 182 100 [101(0.05 + Re )] VT where we have substituted re = = IE 25 = 50 . The value of Re is found from the 0.5 equation above to be

The overall voltage gain can be found from RC RL Rin Rin + Rsig re + Re

Gv = −0.99 ×

18 20 10 10 + 10 0.05 + 0.0672

= −40 V/V Ex: 6.44 Refer to Fig. 6.60(a). Rin = 50 = re RE re re = 50 =

= 40 V/V vˆ o = 40 vˆ sig = 40 × 10 mV = 0.4 V Ex: 6.45 Refer to Fig. 6.61. Consider first the bias design of the circuit in Fig. 6.61(a). Since the required IE = 1 mA, the base current 1 IE = 0.01 mA. For a dc voltage β +1 101 drop across RB of 1 V, we obtain IB =

1V = 100 k 0.01 mA The result is a base voltage of –1 V and an emitter voltage of –1.7 V. The required value of RE can now be determined as

RB =

RE =

−1.7 − (−5) 3.3 = 3.3 k = IE 1 mA

Rin = RB [(β + 1)[re + (RE ro RL )] where ro =

VA 100 V = 100 k = IC 1 mA

Rin = 100 (100 + 1)[0.025 + (3.3 100 1)]

Re = 67.2

Gv = −α

50 × 20(88) 50 + 50

VT IE

⇒ IE = 0.5 mA

Sedra, Smith, Carusone, Gaudet

= 44.3 k Rin 44.3 vi = 0.469 V/V = = v sig Rin + Rsig 44.3 + 50 vo RE r o RL = 0.968 V/V = vi re + (RE ro RL ) vo Gv ≡ = 0.469 × 0.968 = 0.454 V/V v sig RB Rsig Rout = ro RE re + β +1 100 50 = 100 3.3 0.025 + 101 = 320

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–1

Solutions to End-of-Chapter Problems 6.1 Coordinates of point A: v GS = Vt = 0.5 V and v DS = VDD = 5 V.

vDS VDD 2 V

To obtain the coordinates of point B, we first use Eq. (6.6) to determine VGS B as √ 2k n RD VDD + 1 − 1 VGS B = Vt + k n RD √ 2 × 10 × 20 × 5 + 1 − 1 = 0.5 + 10 × 20

Q

VDSQ

0.2 V

= 0.5 + 0.22 = 0.72 V

The vertical coordinate of point B is VDS B , VDS B = VGS B − Vt = VOV B = 0.22 V

B

VDSB

VOV vˆ gs

Vt 6.2 Bias point Q: VOV = 0.2 V and VDS = 1 V. IDQ = =

VDSB

1 2 k n VOV 2 VDS Q = VDS B + 0.2 V

1 × 10 × 0.04 = 0.2 mA 2

RD =

VDD − VDS 5−1 = 20 k = IDQ 0.2

Coordinates of point B: Equation (6.6): VGS B = Vt + √ = 0.5 +

√

2k n RD VDD + 1 − 1 k n RD

2 × 10 × 20 × 5 + 1 − 1 10 × 20

(1) as indicated in the figure above. Now, VDS B is given by Eq. (6.8) [together with Eq. (6.7)], √ 2k n RD VDD + 1 − 1 (2) VDS B = k n RD From the figure we see that VDS B = VOV + vˆgs where VOV = 0.2 V (given) and vˆgs =

= 0.5 + 0.22 = 0.72 V Equations (6.7) and (6.8): √ 2k n RD VDD + 1 − 1 = 0.22 V VDS B = k n RD Av = −k n RD VOV = −10 × 20 × 0.2 = −40 V/V The lowest instantaneous voltage allowed at the output is VDS B = 0.22 V. Thus the maximum allowable negative signal swing at the output is VDSQ − 0.22 = 1 − 0.22 = 0.78 V. The corresponding peak input signal is vˆgs =

0.78 0.78 V = = 19.5 mV | Av | 40

6.3 (a) To obtain maximum gain while allowing for a −0.2-V signal swing at the output, we bias the MOSFET at point Q where

Sedra, Smith, Carusone, Gaudet

vGS

=

0.2 V | Av |

0.2 0.2 1 = = k n RD VOV k n RD × 0.2 k n RD

Thus, VDS B = 0.2 +

1 (3) k n RD Substituting for VDS B from Eq. (2), we obtain √ 2k n RD VDD + 1 − 1 1 = 0.2 + k n RD k n RD

Substituting VDD = 2 V, rearranging the equation to obtain a quadratic equation in k n RD , and solving the resulting quadratic equation results in k n RD = 79.05 which can be substituted into Eq. (3) to obtain VDS B = 0.2127 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–2

The value of VDS at the bias point can now be found from Eq. (1) as VDS Q = 0.2127 + 0.2 = 0.4127 V

| Av | = =

(b) The gain achieved can be found as

= −79.05 × 0.2 = −15.81 V/V 0.2 0.2 = = 12.7 mV | Av | 15.81

(c) ID = 100 µA VDD − VDS Q RD = ID

VCC 2VT

VCC

2 − 0.4127 = 15.87 k 0.1 79.05 = 4.98 mA/V2 (d) k n = 15.87 4.98 W = = 12.45 L 0.4

1.0 1.5

2.0

3.0 (V)

VCC − 0.3 0.7 1.2

1.7

2.7 (V)

68

108 (V/V)

| Av max |

28

VCC /2

0.5 0.75 1.0

| Av |

20

1.5 (V)

40

60 (V/V)

VCC − 0.7 0.3 0.8

1.3

2.3 (V)

52

92 (V/V)

| Av |

VCC − VCE Av = − VT

48 30

6.4 Refer to Fig. 6.6.

12

32

6.6 See figure at the bottom of this page.

5−1 = −160 V/V 0.025 The transistor enters saturation when v CE ≤ 0.3 V, thus the maximum allowable output voltage swing is 1 − 0.3 = 0.7 V. The corresponding maximum input signal permitted vˆ be is =−

0.7 V 0.7 = = 4.4 mV | Av | 160

Av = −

IC R C 0.2 × 5 = −40 V/V =− VT 0.025

6.7 (a) Using Eq. (6.23) yields | Av max | =

VCC − 0.3 3 − 0.3 = = 108 V/V VT 0.025

(b) Using Eq. (6.22) with Av = −60 yields −60 = −

3 − VCE VCC − VCE =− VT 0.025

⇒ VCE = 1.5 V

6.5 To determine | Av max |, we use Eq. (6.23), VCC − 0.3 | Av max | = VT Then, for VCE =

VT

The results are as follows:

=

vˆ be =

VCC 2

Finally, if a negative-going output signal swing of 0.4 is required, to obtain maximum gain the transistor must be biased at VCE = 0.4 + 0.3 = 0.7 V and the gain achieved becomes VCC − 0.7 . | Av | = VT

Av = −k n RD VOV

vˆgs =

VCC −

(c) IC = 0.5 mA IC RC = VCC − VCE = 3 − 1.5 = 1.5 V

VCC we obtain 2

RC =

1.5 = 3 k 0.5

This figure belongs to Problem 6.6. 3 V

1.5 V

10 k

5 k

Thévenin vO

vI

Sedra, Smith, Carusone, Gaudet

0.2 mA

10 k

vO vI

0.2 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–3

(d) IC = IS eVBE /VT

iB (A)

0.5 × 10−3 = 10−15 eVBE /0.025 ⇒ VBE = 0.673 V

IB

(e) Assuming linear operation around the bias point, we obtain

ib 1 A

6 A 5 A 4 A

v ce = Av × v be = −60 × 5 sin ωt = −300 sin ωt, mV

0

= −0.3 sin ωt, V −v ce (f) ic = = 0.1 sin ωt, mA RC 0.5 mA IC = = 0.005 mA β 100 ic 0.1 ib = = sin ωt = 0.001 sin ωt, mA β 100 vˆbe (h) Small-signal input resistance ≡ iˆb

The value of RD required can now be found as VDD − VDS B RD = ID

(g) IB =

B

5 − 0.5 = 7.2 k 0.625 If the transistor is replaced with another having twice the value of k n , then ID B will be twice as large and the required value of RD will be half that used before, that is, 3.6 k. =

5 mV = 5 k 0.001 mA

=

t

(i)

vBE

6.9 From Eq. (6.18):

vbe VBE

5 mV

VDD − VOV B | Av max | = VOV B /2 2 − VOV B 14 = VOV B /2 ⇒ VOV B = 0.25 V

0.673 V

0

t

vCE VCE

vce

Now, using Eq. (6.15) at point B, we have Av B = −k n VOV B RD

1.8 V 1.5 V

Thus,

0.3 V

1.2 V 0 iC (mA) IC

−14 = −k n RD × 0.25 ⇒ k n RD = 56

t

0.1 mA

0.6 mA 0.5 mA 0.4 mA

0

Thus, 1 1 2 = × 5 × 0.52 = 0.625 mA ID B = k n VDS B 2 2

Sedra, Smith, Carusone, Gaudet

Q

Thus, 12 = 0.214 V VOV Q = 56

t

6.8 VDS B = VOV B = 0.5 V

To obtain a gain of −12 V/V at point Q: −12 = −k n RD VOV Q = −56VOV

To obtain the required VDS Q , we use Eq. (6.17), VDD − VDS Q Av = − VOV /2 Q

−12 = −

2 − VDS Q 0.214/2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–4 ⇒ VDS Q = 0.716 V

iD =

1 × 1 × (1.39 − 1)2 = 0.076 mA 2

and 6.10

v DS = 5 − 0.076 × 24 = 3.175 V

VDD

and v GS = 1.5 + 0.11 = 1.61 V at which v DS = 0.61 V.

iD RD

Thus, the large-signal gain is vDS

ⵑ

0.61 − 3.175 = −11.7 V/V 1.61 − 1.39 whose magnitude is slightly less (−2.5%) than the incremental or small-signal gain (−12 V/V). This is an indication that the transfer characteristic is not a straight line.

vgs

VGS

6.11 RD = 20 k

(a) Coordinates of point A:

k n = 200 µA/V2

VGS |A = Vt = 1 V, VDS |A = VDD = 5 V

VRD = 1.5 V VGS = 0.7 V

Coordinates of point B can be obtained using Eqs. (6.6)–(6.8): √ VOV |B = √ =

2kn RD VDD + 1 − 1 kn RD

Av = −10 V/V Av = −k n VOV RD 1 2 k n VOV RD 2 −2 −10 = = VOV 1.5

VRD = ID RD =

2 × 1 × 24 × 5 + 1 − 1 = 0.61 V 1 × 24

Av VRD

VGS |B = Vt + VOV |B = 1 + 0.61 = 1.61 V

∴ VOV = 0.30 V

VDS |B = VOV |B = 0.61 V

Vt = VGS −VOV = 0.40 V kn =

(b) At Q, VOV = 0.5 V, thus

Av −10 = VOV RD −0.3 × 20

= 1.67 mA/V2 VGS |Q = 1 + 0.5 = 1.5 V ID =

1 1 2 kn VOV = × 1 × 0.52 = 0.125 mA 2 2

VDS |Q = VDD − ID RD

k n = k n ∴

W = 1.67 mA/V2 L

W = 8.33 L

= 5 − 0.125 × 24 = 2 V Av = −kn VOV RD = −1 × 0.5 × 24 = −12 V/V

(c) From part (a) above, the maximum instantaneous input signal while the transistor remains in saturation is 1.61 V and the corresponding output voltage is 0.61 V. Thus, the maximum amplitude of input sine wave is (1.61 − 1.5) = 0.11 V. That is, v GS ranges from 1.5 − 0.11 = 1.39 V, at which

Sedra, Smith, Carusone, Gaudet

6.12 At sat/triode boundary v GS B = VGS + vˆgs v DS B = VDS − vˆo vˆo = max downward amplitude , we get vˆo − Vt v DS B = v GS B − Vt = VGS + | Av | = VDS − vˆo VOV +

vˆo = VDS − vˆo | Av |

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–5

vˆo =

VDS − VOV 1 + | A1v |

(1)

v O = VDD − Vt + Vt

For VDD = 5 V, VOV = 0.5 V, and −v I

W = 1 mA/V2 , we use L −2(VDD − VDS ) Av = VOV k n

Q.E.D.

50 √ ( W/L )1 0.5 = 10, = For

5 ( W/L )2 0.5 √ Av = − 10 = −3.16 V/V

and Eq. (1) to obtain vˆo

vˆi

V DS

Av

1V

−16

471 mV 29.4 mV

1.5 V −14

933 mV 66.7 mV

−12 1385 mV

115 mV

2.5 V −10 1818 mV

182 mV

2V

( W/L )1 ( W/L )2

( W/L )1 ( W/L )2

6.14

5 V

For VDS = 1 V, Av = −16 = −k n VOV RD ∴ RD = 32 k

RC 1 k

IC

ID RD = 4 V, ID = 0.125 mA

vO v be

6.13

VBE

VDD

Q2 For IC = 0.5 mA, we have

vO vI

Av = −

Q1 iD

IC R C 0.5 = −20 V/V =− VT 0.025

VCE = VCC − IC RC = 5 − 0.5 = 4.5 V

given Vt1 = Vt2 = Vt W 1 [ VDD − v O − Vt ]2 For Q2 , iD = k n 2 L 2 For Vt ≤ v I ≤ v O + Vt , 1 W [ v I − V t ]2 for Q1 , iD = k n 2 L 1 Equating iD1 and iD2 W [ VDD − v O + Vt ]2 L 2 W [ v I − V t ]2 = L 1 ( W/L)1 [ VDD − v O − Vt ] = · [ v I − Vt ] ( W/L )2

Sedra, Smith, Carusone, Gaudet

max +v O = 5 − 4.5 = 0.5 V max −v O = 4.5 − 0.3 = 4.2 V Similarly → IC (mA)

VCE (V)

Av (V/V)

0.5

4.5

–20

0.5

4.2

1.0

4.0

–40

1.0

3.7

2.5

2.5

–100

2.5

2.2

4.0

1.0

–160

4.0

0.7

4.5

0.5

–180

4.5

0.2

Microelectronic Circuits, 8th International Edition

POS Neg vO (V) vO (V)

© Oxford University Press 2021

Chapter 6–6

but we have to make sure that the amplifier can support a positive peak amplitude of P, that is,

6.15

| Av |VT ≥ P, otherwise the transistor cuts off at or near the positive peaks. In the results obtained, tabulated below, VCEsat = 0.3 V and VCC is the nearest 0.5 V to VCCmin . Case Av (V/V) P (V) | Av |VT VCCmin VCC a

−20

0.2

0.5

1.0

1.0

b

−50

0.5

1.25

2.05

2.5

c

−100

0.5

2.5

3.3

3.5

On the verge of satuation

d

−100

1.0

2.5

3.8

4.0

VCE − vˆce = 0.3 V

e

−200

1.0

5.0

6.3

6.5

For linear operation, v ce = Av v be

f

−500

1.0

12.5

13.8 14.0

VCE − | Av vˆbe | = 0.3

g

−500

2.0

12.5

14.8 15.0

Av = −

IC R C VCC − VCE =− VT VT

( 5 − IC RC ) − | Av | × 5 × 10−3 = 0.3 But | Av | =

IC R C VT

Thus,

v CE = VCC − RC iC

IC RC = | Av |VT and 5 − | Av |VT − | Av | × 5 × 10−3 = 0.3 | Av |( 0.025 + 0.005 ) = 5 − 0.3 | Av | = 156.67. Note AV is negative. ∴ Av = −156.67 V/V Now we can find the dc collector voltage. Referring to the sketch of the output voltage, we see that VCE = 0.3 + | Av | 0.005 = 1.08 V 6.16 To obtain an output signal of peak amplitude P volts and maximum gain, we bias the transistor at VCE = VCEsat + P

VCE = VCC − RC IC dv CE Av = dv BE v BE =VBE , v CE =VCE VCE VBE /VT 1 e = −RC IS 1 + VA VT dv CE 1 −RC IS eVBE /VT dv BE VA 1 1 IC A = −RC IC − RC VCE VA v VT 1+ VA Thus, Av =

−IC RC /VT IC RC 1+ VA + VCE

Q.E.D.

Substituting IC RC = VCC − VCE , we obtain

The resulting gain will be Av = −

6.17 See figure on next page. v CE 6.18 iC = IS ev BE /VT 1 + VA VCE IC = IS eVBE /VT 1 + VA

VCC − VCE VT

Av = −

which results in VCC of

(VCC − VCE )/VT VCC − VCE 1+ VA + VCE

Q.E.D.

For VCC = 5 V, VCE = 3 V, and VA = 100 V,

VCC = VCE + | Av |VT Thus the minimum required VCC will be

Av (without the Early effect) = −

VCCmin = VCEsat + P + | Av |VT

= −80 V/V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

5−3 0.025

© Oxford University Press 2021

Chapter 6–7

This figure belongs to Problem 6.17. vO 0.5 V 0 B

vI

vI 0.3 V

vO

RC

5 V

Av (with the Early effect) =

−80 2 1+ 100 + 3

= −78.5 V/V 6.19 IC =

5 V

A

VCC − VCE 3−1 = = 1 mA RC 2

VCC − VCE 2 = −80 V/V =− Av = − VT 0.025 Using the small-signal voltage gain with v BE = +5 mV, we have v O = Av × v BE = −80 × 5 mV = −0.4 V

10% and overestimates | v O | for negative v BE by about 10%. IC 6.20 Av = − RC VT But Av ≡

v O − iC RC = = −gm RC v BE v BE

Thus, gm = IC /VT For a transistor biased at IC = 0.5 mA, we have

Using the exponential characteristic yields

gm =

iC = IC ev BE /VT

6.21

0.5 = 20 mA/V 0.025

= 1 × e5/25 = 1.22 mA Thus, iC = 0.22 mA and v O = − iC RC = −0.22 × 2 = −0.44 V Repeating for v BE = −5 mV as follows. Using the small-signal voltage gain:

Peak-to-peak v C swing = 4 − 1 = 3 V

v O = −80 × −5 = +0.4 V

For point Q at VCC /2 = 2.5 V, we obtain

Using the exponential characteristic:

VCE = 2.5 V,

iC = IC ev BE /VT

IB = 25 µA

= 1 × e−5/25 = 0.82 mA

IB =

Thus, iC = 0.82 − 1 = −0.18 mA and v O = 0.18 × 2 = 0.36 V

⇒ VBB = IB RB + 0.7 = 2.5 + 0.7 = 3.2 V

v BE

v O (exp) v O (linear)

IC = 2.5 mA

VBB − 0.7 = 25 µA RB

+5 mV −440 mV

−400 mV

6.22 See the graphical construction that follows. For this circuit:

−5 mV +360 mV

+400 mV

VCC = 10 V,

β = 100,

RC = 1 k,

VA = 100 V,

Thus, using the small-signal approximation underestimates | v O | for positive v BE by about

Sedra, Smith, Carusone, Gaudet

IB = 50 µA (dc bias),

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–8

At v CE = 0, iC = βiB

Thus the collector current varies as follows:

∴ iC = 50 × 100

2.91 mA

= 5 mA

8.15 mA

Given the base bias current of 50 µA, the dc or bias point of the collector current IC , and voltage VCE can be found from the intersection of the load line and the transistor line L1 of iB = 50 µA. Specifically:

5.24 mA

Eq. of L1 ⇒ iC = IC ( 1 + v CE /VA )

i 5.99 mA, peak to peak

2.16 mA 3.08 mA

And the collector voltage varies as follows:

= 5 ( 1 + v CE /100 )

3.08 V

= 5 + 0.05v CE

7.84 V

VCC − v CE Load line ⇒ iC = = 10 − v CE RC

v 5.99 V

4.76 V

∴ 10 − v CE = 5 + 0.05v CE

1.85 V

VCE = v CE = 4.76 V

2.91 V

IC = iC = 10 − v CE = 5.24 mA Now for a signal of 30-µA peak superimposed on IB = 50 µA, the operating point moves along the load line between points N and M. To obtain the coordinates of point M, we solve the load line and line L2 to find the intersection M, and the load line and line L3 to find N: For point M:

6.23 (a) VOV = 0.7 − 0.5 = 0.2 V ID =

1 W 2 1 k V = × 0.4 × 12.5 × 0.22 2 n L OV 2

= 0.1 mA VDS = 2 − 0.1 × 15 = 0.5 V (b)

iC = 8 + (8/100)v CE and iC = 10 − v CE ∴ iC M = 8.15 mA, v CE M = 1.85 V

gm =

2ID 2 × 0.1 = 1 mA/V = VOV 0.2

(c)

For point N:

Avo = −gm RD

iC = 2 + 0.02v CE and iC = 10 − v CE v CE N = 7.84 V, iC N = 2.16 mA

= −1 × 15 = −15 V/V (d) v ds = Avo × v gs = −15 × 0.015 sin wt v ds = −0.225 sin wt, V vˆ DS = VDS + 0.225 = (VDD − RD ID ) + 0.225 = (2 − 15 × 0.1) + 0.225 = 0.5 + 0.225 = 0.725 V vˇ DS = 0.5 − 0.225 = 0.275 V (e) From Eq. (6.28), 1 iD = 0.1 + kn VOV v gs + kn v 2gs , mA 2 = 0.1 + (0.4 × 12.5) × 0.2 × 0.015 sin wt +

1 × 0.4 × 12.5 × 0.0152 sin2 wt 2

= 0.1 + 0.015 sin wt + 2.5 1 1 × 0.0152 − cos 2 wt 2 2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–9

Thus, the second harmonic distortion is =

2.5 ×

6.25

VDD

× 0.0152 × 100 0.015 1 2

RD

= 1.9%

vd

In addition to the component at (2ω), there is a dc shift in the value of iD given by −4

2.5 × 0.015 × 0.5 = 2.8 × 10 2

vi

mA = 0.28 µA

vs

Compared to ID = 100 µA, this shift is negligible.

RS

VSS

6.24 Refer to Fig. 6.10. Av = −gm RD

v i = gm v gs

− 10 = −gm × 10

vd

D

To obtain maximum gain while allowing for a 0.2-V peak signal at the drain, we bias so that

RD

gmvgs

VDS = 0.2 + VOV

(1) G

vi

which indicates that the minimum signal at the drain will be equal to VOV . Since we can express the gain magnitude of

vgs

VDD − VDS 0.5VOV

1 gm

S vs RS

Thus, 10 =

1 + RS gm

v d = −gm v gs RD

⇒ gm = 1 mA/V

| Av |=

1.8 − VDS 0.5VOV

v s = +gm v gs RS

⇒ VDS = 1.8 − 5VOV

(2)

Solving Eqs. (1) and (2) together yields VOV = 0.27 V

∴

vs RS +gm RS = = 1 vi 1 + gm RS + RS gm

vd −RD −gm RD = = 1 vi 1 + gRS + RS gm

Thus, VDS = 0.47 V 1.8 − 0.47 = 0.133 mA 10 1 2 ID = kn VOV 2 2 × 0.133 ⇒ kn = = 3.65 mA/V2 0.272 W kn = kn L W 3.65 = 0.38 × L W = 9.6 ⇒ L ID =

Sedra, Smith, Carusone, Gaudet

6.26 Vt = 0.8 V VA = 20 V Given VDS = VGS = 1 V. Also, ID = 0.2 mA. VOV = 0.2 V, gm = ro =

2ID = 2 mA/V VOV

VA = 100 k ID

vo = −gm ( RG RL ro ) = −26.1 V/V vi

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–10

Since the drain voltage (+7 V) is higher than the gate voltage (+5 V), the transistor is operating in saturation.

VDD

From the circuit

I 200 µA RG 22 M

vo

RL 15 k

vi

For ID = 1 mA:

VOV increases by √

VD = VDD − ID RD = 15 − 0.5 × 16 = +7 V, as assumed

√ 1 = 5 to 0.2

Finally, VGS = 1.5 V, thus VOV = 1.5 − Vt = 1.5 − 1 = 0.5 V 1 1 2 = × 4 × 0.52 = 0.5 mA ID = k n VOV 2 2 which is equal to the given value. Thus the bias calculations are all consistent. 2ID 2 × 0.5 = 2 mA/V = (b) gm = VOV 0.5 ro =

5 × 0.2 = 0.45 V.

VA 100 = 200 k = ID 0.5

(c) See Fig. 2 on next page

VGS = VDS = 1.25 V gm = 4.47 mA/V, ro = 20 k and vo = −38.3 V/V vi 6.27 (a) Open-circuit the capacitors to obtain the bias circuit shown in Fig. 1, which indicates the given values.

15 V

(d) Rin = 10 M 5 M = 3.33 M Rin 3.33 v gs = = v sig Rin + Rsig 3.33 + 0.2 = 0.94 V/V vo = −gm (200 16 16) v gs = −2 × 7.69 = −15.38 V/V vo v gs vo = × = −0.94 × 15.38 v sig v sig v gs = −14.5 V/V

0.5 mA 10 M

16 k

6.28

5 V

7 V 1.5 V

5 M

0.5 mA

0.5 mA

5 k

7 k VC

Figure 1

vBE

From the voltage divider, we have VG = 15

5 =5V 10 + 5

With v BE = 0.700 V VC = VCC − RC IC

From the circuit, we obtain

= 5 − 5 × 0.5 = 2.5 V

VG = VGS + 0.5 × 7

For v BE = 705 mV ⇒ v be = 5 mV

= 1.5 + 3.5 = 5 V which is consistent with the value provided by the voltage divider.

Sedra, Smith, Carusone, Gaudet

iC = IC ev be /VT = 0.5 × e5/25 = 0.611 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–11

This figure belongs to Problem 6.27, part (c).

Rsig 200 k vo

vsig

vgs 10 M

gmvgs

5 M

200 k

16 k

16 k

Rin

Figure 2

v C = VCC − RC iC = 5 − 5 × 0.611 = 1.95 V

6.31 IC = 1 mA,

v ce = v C − VC = 1.95 − 2.5 = −0.55 V v ce 0.55 V =− Voltage gain, Av = v be 5 mV = −110 V/V Using small-signal approximation, we write Av = −gm RC where IC 0.5 mA = = 20 mA/V VT 0.025 V

gm =

IC 1 mA = 40 mA/V = VT 0.025 V

rπ =

β 100 = 2.5 k = gm 40 mA/V

ro =

VA 100 V = 100 k = IC 1 mA

α=

β 100 = = 0.99 β +1 100 + 1

re =

VT αVT 0.99 × 25 mV = 24.75 = = IE IC 1 mA

6.32 Refer to Fig. P6.32: VC = 3 − 0.2 × 10 = 1 V re =

IC 1 mA = 40 mA/V = 6.29 gm = VT 0.025 V

VT 25 mV = 125 = IE 0.2 mA

Replacing the BJT with the T model of Fig. 6.26(b), we obtain the equivalent circuit shown below.

α 0.99 re = = 25 gm 40 mA/V rπ =

VA = 100 V

gm =

Av = −20 × 5 = −100 V/V Thus, the small-signal approximation at this signal level (v be = 5 mV) introduces an error of −9.1% in the gain magnitude.

β = 100,

β 100 = 2.5 k = gm 40 mA/V

Av = −gm RC = −40 × 5 = −200 V/V vˆo = | Av |vˆbe = 200 × 5 mV = 1 V 6.30 Transistor α

a

b

1.000 0.990

β

∞

100

IC (mA)

1.00

0.99

IE (mA)

1.00

c

d

0.980

1

50

∞

e

f

g

0.990 0.900 0.940 100

9

15.9

1.00 1.00 0.248

4.5

17.5

1.00

1.02 1.00 0.25

5

18.6

IB (mA)

0

0.010

0.020

0

0.002

0.5

1.10

gm (mA/V)

40

39.6

40

40

9.92

180

700

re ()

25

25

24.5

25

100

5

1.34

rπ ()

∞

50

22.7

2.525 k 1.25 k ∞ 10.1 k

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–12

This figure belongs to Problem 6.31. C

B rp

vp

E rp 2.5 k,

C rp

ro

gmvp

ib B

E ro 100 k,

gm 40 mA/V

gmvp vp

b 100 C

C

B

ro

bib

ro

ai

B

ro

re

re i

E

E re 24.75 , gm 40 mA/V

ro 100 k, a 0.99

vo vo vπ = v sig v π v sig

v c = −ie × 10 k where vi vi ie = − = − re 0.125 k

= −gm RC

Thus,

= −gm rπ

vc 10 k = vi 0.125 k

=−

= 80 V/V

rπ rπ + Rsig

RC rπ + Rsig

βRC rπ + Rsig

Q.E.D.

6.34 Replacing the BJT with the T model of Fig. 6.26(b), we obtain the circuit shown below. 6.33 See figure below. vπ Rin ≡ = rπ ib vπ rπ = v sig rπ + Rsig

C

ix

aie B

v o = −gm v π RC vo = −gm RC vπ

v x ie re

The overall voltage gain can be obtained as follows:

E v r x ix

This figure belongs to Problem 6.33.

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–13

Since v x appears across re and ix = ie = small-signal resistance r is given by vx vx r≡ = = re ix ie

vx , the re

6.35 Refer to Fig. P6.53. The transistor is biased at IE = 0.33 mA. Thus re =

5 V

RE

Rsig 50

vsig

VT 25 mV = = 75 IE 0.33 mA

Rin re 50

vo RC

Replacing the BJT with its T model results in the following amplifier equivalent circuit. 5 V

To obtain maximum gain and the largest possible signal swing at the output for v eb of 10 mV, we select a value for RC that results in VC + | Av | × 0.01 V = +0.4 V which is the highest allowable voltage at the collector while the transistor remains in the active region. Since VC = −5 + IC RC −5 + 0.5RC The input resistance Rin can be found by inspection to be

then −5 + 0.5RC + gm RC × 0.01 = 0.4 Substituting gm = 20 mA/V results in

Rin = re = 75 To determine the voltage gain (v o /v i ) we first find ie : vi vi vi ie = − =− =− Rsig + re 150 0.15 k The output voltage v o is given by v o = −α ie (RC RL )

RC = 7.7 k The overall voltage gain achieved is vo Rin = × gm RC v sig Rin + Rsig 50 × 20 × 7.7 50 + 50 = 77 V/V =

= −0.99 ie × (12 12) = −0.99 × 6ie = −0.99 × 6 ×

−v i 0.15

Thus, vo = 39.6 V/V vi

6.37 Substituting v gs = Vgs sin ωt in Eq. (6.28), iD =

1 k n (VGS − Vt )2 + k n (VGS − Vt )Vgs sin ωt 2 1 2 + k n Vgs sin2 ωt 2

1 k n (VGS − Vt )2 + k n (VGS − Vt )Vgs sin ωt 2 1 1 2 1 + k n Vgs ( − cos 2 ωt) 2 2 2 Second-harmonic distortion 1 2 k n Vgs 4 = × 100 k n (VGS − Vt )Vgs =

6.36 re = 50 =

VT IE

⇒ IE = 0.5 mA Thus, 5 − VE = 0.5 mA RE where VE 0.7 V ⇒ RE = 8.6 k

Sedra, Smith, Carusone, Gaudet

=

1 Vgs × 100 4 VOV

Q.E.D.

For Vgs = 10 mV, to keep the second-harmonic distortion to less than 1%, the minimum overdrive voltage required is

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–14

VOV =

1 0.01 × 100 × = 0.25 V 4 1

6.38 (a) ID =

1 × 5(0.6 − 0.4)2 = 0.1 mA 2 VDS = VDD − ID RD = 1.8 − 0.1 × 10 = 0.8 V =

(b) gm = k n VOV = 5 × 0.2 = 1 mA/V (c) Av = −gm RD = −1 × 10 = −10 V/V

ro =

VA =

2ID = 0.25 V gm

∴ VGS = VOV + Vt = 0.75 V

1 k n (VGS − Vt2 ) 2

(d) λ = 0.1 V−1 ,

VOV =

1 = 10 V λ

6.41 For the NMOS device: W 2 1 ID = 100 = μn Cox VOV 2 L 1 10 2 = × 400 × × VOV 2 0.5 ⇒ VOV = 0.16 V 2ID 2 × 0.1 mA = 1.25 mA/V = VOV 0.16

gm =

VA = 5L = 5 × 0.5 = 2.5 V

VA 10 = = 100 k ID 0.1

ro =

Av = −gm (RD ro ) = −1(10 100) = −9.1 V/V 6.39 (a) The bias voltage VGS = 0.75 V. Thus VOV = 0.75 − 0.5 = 0.25 V W gm = kn VOV = kn VOV L

VA 2.5 = = 25 k ID 0.1

For the PMOS device: W 2 1 ID = 100 = μp Cox VOV 2 L 10 1 2 × VOV = × 100 × 2 0.5 ⇒ VOV = 0.316 V 2ID 2 × 0.1 = 0.63 mA/V = VOV 0.316

gm =

VA = 6L = 6 × 0.5 = 3 V

= 0.4 × 20 × 0.25 = 2 mA/V

VA 3 = 30 k = ID 0.1

1 1 2 = × 0.4 × 20 × 0.252 ID = kn VOV 2 2

ro =

= 0.25 mA

6.42 (a) Using the exponential characteristic: ic = IC ev be /VT − IC

ro =

VA 5V = 20 k = ID 0.25 mA

giving

(b) The bias current ID = 0.5 mA. Thus, 0.5 =

1 2 2 = 0.5 × 0.4 × 20 × VOV kn VOV 2

⇒ VOV = 0.35 V gm =

2ID 2 × 0.5 = 2.9 mA/V = VOV 0.35

VA 5V = 10 k ro = = ID 0.5 mA

6.40 Given μn Cox = 250 µA/V2 , Vt = 0.5 V, L = 0.5 µm For gm = 2 mA/V2 and ID = 0.25 mA,

W W = 32 gm = 2μn Cox ID ⇒ L L ∴ W = 16 µm

Sedra, Smith, Carusone, Gaudet

ic = ev be /VT − 1 IC

v be ic /I C ic /I C Error (mV) Exponential Small signal (%) +1

+0.041

+0.040

–2.4

–1

–0.039

–0.040

+2.4

+2

+0.083

+0.080

–3.6

–2

–0.077

–0.080

+3.9

+5

+0.221

+0.200

–9.7

–5

–0.181

–0.200

+10.3

+8

+0.377

+0.320

–15.2

–8

–0.274

–0.320

+16.8

+10

+0.492

+0.400

–18.7

–10

–0.330

–0.400

+21.3

+12

+0.616

+0.480

–22.1

–12

–0.381

–0.480

+25.9

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–15

(b) Using small-signal approximation: ic = gm v be = Thus,

IC · v be VT

Thus, the extreme values of rπ are and

ic v be = IC VT

See table on the previous page. For signals at ±5 mV, the error introduced by the small-signal approximation is 10%. The error increases to above 20% for signals at ±10 mV. 6.43 At IC = 0.5 mA, IC 0.5 mA = 20 mA/V gm = = VT 0.025 V β 100 = 5 k = rπ = gm 20 mA/V VT αVT re = = IE IC where 100 β = = 0.99 β +1 100 + 1 0.99 × 25 mV re = 50 0.5 mA At IC = 50 µA = 0.05 mA, IC 0.05 = 2 mA/V gm = = VT 0.025 β 100 = 50 k = rπ = gm 2 mA/V αVT 0.99 × 25 mV 500 re = = IC 0.5 mA α=

6.44 For gm = 30 mA/V, IC gm = ⇒ IC = gm VT = 30×0.025 = 0.75 mA VT β β = rπ = gm 30 mA/V For rπ ≥ 3 k, we require β ≥ 90

150 = 4.7 k. 32

6.46 VCC = 3 V,

3−1 = 1 mA 2 IC 1 mA gm = = 40 mA/V = VT 0.025 V

IC =

v be = 0.005 sin ωt ic = gm v be = 0.2 sin ωt, mA iC (t) = IC + ic = 1 + 0.2 sin ωt, mA v C (t) = VCC − RC iC = 3 − 2(1 + 0.2 sin ωt) = 1 − 0.4 sin ωt, V iB (t) = iC (t)/β = 0.01 + 0.002 sin ωt, mA Av =

vc 0.4 = −80 V/V =− v be 0.005

6.47 Since Vˆ be is the maximum value for acceptable linearity, the largest signal at the collector will be obtained by designing for maximum gain magnitude. This in turn is achieved by biasing the transistor at the lowest VCE consistent with the transistor remaining in the active mode at the negative peak of v o . Thus VCE − | Av |Vˆ be = 0.3 where we have assumed VCEsat = 0.3 V. Since VCE = VCC − IC RC and | Av | = gm RC =

VCC − IC RC − β gm

IC RC VT

Vˆ be IC RC = 0.3 V VT

which can be manipulated to yield

where IC gm = VT Nominally, gm = 40 mA/V. However, IC varies by ±20%, so gm ranges from 32 mA/V to 48 mA/V. Thus 50 to 150 rπ = 32 to 48 mA/V

Sedra, Smith, Carusone, Gaudet

VC = 1 V, RC = 2 k

then

That is, βmin = 90.

6.45 rπ =

50 = 1.04 k 48

IC RC =

VCC − 0.3 Vˆ be 1+ VT

(1)

Since the voltage gain is given by Av = −

IC R C VT

then

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–16 VCC − 0.3 VT + Vˆ be

Av = −

6.50 The large-signal model of Fig. 4.5(d) is shown in Fig. 1.

For VCC = 3 V and Vˆ be = 5 mV,

iB B

3 − 0.3 = 2.25 V IC RC = 5 1+ 25 Thus,

C

VCE = VCC − IC RC

E

= 3 − 2.25 = 0.75 V

Figure 1

Vˆ o = VCE − 0.3 = 0.75 − 0.3 = 0.45 V 3 − 0.3 = −90 V/V 0.025 + 0.005

Av = − Check:

Av = −gm RC = −

IC R C 2.25 =− = −90 V/V VT 0.025

Vˆ o = | Av | × Vˆ be = 90 × 5 = 450 mV = 0.45 V 6.48

For v BE undergoing an incremental change v be from its equilibrium value of VBE , the current iB changes from IB by an increment ib , which is related to v be by the incremental resistance of DB at the bias current IB . This resistance is given by VT /IB , which is rπ . The collector current βiB changes from βIB to β(IB + ib ). The incremental changes around the equilibrium or bias point are related to each other by the circuit shown in Fig. 2, which is the hybrid-π model of Fig. 7.25(b). Q.E.D.

gmvbe B

ib

ib B

vbe

biB

DB (ISB IS/b)

vBE

vbe

vbe

C

bib

rp

re

E

E Rin

Figure 2

v be − gm v be re 1 = v be − gm re

6.51 The large-signal T model of Fig. 4.5(b) is shown below in Fig. 1.

ib =

C

Since re =

α gm

ib = v be

aiE g

m

− gm

B

α 1−α = gm v be α gm v be = β Rin ≡

v be β = = rπ ib gm

vBE

iE

(I

DE

SE

IS a

)

E Figure 1 Q.E.D.

6.49 Refer to Fig. 6.27. v be α = v be ic = αie = α re re = gm v be

If iE undergoes an incremental change ie from its equilibrium or bias value IE , the voltage v BE will correspondingly change by an incremental amount v be (from its equilibrium or bias value VBE ), which is related to ie by the incremental

Q.E.D.

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–17

resistance of diode DE . The latter is equal to VT /IE , which is re .

C

The incremental change ie in iE gives rise to an incremental change αie in the current of the controlled source.

aie ib

B

The incremental quantities can be related by the equivalent circuit model shown in Fig. 2, which is the small-signal T model of Fig. 6.27(b). Q.E.D.

ie vi

re E Re

C

vo

Rin

aiE B

ie re

vbe

| Av | = gm RC = 50 × 2 = 100 V/V For v ce being 1 V peak to peak, 1V = 0.01 V peak to peak = 100 v be rπ

1 β +1

Q.E.D.

vo Re = vi Re + r e

Q.E.D.

6.54 Refer to Fig. P6.54. α=

200 β = = 0.995 β +1 201

IC = α × IE = 0.995 × 10 = 9.95 mA VC = IC RC = 9.95 × 0.1 k = 0.995 V 1 V

where

Replacing the BJT with its hybrid-π model results in the circuit shown below. IC 10 mA = 400 mA/V gm = VT 0.025 V

β 100 = = 2 k gm 50

Thus, ib =

1−α =

From the equivalent circuit, we see that v o and v i are related by the ratio of the voltage divider formed by re and Re :

6.52 v ce = | Av |v be

rπ =

But

Rin = (β + 1)(re + Re )

Figure 2

ib =

re + Re 1−α

Thus,

E

v be

Rin =

0.01 V = 0.005 mA peak to peak 2 k

rπ =

β 200 = 0.5 k = gm 400

Rib = rπ = 0.5 k Rin = 10 k 0.5 k = 0.476 k

6.53 Refer to Fig. P6.53. Replacing the BJT with the T model of Fig. 6.27(b) results in the following amplifier equivalent circuit: vi vi = Rin ≡ ib (1 − α)ie

Rin 0.476 vπ = 0.322 V/V = = v sig Rin + Rsig 0.476 + 1 vo = −gm RC = −400 × 0.1 = −40 V/V vπ vo = −40 × 0.322 = −12.9 V/V v sig

From the circuit we see that vi ie = r e + Re

For

Thus,

v o = ±0.4 V/V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–18

This figure belongs to Problem 6.54.

Rsig 1 k

vb

vo

vsig

vp

10 k

v sig =

RC

gmvp

Rib

Rin

vb = vπ =

rp

±0.4 = ∓0.01 V = ∓10 mV −40

This current is obtained by raising RE to the value found from 10 − 0.7 = 0.33 mA IE = RE

±0.4 = ∓31 mV −12.9

⇒ RE = 28.2 k 6.55 The largest possible voltage gain is obtained when RL → ∞, in which case IC VA vo = −gm ro = − v sig VT IC

Note that the dc voltage at the collector remains unchanged. The voltage gain now becomes αRC 0.99 × 14.1 vo = 186 V/V = = vi re 0.075

VA =− VT For VA = 25 V,

28.2 = 2.82. 10 The collector resistor becomes 2.82 × 5 = 14.1 k

Thus, n =

vo 25 =− v sig 0.025

= −1000 V/V For VA = 125 V,

vo 125 =− v sig 0.025

6.57 Refer to Fig. P6.57. Since β is very large, the dc base current can be neglected. Thus the dc voltage at the base is determined by the voltage divider,

= −5000 V/V

VB = 5 6.56 Refer to Fig. 6.31(a):

and the dc voltage at the emitter will be

Rin re

VE = VB − 0.7 = 1.8 V

To obtain an input resistance of 75 , re = 75 =

100 = 2.5 V 100 + 100

The dc emitter current can now be found as VE 1.8 = 0.5 mA = IE = RE 3.6

VT IE

Thus,

and

25 mV = 0.33 mA IE = 75

IC IE = 0.5 mA

This figure belongs to Problem 6.55. B

vsig

vp

rp

gmvp

ro

C

vo

RL very high

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–19

This figure belongs to Problem 6.57.

vo2 aie

RC

ie re vo1

vi

100 k

100 k

RE

Replacing the BJT with the T model of Fig. 6.27(b) results in the equivalent circuit model for the amplifier (see figure above). vi ie = RE + r e v o1

6.58 (a)

Rin = 0.95 Rin + 100 ⇒ Rin = 1.9 M

RE = ie R E = v i RE + r e

v o1 RE = vi RE + r e

v o2 αRC =− vi RE + r e

(b) With RL = 2 k, v o = Av o v i

Q.E.D.

v o2 = −αie RC = −α

vi RC RE + r e

2 2 + Ro

With RL = 1 k, v o = Av o v i

Q.E.D.

1 1 + Ro

Thus the change in v o is 2 1 v o = Av o v i − 2 + Ro 1 + Ro

For α 1, re =

Rin = 0.95 Rin + Rsig

VT 25 mV = 50 = IE 0.5 mA

To limit this change to 5% of the value with RL = 2 k, we require 2 2 1 = 0.05 − 2 + Ro 1 + Ro 2 + Ro

v o1 3.6 = 0.986 V/V = vi 3.6 + 0.05 v o2 3.3 = 0.904 V/V =− vi 3.6 + 0.05

⇒ Ro =

If v o1 is connected to ground, RE will in effect be short-circuited at signal frequencies, and v o2 /v i will become

1 k = 111 9

(c) Gv = 10 =

v o2 αRC 3.3 = −66 V/V =− =− vi re 0.05

=

RL Rin Av o Rin + Rsig RL + Ro

1.9 2 × Av o × 1.9 + 0.1 2 + 0.111

This figure belongs to Problem 6.58.

Rsig

Ro

vsig

vi

Rin

Avovi

RL

Sedra, Smith, Carusone, Gaudet

vo

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–20

1 2 × 400 × 10 × VOV 2 ⇒ VOV = 0.4 V

⇒ Av o = 11.1 V/V

320 =

The values found about are limit values; that is, we require

2ID 2 × 0.32 = 1.6 mA/V = VOV 0.4

Rin ≥ 1.9 M

gm =

Ro ≤ 111

Av o = −gm RD = −1.6 × 10 = −16 V/V

Av o ≥ 11.1 V/V

Ro = RD = 10 k

6.59 Refer to Fig. 1. v o = Gm Ro Avo = v i RL =∞

Gv = Av o

RL RL + Ro

= −16 ×

10 = −8 V/V 10 + 10

Peak value of v sig =

= 10 mA/V × 100 k = 1000 V/V vi vo vo = × Gv = v sig v sig vi

0.2 V = 25 mV. 8

6.61 (a) See figure below. (b) gm1 = gm2 =

Rin = × Gm (RL Ro ) Rin + Rsig

2ID 2 × 0.3 = = 3 mA/V VOV 0.2

RD1 = RD2 = 10 k

100 × 10 × (100 100) = 100 + 100

RL = 10 k v gs2 vo × Gv = v gs1 v gs2

= 0.5 × 10 × 50 = 250 V/V

= −gm1 RD1 × −gm2 (RD2 RL ) 6.60 Rin = ∞

= 3 × 10 × 3 × (10 10)

W 2 1 ID = μn Cox VOV 2 L

= 450 V/V

This figure belongs to Problem 6.59.

Rsig 100 k vsig

vi

Rin 100 k

Gmvi

RL vo

Ro 100 k

Gm 10 mA/V

RL 100 k

Figure 1 This figure belongs to Problem 6.61.

Rsig 200 k vsig

vgs1

Sedra, Smith, Carusone, Gaudet

gm1vgs1

RD1 vgs2

gm2vgs2

Microelectronic Circuits, 8th International Edition

RD2

RL vo 10 k

© Oxford University Press 2021

Chapter 6–21

6.62 gm =

IC 0.5 mA = = 20 mA/V VT 0.025 V

β 100 = = 5 k rπ = gm 20 mA/V Rin = rπ = 5 k Ro = RC = 10 k

Ro = RC = 15 k Av = Av o

RL RL + Ro

= −60 ×

15 = −30 V/V 15 + 15

Gv =

Av o = −gm RC = −20 × 10 = −200 V/V A v = Av o

RL 10 = −200 × RL + Ro 10 + 10

= −100 V/V Gv = =

Rin Av Rin + Rsig

=

Rin × Av Rin + Rsig

25.25 × (−30) = −21.5 V/V 25.25 + 10

vˆπ = 5 mV ⇒ vˆb = vˆπ

r e + Re re

50 + 200 = 25 mV 50 Rsig + Rin = vˆ b Rin

⇒ vˆ b = 5 ×

5 × −100 5 + 10

vˆsig

= −33.3 V/V

= 25 ×

For vˆπ = 5 mV, vˆsig can be found from Rin 5 = vˆsig × vˆπ = vˆsig × Rin + Rsig 5 + 10

10 + 25.25 = 35 mV 25.25

vˆo = vˆsig × | Gv | = 21.5 × 35 753 mV

⇒ vˆsig = 15 mV Correspondingly, vˆo will be vˆo = Gvvˆsig = 15 × 33.3 = 500 mV = 0.5 V

6.65 Adding a resistance of 100 in series with the 100- Rsig changes the input voltage divider ratio from 1/gm 1/gm to (1/gm ) + 100 1/gm + 200

6.63 Including Rs reduced the gain by a factor of 2, thus

Since this has changed the overall voltage gain from 12 to 10, then

1 + gm Rs = 2

(1/gm ) + 200 12 = , where gm is in A/V 10 (1/gm ) + 100

⇒ gm =

1 1 = 5 mA/V = Rs 0.2

The gain without Rs is −20 V/V. Thus, 20 = 4 k. To obtain a −gm RD = −20 ⇒ RD = 5 gain of −16 V/V, we write 16 =

20 20 = 1 + gm Rs 1 + 5Rs

0.2 A/V = 2.5 mA/V 80 For ID = 0.25 mA ⇒ gm =

2.5 =

2ID 2 × 0.25 = VOV VOV

⇒ VOV = 0.2 V

⇒ Rs = 50 6.64 gm = re

IC 0.5 = 20 mA/V = VT 0.025

1 = 50 gm

6.66 Refer to the circuit in Fig. P6.66. Since Rsig re , most of isig flows into the emitter of the BJT. Thus ie isig

Rin = (β + 1)(re + Re )

and

= 101(50 + 200) = 25.25 k

ic = αie isig

Av o

αRC 0.99 × 15 =− =− r e + Re 0.25

−60 V/V

Sedra, Smith, Carusone, Gaudet

Thus, v o = ic RC = isig RC

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–22

6.67

IE =

VT 25 mV = = 1.25 mA re 20

At the peak of the output sine wave, we have

Rsig

ˆie = vˆo = 0.5 = 0.25 mA RL 2

i

Thus,

i vsig

1 gm

RL

50 mV (peak) 0.5 V (peak)

iEmax = 1.25 + 0.25 = 1.5 mA and iEmin = 1.25 − 0.25 = 1.0 mA From the figure, we have Gv =

vo = v sig

From the figure above, we have 1 = 0.1 × RL gm

=

= 0.1 × 2 = 0.2 k

RL RL + r e +

2 2 + 0.02 +

200 101

Rsig β +1

= 0.5 V/V

gm = 5 mA/V gm = 2k n ID 5 = 2 × 5 × ID

Thus,

ID = 2.5 mA

6.69 Refer to Fig. 6.48(b). To find Gvo , we remove RL . The result is v o = v sig and

vˆsig =

At the peak of the sine wave,

vˆo 0.5 V =1V = Gv 0.5 V/V

Gvo = 1 V/V

0.5 V = 0.25 mA, thus 2 k iDmax = ID + 0.25 = 2.75 mA

To obtain Rout , we write by inspection

iDmin = ID − 0.25 = 2.25 mA

Rout = re2 +

id =

vˆsig = vˆgs + vˆo = 0.05 + 0.5 = 0.55 V

We can find Gv as Gv = Gvo

6.68 Rsig/(b + 1)

re

vsig

RL

vbe vo

re1 + Rsig /(β1 + 1) β2 + 1

RL RL + Rout

To obtain numerical values, we use IC2 = 10 mA. Thus, IE2 IC2 = 10 mA, and VT 25 mV = 2.5 = IE2 10 mA 10 IE2 = 0.1 mA = β +1 101

re2 = IE1

re1 =

VT 25 = 250 = IE1 0.1 6

vˆo = 0.5 V

Rout = 2.5 +

RL = 2 k

250 + 10 101 101

= 105

vˆbe = 5 mV

1000 = 0.9 V/V 1000 + 105

From the figure above we see that

Gv = 1 ×

re 5 mV = RL 500 mV

6.70 (a) Refer to Fig. P6.70.

⇒ re =

RL = 20 100

Sedra, Smith, Carusone, Gaudet

−ic RC vc = v sig ib RB + ie (re + RE )

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–23

=−

RC ie (re + RE ) RB + ib

ic ib

= −β

ii =

io v o Rsig + Rin = ii v sig RL

RC RB + (β + 1)(re + RE )

= Gv

ve ie RE = v sig ib RB + ie (re + RE ) =

RB + r e + RE β +1

(b)

vc

Now, setting RL = ∞ in the equivalent circuit in Fig. 1(b), we can determine Gv o from Rin Av o Gv o = R +R

RC

ie = −

RE

in

sig RL =∞

Denoting Rin with RL = ∞ as Ri , we can express Gv o as Ri Av o Q.E.D. Gv o = Ri + Rsig From the equivalent circuit in Fig. 1(a) on the next page, the overall voltage Gv can be obtained as RL Q.E.D. Gv = Gv o RL + Rout

Y

v sig r e + RE

v c = −ic RC = −αie RC vc −ic RC RC = =α v sig ie (re + RE ) r e + RE

6.73 Refer to Fig. P6.73. To determine Rin , we simplify the circuit as shown in Fig. 1 on the next page, where vi vi = R1 Rin , where Rin ≡ Rin ≡ ii if v i = if Rf + (if − gm v i )(R2 RL )

6.71 See figure below. Gv =

20 + 100 = 4762 A/A 2

6.72 See figure on the next page. v o Gv o = v sig RL = ∞

ic

vsig

Rsig + Rin RL

= 79.4 ×

RE

ie

v sig Rsig + Rin

RL Rin Av o Rin + Rsig RL + Ro

Thus, v i [1 + gm (R2 RL )] = if [Rf + (R2 RL )] Rin ≡

2 100 × 100 × = 100 + 20 2 + 0.1

vi Rf + (R2 RL ) = if 1 + gm (R2 RL )

and

= 79.4 V/V

Rin = R1 Rin Rf + (R2 RL ) = R1

1 + gm (R2 RL )

vo io = RL

Q.E.D.

This figure belongs to Problem 6.71.

Rsig 20 k

Ro 200

vsig

vi

Rin 200 k

Avovi

RL vo 2 k

Avo 100

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–24

This figure belongs to Problem 6.72.

Rsig

Rout

vsig

Rin

vi

Gv ovsig

RL

vo

(a) Ro

Rsig vsig

Rin

vi

Av ovi

RL

vo

(b) Figure 1

This figure belongs to Problem 6.73.

Figure 1 To determine Av o , we open-circuit RL and use the circuit in Fig. 2, where

Rf

if

vi

gmvi

R1

Figure 2 if = gm v i +

vo R2

vo Rf + v o v i = if Rf + v o = gm v i + R2 Rf v i (1 − gm Rf ) = v o 1 + R2

Sedra, Smith, Carusone, Gaudet

vo R2 R2

Thus, Av o ≡

vo 1 − gm Rf = Rf vi 1+ R2

which can be manipulated to the form

vo

Av o = −gm R2

1 − 1/gm Rf 1 + (R2 /Rf )

Q.E.D.

Finally, to obtain Ro we short-circuit v i in the circuit of Fig. P6.73. This will disable the controlled source gm v i . Thus, looking between the output terminals (behind RL ), we see R2 in parallel with Rf , Ro = R2 Rf

Q.E.D.

For R1 = 100 k, Rf = 1 M, gm = 100 mA/V R2 = 100 and RL = 1 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–25

Rin = 100

1000 + (0.1 1) = 100 99.1 1 + 100(0.1 1)

= 49.8 k Without Rf present (i.e., Rf = ∞), Rin = 100 k. Av o = −100 × 0.1

1 − (1/100 × 1000) 0.1 1+ 1000

−10 V/V Without Rf , Av o = −10 V/V and Ro = 0.1 1000 0.1 k = 100

6.76 | Gv | =

RL (Rsig /β) + (1/gm )

RL = 10 k, Rsig = 10 k, gm = 1 = 40 mA/V 0.025 Nominal β = 100 =

(a) Nominal | Gv | =

(b) β = 50, | Gv | =

Thus the only parameter that is significantly affected by the presence of Rf is Rin , which is reduced by a factor of 2!

= 44.4 V/V

RL Rin Av o Rin + Rsig RL + Ro

10 (10/150) + 0.025

β = 150, | Gv | =

Thus, | Gv | ranges from 44.4 V/V to 109.1 V/V.

1 49.8 × −10 × 49.8 + 100 1 + 0.1

= −3 V/V Without Rf , Gv =

10 (10/50) + 0.025

= 109.1 V/V

With Rf , Gv =

10 (10/100) + 0.025

= 80 V/V

Without Rf , Ro = 100 .

Gv =

IC VT

(c) For | Gv | to be within ±20% of nominal (i.e., ranging between 64 V/V and 96 V/V), the corresponding allowable range of β can be found as follows: 64 =

1 100 × −10 × = −4.5 V/V 100 + 100 1 + 0.1

6.74 Rsig = 1 M, RL = 10 k

10 (10/βmin ) + 0.025

⇒ βmin = 76.2 96 =

10 (10/βmax ) + 0.025

gm = 2 mA/V, RD = 10 k

⇒ βmax = 126.3

Gv = −gm (RD RL )

(d) By varying IC , we vary the term 1/gm in the denominator of the | Gv | expression. If β varies in the range 50 to 150 and we wish to keep | Gv | within ±20% of a new nominal value of | Gv | given by

= −2(10 10) = −10 V/V

6.75 RD = 2RL = 30 k

| Gv nominal =

VOV = 0.25 V

10 (10/100) + (1/gm )

then

Gv = −gm (RD RL ) −10 = −gm (30 15)

0.8 | Gv nominal =

⇒ gm = 1 mA/V

That is,

gm = 1=

10 (10/50) + (1/gm )

10 8 = 0.1 + (1/gm ) 0.2 + (1/gm )

2ID VOV

2 × ID 0.25

⇒

1 = 0.3 or gm = 3.33 mA/V gm

⇒ ID = 0.125 mA = 125 µA

| Gv nominal =

If RD is reduced to 15 k, Gv = −gm (RD RL )

| Gv min =

= −1 × (15 15) = −7.5 V/V

= 20 V/V (−20% of nominal)

Sedra, Smith, Carusone, Gaudet

10 = 25 V/V 0.1 + 0.3

10 0.2 + 0.3

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–26

We need to check the value obtained for β = 150, | Gv max =

6.79 The gain magnitude is reduced by a factor of (1 + gm Rs ). Thus, to reduce the gain from −10 V/V to −5 V/V, we write

10 = 27.3 V/V 10/150 + 0.3

2 = 1 + gm R s

which is less than the allowable value of 1.2 | Gv nominal = 30 V/V. Thus, the new bias current is

⇒ Rs =

IC = gm × VT = 3.33 × 0.025 = 0.083 mA | Gv nominal = 25 V/V

6.80 Rin = (β + 1)(re + Re )

1 1 = = 0.5 k gm 2

15 = 75(re + Re ) 15 k = 200 75 Rin re vˆπ = vˆsig Rin + Rsig re + Re re 15 5 = 150 × 15 + 30 re + Re re = 0.1 ⇒ r e + Re r e + Re =

6.77 (a) See figure below. (b) RC1 = RC2 = 10 k

Rsig = 10 k

RL = 10 k gm1 = gm2 =

IC 0.25 mA = 10 mA/V = VT 0.025 V

rπ 1 = rπ 2 =

β 100 = 10 k = gm 10

But re + Re = 200 , thus re = 20

vπ1 rπ 1 10 = 0.5 V/V = = v sig rπ 1 + Rsig 10 + 10

which requires a bias current IE of

vπ2 = −gm1 (RC1 rπ 2 ) = −10(10 10) vπ1

IE =

VT 25 mV = 1.25 mA = re 20

= −50 V/V

IC IE = 1.25 mA

vo = −gm2 (RC2 RL ) vπ2

Re = 180 Gv =

= −10(10 10) = −50 V/V vo vo vπ2 vπ1 = × × v sig vπ2 vπ1 v sig

Rin Rin + Rsig

−α × Total resistance in collector Total resistance in emitter −0.99 × 6 15 × = 15 + 30 0.2 ×

= −50 × −50 × 0.5 = 1250 V/V

−10 V/V

6.78 gm effective =

vˆ0 = 0.15 × | Gv | = 1.5 V gm 1 + gm Rs

6.81 Using Eq. (6.113), we have

5 2= 1 + 5Rs

Gv = −β

⇒ Rs = 0.3 k = 300 −

R C RL Rsig + (β + 1)(re + Re )

RC RL (Rsig /β) + (re + Re )

This figure belongs to Problem 6.77.

Rsig 10 k vsig

vp1

rp1

vp2 gm1vp1

Sedra, Smith, Carusone, Gaudet

rp2

vo

gm2vp2

RC1

Microelectronic Circuits, 8th International Edition

RC2

RL

© Oxford University Press 2021

Chapter 6–27

| Gv | =

10 (10/β) + 0.025 + Re

= 5.6 V/V For Rin = Rsig = 0.25 k

Without Re , | Gv | =

10 (10/β) + 0.025

For the nominal case, β = 100, Gv

nominal

=

10 = 80 V/V 0.1 + 0.025

For β = 50, Gv

low

=

10 = 44.4 V/V 0.2 + 0.025

For β = 150, Gv

high

To limit the range of | Gv | to ±20% of a new nominal value, we connect a resistance Re and find its value as follows. With Re , 10 Gv = nominal (10/100) + 0.025 + Re 10 0.125 + Re

low

=

10 0.225 + Re

To limit this value to −20% of Gv nominal , we use 10 10 = 0.8 × 0.225 + Re 0.125 + Re ⇒ Re = 0.275 k = 275 With this value of Re , 10 Gv = 25 V/V = nominal 0.125 + 0.275 10 Gv = low 0.225 + 0.275 = 20 V/V (−20% of nominal) Gv

high

=

10 (1/15) + 0.025 + 0.275

= 27.3 V/V (+9.1% of nominal)

6.82 Rin = Gv = =

1 1 = 0.2 k = gm 5 mA/V

Rin × gm (RD RL ) Rin + Rsig

0.2 × 5(5 5) 0.2 + 0.25

Sedra, Smith, Carusone, Gaudet

and, with α 1, IC

VT 25 mV = = 0.5 mA re 50

gm = IC /VT = 20 mA/V Gv =

Rin gm (RC RL ) Rin + Rsig

Gv =

50 × 20 × (5 5) 50 + 50

= 25 V/V

6.84 Rin = re =

Now, β = 50, Gv

6.83 For Rin = Rsig = 50 , re = 50

10 = 109.1 V/V = (1/15) + 0.025

Thus, | Gv | ranges from 44.4 V/V to 109.1 V/V with a nominal value of 80 V/V. This is a range of −44.5% to +36.4% of nominal.

=

1 = 0.25 ⇒ gm = 4 mA/V gm Since gm = 2k n ID , then to change gm by a 4 factor = 0.8, ID must be changed by a factor of 5 (0.8)2 = 0.64.

VT 25 mV = = 125 IE 0.2 mA

gm =

IC 0.2 mA = 8 mA/V VT 0.025 V

Gv =

Rin gm (RC RL ) Rin + Rsig

=

0.125 × 8(10 10) = 8 V/V 0.125 + 0.5

vˆπ = vˆsig

Rin Rin + Rsig

10 = vˆsig

0.125 0.125 + 0.5

⇒ vˆsig = 50 mV vˆo = Gv vˆsig = 8 × 50 = 400 mV = 0.4 V

6.85 Av =

RL RL + Ro

Av nominal =

2 2 + Ro

Av low =

1.5 1.5 + Ro

Av high =

5 5 + Ro

For Av high = 1.1 Av nominal

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–28 5 1.1 × 2 = 5 + Ro 2 + Ro ⇒ Ro = 0.357 k 2 = 0.85 V/V Av nominal = 2.357 5 Av high = 5.357 = 0.93

(b)

Rsig

ai

B i

re

vsig

(+10% above nominal) 1.5 Av low = 1.5 + 0.357

RL

vbe vo

Rin

= 0.81 (−5% from nominal) 1 = Ro = 0.357 k gm

vˆbe = 10 mV

⇒ gm = 2.8 mA/V

vˆo =

To find ID , we use

=

gm =

2k n ID

⇒ ID = g2m /2k n =

2.82 = 0.392 mA 2 × 10

ID =

1 2 k n VOV 2

0.392 =

1 2 × 10 × VOV 2

⇒ VOV = 0.28 V

RL × vˆbe re

200 × 10 5 = 400 mV = 0.4 V

vˆsig =

(c) Gv o = 1 V/V Rout = re +

re =

VT VT 25 =5 = IE IC 5

(a) Rin = (β + 1) (re + RL ) = 101 × (5 + 200) = 20.7 k

Rsig 10,000 =5+ β +1 101

= 104 Thus, Gv = Gv o =1×

6.86 IC = 5 mA

vˆo 0.4 = 0.615 V = Gv 0.65

RL RL + Rout

200 = 0.66 V/V 200 + 104

which is approximately the same value obtained in (a) above. For RL = 150 , Gv = Gv o =1×

RL RL + Rout

150 = 0.59 V/V 150 + 104

Rin 20.7 vb = = v sig Rin + Rsig 20.7 + 10 Rsig β +1

= 0.67 V/V

6.87 Rout = re +

vo vb vo = × v sig v sig vb

125 = re +

10,000 β +1

(1)

= 0.67 ×

RL RL + r e

225 = re +

20,000 β +1

(2)

= 0.67 ×

200 200 + 5

Subtracting Eq. (1) from Eq. (2), we have

= 0.65 V/V

Sedra, Smith, Carusone, Gaudet

100 =

10,000 β +1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–29

β + 1 = 100

With RL = 500 = 0.5 k,

Substituting in Eq. (1) yields

Gv =

10,000 125 = re + 100 ⇒ re = 25 Gv =

=

(0.5 ro ) +

1 gm

= 0.49

ro 1 = gm 49

Rsig β +1

Substituting in Eq. (2) and solving for ro gives

1000 10, 000 1000 + 25 + 100

ro = 25 k

= 0.89 V/V

6.88 With the Early effect neglected,

Thus 25 1 k = gm 49

Gv = −gm RC = −100 V/V

⇒ gm = 1.96 mA/V

With the Early effect taken into account, the effective resistance in the collector is reduced from RC = 10 k to (RC ro ), where

6.90 Adapting Eq. (6.114) gives Gv = −β

VA 125 V = 125 k = ro = IC 1 mA

=−

(RC ro ) = 10 100 = 9.1 k Thus, Gv becomes Gv = −100 ×

(2)

From Eq. (1), we have

RL RL + r e +

(0.5 ro )

9.1 k 10 k

=−

= −91 V/V

RC RL r o Rsig + (β + 1)re

RC RL r o Rsig β +1 + re β β RC RL r o Rsig 1 + β gm

Thus, 6.89

| Gv | =

i 0

Rsig

vsig

where ro and

ro

G

ro =

i 1 gm

vg

RL

(1)

1 are in kilohms and are given by gm

VA 25 V = IC IC mA

(2)

1 VT 0.025 V = = gm IC IC mA

vo

10 ro 1 0.1 + gm

(3)

I C (mA) 1/gm (k) ro (k) | Gv | (V/V) 0.1

0.250

250

27.5

v g = v sig

0.2

0.125

125

41.2

Noting that ro appears in effect in parallel with RL , v o is obtained as the ratio of the voltage divider formed by (1/gm ) and (RL ro ),

0.5

0.050

50

55.6

1.0

0.025

25

57.1

1.25

0.020

20

55.6

vo vo = = Gv = v sig vg

(RL ro ) (RL ro ) +

With RL removed, ro Gv = = 0.98 1 ro + gm

Sedra, Smith, Carusone, Gaudet

1 gm

Q.E.D.

(1)

Observe that initially | Gv | increases as IC is increased. However, above about 1 mA this trend reverses because of the effect of ro . From the table we see that gain of 50 is obtained for IC between 0.2 and 0.5 mA and also for IC above 1.25 mA. Practically speaking, one normally uses

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–30

the low value to minimize power dissipation. The required value of IC is found by substituting for ro and 1/gm from Eqs. (2) and (3), respectively, in Eq. (1) and equating Gv to 50. The result (after some manipulations) is the quadratic equation. IC2

− 2.25IC + 0.625 = 0

and determine RG2 from 4V RG2 = RG1 5V ⇒ RG2 = 0.8RG1 = 0.8 × 22 = 17.6 M

The two roots of this equation are IC = 0.325 mA and 1.925 mA; our preferred choice is IC = 0.325 mA.

Using only two significant figures, we have

6.91

Note that this will cause VG to deviate slightly from the required value of 5 V. Specifically,

VDD 9 V

VG = VDD

ID RD

RG1

VD VG VS RS

RG2

RG2 = 18 M

ID = 0.2 mA

=9×

RG2 RG2 + RG1

18 = 4.05 V 22 + 18

It can be shown that the resulting ID will be ID = 0.204 mA, which is very close to the desired 0.2 mA. Since VD = VDD − ID RD +6 V and VG 4 V, and the drain voltage can go down to VG − Vt = 3.3 V, the drain voltage is 2.7 V above the value that causes the MOSFET to leave the saturation region.

6.92

5 V

1 2 ID = k n VOV 2 0.2 =

1 2 × 5 × VOV 2

RD

⇒ VOV = 0.28 V

VG 0

ID

0

VD

VGS = Vt + VOV = 0.7 + 0.28 = 0.98 1 V Now, selecting VS =

VDD = 3 V, 3

RG 10 M

ID RS = 3 RS =

RS

3 = 15 k 0.2

5 V

Also, ID RD =

For ID = 0.5 mA

VDD =3V 3

⇒ RD =

0.5 =

3 = 15 k 0.2

1 2 k n VOV 2

1 2 × 1 × VOV 2 ⇒ VOV = 1 V =

VG = VS + VGS =3+1=4V

VGS = Vt + VOV = 1 + 1 = 2 V

Thus the voltage drop across RG1 (5 V) is larger than that across RG2 (4 V). So we select

Since

RG1 = 22 M

which leads to

Sedra, Smith, Carusone, Gaudet

VG = 0 V,

VS = −VGS = −2 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–31

RS =

VS − (−5) −2 + 5 = 6 k = IC 0.5

6.94

VDD 10 V

VD is required to be halfway between cutoff (+5 V) and saturation (0 − Vt = −1 V). Thus

IG

VD = +2 V

R1

and RD =

5−2 = 6 k 0.5

6.93

RS ID

VS

VG

R2

VDD

RG1

RD ID ID = 1 mA and VD = 3 V

VG 5 V ID RG2

VD 3 V ID 1 mA RD

RS 3 k

Thus, RD =

VD 3V = 3 k = ID 1 mA

For the transistor to operate 1 V from the edge of saturation VD = VG + |Vt | − 1 Thus,

VS = ID RS = 3ID

3 = VG + |Vt | − 1

VGS = 5 − VS = 5 − 3ID

VG + |Vt | = 4 V

1 ID = k n (VGS − Vt )2 2 =

(a) |Vt | = 1 V and k p = 0.5 mA/V2 VG = 3 V

1 × 2(5 − 3ID − 1)2 2

R2 =

VG 3V = = 0.3 M IG 10 µA

R1 =

VDD − VG 7V = 0.7 M = IG 10 µA

= 16 − 24ID + 9ID2 9ID2

− 25I + 16 = 0

VD = 3 V

ID = 1.78 mA or 1 mA

RD = 3 k

The first answer is physically meaningless, as it would result in VS = 5.33 V, which is greater than VG , implying that the transistor is cut off. Thus, ID = 1 mA. If a transistor for which k n = 3 mA/V2 is used, then ID =

1 k p (VSG − |Vt |)2 2 1 1 = × 0.5(VSG − 1)2 2 ⇒ VSG = 3 V

ID =

VS = VG + 3 = 3 + 3 = 6 V

1 × 3(5 − 3ID − 1)2 2

RS =

= 1.5(16 − 24I + 9ID2 )

VDD − VS ID

10 − 6 = 4 k 1

9ID2 − 24.67ID + 16 = 0

=

whose physically meaningful solution is

(b) |Vt | = 2 V and k p = 1.25 mA/V2

ID = 1.05 mA

VG = 4 − |Vt | = 2 V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–32

R2 =

VG 2V = = 0.2 M IG 10 µA

RG1 1.2 = ⇒ RG1 = 20.3 M RG2 1.3

R1 =

VDD − VG 8V = 0.8 M = IG 10 µA

Specifying all resistors to two significant digits, we have RD = 2.5 k, RG1 = 22 M, and RG1 = 20 M.

VD = 3 V RD = 3 k 1 k p (VSG − |Vt |)2 2 1 1 = × 1.25(VSG − 2)2 2 VSG = 3.265 V

6.96

ID =

VCC 9 V 0.05 mA

0.5 mA 3V RC 3V

R1

VS = VG + 3.265 = 2 + 3.265 = 5.265 V 10 − 5.265 = 4.7 k RS = 1

VB

IB

IR2 R2

6.95

RE

3V

VDD 5 V ~ – 1 mA RD RG1

IE =

VD IG

1 mA

RG2

ID =

1 2 k n VOV 2

1=

1 2 × 8VOV 2

0.5 = 0.506 mA 0.988 VE 3V = = 5.93 k RE = IE 0.506 mA 3 3V = 6 k = RC = IC 0.5 mA IE =

VB = VE + 0.7 = 3.7 V 9 − 3.7 = 106 k R1 = 0.05 0.5 IC = = 0.0063 mA IB = β 80

⇒ VOV = 0.5 V Since the transistor leaves the saturation region of operation when v D < VOV , we select VD = VOV + 2

IR2 = 0.05 − IB = 0.05 − 0.0063 = 0.0437 mA VB 3.7 = 84.7 k R2 = = IR2 0.0437 Consulting the table of 5-% resistors in Appendix J we select:

VD = 2.5 V Since IG ID , we can write VDD − VD 5 − 2.5 = 2.5 k = RD = ID 1 VGS = Vt + VOV = 0.8 + 0.5 = 1.3 V Thus the voltage drop across RG2 is 1.3 V and that across RG1 is (2.5 − 1.3) = 1.2 V. Thus RG2 is the larger of the two resistances, and we select RG2 = 22 M and find RG1 from

Sedra, Smith, Carusone, Gaudet

β 80 IC where α = = = 0.988 α β +1 81

RE = RC = 6.2 k R1 = 100 k

R2 = 82 k

This resistor combination will result in

IE =

2 9 R1R+R − 0.7 2

R2 RE + ( RR11+R )/(β + 1) 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–33

which yields

For β = ∞:

IE = 0.5 mA

IE =

IC = αIE = 0.49 mA

IC = αIE = 0.52 mA

VE = IE RE = 0.5 × 6.2 = 3.1 V

VE = −1.5 + 0.53 × 1.5 = −0.7 V

VB = 3.8 V

VB = 0

VC = VCC − RC IC = 9 − 6.2 × 0.49 6 V

VC = 1.5 − 2.4 × 0.52 = 0.25 V which is 0.95 V above VE . For β = 50:

1.5 V

6.97

IE =

RC

IC

1.5 − 0.7 = 0.53 mA 1.5

1.5 − 0.7 = 0.49 mA 1.5 + 7.5 51

IC = αIE = 0.48 mA

VC

VE = −1.5 + 0.49 × 1.5 = −0.77 V

IB

VB = VE + 0.7 = −0.07 V IB RB = 0.07 V

VE RB

IE RE = 0.49 × 1.5 = 0.7 V

IE

Thus,

RE

IB RB 0.1 IE RE

1.5 V

IE =

1.5 − 0.7 RE +

RB β+1

=

VC = 1.5 − 2.4 × 0.48 = 0.35 V which is 1.1 V above VE .

0.8 RE +

RB 51

(1) 6.98

For,

3 V 1.01 mA

IB RB = 0.1 IE RE RB = 0.1 ×

Substituting in Eq. (1) 0.5 IC = = 0.51 mA and RB = 5.1RE IE = α 50/51 gives 0.51 =

RC

IE × RE = 0.1 × 51 × RE = 5.1RE IB

0.01 mA 1.5 V 1 mA RB

0.8 RE + 5.1 R 51 E

⇒ RE = 1.43 k

Figure 1

RB = 5.1RE = 7.3 k VE = −1.5 + IE RE = −1.5 + 0.51 × 1.43 = −0.77 V VC = VE + 1 = +0.23 V RC =

1.5 − 0.23 = 2.54 k 0.5

Consulting the table of 5% resistors in Appendix J, we select RE = 1.5 k RC = 2.4 k RB = 7.5 k

Sedra, Smith, Carusone, Gaudet

(a) From the circuit diagram of Fig. 1, we can write 3 − 1.5 1.5 k RC = 1.01 mA 1.5 = 0.01RB + VBE = 0.01RB + 0.7 ⇒ RB = 80 k (b) Selecting 5% resistors, we have

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–34

RC = 1.5 k

= 1.1 V

RB = 82 k

IC + 0.01 =

IE =

VCC − VBE RB RC + β +1

3 − 1.1 3 − 1.1 = 1.29 = RC 1.47

IC = 1.28 mA

3 − 0.7 = 0.99 mA 82 1.5 + 101 IC = αIE = 0.99 × 0.99 = 0.98 mA =

6.99 Refer to the circuit in Fig. P6.100. Replacing VCC together with the voltage divider (R1 , R2 ) by its Thévenin equivalent results in the circuit shown below.

VC = 3 − 1.5 × 0.99 = 1.52 V (c) β = ∞: I C = IE =

IO aIE

VCC − VBE 3 − 0.7 = = 1.53 mA RC 1.5

IE /(b1)

VC = 0.7 V

RB

(d) From the circuit diagram of Fig. 2, we can write

IE

VBB

3 V

RE

RC

where

IC 2IB 2IB RB1

1.5 V IC 1 mA

RB = (R1 R2 )

IB

Now,

RB2

VBB = Figure 2 IE = IB =

1 IC = = 0.01 mA β 100

RB2 =

R2 R1 + R2

and

0.7 V IB2 IB

VBB = VCC

IE RB + VBE + IE RE β +1

VBB − VBE RE + (R1 R2 )/(β + 1)

IC = αIE =α

0.7 0.7 = IB2 0.01

VCC [R2 /(R1 + R2 )] − VBE RE + (R1 R2 )/(β + 1)

Q.E.D.

= 70 k 6.100

1.5 = 2IB RB1 + 0.7

5 V

0.8 = 2 × 0.01 × RB1 RB1 = 40 k RC =

3 − 1.5 1.5 = 1.47 k = IC + 2IB 1.02

IE

50.7 R

R 0.7 V

For β = ∞: IB = 0,

IB2 =

0.7 0.7 = = 0.01 mA RB2 70

IB1 = IB2 = 0.01 mA VC = 0.01RB1 + 0.7 = 0.01 × 40 + 0.7

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

IO

© Oxford University Press 2021

Chapter 6–35

IO = αIE 0.5 mA

VS = ID RS = 0.62ID

IE = 0.5 mA

VGS = VG − VS = 5.07 − 0.62ID

⇒R=

5 − 0.7 = 8.6 k 0.5

1 k n (VGS − Vt )2 2

ID =

1 k n (5.07 − 0.62ID − Vt )2 2

v Cmax = 0.7 − VECsat = 0.7 − 0.3

=

= +0.4 V

For k n = 0.3 mA/V2 and Vt = 1, 1 × 0.3(4.07 − 0.62ID )2 2

ID =

VDD 15 V

6.101

= 0.15(4.072 − 2 × 4.07 × 0.62ID + 0.622 ID2 )

10 M

RD

RG1

0.058ID2 − 1.757ID + 2.488 = 0 which results in ID = 28.8 mA, or 1.49 mA The first value does not make physical sense. Thus,

5.1 M

VG = VDD = 15 ×

RS

RG2

ID = 1.49 mA 1.5 mA which is the maximum value. The minimum value can be obtained by using k n = 0.2 mA/V2 and Vt = 1.5 V in Eq. (1),

RG2 RG1 + RG2

1 × 0.2(3.57 − 0.62ID )2 2

ID =

5.1 = 5.07 V 10 + 5.1

= 0.1(3.572 − 2 × 3.57 × 0.62ID + 0.622 ID2 )

k n = 0.2 to 0.3 mA/V2

0.038 ID2 − 1.442ID + 1.274 = 0

Vt = 1.0 V to 1.5 V

which results in

With RS = 0,

ID = 37 mA or 0.91 mA

1 k n (VG − Vt )2 2 is obtained with Vtmin and k nmax :

ID = IDmax

1 × 0.3(5.07 − 1)2 = 2.48 mA 2 is obtained with Vtmax and k nmin :

IDmax = IDmin

Here again, the physically meaningful answer is ID = 0.91 mA, which is the minimum value of ID . Thus with a 0.62-k resistance connected in the source lead, the value of ID is limited to the range of 0.91 mA to 1.5 mA. 6.102

VDD

1 IDmin = × 0.2(5.07 − 1.5)2 = 1.27 mA 2 With RS installed and Vt = 1 V, k n = 0.3 mA/V2 , we require ID = 1.5 mA:

RD ID

1 × 0.3(VGS − 1)2 2 ⇒ VGS = 4.16 V

1.5 =

VG 5 V VS 2 V ID

Since VG = 5.07 V,

RS 2 k

VS = VG − VGS = 5.07 − 4.16 = 0.91 V Thus, RS =

VS 0.91 = 607 = ID 1.5

ID =

From Appendix J, the closest 5% resistor is 620 . With RS = 620 ,

Sedra, Smith, Carusone, Gaudet

2V = 1 mA 2 k

But ID =

1 k n (VGS − Vt )2 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–36

1=

1 × 2(VG − VS − Vt )2 2

1 = (5 − 2 − Vt )

6.104

VDD

2

Vt = 2 V

RD ID

If Vt = 1.5 V, then we have 0V

VS = ID RS = 2ID VGS = VG − VS = 5 − 2ID ID =

VGS

RG

ID

RS

1 × 2(5 − 2ID − 1.5)2 2

4ID2 − 15ID + 12.25 = 0

VSS

ID = 1.2 mA

(a) VGS + ID RS = VSS

VS = 2.4 V

But ID =

6.103 ID = 0.5 mA =

1 × 4(VGS − 1)2 2

1 W kn (VGS − Vt )2 2 L

= K(VGS − Vt )2

⇒ VGS = Vt +

5 V

Thus,

ID

RG

Vt +

RD VD

VS

ID RS

ID K

ID + ID RS = VSS K

Differentiating relative to K, we have ID 1 ∂ID 1 ∂ID − 2 + RS =0 0+ √ K ∂K 2 ID /K K ∂K ∂ID K 1 = √ ∂K ID 1 + 2 KID RS ID ⇒ SK = 1/[1 + 2 KID RS ]

⇒

K = ±0.1, and Vt = 1 V. K ID We require ID = 100 µA and = ±0.01. ID Thus, (b) K = 100 µA/V2 ,

5 V ⇒ VGS = 1.5 V Since VG = 0 V, VS = −1.5 V, and −1.5 − (−5) = 7 k 0.5 Maximum gain is obtained by using the largest possible value of RD , that is, the lowest possible value of VD that is consistent with allowing negative voltage signal swing at the drain of 1 V. Thus

RS =

SKID =

ID /ID 0.01 = = 0.1 K/K 0.10

Substituting in the expression derived in (a), 0.1 =

1 √ 1 + 2 0.1 × 0.1RS

⇒ RS = 45 k

VD − 1 = v Dmin = VG − Vt = 0 − 1

To find VGS ,

⇒ VD = 0 V

ID = K(VGS − Vt )2

where we have assumed that the signal voltage at the gate is small. Now,

100 = 100(VGS − 1)2

VD = 0 = VDD − ID RD 0 = 5 − 0.5 × RD ⇒ RD = 10 k

Sedra, Smith, Carusone, Gaudet

Q.E.D.

VGS = 2 V VGS + ID RS = VSS 2 + 0.1 × 45 = 6.5 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–37

(c) For VSS = 5 V and VGS = 2 V, ID RS = 3 V 3 = 30 k 0.1 1 1 = = √ 7 1 + 2 0.1 × 0.1 × 30

RS = SKID

1 1 K ID = × ±10% = ±1.4% = × ID 7 K 7

∂ID =− ∂Vt SVIDt =

1 1 + RS √ 2k n ID

∂ID Vt Vt = − ∂Vt ID ID + ID RS 2k n

But

1 2 ⇒ VOV = ID = k n VOV 2

2ID kn

Thus SVIDt = −

6.105 (a) With a fixed VGS , 1 k n (VGS − Vt )2 2

ID =

∂ID Vt k n (VGS − Vt )Vt =− ∂Vt ID ID

k n (VGS − Vt )Vt 1 k n (VGS − Vt )2 2 2Vt 2Vt =− =− VGS − Vt VOV

Vt = ±5%, and Vt ID to ∓5% we VOV = 0.25 V, to limit ID require SVIDt = 1 Thus

=−

−1 = − Q.E.D.

Vt = ±5%, and For Vt = 0.5 V, Vt VOV = 0.25 V, we have Vt ID = SVIDt ID Vt

2 × 0.5 0.25 + 2ID RS

⇒ ID RS = 0.375 V For ID = 0.1 mA, 0.375 = 3.75 k 0.1 VG = VGS + ID RS = 0.5 + 0.25 + 0.375 = 1.125 V RS =

6.106

2 × 0.5 × ±5% 0.25 = ∓20%

VDD 5 V

=−

(b) For fixed bias at the gate VG and a resistance RS in the source lead, we have

RG

VG = VGS + ID RS where VGS is obtained from

10 M

1 k n (VGS − Vt )2 2 2ID ⇒ VGS = Vt + kn ID =

Thus Vt +

Q.E.D.

For Vt = 0.5 V,

∂ID = −k n (VGS − Vt ) ∂Vt SVIDt ≡

2Vt VOV + 2ID RS

0

RD 5 k ID ID

VDS

VGS

VGS = VDD − ID RD

= 5 − 5ID

2ID + ID RS = VG kn

(a) Vt = 0.7 V and k n = 5 mA/V2

Differentiating relative to Vt , we have 2 ∂ID ∂ID 1 + RS =0 1+ √ ∂Vt 2 2ID /k n k n ∂Vt 1 ∂ID + RS = −1 √ ∂Vt 2k n ID

Sedra, Smith, Carusone, Gaudet

1 k n (VGS − Vt )2 2 1 ID = × 5(5 − 5 ID − 0.7)2 2 ID =

⇒ ID2 − 1.736ID + 0.7396 = 0

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–38

ID = 0.99 mA or 0.75 mA

ICmax = 1 × e(0.710−0.699)/0.025

The first root results in VGS = −0.05 V, which is physically meaningless. Thus

= 1.55 mA

ID = 0.75 mA VG = VD = 5 − 5 × 0.75 = 1.25 V (b) Vt = 1.5 V and k n = 10 mA/V2

and ICmin = 1 × e(0.710−0.721)/0.025 ICmin = 0.64 mA VCE will range from

1 ID = × 10(5 − 5ID − 1.5)2 2

VCEmin = 3 − 1.55 × 2 = −0.1 V

⇒ ID2 − 1.408ID + 0.49 = 0

which is impossible, implying that the transistor will saturate at this value of dc bias!

ID = 0.78 mA or 0.63 mA The first root can be shown to be physically meaningless, thus ID = 0.63 mA VG = VD = 5 − 5 × 0.63 = 1.85 V

VCEmax = 3 − 0.64 × 2 = 1.72 V It should be clear that this biasing arrangement is useless, since even the small and inevitable tolerances in RB1 and RB2 caused such huge variations in IC that in one extreme the transistor left the active mode of operation altogether!

6.107 6.109

5V

VCC 3 V RD 0

RG

ID

RC 2 k

VD VG VGS

VCE

RB

IB ID = 0.2 =

1 × 10(VGS − Vt )2 2

⇒ VGS = 1.2 V

To obtain IC = 1 mA, we write

5 − 1.2 = 19 k RD = 0.2

IB =

1 mA IC = = 0.01 mA β 100

Thus, RB =

6.108 Refer to Fig. 6.53(a). RB2 × 3 = 0.710 RB1 + RB2 RB1 ⇒ = 3.225 RB2 Given that RB1 and RB2 are 1% resistors, the maximum and minimum values of the ratio RB1 /RB2 will be 3.225 × 1.02 = 3.2899 and 3.225 × 0.98 = 3.1608. The resulting VBE will be 0.699 V and 0.721 V, respectively. Correspondingly, IC will be

Sedra, Smith, Carusone, Gaudet

VCC − VBE 3 − 0.7 = 230 k IB 0.01

Since β ranges from 50 to 150 and IB is fixed at 0.01 mA, the collector current IC will range from 0.01 × 50 = 0.5 mA to 0.01 × 150 = 1.5 mA. Correspondingly, VCE will range from (3 − 0.5 × 2) = 1 V to (3 − 1.5 × 2) = 0 V. The latter value implies that the high-β transistor will leave the active region of operation and saturate. Obviously, this bias method is very intolerant of the inevitable variations in β. Thus it is not a good method for biasing the BJT.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–39

6.110

6.111 Refer to Fig. 6.54.

VCC 9 V 0.06 mA

0.6 mA RC

R1

3V

VCE 3 V RE

R2

3V

Initial design: β = ∞ RC = RE =

3V = 5 k 0.6

3.7 = 61.7 k 0.06

R1 = 150 − 61.7 = 88.3 k Using 5% resistors from Appendix J, and selecting R1 and R2 so as to obtain a VBB that is slightly higher than 3.7 V, we write R1 = 82 k and R2 = 62 k RE = 5.1 k and RC = 5.1 k

IE =

VBB − VBE IE nominal = RB RE + 101 VBB − VBE IE high = RB RE + 151 VBE − VBE IE low = RB RE + 51 Let’s constrain IE low to be equal to IE nominal × 0.95 and then check IE : VBB − VBE VBB − VBE = RB RB RE + RE + 101 51 RB /RE 1+ 101 0.95 = RB /RE 1+ 51 RB ⇒ = 5.73 RE

0.95

VB = VE + VBE = 3 + 0.7 = 3.7 V

VBB = VCC

VBB − VBE RB RE + β +1

high

9 = 150 k R1 + R2 = 0.06

R2 =

(a) IE =

R2 62 = 3.875 =9× R1 + R2 62 + 82

VBB − VBE RB RE + β +1

For this value,

VBB − VBE IE nominal = 0.946 RE VBB − VBE = 0.95 IE nominal IE low = 0.90 RE VBE − VBE = 1.02 IE nominal IE high = 0.963 RE Thus, the maximum allowable ratio is

where

RB = 5.73 RE

RB = R1 R2 = 62 82 = 35.3 k

(b) IE =

3.875 − 0.7 = 0.58 mA 35.3 5.1 + 91

IE RE =

IE =

VE = 0.58 × 5.1 = 3.18 VB = 3.88 V IC = αIE =

90 × 0.58 = 0.57 mA 91

VC = 6.1 V IR2 = IB =

VB 3.88 = 0.063 mA = R2 62 0.58 IE = = 0.006 mA and β +1 91

IR1 = 0.069 mA

Sedra, Smith, Carusone, Gaudet

VBB − VBE RB /RE RE 1 + β +1 VBB − VBE 5.73 1+ β +1

VBB − VBE VCC = 5.73 3 1+ 101 ⇒ VBB = VBE + 0.352VCC (c) VCC = 5 V VBB = 0.7 + 0.352 × 5 = 2.46 V RE =

VCC /3 5/3 = 3.33 k = IE 0.5

RB = 5.73 × RE = 19.08 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–40

VBB = VCC 2.46 = 5

R2 R1 + R2

To maximize gain while allowing for ±1 V signal swing at the collector, design for the lowest possible VC consistent with

R2 R1 + R2

VC − 1 = −0.7 + VCEsat

R1 R2 2.46R1 = 5 = 5RB R1 + R2

= −0.7 + 0.3 = −0.4 V

= 5 × 19.08

VC = 0.6 V

⇒ R1 = 38.8 k 1 1 = 37.5 k − R2 = 1 RB R1

RC =

(d) VCE = VCC − RC IC 1 = 5 − RC × 0.99 × 0.5 ⇒ RC = 8.1 k Check design: VBB = VCC

R2 37.5 =5× R1 + R2 37.5 + 38.8

= 2.46 V

VCC − VC 3 − 0.6 = 6.2 k = IC 0.39

As temperature increases from 25◦ C to 125◦ C, (i.e., by 100◦ C), VBE decreases by 2 mV × 100 = 0.2 V − 200 mV. Thus IE increases by = RE 0.2 V = 0.035 mA to become 0.435 mA. The 5.75 k collector current becomes IC =

β × 0.435 β +1

where β is the increased value of 150,

RB = R1 R2 = 37.5 38.8 = 19.07 k 2.46 − 0.7 = 0.5 mA IE nominal = 19.07 3.33 + 101 2.46 − 0.7 = 0.475 mA IE low = 19.07 3.33 + 51 which is 5% lower than IE , and nominal

IE

2.46 − 0.7 = 0.509 mA = high 19.07 3.33 + 151 which is 1.8% higher than IE .

150 × 0.435 mA 151 = 0.432 mA IC =

Thus, IC = 0.432 − 0.39 = 0.042 mA for a percentage increase of 0.042 IC × 100 = 10.8% × 100 = IC 0.39

6.113

3 V

nominal

IE

6.112

3 V

IE/(b 1)

RC VC

RC

IC VC

RB

IE

0.7 V RE

0.4 mA

3 V RE =

−0.7 − (−3) 0.4

= 5.75 k

Sedra, Smith, Carusone, Gaudet

VC = VCEsat + 1 V = 1.3 V 3 − 1.3 = 0.5 mA IE = RC ⇒ RC = 3.4 k IB =

0.5 IE = 0.005 mA β +1 101

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–41

VC = VBE + IB RB

6.115

1.3 = 0.7 + 0.005 × RB

VCC

⇒ RB = 120 k

R1

I IO

Standard 5% resistors: RC = 3.3 k

0

RB = 120 k

Q1

If the actual BJT has β = 50, then IE =

VCC − VBE 3 − 0.7 = 0.41 mA = RB 120 RC + 3.3 + β +1 51

Q3

VB

I

VE3 Q2

VC = 3 − IE RC = 3 − 0.41 × 3.3 = 1.65 V

IO

RE

R2

Allowable negative signal swing at the collector is as follows: VC − VCEsat = 1.65 − 0.3 = 1.35 V An equal positive swing is just possible. For β = 150: 3 − 0.7 = 0.56 mA IE = 120 3.3 + 151 VC = 3 − IE RC = 3 − 0.56 × 3.3 = 1.15 V Allowable negative signal swing at the collector = 1.15 − 0.3 = 0.85 V. An equal positive swing is possible.

I=

VCC − VBE1 − VBE2 R1 + R2

VB = IR2 + VBE2 + VBE1 VE3 = VB − VBE3 VE3 = IR2 + VBE2 + VBE1 − VBE3 = (VCC − VBE1 − VBE2 ) +VBE2 − VBE3 IO =

6.114

VC RB

IC

VE3 1 R2 (VCC − VBE1 − VBE2 ) = RE RE R1 + R2 + VBE1 + VBE2 − VBE3

Now, for R1 = R2 and the currents in all junctions equal,

I

IB

R2 + VBE1 R1 + R2

VBE1 = VBE2 = VBE3 = VBE 1 1 (VCC − 2VBE ) × + VBE IO = RE 2 IO =

VCC 2RE

Q.E.D.

Thus, IC = 1 mA

IO R E =

I = IC + IB

1 IC =1 1+ = IC + β β I = 1.01 mA

VB =

VCC 2

VCC + VBE 2

I = (VB − 2VBE )/R2 =

VCC − VBE 2

VC = 1.5 V = IB RB + VBE

But since I must be equal to IO , we have

1.5 = 0.01 × RB + 0.7

VCC /2 − VBE VCC = 2RE R2

RB = 80 k

Thus,

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

R2

© Oxford University Press 2021

Chapter 6–42 R1 = R2 = RE

VCC − 2VBE VCC

indeed operating in saturation. (Equivalent: VD = 2.5 V is higher than VG − Vt = 1.25 V.)

For VCC = 10 V and VBE = 0.7 V, 10 − 1.4 R1 = R2 = RE = 0.86RE 10

1 2 k n VOV 2 1 0.25 = k n × 0.252 2

ID =

To obtain IO = 0.5 mA,

⇒ k n = 8 mA/V2

10 VCC = 0.5 = 2RE 2RE ⇒ RE = 10 k

(b) The amplifier small-signal equivalent-circuit model is shown in Fig. 2.

R1 = R2 = 8.6 k

Rin = RG1 RG2 = 750 k 500 k = 300 k

6.116 (a) Refer to Fig. 6.57(a). The dc circuit can be obtained by opening all coupling and bypass capacitors, resulting in the circuit shown in Fig. 1.

gm =

2ID 2 × 0.25 = = 2 mA/V VOV 0.25

ro =

VA 50 = 200 k = ID 0.25

Gv = − =−

Rin gm (ro RD RL ) Rin + Rsig

300 × 2 × (200 10 10) 300 + 100

= −7.3 V/V (c) VG = 2 V,

VD = 2.5 V

vˆG = 2 + vˆgs ,

vˆD = 2.5 − | Av |vˆgs

where | Av | = gm (ro RD RL ) = 9.8 V/V To remain in saturation, vˆD ≥ vˆG − Vt 2.5 − 9.8vˆgs ≥ 2 + vˆgs − 0.75

Figure 1

This is satisfied with equality at

See analysis on figure.

2.5 − 1.25 = 0.116 V 10.8

VGS = 2 − 1 = 1 V

vˆgs =

VOV = VGS − Vt = 1 − 0.75 = 0.25 V

The corresponding value of vˆsig is 300 + 100 = 0.154 V vˆsig = vˆgs 300

Since VD at 2.5 V is 1.2 V higher than VS + VOV = 1 + 0.25 = 1.25 V, the transistor is

The corresponding amplitude at the output will be

This figure belongs to Problem 6.116.

Rsig 100 k vsig

RG1

RG2

vgs gmvgs

ro

RD 10 k

RL 10 k

vo

Rin RG RG1 // RG2 750 // 500 300 k Figure 2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–43

| Gv |vˆsig = 7.3 × 0.154 = 1.13 V (d) To be able to double vˆsig without leaving saturation, we must reduce vˆgs to half of what would be its new value; that is, we must keep vˆgs unchanged. This in turn can be achieved by connecting an unbypassed Rs equal to 1/gm , Rs =

1 = 500 2 mA/V

Since vˆgs does not change, the output voltage also will not change, thus vˆo = 1.13 V.

vz = vx

(RS ro ) (RS ro ) +

1 gm

(9.5 100)

= 0.946 V/V 1 2 Looking into terminal Z, we see Ro : =

(9.5 100) +

1 gm

Ro = RS r o

1 = 473 2 (e) If X is grounded, the circuit becomes a CG amplifier. = 9.5 100

6.117 (a) DC bias: Refer to the circuit in Fig. P6.117 with all capacitors eliminated: RD

Rin at gate = RG = 10 M VG = 0, thus VS = −VGS , where VGS can be obtained from 1 2 ID = k n VOV 2 1 2 0.4 = × 5 × VOV 2 ⇒ VOV = 0.4 V

vy

vsg

Rsig 100 k

VGS = Vt + 0.4 = 0.8 + 0.4 = 1.2 V VS = −1.2 V −1.2 − (−5) RS = = 9.5 k 0.4 To remain in saturation, the minimum drain voltage must be limited to VG − Vt = 0 − 0.8 = −0.8 V. Now, to allow for 0.8-V negative signal swing, we must have VD = 0 V

The figure shows the circuit prepared for signal calculations. 1 v sg = isig × Rsig RS

gm 1 (k) = 50 × 10−3 (mA) 100 9.5

2 = 0.024 V v y = (gm RD )v sg

and 5−0 = 12.5 k 0.4 2ID 2 × 0.4 = 2 mA/V (b) gm = = VOV 0.4 RD =

VA 40 ro = = 100 k = ID 0.4 (c) If terminal Z is connected to ground, the circuit becomes a CS amplifier, Gv = −

isig 50 A

RS

vy RG = × −gm (ro RD RL ) v sig RG + Rsig

10 × 2 × (100 12.5 10) =− 10 + 1

= (2 × 12.5) × 0.024 = 0.6 V 6.118 (a) Refer to the circuit of Fig. P6.118(a): Av o ≡

Ro =

v o1 = vi

10 1 10 + gm

=

10 10 +

1 10

= 0.99 V/V

1

10 k = 0.1 10 = 99 gm

(b) Refer to Fig. P6.118(b): Rin = 10 k

1 = 10 0.1 = 99 gm

= −9.6 V/V

5 2 vo = = 10(5 2) = 14.3 V/V v i2 1/gm

(d) If terminal Y is grounded, the circuit becomes a CD or source-follower amplifier:

(c) v i2 = (Av o v i )

Sedra, Smith, Carusone, Gaudet

Rin Rin + Ro

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–44

= 0.99 × v i ×

vo 40 =− = −104 V/V vπ 0.385

99 99 + 99

0.5v i

But vo = −gm (RC RL ) vπ

v o = 14.3 × v i2 = 14.3 × 0.5v i vo = 7.15 V/V vi

−104 = −73.4 (RC 2) (RC 2) = 1.416

6.119 Refer to the circuit of Fig. 6.58. DC design: VB = 5 V,

VBE = 0.7 V

VE = 4.3 V For IE = 2 mA,

RE =

IR2 = 0.2 mA, IB =

VE 4.3 = = 2.15 k IE 2

RB2 =

5 = 25 k 0.2

2 IE = 0.02 mA β +1 101

IR1 = IR2 + IB = 0.2 + 0.02 = 0.22 mA RB1

VCC − VB 15 − 5 = 45.5 k = = IR1 0.22

Choosing 5% resistors: RE = 2.2 k,

RB1 = 47 k,

RB2 = 24 k

RC = 4.86 k We can select either 4.7 k or 5.1 k. With 4.7 k, the gain will be vo = −0.385 × 73.4 × (4.7 2) = −39.6 V/V v sig which is slightly lower than the required −40 V/V, and we will obtain VC = 15 − 4.7 × 1.84 = 6.4 V allowing for about 2 V of negative signal swing at the collector. If we choose 5.1 k, the gain will be vo = −0.385 × 73.4 × (5.1 2) = −40.6 V/V v sig which is slightly higher than the required gain, and we will obtain VC = 15 − 5.1 × 1.84 = 5.6 V which allows for only 1.2-V negative signal swing.

For these values, IE =

VBB − VBE RB RE + β +1

6.120 Refer to the circuit of Fig. P6.120. (a) IE = 0.2 mA. Writing a loop equation for the base–emitter circuit results in

where RB2 24 = 15 × VBB = VCC RB1 + RB2 24 + 47 = 5.07 V RB = RB1 RB2 = 47 24 = 15.89 k 5.07 − 0.7 = 1.85 mA 15.89 2.2 + 101 VB = IE RE + VBE = 1.85 × 2.2 + 0.7 = 4.8 V

IE =

IC = αIE = 0.99 × 1.85 = 1.84 mA gm =

IC 1.84 = = 73.4 mA/V VT 0.025

β 100 = = 1.36 k rπ = gm 73.4

IB Rsig + VBE + IE RE = 3 IE Rsig + VBE + IE RE = 3 β +1 0.2 × 5 + 0.7 + 0.2RE = 3 101 ⇒ RE = 11.5 k (b) IC = αIE 0.2 mA VC = 0.5 = 3 − 0.2RC ⇒ RC = 12.5 k (c) gm =

IC 0.2 mA = 8 mA/V = VT 0.025 V

rπ =

β 100 = = 12.5 k gm 8

Rin 1.25 vπ = = = 0.385 V/V v sig Rin + Rsig 1.25 + 2

Gv =

vo 12.5 × 8 × (12.5 10) =− v sig 12.5 + 5

For an overall gain of −40 V/V,

= −31.7 V/V

Rin = RB1 RB2 rπ = 47 24 1.36 = 1.25 k

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–45

This figure belongs to Problem 6.120.

Rsig vo

vsig

rp

vp gmvp

RC

RL

6.121 Refer to the circuit in Fig. P6.121:

vo

IE = 0.2 mA

100 k

re =

VT 25 mV = = 125 IE 0.2 mA

gm =

IC 0.2 mA = 8 mA/V VT 0.025 V

ie

ie

5 k

re 50 Rsig 50

Note that the emitter has a resistance Re = 250 . Rin = 100 k (β + 1)(re + Re )

v sig

= 100 [101 × (0.125 + 0.25)] = 100 37.9 = 27.5 k Rin 27.5 vb = 0.733 V/V = = v sig Rin + Rsig 27.5 + 10 vo Total resistance in collector = −α vb Total resistance in emitter 10 10 = −13.33 V/V 0.125 + 0.25 vo Gv = = −0.733 × 13.33 = −9.8 V/V v sig −

For v be to be limited to 5 mV, the signal between r e + RE = base and ground will be vˆb = vˆbe re 0.125 + 0.25 = 15 mV. The limit on v sig 5 0.125 can be obtained by dividing the 15 mV by v b /v sig , 15 mV = 20.5 mV 0.733 Correspondingly, at the output we have

vˆsig =

vˆo = | Gv |vˆsig = 9.8 × 20.5 = 201 mV 0.2 V

Rin re 50 −v sig −v sig = 100 0.1 k At the output node, =

v o = −αie (5 100) v sig (5 100) =α 0.1 vo 5 100 =α 47.6 V/V v sig 0.1 6.123 (a) IE =

3 − 0.7 100 1+ β +1

β = 50: 2.3 = 0.78 mA 100 1+ 51 VE = IE RE = 0.78 V IE =

VB = VE + 0.7 = 1.48 V 6.122 Refer to the circuit in Fig. P6.122. The dc emitter current is equal to 0.5 mA, and IC = αIE 0.5 mA; also, re =

VT 25 mV = 50 = IE 0.5 mA

Rin = re = 50 ie =

−v sig −v sig = re + Rsig 50 + 50

Sedra, Smith, Carusone, Gaudet

β = 200: 2.3 = 1.54 mA 100 1+ 201 VE = IE RE = 1.54 V

IE =

VB = VE + 0.7 = 2.24 V (b) Rin = 100 (β + 1)[re + (1 1)]

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–46

= 100 (β + 1)(re + 0.5)

ID =

β = 50:

1.5 = 3 k 0.5 (b) Gv = −gm RD ⇒ RS =

VT 25 mV = 32.1 = re = IE 0.78 mA Rin = 100 [51 × (0.0321 + 0.5)] = 21.3 k

where gm =

β = 200: re =

2.5 − 1 = 0.5 mA RS

2ID 2 × 0.5 = 4 mA/V = VOV 0.25

Thus,

VT 25 mV = 16.2 = IE 1.54 mA

Rin = 100 [201 × (0.0162 + 0.5)] = 50.9 k

−12 = −4RD ⇒ RD = 3 k (c) v G = 0 V (dc) + v sig v Gmin = −vˆsig

Rin vb = (c) v sig Rin + Rsig

vˆD = VD + | Gv |vˆsig

vo (1 1) 500 = = (re in ) vb (1 1) + re 500 + re

where

β = 50:

To remain in saturation,

vb 21.3 = 0.68 V/V = v sig 21.3 + 10

vˆD ≤ vˇG + |Vtp |

VD = −2.5 + ID RD = −2.5 + 0.5 × 3 = −1 V

−1 + 12 vˆsig ≤ −vˆsig + 0.75

vo 500 = 0.94 V/V = vb 500 + 32.1 vo = 0.68 × 0.94 = 0.64 V/V v sig

and the corresponding output voltage

β = 200:

vˆd = | Gv |vˆsig = 1.62 V

vb 50.9 = 0.836 V/V = v sig 50.9 + 10

(d) If vˆsig = 50 mV, then

vo 500 = 0.969 V/V = vb 500 + 16.2 vo = 0.836 × 0.969 = 0.81 V/V v sig

Satisfying this constraint with equality gives vˆsig = 0.135 V

VD + | Gv |vˆsig = −vˆsig + |Vtp | where VD = −2.5 + ID RD = −2.5 + 0.5RD and | Gv | = gm RD = 4RD

6.124 Refer to the equivalent circuit in Fig. 6.57(c). Gv = − =−

Rin gm (RD RL ro ) Rin + Rsig

Thus −2.5 + 0.5RD + 4RD vˆsig = −vˆsig + |Vtp | −2.5 + 0.5RD + 4RD × 0.05 = −0.05 + 0.75 0.4RD = 3.15

RG gm (RD RL ro ) RG + Rsig

⇒ RD = 4.6 k

10 × 5 × (10 50 20) =− 10 + 1

Gv = −gm RD = −4 × 4.6 = −18.4 V/V

= −26.7 V/V 6.126 Refer to Fig. P7.126. 6.125 Refer to Fig. P6.118.

Ri2 =

(a) DC bias: |VOV | = 0.25 V ⇒ VSG = |Vtp | + |VOV | = 1 V Since VG = 0 V, VS = VSG = +1 V, and

Sedra, Smith, Carusone, Gaudet

1 = 50 gm2

1 A/V = 20 mA/V 50 If Q1 is biased at the same point as Q2 , then ⇒ gm2 =

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–47

gm1 = gm2 = 20 mA/V

= 0.1066 mA

id1 = gm1 × 5 (mV)

The current in the voltage divider is

= 20 × 0.005 = 0.1 mA

I=

v d1 = id1 × 50

Thus the current through RD will be (0.1066 + 0.0016) 0.108 mA and

= 0.1 × 50 = 5 mV

VDD − VD 10 − 4 = = 55.6 k 0.108 0.108 2ID 2 × 0.107 = 1.07 mA/V = (b) gm = VOV 0.2

v o = id1 RD = 1 V

RD =

1V = 10 k RD = 0.1 mA

ro =

6.127 (a) DC bias:

VA 60 = = 561 k ID 0.107

(c) Replacing the MOSFET with its hybrid-π model, we obtain the small-signal equivalent circuit of the amplifier, shown in Fig. 2.

VDD 10 V

Node equation at the output:

RD

vo v o − v gs vo + + + gm v gs = 0 RD ro R2 1 1 1 1 = −gm 1 − v gs + + vo RD ro R2 gm R2

R2 2 M VD ID VGS

VD 4 = = 1.6 µA = 0.0016 mA R1 + R2 2.5

Thus,

R1 0.5 M

A

1 v gs v o = − gm (RD ro R2 ) 1 − gm R2

(1)

Next, we express v gs in terms of v sig and v o using superposition:

Figure 1 VGS = Vt + VOV

R2 R1 + vo R1 + R2 R1 + R2

= 0.6 + 0.2 = 0.8 V

v gs = v sig

From the voltage divider (R1 , R2 : see Fig. 1), we can write R1 0.5 = VD VGS = VD R1 + R2 0.5 + 2

Substituting for v gs from Eq. (2) into Eq. (1) yields v o = −Av sig

VD = 5VGS = 5 × 0.8 = 4 V VDS 1 2 1+ ID = k n VOV 2 VA 4 1 ID = × 5 × 0.22 1 + 2 60

(2)

R2 R1 − Av o R1 + R2 R1 + R2

where

A = gm (RD ro R2 ) 1 −

1 gm R2

Thus,

This figure belongs to Problem 6.127(c).

(vo vgs)⁄R2 R2 R1

vsig

vgs

vo

gmvgs RD

ro

Figure 2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–48 vo 1 + A vo = v sig

R1 R1 + R2

= −A

R2 v sig R1 + R2

VGS =

R2 R1 + R2 R1 1+A R1 + R2

Thus

−A

VD = 2VGS = 1.6 V

I RD

Substituting numerical values yields vo = v sig 2/0.5 − 1 + (2/0.5) 1+ 1.07(55.6 561 2000)(1 − 1/1.07 × 2000) 1+

5 52.8

VD 1.6 V = = 1.6 µA 1 M 1 M = 0.1 + 0.0016 0.102 mA

Idivider =

Q.E.D.

4

1 × 5 × 0.22 = 0.1 mA 2

=

R2 /R1 1 + R2 /R1 1+ gm (RD ro R2 )(1 − 1/gm R2 )

= −

1 2 k n VOV 2

ID =

−R2 /R1 = 1 + R2 /R1 1+ A Thus, vo =− v sig

R1 0.5 VD = VD = 0.5VD R1 + R2 0.5 + 0.5

RD =

VDD − VD 10 − 1.6 = 82.4 k = IRD 0.102

(b) gm =

2ID 2 × 0.1 = 1 mA/V = VOV 0.2

(c) Replacing the MOSFET with its T model results in the amplifier equivalent circuit shown in Fig. 2. At the output node, v o = i[RD (R1 + R2 )] v o = iRD

(1)

vo

= −3.65 V/V Note that the gain is nearly equal to −R2 /R1 = − 4, which is the gain of an op amp connected in the inverting configuration.

R2 0 bvo i

6.128 (a) DC bias:

1/gm

R1

VDD 10 V

vsig Rin

RD R2 0.5 M

Figure 2 VD ID

R1 0.5 M

RD

i

where RD = RD (R1 + R2 ). The voltage at the gate is a fraction β of v o with R1 R1 + R2

β=

VGS

Now, the current i can be found from

i=

Figure 1

(2)

Substituting for i from Eq. (2) into Eq. (1) yields

VOV = 0.2 V

v o = (gm v sig − βgm v o )RD

VGS = Vt + VOV

Thus

= 0.6 + 0.2 = 0.8 V From the voltage divider (R1 , R2 : see Fig. 1), we can write

Sedra, Smith, Carusone, Gaudet

v sig − βv o = gmv sig − βgmv o 1/gm

vo gm RD = v sig 1 + βgm RD

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–49

=

=

1/β 1/β 1+ gm RD

where VBB = VCC

1 + (R2 /R1 ) 1 + R2 /R1 1+ gm RD

RB = R1 R2 = 15 27 = 9.643 k

(3)

Q.E.D.

0.99(5.357 − 0.7) = 1.85 mA 9.643 2.4 + 101 IC 1.85 mA gm = = = 74 mA/V VT 0.025 V IC =

The input resistance Rin can be obtained as follows: v sig Rin = i Substituting for i from Eq. (1) yields v sig Rin = R vo D v sig and replacing by the inverse of the gain vo expression in Eq. (3) gives 1 1 Rin = RD + gm RD 1 + (R2 /R1 ) R1 1 1 + gm RD Q.E.D. Rin = gm R1 + R2

rπ =

vo = v sig

1 + (0.5/0.5) 1 + (0.5/0.5) 1+ 1 × (82.4 1000)

=

2 2 76.13

Rin = R1 R2 rπ = RB rπ = 9.643 1.35 = 1.18 k Rin 1.18 vπ = 0.371 V/V = = v sig Rin + Rsig 1.18 + 2 vo = −gm (RC RL ) vπ = −74(3.9 2) = −97.83 vo = −0.371 × 97.83 = −36.3 V/V v sig

= 1.95 V/V 6.130 Refer to the circuit of Fig. P6.129:

R2 Note that the gain 1 + = 2, similar to that R1 of an op amp connected in the noninverting configuration! 0.5 1 1 + 1 × (82.4 1000) Rin = 1 0.5 + 0.5

IC =

α(VBB − VBE ) RB RE + β +1

where VBB = VCC

= 39.1 k

R2 47 = 5.465 V = 15 × R2 + R1 47 + 82

RB = R1 R2 = 47 82 = 29.88 k 0.99(5.465 − 0.7) = 0.63 mA 29.88 7.2 + 101 IC 0.63 = 25.2 mA/V gm = = VT 0.025 IC =

6.129 Refer to the circuit of Fig. P6.129. IC =

β 100 = 1.35 k = gm 74

Replacing the BJT with its hybrid-π model results in the equivalent circuit shown at the bottom of the page:

(d) Substituting numerical values:

1+

R2 15 = 5.357 V = 15 × R2 + R1 15 + 27

α(VBB − VBE ) RB RE + β +1

This figure belongs to Problem 6.129.

Rsig vo

vsig

RB1

RB2

vp

rp

gmvp

RC

RL

Rin

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–50

rπ =

β 100 = = 4 k gm 25.2

To maximize the open-circuit voltage gain between base and collector while ensuring that the instantaneous collector voltage does not fall below (v B − 0.4) when v be is as high as 5 mV, we impose the constraint

Rin = R1 R2 rπ = RB rπ = 29.88 4 = 3.5 k 3.5 vπ = 0.636 V/V = v sig 3.5 + 2

VC − | Av o | × 0.005 = VB + 0.005 − 0.4

vo = −gm (RC RL ) vπ

where

= −25.2(12 2) = −43.2 V/V vo = −0.636 × 43.2 = −27.5 V/V v sig

= 5 − 0.99 × 0.22RC

VC = VCC − IC RC 5 − 0.22RC

Comparing the results above to those of Problem 6.124, we see that raising the resistance values has indeed resulted in increasing the transmission from source to transistor base, from 0.371 V/V to 0.636 V/V. However, because IC has decreased and gm has correspondingly decreased, the gain from base to collector has decreased by a larger factor (from 97.83 V/V to 43.2 V/V), with the result that the overall gain has in fact decreased (from 36.3 V/V to 27.5 V/V). Thus, this is not a successful strategy!

| Av o | = gm RC =

0.99 × 0.22 RC = 8.7RC 0.025

and VB = −

0.22 × 90 = −0.2 V 101

Thus, 5 − 0.22RC − 8.7RC × 0.005 = −0.2 − 0.395 ⇒ RC = 21.2 k Selecting 5% resistors, we find RB = 91 k RC = 22 k

6.131 Refer to the circuit of Fig. P6.131.

and specifying I to one significant digit gives

DC voltage drop across RB = 0.2 V,

I = 0.2 mA αIC 0.2 gm = = 8 mA/V VT 0.025

IB RB = 0.2 V I RB = 0.2 V β +1 IRB = 0.2 × 101

(1)

rπ =

Rin = RB rπ = 10 k RB

VT = 10 IB

Gv = −

6.132 Refer to the circuit of Fig. P6.132. (a) DC analysis of each of the two stages: VBB = VCC (2)

Substituting for IRB from Eq. (1) yields

I=

0.2 × 101 = 0.22 mA 90

Sedra, Smith, Carusone, Gaudet

11 × 8(22 20) 20 + 11

= −29.7 V/V

0.025 × 101 RB × I = 10 0.025 × 101 RB + I 0.025 × 101RB = 10 IRB + 0.025 × 101

0.025RB = 10 0.225 ⇒ RB = 90 k

β 100 = 12.5 k = gm 8

Rin = RB rπ = 91 12.5 = 11 k

0.025 = 10 RB

I/(β + 1) 0.025 × 101 RB

= 10 I

0.025 × 101RB = 10 0.2 × 101 + 0.025 × 101

Av o = −gm RC = −8 × 22 = −176 V/V

= 4.8 V

R2 47 = 15 R1 + R2 100 + 47

RB = R1 R2 = 100 47 = 32 k IE =

VBB − VBE RB RE + β +1

4.8 − 0.7 = 0.97 mA 1 mA 32 3.9 + 101 IC = αIE 1 mA =

VC = VCC − IC RC = 15 − 1 × 6.8 = 8.2 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–51

This figure belongs to Problem 6.132. B1

B2

From Fig. 1 we see that

(b) See figure above. gm =

IC = 40 mA/V VT

rπ =

β = 2.5 k gm

IC = 0.495 mA VC = IB × 200 k + IE × 0.2 k + VBE = 0.005 × 200 + 0.5 × 0.2 + 0.7

Rsig = 5 k, R1 = 100 k, R2 = 47 k, RC = 6.8 k, RL = 2 k

= 1.18 V (b)

(c) Rin1 = R1 R2 rπ = RB rπ = 32 2.5

vo

vo vi 200

= 2.32 k Rin 2.32 v b1 = 0.32 V/V = = v sig Rin + Rsig 2.32 + 5

i

200 k

(d) Rin2 = R1 R2 rπ = Rin1 = 2.32 k v b2 v b2 = = −gm (RC Rin2 ) v b1 vπ1 = −40(6.8 2.32) = −69.2 V/V vo vo = = −gm (RC RL ) (e) v b2 vπ2

vi

= −40(6.8 2) = −61.8 V/V vo v b2 v b1 vo = × × = −61.8 (f) v sig v b2 v b1 v sig

re 50

i

20 k

200

Figure 2

× − 69.2 × 0.32 = 1368.5 V/V From Fig. 2, we have gm =

6.133 (a)

IC 0.495 20 mA/V = VT 0.025

VT = 50 IE vi vi i= = r e + Re 50 + 200 vi vi = = = 4 v i , mA 250 0.25 k Node equation at the output:

re =

0.5 mA VC 0.495 mA

200 k

0.005 mA

0.5 mA 200

Figure 1

Sedra, Smith, Carusone, Gaudet

vo − vi vo + αi + =0 20 200 vo vi vo + 0.99 × 4v i + − =0 20 200 200 1 1 1 vo + = −v i 4 × 0.99 − 20 200 200 vo = −71.9 V/V vi

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–52

9 V

6.134 Refer to the circuit in Fig. P6.134. IE =

=

3 − 0.7 100 3.3 + β +1

4.5 V 20 k

2.3 = 0.54 mA 100 3.3 + 101

IE 2 k

VT 25 mV re = = = 46.3 IE 0.54 mA Rin = (β + 1)[re + (3.3 2)] Figure 1

= 101 × (0.0463 + 1.245) = 130.4 k

IC = αIE = 0.99 × 1.73 mA

Rin 130.4 vb = = v sig Rin + Rsig 130.4 + 100

= 1.71 mA IC = 68.4 mA/V gm = VT

= 0.566 V/V vo 3.3 2 1.245 = = vb (3.3 2) + re 1.245 + 0.0463 = 0.964 V/V vo = 0.566 × 0.964 = 0.55 V/V v sig vo io = 2 k vi vb ii = = Rin 130.4 k

re =

VT 25 mV = 14.5 = IE 1.73 mA

= 0.0145 k rπ = (β + 1)re = 101 × 0.0145 = 1.46 k (b) Replacing the BJT with its T model (without ro ) and replacing the capacitors with short circuits results in the equivalent-circuit model shown in Fig. 2.

io vo 130.4 = 0.964 × 65.2 = × ii vb 2 = 62.9 A/A

100 Rout = 3.3 re + β +1 100 = 3.3 0.0463 + 101 = 0.789 k = 789 6.135 Refer to the circuit in Fig. P6.135. For dc analysis, open-circuit the two coupling capacitors. Then replace the 9-V source and the two 20-k resistors by their Thévenin equivalent, namely, a 4.5-V source and a 10-k series resistance. The latter can be added to the 10-k resistor that is connected to the base. The result is the circuit shown in Fig. 1, which can be used to calculate IE . (a) IE =

4.5 − 0.7 20 2+ β +1

3.8 = 1.73 mA = 20 2+ 101

Sedra, Smith, Carusone, Gaudet

Figure 2 From Fig. 2 we see that re (10 2) v e = i e + ie 10

re v b = v e + ie re = ie (10 2) 1 + + ie r e 10 re ii = (1 − α)ie + ie 10 =

re ie + ie β +1 10

We can now obtain Rin from

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–53 re + re (10

2) 1 + vb 10 Rin ≡ = 1 re ii + β + 1 10 re + (β + 1)re (β + 1)(10 2) 1 + 10 = re 1 + (β + 1) 10 =

101 × (10 2) × (1 + 0.00145) + 101 × 0.0145 1 + 101 × 0.00145

=

168.577 + 1.4645 = 148.3 k 1 + 0.14645

vb Rin 148.3 = = = 0.937 v sig Rin + Rsig 148.3 + 10 re (10 2) 1 + i e ve vo 10 = = r e vb vb (10 2) + ie re ie 1 + 10 =

The reduced Rin will result in a reduction in v b /v sig , vb Rin 18.21 = = v sig Rin + Rsig 28.21 = 0.646 V/V vo 2 = = 0.993 vb 2 + 0.0145 vo = 0.646 × 0.993 Gv ≡ v sig = 0.64 V/V which is much reduced relative to the value obtained with bootstrapping.

6.136

1.00145 × (10 2) 1.00145 × (10 2) + 0.0145 0.5 M

= 0.991 V/V Gv ≡

5 V

Q1

vo = 0.937 × 0.991 = 0.93 V/V v sig

(c) When CB is open-circuited, the equivalent circuit becomes that shown in Fig. 3.

100 51

2 A

100 A 0.1 mA

2.5 V

Q2 0.05 mA 50 A 50 A

5 mA

(a) Applying Thévenin’s theorem to the base-biasing circuit of Q1 results in the dc circuit shown below. From our partial analysis on the figure, we can write IE1 = 0.1 mA IE2 = 5 mA

Figure 3 Thus,

VB1 can be obtained as

Rin = 20 k Rib

VB1 = 2.5 − 2 µA × 0.5 M = 1.5 V

= 20 k (β + 1)(re + 2)

and VB2 can be found as VB2 = VB1 − 0.7 = 0.8 V

= 20 101 × 2.0145 = 18.21 k which is greatly reduced because of the absence of bootstrapping. The latter causes the lower node of the 10-k base-biasing resistor to rise with the output voltage, thus causing a much reduced signal current in the 10-k resistor and a correspondingly larger effective resistance across the amplifier input.

Sedra, Smith, Carusone, Gaudet

(b) Refer to the circuit in Fig. P6.136. With a load resistance RL = 1 k connected to the output terminal, the voltage gain v o /v b2 can be found as vo RL = v b2 RL + re2 where re2 =

25 mV =5 5 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 6–54 vo 1000 = = 0.995 V/V v b2 1000 + 5

6.137 We need to raise fH by a factor of

Rib2 = (β2 + 1)(re2 + RL )

2 MHz = 4. Thus 500 kHz

= 101 × 1.005 = 101.5 k

1 + gm Re = 4

(c) Rin = 1 M 1 M (β + 1)(re1 + Rib2 ) where

⇒ Re =

3 gm

VT 25 mV = 250 = 0.25 k = IE1 0.1 mA

Since

Rin = 0.5 M [51 × (0.25 + 101.5)] k

gm =

IC 1 mA = = 40 mA/V VT 0.025 V

Re =

3 = 75 0.04

re1 =

= 0.5 M 5.2 M = 456 k Rib 101.5 v e1 = = v b1 Rib + re1 101.5 + 0.25 = 0.9975 V/V (d)

v b1 Rin 456 = 0.82 V/V = = v sig Rin + Rsig 456 + 100

(e)

vo = 0.82 × 0.9975 × 0.995 = 0.814 V/V v sig

Sedra, Smith, Carusone, Gaudet

The new value of fL will be fL =

100 Hz 100 = = 25 Hz 1 + gm Re 4

and the midband gain will become | AM | =

100 100 = 25 V/V = 1 + gm Re 4

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 7–1

Chapter 7 Solutions to Exercises within the Chapter Ex: 7.1 In the current source of Example 7.1 (Fig. 7.1) we have IO = 100 µA and we want to reduce the change in output current, IO , corresponding to a 1-V change in output voltage, VO , to 1% of IO . That is, IO =

VO 1V = 0.01IO ⇒ ro2 ro2

= 0.01 × 100 µA = 1 µA ro2 =

1V = 1 M 1 µA

ro2 =

VA × L 20 × L ⇒ 1 M = IO 100 µA

⇒L=

W W 1 1 2 μn Cox VOV2 = k n V2 2 L 2 2 L 2 OV2 µA W 1 60 µA = 200 2 (0.2)2 2 V L 2 W 120 ⇒ = = 15 ⇒ W2 L 2 200 × (0.2)2 I2 =

100 V = 5 µm 20 V/ µm

To keep VOV of the matched transistors the same W of the transistor should as that in Example 7.1, L remain the same. Therefore, 10 µm W = ⇒ W = 50 µm 5 µm 1 µm

= 15 × L2 W2 = 15 µm,

W2 W2 = 6 ⇒ W1 = = 2.5 µm W1 6

W3 = 2 ⇒ W3 = 2 × W1 = 5 µm W1 To allow the voltage at the drain of Q5 to go up to within 0.2 V of positive supply, we need VOV5 = 0.2 V: W 1 I5 = k p V2 2 L 5 OV5 1 µA W 80 µA = 80 2 (0.2)2 ⇒ 2 V L 5 W 2 × 80 = = 50 ⇒ W5 = 50 L5 L 5 80 × (0.2)2 W5 = 50 µm 50 µm W5 = 12.5 µm = 4 ⇒ W4 = W4 4

So the dimensions of the matched transistors Q1 and Q2 should be changed to

Thus:

W = 50 µm and L = 5 µm

W1 = 2.5 µm, W2 = 15 µm, W3 = 5 µm W4 = 12.5 µm, and W5 = 50 µm

Ex: 7.2 For the circuit of Fig. 7.4 we have

Ex: 7.3 VBE = VT ln(IREF /IS )

I2 = IREF

(W/L)2 (W/L)3 , I3 = IREF (W/L)1 (W/L)1

(W/L)5 and I5 = I4 (W/L)4 L1 = L2 = · · · = L5 = 1 µm and IREF = 10 µA, I2 = 60 µA, I3 = 20 µA, I4 = I3 = 20 µA, and I5 = 80 µA, we have I2 = IREF

W2 W2 I2 60 =6 ⇒ = = W1 W1 IREF 10

I3 = IREF

W3 W3 I3 20 =2 ⇒ = = W1 W1 IREF 10

W5 W5 I5 80 ⇒ = = =4 W4 W4 I4 20

To allow the voltage at the drain of Q2 to go down to within 0.2 V of the negative supply voltage, we need VOV2 = 0.2 V:

Sedra, Smith, Carusone, Gaudet

= 0.691 V From Eq. (7.21) we have ⎞ ⎛

Since all channel lengths are equal, that is,

I5 = I4

= 0.025 ln(10−3 /10−15 )

⎟ m ⎟ 1 + VO − VBE m+1⎠ VA2 1+ β ⎞ ⎛ 5 − 0.691 1 ⎟ ⎜ 1+ IO = 1 mA⎝ 1+1⎠ 100 1+ 100 = 1.02 mA ⎜ IO = IREF ⎜ ⎝

IO = 1.02 mA Ro = ro2 =

VA 100 V = 98 k 100 k = IO 1.02 mA

Ex: 7.4 From Eq. (7.23), we have VO − VBE IREF 1+ IO = 1 + (2/β) VA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 7–2

Now, Rin = 1 k, thus

VCC R

IREF

gm1 = 1 mA/V

IO

But

VO Q1

gm1 =

Q2

2(μn Cox )

IO where VBE = VT ln IS 0.5 × 10−3 = 0.025 ln = 0.673 V 10−15 2 − 0.673 IREF 1+ ⇒ 0.5 mA = 1 + (2/100) 50 IREF IREF

1= ⇒

2 × 0.4 ×

W L

W L

W L

ID1 1

× 0.1 1

= 12.5 1

To obtain =5 Ais ideal

(W/L)2 (W/L)1 W ⇒ = 5 × 12.5 = 62.5 L 2

1.02 = 0.497 mA = 0.5 mA 1.0265 VCC − VBE VCC − VBE ⇒R= = R IREF

5=

5 − 0.673 = 8.71 k 0.497 mA VOmin = VCEsat = 0.3 V R=

Ro = ro2 =

VA2 VA2 = ID2 5ID1

Thus,

For VO = 5 V, From Eq. (7.23) we have VO − VBE IREF 1+ IO = 1 + (2/β) VA 5 − 0.673 V 0.497 1+ = 0.53 mA IO = 1 + (2/100) 50

40 k =

VA2 5 × 0.1

⇒ VA2 = 20 V But L2 VA2 = VA2

Ex: 7.5 I1 = I2 = · · · = IN = IC |QREF

20 = 20 × L2

At the input node,

⇒ L2 = 1 µm

IREF = IC |QREF + IB |QREF + IB1 + · · · + IBN

Selecting L1 = L2 , then

= IC |QREF + (N + 1) IB |QREF

L1 = L2 = 1 µm

= IC |QREF + ⇒ IC |QREF =

(N + 1) IC |QREF β

W1 = 12.5 µm W2 = 62.5 µm

IREF N+1 1+ β

The actual short-circuit current-transfer ratio is given by Eq. (7.31),

Thus, IREF I1 = I2 = · · · = IN = N+1 1+ β

Ais = Q.E.D.

For β = 100, to limit the error to 10%, 0.1 =

N+1 N+1 = β 100

⇒N=9 Ex: 7.6 Rin

1 gm1

Sedra, Smith, Carusone, Gaudet

5 1+

1 gm1 ro1

Thus the percentage error is Error =

−1/gm1 ro1 × 100% 1 + (1/gm1 ro1 )

Substituting, gm1 = 1 mA/V ro1 =

VA1 V L1 20 × 1 = A = = 200 k ID1 ID1 0.1

Error =

−1/200 × 100 = −0.5% 1 + (1/200)

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 7–3

Ex: 7.7

= 1.24 mA/V

Using Eq. (7.42):

W · ID gm = 2μn Cox L

ro1 = ro2 =

L1 VAn 5 V/µm (0.36 µm) = 18 k = ID1 0.1 mA |L2 |VAp

= ID2 Voltage gain is

For ID = 10 µA, we have gm = 2(387 µA/V2 )(10)(10 µA)

6 V/µm (0.36 µm) = 21.6 k 0.1 mA

Av = −gm1 (ro1 ro2 )

= 0.28 mA/V

Av = − (1.24 mA/V) (18 k 21.6 k)

Using Eq. (7.46): 2μn Cox (W/L) A 0 = VA √ ID 5 V/µm 2(387 µA/V2 )(10)(0.36)2 = √ 10 µA

= −12.2 V/V Ex: 7.9 VCC

VBIAS

A0 = 50 V/V Since gm varies with

Q2 I

1 ID and A0 with √ , ID

for

100 ID = 100 µA ⇒ gm = 0.28 mA/V 10 = 0.88 mA/V 10 1/2 = 15.8 V/V A0 = 50 100

Q1

vi

1/2

vo

Rin

IC1 = I = 100 µA = 0.1 mA IC1 0.1 mA = 4 mA/V = VT 25 mV β 100 Rin = rπ 1 = 1 = = 25 k gm 1 4 mA/V

gm1 =

For ID = 1 mA, we have 1/2 1 gm = 0.28 mA/V = 2.8 mA/V 0.010 0.010 1/2 = 5 V/V A0 = 50 1

VA 50 V = = 500 k I 0.1 mA |VA | 50 V = = 500 k ro2 = I 0.1 mA A0 = gm1 ro1 = (4 mA/V) (500 k) = 2000 V/V ro1 =

Av = −gm1 (ro1 ro2 ) = −(4 mA/V) × Ex: 7.8

(500 k 500 k) = −1000 V/V

VDD Q3

Ex: 7.10 Refer to Fig. 1,

Q2

i

vO IREF

⫹

Q1

⫹

vI ⫺ RL

Since all transistors have the same 7.2 µm W = , L 0.36 µm

⫺

vsig ⫹ ⫺

we have IREF = ID3 = ID2 = ID1 = 100 µA

W gm1 = 2μn Cox ID1 L 1

7.2

= 2 387 µA/V2 (100 µA) 0.36

Sedra, Smith, Carusone, Gaudet

i

Rsig

vo

Rin

vo = iRL vsig = i(Rs + Rin ) Thus,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 7–4 RL vo = vsig Rs + Rin

For RL = βro ,

Q.E.D.

Rin = rπ ( Ex: 7.11 Since gm ro 1, we use Eq. (7.52), 1 RL + Rin gm gm ro 0

ro

Rin

1 gm

2 gm

Rin = rπ

∞

ro

1 rπ 2

rπ rπ =

For RL = ∞,

(gm ro )ro ∞

RL

1 βro 1 + ) = rπ ( + rπ ) gm gm ro gm

Summary:

Ex: 7.12 For gm ro 1, we use Eq. (7.57), Rout ro + (gm ro )Rs

RL

0

ro

βro

∞

Rin

re

2re

1 rπ 2

rπ

to obtain ro

(gm ro )ro

∞

Rs

0

Rout

ro (gm ro )ro (gm ro )2 ro ∞

Ex: 7.13 Substituting Rs = ro in Eq. (7.53) gives

Ex: 7.16 Using Eq. (7.69), Rout ro + (gm ro )(Re rπ ) we obtain Re

0

re

Rout ro 2ro 1 1+ gm ro Ais = 2 1+ gm ro

Ex: 7.17 Using Eq. (7.65) with Re = ro ,

Substituting the given values of (gm ro ) we obtain

gm ro

20

50

100

Ais (A/A)

0.95

0.98

0.99

Ex: 7.14 Av o remains unchanged at gm ro . With a load resistance RL connected, Av = A v o = (gm ro )

ro ∞ β + 1 ro (β + 1)ro (β + 1)ro 2 rπ

RL RL + Ro

1+ Ais =

1+

1 1 + gm ro gm (ro rπ ) 1+

=

1+

1 gm ro

1 gm ro

1 gm ro = 2 1 1+ + gm ro β 1+

2 1 + gm ro gm rπ

For β = 100 and gm ro = 100, 1+

1 100

= 0.98 A/A 1 2 + 100 100 For β = 100 and gm ro = 1000, Ais =

RL RL + (1 + gm Rs )ro

1+

1+

1 1000

= 0.99 A/A

Ex: 7.15 Since gm ro 1 we use Eq. (7.62), 1 RL Rin rπ + gm gm ro

Ais =

For RL = 0,

Ex: 7.18 Ro = [1 + gm (Re rπ )]ro

1 Rin rπ = re gm

where

For RL = ro , Rin = rπ (

1 1 2 + ) = rπ 2re gm gm gm

Sedra, Smith, Carusone, Gaudet

1+

1 2 + 1000 100

gm = 4 mA/V, rπ =

β = 25 k, gm

Re = 1 k, and ro =

VA 20 = = 200 k IC 0.1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 7–5

Thus, the allowable range of vO is from 0.5 V to 1.3 V.

Thus, Ro = [1 + 4(1 25)] × 200 1 M

Ex: 7.21 Refer to Fig. 7.31.

Without emitter degeneration,

gm1 = gm2 = gm3 = gm4 =

Ro = ro = 200 k

= 2 mA/V

2 × 0.2 2 ID = |VOV | 0.2

|VA | 2 = 10 k = ID 0.2

0.55 µm Ex: 7.19 If L is halved L = and 2 |VA | = VA · L, we obtain 0.55 µm = 1.375 V |VA | = 5 V/µm 2

Ron = (gm2 ro2 )ro1 = (2 × 10) × 10 = 200 k

|VA | |VA | 2 (1.375 V)2 · Ro = = |VOV |/2 ID (0.3 V) (100 µA)

Av = −gm1 Ro = −2 × 100 = −200 V/V

= 126 k Since ID = W = L

W 1 VSD |VOV |2 1 + μp Cox |VA | 2 L

2 (100 µA) 0.3 V 90 µA/V2 (0.3 V)2 1 + 1.375 V

ro1 = ro2 = ro3 = ro4 =

Rop = (gm3 ro3 )ro4 = (2 × 10) × 10 = 200 k Ro = Ron Rop = 200 200 = 100 k

Ex: 7.22 gm1 = gm2 = gm ID 0.1 mA = 1 mA/V = VOV (0.2/2) V 2 ro1 = ro2 = ro =

=

W = 20.3 L

VA 2V = 20 k = ID 0.1 mA

so, gm ro = 1 mA/V (20 k) = 20 (a) For RL = 20 k,

Ex: 7.20 VDD ⫽ ⫹1.8 V

VG4 ⫽ 1.1 V

Q4

Rin2 =

RL + ro2 20 k + 20 k = 1.9 k = 1 + gm2 ro2 1 + 20

∴ Av1 = −gm1 (ro1 Rin2 ) = −1 mA/V (20 1.9) = −1.74 V/V

VG3 ⫽ 0.8 V

or

Q3

If we use the approximation of Eq. (7.84),

vO VG2 ⫽ 1.0 V

Q2

Rin2 ≈

RL 1 20 k 1 + = + = 2 k gm2 ro2 gm2 20 1 mA/V

then Av1 = −1 mA/V (20 k 2 k) = −1.82 V/V

Q1 VI ⫽ 0.7 V

Continuing, from Eq. (7.81), Av = −gm1 (gm2 ro2 ro1 ) RL Av = −1 mA/V {[(20) (20 k)] 20 k} = −19.0 V/V

If all transistors are matched and are obviously operating at the same ID , then all |VOV | will be equal and equal to that of Q1 , namely, |VOV | = 0.7 − 0.5 = 0.2 V To keep Q2 in saturation, v Omin = VG2 − |Vt | = 0.5 V To keep Q3 in saturation, v Omax = VG3 + |Vt | = 1.3 V

Sedra, Smith, Carusone, Gaudet

Av 2 =

Av −19.0 = 10.5 V/V = Av 1 −1.82

(b) Now, for RL = 400 k, Rin2

1 RL 1 400 k + + = gm2 ro2 gm2 20 1 mA/V

= 21 k Av 1 = −1 mA/V (20 k 21 k) = −10.2 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 7–6

Av = −1 mA/V [(20) (20 k)] 400 k

Av = −4186 V/V

= −200 V/V

Av max occurs when ro1 and ro4 are rπ .

Av 2 =

Av −200 = = 19.6 V/V Av 1 −10.2

Then Ron = (gm2 ro2 ) rπ2 = β 2 ro2

Ex: 7.23 Referring to Fig. 7.34,

Ron = 100 (25 k) = 2.5 M

Rop = (gm3 ro3 ) (ro4 rπ 3 ) and

Rop = (gm3 ro3 ) rπ3 = β 3 ro3

Ron = (gm2 ro2 ) (ro1 rπ 2 )

Rop = 50 (20 k) = 1 M

The maximum values of these resistances are obtained when ro rπ and are given by Ron = (gm2 ro2 ) rπ 2 max Rop = (gm3 ro3 ) rπ 3

Finally,

max

Av max = −(8 mA/V) (2.5 M 1.0 M) Av max = −5714 V/V Ex: 7.25 2ID 2 × 0.2 gm = = 2 mA/V = VOV 0.2

Since gm rπ = β, Ron = β 2 ro2 max Rop = β 3 ro3

gmb = χ gm = 0.2 × 2 = 0.4 mA/V

Since Av = −gm1 Ron Rop ,

|Av max | = gm1 β 2 ro2 β 3 ro3

RL = ro1 ro3

ro1 = ro3 =

max

Ex: 7.24 For the npn transistors, gm1 = gm2 =

|IC | 0.2 mA = 8 mA/V = |VT | 25 mV

rπ 1 = rπ2 =

β 100 = 12.5 k = gm 8 mA/V

ro1 = ro2 =

|VA | 5V = 25 k = |IC | 0.2 mA

From Fig. 7.34, Ron = (gm2 ro2 ) (ro1 rπ 2 ) = (8 mA/V) (25 k) (25 k 12.5 k) Ron = 1.67 M For the pnp transistors, gm3 = gm4 =

|IC | 0.2 mA = 8 mA/V = VT 25 mV

rπ 3 = rπ4 =

β 50 = 6.25 k = gm 8 mA/V

ro3 = ro4 =

|VA | 4V = 20 k = |IC | 0.2 mA

VA 5 = = 25 k ID 0.2 1 = 25 25 2.5 k gmb

= 2.083 k RL 2.083 vo = 0.81 V/V = = 1 1 vi RL + 2.083 + gm 2 To obtain Rout we use Eq. (7.95), Rout =

1 1 ro1 ro3 gm gmb

= 0.52.52525 = 0.403 k = 403 Ex: 7.26 Refer to the circuit in Fig. 7.36. All transistors are operating at ID = IREF = 100 µA and equal VOV , found from W 1 2 VOV ID = μn Cox 2 L 3.6 1 2 × 387 × × VOV 2 0.36 ⇒ VOV = 0.227 V 100 =

VGS = 0.227 + 0.5 = 0.727 V VOmin = VG3 − Vt3

Rop = (gm3 ro3 ) (ro4 rπ 3 )

= VGS4 + VGS1 − Vt3

= (8 mA/V) (20 k) (20 k 6.25 k)

Thus,

Rop = 762 k

Av = −gm1 Ron Rop

VOmin = 2VGS − Vt

= − (8 mA/V) (1.67 M 762 k)

= 0.5 + 2 × 0.227 = 0.95 V

Sedra, Smith, Carusone, Gaudet

= Vt + 2 VOV

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 7–7

gm =

2 ID 2 × 0.1 = = 0.88 mA/V VOV 0.227

ro =

VA V L 5 × 0.36 = 18 k = A = ID ID 0.1

Ro = (gm3 ro3 )ro2 = (0.88 × 18) × 18 = 285 k Ex: 7.27 For the Wilson mirror from Eq. (7.97), we have 1 IO = 0.9998 2 IREF 1+ 2 β Thus

|IO − IREF | × 100 = 0.02% IREF

whereas for the simple mirror from Eq. (7.18) we have 1 IO = 0.98 = 2 IREF 1+ β Hence

|IO − IREF | × 100 = 2% IREF

For the Wilson current mirror, we have Ro =

100 × 100 k βro = = 5 M 2 2

Sedra, Smith, Carusone, Gaudet

and for the simple mirror, Ro = ro = 100 k. Ex: 7.28 For the two current sources designed in Example 7.5, we have gm =

mA IC 10 µA = 0.4 = VT 25 mV V

ro =

VA 100 V = 10 M, = IC 10 µA

rn =

β = 250 k gm

For the current source in Fig. 7.40(a), we have Ro = ro2 = ro = 10 M For the current source in Fig. 7.40(b), from Eq. (7.105), we have Rout 1 + gm (RE rn ) ro From Example 7.5, RE = R3 = 11.5 k; therefore, mA Rout 1 + 0.4 (11.5 k 250 k) 10 M V ∴ Rout = 54 M

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–1

Solutions to End-of-Chapter Problems

7.3

7.1 Refer to Fig. 7.1. 1 IO = kn 2 80 =

W L

VDD

VDD

2 VOV

Q1

1 2 × 400 × 10 × VOV 2

Q2 VO

⇒ VOV = 0.2 V VGS = Vt + VOV = 0.5 + 0.2 = 0.7 V VDD − VGS IREF = 80 µA = R 1.8 − 0.7 = 13.75 k ⇒R= 0.08 mA For Q2 to operate in saturation, VO must not decrease below VG2 by more than Vt , thus VOmin = VG − Vt = VGS − Vt = VOV = 0.2 V

IREF

R

IO

Set |VOV | = VDD − VOmax = 1.3 − 1.1 = 0.2 V VG = VDD − Vtp − |VOV | = 1.3 − 0.4 − 0.2 = 0.7 V

7.2 Refer to Fig. 7.1. 1 W kn V2 2 L 1 OV 1 1 2 VOV 10 = × 400 × 2 0.4

IREF = 10 µA =

⇒ VOV = 0.14 V

0.7 V VG = 8.75 k = ID1 80 µA W 1 |VOV |2 ID = μp Cox 2 L

R=

thus 2ID 2 × 80 µA W = = = 50 L μp Cox |VOV |2 80 µA/V2 × 0.22

VGS = Vt + VOV = 0.4 + 0.14 = 0.54 V VDD − VGS IREF 1.2 − 0.54 = 66 k = 0.01 mA

7.4 Refer to the circuit of Fig. P7.4. For Q2 to operate properly (i.e., in the saturation mode) for drain voltages as high as +0.8 V, and provided its width is the minimum possible, we use

Since

|VOV | = 0.2 V

R=

Note that all three transistors Q1 , Q2 , and Q3 will be operated at this value of overdrive voltage. For Q1 ,

IO

60 = 6, = IREF 10 W W =6 L 2 L 1 But L1 = L2 , thus W2 = 6 W1 = 6 × 1 = 6 µm VOmin = VOV = 0.2 V Ro = ro2 =

VA L2 6 × 0.4 = 40 k = IO 0.060 mA

ID1 = IREF = 20 µA W 1 ID1 = μp Cox |VOV |2 2 L 1 W 1 20 = × 100 × × 0.04 2 L 1 W = 10 ⇒ L 1 For L = 0.5 µm,

For

W1 = 5 µm

VO = +0.2 V,

Now, for

IO =

+0.2 V +0.2 V = 0.005 mA or 5 µA = Ro 40 k

Sedra, Smith, Carusone, Gaudet

I2 = 100 µA = 5IREF , we have (W/L)2 =5 (W/L)1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–2 ⇒

W L

The output resistance of the current sink Q5 is

= 5 × 10 = 50 2

ro5 =

W2 = 50 × 0.5 = 25 µm For

=

I3 = 40 µA = 2IREF , we obtain (W/L)3 =2 (W/L)1 W ⇒ = 20 L 3

VA5 V × L = An I5 I5

5 × 0.5 = 31.25 k 80

7.5 Using Eq. (7.22) with m = 1, IO = IREF

W3 = 10 µm We next consider Q4 and Q5 . For Q5 to operate in saturation with the drain voltage as low as −0.8 V, and for it to have the minimum possible W/L, we operate Q5 at

VO − VBE 1+ 2 VA2 1+ β 1

where IREF = 100 µA, β = 100, VA2 = 35 V, and

This is the same overdrive voltage at which Q4 will be operating. Thus, we can write for Q4 ,

IREF IS 100 × 10−6 = 0.025 ln 6 × 10−18

I4 = I3 = 40 µA

= 0.761 V

VOV = 0.2 V

and using

VBE = VT ln

Thus, at VO = 2 V,

W 1 μn Cox V2 2 L 4 OV W 1 40 = × 400 × × 0.22 2 L 4 W ⇒ =5 L 4 ID4 =

IO = 100 ×

2 − 0.761 × 1+ 2 35 1+ 100 1

= 101.5 µA Ro = ro2 =

W4 = 2.5 µm

VA2 35 = 350 k = IOnominal 0.1

For

Finally, since

VO = 1 V, IO =

I5 = 80 µA = 2 I4 , W W =2 L 5 L 4 W ⇒ = 10 L 5

VO 1V = 2.9 µA Ro 350 k

7.6

VDD

VDD

W5 = 5 µm To find the value of R, we use |VSG1 | = |Vtp | + |VOV1 |

Q1

Q2

= 0.5 + 0.2 = 0.7 V R=

IO

0.3 V 1 − |VSG1 | = IREF 0.02 mA

IREF

= 15 k The output resistance of the current source Q2 is ro2 = =

|VAp |×L |VA2 | = I2 I2

5 × 0.5 = 25 k 0.1 mA

Sedra, Smith, Carusone, Gaudet

For identical transistors, the transfer ratio is IO IREF

=

1 = 1 + 2/β

Microelectronic Circuits, 8th International Edition

1 1+

2 50

= 0.96

© Oxford University Press 2021

Chapter 7–3

This figure belongs to Problem 7.7.

7.7 The solution is given in the circuit diagram. Note that the starting point is calculating the current I in the Q1 –R1 –Q2 branch. See figure above.

Vx = 1.3 − 1.187 = 0.113 V When VO = Vx , the Early effect on Q1 and Q2 will be the same, and IO = IREF /(1 + 2/β 2 )

7.8 Refer to Fig. 7.11.

Thus, IO will be

IC1 IREF = 0.1 mA

100 = 1 + (2/1002 ) 99.98 µA, for an error of −0.02 µA or −0.02%.

VBE1

1 mA = 0.7 − 0.025 ln 0.1 mA IC1 β

0.1 = 0.002 mA 100 1 mA = 0.7 − 0.025 ln 0.002 mA

=2× VBE3

1 = 1 + (2/1002 ) 0.9998 mA, for an error of −0.0002 mA or −0.02%. For proper current-source operation, the minimum required voltage at the output is the value needed to keep Q3 in the active region, which is approximately 0.3 V. IREF = 1 mA ⇒ IO =

= 0.642 V IC3 IB1 + IB2 = 2 IB1 = 2 ×

IREF = 100 µA ⇒ IO =

= 0.545 V

7.9 Using Eq. (7.28),

Vx = VBE3 + VBE1 = 1.187 V

Rin = ro1

If IREF is increased to 1 mA,

where

VBE1 = 0.7

VA V L 10 × 0.5 = A = = 50 k ID1 ID1 0.1 mA W ID1 gm1 = 2μn Cox L 1 10 = 2 × 0.5 × × 0.1 = 1.414 mA/V 0.5

IC3 0.02 mA VBE3

ro1 =

1 = 0.6 V = 0.7 − 0.025 ln 0.02

Vx = 1.3 V Thus,

Sedra, Smith, Carusone, Gaudet

1 gm1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–4

1 = 0.71 k gm1

VCSmin = VOV = 0.2 V

Thus,

Ro = ro2 =

(W/L)2 50/0.5 = 5 A/A = (W/L)1 10/0.5

RO = ro2 =

VA L

=

IOnominal IOnominal 5 × 0.54 = 27 k Ro = 0.1 VO − VGS IO = IOnominal 1 + VA2

Rin = 50 0.71 = 0.7 k = 700 Ais =

VA2

VA V L = A ID2 ID2

where

10 × 0.5 = = 10 k 5 × 0.1

VGS = 0.5 + 0.2 = 0.7 V

7.10 Refer to Fig. 7.1.

So,

IO = 10% IO

VO − 0.7 IO = 100 1 + 2.7

VA2 = 5 × 0.54 = 2.7 V

IO = 0.1 × 150 = 15 µA

For

VO = 1.8 − 0.3 = 1.5 V ro =

VO 1.5 V = = 100 k IO 15 µA

VO = 0.2 V, 0.2 − 0.7 = 81.5 µA IO = 100 1 + 2.7

But ro =

VA V L = A IO IO

For

10 × L ⇒ L = 1.5 µm 0.15 ⇒ VA = 15 V 100 =

VO = 0.7 V, 0.7 − 0.7 = 100 µA IO = 100 1 + 2.7

VOV = VDS2min = 0.3 V

For

VGS = Vt + VOV = 0.5 + 0.3 = 0.8 V W VDS 1 2 VOV 1+ ID = k n 2 L VA 1 W 0.8 150 = × 400 × × 0.09 1 + 2 L 15

VO = 1.2 V, 1.2 − 0.7 = 118.5 µA IO = 100 1 + 2.7 For

W = 7.91 L W = 7.91 × 1.5 = 11.9 µm

⇒

R=

VO = 1.7 V, 1.7 − 0.7 = 137 µA IO = 100 1 + 2.7

VDD − VGS 1.8 − 0.8 = 6.7 k = IREF 0.15

7.12 Referring to Fig. P7.12, we obtain

7.11 Refer to Fig. 7.2.

VGS1 = VGS2 , so

1 W (μn Cox ) V2 2 L 1 OV 1.74 1 2 × VOV 25 = × 387 × 2 0.54 ⇒ VOV = 0.2 V

IREF =

IOnominal = IREF

(W/L)2 = IREF (W/L)1

IOnominal = 25 × 4 = 100 µA

Sedra, Smith, Carusone, Gaudet

⇒ ID2 = IREF

ID2 (W/L)2 = ID1 (W/L)1

(W/L)2 (W/L)1

ID3 = ID2

W2 W1

VGS3 = VGS4 , thus

ID4 (W/L)4 = ID3 (W/L)3

⇒ IO = ID4 = IREF

(W/L)2 (W/L)4 . (W/L)1 (W/L)3

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–5

7.13 Referring to the figure, suppose that Q1 has W = 10 µm, Q2 has W = 20 µm, and Q3 has W = 40 µm. VDD

Q1

Q2

I1

Q3

I2

I3

IREF

So three different currents are obtained with double-diode connects. To find VSG , we use the following for the diode-connected transistor(s): W 1 (VSG − |Vtp |)2 ID = μp Cox 2 L and substitute ID = IREF = 100 µA. Thus W 1 (VSG − 0.6)2 100 = × 100 × 2 1 µm 2 ⇒ VSG = 0.6 + W(µm) For the six cases above we obtain (1) W = W1 = 10 µm ⇒ VSG = 1.05 V

(1) With Q1 diode connected, 20 (W/L)2 = 200 µA = 100 µA I2 = IREF 10 (W/L)1 40 = 400 µA I3 = 100 µA 10

(2) W = W2 = 20 µm ⇒ VSG = 0.92 V

(2) With Q2 diode connected, and W = 20 µm, 10 = 50 µA I1 = 100 µA 20 40 = 200 µA I3 = 100 µA 20

(6) W = W1 + W3 = 50 µm ⇒ VSG = 0.80 V

(3) If Q3 with W = 40 µm is diode connected, 10 I1 = 100 µA = 25 µA 40 20 = 50 µA I2 = 100 µA 40 So, with only one transistor diode connected, we can get 25 µA, 50 µA, 200 µA, and 400 µA, or four different currents. Now, if two transistors are diode connected, the effective width is the sum of the two widths. (4) If Q1 and Q2 are diode connected, then Weff = 20 + 10 = 30 µm, so that 40 = 133 µA I3 = 100 µA 30 (5) If Q2 and Q3 are diode connected, then Weff = 20 + 40 = 60 µm, so that 10 = 16.7 µA I1 = 100 µA 60 (6) If Q1 and Q3 are diode connected, Weff = 10 + 40 = 50 µm, so that 20 = 40 µA I2 = 100 µA 50

Sedra, Smith, Carusone, Gaudet

(3) W = W3 = 40 µm ⇒ VSG = 0.82 V (4) W = W1 + W2 = 30 µm ⇒ VSG = 0.86 V (5) W = W2 + W3 = 60 µm ⇒ VSG = 0.78 V

7.14 Nominally, IO = IREF = 1 mA ro2 =

VA2 90 = 90 k = IO 1

ro2 =

VO 10 − 1 ⇒ = 90 ⇒ IO = 0.1 mA IO IO

IO 0.1 = = 10% change IO 1 7.15 (a) If IS = 10−17 A and we ignore base currents, then IREF = IS eVBE /VT so that IREF VBE = VT ln 10−17 For IREF = 10 µA, −5 10 = 0.691 V VBE = 0.025 ln 10−17 For IREF = 10 mA, −2 10 = 0.863 V VBE = 0.025 ln 10−17 So for the range of 10 µA ≤ IREF ≤ 10 mA, 10 µA ≤ IO ≤ 10 mA and, 0.691 V ≤ VBE ≤ 0.863 V (b) Accounting for finite β,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–6

IO = IREF ·

1 m ≤ 80 − 1 − 1 = 7.88 0.9

1 1 + 2/β

For IREF = 10 µA,

Thus, the largest current transfer ratio possible is 7.88.

10 µA = 9.62 µA IO = 2 1+ 50 For IREF = 0.1 mA,

7.17 Using Eq. (7.21), VO − VBE m 1+ IO = IREF m+1 VA2 1+ β

0.1 mA = 0.098 mA 2 1+ 100 For IREF = 1 mA, IO =

This equation can be adapted for the pnp mirror of Fig. P7.17 by substituting m = 1, replacing VO with the voltage across Q3 , namely (3 − VO ), replacing VBE with VEB , and VA2 with |VA |:

1 mA = 0.98 mA 2 1+ 100 For IREF = 10 mA, IO =

IO =

IO = IREF

1 + [(3 − VO − VEB )/|VA |] 1 + (2/β)

(1)

Now, substituting IO = 1 mA, VO = 1 V, β = 50, |VA | = 50 V, and −3 IO 10 = 0.691 V VEB = VT ln = 0.025 ln IS 10−15

10 mA = 9.62 mA 2 1+ 50

7.16

results in 1 × (1 + 0.04) = 1.013 mA 3 − 1 − 0.691 1+ 50 VCC − VEB 3 − 0.691 = 2.28 k R= = IREF 1.013 IREF =

Maximum allowed voltage VO = 3 − 0.3 = 2.7 V. For VO = 2.7 V, Eq. (1) yields 3 − 2.7 − 0.691 50 IO = 1.013 = 0.966 mA 1.04 For VO = −5 V, Eq. (1) yields 1+

IO = mIC1 A node equation at the collector of Q1 yields IREF = IC1 +

IO + IC1 β

3 − (−5) − 0.691 50 = 1.116 mA IO = 1.013 1.04 Thus, the change in IO is 0.15 mA. 1+

Substituting IC1 = IO /m results in IO = IREF

m m+1 1+ β

7.18 Refer to the circuit in Fig. P7.18.

Q.E.D.

V2 = 2.7 − VEB = 2.7 − 0.7 = +2 V

For β = 80 and the error in the current transfer ratio to be limited to 10%, that is, m ≥ 0.9m m+1 1+ β m+1 1 1+ ≤ β 0.9 1 m+1 ≤ −1 β 0.9 1 m≤β −1 −1 0.9

Sedra, Smith, Carusone, Gaudet

V3 = 0 + VEB = +0.7 V Thus, Q3 and Q4 are operating in the active mode, and each is carrying a collector current of I/2. The same current is flowing in Q2 and Q1 ; thus I V1 = −2.7 + R 2 But V1 = −VBE1 = −0.7 Thus, 1 −0.7 = −2.7 + IR 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–7

⇒ IR = 4 V The current I splits equally between Q5 and Q6 ; thus I R = −2.7 + 2 = −0.7 V V4 = −2.7 + 2 I R = −2.7 + 1 = −1.7 V V5 = −2.7 + 2 2 Thus, Q5 and Q6 are operating in the active mode as we have implicitly assumed. Note that the values of V1 , V2 , V3 , V4 , and V5 do not depend on the value of R. Only I depends on the value of R: 4 = 0.4 mA (a) R = 10 k ⇒ I = 10 4 = 0.04 mA (b) R = 100 k ⇒ I = 100

7.20 (a) Figure 1 shows the current conveyor circuit with Y connected to a voltage V, X fed with a current source I, and Z connected to a voltage VZ that keeps Q5 operating in the active mode. Assuming that all transistors are operating in the active mode and that β 1, so that we can neglect all base currents, we see that the current I through Q1 will flow through the two-output mirror Q3 , Q4 , and Q5 . The current I in Q5 will be drawn from Q2 , which forms a mirror with Q1 . Thus VEB2 = VEB1 and the voltage that appears at X will be equal to V. The current in Q5 will be equal to I, thus terminal Z sinks a constant current I.

7.19 There are various ways this design could be achieved, but the most straightforward is the one shown:

Q1 1

R

2

4

8

0.2 mA

0.4 mA

0.8 mA

0.5 mA Q2 1

5

1 mA

10

Figure 1

2 mA

20

(b)

With this scheme, 5 − 0.7 − 0.7 − (−5) = 86 k 0.1 mA and each transistor has EBJ areas proportional to the current required. Multiple, parallel transistors are acceptable. R=

Note: This large value of R is not desirable in integrated form; other designs may be more suitable. Figure 2

Even without knowing exact circuitry, we can find the total power dissipation as approximately PT = PCC + PEE +5 V (0.1 + 0.5 + 1 + 2) mA

Figure 2 shows the special case of V = 0 V. As before, the voltage at X, VX , will be equal to V. Thus

PT = 7.5 mW + 18 mW = 25.5 mW

VX = 0

PT = 5 V (0.1 + 0.2 + 0.4 + 0.8) mA

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–8

That is, a virtual ground appears at X, and thus the current I that flows into X can be found from 5−0 5 − VX = = 0.5 mA I= 10 k 10 This is the current that will be mirrored to the output, resulting in IZ = 0.5 mA.

determine Rin , we obtain the equivalent circuit shown.

C3 aie3

ix

7.21

B3

x C1

IB3 Q3

X IC1

IO1

Q1 b

IO3

IOn

re1

ie1 Q21

IO

IO2

Q22

Q23

aie2

E3

B1 i b1

IE3

C2

re3

aie1

vx

IREF

Out ie3

ib2 B2

ve re2

ie2

Q2n

nIO b

In this equivalent circuit, re1 = re2 = re =

VT αVT = IE3 2(1 − α)IREF

IO1 = IO2 = IO3 · · · = IOn = IO = IC1

re3 =

The emitter of Q3 supplies the base currents for all transistors, so

ie1 = ie2 = ie

IE3 =

ie3 = ib1 + ib2 = 2(1 − α)ie

(n + 1) IO β

IREF = IB3 + IO = IO = IREF

VT αVT αVT = = IE IC IREF

From the figure we obtain (n + 1) IO + IO β (β + 1)

1 1 n+1 (n + 1) 1+ 1+ β (β + 1) β2

ix = αie1 + (1 − α)ie3 = αie + (1 − α) × 2(1 − α)ie = ie [α + 2(1 − α)2 ] But 2(1 − α)2 α. Thus,

For the deviation from unity to be kept ≤ 0.2%

ix αie

n+1 ≤ 0.002 β2

vx = ie3 re3 + ie1 re1

⇒ nmax = 0.002 × 150 − 1 = 44 2

7.22 Refer to Fig. 7.11 and observe that IC1 IREF and IC2 = IC1 ; thus each of Q1 and Q2 is operating at a collector bias current approximately equal to IREF . Transistor Q3 is operating at an emitter bias current IE3 = IB1 + IB2

2(1 − α) IC = 2IB = 2IC /β = α

2(1 − α) IREF = α Replacing each of the three transistors with its T model and applying an input test voltage vx to

Sedra, Smith, Carusone, Gaudet

(1)

αVT αVT + ie 2(1 − α)IREF IREF VT VT vx = αie + IREF IREF = 2(1 − α)ie

Now, using ix = αie from Eq. (1), we have v x = ix ×

2VT IREF

Thus, Rin ≡

vx 2VT = ix IREF

Q.E.D.

For IREF = 100 µA = 0.1 mA, Rin =

2 × 25 mV = 500 0.1 mA

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–9 vd1 1 gm1 = −gm1 × =− vi gm2 gm2

(W/L)2 (W/L)1

7.23 Ais = 4 =

Since L1 = L2 , then

g2

W2 =4 W1 1 1 gm1 gm1

Rin = ro1

vgs2

For

Thus, 2 × 0.4 ×

2=

W L

W L

× 0.2

= 25

vo = id3 RL = gm1 vi

VA V L = A ID2 ID2

20 L 4 × 0.2

W1 = 25 × 0.8 = 20 µm W2 = 4 W1 = 80 µm The actual Ais is given by Eq. (7.31), 1 (W/L)2 (W/L)1 1 + (1/gm1 ro1 )

1 rπ 1 rπ 2 gm1

Since ro1 rπ 1 , Rin

(W/L)2 =4 (W/L)1 VA V L1 20 × 0.8 = 80 k = A = ID1 ID1 0.2

Thus, 4 1+

1 2×80

1 rπ 1 rπ 2 gm1

(1)

Now,

gm1 = 2 mA/V

Ais =

7.25 Replacing Q1 and Q2 with their small-signal hybrid-π models results in the equivalent circuit shown in the figure on the next page. Observe that the controlled source gm1 vπ 1 appears across its controlling voltage vπ 1 ; thus the controlled source can be replaced with a resistance (1/gm1 ). The input resistance Rin can now be obtained by inspection as Rin = ro1

where

ro1 =

W3 RL W2

Thus, the small-signal voltage gain will be vo = gm1 RL (W3 /W2 ) vi

⇒ L = 0.8 µm

Ais =

(W/L)3 W3 = gm1 vi (W/L)2 W2

which flows through RL and produces the output voltage vo ,

1

Thus, 20 =

The signal current in the drain of Q1 , gm1 vi , will be mirror in the drain of Q3 ;

1

RO = ro2 =

1 gm2 S3

id3 = gm1 vi

g2, d2

gm2vgs2

S2

Rin = 500 ⇒ gm1 = 2 mA/V W gm1 = 2μn Cox ID1 L 1

⇒

d2

β1 β = gm1 gm1

rπ 2 =

β2 β = gm2 gm2

Since IC2 = mIC1 , gm2 = mgm1 , thus

= 3.975 A/A

rπ 2 =

indicating an error of −0.6%. 7.24 Refer to Fig. P7.24. Consider first the diode-connected transistor Q2 . From the figure we see that from a small-signal point of view it is equivalent to a resistance 1/gm2 . Thus the voltage gain of Q1 will be

Sedra, Smith, Carusone, Gaudet

rπ 1 =

β mgm1

Substituting for rπ 1 and rπ 2 into Eq. (1) we obtain m+1 ) Rin = 1/ gm1 (1 + β

(2)

The short-circuit output current io is given by io = gm2 vπ 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–10

This figure belongs to Problem 7.25.

C1

B1, B2

C2

ii

vp1 ro1

rp1

io

rp2 vp2

gm1vp1

ro2 gm2vp2

Rin

Ro

Since vπ 2 = vπ 1 = ii Rin , then the short-circuit current gain Ais is given by Ais =

io = gm2 Rin = mgm1 Rin = ii

m m+1 1+ β

For situations where β1 and β2 are large, we can neglect rπ 1 and rπ 2 in Eqs. (1) and (2) to obtain Rin 1/gm1 Ais gm2 /gm1 Finally, by inspection Ro = ro2 For I = 0.1 mA, m = 5, β = 100, and VA = 30 V, gm1 = 4 mA/V ⇒ Rin =

1 = 0.236 k 5+1 4 1+ 100

(b) Replacing the BJT with its hybrid-π model results in the equivalent circuit in Fig. 2 (see next page). Observing that the controlled-source gm vπ appears across its control voltage vπ , we can replace it by a resistance 1/gm , as indicated. Next the two parallel resistances 1/gm and rπ can be combined as 1 × rπ rπ rπ gm = = = re 1 1 + gm rπ β +1 + rπ gm

re =

VT 25 mV = 250 = IE 0.1 mA

7.27 Refer to Fig. 7.13(b).

7.26 (a)

vgs

1 = 1.6 k gm

Thus, the diode-connected BJT has a small-signal resistance re . For the given data,

5 = 4.72 A/A Ais = 5+1 1+ 100 30 = 60 k Ro = 5 × 0.1

g

diode-connected MOS transistor is 1/gm . For the given values, W ID gm = 2μn Cox L √ = 2 × 0.2 × 10 × 0.1 = 0.632 mA/V

d

gm =

IC I 0.5 mA = 20 mA/V = = VT VT 0.025 V

ro =

VA 100 V = 200 k = IC 0.5 mA

g,d 1 gm

gmvgs

s

s Figure 1

β 100 = 5 k = gm 20 mA/V

Av o = −A0 = −gm ro = −20×200 = −4000 V/V Ro = ro = 200 k

Replacing the MOSFET with its hybrid-π model but neglecting ro results in the equivalent circuit in Fig. 1. Observing that the controlled-source gm vgs appears across its control voltage vgs , we can replace it by a resistance 1/gm , as indicated. Thus the small-signal resistance of the

Sedra, Smith, Carusone, Gaudet

Rin = rπ =

To raise Rin by a factor of 5 by changing I, the value of I must be lowered by the same factor to I = 0.1 mA. Now, gm is reduced by a factor of 5 and ro is increased by a factor of 5, keeping Av o unchanged at −4000 V/V. However, Ro will be increased to

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–11

This figure belongs to Problem 7.26, part (b).

Figure 2 This figure belongs to Problem 7.27.

Rsig 5 k

vsig

rp 25 k

vp

vo

RL 100 k

ro 1 M

gmvp

gm 4 mA/V Ro = 5 × 200 k = 1 M

ID =

If the amplifier is fed with a signal source having Rsig = 5 k and a 100-k load resistance is connected to the output, the equivalent circuit shown below results. rπ vo = × −gm (ro RL ) vsig rπ + Rsig =−

25 × 4 (1000 k 100 k) 25 + 5

= −303 V/V

0.1 = ⇒

1 μn Cox 2

W L

2 VOV

1 W × 0.4 × × 0.04 2 L

W = 12.5 L

7.30 Refer to Fig. 7.15(a) and Table K.1. VG = VDD − VSG2 = VDD − (| Vtp | + | VOV |) = 1.8 − (0.5 + 0.2)

7.28 From Eq. (7.46) we see that A0 is inversely proportional to ID . Thus

= 1.1 V

ID = 100 µA

A0 = 50 V/V

ID1 = ID2 = I = 100 µA

ID = 25 µA

A0 = 100 V/V

ID = 400 µA

A0 = 25 V/V

From Eq. (7.42), gm is proportional to ID . Thus changing ID from 100 µA to 25 µA reduces gm by a factor of 2. Changing ID from 100 µA to 400 µA increases gm by a factor of 2. 7.29 A0 =

2VA L

2VA = VOV VOV

2×5×L 0.2 ⇒ L = 0.5 µm 25 =

1 W (μn Cox )( )1 | VOV |2 2 L W 1 100 = × 387 × ( )1 × 0.22 2 L W ⇒ = 12.92 L 1

ID1 =

1 W (μp Cox )( )2 | VOV |2 2 L W 1 100 = × 86 × ( )2 × 0.22 2 L W ⇒ = 58.14 L 2

ID2 =

2ID 2I = gm = VOV VOV

ro1 =

2I 1= 0.2 ⇒ I = 0.1 mA

ro2 =

Sedra, Smith, Carusone, Gaudet

|L | VAn 5L = = 50L k ID1 0.1 | VAp |L

ID2

=

6L = 60L k 0.1

Av = −gm1 (ro1 ro2 )

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–12

where gm1 =

1 W (μn Cox )( )1 | VOV1 |2 2 L 1 W 100 = × 511 × ( )1 × 0.152 2 L W ⇒ ( )1 = 17.4 L ID1 =

2ID1 2 × 0.1 = = 1 mA/V | VOV | 0.2

(ro1 ro2 ) = (50L60L) = 27.3L k Av = −20 V/V

For Q2 and Q3 , 1 W (μp Cox )( )2,3 | VOV |2 2 L 1 W 100 = × 128 × ( )2,3 × 0.152 2 L W W ⇒ ( )2 = ( )3 = 69.4 L L

Thus,

ID2,3 =

− 20 = −1 × 27.3L ⇒ L = 0.73 µm

7.31 Refer to Fig. P7.31. The gain of the first stage is

(b)

Av 1 = −gm1 (ro1 /2)

Av = −gm1 (ro1 ro2 )

where (ro1 /2) is the equivalent resistance at the output of Q1 and includes ro1 in parallel with the output resistance of the current-source load, which is equal to ro1 . Similarly, the gain of the second stage is

where gm1 =

2ID1 2 × 0.1 = 1.33 mA/V = VOV 0.15

ro1 =

5 × 0.4 VA1 V L = 20 k = An = ID1 I 0.1

Now because VAn = |VAp | = |VA | and both Q1 and Q2 are operating at equal currents I, we have

ro2 =

|L | VAp 6 × 0.4 | VA2 | = = 24 k = ID2 I 0.1

ro1 = ro2 = ro

Thus,

The overall voltage gain Av will be

Av = −1.33(2024)

Av = Av 1 Av 2

= −14.5 V/V

Av 2 = −gm2 (ro2 /2)

1 Av = gm1 gm2 ro2 4 If the two transistors are operated at equal overdrive voltages, |VOV |, both will have equal gm , Av =

(c) VOV1 ≤ vO ≤ VDD − | VOV2 | 0.15 V ≤ vO ≤ (1.3 − 0.15) V 0.15 V ≤ vO ≤ 1.15 V

1 (gm ro )2 4

and 2×5 10 2|VA | = = gm ro = |VOV | |VOV | |VOV | 2 10 1 Av = 400 = × 4 |VOV | ⇒ |VOV | = 0.25 V

(d) To keep the gain unchanged, RD = ro2 = 24 k To keep the signal swing at the output unchanged, we must bias Q1 so that VDS1 = midpoint of signal swing, thus VDS1 = 0.65 V

7.32 Refer to Fig. 7.16(a) and to Table K.1 (a) VI = Vtn + VOV1 = 0.4 + 0.15 = 0.55 V For Q1 ,

Sedra, Smith, Carusone, Gaudet

We can now find the required value of the power supply, VDD = VDS1 + ID RD

= 0.65 + 0.1 × 24 = 3.05 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–13

which is much larger than the voltages allowed in the 0.13-µm technology. 7.33

If L is to be an integer multiple of 0.13 µm, then L = 0.39 µm The gain achieved will be −10.6 V/V.

1.3 V

To raise the gain to 20 V/V, ro1 ro2 has to be raised to 40 k, which requires L = 2 × 0.367 = 0.734 Again, to use a multiple of 0.13 µm we select L = 0.78 µm. This doubling of L will mean a doubling of W for each of the three transistors. Thus the area (WL) of each transistor will increase by a factor of 4 and the total gate area of the circuit will increase by the same factor. 7.34 (a) IREF = IC3 =

For an output of 1.1 V, VSD2min = |VOV2 | = 1.3 − 1.1 = 0.2 V, For an output of 0.2 V, VDS1min = 0.2 V,

2.7 − VBE3 50 k

2.7 − 0.7 50 = 0.04 mA

IREF =

⇒ IC2 = 5IC3

thus

IC2 = I = 0.2 mA ⇒ I = 0.2 mA

VOV1 = 0.2 V

2.7 V

Since ID2 = ID3 = ID1 = 50 µA

1 2 μ Cox (W/L) VOV , we have 2 p W W 2ID2

= = L 2 L 3 μp COX |VOV |2

and ID =

2.7 V

vo

50 k

2 (50 µA)

= = 19.5 128 µA/V2 (0.2 V)2

I Q3

Q2

For Q1 , W 2 (50 µA)

= = 4.9 L 1 511 µA/V2 (0.2 V)2 (b) |VA | = 20 V ⇒ ro1 =

Av must be at least −10 V/V and Av = −gm (ro1 ro2 ) gm1 =

2ID 2 × 0.05 = 0.5 mA/V = VOV1 0.2

ro1 ro2 =

10 = 20 k 0.5

Q1

vi

IREF

|VA | 20 = I 0.2

= 100 k 20 = 100 k ro2 = 0.2 Total resistance at the collector of Q1 is equal to ro1 ro2 , thus rtot = 100 k 100 k = 50 k

But L VAn

ro1 =

VA1 5L = = = 100L ID1 ID1 0.05

(c) gm 1 =

ro2 =

|VAp |L |VA2 | 6L = 120L = = ID2 ID2 0.05

rπ 1 =

IC1 0.2 = 8 mA/V = VT 0.025

β 50 = 6.25 k = gm 8

Thus,

(d) Rin = rπ1 = 6.25 k

100L 120L = 20 k

Ro = ro1 ro2 = 100 k 100 k = 50 k

⇒ L = 0.367 µm

Av = −gm 1 Ro = −8 × 50 = −400 V/V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–14

7.35 For I = 10 µA:

gm =

gm =

I 10 µA = 0.4 mA/V = VT 25 mV

rπ =

β 100 = 250 k = gm 0.4 mA/V

ro =

VA 10 V = = 1 M I 10 µA

A0 = gm ro =

2ID ⇒ ID = 0.1 mA 0.2 1 W 2 ID = k n VOV 2 L W 1 × 0.22 0.1 = × 0.4 × 2 L W = 12.5 ⇒ L W = 5 µm 1=

VA 10 V = 400 V/V = VI 0.025 V

For I = 100 µA : 100 µA = 4 mA/V 25 mV 100 = 25 k rπ = 4 mA/V

gm =

ro =

2ID VOV

7.37

10 V = 100 k 100 µA

A0 = 4 mA/V × 100 k = 400 V/V

VCC

The highest instantaneous voltage allowed at the drain is that which results in a voltage equal to (VOV ) across the transistor. Thus

I vo

vOmax = 1.8 − 0.2 = +1.6 V

Q1

7.38 For the npn transistor,

vi

gm =

IC 0.1 mA = 4 mA/V = VT 0.025 V

For the NMOS transistor, gm =

For I = 1 mA: 1 mA = 40 mA/V gm = 25 mV 100 = 2.5 k rπ = 40 mA/V

2 ID 0.25 ⇒ ID = 0.5 mA 4=

10 V = 10 k 1 mA A0 = 40 mA/V × 10 k = 400 V/V

ro =

I

gm

rπ

10 µA 0.4 mA/V 250 k

7.39 gm =

ro

A0

1 M

400 V/V

100 µA 4.0 mA/V 25 k 100 k 400 V/V 1 mA

7.36 A0 =

40 mA/V 2.5 k

2VA L

10 k 400 V/V

2VA 2 × 10 × 0.4 = = = 40 V/V VOV VOV 0.2

Sedra, Smith, Carusone, Gaudet

2 ID VOV

2ID 2 × 0.05 mA = = 0.67 mA/V VOV 0.15 V

From Table K.1 (in Appendix K), VA = 5 V/µm Thus, VA = VA L = 5 × 0.5 = 2.5 V VA 2.5 = = 50 k ID 0.05 A0 = gm ro = 0.67 × 50 = 33.5 V/V ro =

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–15

From Table K.1,

ro =

μn Cox = 511 µA/V2

40 A0 = gm ro = √ V/V (ID in mA) ID

Since,

ID =

W 1 (μn Cox ) 2 L

Rin = ∞

2 VOV

BJT Cell

then, 50 = ⇒

1 × 511 × 2

W L

× 0.152

W = 8.7 × 0.5 = 4.35 µm

2ID 2 × 0.2 = 2 mA/V = VOV 0.2

From Table K.1 (Appendix K), for the 0.18-µm process we have |VA | = 5 V/µm, μn Cox = 387 µA/V2 Thus, for our NMOS transistor whose L = 0.54 µm,

gm (mA/V)

4

40

1.26

4

ro (k)

1000

100

100

10

A0 (V/V)

4000

4000

126

40

Rin (k)

25

2.5

∞

∞

7.42 Using Eq. (7.46), VA 2(μn Cox ) (WL) A0 = ID √ 5 2 × 0.4 × 10 × 0.54 × 0.54 20 = ID ⇒ ID = 0.146 mA

VA = 5 × 0.54 = 2.7 V VA 2.7 V ro = = = 13.5 k ID 0.2 mA A0 = gm ro = 2 × 13.5 = 27 V/V W 2 1 μn Cox VOV 2 L W 1 × 0.22 200 = × 387 × 2 L W ⇒ = 25.8 ⇒ W = 14 µm L

ID =

MOSFET Cell

IC = 0.1 IC = 1 ID = 0.1 ID = 1 mA mA mA mA

Bias current

W = 8.7 L

7.40 gm =

VA 10 V = ID ID

W ID L = 2 × 0.4 × 10ID = 2.83 ID √ ID = 20 µA, gm = 2.83 0.020 = 0.4 mA/V √ ID = 200 µA, gm = 2.83 0.20 = 1.26 mA/V √ gm = 2.53 2.0 = 4 mA/V ID = 2.0 mA, 7.43 gm =

ro =

2μn Cox

VA V L 5 × 0.36 1.8 = A = = ID ID ID ID

A0 = gm ro 7.41 For the BJT cell: IC IC = gm = VT 0.025 V ro =

1.8 = 90 k 0.020 A0 = 0.4 × 90 = 36 V/V ID = 20 µA ro =

VA 100 V = IC IC

A0 = gm ro =

VA 100 V = 4000 V/V = VT 0.025 V

β 100 Rin = rπ = = gm gm For the MOSFET cell: W ID = 2 × 0.2 × 40 × ID gm = 2μn Cox L = 16ID = 4 ID mA/V (ID in mA)

Sedra, Smith, Carusone, Gaudet

1.8 = 9 k 0.20 A0 = 1.26 × 9 = 11.4 V/V

ID = 200 µA ro =

1.8 = 0.9 k 2.0 A0 = 4 × 0.9 = 3.6 V/V

ID = 2.0 mA ro =

7.44 L = 0.36 µm, VOV = 0.25 V, ID = 10 µA (a) gm =

2ID 2 × 10 = = 80 µA/V VOV 0.25

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–16

ro =

VA V L = A ID ID

100 =

From Appendix K, Table K.1, VA = 5 V/µm, 5 × 0.36 = 0.18 M 10 A0 = gm ro = 80 × 0.18 = 14.4 V/V

ro =

and gm increases by a factor of

√ 10 = 3.16 to

gm = 80 × 3.16 = 253 µA/V = 0.253 mA/V and ro decreases by a factor of 10 to 0.18 M = 18 k 10 Thus, A0 becomes ro =

A0 = 0.253 × 18 = 4.55 V/V (c) If the device is redesigned with a new value of W so that it operates at VOV = 0.25 V for ID = 100 µA, gm =

2ID 0.2 mA = 0.8 mA/V = VOV 0.25 V

ro =

VA L 5 × 0.36 = 18 k = ID 0.1

A0 = gm ro = 0.8 × 18 = 14.4 V/V (d) If the redesigned device in (c) is operated √ at 10 µA, VOV decreases by a factor equal to 10 to 0.08 √ V, gm decreases by a factor of 10 to 0.253 mA/V, ro increases by a factor of 10 to 180 k, and A0 becomes 0.253 × 180 = 45.5 V/V which is an increase by a factor of

√

W = 22.2 L Thus, ⇒

W = 22.2 × 0.5 = 11.1 µm

(b) If ID is increased to 100 µA (i.e., √by a factor of 10), VOV increases by a factor of 10 = 3.16 to VOV = 0.25 × 3.16 = 0.79 V

1 W × 400 × × 0.152 2 L

10.

(e) The lowest value of A0 is obtained with the first design when operated at ID = 100 µA. The resulting A0 = 4.55 V/V. The highest value of A0 is obtained with the second design when operated at ID = 10 µA. The resulting A0 = 45.5 V/V. If in any design W/L is held constant while L is increased by a factor of 10, gm remains unchanged but ro increases by a factor of 10, resulting in A0 increasing by a factor of 10.

gm =

2 ID 2 × 0.1 = = 1.33 mA/V VOV 0.15

ro =

VA L 6 × 0.5 = 30 k = ID 0.1

7.46 Refer to Fig. 7.15(a) and Table K.1. VSG2 = |Vtp | + |VOV | = 0.4 + 0.15 = 0.55 V VG = 1.3 − VSG2 = 1.3 − 0.55 = 0.75 V W 1 ID1 = μn Cox V2 2 L 1 OV1 W 1 100 = × 511 × × 0.152 2 L 1 W ⇒ = 17.4 L 1 W 1 |VOV2 |2 ID2 = μp Cox 2 L 2 1 W 100 = × 128 × × 0.152 2 L 2 W ⇒ = 69.4 L 2 Av = −gm1 (ro1 ro2 ) gm1 =

2ID 2 × 0.1 = 1.33 mA/V = VOV 0.15

ro1 =

|VA |L 5 × 0.4 = 20 k = ID 0.1

ro2 =

|VA |L 6 × 0.4 = 24 k = ID 0.1

Av = −1.33 × (20 24) = −14.5 V/V 7.47 (a) Neglecting the dc current in the feedback network and the Early effect, we see from Fig. 1 that ID = 200 µA. Now, using W 1 2 VOV ID = μn Cox 2 L we can determine VOV :

7.45

2VA 2VA L 2 × 6 × 0.5 = 40 V/V = = VOV VOV 0.15 W 1 2 VOV ID = k n 2 L A0 =

Sedra, Smith, Carusone, Gaudet

1 2 × 2 × VOV 2 ⇒ VOV = 0.45 V

0.2 =

VGS = Vt + VOV = 0.5 + 0.45 = 0.95 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–17

= −86 V/V To obtain the maximum allowable negative signal swing at the output, we note that the MOSFET will remain in saturation as long as v DG ≥ −Vt . Thus at the limit vDG = −0.5 V, or equivalently,

200 A

3 M

ID IR

vGmax = 0.5 + vDmin

2 M

VGS + | vˆi | = 0.5 + VDS − | vˆo |

VGS

0.95 +

Figure 1

| vˆo | =

The current in the feedback network can now be found as 0.95 VGS = = 0.475 µA IR = 2 M 2 which indeed is much smaller than the 200 µA delivered by the current source. Thus, we were justified in neglecting IR above. 3+2 = 2.4 V 2 (b) Replacing the MOSFET with its hybrid-π model, we obtain the equivalent circuit shown in Fig. 2. VDS = VGS

| vˆo | = 0.5 + 2.4 − | vˆo | |Av | 0.5 + 2.4 − 0.95 1 1+ |Av |

Substituting |Av | = 86, we obtain | vˆo | = 1.93 V An approximate value of | vˆo | could have been obtained from vOmin = VOV = 0.45 V Thus, VDS − | vˆo | = VOV ⇒ | vˆo | = VDS − VOV = 2.4 − 0.45 = 1.95 V |vˆ o | = 22 mV 86 (c) To determine Rin , refer to Fig. 2, |vˆ i | =

ii

vi

RG2 3 M

RG1 2 M

Rin

vgs

vovi RG2

gmvgs

vgs vi

Figure 2

A node equation at the output node yields vo − vi vo + gm vgs + =0 ro RG2 where vgs = vi . Thus, 1 1 1 = −vi gm − + vo ro RG2 RG2 1 vo (ro RG2 ) = − gm − vi RG2 gm =

2ID 2 × 0.2 = = 0.89 mA/V VOV 0.45

VA 20 = = 100 k ID 0.2 1 vo × (100 3000) = − 0.89 − vi 3000

ro =

Sedra, Smith, Carusone, Gaudet

ro vo

ii =

vi vo − vi − RG1 RG2

vi Av v i − v i − RG1 RG2 1 (1 − Av ) = vi + RG1 RG2 vi 1 (1 − Av ) Rin = =1 + ii RG1 RG2 1 (1 + 86) + = 33.9 k =1 2 3 =

7.48 Refer to Fig. 7.16(a). Ro = 50 k = ro1 ro2 But ro1 = ro2 =

|VA | 5 = IREF IREF

Thus, 50 =

5V 1 × 2 IREF mA

⇒ IREF = 50 µA Av = −gm1 Ro

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–18

−20 = −gm1 × 50

and thus,

⇒ gm1 = 0.4 mA/V

VG = 0

But

Thus,

gm1 =

2μn Cox

W L

1 × 1 × (1 − 0.5)2 = 0.125 mA 2

ID =

ID1 1

(b)

W 0.4 = 2 × 0.4 × 0.05 L 1 W =4 ⇒ L 1 W 1 V2 ID1 = μn Cox 2 L 1 OV1

vgs

G

gmvgs D

R 2gmvgs

vi

vgs

1 2 50 = × 400 × 4 × VOV1 2 ⇒ VOV1 = 0.25 V

vo

gmvgs

If Q2 and Q3 are operated at |VOV | = 0.25 V, W 1 ID2 = ID3 = μp Cox |VOV |2 2 L 2,3 W 1 50 = × 100 × × 0.252 2 L 2,3 W ⇒ = 16 L 2,3

vgs vi Figure 2

From Fig. 2 we see that vo = vi − 2gm vgs R But vgs = vi Thus,

7.49 (a)

Av =

vo = 1 − 2gm R vi

gm =

2ID 2 × 0.125 = 0.5 mA/V = VOV 1 − 0.5

1.0 V

R

VG

Q2 ID2

Av = 1 − 2 × 0.5 × 1000 = −999 V/V (c)

VD

ID1

0

ii

Q1

R

1.0 V Figure 1

vi

vgs

2gmvgs

ro/2 vo

vi

For G and D open, from Fig. 1 that since the dc currents into the gates are zero,

Rin i i vgs vi Figure 3

VD = VG Also, since Q1 and Q2 are matched and carry equal drain currents,

For the circuit in Fig. 3 we can write at the output

ID1 = ID2 = ID

vo vo − vi + 2gm vgs + =0 ro /2 R

VSG2 = VGS1 =

1 [1 − (−1)] = 1 V 2

Sedra, Smith, Carusone, Gaudet

Substituting vgs = vi and rearranging, we obtain

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–19

1 1− vo 2gm R = −2gm 2 1 vi + R ro

Thus, io Rout = Ais isig Rout + RL

But 2gm R 1; thus ro vo −2gm R Av = vi 2

Substituting into Eq. (1) gives vo Rout RL = ( ) Ais vsig Rs Rout + RL

where |VA | 20 = 160 k ro = = ID 0.125

Thus

io

Av = −2 × 0.5(1000 80) = −74.1 V/V Rin =

1 vi vi =R = vo ii (vi − vo )/R 1− vi

RL vo

R 1000 = 13.3 k = = 1 − Av 1 + 74.1 (d)

vi Rin 13.3 = 0.4 V/V = = vsig Rin + Rsig 20 + 13.3

Gv =

vo vi vo = × vsig vsig vi

io

Rs

vsig

= 0.4 × −74.1 = −29.6 V/V

Rin

(e) Both Q1 and Q2 remain in saturation for output voltages that ensure that the minimum voltage across each transistor is equal to |VOV | = 0.5 V. Thus, the output voltage can range from −0.5 V to +0.5 V.

Figure 1 840 vo 840 × 0.976 × = vsig 20 840 + 840 = 20.5V/V

7.50 Using Eq. (7.51), Rin =

An alternative and much simpler approach of finding the voltage gain of the circuit in Fig. 1 is to find the current io from vsig (2) io = Rs + Rin

20 + 840 r o + RL = 21 k gm ro + 1 (2 × 20) + 1

Using Eq. (7.53),

and then the output voltage vo as

1 1 1+ 1+ gm ro 2 × 20 = Ais = 1 1 1 1 1+ + 1+ + gm ro gm Rs 2 × 20 2 × 20 = 0.976 A/A Using Eq. (7.55),

= 840 k

From Fig. 7.18 we see that io = Ais isig

Rout Rout + RL

Sedra, Smith, Carusone, Gaudet

(3)

Combining Eqs. (2) and (3) gives RL 840 vo = 20.5 V/V = = vsig Rs + Rin 20 + 21 7.51 Refer to Fig. P7.51.

Rout = ro + Rs + gm ro Rs = 20 + 20 + 2 × 20 × 20

vo io R L io R L = =( ) vsig isig Rs isig Rs

v o = io R L

(1)

ID = 0.2 mA VOV = 0.2 V gm =

2ID 2 × 0.2 = 2 mA/V = VOV 0.2

ro =

VA 20 = 100 k = ID 0.2

Rout = ro + Rs + gm ro Rs 500 = 100 + Rs (1 + 2 × 100) ⇒ Rs =

400 2 k 201

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–20

VBIAS = ID RS + VGS

(e)

= ID RS + Vt + VOV

RL 200 vo = = = 90.9 V vi Rin 2.2 vo = 90.9 × 0.98 = 89 V/V vsig

= 0.2 × 2 + 0.5 + 0.2 = 1.1 V

7.52 Refer to Fig. P7.52. To obtain maximum output resistance, we use the largest possible Rs consistent with ID Rs ≤ 0.2 V. Thus 0.2 V = 2 k 0.1 mA Now, for Q2 we have Rs =

gm = 2 mA/V

and

VA = 5 V

Thus, ro =

VA 5V = 50 k = ID 0.1 mA

(f) The value of vo can range from VBIAS − Vt = 1.03 − 0.8 = 0.23 V to (VDD − VOV2 ). Since ID2 = ID1 and k n = k p , then VOV2 = VOV1 . Thus the maximum value of vo is 3.3 − 0.224 = 3.076 V. Thus the peak-to-peak value of vo is 3.076 − 0.23 = 2.85 V. Correspondingly, the peak-to-peak value of vsig will be 2.85 vsig (peak to peak) = = 32 mV 89 7.54

Rout = ro + Rs + gm ro Rs

IC = 0.1 mA β = 100 VA = 20 V

= 50 + 2 + 2 × 50 × 2

IC 0.1 mA = 4 mA/V = VT 0.025 V β 100 rπ = = = 25 k gm 4

gm =

= 252 k

7.53 Refer to Fig. P7.53.

ro =

(a) ID1 = ID2 = ID3 = 100 µA Using ID1 =

vi Rin 2.2 = = = 0.98 V/V vsig Rin + Rs 2.2 + 0.05

1 2 k (W/L)1 VOV1 , we obtain 2 n

1 2 × 4 × VOV1 2 ⇒ VOV1 = 0.224 V

VA 20 = 200 k = IC 0.1

gm ro = 4 × 200 = 800 V/V Using Eq. (7.61),

0.1 =

Rin = rπ

VGS1 = Vt + VOV1 = 0.8 + 0.224 = 1.024 V VBIAS = VGS + ID1 Rs = 1.024 + 0.1 × 0.05 = 1.03 V (b) gm1 =

2 ID1 2 × 0.1 = = 0.9 mA/V VOV1 0.224

= 25

Using Eq. (7.65), Ais =

|VA | 20 = 200 k = ro = ID 0.1

=

Rin = =

r o + RL 1 + gm ro

1+

1 gm ro

1+ 1+

1 800

1 gm ro

+

1 gm (Re rπ )

1 800

+

1 4(2525)

= 0.98 A/A

Rout = ro + (Re rπ ) + (gm ro )(Re rπ ) = 200 + (2525) + 800(2525)

(d) Rout = ro + Rs + gm ro Rs

= 10.2 M

= 200 + 0.05 + 0.9 × 200 × 0.05

Sedra, Smith, Carusone, Gaudet

1+

Using Eq. (7.67),

200 + 200 = 2.2 k 1 + 0.9 × 200

= 209 k

200 + 1000 = 251.5 800 + 1

= 1.4 k

All transistors are operating at ID = 0.1 mA and have |VA | = 20 V. Thus all have equal values for ro :

(c) For Q2 , RL = ro2 = 200 k,

r o + RL gm ro + 1

vo io R L RL io = =( )( ) vsig isig Re Re isig

Microelectronic Circuits, 8th International Edition

(1)

© Oxford University Press 2021

Chapter 7–21

From Fig. 7.25(b),

gm =

Rout Rout + RL Rout = Ais Rout + RL

100 Re Rout =1+ ro 5 + Re

io = Ais isig ⇒

io isig

IC 0.5 mA = = 20 mA/V, thus VT 0.025 V (1)

where Re is in kilohms.

Substituting in Eq. (1), we obtain

(a) For Rout = 5 ro , Eq. (1) gives Re = 0.208 k = 208 (b) For Rout = 10 ro , Eq. (1) gives Re = 0.495 k 500 .

vo RL Rout = ( )Ais ( ) vsig Re Rout + RL

(c) For Rout = 50 ro , Eq. (1) gives Re = 4.8 k. From Eq. (1) we see that the maximum value of Rout /ro is obtained with Re = ∞ and its value is 101, which is (β + 1).

1000 10.2 = × 0.98 × 25 10.2 + 1 = 35.7 V/V

7.57 7.55 Refer to Fig. P7.55. 5 − 0.7 I = IC = αIE = 0.99 × 1 mA 4.3 VA 100 V = 100 k = ro = IC 1 mA

(a) Refer to Fig. 7.18(a) and note that the current io that flows through RL is equal to the current that enters the source of the transistor, thus io = isig

Rout = ro + (RE rπ )(1 + gm ro )

and the current gain is

where gm =

IC 1 mA = 40 mA/V = VT 0.025 V

gm ro = 40 × 100 = 4000 rπ =

Rs Rs + Rin

io Rs = isig Rs + Rin (b)

β 100 = = 2.5 k gm 40

20 + 20 r o + RL = = 0.976 k gm ro + 1 2 × 20 + 1

RE = 4.3 k

Rin =

Thus,

io 20 = 0.95 A/A = isig 20 + 0.976

Rout = 100 + (4.3 2.5) × 4001 = 6.4 M

(c) If RL is doubled; that is, increases to 40 k,

For VC = 10 V

Rin =

10 V = 1.6 µA I = 6.4 M A very small change indeed!

and io 20 = 0.93 A/A = isig 20 + 1.46

7.56 Refer to Fig. 7.27. Rout = ro + (Re rπ )(1 + gm ro ) ro + (Re rπ )(gm ro ) Rout = 1 + gm (Re rπ ) ro

Thus, a change in RL of 100% results in a change in the current gain of only 2%. (d) For RL = 200 k,

=1+

gm rπ Re r π + Re

Rin =

=1+

βRe (β/gm ) + Re

and

For our case β = 100,

Sedra, Smith, Carusone, Gaudet

20 + 40 = 1.46 k 2 × 20 + 1

20 + 200 = 5.37 k 2 × 20 + 1

io 20 = 0.79 A/A = isig 20 + 5.37

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–22

Thus, a change of +900% in RL results in only a change of −0.16 or −17% in the current gain.

Using Eq. (7.65),

(e) The fact that the current gain io /isig of the CG amplifier changes very little in the face of large changes in RL indicates that it functions as a current buffer.

1+

Ais = =

1+ 1 gm ro

1+ 1+

1 1000

1 gm ro

+

1 gm (Re rπ )

1 1000

+

1 5(2020)

= 0.98 A/A 7.58 Using Eq. (7.67), Rs = 25 k RL = 1 M Rout = 1 M Rout (gm ro )Rs

= 200 + (2020) + 1000(2020)

Thus,

= 10.2 M

A0 = gm ro

Rout 1 M = = 40 V/V Rs 25 k

2VA VOV

10.2 = 0.96 A/A 10.2 + 0.2

(c) If RL is doubled, it becomes 400 k and the current gain becomes

Thus,

10.2 io = 0.94 A/A = 0.98× isig 10.2 + 0.4

2VA 40 = 0.15 ⇒ VA = 3 V

Thus, in the face of a 100% change in RL , the current gain decreases by 0.02 which is approximately 2.1%:

Since VA = VA L

(d) If RL is increased by a factor of 10 or by 900%, from 200 k to 2 M, the current gain becomes

3=5×L ⇒ L = 0.6 µm

10.2 io = 0.98 × isig 10.2 + 2 = 0.82 A/A

7.59 (a) Refer to Fig. 7.25(b): At the output we see that Rout ) io = Ais isig ( Rout + RL Thus, io Rout = Ais isig Rout + RL

which is a decrease of 0.15 or 15%. (e) A current buffer works to keep the current gain as constant as possible in the face of changes in the load resistance RL . From the results of (c) and (d) above we see that the CB amplifier fulfills this requirement admirably! 7.60 Using Eq. (7.61), Rin = rπ

(b) gm = 5 mA/V ro = 200 k β = 100 Re = 20 k RL = 200 k gm ro = 1000 V/V rπ =

io Rout = Ais isig Rout + RL = 0.98 ×

But, A0 =

Rout = ro + (Re rπ ) + (gm ro )(Re rπ )

β 100 = 20 k = gm 5

Sedra, Smith, Carusone, Gaudet

r o + RL gm ro + 1

= (β + 1)re

r o + RL gm ro + 1

To obtain Rin = 2re , we arrange that r o + RL 2re gm ro + 1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–23

Usually gm ro 1, thus we need

7.62 Using Eq. (7.75), Avo = −A20

r o + RL 2re gm ro

where

that is, RL 1 + 2re gm gm ro Since

(1)

A0 =

2VA 2×3 = = 40 VOV 0.15

Thus,

1 re then Eq. (1) is satisfied when gm

Avo = −1600 V/V

RL = r o 7.63 Refer to Fig. 7.30 Ro = gm3 ro3 ro4

7.61 Refer to the circuit in Fig. P7.55. I=α

2.3 3 − VBE Re Re

For identical transistors, Ro = (gm ro )ro

Thus, Re =

2.3 k I(mA)

(1)

Ro = (gm ro ) (Re rπ )

(2)

gm ro =

Thus,

2 × 42 = 160 0.2 Ro = 1.6 M

1 +

=

1 rπ

I 2.3

1 + βVI T

To obtain the W/L values, W 1 |VOV |2 I = ID = μp Cox 2 L 3,4 1 W 100 = × 100 × × 0.22 2 L 3,4 W ⇒ = 50 L 3,4

Substituting in Eq. (2), 1 I

Q.E.D.

0.1 × Ro =

β β βVT = = gm I/VT I 1 Re

2|VA |2 VOV

(a) I = 0.1 mA

VA VT

Re rπ =

Ro =

|VA | 2|VA | × |VOV | I

IRo =

But,

rπ =

=

VA /VT + βV1 T

1 2.3

That is,

(b) I = 0.5 mA VA 1 R o = ( ) ( VT I + 2.3

1 β

)

Substituting VT = 0.025 V and the given numerical values VA = 25 V and β = 100, we get Ro

2 × 42 = 160 0.2 Ro = 320 k W 1 I = μp Cox |VOV |2 2 L 3,4 W 1 500 = × 100 × |VOV |2 2 L 3,4 W ⇒ = 250 L 3,4

0.5Ro =

1.2 M I(mA)

(3)

We can now use Eqs. (1) and (3) to obtain Re and Ro for each of the given values of I, obtaining I (mA)

0.1

0.2

0.5

1.0

Re (k)

23

11.5

4.6

2.3

7.64 Refer to Fig. 7.28(a).

Ro (M)

12

6

2.4

1.2

gm1 =

Sedra, Smith, Carusone, Gaudet

2ID 2I = VOV VOV

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–24

2I 0.15 ⇒ I = 0.225 mA

3=

VG4 = VDD − VSG4 = 1.3 − 0.55 = 0.75 V

For identical transistors,

To obtain the largest possible signal swing at the output, we maximize the allowable positive signal swing by setting VD4 at its highest possible value of VDD − |VOV | = 1.3 − 0.15 = 1.15 V. This will be obtained by selecting VG3 as follows:

Ro = (gm ro )ro = 150 =

2VA2 2VA VA = VOV I VOV I

VG3 = VD4 − VSG3

2VA2 0.15 × 0.225

Since

⇒ VA = 1.6 V

VSG3 = VSG4 = 0.55 V

VA = VA L

VG3 = 1.15 − 0.55 = 0.6 V

1.6 VA L= = = 0.32 µm VA 5

The highest allowable voltage at the output will be

To obtain W/L, we use W 1 2 VOV ID = I = μn Cox 2 L 1 W × 0.152 225 = × 511 × 2 L

vD3max = VG3 + |Vtp | = 0.6 + 0.4 = 1 V

W = 39.1 L To obtain maximum negative signal swing at the output, we select VG so that the voltage at the drain of Q1 is the minimum permitted, which is equal to VOV (i.e., 0.15 V). Thus ⇒

VG = 0.15 + VGS2 = 0.15 + VOV2 + Vt

Since both Q3 and Q4 carry the same current I = 100 µA and are operated at the same overdrive voltage, |VOV | = 0.15 V, their W/L ratios will be the same and can be found from W 1 |VOV |2 ID = μp Cox 2 L 3,4 1 W 100 = × 128 × × 0.152 2 L 3,4 W ⇒ = 69.4 L 3,4 To obtain Ro , we first find gm and ro of both devices,

= 0.15 + 0.15 + 0.4 = 0.7 V The minimum permitted output voltage is VG − Vt = 0.7 − 0.4 = 0.3 V or 2VOV . 7.65

gm3,4 =

2 × 0.1 2ID = = 1.33 mA/V |VOV | 0.15

ro3,4 =

|VA | 3 = = 30 k ID 0.1

Ro = (gm3 ro3 )ro4

VG4

VDD 1.3 V

= 1.33 × 30 × 30 = 1.2 M

Q4

7.66 Refer to Fig. 7.31. Av = −gm1 Ro

VD4 VG3

−280 = −1 × Ro ⇒ Ro = 280 k gm1 =

Q3

2ID 2I 1 = ⇒ I = gm1 VOV VOV VOV 2

VSG4 = |Vtp | + |VOV |

1 × 1 × 0.25 = 0.125 mA 2 All four transistors are operated at the same value of ID and the same value of |VOV |. Also all have the same channel length and |VA |; thus all ro values are equal. Thus

= 0.4 + 0.15 = 0.55 V

Ron = Rop = 2Ro = 2 × 280 = 560 k

Thus,

560 = (gm ro )ro

Ro I 100 A

Sedra, Smith, Carusone, Gaudet

=

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–25

=

2|VA | |VA | |VOV | I

=

2|VA |2 0.25 × 0.125

−100 = −1.52(Ro RL ) ⇒ Ro RL = 65.8 k 1 1 1 + = Ro RL 65.8

⇒ VA = 2.96 V

1 1 1 1 = − = RL 65.8 141 123

VA 2.96 L= = = 0.6 µm VA 5

⇒ RL = 123 k

For each of the NMOS devices, W 1 V2 ID = μn Cox 2 L 1,2 OV 1 W 125 = × 400 × × 0.252 2 L 1,2 W ⇒ = 10 L 1,2

Rin2 = =

ro2 + RL 1 + gm2 ro2

9 + 123 = 9 k 1 + 1.52 × 9

Rd1 = ro1 Rin2 = 9 9 = 4.5 k A1 = −gm1 Rd1 = −1.52 × 4.5 = −6.84 V/V

For each of the PMOS transistors, W 1 |VOV |2 ID = μp Cox 2 L 3,4 W 1 125 = × 100 × × 0.252 2 L 3,4 W ⇒ = 40 L 3,4

7.68

VCC

7.67

VB4

Q4

VB3

Q3

vo

I 0.2 mA

Ro

Ro

VG2

Q2 5.4/0.36

RL Ro = (gm3 ro3 )(ro4 rπ 3 )

Rin2 ro1

vi

Q1 5.4/0.36

I = 0.2 mA I 0.2 = 8 mA/V = gm3 = VT 0.025 5 VA = = 25 k I 0.2 β 50 = 6.25 k = = gm3 8

ro3 = ro4 = rπ 3 gm1 = gm2 =

2μn Cox

=

2 × 0.387 ×

ro1 = ro2 =

W ID L

5.4 × 0.2 = 1.52 mA/V 0.36

VA V L 5 × 0.36 = 9 k = A = ID ID 0.2

Ro = ro1 + ro2 + gm2 ro2 ro1 = 9 + 9 + 1.52 × 9 × 9

Ro = (8 × 25)(25 6.25) = 1 M 7.69 Refer to Fig. 7.34. When all transistors have equal β and ro , and, since they conduct equal currents, they have equal gm , then Ron = Rop = gm ro (ro rπ )

= 141 k

Ro = Ron Rop =

Av = −gm1 (Ro RL )

Av = −gm Ro

Sedra, Smith, Carusone, Gaudet

1 (gm ro )(ro rπ ) 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–26

This table belongs to Problem 7.71. L = Lmin = 0.18 µm IRo = 8.1 V

L = 2Lmin = 0.36 µm IRo = 32.4 V

L = 3Lmin = 0.54 µm IRo = 72.9 V

gm Ro Av 2WL gm Ro Av 2WL gm Ro Avo 2WL (mA/V) (k) (V/V) (µm2 ) (mA/V) (k) (V/V) (µm2 ) (mA/V) (k) (V/V) (µm2 ) I = 0.01 mA W/L = n

0.1

810 −40.5 0.065 n

I = 0.1 mA W/L = 10 n

1.0

81

I = 1.0 mA W/L = 100 n

10.0

8.1 −40.5

−40.5 0.65n 6.5n

1 = − (gm ro )gm (ro rπ ) 2 1 gm ro rπ = − (gm ro ) 2 rπ + ro

3,240 −162 0.26n

0.1

7,290 −364.5 0.58n

1.0

324

−162

2.6n

1.0

729 −364.5 5.8n

10.0

32.4 −162

26n

10.0

72.9 −364.5

58n

Substituting |VA | = |VA | L IRo =

1 1 = − (gm ro ) 1 1 2 + gm ro gm rπ Substituting gm ro =

0.1

2|VA |2 2 L |VOV |

Q.E.D.

Now, for

|VA | and gm rπ = β, VT

L = 0.18 µm,

IRo =

2 × 52 × 0.182 0.2

IRo =

2 × 52 × 0.362 0.2

= 8.1 V

|VA |/VT 1 Av = − 2 (VT /|VA |) + (1/β)

L = 0.36 µm,

For |VA | = 5 V and β = 50 we obtain

= 32.4 V

5/0.025 1 = −4000 V/V Av = − 2 (0.025/5) + (1/50)

2 × 52 ×0.542 = 72.9 V 0.2 To fill out the table we use 2I 2I 2ID = = = 10I gm = |VOV | |VOV | 0.2

7.70 50 = gm2 ro2 = A02 =

2VA VOV

L = 0.54 µm,

IRo =

Av = gm (Ro /2)

VA = 50 × VOV /2

(a) The price paid is the increase in circuit area.

= 25 × 0.2 = 5 V

(b) As I is increased, gm increases and hence the current-driving capability of the amplifier, and as we will see later, its bandwidth.

VA = VA L 5 = 5 × L ⇒ L = 1 µm 7.71

(c) The circuit with the largest area (58n) as compared to the circuit with the smallest area (0.065n): Av is 364.5/40.5 = 9 times larger; gm is 100 times larger, but Ro is 11.1 times lower.

Refer to Fig. 7.30. Ro = (gm3 ro3 )ro4 For identical transistors, Ro = (gm ro )ro

7.72 VSG4 = |Vtp | + |VOV | = 0.8 + 0.2 = 1 V

|VA | 2|VA | × = |VOV | I

Thus,

Thus,

VG4 = VDD − VSG4 = 3.3 − 1 = 2.3 V

IRo =

2|VA |2 |VOV |

Sedra, Smith, Carusone, Gaudet

To obtain the largest possible signal swing at the output, we maximize the allowable positive

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–27

VDD 3.3 V

ix Ro

VG4

Q4

Q2 VD4 ix VG3

vx

vy ro1

Q3

Q1

Ro I 100 A

signal swing by setting VD4 at its highest possible value of VDD − |VOV | = 3.3 − 0.2 = 3.1 V. This will be obtained by selecting VGS as follows: VG3 = VD4 − VSG3

=

ro1 ro1 + ro2 + gm2 ro2 ro1

For gm2 ro2 1 and gm2 ro1 1,

Since

vy 1 vx gm2 ro2

VSG3 = VSG4 = 1 V VG3 = 3.1 − 1 = 2.1 V the highest allowable voltage at the output will be

7.74

vD3max = VG3 + |Vtp |

VDD 1.3 V

= 2.1 + 0.8 = 2.9 V Since both Q3 and Q4 carry the same current I = 100 µA and are operated at the same overdrive voltage, |VOV | = 0.2 V, their W/L ratios will be the same and can be found from W 1 |VOV |2 ID = μp Cox 2 L 3,4 W 1 100 = × 60 × × 0.22 2 L 3,4 W ⇒ = 83.3 L 3,4 To obtain Ro , we first find gm and ro of both devices, gm3,4 =

2 × 0.1 2ID = = 1 mA/V |VOV | 0.2

ro3,4 =

|VA | 5 = 50 k = ID 0.1

0.7 V

Q4

0.4 V

Q3

vo 0.9 V

Q2

VI vi

Q1

0.6 V

Ro = (gm3 ro3 )ro4 = 1 × 50 × 50 = 2.5 M 7.73 While vx appears across Ro , vy appears across ro1 , Thus, ro1 vy = vx Ro

Sedra, Smith, Carusone, Gaudet

Since all four transistors have equal transconductance parameters, k, and all four have the same bias current, their overdrive voltages will be equal. We can obtain |VOV | by considering either Q1 or Q4 . For Q1 ,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–28

VGS = VI = 0.6 V = Vt + VOV

vi

Thus, VOV = 0.6 − 0.4 = 0.2 V

5 mV 0

Similarly, for Q4 , VSG = VDD − VG4 = 1.3 − 0.7 = 0.6 V Thus, |VOV | = VSG − |Vt | = 0.6 − 0.4 = 0.2 V The maximum allowable voltage at the output is

t

v1 50 mV 0

t

vOmax = VG3 + |Vt3 | = 0.4 + 0.4 = 0.8 V

v2

The minimum allowable voltage at the output is

1V

vOmin = VG2 − |Vt2 | = 0.9 − 0.4 = 0.5 V

0

t

Thus the output voltage can range from 0.5 V to 0.8 V.

v3 7.75 Refer to Fig. P7.75. (a) R1 = ro1 = ro

50 mV

R2 (gm ro )ro R3 =

R2 + r o gm ro2 + ro = ro gm ro gm ro

0

t

(b) i1 = gm vi i2 = i 1

R3 ro 1 = gm vi = gm vi R3 + r o ro + ro 2

i3 = i 1 − i 2 =

1 gm vi 2

Thus, v1 is a 50-mV peak sine wave that is 180◦ out of phase with vi .

1 gm vi 2 1 i5 = i4 = gm vi 2 i6 = 0 (because vsg4 = 0) i4 = i 3 =

i7 = i5 =

1 vˆ 2 = − × 202 × 5 = −1 V 2 Thus, v2 is a 1-V peak sine wave, 180◦ out of phase relative to vi .

1 gm vi 2

1 (c) v1 = −i2 ro = − (gm ro )vi 2 1 v2 = −i4 R2 = − gm (gm ro )ro vi 2 1 = − (gm ro )2 vi 2 1 1 v3 = −i5 R1 = − gm vi ro = − (gm ro )vi 2 2 (d) vi is a 5-mV peak sine wave. 1 vˆ 1 = − × 20 × vi = −10 × 5 = −50 mV 2

Sedra, Smith, Carusone, Gaudet

1 vˆ 3 = − × 20 × 5 = −50 mV 2 Thus, v3 is a 50-mV peak sine wave, 180◦ out of phase relative to vi .

7.76 Refer to Fig. P7.76. Ro = (gm3 ro3 )(gm2 ro2 )ro1 Avo = −gm1 Ro = −(gm1 ro1 )(gm2 ro2 )(gm3 ro3 )

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–29

For identical transistors with equal gm and ro ,

= −gm (Ro βro )

Ro = (gm ro )2 ro Avo = −(gm ro )3 For gm = 2 mA/V and ro = 15 k, Ro = −(30)2 × 15 = 13.5 M Avo = −(30)3 = −2.7 × 104 V/V

7.77 When Eq. (7.86) is applied to the case of identical pnp transistors, it becomes Ro = (gm ro )(ro rπ ) I VT

ro =

|VA | I

β gm

Thus, IRo = = =

=

|VA | |VA | rπ VT r o + r π |VA | |VA | VT 1 + r o rπ |VA | 1 1 + gm ro β

I 0.1 = 4 mA/V = VT 0.025

1.8 VA = = 18 k I 0.1 β 125 = 31.25 k = = gm2 4

Gm = gm1 = 1.41 mA/V Ro = gm2 ro2 (ro1 rπ 2 ) = 4 × 18 × (18 31.25) = 822.3 k Q.E.D.

5 = 200 V (0.025/5) + (1/50) 0.1

Av = −8(24.7 25) × 103

rπ 2

For |VA | = 5 V and β = 50 we obtain

I (mA)

= 24, 691 k

ro2 =

|VA | (VT /|VA |) + (1/β)

IRo =

100 |VA | = = 500 k I 0.2 Ro = 8 × 500 × (500 6.25)

ro =

gm2 =

1 |VA | = 1 1 VT 1 + |VA | β VT =

β 50 = 6.25 k = gm 8

= 1.41 mA/V VA 1.8 = 18 k = ro1 = ID 0.1

|VA | Iro rπ VT r o + r π

|VA | VT

rπ =

7.79 (a) Refer to the circuit in Fig. P7.79(a). W ID gm1 = 2μn Cox L √ = 2 × 0.4 × 25 × 0.1

gm ro = |VA |/VT rπ =

For |VA | = 100 V, β = 50, and I = 0.2 mA we obtain I 0.2 = gm = = 8 mA/V VT 0.025

= −99.4 × 103 −105 V/V

Now, gm =

Av = −gm (Ro RL )

0.5

1

Ro (k) 2000 400 200

Av o = −Gm Ro = −1.41 × 822.3 = −1159 V/V (b) Refer to circuit in Fig. P7.79(b). √ gm1 = gm2 = 2 × 0.4 × 25 × 0.1 = 1.41 mA/V 1.8 VA = = 18 k ro1 = ro2 = I 0.1 Gm = gm1 = 1.41 mA/V Ro = gm2 ro2 ro1

7.78 The output resistance of the cascode amplifier (excluding the load) is Ro = gm ro (ro rπ ) Thus,

Sedra, Smith, Carusone, Gaudet

= 1.41 × 18 × 18 = 457 k Av o = −Gm Ro = −1.41 × 457 = −644 V/V We observe that the circuit with a BTT cascode transistor provides higher gain.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–30

7.80 The dc level shift provided by a source follower is equal to its VGS . Thus, to obtain a dc level shift of 0.9 V, we write

ID = IREF = 50 µA and thus all have the same VOV which can be determined as follows:

VGS = 0.9 V = Vt + VOV

ID =

⇒ VOV = 0.9 − 0.6 = 0.3 V To obtain the required bias current, we use W 1 2 VOV I = ID = μn Cox 2 L 20 1 × 0.2 × × 0.32 2 0.5 I = 0.36 mA = 360 µA

⇒ VOV = 0.14 V and VGS = Vt +VOV = 0.4+0.14 = 0.54 V

=

gm =

1 W 2 (μn Cox )( )VOV 2 L 3.9 1 2 × VOV 50 = × 511 × 2 0.39

2ID 2 × 0.36 = = 2.4 mA/V VOV 0.3

The minimum voltage required at the output is VOmin = 2VGS − Vt = Vt + 2 VOV

gmb = χ gm = 0.2 × 2.4 = 0.48 mA/V VA V L 20 × 0.5 = 27.8 k = A = ro = ID ID 0.36 To determine Av o , we note [refer to Fig. 7.33(b)] that the total effective resistance between the MOSFET source terminal and ground is 1 . Denoting this resistance R, we ro1 ro3 gmb have 1 R = ro ro gmb

= 1.81 k

=

R R+

1 2.4 1 Ro = R// gm 1.81 +

2ID 2 × 0.05 = = 0.7 mA/V VOV 0.14

ro =

VA V L 5 × 0.39 = 39 k = A = ID ID 0.05

Thus

Ro = β3 ro3 /2

ro3 = = 0.81 V/V

= 1.81 k//

100 V VA = = 100 k I 1 mA

Thus, 100 × 100 = 5 M 2 VO 10 V IO = = 2 µA = Ro 5 M Ro =

1 2.4 mA/V

IO 2 µA = 0.002 = IO 1 mA

= 0.339 k When a load resistance of 2 k is connected to the output, the total resistance between the output node and ground becomes R RL = 1.81 2 = 0.95 k. Thus, the voltage gain becomes

7.81

gm =

where

1 gm

1.81

Av =

where

7.82 Refer to Eq. (7.98),

Thus, the open-circuit voltage gain is Av o =

Ro = gm3 ro3 ro2 = (gm ro )ro

Ro = (0.7×39)×39 = 1.1 M

1 0.48

= 27.8 27.8

= 0.4 + 2 × 0.14 = 0.68 V

0.95 1 0.95 + 2.4

7.83 For each of the three transistors, ID = IREF =

1 W 2 (μn Cox )( )VOV 2 L

Thus,

= 0.7 V/V

100 =

Refer to Fig. 7.36. Each of the four transistors is operating at the same drain current,

Sedra, Smith, Carusone, Gaudet

or 0.2%

1 2 × 511 × 16 × VOV 2

⇒ VOV = 0.16 V VGS = Vt + VOV = 0.4 + 0.16 = 0.56 V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–31

The voltage at G1,2 = VGS = 0.56 V

×(2.9 250)

The voltage at G3 = 2VGS = 1.12 V

= 8.2 M A 5-V change in VO gives rise to

VOmin = VG3 − Vt = 1.12 − 0.4 = 0.72 V Ro = (gm3 ro3 )ro2 = (gm ro )ro

IO =

5 = 0.6 µA 8.2

7.85 Refer to Fig. 7.36.

where gm =

2ID 2 × 0.01 = = 1.25 mA/V VOV 0.16

IO = IREF

ro =

VA V L 5 × 4 × 0.13 = 26 k = A = ID ID 0.1

= 20

Ro = (1.25 × 26) × 26 = 845 k

ID1

(W/L)2 (W/L)1

27/0.54 = 200 µA 2.7/0.54 W 1 = μn Cox V2 2 L 1 OV1

1 2.7 2 × 387 × × VOV1 2 0.54 ⇒ VOV1 = 0.14 V 20 =

7.84

VG2 = VGS1 = Vt + VOV1 = 0.5 + 0.14 = 0.64 V

IREF 200 A

Rout IO 20 A

Q1

VBE2 VBE1

VOV4 = VOV1 Thus, VGS4 = VGS1 = 0.64 V

Q2

VG3 = 0.64 + 0.64 = 1.28 V

RE

The lowest voltage at the output while Q3 remains in saturation is VOmin = VG3 − Vt3

= 1.28 − 0.5 = 0.78 V Transistors Q2 and Q3 are operating at the same VOV = 0.14 V.

(a) Assuming β is high so that we can neglect base currents, IREF IO RE = VT ln IO Substituting IO = 20 µA and IREF = 200 µA results in 200 0.02 RE = 0.025 ln 20

Thus, gm2 = gm3 = ro2 = ro3 =

2ID2,3 2 × 0.2 = 2.9 mA/V = VOV2,3 0.14

VA V L 5 × 0.54 = 13.5 k = A = ID ID 0.2

Ro = gm3 ro3 ro2 = 2.9 × 13.5 × 13.5 = 529 k

⇒ RE = 2.88 k

7.86

(b) Rout = (RE rπ 2 ) + ro2 + gm2 ro2 (RE rπ 2 )

From the figure (see next page) we see that

where

Ro = (gm3 ro3 )(gm2 ro2 )ro1

0.02 = 0.8 mA/V gm2 = 0.025 VA 50 = 2500 k = ro2 = IO 0.02

7.87 (a) IO1 = IO2 =

rπ 2 =

β 200 = = 250 k gm 0.8

(b)

Rout = (2.9 250) + 2500 + 0.8 × 2500

Sedra, Smith, Carusone, Gaudet

1 IREF 2 2 1+ 2 β

The figure shows the required circuit. Observe that the output transistor is split into three

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–32

2.5 V

This figure belongs to Problem 7.86.

(gm3ro3) (gm2ro2)ro1

IREF

1.8 V Q2

Q1

IO Q6

Q3

1.1 V

(gm2 ro2)ro1 Q5

Q3

Q2

R

ro1 Q4

IO

0.1 mA

VO

Q1

2.5 V transistors having base–emitter junctions with area ratio 1:2:4. Thus 0.1 0.1 = = 0.0999 mA IO1 = 2 2 1+ 2 1+ 2 β 50

VOmax is limited by Q3 saturating. Thus VOmax = VE3 − VECsat = 1.8 − 0.3 = 1.5 V

0.2 = 0.1998 mA 2 1+ 2 50 0.4 = = 0.3997 mA 2 1+ 2 50

7.89 Replacing each of the transistors in the Wilson mirror of Fig. 7.37(a) with its T model while neglecting ro results in the circuit shown below.

IO2 =

IO4

ix IO1 IREF 0.7 mA

IO2

2

IO3

ie3 Q3

Q5

Q4 1

2

aie2

4 vx

Q1

Q1

VEE

re3 1

Rin

ie2 re2

ie1 re1

Note that the diode-connected transistor Q1 reduces to a resistance re1 . To determine Rin , we have applied a test voltage vx . In the following we analyze the circuit to find ix and hence Rin , as vx Rin ≡ ix Note that all three transistors are operating at equal emitter currents, approximately equal to IREF . Thus

7.88 R=

aie3

(1 a) ie3

1.1 + 2.5 = 36 k 0.1

Sedra, Smith, Carusone, Gaudet

re1 = re2 = re3 =

VT IREF

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–33

Analysis of the circuit proceeds as follows. Since re1 = re2 , we obtain ie2 = ie1

i3

ix

(1)

2

Node equation at node 1:

i3

ie3 + αie2 = ie1 + ie2

i2

1 gm3

i2 1 gm2

i1

Using Eq. (1) yields ie3 = (2 − α)ie1

(2)

vx

1

Node equation at node 2: ix = αie2 + (1 − α)ie3

1 gm1

Using Eqs. (1) and (2) yields ix = ie1 [α + (1 − α)(2 − α)ie3 ] ix = ie1 [(2 − 2α + α 2 )]

Rin (3)

Finally, vx can be expressed as the sum of the voltages across re3 and re1 ,

A node equation at node 2 gives ix + i 3 = i 2 + i 3

vx = ie3 re + ie1 re

Thus

Using Eq. (2) yields vx = ie1 re (3 − α)

(4)

Dividing Eq. (4) by Eq. (3) yields Rin =

3−α vx = re ix 2 − 2α + α 2 VT IREF

The voltage vx can be expressed as the sum of the voltages across 1/gm3 and 1/gm1 : vx = (i3 /gm3 ) + (i1 /gm1 ) Substituting i3 = i2 and i1 = i2 , gm1 = gm3 = gm , and

For α 1, Rin = 2re = 2

ix = i 2

Q.E.D.

vx = 2 i2 /gm

Thus, for IREF = 0.2 mA,

But i2 = ix ; thus

Rin = 250

vx = 2 ix /gm

7.90 Replacing each of the three transistors in the Wilson current mirror in Fig. 7.38(a) with its T model results in the circuit in the figure. Here, we have applied a test voltage vx to determine Rin , vx Rin ≡ ix Since all three transistors are identical and are operating at the same ID , gm1 = gm2 = gm3 Now from the figure we see that i1 = i 2 and i2 + i 3 = i 2 + i 1 Thus i3 = i 1 = i 2

Sedra, Smith, Carusone, Gaudet

and thus 2 Rin = gm

Q.E.D.

7.91 Refer to circuit in Fig. 7.38(a). (a) Each of the three transistors is operating at ID = IREF . Thus W 1 2 VOV IREF = μn Cox 2 L 1 2 × 400 × 10 × VOV 2 ⇒ VOV = 0.3 V 180 =

VGS = Vtn + VOV = 0.5 + 0.3 = 0.8 V (b) Q1 is operating at VDS = VGS = 0.8 V Q2 is operating at VDS = 2VGS = 1.6 V Thus, IREF − IO =

VDS ro

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 7–34

This figure belongs to Problem 7.92.

where ro =

VA 18 = 100 k = IREF 0.18

IREF

0.8 = 0.008 mA = 8 µA 100 IO = 180 − 8 = 172 µA IREF − IO =

IO

(c) Refer to Fig. 8.38(c). Since Q1 and Q2 are now operating at equal VDS , we estimate IO = IREF = 180 µA.

Q4

Q3

Q3

Q1

(d) The minimum allowable VO is the value at which Q3 leaves the saturation region: VOmin = VG3 − Vt = VGS3 + VGS1 − Vt = 0.8 + 0.8 − 0.5 = 1.1 V

VEE

(e) Diode-connected transistor Q4 has an incremental resistance 1/gm4 . Reference to Fig. 7.38(b) indicates that the incremental resistance of Q4 would appear in series with the gate of Q3 and thus carries zero current. Thus, including Q4 has no effect on the value of Ro , which can be found from Eq. (7.99):

Figure 1

From the circuit we see that the voltage across R is VBE = 0.585 V, thus IO R = VBE

Thus,

0.585 = 58.5 k 0.01 0.01 = 0.4 mA/V (b) gm3 = 0.025 40 = 4000 k ro3 = 0.01 β 100 = 250 k = rπ 3 = gm3 0.4

Ro = 1.2 × 100 × 100 = 12 M

Rout = (R rπ 3 ) + ro3 + gm3 ro3 (R rπ 3 )

(f) For VO = 1 V, we obtain

= (58.5 250) + 4000 + 0.4 × 4000 × (58.5 250)

Ro = gm3 ro3 ro2

R=

where gm3 =

2ID 2 × 0.18 = 1.2 mA/V = VOV 0.3

ro2 = ro3 =

IO =

VA 18 = 100 k = IREF 0.18

VO 1V = 0.08 µA = Ro 12 M

IO = 0.04% IO 7.92 Fig. 1 shows the circuit of the modified BJT Wilson mirror. 7.93 (a) Refer to the circuit in Fig. P7.93. Neglecting the base currents, we see that all three transistors are operating at IC = 10 µA, and thus 1 mA VBE1 = VBE2 = VBE3 = 0.7 − 0.025 ln 10 µA = 0.585 V

Sedra, Smith, Carusone, Gaudet

= 79.9 M 7.94 Refer to the circuit in Fig. P7.94. Since Q1 and Q2 are matched and conducting equal currents I, their VGS values will be equal. Thus from the loop Q1 , Q6 , R, and Q2 , we see that IR = VEB6 I = VT ln IS

Q.E.D.

Now to obtain I = 0.2 mA, we write 1 mA 0.2R = 0.7 − 0.025 ln 0.2 mA ⇒ R = 3.3 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 8–1

Chapter 8

1 k 2 n

ID =

W L

(VOV )2

Solutions to Exercises within the Chapter

VDD

Ex: 8.1 Referring to Fig. 8.3,

RD 5 k

If RD is doubled to 5 k, I VD1 = VD2 = VDD − RD 2 0.4 mA = 1.5 − (5 k) = 0.5 V 2 VCMmax = Vt + VD = 0.5 + 0.5 = +1.0 V

RD 5 k

ID

ID

I 0.8 mA VSS

Since the currents ID1 , and ID2 are still 0.2 mA each, Thus,

VGS = 0.82 V So, VCMmin = VSS + VCS + VGS

VOV

= −1.5 V + 0.4 V + 0.82 V = −0.28 V So, the common-mode range is −0.28 V to +1.0 V Ex: 8.2 (a) The value of v id√that causes Q1 to conduct the entire current is 2 VOV √ → 2 × 0.316 = 0.45 V then, VD1 = VDD − I × RD

2ID = W kn L

2 (0.4 mA) 0.2 mA/V2 (100)

= 0.2 V gm =

0.4 mA × 2 ID = = 4 mA/V VOV /2 0.2 V

ro =

VA 20 V = 50 k = ID 0.4 mA

Ad = gm (RD ro ) Ad = (4 mA/V) (5 k 50 k) = 18.2 V/V

= 1.5 − 0.4 × 2.5 = 0.5 V VD2 = VDD = +1.5 V

Ex: 8.5 With I = 200 µA, for all transistors,

v O = VD2 − VD1 = +1 V (b) For Q2 to conduct the entire current: √ v id = − 2 VOV = −0.45 V then, VD1 = VDD = +1.5 V VD2 = 1.5 − 0.4 × 2.5 = 0.5 V v O = VD2 − VD1 = −1 V (c) Thus the differential output range is +1 V to −1 V. Ex: 8.3 Refer to answer table for Exercise 8.3 where values were obtained in the following way: W W 2 1 I I 2 VOV ⇒ = kn = 2 = 2 2 2 L L k n VOV VOV 2(I/2) I 0.4 = = mA/V VOV VOV VOV √ v id /2 2 = 0.1 → v id = 2 VOV 0.1 = VOV gm =

I 200 µA = = 100 µA 2 2 L = 2(0.18 µm) = 0.36 µm

V L ro1 = ro2 = ro3 = ro4 = A ID ID =

(10 V/ µm) (0.36 µm) = 36 k 0.1 mA W 1 2 VOV , Since ID1 = ID2 = μn Cox 2 L W W 2ID = = 2 L 1 L 2 μn Cox VOV =

2(100 µA) = 12.5 400 µA/V2 (0.2 V)2 W W 2ID = = L 3 L 4 μp Cox |VOV |2 2(100 µA) = 50 100 µA/V2 (0.2)2 gm =

ID (100 µA) (2) = = 1 mA/V, VOV /2 0.2 V

so

0.632 VOV I 0.8 mA Ex: 8.4 ID = = = 0.4 mA 2 2

Sedra, Smith, Carusone, Gaudet

=

Ad = gm1 (ro1 ro3 ) = 1(mA/V) (36 k 36 k) = 18 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 8–2

Ex: 8.6 L = 2 (0.18 µm) = 0.36 µm

V · L All ro = A ID

VCC

Ex: 8.8

2.5 V RC

The drain current for all transistors is I 200 µA ID = = = 100 µA 2 2 (10 V/ µm) (0.36 µm) ro = = 36 k 0.1 mA Referring to Fig. 8.13(a), W 1 2 VOV Since ID = μn Cox for all NMOS 2 L transistors, W W W W = = = L 1 L 2 L 3 L 4 2ID 2(100 µA) = = = 12.5 2 μn Cox VOV 400 µA/V2 (0.2 V)2 W W W W = = = L 5 L 6 L 7 L 8 =

2ID 2(100 µA) = = 50 2 μp Cox VOV 100 µA/V2 (0.2 V)2

IC1

IC2

RC 5 k

I 0.4 mA 2.5 V VEE IC1 = IC2 IE1 = IE2 =

0.4 mA I = 2 2

= 0.2 mA VCMmax VC + 0.4 V = VCC − IC RC + 0.4 V = 2.5 − 0.2 mA (5 k) + 0.4 V = +1.9 V

For all transistors,

VCMmin = −VEE + VCS + VBE

|ID | (0.1 mA) (2) = = 1 mA/V gm = |VOV | /2 (0.2 V)

VCMmin = −2.5 V + 0.3 V + 0.7 V = −1.5 V

From Fig. 8.13(b), Ron = (gm3 ro3 ) ro1 = (1 × 36) × 36

Input common-mode range is −1.5 V to +1.9 V

= 1.296 M

Ex: 8.9 Substituting iE1 + iE2 = I in Eq. (8.45) yields

Rop = (gm5 ro5 )ro7 = (1 × 36) × 36

iE1 =

= 1.296 M Ad = gm1 Ron Rop

0.99 I =

= (1 mA/V) (1.296 M 1.296 M) = 648 V/V

I

e(v B2 −v B1 )/VT

1+

I 1 + e(v B2 −v B1 )/VT 1 −1 v B1 − v B2 = −VT ln 0.99 = −25ln (1/99)

Ex: 8.7

= 25ln (99) = 115 mV Ex: 8.10 (a) The DC current in each transister is 0.5 mA. Thus VBE for each will be 0.5 VBE = 0.7 + 0.025ln 1 = 0.683 V ⇒ v E = 5 − 0.683 = +4.317 V (b) gm =

IC 0.5 = 20 mA/V = VT 0.025

(c) iC1 = 0.5 + gm1 v BE1 = 0.5 + 20 × 0.005 sin (2π × 1000t) = 0.5 + 0.1 sin(2π × 1000t), mA

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 8–3

iC2 = 0.5 − 0.1 sin(2π × 1000t), mA (d) v C1 = (VCC − IC RC ) − 0.1 ×RC sin(2π × 1000t) = (15 − 0.5 × 10) − 0.1 × 10 sin(2π × 1000t) = 10 − 1 sin(2π × 1000t) , V v C2 = 10 + 1 sin(2π × 1000t) , V (e) v C2 − v C1 = 2 · sin(2π × 1000t) , V (f) Voltage gain ≡ =

v C2 − v C1 v B1 − v B2

(that is, half the gain obtained with the output taken differentially), and from Eq. (8.74) we have | Acm |

RD 2RSS

Thus, CMRR ≡

| Ad | = gm RSS | Acm |

Substituting gm = 4 mA/V, RD = 5 k, and RSS = 25 k, |Ad | = 10 V/V

2 V peak = 200 V/V 0.01 V peak

|Acm | = 0.1 V/V CMRR = 100 or 40 dB

Ex: 8.11 The transconductance for each transistor is gm = 2μn Cox (W/L)ID ID = Thus, gm =

Ex: 8.14

VCC

0.8 mA I = = 0.4 mA 2 2

VB3 √

Q3

Q4

2 × 0.2 × 100 × 0.4 = 4 mA/V

vo1

The differential gain for matched v O2 − v O1 RD values is Ad = = gm RD v id

vi1

= 86 dB

vo2 Q2

Q1

vi2

I 200 A

If we ignore the 1% here, then we obtain Ad = gm RD = (4 mA/V) (5 k) = 20 V/V RD RD | Acm | = 2RSS RD 5 (0.01) = 0.001 V/V = 2 × 25 | Ad | 20 CMRR (dB) = 20 log = 20 log | ACM | 0.001

VB4

Q5

VB5

VEE I = 200 µA Since β 1, 200 µA I = = 100 µA 2 2 IC 100 µA = 4 mA/V = gm = = VT 25 mV

IC1 ≈ IC2 ≈ Ex: 8.12 From Exercise 8.11,

gm1 = gm2

gm = 4 mA/V Using Eq. (8.85) and the fact that RSS = 25 k, we obtain 2(4 mA/V) (25 k) (2 gm RSS ) = CMRR = gm 0.01 gm = 20, 000 CMRR (dB) = 20 log10 (20, 000) = 86 dB Ex: 8.13 If the output of a MOS differential amplifier is taken single-endedly, then 1 | Ad | = gm RD 2

Sedra, Smith, Carusone, Gaudet

RC1 = RC2 = RC = ro = =

|VA | IC

10 V = 100 k 100 µA

ro1 = ro2 =

10 VA = = 100 k I/2 0.1

re1 = re2 = re =

VT 25 mV = 0.25 k = IE 0.1 mA

RC r o 100 k 100 k = re 0.25 k = 200 V/V

| Ad | =

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 8–4

Rid = 2rπ ,

rπ =

β 100 = = 25 k gm 4 mA/V

Rid = 2(25 k) = 50 k REE =

=

VA 10 V = = 50 k I 200 µA

If the total load resistance is assumed to be mismatched by 1%, then we have | Acm | =

RC RC 2REE RC

100 × 0.01 = 0.01 V/V 2 × 50

Ad

200

= 20 log10

CMRR (dB) = 20 log10

Acm

0.01

= 86 dB =

Using Eq. (8.96), we obtain RC βro = βREE · RC + 2REE 1+ ro 1+

Ricm

=

2 × 10−3

2

2 2 + 2 × 10−3 + 2 × 10−3

2 3 × 2 × 10−3

= 3.5 mV Ex: 8.16 From Eq. (8.103), we get RC 2 IS 2 + VOS = VT RC IS = 25 (0.02)2 + (0.1)2 = 2.55 mV 100 ∼ 100 = IB = = 0.5 µA 2(β + 1) 2 × 101 β IOS = IB β = 0.5 × 0.1 µA = 50 nA

100 1+ 100 × 100 = 100 × 50 × 100 + 2 × 50 1+ 100 Ricm 1.68 M Ex: 8.15 From Exercise 8.4:

1 I = 0.4 mA 2 W 1 ID = μn Cox V2 2 L n OV

Ex: 8.17 ID =

1 2 × 0.2 × 100 × VOV 2 ⇒ VOV = 0.2 V 0.4 =

2ID 2 × 0.4 = = 4 mA/V VOV 0.2

VOV = 0.2 V

gm1,2 =

Using Eq. (8.97) we obtain VOS due to RD /RD as: RD VOV · VOS = 2 RD

Gmd = gm1,2 = 4 mA/V

0.2 = × 0.02 = 0.002 V i.e 2 mV 2 To obtain VOS due to kn , use Eq. (8.98), kn VOV VOS = 2 kn 0.2 × 0.02 = 0.002 ⇒ VOS = 2 ⇒ 2 mV The offset voltage arising from Vt is obtained from Eq. (8.99): VOS = Vt = 2 mV

ro2 =

VAn 20 = 50 k = ID 0.4

ro4 =

|VAp | 20 = 50 k = ID 0.4

Ro = ro2 ro4 = 50 50 = 25 k Ad = Gm Ro = 4 × 25 = 100 V/V Ex: 8.18 Gmd = gm1,2 = 16 mA/V ro2 = ro4 =

100 VA VA = = 250 k = IC I/2 0.4

Ro = ro2 ro4 = 250 250 = 125 k Ad = Gm Ro = 16 × 125 = 2000 V/V Rid = 2rπ = 2 ×

Finally, from Eq. (8.100) the total input offset is VOS =

1/2 VOV kn 2 VOV RD 2 2 + + (Vt ) 2 RD 2 kn

Sedra, Smith, Carusone, Gaudet

I/2 0.4 mA = VT 0.025 V

β 160 = 20 k =2× gm1,2 16

Ex: 8.19 From Exercise 8.17, we get ID = 0.4 mA VOV = 0.2 V gm1,2 = 4 mA/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 8–5

Gmd = 4 mA/V Ad = 100 V/V

ro6 = 3.2/0.128 = 25 k

Now,

ro7 = 3.2/0.128 = 25 k

RSS = 25 k gm3 = =

√

(e) Eq. (8.140):

2μp Cox

W L

A1 = −gm1 (ro2 ro4 )

ID

= −0.853 (50 50) = −21.3 V/V

p

2 × 0.1 × 200 × 0.4 = 4 mA/V

| Acm | =

1 2gm3 RSS

1 = = 0.005 V/V 2 × 4 × 25

Eq. (8.141): A2 = −gm6 (ro6 ro7 ) = −21.3 V/V

100 | Ad | = CMRR = | Acm | 0.005

Overall voltage gain is

= 20,000 or 20 log 20,000 = 86 dB

A1 × A2 = −21.3 × − 21.3 = 454 V/V

Ex: 8.20 Refer to Fig. (8.37).

Ex: 8.21 Rid = 20.2 k

(a) Using Eq. (8.142), we obtain

Av o = 8513 V/V

I6 =

Ro = 152

(W/L)6 (I/2) (W/L)4

With RS = 10 k and RL = 1 k,

(W/L)6 × 64 ⇒ 128 = 11.1 thus, (W/L)6 = 22.2

Gv =

1 20.2 × 8513 × 20.2 + 10 (1 + 0.152)

= 4943 V/V

Using Eq. (8.143), we get I7 =

(W/L)7 I (W/L)5

Ex: 8.22

ib8 R5 15.7 = = = 0.0492 A/A ic7 R5 + Ri4 15.7 + 303.5

(W/L)7 × 128 88.8 thus, (W/L)7 = 88.8 ⇒ 128 =

ic7 = β 7 = 100 A/A ib7

(b) For Q1 ,

W I 1 V2 = μp Cox 2 2 L 1 OV1 64 = 0.15 V ⇒ VOV1 = 1 × 128 × 44.4 2 Similarly for Q2 , VOV2 = 0.15 V For Q6 ,

Q1 Q2 Q6

ic5 = β 5 = 100 A/A ib5 R1 + R 2 40 ib5 = = = ic2 R1 + R2 + Ri2 40 + 5.05 0.8879 A/A

Thus the overall current gain is ie8 = 101 × 0.0492 × 100 × 0.0126 × 100 i1

2ID VOV

ID

ib7 R3 3 = 0.0126 A/A = = ic5 R3 + Ri3 3 + 234.8

ic2 = β 2 = 100 A/A i1

1 2 128 = × 512 × 22.2VOV6 2 ⇒ VOV6 = 0.15 V (c) gm =

ie8 = β 8 + 1 = 101 A/A ib8

× 0.8879 × 100 VOV

gm

64 µA 0.15 V 0.853 mA/V 64 µA 0.15 V 0.853 mA/V 128 µA 0.15 V 1.71 mA/V

(d) ro2 = 3.2/0.064 = 50 k

= 55,599 A/A and the overall voltage gain is R6 ie8 vo = · v id Ri1 i1 =

3 × 55599 = 8257 V/V 20.2

ro4 = 3.2/0.064 = 50 k

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–1

Solutions to End-of-Chapter Problems

(g) −0.2 V to + 1.0 V

8.1 Refer to Fig. 8.2. W I 1 (a) V2 = μn Cox 2 2 L 1,2 OV 0.08 = ⇒ VOV

= −0.2 V

8.2 Refer to Fig. P8.2. (a) For VG1 = VG2 = 0 V,

1 2 × 0.4 × 10 × VOV 2 = 0.2 V

1 × 0.2 = 0.1 mA 2 W 1 |VOV |2 = k p 2 L

ID1 = ID2 =

VGS = Vtn + VOV = 0.4 + 0.2 = 0.6 V (b) VCM = 0

ID1,2

ID1 = ID2 =

1 × 5 × |VOV |2 2 ⇒ |VOV | = 0.2 V

VD1 = VD2

VSG = |Vtp | + |VOV |

0.1 =

VS = 0 − VGS = −0.6 V I = 0.08 mA 2 = VDD − ID1,2 RD

= 1 − 0.08 × 5 = +0.6 V

= 0.4 + 0.2 = 0.6 V

VO = 0 V

VS = 0 + VSG = +0.6 V

(c) VCM = +0.4 V

VD1 = VD2 = −VSS + ID RD

VS = 0.4 − VGS = 0.4 − 0.6 = −0.2 V

= −0.9 + 0.1 × 4 = −0.5 V

I = 0.08 mA 2 = VDD − ID1,2 RD

ID1 = ID2 = VD1 = VD2

Since for each of Q1 and Q2 , VSD = 0.6 − (−0.5)

= 1 − 0.08 × 5 = +0.6 V

= 1.1 V

VO = 0 V Since VCM = 0.4 V and VD = 0.6 V, VGD = −0.2 V, which is less than Vtn (0.4 V), indicating that our implicit assumption of saturation-mode operation is justified. (d) VCM = −0.1 V VS = −0.1 − VGS = −0.1 − 0.6 = −0.7 V I ID1 = ID2 = = 0.08 mA 2 VD1 = VD2 = VDD − ID1,2 RD

which is greater than |VOV |, Q1 and Q2 are operating in saturation as implicitly assumed. (b) The highest value of VCM is limited by the need to keep a minimum of 0.2 V across the current source, thus VCMmax = +0.9 − 0.2 − VSG = +0.9 − 0.2 − 0.6 = +0.1 V The lowest value of VCM is limited by the need to keep Q1 and Q2 in saturation, thus VCMmin = VD1,2 − |Vtp |

= 1 − 0.08 × 5 = +0.6 V

= −0.5 − 0.4 = −0.9 V

VO = 0 V (e) The highest value of VCM for which Q1 and Q2 remain in saturation is

Thus, −0.9 V ≤ VICM ≤ +0.1 V

VCMmax = VD1,2 + Vtn = 0.6 + 0.4 = 1.0 V

8.3 Refer to the circuit in Fig. P8.3.

(f) To maintain the current-source operating properly, we need to keep a minimum voltage of 0.2 V across it, thus

For v G1 = v G2 = 0 V, ID1 = ID2 =

VSmin = −VSS + VCS = −1 + 0.2 = −0.8 V

To obtain

VCMmin = VSmin + VGS

VD1 = VD2 = +0.1 V

= −0.8 + 0.6

VDD − ID1,2 RD = 0.1

Sedra, Smith, Carusone, Gaudet

0.4 = 0.2 mA 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–2

1 1 × 0.25 = × 0.4 × 2 2

0.9 − 0.2 RD = 0.1 ⇒ RD = 4 k

⇒

For Q1 and Q2 ,

W 1 μn Cox V2 2 L 1,2 OV W 1 0.2 = × 0.4 × 0.152 2 L 1,2 W ⇒ = 44.4 L 1,2

ID1,2 =

For Q3 ,

1 W × 0.4 × 2 L W ⇒ = 88.8 L 3

W 0.252 L

W = 10 L

8.5 For v id = 0.1 V v id /2 2 = 0.04 VOV v id /2 = 0.2 VOV

× 0.152

0.4 =

3

0.1/2 = 0.2 VOV ⇒ VOV = 0.25 V gm =

Since Q3 and Q4 form a current mirror with ID3 = 4ID4 , 1 W W = = 22.2 L 4 4 L 3 VGS4 = VGS3 = Vtn + VOV = 0.4 + 0.15 = 0.55 V 0.9 − (−0.9) − 0.55 R= 0.1 = 12.5 k

2 × (I/2) VOV

I 0.25 ⇒ I = 0.5 mA 1V = 10 Ad = 0.1 V gm RD = 10 2=

10 = 5 k 2 1 W I 2 = μn Cox VOV 2 2 L

⇒ RD =

The lower limit on VCM is determined by the need to keep Q3 operating in saturation. For this to happen, the minimum value of VDS3 is VOV = 0.15 V. Thus,

0.25 = ⇒

VICMmin = −VSS + VOV3 + VGS1,2

W 1 × 0.2 × × 0.252 2 L

W = 40 L

= −0.9 + 0.15 + 0.4 + 0.15

8.6 Since the quiescent power dissipation is

= −0.2 V

P = (VDD + VSS ) × I

The upper limit on VCM is determined by the need to keep Q1 and Q2 in saturation, thus

then the maximum allowable I is 0.3 mW = 0.25 mA I= 1.2 V We shall utilize this value. The value of VOV can be found from √ 2 VOV = 0.2 V

VICMmax = VD1,2 + Vtn = 0.1 + 0.4 = 0.5 V Thus, −0.2 V ≤ VICM ≤ +0.5 V 8.4 Refer to Fig. 8.5. 2(I/2) I = gm = VOV VOV I 0.25 ⇒ I = 0.25 mA 1 W I 2 = μn Cox VOV 2 2 L 1=

Sedra, Smith, Carusone, Gaudet

0.2 ⇒ VOV = √ = 0.14 V 2 The realized value of gm will be gm =

2 × (I/2) VOV

0.25 = 1.8 mA/V 0.14 To obtain a differential gain Ad of 8 V/V, =

Ad = gm RD

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–3

Since I = IREF , Q3 and Q6 are matched and are operating at

8 = 1.8 × RD ⇒ RD = 4.4 k Finally, the required value of W/L can be determined from W 2 I 1 ID = = μn Cox VOV 2 2 L 1 W 0.125 = × 0.5 × × 0.142 2 L W ⇒ = 25.5 L 8.7

Thus,

W 1 × 0.22 × 0.1 × 2 L 6,3 W W ⇒ = = 100 L 3 L 6

0.2 =

Each of Q4 and Q5 is conducting a dc current of (I/2) while Q7 is conducting a dc current IREF = I. Thus Q4 and Q5 are matched and their W/L ratios are equal while Q7 has twice the (W/L) ratio of Q4 and Q5 . Thus, W I 1 V2 = μn Cox 2 2 L 4,5 OV4,5 where

RD vod / 2 vid 2

VOV4,5 = −0.8 − (−1.5) − 0.5 = 0.2 V thus,

W 1 × 0.04 × 0.25 × 2 L 4,5 W W = = 20 ⇒ L 4 L 5

0.1 =

Rs / 2

From symmetry, a virtual ground appears at the midpoint of Rs . Thus, the differential half circuit will be as shown in the figure, and Ad ≡

|VOV | = 1.5 − 0.8 − 0.5 = 0.2 V

v od RD = 1 Rs v id + gm 2

and W = 40 L 7 ro4 = ro5 =

|VAp | 10 = = 100 k I/2 0.1

ro1 = ro2 =

VAn 10 = = 100 k I/2 0.1

Ad = gm1,2 (ro1,2 ro4,5 )

For Rs = 0,

50 = gm1,2 (100 100)

RD = gm RD , Ad = 1/gm

⇒ gm1,2 = 1 mA/V But

as expected. To reduce the gain to half this value, we use

gm1,2 =

1 Rs = 2 gm

1=

⇒ Rs =

2 gm

0.2 |VOV1,2 |

⇒ |VOV1,2 | = 0.2 V

8.8 Refer to Fig. P8.8. The value of R is found as follows: VG6 − VG7 R= IREF =

2(I/2) |VOV1,2 |

The (W/L) ratio for Q1 and Q2 can now be determined from W 1 × 0.22 0.1 = × 0.1 × 2 L 1,2 W W ⇒ = = 50 L 1 L 2

0.8 − (−0.8) = 8 k 0.2

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–4

A summary of the results is provided in the table below.

8.10 VDD 1 V

Transistor W/L I D (mA) |V GS |(V) Q1

50

0.1

0.7

Q2

50

0.1

0.7

Q3

100

0.2

0.7

Q4

20

0.1

0.7

Q5

20

0.1

0.7

Q6

100

0.2

0.7

Q7

40

0.2

0.7

RD 5 k vD1 G1 vid

vGS1

5 k vD2

iD1 iD2 Q2

Q1

vS

vGS2

VSS 1 V

All transistors have the same channel length and are carrying a dc current I/2. Thus all transistors |VA | . Also, all transistors are have the same ro = I/2 operating at the same |VOV | and have equal dc currents, thus all have the same 2(I/2) = I/|VOV |. Thus all transistors have gm = |VOV | equal intrinsic gain gm ro = 2|VA |/|VOV |. Now, the gain Ad is given by Ad = gm (Ron Rop ) 1 gm Ron 2 1 1 = gm (gm ro )ro = (gm ro )2 2 2 Thus, 1 2|VA | 2 Ad = 2 VOV

(a) For iD1 = iD2 = 0.08 mA, v G1 = v G2 Thus, v id = 0 V iD1 = iD2 = 0.08 =

=

⇒ VOV

W 1 2 μn Cox VOV 2 L

1 2 × 0.4 × 10 × VOV 2 = 0.2 V

v GS1 = v GS2 = 0.2 + 0.4 = 0.6 V v S = −0.6 V v D1 = v D2 = VDD − iD1,2 RD = 1 − 0.08 × 5 = 0.6 V

Q.E.D.

To obtain Ad = 500 V/V while operating all transistors at |VOV | = 0.2 V, we use |VA |2 0.04 ⇒ |VA | = 3.16 V

500 = 2

Since |VA | = 5 V/µm, the channel length L (for all transistors) must be

v O = v D2 − v D1 = 0 V (b) For iD1 = 0.12 mA and iD2 = 0.04 mA, W 1 (v GS2 − Vtn )2 iD2 = μn Cox 2 L 1 × 0.4 × 10 × (v GS2 − 0.4)2 2 ⇒ v GS2 = 0.541 V 0.04 =

3.16 = 5 × L

Thus,

L = 0.632 µm

v S = −0.541 V W 1 (v GS1 − Vtn )2 iD1 = μn Cox 2 L

To obtain the highest possible gm , we operate at the highest possible I consistent with limiting the power dissipation (in equilibrium) to 0.5 mW. Thus, I=

G2

I 0.16 mA

8.9 Refer to Fig. 8.13.

= 2(|VA |/|VOV |)2

RD

0.5 mW = 0.28 mA (0.9 + 0.9)V

Sedra, Smith, Carusone, Gaudet

0.12 = =

1 × 0.4 × 10 (v id − v S − Vtn )2 2

1 × 0.4 × 10 (v id + 0.541 − 0.4)2 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–5

Case

iD1 (mA)

iD2 (mA)

vid (V)

vS (V)

vD1 (V)

vD2 (V)

vO

a

0.08

0.08

0

−0.6

+0.6

+0.6

0

b

0.12

0.04

+0.104

−0.541

+0.4

+0.8

+0.4

c

0.16

0

+0.283

−0.4

+0.2

+1.0

+0.8

d

0.04

0.12

−0.104

−0.645

+0.8

+0.4

−0.4

e

0

0.16

−0.283

−0.683

+1.0

+0.2

−0.8

⇒ v id = 0.104 V

= 1 − 0.04 × 5 = 0.8 V

v GS1 = 0.104 − (−0.541) = 0.645 V

v D2 = 1 − 0.12 × 5 = 0.4 V

v D1 = VDD − iD1 RD

v O = v D2 − v D1 = −0.4 V

= 1 − 0.12 × 5 = 0.4 V

(e) iD1 = 0 (Q1 just cuts off) and iD2 = 0.16 mA. This case is the complement of that in (c) above, thus

v D2 = VDD − iD2 RD = 1 − 0.04 × 5 = 0.8 V

v GS1 = Vtn = 0.4 V

v O = v D2 − v D1 = 0.8 − 0.4 = 0.4 V

v GS2 = 0.683 V

(c) iD1 = 0.16 mA and iD2 = 0 with Q2 just cutting off, thus

v S = −0.683 V

v GS2 = Vtn = 0.4 V

v id = −0.683 + 0.4 = −0.283 V √ which is − 2 VOV , as derived in the text.

⇒ v S = −0.4 V 1 × 0.4 × 10 (v GS1 − Vtn )2 2 1 0.16 = × 0.4 × 10 (v id + 0.4 − 0.4)2 2 ⇒ v id = 0.283 V √ which is 2VOV , as derived in the text.

iD1 =

v GS1 = 0.283 − (−0.4) = 0.683 V

v D1 = VDD − iD1 RD = 1 − 0 × 5 = 1 V v D2 = VDD − iD2 RD = 1 − 0.16 × 5 = 0.2 V v O = v D2 − v D1 = −0.8 V Summary A summary of the results is shown in the table above.

v D1 = VDD − iD1 RD

8.11 Refer to Fig. P8.2.

= 1 − 0.16 × 5 = +0.2 V

To determine VOV , consider the case v id = 0,

Note that since v G1 = v id = 0.283 V, Q1 is still operating in saturation, as implicitly assumed.

0.1 =

v D2 = VDD − iD2 RD =1−0×5=1V v O = v D2 − v D1 = 1 − 0.2 = 0.8 V (d) iD1 = 0.04 mA and iD2 = 0.12 mA. Since this split of the current I is the complement of that in case (b) above, the value of v id must be the negative of that found in (b). Thus,

1 × 5 × |VOV |2 2 ⇒ |VOV | = 0.2 V

With v G2 = 0 and v G1 = v id , to steer the current from one side of the differential pair to the other, v id must be the ends of the range √ √ − 2 |VOV | ≤ v id ≤ 2 |VOV | that is, −0.283 V ≤ v id ≤ +0.283 V

v id = −0.104 V

At v id = −0.283 V, Q2 just cuts off, thus

v GS1 = 0.541 V

v S = |Vtp | = 0.4 V

v S = −0.645 V

and

v GS2 = 0.645 V

v SG1 = 0.4 − (−0.283) = 0.683 V

v D1 = VDD − iD1 RD

At this value of v SG1 ,

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–6

1 × 5 × (0.683 − 0.4)2 2 = 0.2 mA

For iD1 = 0.09 mA and iD2 = 0.07 mA, W 1 (v GS2 − Vtn )2 iD2 = μn Cox 2 L

iD1 =

which is the entire bias current.

1 × 0.4 × 10(v GS2 − 0.4)2 2 ⇒ v GS2 = 0.587 V 0.07 =

v D1 = −0.9 + 0.2 × 4 = −0.1 V Thus, v DG1 = −0.1 − (−0.283) = +0.183 which is less than |Vtp |; thus Q1 is still operating in saturation, as implicitly assumed. v D2 = −0.9 V Finally, v O = v D2 − v D1 = −0.8 V

and v S = −0.587 V W 1 (v GS1 − Vtn )2 iD1 = μn Cox 2 L 1 × 0.4 × 10 (v GS1 − 0.4)2 2 ⇒ v GS1 = 0.612 V 0.09 =

At v id = +0.283 V, Q1 just cuts off, thus v SG1 = |Vtp | = 0.4 V and v S = +0.283 + 0.4 = +0.683 V

v id = v S + v GS1 = −0.587 + 0.612

and thus

= 0.025 V

v SG2 = 0.683 V

v D2 = VDD − iD2 RD

which results in 1 iD2 = × 5 (0.683 − 0.4)2 2 = 0.2 mA

= 1 − 0.07 × 5 = 0.65 V v D1 = 1 − 0.09 × 5 = 0.55 V v O = v D2 − v D1 = 0.65 − 0.55 = 0.10 V Voltage gain =

which is the entire bias current. Here, v D1 = −0.9 V

0.10 vO = 4 V/V = v id 0.025

To obtain the complementary split in current, that is, iD1 = 0.07 mA and iD2 = 0.09 mA,

v D2 = −0.9 + 0.2 × 4 = −0.1 V which verifies that Q2 is operating in saturation, as implicitly assumed.

v id = −0.025 V 8.13 From Exercise 8.3 and the accompanying table, we note that |v id |max is proportional to VOV :

Finally, v O = v D2 − v D1 = +0.8 V

0.126 |v id |max = 0.63 = VOV 0.2 Thus, to obtain |v id |max = 220 mV = 0.22 V at the same level of linearity, we use

8.12

VDD 1 V

RD

RD 5 k vD1 G1 vid

vGS1

5 k vD2

iD1 iD2 Q2

Q1

vS

0.22 = 0.35 V 0.63 For this value of VOV , the required (W/L) can be found from W 1 × 0.352 0.2 = × 0.2 × 2 L

VOV =

G2 vGS2

W = 16.3 L The value of gm is ⇒

gm =

2 ID 2 × 0.2 = 1.14 mA/V = VOV 0.35

I 0.16 mA

VSS 1 V

Sedra, Smith, Carusone, Gaudet

8.14 Refer to Eq. (8.23). For v id /2 2 ≤k VOV

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–7 ⇒

v id /2 VOV

≤

√

(1)

k

v id /2 v id /2 2 1− VOV VOV √ √ =I k 1−k

(2)

Q.E.D.

and the corresponding value of v id is found from Eq. (1) as √ Q.E.D. (3) v idmax = 2 kVOV Equations (2) and (3) can be used to evaluate Imax v idmax for various values of k: and I/2 VOV k

v idmax VOV Imax I/2

0.01

0.1

0.2

0.2

0.632

0.894

1 1− 4

v id VOV

2 = 0.1

v id 0.1 VOV

v id 0.1VOV

Thus, Imax = 2 k(1 − k) I/2

v id VOV

⇒

I = I Imax

(b) In Eqs. (8.23) and (8.24) let I I iD1 = + × 2 2 I I − × iD2 = 2 2 where v id /2 2 v id 1− = VOV VOV If v id is such that iD1 =m iD2 then

0.2

0.6

0.8

m=

1+ 1−

⇒=

m−1 m+1

8.15 Switching occurs at √ v id = 2VOV

For m = 1, = 0 and v id = 0

Thus,

For m = 2,

0.3 =

√

2−1 = 2+1 v id 1− VOV

2VOV

=

⇒ VOV = 0.212 V Now, to obtain full current switching at v id = 0.5 V, VOV must be increased to VOV = 0.212 ×

0.5 = 0.353 V 0.3

2 Since ID is proportional to VOV the current ID and hence the bias current I must be increased by the ratio (0.353/0.212)2 , then I must be 0.353 2 I = 200 × = 554.5 µA 0.212

8.16 Equations (8.23) and (8.24): I v id I v id /2 2 iD1 = + 1− 2 2 VOV VOV I v id I v id /2 2 1− iD2 = − 2 2 VOV VOV

v id VOV

2 =

1 3

v id = 0.338VOV For m = 1.1, =

0.1 1.1 − 1 = 0.05 1.1 + 1 2.1

v id VOV

1−

1 4

v id VOV

2 = 0.05

⇒ v id 0.05VOV (8.24)

(a) For 10% increase above the equilibrium value I of , 2 v id /2 2 I v id I 1− = 0.1 × 2 VOV VOV 2

Sedra, Smith, Carusone, Gaudet

1 4

Squaring both sides, we obtain a quadratic v id 2 which can be solved to equation in VOV obtain

Thus, (8.23)

1 3

For m = 1.01 1.01 − 1 = 0.005 1.01 + 1 1 v id 2 v id 1− = 0.005 VOV 4 VOV

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–8

v id 0.005VOV

8.19 (a) Ad = gm RD

For m = 20,

20 = gm × 47

19 m−1 = = 0.905 V = m+1 21

⇒ gm =

Thus,

v id VOV

1 1− 4

v id VOV

20 = 0.426 mA/V 47 2ID 2(I/2) I = = (b) gm = VOV VOV VOV

2 = 0.905

⇒ v id = 1.072VOV

(c) Across each RD the dc voltage is

1 2 × 0.2 × 32VOV 2 = 0.18 V

8.17 0.1 = ⇒ VOV

I 0.2 ⇒ I = 0.085 mA = 85 µA 0.426 =

2 × (0.2/2) = 1.11 mA/V 0.18 VA 10 = 100 k = ro = ID 0.1 gm =

0.085 I RD = × 47 = 2 V 2 2 (d) The peak sine-wave signal of each gate source is 5 mV, thus at each drain the peak sine wave is Ad × 5 = 20 × 5 = 100 mV = 0.1 V

Ad = gm (RD ro )

(e) The minimum voltage at each drain will be

= 1.11 × (10 100) = 10.1 V/V

v Dmin = VDD − RD ID − Vpeak = VDD − 2 − 0.1

8.18 To limit the power dissipation to 1 mW, P = (VDD + VSS )I

v Dmin ≥ v Gmax − Vtn

Thus, the maximum value we can use for I is 1 mW = 0.5 mA 2V Using this value, we obtain I=

I RD 2 0.2 = 1 − 0.25 × RD

VD = VDD −

⇒ RD = 3.2 k Ad = gm RD

For the transistor to remain in saturation

where v Gmax = VCM + Vpeak (input) = 0.5 + 0.005 = 0.505 V Thus, VDD − 2.1 ≥ 0.505 − 0.5 VDD ≥ 2.105 V Thus, the lowest value of VDD is 2.11 V.

10 = gm × 3.2 gm =

10 = 3.125 mA/V 3.2

But 2 × (I/2) I gm = = VOV VOV 3.125 =

0.5 VOV

8.20 For a CS amplifier biased at a current ID and utilizing a drain resistance RD , the voltage gain is |A| = gm RD where gm =

⇒ VOV = 0.16 V

Thus,

To obtain W/L, we use W 1 2 VOV ID = μn Cox 2 L

|A| =

0.25 =

1 W × 0.4 × × 0.162 2 L

W = 48.8 50 ⇒ L

Sedra, Smith, Carusone, Gaudet

2μn Cox 2μn Cox

W ID L W ID RD L

(1)

For a differential pair biased with a current I and utilizing drain resistances RD , the differential gain is Ad = gm RD where

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–9

Thus

8.22

W L

2μn Cox

W I/2 RD L

Ad =

I 2

2μn Cox

gm =

(2)

Q3

Equating the gains from Eqs. (1) and (2), we get I = 2ID

(vod / 2)

That is, the differential pair must be biased at a current twice that of the CS amplifier. Since both circuits use equal power supplies, the power dissipation of the differential pair will be twice that of the CS amplifier.

8.21 Since both circuits use the same supply voltages and dissipate equal powers, their currents must be equal, that is, ID = I where ID is the bias current of the CS amplifier and I is the bias current of the differential pair. The gain of the CS amplifier is | A| = gm RD where gm = Thus,

2μn Cox

| A| =

W L

2μn Cox

ID CS

(1)

ID RD CS

The gain of the differential amplifier is Ad = gm RD where gm = Thus, Ad =

2μn Cox

2μn Cox

W L

W L

I 2 diff

I RD 2 diff

(2)

Equating the gains in Eqs. (1) and (2) and substituting ID = I gives W W 1 = × L CS L diff 2 W W ⇒ =2 L diff L CS If all transistors have the same channel length, each of the differential pair transistors must be twice as wide as the transistor in the CS amplifier.

Sedra, Smith, Carusone, Gaudet

Q1

(a) The figure shows the differential half-circuit. Recalling that the incremental (small-signal) resistance of a diode-connected transistor is given 1 ro , the equivalent load resistance of by gm Q1 will be RD =

1 ro3 gm3

and the differential gain of the amplifier in Fig. P8.22 will be 1 v od = gm1 ro3 ro1 Ad ≡ v id gm3

W L

vid 2

Since both sides of the amplifier are matched, this expression can be written in a more general way as 1 ro3,4 ro1,2 Ad = gm1,2 gm3,4 (b) Neglecting ro1,2 and ro3,4 (much larger than 1/gm3,4 ), gm1,2 Ad gm3,4 2μn Cox (W/L)1,2 (I/2) = 2μp Cox (W/L)3,4 (I/2) μn (W/L)1,2 = μp (W/L)3,4 (c) μn = 4μp and all channel lengths are equal, W1,2 Ad = 2 W3,4 For Ad = 10, W1,2 10 = 2 W3,4 ⇒

W1,2 = 25 W3,4

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–10

8.23 Refer to Fig. P8.23.

But

(a) With v G1 = v G2 = 0,

VOV3 = VC + VOV1,2

v GS1 = v GS2 = VOV1,2 + Vtn

⇒ VC = VOV1,2

Thus 8.24 Refer to Fig. P8.24.

VS1 = VS2 = −(VOV1,2 + Vtn ) (b) For the situation in (a), VDS of Q3 is zero, thus zero current flows in Q3 . Transistor Q3 will have an overdrive voltage of VOV3 = VC − VS1,2 − Vtn = VC + (VOV1,2 + Vtn ) − Vtn = VC + VOV1,2 (c) With v G1 = v id /2 and v G2 = −v id /2 where v id is a small signal, a small signal will appear between drain and source of Q3 . Transistor Q3 will be operating in the triode region and its drain-source resistance rDS will be given by

rDS = μn Cox

1 W VOV3 L 3

(a) With v G1 = v G2 = 0 V, VS1 = VS2 = −VGS1,2 = −(Vt + VOV ) The current through Q3 and Q4 will be zero because the voltage across them (v DS3 + v DS4 ) is zero. Because the voltages at their gates are zero and at their sources are −(Vt + VOV ), each of Q3 and Q4 will be operating at an overdrive voltage equal to VOV . Thus each of Q3 and Q4 will have an rDS given by

rDS3,4 = μn Cox Since

gm1,2 = μn Cox

Thus, 1 W VOV3 L 3

Rs = μn Cox Now,

rDS3,4 =

W VOV1,2 L 1,2 W gm3 = (μn Cox ) VOV3 L 3 W W For = , L 3 L 1,2 gm1,2 W = μn Cox L VOV1,2

1

Rs = gm1,2 = gm1,2 × VOV3 VOV1,2

VOV

(2)

1,2

1 (W/L)1,2 gm1,2 (W/L)3,4

Rs = rDS3 + rDS4 then Rs =

2 (W/L)1,2 gm1,2 (W/L)3,4

VOV1,2 VOV3

=

2 RD 1 1 + Rs + gm1 gm2

Using (3), we obtain

1 gm1,2

But VOV3 = VC + VOV1,2 ⇒ VC = 0 0.5 gm1,2

⇒ VOV3 = 2 VOV1,2

Sedra, Smith, Carusone, Gaudet

(3)

(b) With v G1 = v id /2 and v G2 = −v id /2 where v id is a small signal, v od Ad ≡ v id

Ad =

VOV3 = VOV1,2

(ii) Rs =

and since

Thus,

(d) (i) Rs =

W L

(1)

substituting from (2) into (1) gives

gm1,2 = (μn Cox )

1

1 W VOV L 3,4

1 gm1,2

=

RD 1 (W/L)1,2 + gm1,2 (W/L)3,4

gm1,2 RD (W/L)1,2 1+ (W/L)3,4

8.25 Refer to Fig. P8.25. (a) Since the dc voltages VGS1 and VGS2 are equal, Q1 and Q2 will be operating at the same value of

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–11

VOV and their dc currents ID1 and ID2 will have the same ratio at their (W/L) ratios, that is, ID1 = I/3

0.4 V

(b) Q1 and Q2 will be operating at the same VOV , obtained as follows: W I 1 2 VOV = μn Cox 3 2 L

2I

⇒ VOV =

W 3μn Cox L

=

v od v id

2 × (2I/3) 4I = VOV 3VOV

Since VC1,2 = 0.4 V, VEmax = 0.4 − 0.3 = 0.1 V

VCMmax = 0.574 + 0.1 = 0.674 V The minimum value of VCM is dictated by the need to keep the current source operating properly, i.e., to keep 0.3 V across it, thus

where

gm2 =

The maximum value of VCM is limited by the need to keep Q1 and Q2 in the active mode. This is achieved by keeping v CE1,2 ≥ 0.3 V.

VCMmax = VBE1,2 + VEmax

1 1 + gm1 gm2 2 × (I/3) 2I = VOV 3VOV

(b) Refer to Fig. 8.15(a).

and

2RD

gm1 =

VC1 = VC2 = VCC − IC RC = 1.2 − 9.8 × 10−3 × 82

ID2 = 2I/3

(c) Ad ≡

VE = −0.574 V

VEmin = −1.2 + 0.3 = −0.9 V and VCMmin = VEmin + VBE1,2

8 IRD 2RD = Ad = 3 3 9 VOV + (VOV /I) 2 4

= −0.9 + 0.574 = −0.326 V Thus, the input common-mode range is −0.326 V ≤ VICM ≤ +0.674 V

8.26 Refer to Fig. 8.15(a).

(c) Refer to Fig. 8.15(d).

The current I will split equally between Q1 and Q2 . Thus,

iE1 = 11 µA,

IE1 = IE2 = 0.2 mA

iE2 = 9 µA

iC1 = 10.78 µA,

iC2 = 8.82 µA 10.78 × 10−3 = 0.69 + 0.025 ln 1

IC1 = IC2 = α × 0.2 = 0.99 × 0.2 = 0.198 mA 0.198 VBE1 = VBE2 = 0.7 + 0.025 ln 1

v BE1

= 0.660 V

v BE2 = 0.69 + 0.025 ln

VE1 = VE2 = −1 − 0.66 = −1.66 V

= 0.5717 V

VC1 = VC2 = VCC − IC1.2 RC

Thus,

= 2.5 − 0.198 × 5 = +1.51 V

v B1 = v BE1 − v BE2

= 0.5767 V

8.82 × 10−3 1

= 0.5767 − 0.5717 = 0.005 V 8.27 (a) Refer to Fig. 8.15(a). I IE1 = IE2 = = 10 µA 2 IC1 = IC2 = α × 10 = 0.98 × 10 = 9.8 µA 9.8 × 10−3 VBE1 = VBE2 = 0.690 + 0.025 ln 1

= 5 mV 8.28 Require v O = 1 V when v id = 10 mV and I = 1 mA. Using Eq. (8.48), we obtain 1 (mA) = 0.599 mA 1 + e−10/25

= 0.574 V

iE1 =

Thus,

iE2 = I − iE1 = 1 − 0.599 = 0.401 mA

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–12

v O = v C2 − v C1 = (VCC − iC2 RC ) − (VCC − iC1 RC )

At each collector we expect a signal of 40 × 5 mV = 200 mV. Between the two collectors, the signal will be 400 mV.

= (iC1 − iC2 )RC (iE1 − iE2 )RC

8.32

= 0.198RC

VCC

For v O = 1 V, we have 1 = 5.05 k 0.198 I VC1 = VC2 = VCC − RC 2 = 5 − 0.5 × 5.05 2.5 V RC =

RC

100 mV

With a signal of 10 mV applied, the voltage at one collector rises to 3 V and at the other falls to 2 V. To ensure that the transistors remain in the active region, the maximum common-mode input voltage must be limited to (2 + 0.4) = +2.4 V. 8.29 gm =

β 160 = 40 k =2× gm 8

8.30 Rid = 2rπ = 20 k rπ = 10 k

0

I

VEE

v id = 200 mV appears across (2 re + 2 Re ). Thus the signal across (re + Re ) is 100 mV. Since the signal across re is 5 mV, it follows that the signal across Re must be 100 − 5 = 95 mV and thus

The input resistance Rid is Rid = (β + 1) (2re + 2Re )

⇒ gm = 10 mA/V

= 2(100 + 1) (re + Re )

Ad = 100 = gm RC

= 2 × 101 × (re + 19re )

100 100 RC = = = 10 k gm 10

= 2 × 101 × 20re To obtain Rid = 400 k,

IC I/2 VT VT

400 = 2 × 101 × 20 × re

⇒ I = 2VT gm

⇒ re 100

= 2 × 0.025 × 10 = 0.5 mA

Since VT , re = IE

8.31 v id = 10 mA/V Input signal to half-circuit = 5 mV. For I = 200 µA, the bias current of the half-circuit is 100 µA and, 25 mV = 250 re = 0.1 mA

= −40 V/V

Re 95 mV

(See signal distribution on the figure above.)

100 = 10 gm

Gain of half-circuit = −

100 mV

Re = 19re

β = 10 k gm

gm =

vod

5 mV

5 mV Re 95 mV

IC α × 0.2 = 8 mA/V VT 0.025

Rid = 2rπ = 2

RC

100 =

25 mV IE

⇒ IE = 0.25 mA I = 0.5 mA

10 RC =− re 0.25

Sedra, Smith, Carusone, Gaudet

Re = 19re = 19 × 0.1 = 1.9 k Gain =

α × 2RC 2re + 2Re

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–13

RC r e + Re

8.34

But the gain required is Gain =

2V v od = 10 V/V = v id 0.2 V

RC

Thus, 10 =

RC 0.1 + 1.9

Q1

⇒ RC = 20 k

vod

RC

RL

Q2

The determination of a suitable value of VCC requires information on the required input common-mode range (which is not specified). Suffice it to say that the dc voltage drop across RC is 5 V and that each collector swings ±1 V. A supply voltage VCC = 10 V will certainly be sufficient. Total resistance between collectors Total resistance in the emitter circuit (2RC RL ) =α 2re Ad = α

8.33

RC vod / 2

8.35 Rid = 2rπ = 2

IC 0.2 = 8 mA/V VT 0.025

gm =

vid 2

β gm

2 × 100 = 25 k 8 α(2RC RL ) Rid Gv = Rid + Rsig 2re

Rid =

1 R 2 id Biased at (I/2)

Re

Gv = Both circuits have the same differential half-circuit shown in the figure. Thus, for both Ad =

1 25 × × 8 × (40 40) 25 + 100 2

= 16 V/V

αRC r e + Re

Rid = (β + 1)(2re + 2Re )

8.36 Refer to Fig. P8.36.

= 2(β + 1)(re + Re ) With v id = 0, the dc voltage appearing at the top end of the bias current source will be I Re (a) VCM − VBE − 2 (b) VCM − VBE Since circuit (b) results in a larger voltage across the current source and given that the minimum value of VCM is limited by the need to keep a certain specified minimum voltage across the current source, we see that circuit (b) will allow a larger negative VCM .

Sedra, Smith, Carusone, Gaudet

=

Rid 1 × gm (2RC RL ) Rid + Rsig 2

re =

VT 25 mV = 250 = IE 0.1 mA

vo α × Total resistance in collectors = vi Total resistance in emitters =

0.99 × 25 k 2re + 500

=

0.99 × 25 k 25 V/V 500 + 500

Rin = (β + 1)(2re + 500 ) = 101 × (2 × 250 + 500 ) = 101 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–14

= +0.52 V

8.37

VC2 = VCC − IC2 × RC

VCC 2.5 V IC1

= 2.5 − 0 × 5 = 2.5 V

IC2

RC 5 k

RC 5 k

VC1

VC2

0.5 V

Q1

Observe that Q1 is operating in the active mode, as implicitly assumed, and the current source has a voltage of 2.323 V across it, more than sufficient for its proper operation. (b) With v B1 = −0.5 V, Q1 turns off and Q2 conducts all the bias current (0.4 mA) and thus exhibits a VBE of 0.677 V, thus

Q2

VE = −0.677 V

VE

which indicated that VBE1 = +0.177 V, which is too small to turn Q1 on. Also, note that the current source has a voltage of −0.677 + 2.5 = 1.823 V across it, more than sufficient for its proper operation.

0.4 mA

VEE 2.5 V (a)

VC1 = VCC − IC1 RC = 2.5 − 0 × 5 = 2.5 V

VCC 2.5 V IC1

VC2 = 2.5 − 0.99 × 0.4 × 5 = +0.52 V

IC2

RC 5 k

RC 5 k

VC1

VC2

0.5 V

Q1

Q2

VE 0.4 mA

8.38 Refer to Fig. 8.15(a) and assume the current source I is implemented with a single BJT that requires a minimum of 0.3 V for proper operation. Thus, the minimum voltage allowed at the emitters of Q1 and Q2 is −2.5 V + 0.3 V = −2.2 V. Now, since each of Q1 and Q2 is conducting a current of 0.2 mA, their VBE voltages will be equal: 0.99 × 0.2 VBE1,2 = 0.7 + 0.025 ln 1 = 0.660 V

VEE 2.5 V (b)

Thus, the minimum allowable VCM is VCMmin = −2.2 + 0.660 = −1.54 V

(a) For v B1 = +0.5 V, Q1 conducts all the current I (0.4 mA) while Q2 cuts off. Thus Q1 will have a VBE obtained as follows: 0.99 × 0.4 VBE1 = 0.7 + 0.025 ln 1

The upper limit on VCM is dictated by the need to keep Q1 and Q2 operating in the active mode, thus

= 0.677 V

= +1.91 V

Thus,

Thus, the input common-mode range is

VE = +0.5 − 0.677 = −0.177 V which indicates that VBE2 = +0.177 V, too small to turn Q2 on. VC1 = VCC − IC1 RC = 2.5 − 0.99 × 0.4 × 5

Sedra, Smith, Carusone, Gaudet

VCMmax = 0.4 + VC1,2 = 0.4 + (2.5 − 0.99 × 0.2 × 5)

−1.54 V ≤ VICM ≤ 1.91 V 8.39 The solution is given on the circuit diagram on the next page.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–15

8.42

iE1 − (I/2) iE1 = I I

iE1 − 0.5 I Using Eq. (8.34), we obtain =

1 iE1 = − 0.5 I 1 + e−v id /VT

8.40 Refer to Fig. 8.15(a) with VCC replaced by (VCC + v r ).

We use this equation to obtain the results in the table below. Observe that for v id < 10 mV the proportional transconductance gain is nearly constant at about 10. The gain decreases as v id further increases, indicating nonlinear operation. This is especially pronounced for v id > 20 mV. 8.43 Refer to Fig. 8.14.

I v C1 = (VCC + v r ) − α RC 2 I = (VCC − α RC ) + v r 2 I v C2 = (VCC + v r ) − α RC 2 I = (VCC − α RC ) + v r 2 v od ≡ v C2 − v C1 = 0

v o = v C2 − v C1 = (VCC − iC2 RC ) − (VCC − iC1 RC ) = RC (iC1 − iC2 )

Thus, while v C1 and v C2 will include a ripple component v r , the difference output voltage v od will be ripple free. Thus, the differential amplifier rejects the undesirable supply ripple.

Using Eqs. (8.48) and (8.49) and assuming α 1, so that iC1 iE1 and iC2 iE2 , we get 1 1 v o = IRC − 1 + e−v id /VT 1 + ev id /VT 1 1 − =5 1 + e−v id /VT 1 + ev id /VT This relationship can be used to obtain the data in the table on the next page

8.41 Refer to Fig. 8.14. I (a) VCMmax = VCC − RC 2 (b) For VCC = 2 V and VCMmax = 1 V, 1 1 = 2 − (IRC ) 2 ⇒ IRC = 2 V (c) IB =

I/2 ≤ 2 µA β +1

I ≤ 2 × 101 × 2 = 404 µA Select I = 0.4 mA

The figure shows v o versus v id and the gain versus v id . Observe that the transfer characteristic is nearly linear and the gain is nearly constant for v id ≤ 10 mV. As v id increases, the transfer

then RC =

2 = 5 k 0.4

This table belongs to Problem 8.42.

v id (mV) 2 5 8 10 20 30 40 iE1 /v id (V−1 ) 9.99 9.97 9.92 9.87 9.50 8.95 8.30 I

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–16

This table belongs to Problem 8.43. v id (mV)

2

v o (V)

0.2

0.498 0.987 1.457 1.90 2.311 2.685 3.022 3.320

100

99.7

Gain =

v od v id

5

10

15

98.7

97.1

20

95.0

25

92.4

30

35

89.5

40

86.3

83.0

characteristic bends and the gain is reduced. However, for v id even as large as 20 mV, the gain is only 5% below its ideal value of 100.

8.45 (a) VBE = 0.69 + 0.025 ln

8.44

(b) Using Eq. (8.50), we obtain

0.1 1

= 0.632 V

iC1 = αiE1

I 1 + e−v id /VT

For v id = 20 mV,

0.5 mA B1

1 mA Q1

B2

Q2 2

1.5 mA

iC1 =

200 µA = 138 µA 1 + e−20/25

iC2 = 200 − 138 = 62 µA (c) For v id = 200 mV while iC1 = 138 µA and iC2 = 62 µA: Since iC1 and iC2 have not changed, v BE1 and v BE2 also would not change. Thus, v B1 − v B2 = v BE1 + iE1 Re − iE2 Re − v BE2 = (v BE1 − v BE2 ) + Re (iE1 − iE2 ) 200 = 20 + Re (iC1 − iC2 ) = 20 + Re (138 − 62)

Since Q2 has twice the EBJ area of Q1 , the 1.5-mA bias current will split in the same ratio, that is,

v id = 20 mV → iC1 − iC2 = 76 µA

Thus, iE2 = 1 mA and iE1 = 0.5 mA To equalize the collector currents, we apply a signal v id = v B1 − v B2 IS (v B1 −v E )/VT e (1) α 2IS (v B2 −v E )/VT (2) iC2 = e α where we have denoted the scale current of Q1 by IS and that of Q2 as 2IS . Dividing (1) by (2), we get

For iC1 = iC2 , we obtain

76 µA = 3.8 mA/V 20 mV With Re ,

Gm =

v id = 200 mV → iC1 − iC2 = 76 µA 76 µA = 0.38 mA/V 200 mV Thus, the effective Gm has been reduced by a factor of 10, which is the same factor by which the allowable input signal has been increased while maintaining the same linearity.

Gm =

Now,

1 iC1 = e(v B1 −v B2 )/VT iC2 2

180 mV = 2.37 k 76 µA

(d) Without Re ,

iE2 = 2 iE2

iC1 =

⇒ Re =

25 mV = 100 0.25 mA The 0.2-V differential input signal appears across (2re + 2Re ), thus

8.46 (a) re =

200 mV = 0.2 mA 200 + 2 × 400

v B1 − v B2 = VT ln 2

ie =

= 25 ln 2 = 17.3 mV

v be = 0.2 × 100 = 20 mV

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–17

(b) The total emitter current in one transistor is I + ie = 0.45 mA and in the other transistor 2 I − ie = 0.05 mA. 2 (c) At one collector the signal voltage is −αie RC −ie RC = −0.2 × 10 = −2 V and at the other collector the signal voltage is +2 V. (d) Voltage gain =

4V = 20 V/V 0.2 V

60 = 4 × RC

(3)

Since vˆid 2 to keep Q1 in the active mode, v B1 = VCMmax +

Thus,

vˆid vˆid VCMmax + = 0.4 + VCC − Ad VT + 2 2 vˆid − ⇒ VCMmax = VCC + 0.4 − 2 vˆid Q.E.D. Ad VT + 2

(4)

(b) VCC = 2.5 V, vˆid = 10 mV, Ad = 50 V/V,

⇒ RC = 15 k

VCMmax = 2.5 + 0.4 − 0.005 − 50(25 + 5) × 10−3

I VC1 = VC2 = VCC − RC 2 0.2 × 15 = 2.5 − 2 = +1 V

1.4 V vˆod = Ad × vˆid = 50 × 10 = 500 mV = 0.5 V Using Eq. (2), we obtain

β gm

IRC = 2Ad VT = 2 × 50 × 0.025

100 = 50 k 4 (c) v od = Ad × v id

= 2.5 V

=2×

To limit the power dissipation in the quiescent state to 1 mV, the bias current must be limited to

= 60 × 10 = 600 mV = 0.6 V Thus, there will be ±0.3 V signal swing at each collector. That is, the voltage at each collector will range between 0.7 V and +1.3 V.

I=

1 Pmax = = 0.2 mA VCC + VEE 5

Using this value for I, we get

VCMmax = 0.4 + v Cmin

2.5 = 12.5 k 0.2 (c) To obtain VCMmax = 1 V, we use Eq. (4) to determine the allowable value of Ad ,

= 0.4 + 0.7 = 1.1 V

1 = 2.5 + 0.4 − 0.005 − Ad (25 + 5) × 10−3

(d) To maintain the BJT in the active mode at all times, the maximum allowable VCM is limited to

8.48 (a) Consider transistor Q1 , vˆid I v C1min = (VCC − RC ) − Ad 2 2 where Ad = gm RC

(2)

v B1 ≤ 0.4 + v C1min

8.47 (a) The maximum allowable value of the bias current I is found as 1 mW P = = 0.2 mA I= (VCC + VEE ) 5V We choose to operate at this value of I. Thus IC α(0.2/2) 4 mA/V = gm = VT 0.025 Ad = gm RC

(b) Rid = 2rπ = 2

IRC = Ad VT 2 Substituting from (2) into (1), we obtain vˆid v C1min = VCC − Ad VT + 2

I/2 RC VT

IRC = 2VT Thus,

Sedra, Smith, Carusone, Gaudet

RC =

⇒ Ad = 63.2 V/V (1)

Thus, by reducing VCMmax from 1.4 V to 1 V, we are able to increase the differential gain from 50 V/V to 63.2 V/V. 8.49 Ad = gm RC =

IC RC VT

(I/2) RC VT

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–18

=

IRC 2VT

Ad =

=

4 = 80 V/V 2 × 0.025

=

VA I VA = × 2VT I 2VT

=

20 = 400 V/V 2 × 0.025

I VC1 = VC2 = VCC − RC 2 =5−2=3V

I 2VT

2VA 2VA I I

v C1 = 3 − 80 × 0.005 sin(ωt)

8.51 Refer to Fig. P8.51(a).

= 3 − 0.4 sin(ωt)

I Re = 3VT 2 6VT ⇒ Re = I I RC = 80VT α 2

v C2 = 3 + 0.4 sin(ωt) v O = v C2 − v C1 = 0.8 sin(ωt) The waveforms are sketched in the figure below.

(1)

160VT αI Total resistance in collector circuit Ad = α Total resistance in emitter circuit 2RC RC =α Ad = α 2re + 2Re r e + Re

RC =

(2)

Substituting for RC from (2), for Re from (1), and for re = VT /(I/2), we obtain Ad = =

α(160VT /αI) (2VT /I) + (6VT /I)

160 = 20 V/V 2+6

8.52 See figure on next page. v id Rid = v sig Rid + Rsig

(1)

where Rid = (β + 1)(2re + 2Re ) thus, 2(β + 1)(re + Re ) v id = v sig 2(β + 1)(re + Re ) + Rsig

(2)

v od α × Total resistance between collectors = v id Total resistance in emitters = 8.50 See figure on next page. The circuit together with its equivalent half-circuit is shown in the figure. Ad = gm1,2 (ro1,2 ro3,4 )

2αRC 2re + 2Re

v od αRC = v id r e + Re Using (2) and (3), we get Gv ≡

For ro1,2 = ro3,4 = gm1,2 =

VA 2VA α(I/2) I

IC1,2 I VT 2VT

Sedra, Smith, Carusone, Gaudet

(3)

v od 2α(β + 1)RC = v sig 2(β + 1)(re + Re ) + Rsig

Since α = Gv =

β , α(β + 1) = β, we have β +1

2βRC 2(β + 1)(re + Re ) + Rsig

Microelectronic Circuits, 8th International Edition

(4)

© Oxford University Press 2021

Chapter 8–19

This figure belongs to Problem 8.50.

VCC Q3

VBIAS

Q3

Q4 vod

vid VCM 2

vod /2

Q2

Q1

VCM

vid 2

vid 2

Q1 Biased at I/2 Equivalent half-circuit

I

VEE This figure belongs to Problem 8.52.

VCC

RC

RC

vod

Rsig/2

Rsig/2 Q1

vsig 2

Q2

Re

VCM

vsig 2

Re I

VEE

If v id = 0.5 v sig , then from (1) we obtain

Thus the gain increases from approximately 2 1 RC /(re + Re ) to RC /(re + Re ). 2 3

Rid = Rsig Substituting for Rsig = Rid = 2(β + 1)(re + Re ) into Eq. (4) gives Gv =

1 αRC 2βRC = 4(β + 1)(re + Re ) 2 r e + Re

(5)

If β is doubled to 2β while Rsig remains at its old value, we get Rsig = 2(β + 1)(re + Re )

(6)

then the new value of Gv is obtained by replacing β by 2β in Eq. (4) and substituting for Rsig from (6): Gv =

4βRC 2(2β + 1)(re + Re ) + 2(β + 1)(re + Re )

4RC 2 RC = 6(re + Re ) 3 r e + Re

Sedra, Smith, Carusone, Gaudet

8.53 Refer to Fig. P8.53. vo α × Total resistance in collectors = vi Total resistance in emitters =

0.99 × 20 2re + 2 × 0.25

where VT 25 mV re = = = 250 IE 0.1 mA Thus, 0.99 × 20 vo 20 V/V = vi 2 × 0.25 + 2 × 0.25 Rin = (β + 1)(2re + 2Re )

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–20

= 2 × 101 × (0.25 + 0.25)

CMRR =

= 101 k 8.54 (a) Refer to the circuit in Fig. P8.54. As a differential amplifier, the voltage gain is found from α × Total resistance in collectors vo = vi Total resistance in emitters =

α × RC 2re

=

αRC 2re

| Ad | = 2 × 104 or 86 dB | Acm |

8.56 | Acm | =

RD 2RSS

(W/L) W/L

| Ad | = gm RD CMRR =

| Ad | = 2gm RSS | Acm |

(W/L) W/L

where gm =

(b) The circuit in Fig. P8.54 can be considered as the cascade connection of an emitter follower Q1 (biased at an emitter current I/2) and a common-gate amplifier Q2 (also biased at an emitter current of I/2). Referring to the figure below:

2ID 2(0.1/2) = = 0.5 mA/V VOV 0.2

For CMRR of 80 dB, the CMRR is 104 ; thus 104 = 2 × 0.5 × RSS /0.02 RSS = 200 k For the current source transistor to have ro = 200 k, VA × L 0.1 mA 200 × 0.1 L= = 4 µm 5 200 =

RC vo Q1 vi

Q2 re1 Rin2 re2

8.57 Refer to Fig. P8.57. vo = (a) v id Total resistance across which v o appears α Total resistance in the emitter 2 k =α× re1 + re2 VT , where IE is IE the dc emitter current of each of Q1 and Q2 , we use To determine re1 = re2 = re =

v e1,2 re2 1 = = vi re1 + re2 2 vo αRC = v e1,2 re2

VE = VB − VBE = 0 − 0.7

Thus,

= −0.7 V

vo 1 αRC αRC = × = vi 2 re2 2re

−0.7 − (−5) = 1 mA 4.3 IE = 0.5 mA 2IE =

which is identical to the expression found in (a) above. 8.55 gm = =

√

2μn Cox

W ID L

2 × 5 × 0.1 = 1 mA/V

(b)

| Ad | = gm RD = 1 × 12 = 12 V/V RD RD | Acm | = 2RSS RD =

12 × 0.01 = 6 × 10−4 V/V 2 × 100

Sedra, Smith, Carusone, Gaudet

25 mV = 50 0.5 mA vo 2 k 20 V/V =α× v id 0.1 k

re1 = re2 =

The common-mode half-circuit is shown in the figure, α × 2 k vo =− v icm (0.05 + 8.6) k −0.23 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–21

vo

v

icm

= 0.23 V/V

|v o /v id | 20 = = 86.5 |v o /v icm | 0.23

(c) CMRR = or 38.7 dB

(d) v o = −0.023 sin 2π × 60t + 0.2 sin 2π × 1000 t volts 8.58 (a) gm =

IC 0.1 mA = 4 mA/V VT 0.025 V

Ad = gm RC = 4 × 25 = 100 V/V β 100 =2× = 50 k gm 4 RC RC (c) | Acm | = 2REE RC

(b) Rid = 2rπ = 2

=

25 × 0.01 2 × 500 | Ad | 100 = 4 × 105 = | Acm | 2.5 × 10−4

or 112 dB VA 100 = 1000 k (e) ro = IC 0.1 Ricm βREE

= 100 × 500

α × Total resistance in collector Total resistance in emitter 10 k 10 k (50 + 150) Ad =

= 2.5 × 10−4 V/V (d) CMRR =

25 mV = 50 0.5 mA VA 100 = 200 k = ro = IC 0.5 re =

1 + (RC /βro ) RC + 2REE 1+ ro

5 = 25 V/V 0.2 We have neglected ro because its equivalent value at the output will be ro [1 + (Re /re )] = 200[1 + (150/50)] = 800 k which is much greater than the effective load resistance of 5 k. =

Rid = 2 × (β + 1)(50 + 150 )

1 + (25/(100 × 1000)) 25 + 1000 1+ 1000

25 M

= 2 × 101 × 0.2 (k) = 40.4 k RC RC | Acm | 2RSS RC 10 × 0.02 = 0.001 V/V 200 To obtain Ricm , we use Eq. (8.98): | Acm | =

8.59 Figure (a) shows the differential half-circuit. IE = 0.5 mA, gm =

IC = αIE 0.5 mA

IC 0.5 mA = = 20 mA/V VT 0.025 V

Sedra, Smith, Carusone, Gaudet

Ricm βREE

1 + (RC /βro ) RC + 2REE 1+ ro

where 2REE = 200 k, thus REE = 100 k and

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–22

Ricm = 100 × 100

1 + (10/(100 × 200)) 10 + 200 1+ 200

= 4.88 M

RD ΔVD1

8.60 Refer to Fig. P8.2.

ΔVCM

W 1 |VOV |2 ID = 0.25 mA = μp Cox 2 L

Q1

1/gm

2RSS 40 k

1 × 4 × |VOV |2 2 ⇒ |VOV | = 0.353 V 0.25 =

gm =

2 × 0.25 2ID = = 1.416 mA/V |VOV | 0.353

| Ad | = gm RD = 1.416 × 4 = 5.67 V/V RD RD | Acm | = 2RSS RD =

4 × 0.02 2 × 30

= 1.33 × 10−3 V/V

Substituting VCM = 0.94, Vt = 0.4 V, VD1 = 0.93 V, and VD1 = −2.26VCM results in

8.61 Refer to Fig. P8.61. (a) Assume v id = 0 and the two sides of the differential amplifier are matched. Thus,

0.94 + VCM = 0.4 + 0.93 − 2.26VCM ⇒ VCM = 0.12 V With this change, VCM = 1.06 V and VD1,2 = 0.66 V; thus VCM = Vt + VD1,2 .

ID1 = ID2 = 0.01 mA W 1 2 VOV ID1,2 = μn Cox 2 L

⇒ VOV

VD1 107 = −2.26 V/V =− 1 VCM + 40 0.14 (e) For Q1 and Q2 to enter the triode region VCM + VCM = Vt + VD1 + VD1

CMRR = 4252.5 or 72.6 dB

0.01 =

VD1 RD =− 1 VCM + 2 RSS gm

8.62 The new deliberate mismatch RD /RD cancels the two existing mismatch terms in the expression for Acm given in the problem statement so as to reduce Acm to zero. Thus,

1 2 × 1 × VOV 2 = 0.14 V

VCM = VGS + 0.02 mA × RSS

RD RD × = −0.002 2RSS RD

= Vt + VOV + 0.02 × RSS

RD 5 × = −0.002 2 × 25 RD

= 0.4 + 0.14 + 0.4 = 0.94 V 2ID 2 × 0.01 = 0.14 mA/V = (b) gm = VOV 0.14 Ad = gm RD

⇒

RD = −0.02 or − 2% RD

(Note the sign of the change is usually determined experimentally.)

15 = 0.14 × RD ⇒ RD = 107 k

8.63 It is required to raise the CMRR by 40 dB, that is, by a factor of 100. Thus, the cascoding of the bias current source must raise its output resistance RSS by a factor of 100. Thus the cascode transistor must have A0 = 100. Since

(c) VD1 = VD2 = VDD − ID RD = 2 − 0.01 × 107 = 0.93 V (d) The figure shows the common-mode half-circuit,

Sedra, Smith, Carusone, Gaudet

A0 = gm ro =

2I VA 2VA = VOV I VOV

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–23

1 βro 2 where ro is that of the transistors in the Wilson mirror, then

2VA 0.2 ⇒ VA = 10 V

=

100 =

VA = VA × L ⇒ L = 2 µm

20 VA = = 100 k I 0.2 For the transistors in the differential pair, we have

8.64 REE =

ro =

20 VA = = 200 k I/2 0.1

Ricm βREE

= 1 × 10−4 V/V CMRR =

8.66 See figure on next page. v be1 = 2.5 sin(ωt), mV and v be2 = −2.5 sin(ωt), mV I RC − gm RC × 2.5 × 10−3 sin(ωt) v C1 VCC − 2

50 × 100 = 2.5 M 2 × 100 1+ 200

where gm =

8.65 For the differential-pair transistors, we have IC 0.1 mA

I/2 I mA = VT 0.05 V

Thus, I I ×10×2.5×10−3 sin(ωt) v C1 = 5− ×10− 2 0.05

0.1 = 4 mA/V 0.025 VA 20 = 200 k = ro = IC 0.1 gm =

= 5 − 5I − 0.5I sin(ωt) (V) Similarly,

(a) Ad = gm RC = 4 × 10 = 40 V/V where we have neglected the effect of ro since r o RC . (b) If the bias current is realized using a simple current source, REE = ro | current source =

40 = 4 × 105 1 × 10−4

or 112 dB

1 + (RC /βro ) RC + 2REE 1+ ro

For RC ro , 2REE 1+ Ricm βREE ro =

20 = 100 k 0.2 1 REE = × 100 × 100 = 5 M 2 10 × 0.1 | Acm | = 2 × 5, 000 ro =

10 = 5 × L

20 VA = = 100 k I 0.2

v C2 = 5 − 5I + 0.5I sin(ωt) (V) To ensure operation in the active mode at all times with v CB = 0 V, we use v C1min = 0.005 5 − 5.5I = 0.005

RC RC 2REE RC 10 = × 0.1 2 × 100

⇒ I 0.9 mA

= 5 × 10−3 V/V

Ad = gm RC = 18 × 10 = 180 V/V

| Acm | =

With this value of bias current, we obtain gm =

40 | Ad | = = 8 × 103 CMRR = | Acm | 5 × 10−3 or 78 dB (c) If the bias current I is generated using a Wilson mirror, REE = Ro | Wilson mirror

Sedra, Smith, Carusone, Gaudet

0.9 = 18 mA/V 0.05

At each collector there will be a sine wave of 180 × 2.5 = 450 mV = 0.45 V amplitude superimposed on the dc bias voltage of 5 − 0.45 × 10 = 0.5 V. Between the two collectors there will be a sine wave with 0.9 V peak amplitude. The figure illustrates the waveforms obtained.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–24

This figure belongs to Problem 8.66.

8.67

v o1 v o2 = −100 V/V = +100 V/V v id v id

v o1,2 = −0.1 V/V v icm Rid = 10 k I = 2 mA IE1 = IE2 = 1 mA re1 = re2 = 25

gm1 = gm2 = 40 mA/V rπ 1 = rπ 2 =

β 100 = 2.5 k = gm 40

Since Rid > rπ , we need emitter resistances, as shown in the figure. VCC RC

RC

vo1

vo2 Q1

Q2 Re

Re

vid

I 2 mA

Rid 10 k

10 = (β + 1) (2re + 2Re ) r e + Re =

10 50 2 × 101

Re = 25 αRC v o1 =− v id 2(re + Re ) −αRC −100 = 2(0.025 + 0.025) ⇒ RC 10 k To allow for ±2 V swing at each collector, I VCC − RC − 2 ≥ 0 2 assuming that VCM = 0 V. Thus, 2 VCC = × 10 + 2 = 12 V 2 We can use VCC = 15 V to allow for VICM as high as +3 V. RC | Acm | (to each collector) 2REE For | Acm | = 0.1, 10 0.1 = 2REE ⇒ REE = 50 k

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–25

This is the minimum value of Ro of the bias current source. If the current source is realized by a simple current mirror, we obtain REE = ro =

VBE ’s. Thus their currents must be related by the ratio of their scale currents IS , which are proportional to the junction areas.

VA I

Thus, VA 2 ⇒ VA = 100 V

RC

50 =

vo1

The common-mode input resistance is Ricm

vicm

vicm REE

100 VA = = 100 k ro = I/2 1

8.68 If the output is taken single-endedly, then RC | Acm | = 2REE

CMRRd = 2gm REE /

ie1 =

2 v icm 3 REE

RC RC

v o1 = −ic1 RC −ie1 RC = −

2 RC v icm 3 REE

and v o2 = −

1 RC v icm 3 REE

With the output taken differentially, we have v o = v o2 − v o1 =

CMRRd 2 = CMRRs RC /RC

Acm =

2 = 46 dB RC /RC

RC = 0.01 = 1% RC

8.69 If Q1 has twice the base-emitter junction 2 area of Q2 , the bias current I will split I in Q1 3 1 and I in Q2 . This is because with B1 and B2 3 grounded the two transistors will have equal

Sedra, Smith, Carusone, Gaudet

1 v icm 3 REE

Thus,

Thus,

⇒

REE

With a common-mode input signal v icm applied, as shown in the figure, the current (v icm /REE ) will split between Q1 and Q2 in the same ratio as that of their base-emitter junction areas, thus

ie2 =

If the output is taken differentially, then RC RC | Acm | = 2REE RC

20 log

ie2

and

1 | Ad | = gm RC 2 | Acm | = gm REE CMRRs = | Ad |

| Ad | = gm RC

vicm

ve vicm

where ro is the output resistance of each of Q1 and Q2 ,

= 2.4 M

vo2 Q2

ie1

2

1 + (10/(100 × 100)) 10 + 100 1+ 100

vo

RC

iC2

Q1

1 + RC /βro βREE RC + 2REE 1+ ro

Ricm = 100 × 50

iC1

1 RC v icm 3 REE

15 1 RC 1 = 0.0125 V/V = × 3 REE 3 400

8.70 (a) gm = 2k n ID = k n I Ad = gm RD = k n I RD

I/2 I

= VOV = 1 kn kn 2 RD VOV VOS = 2 RD

(1)

Thus,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–26

VOS =

1 I/k n 2

RD RD

(2)

Equations (1) and (2) can be combined to obtain the following relationship between Ad and VOS : 2kn RD VOS Ad = (3) RD /RD For the numerical values given, Eq. (3) becomes Ad = 4 × 103 VOS

(4)

(b) For each value of VOS we use Eq. (2) to determine I and then Eq. (4) to determine Ad . The results are as follows: VOS (mV) I (mA)

1

2

3

4

5

0.04 0.16 0.36 0.64 1.00

Ad (V/V)

4

8

12

16

20

We observe that by accepting a larger offset we are able to obtain a higher gain. Observe that the gain realized is proportional to the offset voltage one is willing to accept. Also, observe that the bias current required increases by a factor equal to the square of the increase in gain. 8.71 The offset voltage due to Vt is VOS = ±5 mV The offset voltage due to RD is RD 0.3 VOV × 0.02 = 3 mV = VOS = 2 RD 2

= 25 × 0.1 = 2.5 mV 8.73 VOS = VT

IS IS

= 25 × 0.1 = 2.5 mV 8.74 With a resistance RE connected in the emitter of each transistor, we follow a development similar to that used for the MOSFET and the BJT differential amplifiers except that we α . Thus, replace gm by Gm = r e + RE RC αI/2 VOS = Gm RC RC I = (re + RE ) 2 RC VT RC I = + RE 2 I/2 RC which yields, RC 1 VOS = VT + IRE 2 RC

8.75 Refer to Fig. P8.75. (a) RC1 = 1.04 × 5 = 5.20 k RC2 = 0.96 × 5 = 4.80 k To equalize the total resistance in each collector, we adjust the potentiometer so that RC1 + x × 1 k = RC2 + (1 − x) × 1 k

The offset voltage due to (W/L) is 0.3 VOV (W/L) VOS = = × 0.02 = 3 mV 2 (W/L) 2

⇒ x = 0.3 k

The worst-case offset voltage will be when all three components add up,

(b) If the area of Q1 and hence IS1 is 5% larger than nominal, then we have

VOS = 5 + 3 + 3 = 11 mV

IS1 = 1.05IS

The major contribution to the total is the variability of Vt .

and the area of Q2 and hence IS2 is 5% smaller than nominal,

To compensate for a total offset of 11 mV by appropriately varying RD , we need to change RD by RD obtained from RD VOV × 11 mV = 2 RD

IS2 = 0.95IS

⇒

RD 11 × 2 = 0.0733 = RD 300

or 7.33% 8.72 VOS = VT

RC RC

Sedra, Smith, Carusone, Gaudet

5.2 + x = 4.8 + 1 − x

Thus, IE1 = 0.5 × 1.05 = 0.525 mA IE2 = 0.5 × 0.95 = 0.475 mA Assuming α 1, we obtain IC1 = 0.525 mA IC2 = 0.475 mA To reduce the resulting offset to zero, we adjust the potentiometer so that VC1 = VC2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–27

⇒ VCC − (RC1 + x)IC1 = VCC − (RC2 + 1 − x)IC2 IC1 (RC1 + x) = IC2 (RC2 + 1 − x) 0.525(5 + x) = 0.475(5 + 1 − x) ⇒ x = 0.225 8.76 IBmax IBmin =

400 2.5 µA = 2 × 81

400 = 1 µA 2 × 201

IOSmax =

200 200 − 1.5 µA 81 201

(W/L) = 0.04 ⇒ VOS = (W/L)

VOV 2

(W/L) (W/L)

0.224 × 0.04 = 4.5 mV 2 Vt = 5 mV ⇒ VOS = Vt = 5 mV

=

Worst-case VOS = 4.5 + 4.5 + 5 = 14 mV If the three components are independent, VOS = 4.52 + 4.52 + 52 = 8.1 mV

8.79

8.77 gm = 2 k n (W/L)ID = k n (W/L)I

VCC RC

RC

Ad = gm RD

VO

(a1I/2)

2 ID I = gm gm RD VOV = 2 RD

VOV = VOS

(a2I/2)

Q1

Q2 I 2

I 2

For I = 160 µA, we have √ gm = 4 × 0.16 = 0.8 mA/V

I

Ad = 0.8 × 10 = 8 V/V 0.16 = 0.2 V 0.8 0.2 × 0.02 = 2 mV VOS = 2 For I = 360 µA, we have √ gm = 4 × 0.36 = 1.2 mA/V VOV =

The current I splits equally between the two emitters. However, the unequal β’s will mean unequal α’s. Thus, the two collector currents will be unequal, IC1 = α1 I/2

Ad = 1.2 × 10 = 12 V/V 0.36 = 0.3 V 1.2 0.3 × 0.02 = 3 mV VOS = 2 Thus by increasing the bias current, both the gain and the offset voltage increase, and by the same factor (1.5). VOV =

8.78 VOV

I/2 I

= = 1 k n (W/L) k (W/L) 2 n

0.1 = 0.224 V 0.2 × 10 RD RD VOV = 0.04 ⇒ VOS = RD 2 RD

=

0.224 × 0.04 = 4.5 mV = 2

Sedra, Smith, Carusone, Gaudet

IC2 = α2 I/2 and the collector voltages will be unequal, VC1 = VCC − α1 (I/2)RC VC2 = VCC − α2 (I/2)RC Thus a differential output voltage VO develops: VO = VC2 − VC1 1 IRC (α1 − α2 ) 2 The input offset voltage VOS can be obtained by dividing VO by the differential gain Ad : =

Ad = gm RC

I/2 IRC RC = VT 2VT

Thus, VOS = VT (α1 − α2 ) Substituting, we obtain

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–28

α1 =

β1 β1 + 1

IC1

and β2 β2 + 1 β1 β2 = VT − β1 + 1 β2 + 1

α2 = VOS

For IC1

αI = 2

VCE1 VA1 VCE1 VCE2 1+ + 2VA1 2VA2 1+

VCE1 VCE2 1 and 1 we have VA1 VA2 αI 1 VCE2 1 VCE1 − 1+ 2 2 VA1 2 VA2 1 VCE2 αI 1 VCE1 1+ − 2 2 VA2 2 VA1

β1 β2 + β 1 − β 1 β2 − β 2 = VT (β1 + 1)(β2 + 1)

IC2

β1 − β 2 = VT (β1 + 1)(β2 + 1)

We can express IC1 and IC2 as

β1 − β 2 β1 β2 1 1 = VT − β2 β1

VT

αI (1 + ) 2 αI (1 − ) = 2

IC1 = Q.E.D.

For β1 = 50 and β2 = 100, we have 1 1 − = −0.25 mV VOS = 25 100 50

IC2

where =

1 VCE1 VCE2 ( − ) 2 VA1 VA2

(1)

With VOS applied, 8.80 For the MOS amplifier: RD VOV VOS = 2 RD

200 × 0.04 2 = 4 mV

αI 2

= gm (VOS /2)

=

Substituting gm =

For the BJT amplifier: RC VOS = VT RC

VOS = 2VT

= 25 × 0.04 = 1 mV If in the MOS amplifier the width of each device is increased by a factor of 4 while the bias current is kept constant, VOV will be reduced by a factor of 2. Thus VOS becomes VOS = 2 mV 8.81 Since the only difference between the two sides of the differential pair is the mismatch in VA , we can write VCE1 IC1 = IC 1 + VA1 VCE2 IC2 = IC 1 + VA2 IC1 + IC2 = αI VCE1 VCE2 = αI + IC 2 + VA1 VA2 VCE1 VCE2 ⇒ IC = αI 2+ + VA1 VA2

Sedra, Smith, Carusone, Gaudet

αI/2 gives VT

Substituting for from Eq. (1) gives VCE1 VCE2 − VOS = VT VA1 VA2 As a first-order approximation, we can assume VCE1 VCE2 = 10 V and substitute VA1 = 100 V and VA2 = 200 V to determine VOS as 10 10 − VOS = 25 100 200 = 25 × 0.05 = 1.25 mV

8.82 Refer to the figure and note that, IB1 = IB − IB IB2 = IB + IB where IB =

I 2(β + 1)

Microelectronic Circuits, 8th International Edition

(1)

© Oxford University Press 2021

Chapter 8–29

VCC RC

Rs Rs 2

8.83

RC

VO

Q1

Q2

RC

IB1

IB2

I I 2 2

VCC a2I 3

Rs Rs 2

VC1

I I 2 2

I

VO

aI 3

RC

Q1

VEE

Q2 I 3

2I 3

2

VC2

I

IB =

I 2(β + 1)

(2)

Equating the incremental changes in the voltage between each grounded base and emitter, we write I Rs Rs ) − IB Rs − IB ( ) − ( )re IB ( 2 2 2 I Rs Rs ) + IB Rs + IB ( ) + ( )re = IB (− 2 2 2 Substituting for IB from Eq. (1) and for IB from Eq. (2) and collecting terms gives I =

I Rs 2(β + 1) re +

+

VO = IRC

I gm

re1

1 I Rs Rs 2(β + 1)gm re + β+1 α β = , re (β + 1)re

From Fig. (a) we see that the transistor with twice the area (Q1 ) will carry twice the current in the other transistor (Q2 ). Thus 2I , 3 α2I IC1 = , 3 Thus,

IE2 =

I 3

IC2 =

αI 2

VO = VC2 − VC1

1 I Rs Rs 2β 1 + (β+1)r e

1 αIRC 3 To reduce this output voltage to zero, we apply a dc input voltage v id in the direction shown in Fig. (b). The voltage v id is required to produce v od in the direction shown which is opposite in direction to VO and of course |v od | = |VO |, thus VO =

β , thus gm I gm Rs Rs / 1 + = 2β β

Sedra, Smith, Carusone, Gaudet

re2

α2I RC 3 αI VC2 = VCC − RC 3 and the dc offset voltage at the output will be

But, (β + 1)re = rπ = VOS

vid

Q2

VC1 = VCC −

Using Eq. (3) gives

VOS =

Q1

IE1 =

The input offset voltage VOS is obtained by dividing VO by Ad = gm RC , thus

Substituting gm =

Vod

(b)

The offset voltage at the output is

VOS =

RC

Rs 2(β+1)

Neglecting the third term in the denoniminator of the second factor, 1 I Rs (3) I Rs 2(β + 1) re + β+1

VOS =

RC

1 Rs β+1

(a)

Q.E.D.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–30

1 αIRC 3 The gain Ad is found as follows:

Ad v id =

Ad = =

(1)

α × Total resistance in collectors Total resistance in emitters

α × 2RC re1 + re2

8.85 Gm = gm1,2 =

re1 =

3VT 1.5VT VT VT = = = IE1 2I/3 2I I

re2 =

VT VT 3VT = = IE2 I/3 I

For ro2 = ro4 =

2αRC 2αIRC = 4.5 VT /I 4.5 VT

(2)

Substituting in Eq. (1) gives

2 × 5 × 0.5 5 = I I 2.5 1 5 Ro = × = 2 I I Thus, Ad = Gm Ro =

v id = 0.75 VT = 18.75 mV Now, using large signal analysis:

8.86

v id = VB2 − VB1 = (VB2 − VE ) − (VB1 − VE ) IC1 = IS1 e(VB1 −VE )/VT

(3)

IC2 = IS2 e(VB2 −VE )/VT

(4)

where IS1 = 2 IS2 .

2.5 I × = 10 V/V 0.25 I

W I 1 2 = μn Cox VOV 2 2 L

1 2 × 0.4 × 20 × VOV 2 ⇒ VOV = 0.16 V 0.1 =

gm1,2 =

To make IC1 = IC2 ,

2 × (I/2) 2 × 0.1 = 1.25 mA/V = VOV 0.16

IS1 e(VB1 −VE )/VT = IS2 e(VB2 −VE )/VT

ro2 = ro4 =

e(VB2 −VB1 )/VT = 2

= 30 k

|V | × L 6 × 0.5 |VA | = A = I/2 I/2 0.1

Ad = gm1,2 (ro2 ro4 )

VB2 − VB1 = VT ln 2

= 1.25 × (30 30)

Thus,

= 18.8 V/V

v id = 17.3 mV which is reasonably close to the approximate value obtained using small-signal analysis. 8.84 A 2-mV input offset voltage corresponds to a difference RC between the two collector resistances,

8.87 Ad = gm1,2 (ro2 ro4 ) W ID gm1,2 = 2k n L √ √ = 4I = 2 I ro2 = ro4 =

RC RC

2|VA | 2×5 10 |VA | = = = I/2 I I I

√ 10 1 10 = √ Ad = 2 I × × 2 I I

RC 20 ⇒ RC = 1.6 k = 25 ×

Thus a 2-mV offset can be nulled out by adjusting one of the collector resistances by 1.6 k. If the adjustment mechanism raises one RC and lowers the other, then each need to be adjusted by only (1.6 k/2) = 0.8 k.

Sedra, Smith, Carusone, Gaudet

|V |L |VA | = A I/2 I/2

=

thus,

2 = VT

2(I/2) I I = = VOV VOV 0.25

Ro = ro2 ro4

where

Ad =

If a potentiometer is used (as in Fig. P8.84), the total resistance of the potentiometer must be at least 1.6 k. If specified to a single digit, we use 2 k.

10 20 = √ I ⇒ I = 0.25 mA 8.88 Gm = gm1,2 =

Microelectronic Circuits, 8th International Edition

2(I/2) 0.2 = 1 mA/V = VOV1,2 0.2

© Oxford University Press 2021

Chapter 8–31

ro2 =

VAn 20 = = 200 k I/2 0.1

ro4 =

|VAp | 12 = = 120 k I/2 0.1

Since all β and ro are equal, we obtain 1 Ro = (βro ) βro 2

Ro = ro2 ro4 = 200 120 = 75 k Ad = Gm Ro = 1 × 75 = 75 V/V The gain is reduced by a factor of 2 with RL = Ro = 75 k. 8.89 IE1 = IE2 = 0.25 mA

1 βro 3 and 1 Ad = βgm ro Q.E.D. 3 For β = 100 and VA = 20 V, we have =

gm ro =

IC1 = IC2 0.25 mA

IC VA VA 20 = 800 = = VT IC VT 0.025

1 × 100 × 800 = 2.67 × 104 V/V 3

gm1,2

IC1,2 0.25 mA = 10 mA/V = = VT 0.025 V

Ad =

ro =

|VA | 10 V = = 40 k IC 0.25 mA

8.91 (a) For Q1 and Q2 , 1 W I = μn Cox V2 2 2 L 1,2 OV W 1 0.1 = × 0.5 × × 0.152 2 L 1,2 W W ⇒ = = 17.8 L 1 L 2

Rid = 2 rπ = 2

β 100 = 20 k =2× gm 10

Ro = ro2 ro4 = 40 40 = 20 k Gm = gm1,2 = 10 mA/V Ad = Gm Ro = 10 × 20 = 200 V/V If RL = Rid = 20 k, then Gv = 200 × = 200 ×

For Q3 and Q4 , 1 W I = μp Cox |VOV |2 2 2 L 3,4 W 1 0.1 = × 0.125 × × 0.152 2 L 3,4 W W ⇒ = = 71.1 L 3 L 4

RL RL + Ro

20 = 100 V/V 20 + 20

8.90 Refer to Fig. P8.90. Ad = Gmd Ro

(b) Gmd = gm1,2 =

where Gmd = gm1,2

= 1.33 mA/V

I/2 VT

Ad = Gmd Ro

and

60 = 1.33 × Ro

Ro = Ro4 Ro7

⇒ Ro = 45 k

Here Ro4 is the output resistance of the cascode amplifier (looking into the collector of Q4 ), thus

But

Ro4 = gm4 ro4 (ro2 rπ 4 )

and the same VA ). Thus

Ro4 gm4 rπ 4 ro4 = β4 ro4 The resistance Ro7 is the output resistance of the Wilson mirror and is given by

Ro = (β4 ro4 )

Ro = ro2 ro4 and ro2 = ro4 (Q2 and Q4 have the same ID =

Usually rπ 4 ro2 ,

1 Ro7 = β7 ro7 2 Thus

2 (I/2) I 0.2 = = VOV VOV 0.15

ro2 = ro4 = 90 k = |VA | =

I 2

|VA | I/2

I × 90 k = 9 V 2

9 = |VA |L = 15 L

1 β7 ro7 2

Sedra, Smith, Carusone, Gaudet

⇒ L = 0.6 µm (c) v Omin = VCM − Vtn

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–32

= 0 − 0.4 = −0.4 V

Using Eq. (8.137),

v Omax = VDD − |VOV | = 0.8 − 0.15 = 0.65 V

Acm = −

Thus,

1 2gm3 RSS

1 = 0.0167 V/V 2 × 1 × 30 | Ad | 30 CMRR = = = 1800 | Acm | 0.0167 =−

−0.4 V ≤ v O ≤ 0.65 V |VA | 9 = = 45 k I 0.2 The CMRR can be obtained using Eq. (8.138):

(d) RSS =

or 65.1 dB

CMRR = (2gm1,2 RSS )[gm3 (ro ||ro3 )] = (gm RSS )(gm ro )

8.93 Gm = 2 mA/V

= (1.33 × 45)(1.33 × 90)

With RL = ∞,

= 7200 or 77 dB

Ad = Gm Ro and

8.92

v o = Gm Ro v id

Gmd = gm1,2 =

With RL = 20 k,

2(I/2) I = VOV VOV

0.2 = 1 mA/V = 0.2 | VA | 6 ro2 = ro4 = = = 60 k I/2 0.1 Ro = ro2 ro4 = 6060 = 30 k

RSS

Using Eq. (8.134), Rim =

1 1 ro3 = ro gm3 gm1,2

= Gm Ro

RL RL + Ro

20 1 v id = Gm Ro v id 20 + Ro 2

Thus, Ro = 20 k

Ad = Gmd Ro = 1 × 30 = 30 V/V | VA | 6 = = = 30 k I 0.2

v o = Gm Ro v id

Ad (with RL = ∞) = Gm Ro = 2 × 20 = 40 V/V 8.94 For Q1 , Q2 , Q3 and Q4 : W I 1 2 VOV = μn Cox 2 2 L 1 2 × 5 × VOV 2 ⇒ VOV = 0.2 V 0.1 =

1 60 k = 1 mA/V = 0.984 k

VGS = Vt + |VOV | = 0.5 + 0.2 = 0.7 V

Using Eq. (8.135),

For Q5 , Q6 , Q7 , and Q8 : Am =

1 1+

1 gm3 ro3

=

1 1+

1 1×60

ID = 0.2 mA 1 2 × 5 × VOV 2 ⇒ VOV = 0.28 V

= 0.9836 A/A

0.2 =

Using Eq. (8.136),

VGS = 0.5 + 0.28 = 0.78 V

Gmcm − =−

1 1 2RSS gm3 (ro ro3 )

1 1 × 2 × 30 1 × 0.5 × 60

= 5.56 × 10

−4

mA/V

Sedra, Smith, Carusone, Gaudet

From the figure we see that for each transistor to operate at VDS at least equal to VGS , the total power supply is given by VDD + VSS = VDS4 + VDS2 + VDS7 + VDS6 = VGS4 + VGS2 + VGS7 + VGS6

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–33

= 0.7 + 0.7 + 0.78 + 0.78

(b) Ad = gm1,2 (Ro4 Ro6 )

= 2.96 3.0 V

gm1,2 =

Ro6 = gm6 ro6 ro8

VDD

Q3

Since all transistors are operated at a bias current (I/2) and have the same overdrive voltage |VOV | and the same Early voltage, |VA |, all have the same gm = I/|VOV | and the same |VA | = 2|VA |/I. Thus, ro = I/2

Q4

Q1

2(I/2) I = VOV VOV

Ro6 = gm ro2

Q2

Ro4 = gm4 ro4 ro2 = gm ro2 Ad = gm (gm ro2 gm ro2 ) 1 (gm ro )2 2 2|VA | 2|VA | I × = gm ro = |VOV | I |VOV | =

I Q8

Q7

Ad = 2(|VA |/|VOV |)2

Q5

Q.E.D.

For |VOV | = 0.2 V and |VA | = 10 V, we have 10 2 = 5000 V/V Ad = 2 0.2

Q6

VSS 8.96 The currents i1 to i13 are shown on the circuit diagram. Observe that i11 = i7 = i3 (the current that enters a transistor exits at the other end!). Also observe that the mirror Q3 and Q4 is indeed functioning properly as the drain currents 1 of Q3 and Q4 are equal (i12 = i2 = gm v id ). 4 However, the currents in their ro ’s are far from being equal!

8.95

Q7

Q8

Q5

Q6 Ro6 Ro4

Q3

Q4 VBIAS Q1

Q2

I

(a) See figure.

Sedra, Smith, Carusone, Gaudet

vo

There are some inconsistencies that result from the approximations made to obtain the results shown in Fig. P8.96, namely, gm ro 1. Note for instance that although we find the current in ro of 1 Q2 to be gm v id , the voltages at the two ends of 2 1 ro are (gm ro )v id and v id /4; thus the current must 2 1 1 gm ro − ro , which is approximately be v id 2 4 1 gm v id . 2 The purpose of this problem is to show the huge imbalance that exists in this circuit. In fact, Q1 1 3 has |v gs | = v id while Q2 has |v gs | = v id . This 4 4 imbalance results from the fact that the current mirror is not a balanced load. Nevertheless, we know that this circuit provides a reasonably high common-mode rejection.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–34

This figure belongs to Problem 8.96.

i4

i5

1 W I = k n V2 2 2 L A OV 1 (W/L)A I 1+ = 2 2 (W/L)A

ID1 =

8.97

VDD

ID1

Q3 ID3

Q4

ID1 Q1

Similarly, we can show that 1 (W/L)A I ID2 = 1− 2 2 (W/L)A

ID4 IO

The current mirror causes

ID2

ID4 = ID3 = ID1

Q2

Thus, IO = ID4 − ID2 = ID1 − ID2

I

=

I (W/L)A 2 (W/L)A

The input offset voltage is

VSS

VOS =

(a) Let W W W 1 = + L 1 L A 2 L A W W W 1 = − L 2 L A 2 L A Q1 and Q2 have equal values of VGS and thus of VOV , thus W W 1 1 V2 ID1 = k n + 2 L A 2 L A OV W 1 (W/L)A 2 1 1+ VOV = k n 2 L A 2 (W/L)A

IO Gmd

where Gmd = gm1,2 =

2(I/2) I = VOV VOV

Thus, VOS = (VOV /2) (b) ID1 = ID2 =

(W/L)A (W/L)A

Q.E.D.

I 2

ID3 = ID1 If the (W/L) ratios of the mirror transistors have a mismatch (WL)M , the current transfer ratio of

Since, in the ideal case

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–35

the mirror will have an error of [(W/L)M /(W/L)M ]. Thus (W/L)M ID4 = ID3 1 + (W/L)M At the output node, we have IO = ID4 − ID2 (W/L)M = ID3 1 + − ID2 (W/L)M (W/L)M − ID2 = ID1 1 + (W/L)M =

i=

=

αI 2

αI βP

αI αI/2 − 2 2 1+ βP 2/βP 2 1+ βP

To reduce this output current to zero, an input voltage VOS has to be applied with a value of

I (W/L)M 2 (W/L)M

VOS = −

and the corresponding VOS will be IO IO VOS = = Gmd I/VOV VOV (W/L)M = 2 (W/L)M

8.99 Refer to Fig. P8.99,

Thus, VOS = −

Q.E.D.

0.2 × 0.02 = 2 mV (c) VOS |Q1 ,Q2 mismatch = 2 0.2 × 0.02 = 2 mV VOS |Q3 ,Q4 mismatch = 2

Worst-case VOS = 2 + 2 = 4 mV

i αI/βP =− Gmd (αI/2)/VT 2VT βP

Q.E.D.

For βP = 50, VOS = −

2 × 25 = −1 mV 50

If the simple current mirror is replaced by the Wilson current mirror, i =

αI αI/2 − 2 2 1+ 2 βP

which leads to i

8.98 Refer to Fig. P8.90.

αI βP2

(a) VB7 = +5 − VEB6 − VEB7 = 5 − 0.7 − 0.7

and the offset voltage becomes

= +3.6 V

VOS = −

v Omax = VB7 + 0.4 = +4 V (b) The dc bias voltage should be

a negligible quantity!

= 4 − 1.5 = +2.5 V (c) For v O to swing negatively (i.e., below the dc bias value of 2.5 V) by 1.5 V, that is, to +1 V with Q4 remaining in saturation, VBIAS should be VBIAS = v Omin + 0.4 = 1.4 V (d) With VBIAS = 1.4 V, the bias voltage at the collectors of Q1 and Q2 is

8.100 The overdrive voltage, |VOV |, at which Q1 and Q2 are operating is found from 1 I = k p (W/L)|VOV |2 2 2 1 0.1 = × 6.4 × |VOV |2 2 ⇒ |VOV | = 0.18 V Gmd = gm1,2 =

VC1,2 = VBIAS − VBE3,4 = 1.4 − 0.7 = +0.7 V

2(I/2) |VOV |

0.2 = 1.13 mA/V 0.18 |VAp | 10 = = 100 k ro2 = I/2 0.1 =

The upper limit on VCM is 0.4 V above VC1,2 :

Sedra, Smith, Carusone, Gaudet

which for βP = 50 gives VOS = −0.02 mV

VO = v Omax − 1.5

VCMmax = 0.7 + 0.4 = +1.1 V

2VT βP2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–36

ro4 =

|VAnpn | 30 = = 300 k I/2 0.1

For case (a),

Ro = ro2 ro4 = 100 k 300 k = 75 k

CMRR = 2 ×

Ad = Gmd Ro = 1.13 × 75 = 85 V/V

For case (b),

8.101 The CMRR is given by Eq. (8.138):

√ CMRR = 2 2

5 0.2

2

5 0.2

= 1250 or 62 dB 3 = 4.42 × 104

CMRR = (2gm1 RSS )[gm3 (ro ||ro3 )]]

or 93 dB

(a) Current source is implemented with a simple current mirror: |VA | RSS = ro | QS = I 2(I/2) I = gm1,2 = gm3 = VOV VOV

8.102 To determine Gmcm we use Eq. (8.133) and assume to be negligibly small, thus 1 Rim (Am + 1) (Am − 1) − Gmcm = 2RSS 2ro

ro2 = ro3

|VA | 2|VA | = ro4 = = I/2 I

Thus, I 2I |VA | 1 2|VA | × × × × VOV I 2 VOV I 2 VA =2 Q.E.D. VOV CMRR =

(b) Current source is implemented with the modified Wilson mirror in Fig. P8.101: RSS = gm7 ro7 ro9 Transistor Q7 has the same k (W/L) as Q1 and Q2 , but Q7 carries a current I twice that of Q1 and Q2 . Thus √ √ VOV7 = 2VOV1,2 = 2VOV and gm7 =

√ 2I 2I 2I = √ = VOV7 VOV 2VOV

ro7 = ro9 = Thus, RSS =

√

2I

VOV

VA I

The common-mode gain can now be found as | Acm |=| Gmcm | (Rod Rom ) = 2.61 × 10−4 × (5075) = 7.83 × 10−3 V/V The CMRR can be found as CMRR =

√

2 =

| Acm |= Gmcm (Rod Rom )

√ 2 1 I 2VA 2|VA | 2I × × × × CMRR = VOV VOV I 2 VOV I

3 √

VA

= 2 2

Q.E.D. VOV For k (W/L) = 4 mA/V and I = 160 µA, 1 × 4 × |VOV |2 2 ⇒ |VOV | = 0.2 V

0.080 =

For |VA | = 5 V:

Sedra, Smith, Carusone, Gaudet

40 | Ad | = 4 = 4×10−3 V/V CMRR 10

But,

2VA2 VOV I

2

30 | Ad | = | Acm | 7.83 × 10−3

= 3.83 × 103 or 71.7 dB

| Acm |=

and

= 2.61 × 10−4 mA/V

8.103 To obtain CMRR = 80 dB or CMRR = 104 we must design for a common-mode gain | Acm | of

VA I

where RSS = 100 k, Am = 0.98 A/A, Rim = 1.63 k, and ro = Rod = 50 k. Thus, 1.63 1 (0.98 − 1) − (0.98 + 1) Gmcm = 2 × 100 2 × 50

Thus, 4 × 10−3 = Gmcm × (4040) ⇒ Gmcm = 2 × 10−4 mA/V Now, using Eq. (8.133) with assumed negligibly small and ro = Rod , we obtain 2 × 10−4 =

1 × [(0.98 − 1)− 2RSS

0.5 (0.98 + 1)] ⇒ RSS = 81 k 2 × 40

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–37

8.104 Refer to Fig. P8.104.

W 1 V2 μn Cox 2 L 1,2 OV W 1 100 = × 400 × × 0.04 2 L 1,2 W W = = 12.5 ⇒ L 1 L 2 W 1 |VOV |2 ID3 = ID4 = 100 = μp Cox 2 L 3,4 W 1 100 = × 100 × × 0.04 2 L 3,4 W W = = 50 ⇒ L 3 L 4

(a) ID1 = ID2 = 100 =

ID5 = ID7 = ID8 = 200 µA Thus

W 1 μn Cox V2 2 L 5,7,8 OV W 1 = × 400 × × 0.04 2 L 5,7,8 W W W = = = 25 L 5 L 7 L 8 W 1 ID6 = 200 = μp Cox |VOV |2 2 L 6 W 1 200 = × 100 × × 0.04 2 L 6 W = 100 L 6

200 =

The results are summarized in the following table: Transistor Q1 W/L

Q2 Q3 Q4 Q5 Q6 Q7 Q8

12.5 12.5 50 50 25 100 25 25

−0.1 V ≤ VICM ≤ +0.7 V (c) v Omax = VDD − |VOV6 | = 0.9 − 0.2 = +0.7 V v Omin = −VSS + |VOV7 | = −0.9 + 0.2 = −0.7 V Thus −0.7 V ≤ v O ≤ +0.7 V (d) A1 = −gm1,2 (ro2 ro4 ) where 2 × 0.1 = 1 mA/V 0.2 6 |VA | = = 60 k ro2 = ro4 = 0.1 mA 0.1 A1 = −1 × (60 60) = −30 V/V gm1,2 =

A2 = −gm6 (ro6 ro7 ) where 2 × 0.2 = 2 mA/V 0.2 6 |VA | = = 30 k ro6 = ro7 = 0.2 0.2 A2 = −2 × (30 30) = −30 V/V gm6 =

A0 = A1 A2 = 30 × 30 = 900 V/V 8.105 If (W/L)7 becomes 48/0.8, ID7 will become (W/L)7 ID7 = ID8 (W/L)8 (24/0.4) = IREF (20/0.4) = 100 × 1.2 = 120 µA Thus ID7 will exceed ID6 by 20 µA, which will result in a systematic offset voltage, VO = 20 µA(ro6 ro7 )

Ideally, the dc voltage at the output is zero.

where

(b) The upper limit of VICM is determined by the need to keep Q1 and Q2 in saturation, thus

ro6 = 50 k

VICMmax = VD1 + Vt

and ro7 now becomes ro7 =

= VDD − |VSG4 | + Vt

Thus

= 0.9 − |Vt | − |VOV4 | + Vt

VO = 20 × 10−3 × (50 41.7)

= 0.9 − 0.2 = +0.7 V The lower limit of VICM is determined by the need to keep Q5 in saturation, VICMmin = −0.9 + |VOV5 | + |VGS1 |

= 455 mV The corresponding input offset voltage will be VOS =

= −0.9 + 0.2 + 0.2 + 0.4 = −0.1 V Thus

Sedra, Smith, Carusone, Gaudet

5 = 41.7 k 0.12

=

VO A0

455 = 0.73 mV 625

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–38

8.106 Refer to Fig. P8.106. (a) With the inputs grounded and the output at 0 V dc, we have

= −40 × −2070 × 0.976 = 8.07 × 104 V/V

1 × 0.4 = 0.2 mA 2 0.2 mA

IE1 = IE2 = IE3 = IE4

The overall voltage gain can now be obtained as vo A0 ≡ = A1 A2 A3 v id

8.107

IE5 0.5 mA IE6 = 1 mA (b) The short-circuit transconductance of the first stage is Gm = gm1,2 =

Q3

gm5

Q4

ic1 25

ib1

vid

25 Q2

Q1

β = gm5

where

5 k

ib3

IC1,2 0.2 mA = 8 mA/V VT 0.025 V

The voltage gain of the first stage can be obtained by multiplying Gm by the total resistance at the output node of the stage, i.e., the common collectors of Q2 and Q4 and the base of Q5 . Since ro2 = ro4 = ∞, the resistance at this node is equal to the input resistance of Q5 which is Rπ 5 , rπ 5

ic3

5 k

50

50

50

50

Rin

IC5 0.5 = 20 mA/V = = VT 0.025

thus

Rin2 = 2(β + 1)(25 + 25)

100 = 5 k rπ 5 = 20 Thus the voltage gain of the first stage is given by v b5 A1 ≡ = −Gm rπ 5 v id

= 2 × 101 × 50 10 k

= −8 × 5 = −40 V/V The voltage gain of the second stage is v c5 = −gm5 RC A2 ≡ v b5 where RC5 is the total resistance in the collector of Q5 . Since ro5 = ∞, RC5 is simply the input resistance of the emitter follower Q6 , we have RC5 = Ri6 = (β + 1)(re6 + RL )

Effective load of first stage = Rin2 (5 + 5) = 10 10 = 5 k A1 = α

Total resistance between collectors of Q1 and Q2 Total resistance in emitters of Q1 and Q2

5 k = 25 V/V 4 × 50

Rin = (β + 1)(4 × 50 ) = 101 × 200 20 k ic1 = β1 = 100 ib1 ib3 (5 + 5) 10 = 0.5 = = ic1 (5 + 5) + Rin2 10 + 10

where VT 25 mV = re6 = = 25 IE6 1 mA

ic3 = β3 = 100 ib3

Ri6 = (100 + 1)(0.025 + 1) = 103.5 k Thus A2 = −20 × 103.5 = −2070 V/V

Thus ic3 ic3 ib3 ic1 = × × = 100 × 0.5 × 100 ib1 ib3 ic1 ib1 = 5000 A/A

The gain of the third stage is given by A3 =

vo RL 1 = 0.976 V/V = = v5 RL + re6 1 + 0.025

Sedra, Smith, Carusone, Gaudet

8.108 Refer to Fig. 8.38. With R5 replaced with a 1-mA constant-current source with a high output

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–39

resistance, the total resistance in the collector of Q7 now becomes the input resistance of Q8 , which is

Refer to Fig. (a) for the dc analysis. Replacing the 68 k-33 k divider network by its Thévenin equivalent, we obtain

Ri4 = (β + 1)(re8 + R6 )

VBB = −5 V +

= 101 × (0.005 + 3) = 303.5 k Thus the gain of the third stage now becomes A3 = −α

303.5 2.3 + 0.025

33 × 10 V 33 + 68

= −1.73 V RBB = 68 k 33 k = 22.2 k Now, we can determine IE1 from

−130.5 V/V

IE1 =

and the overall voltage gain increases to 130.5 vo = 1.73 × 105 V/V = 8513 × v id 6.42 (b) The output resistance now becomes very large resistance Ro = 3 k re8 + β +1 3 k When the amplifier is loaded with RL = 100 ,

VBB − (−5) − 0.7 RBB 4.7 + β +1

−1.73 + 5 − 0.7 = 0.52 mA 22.2 4.7 + 101 IC1 = α1 × 0.52 = 0.99 × 0.52 0.52 mA

=

The collector current IC1 and the 8.2-k resistor it feeds can be replaced by a Thévenin equivalent as shown in Fig. (b). Thus

RL = RL + Ro 100 1.73 × 105 × 3000 + 100 Gv = 1.73 × 105

5V

Gv = 5581 V/V

3.3 k

If the original amplifier is loaded in RL = 100 ,

0.74 V 8.2 k

100 = 3378 V/V Gv = 8513 × 152 + 100 Thus, although the output resistance of the original amplifier is much lower than that of the modified one, the overall voltage gain realized when the original amplifier is loaded in 100- resistance is much lower than that obtained with the modified design. Thus, replacing the 15.7-k resistance with a constant-current source is an excellent modification to make! 8.109 (a)

IE2 Q2 IC2 (b)

5 − 0.74 − 0.7 8.2 3.3 + 101 = 1.05 mA IE2 =

IC2 1.04 mA The collector current IC2 and the 5.6-k resistance it feeds can be replaced by a Thévenin equivalent as shown in Fig. (c). Thus 0.824 − 0.7 − (−5) 5.6 2.4 + 101 = 2.1 mA IE3 =

VO = −5 + 2.1 × 2.4 = 0 V (b) Rin = 68 k 33 k rπ 1 where rπ 1 =

Sedra, Smith, Carusone, Gaudet

β gm1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–40

ib2 = gm1 v i

8.2 = 16.1v i 8.2 + 2.4

ic2 = β2 ib2 = 100 × 16.1v i = 1610v i Ri3 = (β + 1)(re3 + 2.4 k) = 101(0.0119 + 2.4) = 243.6 k ib3 = ic2 ×

5.6 = 0.0225ic2 5.6 + 243.6

= 0.0225 × 1610v i = 36.18v i ie3 = (β + 1)ib3 = 101 × 36.18 = 3654v i gm1 =

IC1 0.52 = = 20.8 mA/V VT 0.025

100 = 4.81 k 20.8 Rin = 68 k 33 4.81 4 k 5.6 k Rout = 2.4 k re3 + β +1

rπ 1 =

where

v o = ie3 × 2.4 k = 3654 × 2.4v i = 8770v i Thus vo = 8770 V/V vi

8.110 Refer to Fig. 8.37. W6 can be determined using Eq. (8.144):

VT 25 mV = 11.9 = IE3 2.1 mA 5.6 = 2.4 0.0119 + 101

re3 =

(W/L)7 (W/L)6 =2 (W/L)4 (W/L)5

Rout

(50/0.5) (W/0.5)6 =2 (6.25/0.5) (50/0.5)

= 65.5

⇒ W6 = 12.5 µm

(c) Refer to Fig. (d)

For all devices we can evaluate ID as follows: ID8 = IREF = 50 µA ID5 = IREF

(W/L)5 = 4IREF = 200 µA (W/L)8

I = ID5 = 200 µA 1 ID5 = 100 µA 2 = ID1 = 100 µA

ID1 = ID2 = ID3 = ID4

ID6 = ID7 = 4IREF = 200 µA With ID in each device known, we can use W 1 |VOVi |2 IDi = μCox 2 L i to determine |VOVi | and then |VGSi | = |VOVi | + |Vt |

ic1 = gm1 v i = 20.8v i Ri2 = rπ 2 =

β gm2

where gm2 =

IC2 1.04 mA = 41.6 mA = VT 0.025 V

rπ 2 =

100 = 2.4 k 41.6

Sedra, Smith, Carusone, Gaudet

The values of gmi and roi can then be determined from 2IDi gmi = |VOVi | roi =

|VA | IDi

A1 = −gm1 (ro2 ro4 )

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–41

This table belongs to Problem 8.110.

ID (µA)

Q1

Q2

Q3

Q4

Q5

Q6

Q7

Q8

100

100

100

100

200

200

200

50

|VOV | (V)

0.21 0.21 0.21 0.21 0.21 0.21 0.21 0.21

|VGS | (V)

0.61 0.61 0.61 0.61 0.61 0.61 0.61 0.61

gm (mA/V) 0.95 0.95 0.95 0.95 ro (k)

60

60

60

60

1.9

1.9

1.9

0.48

30

30

30

120

= −0.95 × (60 60) = −28.5 V/V A2 = −gm6 (ro6 ro7 ) = −1.9 × (30 30) = −28.5 V/V A0 = A1 A2 = −28.5 × −28.5 = 812 V/V The upper limit of VICM is determined by the need to keep Q5 in saturation, thus VICMmax = VDD − |VOV5 | − |VSG1 | = 0.9 − 0.21 − 0.61 = +0.08 V

If Q3 has a threshold voltage Vt and Q4 has a threshold voltage Vt + Vt then 1 I = k n3 (VGS3 − Vt )2 2 2 ⇒ VGS3 = Vt + I/k n3

ID3 =

ID4 =

1 k n4 (VGS4 − Vt − Vt )2 2

Since k n4 = k n3 and VGS4 = VGS3 , we have

−0.69 V ≤ VICM ≤ +0.08 V

1 k n3 (VGS3 − Vt − Vt )2 2 1 = k n3 ( I/k n3 − Vt )2 2 Vt 2 I 1 1− √ = k n3 2 k n3 I/k n3 Vt 2 I 1− = 2 VOV3

The output voltage range is

The output current of the first stage will be

−VSS + VOV6 ≤ v O ≤ VDD − |VOV7 |

IO = ID2 − ID4 I Vt 2 I 1− = − 2 2 VOV3

The lower limit of VICM is determined by the need to keep Q1 and Q2 in saturation, thus VICMmin = VG3 − |Vt | = −VSS + |VGS3 | − |Vt | = −0.9 + 0.61 − 0.4 = −0.69 V Thus

that is, −0.69 V ≤ v O ≤ +0.69 V The results are summarized in the table at the top of the page. 8.111 (a) Increasing (W/L)1 and (W/L)2 by a factor of 4 reduces |VOV1,2 | by a factor of 2. Thus gm1,2 = 2ID /|VOV1,2 | increase by a factor of 2. (b) A1 is proportional to gm1,2 , thus A1 increases by a factor of 2 and the overall voltage gain increases by a factor of 2. (c) Since the input offset voltage is proportional to |VOV1,2 |, it will decrease by a factor of 2. This, however, does not apply to VOS due to Vt . 8.112 Refer to Fig. 8.37 and let the two input terminals be grounded. Then, ID1 = ID2 =

I 2

Sedra, Smith, Carusone, Gaudet

ID4 =

Vt 1 we obtain VOV3 I 2Vt I 1− IO − 2 2 VOV3 For

=

2(I/2) Vt VOV3

= gm3 Vt

Q.E.D.

The corresponding input offset voltage will be VOS =

IO gm1,2

gm3 Vt gm1,2 gm3 VOS = Vt gm1,2 =

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–42

This figure belongs to Problem 8.113.

VDD 1.2 V Q3

Q4 Q6 Q1

Q2 vo

IREF 200 A Q8

8.113 (a) For Q1 − Q4 to conduct a current of 200 µA, we have 1 W V2 ID1,2 = k n 2 L 1,2 OV 1 W 200 = × 540 × × 0.152 2 L 1,2 W ⇒ = 32.9 L 1,2 W 1 |VOV |2 ID3,4 = k p 2 L 3,4 W 1 200 = × 100 × × 0.152 2 L 3,4 W ⇒ = 178 L 3,4 Transistor Q5 must carry a current of 400 µA, thus W 1 V2 400 = k n 2 L 5 OV 1 W = × 540 × × 0.152 2 L 5 W ⇒ = 65.8 L 5 Similarly, Q7 is required to conduct a current of 400 µA, thus W W = = 65.8 L 7 L 5 Transistor Q8 conducts a current of 200 µA, thus W 1 W = = 32.9 L 8 2 L 5

Sedra, Smith, Carusone, Gaudet

Q7

Q5

Finally, Q6 must conduct a current equal to that of Q7 , that is, 400 µA, thus a W 1 400 = × 100 × × 0.152 2 L 6 W ⇒ = 356 L 6 The results are summarized in the following table: Transistor Q1

Q2

Q3

Q4

Q5

Q6

Q7

Q8

ID (µA)

200 200 200 200 400 400 400 200

W/L

32.9 32.9 178 178 65.8 356 65.8 32.9

(b) The upper limit on VICM is determined by the need to keep Q1 and Q2 in saturation, thus VICMmax = VD1,2 + |Vt | = VDD − |Vt | − |VOV | + |Vt | = 1.2 − 0.15 = 1.05 V The lower limit on VICM is determined by the need to keep Q5 in saturation, thus VICM = |VOV5 | + VGS1,2 = 0.15 + 0.15 + 0.35 = 0.65 V Thus 0.65 V ≤ VICM ≤ 1.05 V (c) 0.15 V ≤ v O ≤ (1.2 − 0.15) that is, 0.15 V ≤ v O ≤ 1.05 V 2 × 0.2 = 2.67 mA/V 0.15 2.4 |VA | = = 12 k = 0.2 mA 0.2

(d) gm1,2 = ro2 = ro4

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–43

A1 = −gm1,2 (ro2 ro4 ) = 2.67(12 12) = −16 V/V 2 × 0.4 = 5.33 mA/V 0.15 2.4 |VA | = = 6 k ro6 = ro7 = 0.4 mA 0.4 A2 = −gm6 (ro6 ro7 )

gm6 =

= −5.33(6 6) = 16 V/V A0 = A1 A2 = −16 × −16 = 256 V/V 8.114 Refer to Fig. 8.38. From Example 8.7, we obtain IC1 = IC2 = 0.25 mA IC7 = 1 mA

A2 =

v o2 3 k Ri3 = −α v o1 2 × 0.025 + 2 × 0.025

From Example 8.14, Ri3 = 234.8 k, thus A2 −

3 234.8 = −29.6 V/V 0.1

which is half the value without the two 25- emitter resistances. The gain of the third stage remains unchanged at −6.42 V/V, and the gain of the fourth stage remains unchanged at 1 V/V. Thus the overall voltage gain becomes

= 20 × −29.6 × −6.42 × 1

IC8 = 5 mA

= 3800.6 V/V

Thus 25 mV = 100 0.25 mA 25 mV = 25 re4 = re5 = 1 mA With 100- resistance in the emitter of each of Q1 and Q2 , we have re1 = re2

Rid = (β + 1)(2re1,2 + 2Re1,2 ) = 101 × (2 × 0.1 + 2 × 0.1) = 40.4 k Thus, Rid increases by a factor of 2. With 25- resistance in the emitter of each of Q4 and Q5 , the input resistance of the second stage becomes Ri2 = (β + 1)(2re4,5 + 2Re4,5 ) = 101 (2 × 0.025 + 2 × 0.025) = 10.1 k Thus, Ri2 is increased by a factor of 2. The gain of the first stage will be v o1 = v id α × Total resistance between the collectors of Q1 and Q2 Total resistance in emitters of Q1 and Q2 40 k 10 k = 20 V/V 2 × 0.1 + 2 × 0.1

Thus the gain of the first stage decreases but only slightly. Of course, the two 100- resistances in the emitters reduce the gain but some of the

Sedra, Smith, Carusone, Gaudet

The gain of the second stage will now be

vo = A1 A2 A3 A4 v id

IC4 = IC5 = 1 mA

reduction is mitigated by the increase in Ri2 , which increases the effective load resistance of the first stage.

which is slightly less than half the gain obtained without the emitter resistances. This is the price paid for doubling Rid . 8.115 The output resistance is mostly determined by R5 . To reduce Ro by a factor of 2, we use 152 R5 Ro = = R6 re6 + 2 β +1 R5 76 = 3000 5 + 101 ⇒ R5 = 7.37 k This change in R5 will affect the gain of the third stage, which will now become A3 = − =−

R5 (β + 1)(re8 + R6 ) R4 + re7

7.37 (101)(0.005 + 3) 2.3 + 0.025

= −3.1 V/V which is about half the original value (not surprising since R5 is about half its original value). Thus, the amplifier gain will be reduced by a factor of 2. To restore the gain of the third stage to its original value, we can reduce R4 . This will, however, change Ri3 and will reduce the gain of the second stage, though only slightly. For instance, to restore the gain of the third stage to −6.42 V/V, we use 6.42 2.3 + 0.025 = R4 + 0.025 3.1 ⇒ R4 = 1.085 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–44

This table belongs to Problem 8.116. Transistor

Q1

Q2

Q3

Q4

Q5

Q6

Q7

Q8

ID

I 2

I 2

I 2

I 2

I 2

I

I 2

I

gm

I |VOV |

I |VOV |

I |VOV |

I |VOV |

I |VOV |

2I |VOV |

I |VOV |

2I |VOV |

ro

2|VA | I

2|VA | I

2|VA | I

2|VA | I

2|VA | I

|VA | I

2|VA | I

|VA | I

This figure belongs to Problem 8.116, part (b).

Q5

Q3

id

id

vid 2

Q1

id

3 k 112 k = −58.4 V/V 30 which is a slight decrease in magnitude from the original value of −59.2 V/V.

A2 = −

8.116 (a) Refer to Fig. P8.116. The current ID in each of the eight transistors can be found by inspection. Then, gm of each transistor can be determined as 2ID /|VOV | and ro as |VA |/ID . The results are given in the table at the top of the page. (b) See figure above. Observe that at the output node the total signal current is 4id where v id id = gm1,2 2 I v id = 2|VOV | and since the output resistance is 1 |VA | 2 I

then 1 |VA | I × × v id 2|VOV | 2 I

Thus Ad ≡

vo |VA | = v id |VOV |

2id

id Q2

vid 2

vo 4vid 2id Q8

Now Ri3 = 101 × (1.085 + 0.025) = 112 k and the gain of the second stage becomes

v o = 4 id R o = 4 ×

Q6

0V

Q7

Ro = ro6 ro8 =

Q4

Q.E.D.

Sedra, Smith, Carusone, Gaudet

(c) See figure on next page. With v icm applied to both input terminals, we can replace each of Q1 and Q2 with an equivalent circuit composed of a controlled current, v icm /2RSS in parallel with a very large output resistance (Ro1 and Ro2 which are equal). The resistances Ro1 and Ro2 will be much larger than the input resistance of each of the mirrors Q3 − Q5 and Q4 − Q6 and thus we can neglect Ro1 and Ro2 altogether. The short-circuit output current of the Q4 − Q6 mirror will be 1 v icm gm6 1− io6 = gm4 gm4 ro4 2RSS v icm |VOV | = 1− 2|VA | RSS and the output resistance will be ro6 . The short-circuit output current of the Q3 − Q5 mirror will be 1 v icm gm5 1− io5 = gm3 gm3 ro3 2RSS v icm |VOV | = 1− 2|VA | 2RSS and the output resistance will be ro5 . Since ro5 is much larger than the input resistance of the Q7 − Q8 mirror ( 1/gm7 ), most of io5 will flow into Q7 , resulting in an output short-circuit current io8 :

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 8–45 1 io5 gm7 ro7 1 io5 =2 1− gm7 ro7 v icm |VOV | 1 = 1− 1− 2|VA | gm7 ro7 RSS

io8 =

gm8 gm7

1−

1 |VA |/I 1 2 | Acm | = |VA |/I [I/|VOV | ] [2|VA |/I]

1 VOV

1 1 |VOV | =

| Acm | = × 2 2 |VA | 4 VA

2

VA

CMRR = 4

Q.E.D. VOV

and the output resistance is ro8 . Thus, at the output node we have a net current 1 v icm |VOV | io6 − io8 = 1 − 2|VA | gm7 ro7 RSS 1 v icm gm7 ro7 RSS

(e) The upper limit on VICM is determined by Q1 and Q2 remaining in saturation, thus VICMmax = VDD − |VSG | + |Vt | = VDD − |VOV | The lower limit on VICM is determined by the need to keep the bias current source in saturation, i.e. maintaining a minimum voltage across it of |VOV |, thus

This current flows into the output resistance (ro6 ro8 ) and thus produces an output voltage vo =

1 ro6 ro8 v icm RSS gm7 ro7

VICMmin = −VDD + |VOV | + |VGS |

and the common-mode gain becomes

= −VDD + 2|VOV | + |Vt |

ro6 ro8 1 | Acm | = RSS gm7 ro7

Thus

(d) RSS =

Q.E.D.

−VDD + |Vt | + 2|VOV | ≤ VICM ≤ VDD − |VOV |

|VA | I

The output linear range is

VA Ad = | | VOV

−VDD + |VOV | ≤ v O ≤ +VDD − |VOV |

This figure belongs to Problem 8.116, part (c).

Q5

Q3

Q6

Q4

io6

io5

ro6 vo ro8

ro5 Ro1

vicm 2RSS

vicm 2RSS

Ro2

1 ( gm7 // ro7 (

io8

Q7

Sedra, Smith, Carusone, Gaudet

Q8

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 9–1

fT =

Chapter 9 Solutions to Exercises within the Chapter Ex: 9.1 Cox =

ox 3.45 × 10−11 F/m = tox 3 × 10−9 m

= 11.5 × 10−3 F/m2 Cov = WLov Cox = 1.5 × 0.03 × 11.5 = 0.52 f F 2 Cgs = WLCox + Cov 3 2 = × 1.5 × 0.15 × 11.5 + 0.52 3 = 2.25 fF

Ex: 9.4 |β| = 10 at f = 50 MHz fT = 10 × 50 = 500 MHz gm Cπ + Cμ = 2π fT =

40 × 10−3 2π × 500 × 106

= 12.7 pF Cπ = 12.7 − 2 = 10.7 pF Ex: 9.5 Cπ = Cde + Cje

Cgd = Cov = 0.52 f F Csb0 1 = 0.9 f F = VSB 0.2 1+ 1+ V0 0.9 Cdb0 1 = 0.7 f F = = VDB 0.8 + 0.2 1+ 1+ V0 0.9

Csb =

Ex: 9.2 gm = 2k n (W/L)ID = 2 × 0.5 × (1.5/0.15) × 0.1 = 1 mA/V gm fT = 2π(Cgs + Cgd ) =

40 × 10−3 = 7.47 GHz 2π(0.84 + 0.012) × 10−12

Thus,

= 11.5 f F/µm2

Cdb

=

gm 2π(Cπ + Cμ )

1 × 10−3 2π(2.25 + 0.52) × 10−15

10.7 = Cde + 2 ⇒ Cde = 8.7 pF Since Cde is proportional to gm and hence IC , at IC = 0.1 mA, Cde = 0.87 pF and Cπ = 0.87 + 2 = 2.87 pF fT =

4 × 10−3 = 130.7 MHz 2π(2.87 + 2) × 10−12

Ex: 9.6 AM = −gm RL where gm = 2 mA/V RL = RL ro = 1020 = 6.67 k

fT = 57.5 GHz

Thus, Ex: 9.3 Cde = τF gm

AM = −2 × 6.67 = −13.3 V/V

where

Ceq = (1 + gm RL )Cgd

τF = 20 ps

= (1 + 13.3) × 5 = 71.5 fF

IC 1 mA = 40 mA/V = gm = VT 0.025 V Thus, Cde = 20 × 10−12 × 40 × 10−3 = 0.8 pF Cje 2Cje0 = 2 × 20 = 40 f F Cπ = Cde + Cje = 0.8 + 0.04 = 0.84 pF Cμ =

Cμ0 VCB m 1+ V0c

20 = = 12 f F 2 0.33 1+ 0.5

Sedra, Smith, Carusone, Gaudet

Cin = Cgs + Ceq = 20 + 71.5 = 91.5 fH =

1 2π Cin Rsig

1 2π × 91.5 × 10−15 × 20 × 103 = 86.9 MHz 1 Ex: 9.7 fH = 2π Cin Rsig =

For fH ≥ 100 MHz, Cin ≤

1 = 79.6 fF 2π × 100 × 106 × 20 × 103

But, Cin = Cgs + Ceq = 20 + Ceq

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 9–2

Thus,

1 × 20 = 10 V/V 2

Ex: 9.12 | AM | =

Ceq ≤ 59.6 fF Cgd (1 + Cgd ≤

gm RL )

Rgs = 20 k ≤ 59.6

Rgd = Rsig (1 + gm RL ) + RL

59.6 = 2.8 fF 1 + 20

= 20(1 + 10) + 5 = 225 k RCL = RL = 5 k

Ex: 9.8 Refer to Example 9.2.

τgs = Cgs Rgs = 20 × 10−15 × 20 × 103 = 400 ps

To move ft from 10.6 GHz to 3 GHz, fH must be moved from 530.5 MHz to 3 = 150.1 MHz fH = 530.5 × 10.6 Since,

τgd = Cgd Rgd = 5 × 10−15 × 225 × 103 = 1125 ps

fH =

1 2π(CL + Cgd )RL

τH = τgs + τgd + τCL = 400 + 1125 + 125 = 1650 ps

then CL + Cgd

τCL = CL RCL = 25 × 10−15 × 5 × 103 = 125 ps

1 = = 2π × 150.1 × 106 × 10 × 103

fH =

1 2π τH

1 = 96.5 MHz 2π × 1650 × 10−12

106 fF

=

CL = 106 − 5 = 101 fF

GB = 10 × 96.5 = 0.965 GHz

Ex: 9.9 To reduce the midband gain to half the value found, we reduce RL by the same factor, thus 4.76 = 2.38 k RL = 2 But, RL = RL ro 2.38 = RL 100 ⇒ RL = 2.44 k

2μn Cox

W ID L

Since ID is increased by a factor of 4, gm doubles: gm = 2 × 2 = 4 mA/V Since RL is ro /2, increasing ID by a factor of four results in ro and hence RL decreasing by a factor of 4, thus 1 × 10 = 2.5 k 4

| AM | = gm RL = 4 × 2.5 = 10 V/V

= 7 + 1(1 + 40 × 2.38)

Rgs = Rsig = 20 k

= 103.2 pF

Rgd = Rsig (1 + gm RL ) + RL

1 2π Cin Rsig

= 20(1 + 10) + 2.5

1 = 2π × 103.2 × 10−12 × 1.67 × 103 = 923 MHz Thus, by accepting a reduction in gain by a factor of 2, the bandwidth is increased by a factor of 923/480 = 1.9, approximathely the same factor as the reduction in gain. 1000 Ex: 9.10 T(s) = s 1+ 2π × 105 GB = 1000 × 100 × 103 = 108 Hz Ex: 9.11 | AM | =

Ex: 9.13 gm =

RL =

Cin = Cπ + Cμ (1 + gm RL )

fH =

gm RL

= 2 × 10 = 20 V/V

= 222.5 k RCL = RL = 2.5 k τH = τgs + τgd + τCL = Cgs Rsig + Cgd Rgd + CL RCL = 20 × 10−15 × 20 × 103 + 5 × 10−15 × 222.5 × 103 + 25 × 10−15 × 2.5 × 103 = 400 + 1112.5 + 62.5 = 1575 ps fH =

1 = 101 MHz 2π × 1575 × 10−12

GB = | AM | fH

GB = | AM | fH

= 20 × 56.8 = 1.14 GHz

= 10 × 101 = 1.01 GHz

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 9–3

Ex: 9.14 (a) gm = 40 mA/V

Ex: 9.15 Rin =

RL + r o 1 + gm ro

rπ =

200 = 5 k 40

=

ron =

130 VAn = = 130 k I 1

Gv =

rop =

|VAp | 50 = = 50 k I 1

Rgs = Rsig Rin = 20 20 = 10 k

RL = ron rop = 130 50 = 36.1 k AM = −

800 + 20 20 k 1 + 40 RL 800 = 20 V/V = Rsig + Rin 20 + 20

Rgd = RL Ro = 800 840 = 410 k

rπ gm RL rπ + Rsig

τH = Cgs Rgs + (Cgd + CL )Rgd

5 × 40 × 36.1 =− 5 + 36

= 20 × 10−15 × 10 × 103 + (5 + 25) × 10−15 × 410 × 103

= −176 V/V

= 200 + 12,300

(b) Cin = Cπ + Cμ (1 + gm RL )

= 12,500 ps

= 16 + 0.3(1 + 40 × 36.1)

1 = 12.7 MHz 2π × 12,500 × 10−12

= 450 pF

fH =

Rsig = rπ Rsig

Thus, while the midband gain has been increased substantially (by a factor of 21), the bandwidth has been substantially lowered (by a factor of 20.7). Thus, the high-frequency advantage of the CG amplifier is completely lost!

= 5 36 = 4.39 k fH = =

1 2π Cin Rsig

1 2π × 450 × 10−12 × 4.39 × 103

= 80.6 kHz (c) Rπ = Rsig = 4.39 k

Acascode = −gm (RL Ro )

= 4.39(1 + 40 × 36.1) + 36.1

= −gm (ro gm ro ro )

= 6.38 M RCL =

−gm ro = −40 V/V

= 36.1 k

Thus,

τH = Cπ + Cμ Rμ + CL RCL = 16 × 4.39 + 0.3 × 6.38 × 10 + 5 × 36.1 3

= 70.2 + 1914 + 180.5

Rgs = Rsig Rgd = Rsig (1 + gm RL ) + RL

40 × 10−3 = 21.2 GHz 2π × 0.3 × 10−12

(e) GB = 175 × 73.5 = 12.9 MHz

Sedra, Smith, Carusone, Gaudet

(b) For the CS amplifier,

where

1 2π × 2164.7 × 10−9

= 73.5 kHz gm (d) fZ = 2π Cμ =

Acascode =2 ACS τH = Cgs Rgs + Cgd Rgd

= 2164.7 ns fH =

1 = −gm (ro ro ) = − gm ro 2 1 = − × 40 = −20 V/V 2

Rμ = Rsig (1 + gm RL ) + RL

RL

Ex: 9.16 (a) ACS = −gm (RL ro )

Rsig (1 + gm RL ) 1 = Rsig 1 + gm ro 2 1 = Rsig 1 + × 40 = 21Rsig 2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 9–4

τH = Cgs Rsig + Cgd × 21Rsig

AM = −244 V/V

= Cgs Rsig + 0.25Cgs × 21Rsig

Rsig = rπ 1 Rsig

= 6.25Cgs Rsig

= 5 36 = 4.39 k

fH =

Rπ 1 = Rsig = 4.39 k

1 2π × 6.25Cgs Rsig

Rc1 = ro1 Rin2

For the cascode amplifier, τH Rsig [Cgs1 + Cgd1 (1 + gm1 Rd1 )] where Rd1 = ro1 Rin2 = ro

= 10.6 k τH = Cπ 1 Rπ 1 + Cμ1 Rμ1 + Cπ 2 Rc1 + (CL + Cμ2 )(RL Ro ) = 16 × 4.39 + 0.3 × 10.6 + 16 × 0.035

2ro 2ro ro = = 2 + gm ro 2 + 40 21 gm ro

τH = Cgs Rsig 1 + 0.25 1 + 21 40 = Cgs Rsig 1 + 0.25 1 + 21 =

1 2π × 1.73 Cgs Rsig

= 70.24 + 3.18 + 0.56 + 264.5 = 338.5 ns fH =

1 = 470 kHz 2π × 338.5 × 10−9

Thus, in comparison to the CE amplifier of Exercise 9.19, we see that | AM | has increased from 175 V/V to 242 V/V, fH has increased from 73.5 kHz to 470 kHz, and ft has increased from 12.9 MHz to 113.8 MHz.

Thus, 6.25 fH (cascode) = = 3.6 fH (CS) 1.73 ft (cascode) = 2 × 3.6 = 7.2 ft (CS)

Ex: 9.17 gm = 40 mA/V rπ =

+ (5 + 0.3)(50 26,000)

ft = | AM | fH = 244 × 470 = 113.8 MHz

= 1.73Cgs Rsig

(c)

Rμ1 = Rsig (1 + gm1 Rc1 ) + Rc1 = 4.39(1 + 40 × 0.035) + 0.035

ro + ro gm ro

2 ro 2 gm = ro = 2 gm + ro gm

fH =

= 130 k 35 35

To have fH equal to 1 MHz, τH =

1 1 = = 159.2 ns 2π fH 2π × 1 × 106

Thus,

β 200 = 5 k = gm 40

159.2 = 70.24 + 3.18 + 0.56 + (CL + Cμ )(50 26,000)

Rin = rπ = 5 k

⇒ CL + Cμ = 1.71 pF

A0 = gm ro

Thus, CL must be reduced to 1.41 pF.

= 40 × 130 = 5200 V/V Ro1 = ro1 = 130 k Rin2 = rπ 2

Ex: 9.18 From Eq. (9.103), we obtain

ro2 + RL gm2 ro2

Rgs =

130 + 50 5200 = 35 =5

Rgd = Rsig

Ro β2 ro2 = 200 × 130 = 26 M AM = −

rπ gm (Ro RL ) rπ + Rsig

5 =− 40(26,000 50) 5 + 36

Sedra, Smith, Carusone, Gaudet

Rsig + RL Rsig RL + = gm RL + 1 gm RL + 1 gm RL + 1

RCL =

RL gm RL + 1

Ex: 9.19 From Example 9.8, we get τH = b1 = 104 ps

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 9–5

1 2π τH

fH = =

fP1 = 67.2 MHz

1 = 1.53 GHz 2π × 104 × 10−12

This is lower than the exact value found in Example 9.8 (i.e., 1.86 GHz) by about 18%, still not a bad estimate!

β 100 = 2.5 k = gm 40

W 1 V2 μn Cox 2 L 1,2 OV

2ID 2 × 0.4 = = 4 mA/V VOV 0.2

(b) Ad = gm (RD ro )

= RL ro = 1 100 = 0.99 k RL RL + re +

=

Ex: 9.21 (a) ID1,2 =

gm =

Rsig = 1 k

AM =

fH fP1 = 67.2 MHz

0.4 =

re = 25

RL

Since fP1 fP2 ,

1 2 × 0.2 × 100VOV 2 ⇒ VOV = 0.2 V

Ex: 9.20 gm = 40 mA/V

rπ =

fP2 = 562 MHz

Rsig

where VA 20 = 50 k = ro = ID 0.4 Ad = 4(5 50) = 4 × 4.545

β +1

0.99 = 0.97 V/V 0.99 + 0.025 + (1/101)

gm Cπ + Cμ = 2π fT

= 18.2 V/V (c) fH = =

40 × 10−3 = 2π × 400 × 106

1 2π(CL + Cgd + Cdb )(RD ro )

1 2π(100 + 10 + 10) × 10−15 × 4.545 × 103

= 292 MHz

= 15.9 pF Cμ = 2 pF

(d) τgs = Cgs Rsig = 50 × 10 = 500 ps

τgd = Cgd Rgd = Cgd Rsig (1 + gm RL ) + RL

Cπ = 13.9 pF

= 10 [10(1 + 18.2) + 4.545]

fZ =

1 1 = 2π Cπ re 2π × 13.9 × 10−12 × 25

= 458 MHz

R Rsig RL Cπ + Cμ 1 + L Rsig + Cπ + CL 1 + re rπ b1 = R Rsig 1+ L + re rπ

0.99 × 1 + (13.9 + 0)0.99 13.9 + 2 1 + 0.025 = 1 0.99 + 1+ 0.025 2.5

= 1965.5 ps τCL = (CL + Cdb )RL = 110 × 4.545 = 500 ps τH = τgs + τgd + τCL = 500 + 1965.5 + 500 = 2965.5 ps fH =

1 2π × 2965.5 × 10−12

= 53.7 MHz Ex: 9.22 fZ =

= 2.66 × 10−9 s

1 2π RSS CSS

1 2π × 75 × 103 × 0.4 × 10−12

Cπ Cμ RL Rsig 13.9 × 2 × 0.99 × 1 = b2 = 1 0.99 Rsig RL + 1+ 1+ + 0.025 2.5 re rπ

=

= 0.671 × 10−18

Thus, the 3-dB frequency of the CMRR is 5.3 MHz.

= 5.3 MHz

ωP1 and ωP2 are the roots of the equation 1 + b1 s + b2 s2 = 0

Ex: 9.23 Ad = gm1,2 (ro2 ro4 )

Solving we obtain,

where

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 9–6

0.5 = 20 mA/V 0.025 100 = 200 k ro2 = ro4 = 0.5 Ad = 20(200 200) = 2000 V/V gm1,2 =

The dominant high-frequency pole is that introduced at the output node, fH = =

where Rs =

2 = 1 k gm

20 × 1 20 + 20 Rgs = 20 1+2× 20 + 20 20 + 1 +

= 10.75 k

1 2π CL (ro2 ro4 )

τH = Cgs Rgs + Cgd Rgd + CL RCL

1 2π × 2 × 10−12 × 100 × 103

= 0.8 MHz

= 20 × 10.75 + 5 × 235 + 5 × 15 = 215 + 1175 + 75 = 1465 ps fH =

Ex: 9.24 (a) AM = −gm RL

1 = 109 MHz 2π × 1465 × 10−12

GB = 10 × 109 = 1.1 GHz

where RL = RL ro = 20 20 = 10 k

Ex: 9.25 Refer to Fig. 9.35(b).

AM = −2 × 10 = −20 V/V

AM =

τH = Cgs Rgs + Cgd Rgd + CL RL

= Cgs Rsig + Cgd Rsig (1 + gm RL ) + RL + CL RL

where

= 20 × 20 + 5 [20(1 + 20) + 10] + 5 × 10 = 400 + 2150 + 50 = 2600 ps fH = =

1 2π τH

gm = 20 mA/V 100 = 5 k 20 1 10 × × 20 × 10 = 50 V/V AM = 10 + 10 2

rπ =

fP1 =

1 2π × 2600 × 10−12 =

= 61.2 MHz GB = | AM | fH

1 Cπ 2π + Cμ (2rπ Rsig ) 2

1 6 2π + 2 × 10−12 (10 10) × 103 2

= 6.4 MHz

= 20 × 61.2 = 1.22 GHz (b) Gm =

2rπ 1 × × gm RL 2rπ + Rsig 2

gm 2 = 0.67 mA/V = 1 + gm Rs 1+2

fP2 =

1 2π Cμ RL

1 2π × 2 × 10−12 × 10 × 103

Ro ro (1 + gm Rs )

=

= 20 × 3 = 60 k

= 8 MHz

RL = RL Ro = 20 60 = 15 k AM = −Gm RL = −0.67 × 15 = −10 V/V Rgd = Rsig (1 + Gm RL ) + RL = 235 k RCL = RL = 15 k Rsig + Rs + Rsig Rs /(ro + RL ) ro 1 + gm Rs r o + RL

Sedra, Smith, Carusone, Gaudet

s 1+ ωP1

50 s 1+ ωP2

50 |T( jω)| = 2 2 ω ω 1+ 1+ ωP1 ωP2

= 20(1 + 10) + 15

Rgs =

T(s) =

√ At ω = ωH , |T| = 50/ 2, thus ωH 2 ωH 2 1+ 2= 1+ ωP1 ωP2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 9–7 =1+ ωH4 2 2 ωP1 ωP2

ωH ωP1

2

+

+ ωH2

ωH ωP2

2

1 1 + 2 2 ωP1 ωP2

ωH ωP1

+

2

ωH ωP2

2

−1=0

1 fH4 1 2 −1=0 + f + H 2 2 2 2 fP1 fP2 fP1 fP2 1 fH4 1 2 −1=0 + f + H 6.42 × 82 6.42 82 ⇒ fH = 4.6 MHz (Exact value) Using Eq. (9.150), an approximathe value for fH can be obtained: 1 1 fH 1/ 2 + 2 fP1 fP2 = 1/

1 1 + 2 = 5 MHz 6.42 8

= 8 Hz fZ = =

1 2π CS RS

1 2π × 1 × 10−6 × 10 × 103

= 15.9 Hz Since the highest-frequency pole is fP2 = 334.2 and the next highest-frequency singularity is fZ at 15.9 Hz, the lower 3-dB frequency fL will be fL fP2 = 334.2 Hz Ex: 9.27 Refer to Fig. 9.42.

τC1 = CC1 Rsig + (RB rπ ) = 1 × 10−6 [5 + (100 2.5)] × 103 = 7.44 ms RB Rsig τCE = CE RE re + β +1 β = gm rπ = 40 × 2.5 = 100

RG gm (RD RL ) Ex: 9.26 AM = − RG + Rsig =−

10 × 2(10 10) 10 + 0.1

τCE = 0.071 ms

= −9.9 V/V fP1 = =

τC2 = CC2 (RC + RL )

1 2π CC1 (Rsig + RG )

= 1 × 10−6 (8 + 5) × 103 = 13 ms

1 2π × 1 ×

10−6 (0.1

+ 10) ×

106

= 0.016 Hz gm + 1/RS fP2 = 2π CS =

1 2π CC2 (RD + RL )

1 = 2π × 1 × 10−6 (10 + 10) × 103

Sedra, Smith, Carusone, Gaudet

fL =

1 2π

1 = 2π

(2 + 0.1) × 10−3 = 334.2 Hz 2π × 1 × 10−6

fP3 =

re 1/gm = 25 100 5 × 103 τCE = 1 × 10−6 5 0.025 + 101

1 1 1 + + τC1 τCE τC2

1 1 1 + + × 103 7.44 0.071 13

= 2.28 kHz 1 fZ = 2π CE RE =

1 = 31.8 Hz 2π × 1 × 10−6 × 5 × 103

Since fZ is much lower than fL it will have a negligible effect on fL .

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–1

Solutions to End-of-Chapter Problems 9.1 Cox = =

ox tox

fT = =

3.45 × 10−11 F/m = 0.86 × 10−2 F/m2 4 × 10−9 m

= 0.86 × 10−2 × 10−12 F/µm2 = 8.6 fF/µm2 k n = μn Cox

2.6 × 10−3 = 15.4 GHz 2π(23.7 + 3.1) × 10−15

2ID 2 × 0.2 = 1.33 mA/V = VOV 0.3 gm fT = 2π(Cgs + Cgd ) 9.2 gm =

=

= 450 × 108 (µm2 /V·s)

gm 2π(Cgs + Cgd )

1.33 × 10−3 = 7.1 GHz 2π × (25 + 5) × 10−15

× 8.6 × 10−15 F/µm2 = 387 µA/V2 W 1 2 VOV ID = k n (1 + λVDS ) 2 L 1 2 200 = × 387 × 40 × VOV (1 + 0.1 × 0.7) 2 ⇒ VOV = 0.155 V 2ID 2 × 0.2 = 2.6 mA/V = VOV 0.155 γ χ= 2 2φf + VSB gm =

0.5 = √ = 0.23 2 0.85 + 0.35 gmb = χ gm = 0.6 mA/V ro =

| VA | 1 1 = 50 k = = ID | λ|ID 0.1 × 0.2

Cgs =

2 WLCox + WLov Cox 3

9.3 fT =

gm 2π(Cπ + Cμ )

where IC 0.5 mA = 20 mA/V = VT 0.025 V

gm =

Cπ = 8 pF Cμ = 1 pF Thus, fT =

20 × 10−3 = 353.7 MHz 2π × (8 + 1) × 10−12

fβ =

353.7 fT = = 3.54 MHz β 100

9.4 For f fβ , |β|

fT f

At f = 50 MHz and IC = 0.2 mA, fT 50 ⇒ fT = 500 MHz

=

2 × 12 × 0.3 × 8.6 + 12 × 0.03 × 8.6 3 = 20.64 + 3.1 = 23.7 fF

|β| = 10 =

Cgd = WLov Cox = 12 × 0.03 × 8.6

At f = 50 MHz and IC = 1.0 mA,

= 3.1 fF

|β| = 12 =

Csb =

=

Csb0 | VSB | 1+ V0

5 1+

Cdb

0.35 0.9

= 4.2 fF

5 1+

1.05 0.9

Now, fT =

gm 2π(Cπ + Cμ )

where

Cdb0 = | VDB | 1+ V0

=

fT 50 ⇒ fT = 600 MHz

= 3.4 fF

Cπ = Cde + Cje = τF gm + Cje Cμ = 0.1 pF At IC = 0.2 mA, gm = 500 × 106 =

Sedra, Smith, Carusone, Gaudet

0.2 = 8 mA/V, thus 0.025

8 × 10−3 2π(Cπ + 0.1) × 10−12

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–2

⇒ Cπ = 2.45 pF −3

τF × 8 × 10

+ Cje = 2.45 × 10

At IC = 1 mA, gm = 600 × 106 =

−12

(1)

1 = 40 mA/V, thus 0.025

9.7 fT =

40 × 10−3 2π(Cπ + 0.1) × 10−12 (2)

Solving (1) together with (2) yields τF = 252 ps

gm 2π(Cπ + Cμ )

10 × 109 =

Substituting from (2) and (3) into (1), we obtain W VOV μn Cox L fT = 2 2π × WLCox 3 3μn VOV = 4π L2 We note that for a given channel length, fT can be increased by operating the MOSFET at a higher VOV .

−3

40 × 10 2π(Cπ + 0.1) × 10−12

⇒ Cπ = 0.54 pF rπ =

β 120 = 3 k = gm 40

fβ =

fT 10 × 109 = 83.3 MHz = β0 120

9.6 fT =

2 WLCox + WLov Cox 3 If the overlap component is small, we get Cgs =

2 WLCox (2) 3 The transconductance gm can be expressed as W VOV gm = μn Cox (3) L

IC 1 mA = 40 mA/V = VT 0.025 V

gm =

(1)

Cgs

Cje = 0.43 pF 9.5 fT =

gm 2π(Cgs + Cgd )

For Cgs Cgd gm fT 2π Cgs

⇒ Cπ = 10.51 pF τF × 40 × 10−3 + Cje = 10.51 × 10−12

Also, √we observe that fT is inversely proportional to L WL; thus faster operation is obtained from smaller devices.

gm 2π(Cgs + Cgd )

For L = 0.5 µm and μn = 450 cm2 /V·s, we have

For Cgs Cgd gm fT 2π Cgs

(1)

2 WLCox + WLov Cox 3 If the overlap component (WLov Cox ) is small, we get Cgs =

2 WLCox 3 The transconductance gm is given by W ID gm = 2μn Cox L

Cgs

(2)

3 × 450 × 108 × 0.2 4π × 0.52

= 5.73 GHz VOV = 0.4 V ⇒ fT =

3 × 450 × 108 × 0.4 4π × 0.52

= 11.46 GHz 9.8 A0 =

2VA 2VA L = VOV VOV

2×5×L = 50L, V/V (L in µm) 0.2 3μn VOV 3 × 400 × 108 × 0.2 = fT 2 4π L 4π L2 1.91 fT = 2 , GHz (L in µm) L The expressions for A0 and fT can be used to obtain their values for different values of L. The results are given in the following table. A0 =

(3)

Substituting from (2) and (3) into (1), we get W ID 2μn Cox L fT = 2 2π × WLCox 3 μn ID 1.5 = Q.E.D. π L 2Cox WL We observe thatfor a given device, fT is proportional to ID ; thus to obtain faster operation the MOSFET is operated at a higher ID .

Sedra, Smith, Carusone, Gaudet

VOV = 0.2 V ⇒ fT =

L

Lmin 2Lmin 3Lmin 4Lmin 5Lmin 0.13 µm 0.26 µm 0.39 µm 0.52 µm 0.65 µm

A0 (V/V)

6.5

13

19.5

26

32.5

fT (GHz)

113

28.3

12.6

7.1

4.5

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–3

This figure belongs to Problem 9.9.

rx B

Cm

B'

C

Vp

rp

Cp

gm Vp

ro

E This figure belongs to Problem 9.10.

Cm B

C

Vp

rp

Cp

gm Vp

ro

E 9.9 See figure above. Cπ = Cde + Cje

fT =

where Cde is proportional to IC . =

At IC = 0.5 mA,

Cπ = 3 + 2 = 5 pF.

1 × 6 = 3 pF, and 2

Also, at IC = 0.25 mA, gm = 10 mA/V. Thus fT at IC = 0.25 mA is fT =

10 × 10−3 = 265.3 MHz 2π(5 + 1) × 10−12

9.10 See figure above. gm =

IC 1 mA = 40 mA/V = VT 0.025 V

rπ =

β0 100 = 2.5 k = gm 40

ro =

VA 50 = 50 k = IC 1

Cde = τF gm = 30 × 10−12 × 40 × 10−3 = 1.2 pF Cje0 = 20 fF Cπ = Cde + 2Cje0 = 1.2 + 2 × 0.02 = 1.24 pF Cμ =

Cµ0 VCB m 1+ V0c

Cμ =

30 1+

2 0.75

9.11 For f fβ , fT f 5000 MHz 50 = f ⇒ f = 100 MHz fT 5000 MHz = 25 MHz = fβ = β0 200 |β|

9.12 To complete the table we use the following relationships: VT 25 mV re = = IE IE (mA) IC αIE IE IE (mA) = = gm = VT VT VT 0.025 V β0 , k rπ = gm (mA/V) gm fT = 2π(Cπ + Cμ ) 1 2π(Cπ + Cμ )rπ fT fβ = β0

fβ =

0.5 = 15.7 fF

Sedra, Smith, Carusone, Gaudet

40 × 10−3 2π(1.24 + 0.016) × 10−12

= 5.1 GHz

8 = Cde + 2 ⇒ Cde = 6 pF At IC = 0.25 mA, Cde =

gm 2π(Cπ + Cμ )

The results are summarized in the table on the next page.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–4

This table belongs to Problem 9.12. Transistor IE (mA) re () gm (mA/V) rπ (k)

fT (MHz) Cμ (pF) Cπ (pF) fβ (MHz)

(a)

2

12.5

80

12.5

100

500

2

23.5

5

(b)

1

25

40

3.13

125

500

2

10.7

4

(c)

1

25

40

2.5

100

500

2

10.7

5

(d)

10

2.5

400

0.25

100

500

2

125.3

5

(e)

0.1

250

4

25

100

150

2

2.2

1.5

(f)

1

25

40

0.25

10

500

2

10.7

50

(g)

1.25

20

50

0.2

10

800

1

9

80

9.13 The frequency of the pole is 1 1 = 2π RC 2π × 10 × 103 × 10 × 10−12

fP =

β0

= 1.59 MHz

1 1 + sCin Rsig

A Vo (s) =− Vsig (s) 1 + sCin Rsig (c) DC gain = 40 dB = 100 V/V,

9.14 AM = −gm (ro RL )

⇒ A = 100 V/V 1 f3dB = 2π Cin Rsig

= −4(2020) = −40 V/V fH =

=

1 where Cin = Cgs + Cgd (1 + gm RL ) 2π Rsig Cin

100 × 103 =

1 2π Cin × 1 × 103

= 25 + 5(1 + 40)

⇒ Cin = 1591.5 pF

= 230 fF

C=

fH =

1 2π × 20 × 103 × 230 × 10−15

= 34.6 MHz gm fZ = 2π Cgd =

1591.5 Cin = = 15.8 pF A+1 101

(d) The Bode plot is shown in the figure below. From the figure we see that the gain reduces to unity two decades higher than f3dB , that is at 10 MHz.

4 × 10−3 2π × 5 × 10−15

= 127.3 GHz 9.15 (a) Vo = −AVi If the current flowing through Rsig is denoted Ii , we obtain Ii sC(Vi − Vo ) = Yin = Vi Vi Vo = sC 1 − Vi = sC(1 + A)

9.16 Refer to Example 9.1. If the transistor is replaced with another whose W is half that of the original transistor, we obtain

Thus, Cin = C(1 + A) (b)

Vi (s) = Vsig (s)

1/sCin 1 Rsig + sCin

Sedra, Smith, Carusone, Gaudet

W2 =

1 W1 2

Since Cgs =

2 WLCox + WLov Cox 3

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–5

then

fH2 =

Cgs2 =

1 Cgs1 = 10 fF 2

=

Also,

1 2π Cin2 Rsig

1 2π × 47.75 × 10−15 × 20 × 103

Cgd = WLov Cox

= 166.7 MHz

thus,

in comparison to fH1 = 63.7 MHz

Cgd2 = Since

gm =

| AM2 | = gm2 RL2

1 Cgd1 = 2.5 fF 2

= 1.41 × 10 = 14.1 V/V

W 2μn Cox ID L

in comparison to | AM1 | = 20 V/V.

then

GB2 = 14.1 × 166.7 = 2.35 GHz

1 gm2 = √ gm1 = 1.41 mA/V 2 Since W 1 2 VOV ID = μn Cox 2 L

in comparison to GB1 = 20 × 63.7 = 1.27 GHz.

then VOV2 =

9.17 AM = −gm RL = −2 × 20 = −40 V/V 1 fH = 2π(CL + Cgd )RL

√ 2VOV1

Finally,

=

ro2 = ro1 = 20 k

1 2π(100 + 10) × 10−15 × 20 × 103

= 72.3 MHz

Thus,

ft =| AM | fH

RL2 = RL1 = 10 k

= 40 × 72.3 = 2.89 GHz

Thus, Cin2 = Cgs2 +

(gm2 RL2

fZ =

+ 1)Cgd2

gm 2 × 10−3 = 2π Cgd 2π × 10 × 10−15

= 10 + (1.41 × 10 + 1) × 2.5

= 31.8 GHz

= 47.75 fF

A Bode plot for the gain magnitude is shown in the figure. Since from Eq. (9.43),

This should be compared to Cin1 = 125 fF. Thus,

A(s) = −gm RL

This figure belongs to Problem 9.17.

1 − s(Cgd /gm ) 1 + s(CL + Cgd )RL

Gain, dB

40 32 dB 30

20 dB/decade

20 10 0 10 20

fH

ft

72.3 kHz 2.89 GHz

fZ

f

31.8 GHz

20.8 dB

30

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–6

we find that as s → ∞, the gain magnitude becomes 1 10 Cgd = = | A |= CL + Cgd 100 + 10 11

Finally,

or −20.8 dB as indicated in the figure.

Cin = Cπ + Cμ (1 + gm RL )

1 2π Rsig Cin

fH =

where = 4.6 + 0.5(1 + 102.4)

9.18

gm RL

= 56.3 pF

= 50

Cin = Cπ + Cμ (1 + gm RL )

Thus,

= 10 + 1(1 + 50)

fH =

= 61 pF

= 988 kHz

1 fH = 2π Cin Rsig

1 2π × 2.86 × 103 × 56.3 × 10−12

9.20 Rin =

1 = 2π × 61 × 10−12 × 5 × 103

=

= 522 kHz

9.19 Figure (a) on the next page shows the high-frequency equivalent circuit. A simplified version is shown in Fig. (b) on the next page, from which we get

100 k = 1000 k = 1 M 1 − 0.9

9.21 Using Miller’s theorem, we obtain Zin =

Vo Rin =− (gm RL ) Vsig Rin + Rsig

For

gm =

IC 0.8 mA = 32 mA/V = VT 0.025 V

Zin =

rπ =

β 160 = 5 k = gm 32

Zout

AM =

Z=

where

Rin = RB rπ = RB1 RB2 rπ

R 1−K

Z , 1−A

Zout =

Z 1−

1 A

1 jωC 1 ⇒ Cin = C(1 − A) jωC(1 − A) 1 1 ⇒ Cout = C 1 − = 1 A jωC 1 − A

= 68275 = 4 k

(a) A = −1000 V/V, C = 1 pF

Rsig = Rsig Rin = 104 = 2.86 k

Cin = 1(1 + 1000) = 1001 pF 1 = 1.001 pF Cout = 1 1 + 1000

RL = RL RC = 104.7 = 3.2 k gm RL = 32 × 3.2 = 102.4 AM = −

4 × 102.4 4 + 10

= −29.3 V/V Next we determine Cπ , gm fT = 2π(Cπ + Cμ )

(b) A = −10 V/V, C = 10 pF Cin = 10(1 + 10) = 110 pF 1 = 11 pF Cout = 10 1 + 10 (c) A = −1 V/V,

C = 10 pF

Cin = 10(1 + 1) = 20 pF

32 × 10−3 10 = 2π(Cπ + Cμ )

Cout = 10(1 + 1) = 20 pF

⇒ Cπ + Cμ = 5.1 pF

(d) A = +1 V/V, C = 10 pF

Since,

Cin = C(1 − 1) = 0

Cμ = 0.5 pF

Cout = C(1 − 1) = 0

Cπ = 5.1 − 0.5 = 4.6 pF

(e) A = +10 V/V,

9

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

C = 10 pF

© Oxford University Press 2021

Chapter 9–7

This figure belongs to Problem 9.19.

Rsig

Cm

B

Vsig

RB1

rp

RB2

Vp Cp

gmVp

RC

RL

Vo

RB Rin (a) Rsig

Cm

B

Vsig

Vp Cp

Vsig Vsig

RL

gmVp

Vo

Rin Rin Rsig

Rin RB rp RB1 RB2 rp

Rsig Rsig Rin RL RC RL (b)

Cin = 10(1 − 10) = −90 pF 1 = 9 pF Cout = 10 1 − 10

Rsig

1 MHz,

Sedra, Smith, Carusone, Gaudet

The gain falls off at the rate of 20 dB/decade. For the gain to reach 0 dB (unity), the gain has to fall by 60 dB. This requires three decades or a factor of 1000, thus funity gain = 795 × 1000 = 795 MHz

9.25 fH =

1 2π Cin Rsig

For fH ≥ 6 MHz

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–8

1 1 = 2π fH Rsig 2π × 6 × 106 × 1 × 103

Cin ≤

To double fH by changing Rin , Eq. (2) indicates that Rsig must be halved:

Cin ≤ 26.5 pF

Rsig = 25 k

But,

which requires Rin to be changed to Rin2 ,

Cin = Cgs + (1 + gm RL )Cgd

25 k = 100 Rin2

= 5 + (1 + gm RL ) × 1, pF

⇒ Rin2 = 33.3 k

=6+

gm RL ,

pF

This change will cause | AM | to become

For Cin ≤ 26.5 pF we have gm RL

≤ 20.5

RL ≤

20.5 = 4.1 k 5

Corresponding to | AM | =

gm RL

RL

| AM2 | =

33.3 × 3 × 4.1 33.3 + 100

= 3.1 V/V which is about half the original value. = 4.1 k, we have

To double fH by changing RL , Eq. (2) indicates that Cin must be halved:

= 20.5 V/V

= 20.5 × 6 = 123 MHz

1 × 3.66 = 1.83 pF 2 Using Eq. (1), we obtain

If fH = 2 MHz, we obtain

1.83 = 1 + 0.2(1 + gm RL2 )

Cin = 26.5 × 3 = 79.5 pF

⇒ gm RL2 = 3.15

gm RL

Thus,

GB = | AM | fH

Cin2 =

= 79.5 − 6 = 73.5

Thus,

RL2 = 1.05 k

| AM | = 73.5 V/V

and RL2 can be found from

GB = 73.5 × 2 = 147 MHz

1.05 = RL 8 k 50 k

9.26 AM = −

⇒ RL = 1.24 k

Rin gm RL Rin + Rsig

and the midband gain becomes

where RL

| AM2 | =

= RD RL r o

which is about a quarter of the original gain. Clearly, changing Rin is the preferred course of action!

= 8 k 10 k 50 k = 4.1 k AM = −

100 × 3 × 4.1 100 + 100

= −6.1 V/V Cin = Cgs + Cgd (1 + gm RL )

(1)

= 1 + 0.2(1 + 3 × 4.1) = 3.66 pF fH =

1 2π Cin Rsig

(2)

Rsig = Rsig Rin 1 2π × 3.66 × 10−12 × 50 × 103

Sedra, Smith, Carusone, Gaudet

= 1 mA/V | VA1 | 5 = 50 k = ro1 = ID1 0.1 ro2 =

| VA2 | 5 = 50 k = ID2 0.1

RL = ro1 ro2 = 50 k 50 k

= 100 k 100 k = 50 k

= 870 kHz

9.27 gm = 2μn Cox (W/L)1 ID1 √ = 2 × 0.4 × 12.5 × 0.1

The total resistance at the output node, RL , is given by

where

fH =

100 × 3.15 = 1.6 V/V 100 + 100

= 25 k AM = −gm1 RL = −1 × 25 = −25 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–9

1 2π Cin Rsig

fH =

To reduce ft to 1 GHz, an additional capacitance of 0.159 pF must be connected to the output node. (Doubling the effective capacitance at the output node reduces ft by a factor of 2.)

where Cin = Cgs + Cgd (1 + gm1 RL ) = 30 + 5(1 + 25) = 160 fF Thus, fH =

1 2π × 160 × 10−15 × 20 × 103

= 49.7 MHz fZ =

gm 1 × 10−3 = 2π Cgd 2π × 5 × 10−15

= 31.8 GHz

| VA2 | 19.2 = 192 k = ID2 0.1

RL = ro1 ro2 = 128 192 = 76.8 k AM = −gm1 RL

2 × 10−3 109 = 2π(CL + Cgd )

= −1.06 × 76.8 = −81.4 V/V CL = Cdb1 + Cdb2 + Cgd2

⇒ CL + Cgd = 318.3 fF Since the gain levels off at high frequencies at: −20 dB, then fZ = 10 ft But fZ CL + Cgd = ft Cgd

= 20 + 36 + 15 = 71 fF fH =

1 2π(CL + Cgd1 )RL

fH =

1 2π(71 + 15) × 10−15 × 76.8 × 103

= 24.1 MHz fZ =

Thus Cgd = 0.1 CL + Cgd

gm1 1.06 × 10−3 = 2π Cgd1 2π × 0.015 × 10−12

= 11.2 GHz

⇒ Cgd = 31.83 fF

9.31 AM = −

and

rπ RB gm RL RB + Rsig rπ + Rsig

where

CL = 318.3 − 31.83 = 286.5 fF Since, ft =| AM | fH

RL = ro RL = 100 k 10 k = 9.1 k

then, 109 = 20 MHz 50

9.29 ft =

gm 2π(CL + Cgd )

CL + Cgd

gm = 2π ft

=

= 1.06 mA/V VA1 12.8 ro1 = = = 128 k ID1 0.1 ro2 =

9.28 Refer to Fig. 9.12(c) gm ft = 2π(CL + Cgd )

fH =

9.30 Refer to Fig. P9.27. To determine gm1 we use W ID1 gm1 = 2μn Cox L 1 100 × 0.1 = 2 × 0.090 × 1.6

2 × 10−3 = 0.159 pF 2π × 2 × 109

Sedra, Smith, Carusone, Gaudet

and rπ = β/gm = 100/40 = 2.5 k AM = −

2.5 × 40 × 9.1 2.5 + 10

= −72.8 V/V fH =

1 2π Cin Rsig

where

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–10

2.5 = −10 V/V 25 GB 6.37 MHz = = 637 kHz fH = | AM | 10 V/V

Cin = Cπ + Cμ (1 + gm RL )

AM = −100 ×

= 10 + 0.5 × (1 + 40 × 9.1) = 192.5 pF and Rsig = rπ Rsig = 2.5 10 = 2 k fH =

1 2π × 192.5 × 10−12 × 2 × 103

= 413 kHz

The Bode plots are shown in the figure. 9.32 (a) AM = rπ − gm RL rπ + Rsig

If the midband gain is unity,

For Rsig rπ ,

This is obtained when RL is

fH = GB = 6.37 MHz

rπ gm RL = −βRL /Rsig AM − Rsig

Q.E.D.

(b) Cin = Cπ + (gm RL + 1)Cμ For

gm RL

1 and

gm RL Cμ

1 = 100 ×

RL 25

⇒ RL = 0.25 k = 250

Cπ ,

Cin gm RL Cμ

9.33

1 fH = 2π Cin Rsig

From the figure we see that the controlled current-source gm Vπ appears across its control voltage Vπ , thus we can replace the current source with a resistance 1/gm . Now, the parallel equivalent of rπ and 1/gm is

where Rsig = rπ Rsig

rπ rπ (1/gm ) rπ = = = re 1 gm rπ + 1 β +1 rπ + gm

rπ Thus, fH

1 2π gm RL Cμ rπ

fH =

1 2π Cμ βRL

Thus, the equivalent circuit simplifies to that of re in parallel with Cπ , Q.E.D.

(c) GB = | AM | fH =β

1 RL 1 = Rsig 2π Cμ βRL 2π Cμ Rsig

Q.E.D.

Zi (jω) will have a 45◦ phase at

For Rsig = 25 k and Cμ = 1 pF, GB =

1 re = 1 1 + sCπ re + sCπ re re Zi (jω) = 1 + jωCπ re

Zi (s) =

1 2π × 1 × 10−12 × 25 × 103

= 6.37 MHz

For β = 100,

ω45 Cπ re = 1 ⇒ ω45 =

(i) RL = 25 k:

Now,

25 = −100 V/V 25 6.37 MHz GB = = 63.7 kHz fH = | AM | 100 V/V

fT =

AM = −100 ×

(ii)

RL

= 2.5 k:

Sedra, Smith, Carusone, Gaudet

1 Cπ re

gm 2π(Cπ + Cμ )

At high bias currents, Cπ Cμ and

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–11

fT

gm 2π Cπ

Since gm 1/re , we have fT

1 2π Cπ re

Thus, f45◦ fT = 400 MHz If the bias current is reduced to the value that results in Cπ Cμ , gm gm = fT = 2π × 2Cπ 4π Cπ Again, gm fT

Figure 2

1 , thus re

For IC = 200 µA = 0.2 mA and VA = 100 V,

1 4π Cπ re

It follows that in this case, 1 f45◦ = fT = 200 MHz 2 9.34 Figure 1 shows the amplifier high-frequency equivalent circuit. A node equation at the output provides 1 + sCL Vo + gm Vπ + sCμ (Vo − Vπ ) = 0 ro

Vo 1 − s(Cμ /gm ) = −gm ro Vi 1 + s(CL + Cμ )ro

IC 0.2 mA = 8 mA/V = VT 0.025 V

ro =

VA 100 = 500 k = IC 0.2

DC gain = −gm ro = 8 × 500 = −4000 V/V f3dB = =

1 2π(CL + Cμ )ro

1 2π(1 + 0.2) × 10−12 × 500 × 103

= 265.3 kHz gm fZ = 2π Cμ

Replacing Vπ by Vi and collecting terms results in 1 Vo + s(CL + Cμ ) = −Vi (gm − sCμ ) ro ⇒

gm =

=

Q.E.D.

8 × 10−3 = 6.37 GHz 2π × 0.2 × 10−12

This figure belongs to Problem 9.34, part (a).

Cm Vi

rp

Vp

Cp

gmVp

ro

CL

Vo

Figure 1

B

C

Vp

B

Cm rp

Cp

gm Vp

re

Cp

E Zi (s)

Sedra, Smith, Carusone, Gaudet

E Zi (s)

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–12

ft = | Adc | f3dB

Vo = 2Vi =

= 4000 × 265.3 = 1.06 GHz

Thus,

2 Vsig sCR

The Bode plot is shown in Figure 2.

Vo 2 = Vsig sCR

9.35

which is the transfer function of an ideal noninverting integrator.

(a) Refer to Fig. 1. Vi − 2Vi Vi Vi − Vo = =− R R R Thus,

9.36

Vi = −R Rin ≡ Ii

Vo = AVi

Ii =

From the figure we see that

From Miller’s theorem, we have

(b) Replacing the signal source with its equivalent Norton’s form results in the circuit in Fig. 2. Observe that Req = ∞ when Rsig = R. In this case, IL =

(1)

Ii R 10 k 2

Vsig Vsig = Rsig R

R1 1 k Ii

Vi

Vsig

Figure 2

A

Vo

Rin

Rin =

0

Vi Ii

R2 R2 R2 = = Vo Vo 1+A 1+ 1− Vi −Vi

(2)

Using the voltage divider rule at the input, we get 1 , (c) If ZL = sC V i = I L ZL

−Vi = Vsig

Rin Rin + R1

⇒ Vi = −Vsig

Vsig 1 × = R sC 1 Vsig = sCR and

Rin Rin + R1

(3)

For each value of A we use Eq. (2) to determine Rin , Eq. (3) to determine Vi (for Vsig = 1 V), Eq. (1) to determine Vo , and finally we calculate the

This figure belongs to Problem 9.35.

Ii

Rsig

2

IL Vsig

0

ZL

Vo

Vi

(ViVo)/R

R

Rin Figure 1

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–13

value of Vo /Vsig . The results are given in the table below. A (V/V)

Rin (k)

Vi (V)

Vo (V)

10

9.091 × 10−1

−0.476

100

9.901 × 10−2

−0.0900

1000

9.990 × 10−3

−9.89 × 10−3

Vo /Vsig (V/V)

−4.76 −4.76 −9.00 −9.00 −9.89 −9.89

10,000 9.999 × 10−4 −9.989 × 10−4 −9.99 −9.99

(e) Gv =

Rin Vo × Rin + Rsig Vb

2.14 × −10.8 = −7.4 V/V 2.14 + 1 Vo (f) Cin = Cπ + 1 + Cμ V =

b

where Cπ + Cμ =

gm 21.6 × 10−3 = 2π fT 2π × 600 × 106

Cπ + Cμ = 5.73 pF Cπ = 5.73 − 0.8 = 4.93 pF Cin = 4.93 + (1 + 10.8) × 0.8

9.37 VCC 1.5 V

= 14.37 pF (g) Rsig = Rin Rsig

IE

RC

RB IB

fH = IC

=

VBE

= 2.14 k 1 k = 0.68 k

IE

1 2π Cin Rsig

1 2π × 14.37 × 10−12 × 0.68 × 103

= 16.3 MHz 9.38 AM = 40 dB ⇒ 100 V/V

(a) For the dc analysis, refer to the figure. VCC = IE RC + IB RB + VBE 1.5 = IE × 1 +

IE × 47 + 0.7 β +1

1.5 − 0.7 = 0.546 mA 47 1+ 101 IC = αIE = 0.99 × 0.546 = 0.54 mA

⇒ IE =

(b) gm =

IC 0.54 mA = 21.6 mA/V = VT 0.025 V

A(s) = 100 1+

Since fP1 fP2 fZ , we have f3dB fP1 = 2 MHz The Bode plot is shown in the figure. Gain (dB) 50 40

β 100 = 4.63 k rπ = = gm 21.6

30

Vo (c) = −gm (RC RL ) Vb

10 0 10

(d) Using Miller’s theorem, the component of Rin due to RB can be found as

20

Rin1 =

47 k = 4 k 1 − (−10.8)

20 dB/decade

20

= −21.6(1 1) = −10.8 V/V

RB = 1 − (Vo /Vb )

1 + s/(2π × 200 × 106 ) s s 1 + 2π × 2 × 106 2π × 20 × 106

40 dB/decade

2

200

f (MHz) 20 dB/decade

30

9.39 τH = Cgs Rgs + Cgd Rgd + CL RCL where

Rin = Rin1 rπ

Cgs = 30 fF

= 4 4.63 = 2.14 k

Rgs = Rsig = 10 k

Sedra, Smith, Carusone, Gaudet

20

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–14

Cgd = 5 fF

=

Rgd = Rsig (1 + gm RL ) + RL

1 = 652 kHz 2π × 244 × 10−9

The estimate obtained using the open-circuit time constants is more appropriate as it takes into account the effect of CL . We note from Eq. (1) that τCL is 16.4% of τH , thus CL has a significant effect on the determination of fH .

= 10(1 + 2 × 20) + 20 = 430 k CL = 30 fF RCL = RL = 20 k Thus,

9.41 τgs = Cgs Rgs = Cgs Rsig

τH = 30 × 10 + 5 × 430 + 30 × 20

= 5 × 10 = 50 ns

= 3050 ps

Rgd = Rsig (1 + gm RL ) + RL

fH =

= 10(1 + 5 × 10) + 10

1 2π τH

= 520 k

1 = 2π × 3050 × 10−12

τgd = Cgd Rgd = 1 × 520 = 520 ns RCL = CL RL

= 52.2 MHz fZ =

gm 2 × 10−3 = = 63.7 GHz 2π Cgd 2π × 5 × 10−15

= 5 × 10 = 50 ns τH = τgs + τgd + τCL = 50 + 520 + 50 = 620 ns fH =

9.40 AM = −gm RL = −4 × 20 = −80 V/V

=

Cin = Cgs + Cgd (gm RL + 1) = 10.1 pF

Thus, the total contribution of Rsig is 50 + 510 = 560 ns 560 = 90.3% of τH . To double fH , we 620 must reduce τH to half of its value: which is

1 = 2π × 10.1 × 10−12 × 20 × 103 = 788 kHz

Now,

τgs = Cgs Rgs = Cgs Rsig

τH = Rsig [Cgs + Cgd (1 + gm RL )]

= 2 × 20 = 40 ns

+ Cgd RL + CL RL

Rgd = Rsig (1 + gm RL ) + RL

310 = Rsig [5 + 1(1 + 50)] + 1 × 10 + 5 × 10

= 20(1 + 80) + 20 = 1640 k

⇒ Rsig = 4.46 k

τgd = Cgd Rgd = 0.1 × 1640 = 164 ns τCL = CL RL

9.42 AM = −

= 2 × 20 = 40 ns

1 2π τH

Sedra, Smith, Carusone, Gaudet

1 × 620 = 310 ns 2

τH =

Using the open-circuit time constants, we get

fH =

1 = 257 kHz 2π × 620 × 10−9

Cgd [Rsig (1 + gm RL )] = 1 × 10(1 + 50) = 510 ns

Using the Miller approximation, we obtain 1 fH 2π Cin Rsig

= 40 + 164 + 40 = 244 ns

(1)

The interaction of Rsig with the input capacitance contributes all of τgs (50 ns) and a significant part of τgd , namely

= 2 + 0.1(80 + 1)

τH = τgs + τgd + τCL

1 2π τH

(1)

rπ gm RL rπ + Rsig

where rπ =

β 100 = 2.5 k = gm 40

Thus,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–15

AM = −

2.5 × 40 × 5 2.5 + 1

|A|, dB

= −143 V/V

60

3 dB

Using the Miller effect, we obtain 40

Cin = Cπ + Cμ (1 + gm RL ) = 10 + 0.3(1 + 40 × 5)

20 dB/decade

20

= 70.3 pF Rsig = rπ Rsig

0

= 2.5 1 = 0.71 k

f 10 kHz 100 kHz 1 MHz 10 MHz 100 MHz

0 5.7°

1 fH = 2π Cin Rsig

10 kHz 100 kHz 1 MHz

f

45°

1 = 2π × 70.3 × 10−12 × 0.71 × 103

90°

5.7°

= 3.2 MHz Using the open-circuit time constants, we get

Figure 1

τπ = Cπ Rπ = Cπ Rsig = 10 × 0.71 = 7.1 ns τμ = Cμ [Rsig (gm RL + 1) + RL ] = 0.3[0.71(40 × 5 + 1) + 5]] = 44.3 ns τCL = CL RL = 3 × 5 = 15 ns τH = τπ + τμ + τCL = 7.1 + 44.3 + 15 = 66.4 ns fH = =

Figure 2

1 2π τH

1 = 2.47 MHz 2π × 66.4 × 10−9

GB = 100 MHz

This is a more realistic estimate of fH as it takes into account the effect of CL .

9.44 AM = −

1 9.43 (a) A(s) = 1000 1 + s/(2π × 105 )

=−

(b) Figure 1 shows the Bode plot for the gain magnitude and phase.

RG gm RL RG + Rsig

0.5 × 5 × 10 0.5 + 0.1

= −41.7 V/V τgs = Cgs Rgs

(c) GB = 1000 × 100 kHz = 100 MHz

= Cgs Rsig

(d) The unity-gain frequency ft is

= Cgs (Rsig RG )

ft = 100 MHz

= 1 × 10−12 × (100 500) × 103

(e) Figure 2 shows the magnitude response when a second pole at 1 MHz appears in the transfer function. The unity-gain frequency ft now is ft = 10 MHz which is different from the gain-bandwidth product,

Sedra, Smith, Carusone, Gaudet

= 83.3 ns τgd = Cgd Rgd = Cgd [Rsig (1 + gm RL ) + RL ] = 0.2×10−12 [(100 500)(1+5×10)+10]×103 = 851.7 ns

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–16

τCL = CL RL −12

= 20 × 10

At node 3: × 10 × 10

Req3 = Ro2 RL = 2 k 1 k = 0.67 k

3

= 200 ns

Ceq3 = Co2 + CL = 2 + 7 = 9 pF

τH = τgs + τgd + τCL

fP3 =

1 2π Ceq3 Req3

fP3 =

1 2π × 9 × 10−12 × 0.67 × 103

= 83.3 + 851.7 + 200 = 1135 ns 1 fH = 2π τH =

= 26.4 MHz Thus, the three poles have frequencies 1.59 MHz, 7.94 MHz, and 26.4 MHz. Since the frequency of the second pole is more than two octaves higher than that of the first pole, the 3-dB frequency will be mostly determined by fP1 ,

1 2π × 1135 × 10−9

= 140 kHz

f3dB fP1 = 1.59 MHz 9.45 The figure below shows the equivalent circuit of the two-stage amplifier where we have modeled each stage as a transconductance amplifier. At node 1:

9.46 To lower fH from 56.8 MHz (see Example 9.5) to 40 MHz, τH must be increased to

Req1 = Rsig Rin1

τH =

= 10 k 10 k = 5 k Ceq1 = Cin1 + Csig = 10 + 10 = 20 pF Thus, fP1 = fP1 =

1 1 = 2π fH 2π × 40 × 106

= 3979 ps Now, τH = τgs + τgd + τCL

1

3979 = 400 + 2150 + τCL

2π Ceq1 Req1 1 2π × 20 × 10−12 × 5 × 103

⇒ τCL = 1429 ps But,

= 1.59 MHz

τCL = CL RL

At node 2:

1429 = CL × 10

Req2 = Ro1 Rin2 = 2 k 10 k

⇒ CL = 142.9 fF

= 1.67 k Ceq2 = Co1 + Cin2

Thus, the original CL of 25 fF must be increased by

= 2 + 10 = 12 pF

142.9 − 25 = 118 fF

fP2 = =

1 2π Ceq2 Req2

9.47 RL = ro RL = 20 k 12 k

1 2π × 12 × 10−12 × 1.67 × 103

= 7.94 MHz

= 7.5 k τgs = Cgs Rgs = Cgs Rsig = 0.2 × 100 = 20 ns

This figure belongs to Problem 9.45.

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–17

τgd = Cgd [Rsig (gm RL + 1) + RL ]

RL = ro /2

= 0.2[100(1.5 × 7.5 + 1) + 7.5]

Cgs = Cgd = 0.1 pF

= 246.5 ns

Using the Miller effect, we get

(a) CL = 0

Cin = Cgs + Cgd (gm RL + 1) 1 = 0.1 + 0.1 gm ro + 1 2 1 = 0.1 gm ro + 2 2

τCL = 0 τH = τgs + τgd = 20 + 246.5 = 266.5 ns fH =

1 = 597 kHz 2π × 266.5 × 10−9

(b) CL = 10 pF τCL = CL RL = 10 × 7.5 = 75 ns τH = τgs + τgd + τCL

=

1 1 2π × 0.1 gm ro + 2 (ro /2) × 10−9 2

=

1 1 gm ro + 2 ro × 10−9 0.1π 2

= 20 + 246.5 + 75 = 341.5 ns fH = =

1 2π Cin Rsig

fH =

1 2π τH

1 = 466 kHz 2π × 341.5 × 10−9

where ro is in k.

(c) CL = 50 pF

For initial design,

τCL = CL RL = 50 × 7.5 = 375 ns

gm = 2 mA/V, ro = 20 k

τH = τgs + τgd + τCL = 20 + 246.5 + 375 = 641.5 ns 1 fH = 2π τH =

1 1 × 40 + 2 × 20 × 10−9 0.1π 2

fH =

= 7.23 MHz

1 = 248 kHz 2π × 641.5 × 10−9

(i) For the case, I is reduced by a factor of 4: Since W gm = 2μn Cox ID L

Using the Miller approximation, since CL is not taken into account, then for all three cases we obtain

reducing ID by a factor of 4 reduces gm by a factor of 2,

Cin = Cgs + Cgd (gm RL + 1)

gm = 1 mA/V

= 0.2 + 0.2(1.5 × 7.5 + 1)

Since

= 2.65 pF

ro =

1 fH = 2π Cin Rsig 1 = 2π × 2.65 × 10−12 × 100 × 103 = 600 kHz which is very close to the estimate obtained using the method of open-circuit time constants for the case CL = 0. However, as CL is increased, the estimate obtained using the Miller approximation becomes less and less realistic, which is due to the fact that it does not take CL into account.

VA , ID

reducing ID by a factor of 4, increases ro by a factor of 4, ro = 80 k Thus, fH =

1 1 × 80 + 2 × 80 × 10−9 0.1π 2

= 0.95 MHz

9.48 CS amplifier with:

(ii) For the case, I is increased by a factor of 4, gm increases by a factor of 2,

Rsig = ro /2

gm = 4 mA/V

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–18

and ro decreases by a factor of 4,

= 239 MHz

ro = 5 k

fP2 =

Thus, fH =

1

1 0.1π × 20 + 2 × 5 × 10−9 2

=

1 2π(CL + Cgd )RL

1 2π(2 + 0.2) × 10−12 × 10 × 103

= 7.23 MHz Since fP1 fP2 , fP2 will be dominant and

= 53.1 MHz

fH fP2 = 7.23 MHz 9.49 Rsig = 5 k rπ =

β 100 = = 5 k gm 20

9.51 ro =

VA 10 = 100 k = ID 0.1

Rsig = rπ Rsig = 5 k 5 k

Rsig = ro /2 = 50 k

= 2.5 k

RL = ro = 100 k

τH = Cπ Rπ = Cπ Rsig

gm ro = 1.5 × 100 = 150

= 10 × 2.5 = 25 ns

Rin =

τμ = Cμ Rμ

r o + RL 100 + 100 = 1.32 k = 1 + gm ro 1 + 150

Ro = Rsig + ro + gm ro Rsig

= Cμ [Rsig (1 + gm RL ) + RL ]

= 50 + 100 + 150 × 50 = 7650 k

= 1 × [2.5(1 + 20 × 5) + 5]

τgs = Cgs (Rin Rsig )

= 257.5 ns

τgs = 0.2(1.32 50)

τCL = CL RL

= 0.26 ns

= 10 × 5 = 50 ns

Rgd = RL Ro

τH = τπ + τμ + τCL

= 100 7650 = 98.7 k

= 25 + 257.5 + 50 = 332.5 ns 1 1 = = 479 kHz 2π τH 2π × 332.5 × 10−9 rπ gm RL AM = − rπ + Rsig fH =

5 × 20 × 5 =− 5+5

τgd = (Cgd + CL + Cdb )Rgd = (0.015 + 0.03 + 0.02) × 98.7 = 6.42 ns τH = τgs + τgd = 0.26 + 6.42 = 6.68 ns 1 fH = 2π τH

= −50 V/V

1 = 23.8 MHz 2π × 6.68 × 10−9

9.50 Refer to Fig. 9.18(c).

=

1/gm Vo = gm RL 1 Vsig + Rsig gm

9.52 Ro = ro2 + ro1 + (gm2 ro2 )ro1 = 2ro + gm ro2

1/5 × 5 × 10 = 1 +1 5 = 8.3 V/V 1 fP1 = 1 2π Cgs Rsig gm fP1

= 2 × 20 + 2 × 20 × 20 = 840 k Av = −gm1 (Ro RL ) = −2(840 1000) = −913 V/V Using Eq. (9.92), we obtain

1 = 1 × 103 2π × 4 × 10−12 1 5

Sedra, Smith, Carusone, Gaudet

τH = Rsig [Cgs1 + Cgd1 (1 + gm1 Rd1 )] + Rd1 (Cgd1 + Cdb1 + Cgs2 )

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–19

+ (RL Ro )(CL + Cdb2 + Cgd2 )

f3dB =

ft 289.4 = = 57.9 kHz | AM | 5000

where we have assumed that CL does not include Cdb2 . Now,

If the cascode transistor is removed,

Rd1 = ro1 Rin2

AM = −gm (ro RL )

Rin2 = =

ro2 + RL 1 + gm2 ro2

where RL = ro RL

20 + 1000 = 24.9 k 1 + 2 × 20

VA 10 V = 50 k = ID 0.2 mA

ro =

Rd1 = 20 24.9 = 11.1 k

RL = (gm ro )ro = (2 × 50) × 50 = 5000 k

Thus,

RL = 505000 = 49.5 k

τH = 100[20 + 5(1 + 2 × 11.1)]

AM = −2 × 49.5 = −99 V/V

+ 11.1(5 + 5 + 20) + (1000 840)(20 + 5 + 5) = 13600 + 333 + 13696

f3dB = =

= 27, 629 ps = 27.63 ns 1 = 5.76 MHz fH = 2π × 27.63 × 10−9

1 2π(CL + Cgd )RL

1 2π(1 + 0.1) × 10−12 × 49.5 × 103

= 2.9 MHz ft =| AM | f3dB = 99 × 2.9

9.53 20 log | AM | = 74 dB

= 287.1 MHz

⇒ | AM | = 5000

which is almost equal to ft of the cascode. Thus, removing the cascode transistor lowers | AM | by a factor of almost 50 and increases f3dB by the same factor, keeping ft constant.

Ro (gm ro )ro RL = Ro | AM | = gm (RL Ro ) 5000 =

1 gm Ro 2

9.54 Refer to Fig. 9.22. gm =

1 = (gm ro )2 2 ⇒ gm ro = 100

β 100 = 2.5 k = gm 40 rπ AM = − gm (βro RL ) rπ + rx + Rsig rπ =

2VA = 100 VOV 2 × 10 = 0.2 V 100 W 1 2 ID = μn Cox VOV 2 L ⇒ VOV =

=−

Rsig = rπ (rx + Rsig ) = 2.5 (0.05 + 5) = 1.67 k Rπ 1 = Rsig = 1.67 k ro2 + RL Rc1 = ro1 re2 ro2 + RL /(β2 + 1) ⎤ ⎡

where

ft =

2ID 2 × 0.2 = 2 mA/V = VOV 0.2

2 × 10−3 2π(1 + 0.1) × 10−12

= 289.4 MHz

Sedra, Smith, Carusone, Gaudet

2.5 40(100 × 100 2) 2.5 + 0.05 + 5

−26.5 (V/V)

1 = × 0.2 × 50 × 0.22 2 = 0.2 mA gm ft = 2π(CL + Cgd )

gm =

IC 1 mA = = 40 mA/V VT 0.025 V

100 + 2 ⎥ ⎢ = 100 ⎣0.025 2 ⎦ 100 + 101 = 25.5 Rμ1 = Rsig (1 + gm1 Rc1 ) + Rc1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–20

= 1.67(1 + 40 × 0.0255) + 0.0255

fT =

= 3.4 k

gm 2π(Cπ + Cμ )

40 × 10−3 2π(10 + 1) × 10−12

Ro = β2 ro2 = 100 × 100 = 10,000 k

=

τH = Cπ 1 Rπ 1 + Cμ1 Rμ1 + (Ccs1 + Cπ 2 )Rc1

= 579 MHz

+ (CL + Ccs2 + Cμ2 )(RL Ro ) 9.56 Refer to Fig. 9.18 with

= 10 × 1.67 + 2 × 3.4 + (0 + 10) × 0.0255

RL = r o

+ (0 + 0 + 2)(2 10,000)

Rsig = ro /2

= 16.7 + 6.8 + 0.255 + 4 = 27.8 ns fH =

CL = Cgs

1 1 = 2π τH 2π × 27.8 × 10−9

Now,

= 5.7 MHz

Rin =

9.55 Replacing the BJT with its high-frequency T model while neglecting ro results in the equivalent circuit shown in the figure below.

=

and the other at the output, given by 1 2π(CL + Cμ )RL

Q.E.D.

(b) IC = 1 mA,

2Cgs gm

Ro = ro + Rsig + gm ro Rsig

1 re = 25 gm

1 1 Ro = ro + ro + gm ro ro 2 2

1 2π × 10 × 10−12 (1 0.025) × 103

= 652.5 MHz fP2 =

2 ro = 1 gm gm ro 2

τgs = Cgs Rgs

IC 1 mA = 40 mA/V gm = = VT 0.025 V

fP1 =

2 ro 2 gm

ro 2 × ro 2 gm = = 2 ro 1 + gm ro + 2 2 gm 2

1 = 2π Cπ (Rsig re )

fP2 =

2ro 2 = gm ro gm

Rgs = Rsig Rin

(a) There are two separate poles, one at the input given by fP1

r o + RL ro + ro = 1 + gm ro 1 + gm ro

1 gm ro2 2

Rgd = RL Ro

1 2π(1 + 1) × 10−12 × 10 × 103

= ro

= 7.96 MHz

1 gm ro2 2

ro

Since fP2 fP1 , fP2 will be dominant and

τgd = (CL + Cgd )Rgd

fH fP2 = 7.96 MHz This figure belongs to Problem 9.55.

Rsig

E

C

Vo

Vsig

re

RL

Cp Vp

(CLCm)

gmVp

B

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–21

This figure belongs to Problem 9.58. S Isig

Csig

Cgs

1 gm

Vgs

gmVgs

RL

Vo (Cgd+CL)

= (Cgs + Cgd )ro

9.58 The figure above shows the highfrequency equivalent circuit of the amplifier where the transistor has been replaced with its T model with ro neglected. The transfer function Vo /Isig can be found as follows:

τH = τgs + τgd =

2Cgs + (Cgs + Cgd )ro gm

Since gm ro 1, we obtain

Vgs = −Isig /Yin

τH (Cgs + Cgd )ro

where

and

Yin = gm + s(Csig + Cgs )

1 fH = 2π τH

thus Vgs = −

1 fH = 2π(Cgs + Cgd )ro Since

1 RL

1 + s(CL + Cgd )

(1)

(2)

Combining Eqs. (1) and (2), we get

then fT gm ro

Isig gm + s(Csig + Cgs )

Vo = −gm Vgs

gm fT = 2π(Cgs + Cgd )

fH =

Vo = Isig 1+s

Q.E.D.

Csig +Cgs gm

RL

[1 + s(CL + Cgd )RL ]]

Thus, the two poles have the frequencies gm ωP1 = Csig + Cgs 9.57 Refer to Example 9.9. To reduce fH to 200 MHz, τH must become 1 1 τH = = 2π fH 2π × 200 × 106 = 795.8 ps Since τgs remains constant at 26.6 ps, τCL must be increased to τCL = 795.8 − 26.6 = 769.2 ps

1 (CL + Cgd )RL

9.59 (a) AM = −gm RL where RL = RL ro = 20 20 = 10 k AM = −4 × 10 = −40 V/V

τCL = (Cgd + CL )Rgd

τgs = Cgs Rgs

thus,

= Cgs Rsig = 2 × 20 = 40 ns

769.2 = (5 + CL ) × 18.7

Rgd = Rsig (1 + gm RL ) + RL

⇒ CL + 5 = 41.1 fF

= 20(1 + 4 × 10) + 10

CL = 36.1 fF Thus, the amount of additional capacitance to be connected at the output is

Sedra, Smith, Carusone, Gaudet

ωP2 =

Thus,

But,

36.1 − 25 = 11.1 fF

and

= 830 k τgd = Cgd Rgd = 0.3 × 830 = 249 ns τCL = CL RL = 1 × 10 = 10 ns

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–22

τH = τgs + τgd + τCL

gm ro

= 40 + 249 + 10 = 299 ns

Thus, the gain increases by a factor of 2.

1 fH = 2π τH

fH =

=

1 = 532 kHz 2π × 299 × 10−9

where Cin = Cgs + Cgd (1 + gm Rd1 )

GB ≡ | AM | fH = 40 × 532 = 21.3 MHz

Rd1 = ro Rin2

(b) AM = −gm1 (Ro RL )

= ro

where Ro = ro1 + ro2 + gm2 ro2 ro1

= ro

AM = −4(1640 20) = −79 V/V ro2 + RL 1 + gm2 ro2

= Cgs + 3Cgd

20 + 20 0.49 k = 1 + 4 × 20

τH = Rsig [Cgs1 + Cgd1 (1 + gm1 Rd1 )] + Rd1 (Cgd1 + Cdb1 + Cgs2 ) + (RL Ro )(CL + Cgd2 )

Thus,

= 20[2 + 0.3(1 + 4 × 0.48)]

1 Cgs + (gm ro )Cgd 2 N Cgs + 3Cgd

+ 0.48(0.3 + 0.2 + 2) + (20 1640)(1 + 0.3)

1 gm ro 2 ⇒ gm ro = 100

1 = 1.89 MHz 2π × 84.4 × 10−9

GB ≡ | AM | fH = 79 × 1.89 = 149 MHz

N=

Note the increase in gain, bandwidth, and GB.

=

9.60 (a) For the CS amplifier,

1 1 gm ro + 1 Rsig 2π Cgs + Cgd 2

| AM | = gm (Ro ro )

1+5 = 4.6 1 + 0.3

100 =

1 2π Cin Rsig

For the cascode amplifier,

1 × 100 × 0.1Cgs 2 Cgs + 3 × 0.1Cgs

Cgs +

(c) gm ro =

1 gm ro 2

Q.E.D.

(b) 50 =

τH = 57.5 + 1.2 + 25.7 = 84.4 ns

=

(2)

From (1) and (2), the ratio N of fH of the cascode amplifier to fH of the CS amplifier is 1 Cgs + Cgd gm ro + 1 2 N= Cgs + 3Cgd

Using Eq. (9.92), we obtain

fH =

1 2π(Cgs + 3Cgd )Rsig

fH =

Rd1 = ro1 Rin2 = 20 0.49 = 0.48 k

| AM | = gm (ro ro ) =

ro + ro gm ro

2 2 gm gm 2 Cin = Cgs + Cgd 1 + gm × gm

= 2 × 20 + 4 × 20 × 20 = 1640 k

Rin2 =

RL + r o gm ro

Rd1 = ro

= 2ro + gm ro2

fH =

1 2π Cin Rsig

(1)

2VA VOV

2 × 10 VOV

⇒ VOV = 0.2 V W 1 2 VOV ID = μn Cox 2 L 1 × 0.4 × 10 × 0.22 2 = 0.08 mA = 80 µA =

= gm [(gm ro )ro ro ]

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–23

9.61 (a) From Fig. 9.29, we have gm ft = 2π(CL + Cgd ) √ 2μn Cox (W/L) = ID Q.E.D. 2π(CL + Cgd ) (b) gm = 2μn Cox (W/L) ID ID VOV = 1 μn Cox (W/L) 2 VA ro = ID Ro = (gm ro )ro

= 16 MHz fP2 = (1) (2) (3)

1 2π(0 + 0 + 2) × 10−12 × 2 × 103

40 MHz fH = 1

1 1 + 2 2 fp1 fp2

1 1 =1 + 2 = 14.9 MHz 162 40 (ii) Rsig = 10 k

(4)

Rsig = rπ Rsig = 2.5 10 = 2 k

(5)

fP1 =

AM = −gm (Ro RL )

1 2π × 2 × 103 (10 + 4) × 10−12

= 5.7 MHz

= −gm (Ro Ro ) 1 = − gm Ro 2 Substituting

(6)

fP2 = 40 MHz 1 1 + 2 = 5.6 MHz fH = 1 5.72 40

μn Cox = 0.4 mA/V2 , W/L = 20, CL = 20 fF, Cgd = 5 fF and VA = 10 V in Eqs. (1)–(6), we obtain the results in the table on the next page.

9.63 Refer to Fig. 9.22. IC = 0.1 mA 0.1 mA = 4 mA/V 0.025 V β 100 = 25 k = rπ = gm 4 gm =

9.62 (a) Gain from base to collector of Q1 = −1. Thus, Cin = Cπ 1 + Cμ1 (1 + 1)

ro =

= Cπ 1 + 2Cμ1 fP1 = =

1 = 0.25 k gm rπ gm (βro RL ) AM = − rπ + Rsig

1 2π Rsig Cin 1

2π Rsig (Cπ 1

+ 2Cμ )

re Q.E.D.

At the output node, the total capacitance is (CL + Cc2 + Cμ2 ) and since ro is large, Ro will be very large, thus the total resistance will be RL . Thus the pole introduced at the output node will have a frequency fP2 , fP2 =

1 2π(CL + Cc2 + Cμ2 )RL

Q.E.D.

(b) I = 1 mA 1 mA = 40 mA/V 0.025 V β 100 = 2.5 k = rπ = gm 40

rπ gm (βro βro ) rπ + rπ

1 = − βgm ro 4 1 = − × 100 × 4 × 1000 4 = −100,000 V/V 1 rπ = 12.5 k 2

Rπ 1 = Rsig = 12.5 k

(i) Rsig = 1 k Rsig = rπ Rsig = 2.5 1 = 0.71 k 1 = 2π × 0.71 × 103 (10 + 2 × 2) × 10−12

Sedra, Smith, Carusone, Gaudet

=−

Rsig = rπ Rsig = rπ rπ =

gm =

fP1

VA 100 V = 1000 k = ID 0.1 mA

Rc1 = ro rπ

r o + RL gm ro

= 1000 25

1000 + 100 × 1000 4 × 1000

= 1000 25 25.25 = 1000 12.56 = 12.4 k

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–24

This table belongs to Problem 9.61, part (b). ID (mA) ft (GHz) VOV (V) gm (mA/V) ro (k) Ro (M) AM (V/V) fH (MHz) 0.1

8

0.16

1.26

100

12.6

−7938

1

0.2

11.5

0.22

1.80

50

4.5

−4050

2.8

0.5

18

0.35

2.83

20

1.13

−1600

11.3

Rμ1 = Rsig (1 + gm1 Rc1 ) + Rc1

RL = RL ro

= 12.5(1 + 4 × 12.4) + 12.4 = 645 k Ro = βro = 100 × 1000 = 100 M τH = Cπ 1 Rπ 1 + Cμ1 Rμ1 + (Ccs1 + Cπ 2 )Rc1

= 2 20 = 1.82 k AM =

+ (CL + Cc2 + Cμ2 )(RL Ro )

1 × 109 =

1.82 1.82 +

Ro = r o

To determine Cπ , we use gm fT = 2π(Cπ + Cμ )

fZ =

4 × 10−3 2π(Cπ + Cμ )

=

⇒ Cπ + Cμ = 0.64 pF

1 gmb

1 5

= 0.9 V/V

1 1 = 20 0.2 k = 200 gm 5

gm 2π Cgs

5 × 10−3 = 398 MHz 2π × 2 × 10−12

τH = 0.54 × 12.5 + 0.1 × 645 + 0.54 × 12.4

Next, we evaluate b1 and b2 : Cgs Cgs + CL Rsig + R b1 = Cgd + gm RL + 1 gm RL + 1 L

1 × 100 × 1000 2 = 6.8 + 64.5 + 6.7 + 5000 ns

= 5.96 + 0.54 = 6.50 × 10−9 s

Cπ = 0.64 − 0.1 = 0.54 pF

+ 0.1 ×

Obviously the last term, which is due to the pole at the output node, is dominant. The frequency of the output pole is fP =

1 = 31.8 kHz 2π × 5000 × 10−9

fH fP = 31.8 MHz Because the other poles are at much higher frequencies, an estimate of the unity-gain frequency can be found as ft = | AM | fP = 105 × 31.8 × 103 = 3.18 GHz This estimate of ft is not very good (too high!). The other three poles have frequencies much lower than 3.18 GHz and will cause the gain to decrease faster, reaching the 0 dB value at a frequency lower than 3.18 GHz. Also note that fT of the BJTs is 1 GHz and the models we use for the BJT do not hold at frequencies approaching fT .

9.64 AM =

b2 = =

2+1 2 20 + × 1.82 5 × 1.82 + 1 5 × 1.82 + 1

(Cgs + Cgd )CL + Cgs Cgd Rsig RL gm RL + 1

(2 + 0.1) × 1 + 2 × 0.1 × 20 × 1.82 5 × 1.82 + 1

= 8.3 × 10−18 √ √ b2 8.3 = 0.44 Q= = b1 6.5 Thus, the poles are real and their frequencies can be obtained by finding the roots of the polynomial (1 + b1 s + b2 s2 ) = 1 + 6.5 × 10−9 s + 8.3 × 10−18 s2 which are ωP1 = 0.21 × 109 rad/s and ωP2 = 0.57 × 109 rad/s Thus, ωP1 = 33.4 MHz 2π ωP2 = 90.7 MHz = 2π

fP1 =

RL RL +

= 0.1 +

1 gm

fP2

where

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–25

Since the two poles are relatively close to each other, an estimate of fH can be obtained using 1 1 + 2 fH = 1 2 fP1 fP2 = 31.6 MHz

=

1 2π Cπ re

1 2π × 3.1 × 10−12 × 25

2 GHz b1 =

9.65 Refer to Fig. 9.23(c). Replacing Cgs with an input capacitance between G and ground, we get Ceq = Cgs (1 − K) where K=

fZ =

gm RL 1 + gm RL

Cπ + Cμ 1 +

RL

re

Rsig + Cπ + CL 1 +

Rsig

rπ

RL

Rsig R 1+ L + re rπ 0.95 3.1 + 0.1 1 + × 1 + (3.1 + 0) × 0.95 0.025 × 10−9 = 1 0.95 + 1+ 0.025 2.5

= 0.25 × 10−9

then Ceq = Cgs /(1 +

gm RL )

b2 =

and the total input capacitance becomes Cin = Cgd + Ceq

=

Cgs = Cgd + 1 + gm RL The frequency of the input pole is 1

fP1 =

2π Rsig Cgd +

Cgs 1 + gm RL

fH fP1

(0 + 3.1 × 0.1) × 0.95 × 1 × 10−18 1 0.95 + 1+ 0.025 2.5

= 7.5 × 10−21 √ √ b2 7.5 × 10−21 Q= = = 0.346 b1 0.25 × 10−9 Thus, the poles are real and their frequencies can be found as the roots of the polynomial (1 + b1 s + b2 se ) = 1 + 0.25 × 10−9 s + 7.5 × 10−21 s2 s s 1+ = 1+ ωP1 ωP2

9.66 Refer to Fig. 9.25. IC = 1 mA gm = 40 mA/V, re = 25

⇒ ωP1 = 4.65 × 109 rad/s

100 = 2.5 k rπ = 40 VA 20 = 20 k = ro = IC 1 gm fT = 2π(Cπ + Cμ ) 2 × 109 =

[(Cπ + Cμ )CL + Cπ Cμ ]RL Rsig Rsig R 1+ L + re rπ

ωP2 = 28.7 × 109 rad/s Thus, fP1 = 740 MHz fP2 = 4.6 GHz

40 × 10−3 2π(Cπ + Cμ )

Thus, f3dB fP1 = 740 MHz

⇒ Cπ + Cμ = 3.2 pF Cπ = 3.2 − 0.1 = 3.1 pF

9.67 fH fP1

RL = RL ro = 1 20 = 0.95 k AM =

AM =

RL Rsig RL + re + β +1 0.95 0.95 + 0.025 +

1 101

= 0.96 V/V

Sedra, Smith, Carusone, Gaudet

where b1 = Cgd +

1 2π b1

Cgs Cgs + CL R R + sig gm RL + 1 gm RL + 1 L

For CL = 0, b1 = Cgd Rsig +

Cgs (Rsig + RL ) gm RL + 1

For Rsig RL ,

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–26

Cgs Rsig gm RL + 1 Cgs Rsig = Cgd + gm RL + 1

b1 Cgd Rsig +

RL 100

⇒

2

RL +1=0 −4 100

This equation results in two solutions, RL = 27 k and RL = 373 k

and

For the given numerical values,

The second answer is not very practical as it implies the transistor is operating at gm = 9/373 = 0.024 mA/V, a very small transconductance!. We will pursue only the first answer. Thus,

fH =

RL = 27 k

1 2π Rsig Cgd +

fH =

Cgs gm RL + 1

2π × 100 × 103 0.1 +

Q.E.D.

1

12 × 10−12 5 × (2 20) + 1

= 1.23 MHz

gm = 0.33 mA/V and the 3-dB frequency is found using Eq. (9.110): f3dB = f0 =

9.68 For a maximally flat response we have 1 Q= √ 2

ω3dB =

ω0 = ω3dB = 2π × 106 rad/s

=

Thus, the transfer function will be

=

s2

31.58 × 10 s2 + s 8.886 × 106 s + 39.48 × 1012 12

9.69 With gmb = 0 and ro large, we obtain RL RL

9.70 I = 0.4 mA W 1 2 VOV (a) ID = μn Cox 2 L 1 2 × 0.4 × 16VOV 2 ⇒ VOV = 0.25 V 0.2 =

gm =

and gm RL gm RL + 1

(b) ro =

gm RL gm RL + 1

0.9 =

RD ro = 10 100

√ Now, for a maximally-flat response, Q = 1/ 2. Using the expression for Q in Eq. (9.111), we get √ gm RL + 1 [(Cgs + Cgd )CL + Cgs Cgd ]Rsig RL [Cgs + Cgd (gm RL + 1)]Rsig + (Cgs + CL )RL

√

√ 9 + 1 [(10 + 1)10 + 10 × 1] × 100 × RL [10 + 1(9 + 1)] × 100 + (10 + 10)RL √ √ 1 3 RL √ = RL 2 10 1 + 100

Sedra, Smith, Carusone, Gaudet

VA 20 = ID 0.2

= 100 k

⇒ gm RL = 9

1 √ = 2

2ID 2 × 0.2 = VOV 0.25

= 1.6 mA/V

For AM = 0.9,

Q=

9+1 100 × 27[(10 + 1) × 10 + 10 × 1] × 106 × 10−24

f3dB = 884 kHz

0.8(2π × 106 )2 √ + s 2 × 2π × 106 + (2π × 106 )2

AM =

gm RL + 1 Rsig RL [(Cgs + Cgd )CL + Cgs Cgd ]

= 5.55 Mrad/s

dc gain × ω02 Vo (s) = ω0 Vi (s) + ω02 s2 + s Q =

1 √ 2π b2

= 9.1 k Ad = gm (RD ro ) = 1.6 × 9.1 = 14.5 V/V (c) Rsig small and the frequency response is determined by the output pole: fP2 = =

1 2π(CL + Cgd + Cdb )(RD ro )

1 2π(100 + 5 + 5) × 10−15 × 9.1 × 103

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–27

= 159 MHz

1 2π CL Ro

fP1 =

fH 159 MHz =

(d) Rsig = 40 k τgs = Cgs Rgs

=

= Cgs Rsig

1 2π CL (ro2 ro4 ) 1 2π × 0.2 ×

10−12 (100

100) × 103

1 = 15.9 MHz 2π × 0.2 × 50 × 10−9 gm3 fP2 = 2π Cm

= 40 × 10−15 × 40 × 103

=

= 1.6 ns Rgd = Rsig (gm RL + 1) + RL = 40(1.6 × 9.1 + 1) + 9.1

where

= 631.5 k

gm3 =

2ID I = = 1 mA/V | VOV | | VOV |

fP2 =

1 × 10−3 = 1.59 GHz 2π × 0.1 × 10−12

τgd = Cgd Rgd = 5 × 631.5 = 3.16 ns τCL = (CL + Cdb )RL

fZ =

= (100 + 5) × 9.1 = 955.5 ps = 0.96 ns

2gm3 = 2fP2 = 2 × 1.59 = 3.18 GHz 2π Cm

τH = τgs + τgd + τCL

9.73 At low frequencies,

= 1.6 + 3.16 + 0.96 = 5.72 ns

Ad = 100 V/V

fH = =

Acm = 0.1 V/V

1 2π τH

CMRR =

1 2π × 5.72 × 10−9

Ad = 1000 or 60 dB Acm

The first pole of CMRR is coincident with the zero of the common-mode gain,

= 27.8 MHz

fP1 = 1 MHz 9.71 The common-mode gain will have a zero at fZ = =

1 2π RSS CSS

The second pole is coincident with the dominant pole of the differential gain, fP2 = 10 MHz

1 2π × 100 × 103 × 1 × 10−12

= 1.59 MHz

A sketch for the Bode plot for the gain magnitude is shown in the figure. CMRR (dB)

Thus, the CMRR will have two poles, one at fZ , i.e. at 1.59 MHz, and the other at the dominant pole of Ad , 20 MHz. Thus, the 3-dB frequency of CMRR will be approximately equal to fZ ,

60

20 dB/decade

f3dB = 1.59 MHz 40 40 dB/decade

9.72 gm1,2 = =

2ID1,2 I = VOV1,2 VOV1,2

20

0.2 mA = 1 mA/V 0.2 V

ro2 = ro4 =

0 0.1

| VA | 10 V = 100 k = ID 0.1 mA

1 fP1

10 fP2

100

1000 f, MHz

Ad = gm1,2 (ro2 ro4 )

9.74 The common-mode gain will have a zero at

= 1(100 100) = 50 V/V

fZ =

Sedra, Smith, Carusone, Gaudet

1 2π × 1 × 106 × 1 × 10−12

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–28

= 159 kHz Thus, the CMRR will have two poles: The first will be coincident with the zero of Acm ,

RC

fP1 = 159 kHz

Vo /2

and the second will be coincident with the pole of Ad ,

Vid 2

Rsig

fP2 = 2 MHz

40 VA = I 0.1

9.75 RSS = = 400 k

= −49.8 V/V

CSS = 100 fF fZ =

(c) Cin = Cπ + Cμ (1 + gm RC )

1 2π CSS RSS

= 2.7 + 0.5(1 + 10 × 10)

1 = = 4 MHz 2π × 100 × 10−15 × 400 × 103

= 53.2 pF

If VOV of the current source is reduced by a factor of 2 while I remains unchanged, (W/L) must be increased by a factor of 4. Assume L remains unchanged, W must be increased by a factor of 4. Since CSS is proportional to W, its value will be quadrupled:

fH =

CSS = 400 fF The output resistance RSS will remain unchanged. Thus, fZ will decrease by a factor of 4 to become

1 2π Cin Rsig

where Rsig = rπ (Rsig + rx ) = 10 10.1 5 k fH =

1 = 598 kHz 2π × 53.2 × 10−12 × 5 × 103

GB = | Ad | fH = 49.8 × 598 = 29.8 MHz

fZ = 1 MHz 9.77 Ad = gm1,2 (ro2 ro4 ) 9.76 gm =

IC 0.25 mA = 10 mA/V = VT 0.025 V

β 100 = 10 k = gm 10 gm fT = 2π(Cπ + Cμ )

=

ro2 = ro4 =

gm 2π fT

10 × 10−3 2π × 500 × 106

= 3.2 pF

(a) The figure shows the differential half-circuit and its high-frequency equivalent circuit. Vo rπ =− gm RC (b) Ad ≡ Vid rπ + rx + Rsig 10 × 10 × 10 10 + 0.1 + 10

Sedra, Smith, Carusone, Gaudet

Ad =

| VA | | VOV |

fP1 =

1 2π CL Ro

where

Cπ = 3.2 − 0.5 = 2.7 pF

=−

2ID I = | VOV | | VOV |

2| VA | | VA | = I/2 I 2| VA | 2| VA | I Ad = | VOV | I I

rπ =

Cπ + Cμ =

gm1,2 =

Ro = ro2 ro4 = gm3 fP2 = 2π Cm

| VA | I fP1 = I 2π CL | VA |

(1)

where gm3 =

I 2ID = | VOV | | VOV |

Cm =

CL 4

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–29

This figure belongs to Problem 9.76.

Rsig

Cm

rx

Vo/2

Vid 2

Vp

rp

Cp

RC gmVp

fP2 = fZ =

4I 2π CL | VOV |

(2)

2gm3 8I = 2fP2 = 2π Cm 2π CL | VOV |

= 6.37 GHz (3)

Q.E.D.

Since fP2 = 4 Ad fP1 , the unity-gain frequency ft is equal to GB, thus ft = Ad fP1 =

I | VA | | VOV | 2π CL | VA |

I/| VOV | ft = 2π CL gm = 2π CL

fZ = 2fP2 = 12.7 GHz A sketch of the Bode plot for | Ad | is shown in the figure.

Dividing (2) by (1), we obtain fP2 | VA | = 4 Ad =4 fP1 | VOV |

fP2 = 4Ad fP1 = 4 × 100 × 15.9

9.78 See figure below. The mirror high-frequency equivalent circuit is shown in the figure. Note that we have neglected ro . The model of the diode-connected transistor Q1 reduces to re1 in parallel with Cπ 1 . To obtain the current-transfer function Io (s)/Ii (s), we first determine Vπ in terms of Ii . Observe that the short-circuit at the output causes Cμ2 to appear in parallel with Cπ 1 and Cπ 2 . Thus,

Q.E.D.

Vπ =

For the numerical values given, we have

Ii (s)

20 = 100 V/V Ad = 0.2 gm =

I VOV

=

0.2 = 1 mA/V 0.2

fP1

I = 2π CL | VA |

=

0.2 × 10−3 2π × 100 × 10−15 × 20

= 15.9 MHz

1 1 + re1 rπ 2

1

(1)

+ s(Cπ 1 + Cπ 2 + Cμ2 )

At the output node we have Io (s) = gm2 Vπ − sCμ2 Vπ

(2)

Combining Eqs. (1) and (2) gives gm2 − sCμ2 Io (s) = 1 1 Ii (s) + s(Cπ 1 + Cπ 2 + Cμ2 ) + re1 rπ 2 Since the two transistors are operating at approximately equal dc bias currents, their small-signal parameters will be equal, thus Io (s) gm − sCμ = 1 1 Ii (s) 1+ + s(2Cπ + Cμ ) re β +1 =

gm re 1 − s(Cμ /gm ) 1 1 1+ 1 + s (2Cπ + Cμ )re / 1 + β +1 β +1

= ft = 15.9 × 100 = 1.59 GHz

Sedra, Smith, Carusone, Gaudet

1 − s(Cμ /gm ) 1 1 1+ 1 + s (2Cπ + Cμ )re / 1 + β +1 β +1 α

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–30

This figure belongs to Problem 9.78.

sCm2Vp

Vp re1

Ii (s)

Cp1

Cp2

rp2 gm2Vp

=

1 − s(Cμ /gm ) 1 + s (2Cπ + Cμ )re / 1 +

1 1 + 2/β

= 40 40 = 20 k 1 β +1

Thus we see that the low-frequency transmission is Io (0) = Ii

1 1+

2 β

τCL = CL RL = 1 × 20 = 20 ns τH = τgs + τgd + τCL

IC 1 mA = 40 mA/V = VT 0.025 V

re 25

40 × 10 2π × 500 × 106

=

Cπ = 12.7 − 2 = 10.7 pF 1 2π(2 × 10.7 + 2) × 10−12 × 25]

where RL = RL ro

Sedra, Smith, Carusone, Gaudet

5 = 1.67 mA/V 1 + 5 × 0.4

Ro = ro (1 + gm Rs ) = 40(1 + 5 × 0.4) = 120 k AM = −Gm RL = −1.67 × 30 = −50 V/V Rgd = Rsig (1 + Gm RL ) + RL

40 × 10−3 = 3.18 GHz 2π × 2 × 10−12

9.79 (a) AM = −gm RL

1 = 603 kHz 2π × 264 × 10−9

RL = RL Ro = 40 120 = 30 k

1 × 1012 2π × 23.4 × 25

= 272 MHz gm fZ = 2π Cμ =

fH =

(b) With Rs = 400 , gm Gm = 1 + gm Rs

= 12.7 pF

=

= 40 + 204 + 20 = 264 ns

GB = 100 × 603 = 60.3 MHz

gm 2π fT −3

fP =

Rgd = Rsig (1 + gm RL ) + RL

= 204 ns

For the numerical values given,

=

= 2 × 20 = 40 ns

τgd = Cgd Rgd = 0.1 × 2040

and the zero is at gm fZ = 2π Cμ

Cπ + Cμ =

τgs = Cgs Rgs = Cgs Rsig

= 2040 k

1 2π(2Cπ + Cμ )re

gm =

AM = −5 × 20 = −100 V/V

= 20(1 + 5 × 20) + 20

as expected. The pole is at fP , fP

Io (s)

Cm2

C1, B1, B2

= 20(1 + 1.67 × 30) + 30 = 1050 k τgd = Cgd Rgd = 0.1 × 1050 = 105 ns τCL = CL RCL = CL RL

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–31

AM = −Gm RL

= 1 × 30 = 30 ns Rgs =

Rsig + Rs + Rsig Rs /(ro + RL ) ro 1 + gm Rs r o + RL

20 =

5 [20 20(1 + gm Rs )] 1 + gm Rs

4(1 + gm Rs ) =

20 + 0.4 + 20 × 0.4/(40 + 40) = 40 1 + 5 × 0.4 40 + 40

20 × 20(1 + gm Rs ) 20 + 20(1 + gm Rs )

⇒ 1 + gm Rs = 4 ⇒ Rs =

= 10.25 k τgs = Cgs Rgs = 2 × 10.25 = 20.5 ns

3 = 0.6 k = 600 gm

τH = τgs + τgd + τCL

9.81 (a) Refer to Fig. P9.81(a). Since the total resistance at the drain is ro , we have

= 20.5 + 105 + 30 = 155.5 ns

AM = −gm ro

1 fH = 2π τH

τgs = Cgs Rgs = Cgs Rsig

=

Q.E.D.

Rgd = Rsig (1 + gm RL ) + RL

1 = 1.02 MHz 2π × 155.5 × 10−9

= Rsig (1 + gm ro ) + ro

GB = 51.2 MHz

τgd = Cgd Rgd = Cgd [Rsig (1 + gm ro ) + ro ] τCL = CL RL = CL ro

9.80 (a) GB = | AM | fH

Thus,

1 = 2π Cgd Rsig

τH = τgs + τgd + τCL

=

= Cgs Rsig + Cgd [Rsig (1 + gm ro ) + ro ]

1 2π × 0.2 ×

10−12

× 100 ×

103

+ CL ro

= 7.96 MHz

For the given numerical values,

(b) | AM | = 20 V/V fH =

AM = −1 × 20 = −20 V/V

7.96 = 398 kHz 20

τH = 20 × 20 + 5[20(1 + 1 × 20) + 20] + 10 × 20

(c) A0 = gm ro

= 400 + 2200 + 200 = 2800 ps = 2.8 ns

100 = 5 × ro

fH =

⇒ ro = 20 k Gm =

Q.E.D.

gm 5 = 1 + gm Rs 1 + gm Rs

1 1 = 2π τH 2π × 2.8 × 10−9

= 56.8 MHz

Ro = ro (1 + gm Rs ) = 20(1 + gm Rs )

GB = 20 × 56.8 = 1.14 GHz

RL = RL Ro = 20 20(1 + gm Rs )

(b) From Fig. 1 we see that

This figure belongs to Problem 9.81, part (b).

Rsig

Vsig

Vg1 1 gm1

Q1

Vo

ro1

Vg2

Q2

ro2

Figure 1

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–32

Vg1 =1 Vsig

Vg1 = Ix Rsig Vs1 = Vg1 − Vx = Ix Rsig − Vx

ro1 Vg2 = 1 Vg1 + ro1 gm1

Node equation at S1 , Ix = gm1 Vx −

Vo = −gm2 ro2 Vg2

Ix Rsig − Vx ro1 Rsig 1 = Vx gm1 + Ix 1 + ro1 ro1 = gm1 Vx −

Thus, ro1 × −gm2 ro2 1 + ro1 gm1

AM = 1 ×

ro1 =− (gm2 ro2 ) 1/gm1 + ro1

Thus,

Q.E.D.

Rgs1 ≡

Next we evaluate the open-circuit time constants. Refer to Fig. 2.

Q1

Cgs1

Cgd2

Rgs2 =

Q2

G2

Rsig + ro1 1 + gm1 ro1

Cgs2 : Capacitor Cgs2 sees the resistance between G2 and ground, which is the output resistance of source follower Q1 ,

Cgd1 Rsig G 1

Rsig + ro1 Vx = Ix 1 + gm1 ro1

τgs1 = Cgs1 Rgs1 = Cgs1

CL

1 ro1 gm1

Thus,

Cgs2

τgs2 = Cgs2

Figure 2 Cgd1 : Capacitor Cgd1 is between G1 and ground and thus sees the resistance Rsig , Rgd1 = Rsig τgd1 = Cgd1 Rsig Cgs1 : To find the resistance Rgs1 seen by capacitor Cgs1 , we replace Q1 with its hybrid-π equivalent circuit with Vsig set to zero, Cgd1 = 0, and Cgs1 replaced by a test voltage Vx . The resulting equivalent circuit is shown in Fig. 3.

Rsig

Vs1 ro1

D1

Vg1 G1

1 ro1 gm1

Cgd2 : Transistor Q2 operates as a CS amplifier with an equivalent signal-source resistance equal to the output resistance of the source follower Q2 , 1 that is, ro1 and with a gain from gate to gm1 drain of gm2 ro2 . Thus, the formula for Rgd in a CS amplifier can be adapted as follows: 1 ro1 (1 + gm2 ro2 ) + ro2 Rgd2 = gm1 and thus,

τgd2 = Cgd2

1 ro1 (1 + gm2 ro2 ) + ro2 gm1

CL : Capacitor CL sees the resistance between D2 , and ground which is ro2 , τCL = CL ro2

Ix Vx

ro1

gm1Vx

Summing τgd1 , τgs1 , τgs2 , τgd2 and τCL gives τH in the problem statement. Q.E.D. For the given numerical values:

S1

Ix

Vs1

AM = −

20 (1 × 20) 1 + 20

= −19 V/V

Figure 3

τgd1 = Cgd1 Rsig = 5 × 20 = 100 ps Analysis of the circuit in Fig. 3 proceeds as follows:

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–33

τgs1 = Cgs1

Rsig + ro1 1 + gm ro1

= 6.37 MHz Since fP2 10fP1 , the pole at fP1 will dominate and

20 + 20 = 38 ps 1 + 1 × 20 1 = Cgs2 ro1 gm1

= 20 τgs2

fH fP1 = 637 kHz

= 20 × (1 20) = 19 ps 1 ro1 (1 + gm ro2 ) + ro2 τgd2 = Cgd2 gm1 = 5[(1 20)(1 + 20) + 20]

9.83 Refer to Eqs. (9.133)–(9.136). For our case, gm (1) Gm = 1 + gm Rs Ro = very large RL = RL Ro = RL

= 200 ps

−gm RL 1 + gm Rs

τCL = CL ro2 = 10 × 20 = 200 ps

AM = −Gm RL =

τH = 100 + 38 + 19 + 200 + 200 = 557 ps

Refer to Eqs. (9.135) – (9.138). For our case,

1 fH = 2π τH

Rgd = Rsig (1 + Gm RL ) + RL Rsig + Rs 1 + gm Rs

1 = 2π × 557 × 10−12

Rgs =

= 286 MHz

τH = Cgs Rgs + Cgd Rgd

GB = 19 × 286 = 5.43 GHz Thus, while the dc gain remained approximately the same both fH and GB increased by a factor of about 5!

9.82 gm = =

For the numerical values given: (a) Rs = 0

τH = 10 × 100 + 2 × 2605 = 6210 ns fH =

1 1 gm RD = × 1 × 50 = 25 V/V 2 2 The high-frequency analysis can be performed in an analogous manner to that used in the text for the bipolar circuit. Refer to Fig. 9.35(b) and adapt the circuit for the MOS case. Thus, 1 = Cgs + Cgd 2π Rsig 2

1 = 25.6 kHz 2π × 6210 × 10−9

GB = 25.6 × 25 = 641 kHz (b) Rs = 100 Gm =

5 = 3.33 mA/V 1 + 5 × 0.1

AM = −3.33 × 5 = −16.7 V/V Rgd = 100(1 + 3.33 × 5) + 5 = 1771.7 k

1 = 4 2π × 100 × 103 + 0.5 × 10−12 2 = 637 kHz

Rgs =

100 + 0.1 = 66.7 k 1 + 5 × 0.1

τH = 10 × 66.7 + 2 × 1771.7 = 4211 ns fH =

and

1 = 37.8 kHz 2π × 4211 × 10−9

GB = 631 kHz

1 2π RD Cμ

(c) Rs = 200

1 2π × 50 × 103 × 0.5 × 10−12

Sedra, Smith, Carusone, Gaudet

1 2π τH

Rgd = 100(1 + 5 × 5) + 5 = 2605 k

=

=

(4)

AM = −gm RL = −5 × 5 = −25 V/V

I 0.2 mA = 1 mA/V = VOV 0.2 V

fP2 =

(3)

Gm = gm = 5 mA/V

2ID 2(I/2) = VOV VOV

Vo RD = Vsig 2/gm

fP1

fH =

(2)

Gm =

5 = 2.5 mA/V 1 + 5 × 0.2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–34

AM = −2.5 × 5 = −12.5 V/V

τgs = Cgs Rgs =

Rgd = 100(1 + 2.5 × 5) + 5 = 1355 k 100 + 0.2 = 50.1 k Rgs = 1 + 5 × 0.2 τH = 10 × 50.1 + 2 × 1355 = 3211 ns 1 = 49.6 kHz 2π × 3211 × 10−9

fH =

GB = 49.6 × 12.5 = 620 kHz A summary of the results is provided in the following table:

Cgs Rsig 1 + (k/2)

Rgd = Rsig (1 + Gm RL ) + RL Utilizing the expressions for RL and Gm RL derived earlier, we obtain A0 1+k + ro Rgd = Rsig 1 + 2+k 2+k τgd = Cgs Rgd = A0 1+k + Cgd ro Cgd Rsig 1 + 2+k 2+k τCL = CL RL

Rs = 0 Rs = 100 Rs = 200

= CL ro

1+k 2+k

| AM | (V/V)

25

16.7

12.5

fH (kHz)

25.6

37.8

49.6

Thus,

GB (kHz)

641

631

620

τH = τgs + τgd + τCL Cgs Rsig A0 + Cgd Rsig 1 + 1 + (k/2) 2+k 1+k 1+k + CL ro +Cgd ro 2+k 2+k Cgs Rsig A0 = + Cgd Rsig 1 + 1 + (k/2) 2+k 1+k Q.E.D. +(CL + Cgd )ro 2+k =

Observe that increasing Rs trades off gain for bandwidth while GB remains approximately constant. 9.84 Gm =

gm gm = 1 + gm Rs 1+k

RL = RL Ro = ro ro (1 + gm Rs ) = ro ro (1 + k) =

ro × ro (1 + k) ro + ro (1 + k)

= ro

9.85 Substituting the given numerical values in the expressions for AM and τH given in the statement for Problem 9.84 and noting that A0 = gm ro = 5 × 40 = 200, we obtain

1+k 2+k

AM = −Gm RL = −

gm ro 2+k

AM = −

Thus, −A0 AM = 2+k

τH = Q.E.D.

Rsig + Rs + Rsig Rs /(ro + RL ) Rgs = ro 1 + gm Rs r o + RL Rsig + Rs + Rsig Rs /2ro 1 1 + gm Rs 2 For Rsig Rs , =

Rgs

Rsig (1 + Rs /2ro ) 1 + (k/2)

For ro Rs , Rgs

Rsig 1 + (k/2)

Sedra, Smith, Carusone, Gaudet

200 V/V 2+k

400 1+k 40 +2+ + 44 , ns 1 + (k/2) 2+k 2+k

fH = 1/2π τH GB = |AM |fH To obtain fH = 2 MHz, we see from the table (next page) that k = 11 Thus, 1 + gm Rs = 11 ⇒ Rs =

10 = 2 k 5

The gain achieved is | AM | = 15.4 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–35

k

| AM |, V/V τH ns fH (MHz) GB (MHz)

Vb2 (Rin2 ro1 ) = Vb1 (Rin2 ro1 ) + re1

0

100

264

0.603

60.3

1

66.7

191.3

0.832

55.6

2

50

155

1.03

51.5

3

40

133.2

1.19

47.6

4

33.3

118.7

1.34

44.6

5

28.6

108.3

1.47

42.0

6

25

100.5

1.58

39.5

7

22.2

94.4

1.69

37.5

8

20

89.6

1.78

35.6

9

18.2

85.7

1.86

33.9

To determine fH we use the method of open-circuit time constants. Figure 9.33(b) shows the circuit with Vsig = 0 and the four capacitances indicated. Again, recall that here RL = ro2 . Also, in our present circuit there is a capacitance CL at the output.

10

16.7

82.3

1.93

32.2

Capacitance Cμ1 sees a resistance Rμ1 ,

11

15.4

79.6

2.00

30.8

Rμ1 = Rsig Rin

12

14.3

77.2

2.06

29.5

= 10 k 2.5 M 10 k

13

13.3

75.1

2.12

28.2

14

12.5

73.3

2.17

27.1

To find the resistance Rπ 1 we refer to the circuit in Fig. 9.33(c) where Rin2 is considered to include ro2 ,

15

11.8

71.6

2.22

26.2

=

25 1000 = 0.99 1 V/V (25 1000) + 0.25

Vo = −gm2 ro2 = −4 × 1000 = −4000 V/V Vb2 Thus, AM =

Vo = −4000 V/V Vsig

Rin2 = 25 k 1000 k = 24.4 k We use the formula for Rπ 1 given in Example 9.10:

9.86 At an emitter bias current of 0.1 mA, Q1 and Q2 have

Rπ 1 =

gm = 4 mA/V re = 250 rπ =

Rπ 1 =

β 100 = 25 k = gm 4

10 + 24.4 = 347 10 24.4 + 1+ 25 0.25

Capacitance Cπ 2 sees a resistance Rπ 2 :

VA 100 = 1000 k = IC 0.1 gm Cπ + Cμ = 2π fT ro =

Rπ 2 = Rin2 Rout1 Rsig = rπ 2 ro1 re1 + β1 + 1 10 = 25 1000 0.25 + 101

4 × 10−3 = 3.2 pF 2π × 200 × 106 Cμ = 0.2 pF =

= 344

Cπ = 3 pF To determine Rin and the voltage gain AM , refer to the circuit in Fig. 9.33(a). Here, however, RL is ro2 . Rin2 = rπ 2 = 25 k

Capacitance Cμ2 sees a resistance Rμ2 : Rμ2 = (1 + gm2 ro2 )(Rin2 Rout1 ) + ro2 = (1 + 4 × 1000) × 0.344 + 1000 = 2376 k

Rin = (β1 + 1)[re1 + (ro1 Rin2 )] = 101[0.25 + (1000 25)]

We can determine τH from τH = Cμ1 Rμ1 + Cπ 1 Rπ 1 + Cμ2 Rμ2

2.5 M Rin 2.5 M Vb1 1 V/V = = Vsig Rin + Rsig 2.5 M + 10 k

Sedra, Smith, Carusone, Gaudet

Rsig + Rin2 Rsig Rin2 1+ + rπ 1 re1

+ C π 2 Rπ 2 + C L r o = 0.2 × 10 + 3 × 0.347 + 0.2 × 2376

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–36

+ 3 × 0.344 + 1 × 1000

fP1 = 2π Rsig

τH = 2 + 1 + 475.2 + 1 + 1000 = 1479.2 ns Observe that there are two dominant capacitances: the most significant is CL and the second most significant is Cμ2 . 1 fH = 2π τH =

2π × 20 × 103

IC 1 mA = 40 mA/V VT 0.025 V

re 25

1 2π(Cgd + CL )RL

fP2 =

1 2π × (0.1 + 1) × 10−12 × 20 × 103

= 7.2 MHz Thus,

β 120 = 3 k = gm 40

fP1 = fP2 = 7.2 MHz The midband gain AM is obtained as

Rin = 2rπ = 6 k Vo αRL Rin = Vsig Rin + Rsig 2re

AM =

10 6 × = 66.7 V/V 6 + 12 2 × 0.025 gm Cπ + Cμ = 2π fT

=

40 × 10−3 = = 12.7 pF 2π × 500 × 106

Rsig = Rsig 2rπ = 12 6 = 4 k

fP1 = 2π Rsig

Vo (s) = Vsig (s)

1 Cπ + Cμ 2

= 2π × 4 × 103

1 12.2 + 0.5 × 10−12 2

= 6 MHz 1 2π RL Cμ

=

50 1+

s 2π × 7.2 × 106

50

ω 2π × 7.2 × 106 Vo 50 At ω = ω3dB , = √ , thus Vi 2 2 √ ω3dB 2=1+ 2π × 7.2 × 106 √ f3dB = 2 − 1 × 7.2 MHz sig

Cπ = 12.2 pF

RL 1 = gm RL 2/gm 2

1 × 5 × 20 = 50 V/V 2 Thus, the amplifier transfer function is

Vo V

Cμ = 0.5 pF

fP2 =

1 2 + 0.1 × 10−12 2

= 7.2 MHz,

=

rπ =

fP1 =

and a pole at the output with frequency fP2 ,

1 = 107.6 kHz 2π × 1479.2 × 10−9

9.87 gm =

1 Cgs + Cgd 2

2

2

1+

= 4.6 MHz 9.89 (a) For each of Q1 and Q2 , gm =

2 × 0.1 2ID = = 1 mA/V | VOV | 0.2

= 31.8 MHz

ro =

| VA | 10 = 100 k = ID 0.1

Thus, fP1 is the dominant pole and

gm ro = 100

=

1 2π × 10 × 103 × 0.5 × 10−12

fH fP1 = 6 MHz 9.88 Using an approach analogous to that utilized for the BJT circuit (Fig. 9.35), we see that there is a pole at the input with frequency fP1 :

Sedra, Smith, Carusone, Gaudet

Since the output resistance of each of the current sources is equal to ro , the gain of each stage is (gm ro /2) and the overall voltage gain is Vo = (gm ro )2 = 2500 V/V Vsig

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–37

(b) τgs1 = Cgs Rsig

The dc analysis is shown in Fig. 1. It is based on VS1 = VBE2 = 0.7 V. Neglecting IB2 , we obtain

= 20 × 10 = 200 ps

ID1 =

1 Rgd1 = Rsig (1 + gm1 ro1 ) + (ro1 /2) 2 = 10(1 + 50) + 50

0.7 V 0.1 mA 6.8 k

Q.E.D.

1 2 k (W/L)VOV 2 n 1 2 0.1 = × 2 × VOV 2 ⇒ VOV 0.3 V

ID1 =

= 560 k τgd1 = Cgd1 Rgd1 = 5 × 560 = 2800 ps At the drain of Q1 we have (Cdb1 + Cgs2 ) and the resistance seen is ro /2:

VGS = Vt + VOV = 1 + 0.3 = 1.3 V

τd1 = (Cdb1 + Cgs2 )(ro /2)

VG1 = 0.7 + 1.3 = 2 V

= (5 + 20) × 50 = 1250 ps r 1 o1 1 + gm2 ro2 + (ro2 /2) Rgd2 = 2 2

VC1 = VG1 = 2 V

= 50(1 + 50) + 50 = 2600 k

τd2 = Cdb2 (ro2 /2)

τH = τgs1 + τgd1 + τd1 + τgd2 + τd2

Q.E.D.

2ID1 2 × 0.1 = 0.67 mA/V = VOV 0.3

Cgs = Cgd = 1 pF gm2 =

= 5 × 50 = 250 ps

IC 1 mA = 40 mA/V = VT 0.025 V

β 200 = 5 k = gm2 40 gm2 Cπ 2 + Cμ2 = 2π fT2

rπ 2 =

= 200 + 2800 + 1250 + 13, 000 + 250 = 17,500 ps = 17.5 ns

=

5−2 = 1 mA 3

(b) gm1 =

τgd2 = Cgd2 Rgd2 = 5 × 2600 = 13,000 ps

fH =

IC2 =

1 2π τH

=

1 = 9.1 MHz 2π × 17.5 × 10−9

40 × 10−3 = 10.6 pF 2π × 600 × 106

Cμ2 = 0.8 pF Cπ 2 = 9.8 pF

9.90 (a)

(c) Q1 acts as a source follower, thus

5 V 52 1 mA 3 3 k RG 10 M 0 VG1 2 V

VGS 1.3 V 0.7 V 0.1 mA

1 mA

Q1

(6.8 5) = 0.66 V/V 1.5 + (6.8 5)

Vo = −gm2 (3 k 1 k) Vb2 = −40(3 1) = −30 V/V Thus,

0.1 mA 0 Q2 6.8 k

=

Neglecting RG , we obtain

2 V

Vb2 6.8 k rπ 2 = 1 Vi + (6.8 k rπ 2 ) gm1

Vo = 0.66 × −30 Vi −20 V/V Using Miller’s theorem, the input resistance Rin is found as

Figure 1

Sedra, Smith, Carusone, Gaudet

Rin =

RG 10 M = Vo 1 − (−20) 1− Vi

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–38

This figure belongs to Problem 9.90, part (d).

Rsig Rsig 100 k

Ix

Vx

⬅

gm1Vx

Cgs Ix Rs 6.8 k // rp2

Rs

Figure 2 τμ2 = Cμ2 Rμ2 = 0.8 × 31.4 = 25.1 ns

= 476 k Rin Vi = Vsig Rin + Rsig =

τH = τgd + τgs + τπ 2 + τμ2 = 82.6 + 35.1 + 9.7 + 25.1

476 = 0.83 V/V 476 + 100

= 152.5 ns

Vo = 0.83 × −20 = −16.5 V/V Vsig

fH =

(c) τgd1 = Cgd1 (Rin Rsig )

=

= 1 × 10−12 (476 100) × 103 = 82.6 ns To determine the resistance Rgs seen by Cgs , refer to Fig. 2. We can show that Rsig + Rs Vx = Rgs ≡ Ix 1 + gm1 Rs

9.91 All transistors are operating at IE = 0.5 mA. Thus, gm 20 mA/V

rπ =

β 100 = 5 k = gm 20

ro = very high (neglect)

Rs = 6.8 k rπ 2

Cπ + Cμ =

= 6.8 5 = 2.88 k 100 + 2.88 = 35.1 k 1 + 0.67 × 2.88

τgs = Cgs Rgs = 1 × 10−12 × 35.1 × 103 = 35.1 ns 1 τπ 2 = Cπ 2 rπ 2 6.8 k gm1 = 9.8 × 10−12 × 0.99 × 103 = 9.7 ns 1 Rμ2 = 6.8 k rπ 2 [1 + gm2 (3 1)] gm1 + (3 1) 3 = (1.5 6.8 5) 1 + 40 × + 0.75 4 = 31.4 k

1 = 1 MHz 2π × 152.5 × 10−9

re 50

where

Rgs =

1 2π τH

gm 20 × 10−3 = 2π fT 2π × 400 × 106

= 8 pF Cμ = 2 pF Cπ = 6 pF (a) CE amplifier rπ gm RL AM = − rπ + Rsig =−

5 × 20 × 10 = −66.7 V/V 5 + 10

Since CL = 0, we can obtain a good estimate of fH using the Miller approximation: Cin = Cπ + Cμ (gm RL + 1) = 6 + 2(20 × 10 + 1) = 408 pF

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–39

1 2π Rsig Cin

fH = =

1 2π × 10 × 103 × 408 × 10−12

= 39 kHz (b) This is a cascode amplifier. Refer to Fig. 9.22 for the analysis equations.a rπ gm (βro RL ) AM = − rπ + Rsig −

Since the two poles are relatively close to each other, we use the root-sum-of-the-squares formula to obtain an estimate for fH : 1 1 fH = 1/ 2 + 2 fP1 fP2 1 1 = 1/ + 2 = 5 MHz 8 6.42 AM =

rπ gm RL rπ + Rsig

= −66.7 V(V) (same as the CE in (a))

=

Rsig = rπ Rsig = 5 10 = 3.33 k τπ 1 = Cπ 1 Rπ 1 = 6 × 3.33 = 20 ns 1 = re2 = 50 gm2

= 3.33(1 + 2) + 0.05 = 10.05 k τμ1 = Cμ1 Rμ1 = 2 × 10.05 = 20.1 ns τc1 = Cπ 2 Rc1 = 6 × 0.05 = 0.3 ns τμ2 = Cμ2 RL = 2 × 10 = 20 ns τH = 20 + 20.1 + 0.3 + 20 = 60.4 ns 1 1 = 2π τH 2π × 60.4 × 10−9

= 2.6 MHz (c) This is a CC-CB cascade similar to the circuit analyzed in Fig. 9.35. There are two poles: one at the input, 1

2π(Rsig 2rπ ) fP1 = =

Cπ + Cμ 2

2 × 0.05 +

10 101

50 V/V

and one at the output, 1 2π RL Cμ

1 2π × 10 × 103 × 2 × 10−12

= 8 MHz

Sedra, Smith, Carusone, Gaudet

Rin = (β1 + 1)(re1 + rπ 2 )

= 0.98 V/V rπ 2 5 Vb2 = 0.99 V/V = = Vb1 rπ 2 + re1 5 + 0.05 Vo = −gm2 RL = −20 × 10 = −200 V/V Vb2 AM =

Vo = −0.98 × 0.99 × 200 Vsig

= −194 V/V Rμ1 = Rsig Rin = 10 510 = 9.81 k τμ1 = Cμ1 Rμ1 = 2 × 9.81 = 19.6 ns Rπ 1 =

=

Rsig + Rin2 Rsig Rin2 1+ + rπ 1 re1

10 + 5 = 0.15 k 5 10 + 1+ 5 0.05

τπ 1 = Cπ 1 Rπ 1 = 6 × 0.15 = 0.9 ns Rπ 2 = Rin2 Rout1 Rsig + re1 = rπ 2 β1 + 1 10 + 0.05 = 0.15 k =5 101

= 6.4 MHz

=

1 2π(10 10) × 103 (3 + 2) × 10−12

1 2π × 5 × 5 × 10−9

fP2 =

10

Vb1 Rin 510 = = Vsig Rin + Rsig 510 + 10

= 3.33(1 + 40 × 0.05) + 0.05

fP1 =

Rsig β +1

= 101(0.05 + 5) = 510 k

Rμ1 = Rsig (1 + gm1 Rc1 ) + Rc1

fH =

2re +

(d) This is a CC-CE cascade similar to the circuit analyzed in Example 9.10.

Rπ 1 = Rsig

Rc1 = rπ 2

RL

τπ 2 = Cπ 2 Rπ 2 = 6 × 0.15 = 0.9 ns Rμ2 = (1 + gm2 RL )(Rin2 Rout1 ) + RL

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–40 10 + 0.05 + 10 = (1 + 20 × 10) 5 101

fP1 ≤ 10 Hz thus we select CC1 so that

= 39.1 k

1 ≤ 10 2π CC1 (RG + Rsig )

τμ2 = Cμ2 Rμ2 = 2 × 39.1 = 78.2 ns τH = 19.6 + 0.9 + 0.9 + 78.2 = 99.6 ns

CC1 ≥

1 = 1.6 MHz fH = 2π × 99.6 × 10−9

1 = 18.4 nF 2π × 10 × (667 + 200) × 103

⇒ CC1 = 20 nF

(e) This is a folded cascode amplifier (i.e., one in which the cascode transistor is pnp). The analysis is identical to that for (b) above. AM = −66.7 V/V

9.93 Refer to Fig. 9.37. AM = −

fH = 2.6 MHz

RG × gm (RD RL ) RG + Rsig

(f) This is a CC-CB cascade. The analysis is identical to that for case (c) above.

where

AM = 50 V/V

= 8.246 M

fH = 5 MHz

Rsig = 100 k, gm = 5 mA/V, RD = 4.7 k and RL = 10 k.

Summary of Results

Thus,

Case

Configuration

RG = RG1 RG2 = 47 M 10 M

AM (V/V)

fH GB (MHz) (MHz)

AM = −

8.246 × 5(4.7 10) 8.246 + 0.1

= −15.8 V/V 1 2π CC1 (RG + Rsig )

a

CE

−66.7

0.039

2.6

b

Cascode

−66.7

2.6

173

c

CC-CB

50

5

250

=

d

CC-CE

−194

1.6

310

= 1.9 Hz

Folded Cascode −66.7

2.6

173

fP2 =

5

250

e f

CC-CB

50

fP1 =

= 9.92 Refer to Fig. 9.37(b). Vg = Vsig

1 sCC1

=

where RG = RG1 RG2 = 2 M 1 M = 667 k

Rsig = 200 k Vg RG = Vsig RG + Rsig

1 2π CS RS

1 = 8 Hz 2π × 10 × 10−6 × 2 × 103

fP3 = =

and

gm + 1/RS 2π CS

5 × 10−3 + 0.5 × 10−3 = 87.5 Hz 2π × 10 × 10−6

fZ =

RG RG + Rsig +

1 2π × 0.01 × 10−6 (8.246 + 0.1) × 106

1 2π CC2 (RD + RL )

1 = 10.8 Hz 2π × 1 × 10−6 (4.7 + 10) × 103

Since s s+

1 CC1 (RG + Rsig )

fP2 fP1 , fP3 , fZ , fL fP2 = 87.5 Hz

Thus, fP1 =

1 2π CC1 (RG + Rsig )

We required

Sedra, Smith, Carusone, Gaudet

9.94 See figure on next page. Replacing the MOSFET with its T model results in the circuit shown in the figure.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–41

This figure belongs to Problem 9.94.

CC2 Vo Id Is Rsig

Vsig

AM = − =−

Is

RG × gm (RD RL ) RG + Rsig

1/gm

fP2 =

3 × 10−3 = 95.5 Hz 2π × 5 × 10−6

fZ = 0 (dc) fP3 =

To minimize the total capacitance we select CS so as to place fP2 (usually the highest-frequency low-frequency pole) at 100 Hz. Thus, gm 100 = 2π CS 3 × 10−3 2π CS

⇒ CS = 4.8 µF Select CS = 5 µF. Of the two remaining poles, the one caused by CC2 has associated relatively low-valued resistances (RD and RL are much lower than RG ), thus to minimize the total capacitance we place fP3 at 10 Hz and fP1 at 1 Hz. Thus,

≡

RL

CS

2 × 3(20 10) 2 + 0.5

10 =

Io

RG

= −16 V/V

=

RD

0

CC1

1 2π CC2 (RD + RL )

1 = 5.3 Hz 2π × 1 × 10−6 (20 + 10) × 103

Since fP2 fP1 and fP3 , we have fL fP2 = 95.5 Hz 9.95 The amplifier in Fig. P9.95 will have the equivalent circuit in Fig. 9.41(b) except with RE = ∞ (i.e., omitted). Also, the equivalent circuits in Fig. 9.42 can be used to determine the three short-circuit time constants, again with RE = ∞. Since the amplifier is operating at IC IE = 100 µA = 0.1 mA and β = 100, re =

25 mV = 250 0.1 mA

gm =

0.1 mA = 4 mA/V 0.025 V

rπ =

β 100 = 25 k = gm 4

Using the equivalent circuit in Fig. 9.42(b), we get RB Rsig τCE = CE re + β +1

1 2π CC2 (20 + 10) × 103

⇒ CC2 = 0.53 µF Select CC2 = 1 µF.

To make CE responsible for 80% of fL , we use

1 1= 2π CC1 (RG + Rsig )

1 = 0.8 ωL = 0.8 × 2π fL τCE

1=

1 2π CC1 (2 + 0.5) × 106

⇒ τCE =

⇒ CC1 = 63.7 nF Select CC1 = 100 nF = 0.1 µF. With the selected capacitor values, we obtain fP1 =

1 = 0.64 Hz 2π × 0.1 × 10−6 × 2.5 × 106

Sedra, Smith, Carusone, Gaudet

1 2 ms 0.8 × 2π × 100

Thus, (200 20) × 103 CE 250 + = 2 × 10−3 101 ⇒ CE = 4.65 µF Select, CE = 5 µF.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–42

Using the information in Fig. 9.42(a), we determine τC1 as

which is lower (hence more conservative) than the required value of 100 Hz.

τC1 = CC1 [(RB rπ ) + Rsig ]

Ctotal = 5 + 0.5 + 0.5 = 6.0 µF

To make the contribution of CC1 to the determination of fL equal to 10%, we use 1 = 0.1ωL = 0.1 × 2π fL τC1 ⇒ τC1 =

9.96

Vo

1 = 15.92 ms 0.1 × 2π × 100

aIe

Thus,

RC

CC1 [(200 25) × 103 + 20 × 103 ] = 15.92 × 10−3 ⇒ CC1 = 0.38 µF Select CC1 = 0.5 µF.

Ie

Vsig

re Re

For CC2 we use the information in Fig. 9.42(c) to determine τC2 :

CE

τC2 = CC2 (RC + RL ) To make the contribution of CC2 to the determination of fL equal to 10%, we use 1 = 0.1ωL = 0.1 × 2π fL τC2 ⇒ τC2

1 = = 15.92 ms 0.1 × 2π × 100

Figure 1

Replacing the BJT with its T model results in the circuit shown in Fig. 1. (a) Ie =

Thus, CC2 (20 + 10) × 103 = 15.92 × 10−3

Vsig r e + Re +

⇒ CC2 = 0.53 µF

Vo = −αIe RC

Although, to be conservative we should select CC2 = 1 µF; in this case we can select

Thus,

CC2 = 0.5 µF because the required value is very close to 0.5 µF and because we have selected CC1 and CE larger than the required values. The resulting fL will be 1 1 1 1 + + fL = 2π τCE τC1 τC2 (200 20) × 103 τCE = 5 × 10−6 × 250 + 101

αRC Vo =− Vsig r e + Re

1 sCE

s s+

1 CE (re + Re )

From this expression we obtain AM = −

αRC r e + Re

(2)

and fL = f P =

1 2π CE (re + Re )

= 2.15 ms

(b) From Eq. (2) we see that

τC1 = 0.5 × 10−6 [(200 25) × 103 + 20 × 103 ]

| AM | =

αRC re

= 21.1 ms τC2 = 0.5 × 10−6 (20 + 10) × 103 = 15 ms 3 10 1 103 103 fL = + + 2π 2.15 21.1 15 = 92.2 Hz

Sedra, Smith, Carusone, Gaudet

(1)

(3)

1 1+

Re re

Thus, including Re reduces the gain magnitude by Re the factor 1 + . re (c) From Eq. (3), we obtain fL =

1 2π CE re

1 1+

Re re

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–43

Thus, including Re reduces fL by the factor Re 1+ . This is the same factor by which the re magnitude of the gain is reduced. Thus, Re can be used to tradeoff gain for decreasing fL (that is, increasing the amplifier bandwidth). (d) I = 0.25 mA, RC = 10 k, CE = 10 µF 25 mV VT = = 100 I 0.25 mA

re =

For Re = 0: | AM | =

αRC 10 k = 100 V/V re 100

1 = 159.2 Hz 2π × 10 × 10−6 × 100

fL =

To lower fL by a factor of 10, we use 1+

Re = 10 re

which can be manipulated to the form RL Vo = Vsig RL + Rsig s +

s 1 C(RL +Rsig )

(1)

(b) At very high frequency, i.e., as s → ∞, Vo RL 4 = 0.8 V/V = = Vsig RL + Rsig 4+1 which is the voltage-divider ratio obtained when C acts as a short circuit. (c) From Eq. (1), the 3-dB frequency can be found as the frequency of the real pole, that is, 1 ω3dB = C(RL + Rsig ) 1 = 1 × 10−6 (4 + 1) × 103 = 200 rad/s 200 f3dB = = 31.8 Hz 2π (d) From Eq. (1), at s = 0, Vo =0 Vsig

⇒ Re = 900 The gain now becomes | AM | =

100 100 = 10 V/V = Re 10 1+ re

9.98 Refer to Fig. 9.37(b). Vo = −Id

See Fig. 2 for the Bode plot.

RD × RL 1 RD + + RL sCC2

Vo RD RL =− Id RD + RL

Gain (dB) Re 0 40 dB Re 900

fP3

20 dB

s s+

1 CC2 (RD + RL )

1 = 2π CC2 (RD + RL )

where RD = 10 k and RL = 10 k f (Hz) (log scale)

15.9 Hz 159 Hz

Figure 2

To make fP3 ≤ 10 Hz, 1 ≤ 10 2π CC2 (RD + RL )

9.97 The circuit is shown in the figure below.

⇒ CC2 ≥

1 = 0.8 µF 2π × 10 × (10 + 10) × 103

Select, CC2 = 0.8 µF. Rsig

Vsig

C

RL

9.99 Refer to Fig. 9.37(b).

Vo

Is =

Vg 1 + ZS gm

gm Vg YS YS + gm 1 gm + sCS Is RS = 1 Vg gm + + sCS RS

Is = (a) Using the voltage divider rule, we find the transfer function Vo RL = 1 Vsig RL + Rsig + sC

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–44

= gm

s + 1/CS RS gm + 1/RS s+ CS

100 =

2 × 10−3 + (1/4.5 × 103 ) 2π CS

⇒ CS = 3.54 µF

Thus, fP2

(c) fZ =

gm + 1/RS = 2π CS

1 = 2π CS RS

1 = 10 Hz 2π × 3.54 × 10−6 × 4.5 × 103

1 fZ = 2π CS RS

(d) Since fP fZ ,

where

fL fP = 100 Hz

gm = 5 mA/V and RS = 1.8 k

(e) The Bode plot for the gain is shown in Fig. 2.

To make fP2 ≤ 100 Hz, gm + 1/RS ≤ 100 2π CS

Gain, dB 20 dB/decade

5 × 10−3 + (1/1.8 × 103 ) ⇒ CS ≥ = 8.8 µF 2π × 100

26 dB 20 log 20 26 dB

Select CS = 10 µF. 6 dB

Thus, f 10 Hz f 100 Hz

5 × 10−3 + (1/1.8 × 103 ) = 88.4 Hz 2π × 10 × 10−6

fP2 =

f (log scale)

Figure 2

and fZ =

1 = 8.84 Hz 2π × 10 × 10−6 × 1.8 × 103

Observe that the dc gain is 6 dB, i.e. 2 V/V. This makes perfect sense since from Fig. 1 we see that at dc, capacitor CS behaves as open circuit and the gain becomes

9.100 DC gain = −

Vo

10 k RD = − 1 1 + RS + 4.5 gm 2

= −2 V/V

Id I s RD

9.101 Refer to Fig. 9.41.

1 gm

Is Vi

In the midband, Rin = RB1 RB2 rπ

RS

CS

Figure 1

Replacing the MOSFET with its T model results in the circuit shown in Fig. 1. (a) AM ≡

Vo = −gm RD Vi

where RB1 = 33 k, RB2 = 22 k gm =

IC 0.3 mA = 12 mA/V = VT 0.025 V

re =

VT 25 mV = 83.3 = IE 0.3 mA

rπ =

β 120 = 10 k = gm 12

Thus,

−20 = −2 × RD

Rin = 33 22 10 = 5.7 k

⇒ RD = 10 k

AM = −

(b) fP =

gm + 1/RS 2π CS

Sedra, Smith, Carusone, Gaudet

Rin gm (RC RL ) Rin + Rsig

where

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–45

Rsig = 5 k, RC = 4.7 k, and RL = 5.6 k

⇒ CC1 = 3 µF

Thus,

τC2 = CC2 (RC + RL )

AM = −

5.7 × 12(4.7 5.6) 5.7 + 5

= −16.3 V/V Using the method of short-circuit time constants and the information in Fig. 9.42, we obtain τC1 = CC1 [(RB rπ ) + Rsig ]

1 = 0.1ωL = 0.1 × 2π fL τC2 = 0.1 × 2π × 50 ⇒ τC2 = 31.8 ms Thus,

= CC1 (Rin + Rsig ) = 1 × 10−6 (5.7 + 5) × 103 = 10.7 ms RB Rsig τCE = CE RE re + β +1 = 20×

(33 22 5) × 103 3.9 × 103 83.3 + 121

10−6

For CC2 to contribute 10% of fL , we use

= 2.2 ms

CC2 (4.7 + 5.6) × 103 = 31.8 × 10−3 ⇒ CC2 = 3.09 µF 3 µF Finally,

RB Rsig τCE = CE RE re + β +1 For CE to contribute 80% of fL , we use 1 = 0.8ωL = 0.8 × 2π fL τCE

τC2 = CC2 (RC + RL ) = 1 × 10−6 (4.7 + 5.6) × 103 = 10.3 ms 1 1 1 1 + + fL 2π τC1 τCE τC2 1 1 1 1 = + + 2π 10.7 × 10−3 2.2 × 10−3 10.3 × 10−3 = 102.7 Hz

= 0.8 × 2π × 50 ⇒ τCE = 3.98 ms Thus, (13.2 5) × 1000 CE 3900 83.3 + 121 = 3.98 × 10−3 ⇒ CE = 36.2 µF

9.102 Refer to the data given in the statement for Problem 9.8. RB = RB1 RB2 = 33 k 22 k = 13.2 k IC IE 0.3 mA IC 0.3 mA = 12 mA/V = gm = VT 0.025 V re =

VT 25 mV = = 83.3 IE 0.3 mA

rπ =

β 120 = 10 k = gm 12

9.103 Using the information in Fig. 9.42, we get τC1 = CC1 [(RB rπ ) + Rsig ] = CC1 [(10 1) + 5] × 103 = CC1 × 5.91 × 103 RB Rsig τCE = CE RE re + β +1 where rπ 1000 = 10 β +1 101 τCE= (10 5) × 1000 3 CE 1.5 × 10 10 + 101 re =

From Fig. 9.40, we have τC1 = CC1 [(RB rπ ) + Rsig ] For CC1 to contribute 10% of fL , we use 1 = 0.1ωL = 0.1 × 2π fL τC1

= CE × 41.8 For CC1 and CE to contribute equally to fL ,

= 0.1 × 2π × 50

τC1 = τCE

⇒ τC1 = 31.8 ms

Thus,

Thus, CC1 [(13.2 10) + 5] × 103 = 31.8 × 10−3

Sedra, Smith, Carusone, Gaudet

CC1 × 5.91 × 103 = CE × 41.8 CE = 141.4 ⇒ CC1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 9–46

This figure belongs to Problem 9.104, part (a). CC Vo aIe Rsig

RC

RL

Vsig

Vb I e

re CE

Zin

Figure 1

9.104 Replacing the BJT with its T model results in the equivalent circuit shown in Fig. 1 above. (a) At midband, CE and CC act as short circuits. Thus

(e) To minimize the total capacitance we choose to make the pole caused by CE the dominant one and make its frequency equal to fL = 100 Hz,

Rin = (β + 1)re

CE

Vo (β + 1)re =− gm (RC RL ) Vsig (β + 1)re + Rsig =−

(b) Because the controlled current source αIe is ideal, it effectively separates the input circuit from the output circuit. The result is that the poles caused by CE and CC do not interact. The pole due to CE will have frequency ωPE :

1

CE re +

Rsig β +1

⇒ CE = 12.83 µF Placing the pole due to CC at 10 Hz, we obtain

β(RC RL ) (β + 1)re + Rsig

ωPE =

1 10, 000 25 + 101

2π × 100 =

2π × 10 =

1 CC (10 + 10) × 103

⇒ CC = 0.8 µF (f) A Bode plot for the gain magnitude is shown in Fig. 2.

and the pole due to CC will have a frequency ωPC ωPC =

1 CC (RC + RL )

(c) The overall voltage transfer function can be expressed as s s Vo = AM Vsig s + ωPE s + ωPC (d) re = AM = −

Figure 2

VT 25 mV = 25 = IE 1 mA 100(10 k 10 k) 101 × 25 × 10−3 k + 10 k

= −40 V/V

Sedra, Smith, Carusone, Gaudet

The gain at fP2 (10 Hz) is 12 dB. Since the gain decreases by 40 dB/decade or equivalently 12 dB/octave, it reaches 0 dB (unity magnitude) at f = fPC /2 = 5 Hz.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–1

(c) Aβ = 104 × 10−3 = 10

Chapter 10 Solutions to Exercises within the Chapter

1 + Aβ = 11 or 20.8 dB Af =

104 A = = 909.1 V/V 1 + Aβ 11

1 = 0.1 V/V Ex: 10.1 (b) β = 1/ Af ideal = 10

(d)

R1 0.1 = R1 + R2

1000 =

⇒

⇒ β = 0.0009

R2 =9 R1

0.0009 =

(c) A = 100 β = 0.1 Aβ = 10

⇒

1 + Aβ = 11 or 20.8 dB

(e)

Af =

100 A = = 9.091 V/V 1 + Aβ 11

Af differs from the ideal value of 10 V/V by −9.1%. (d) 10 =

100 A = 1 + Aβ 1 + 100β

Vo = Af Vs = 10 × 1 = 10 V Vf = βVo = 0.09 × 10 = 0.9 V Vi = Vs − Vf = 1 − 0.9 = 0.1 V

80 1 + 80 × 0.09

= 9.76 V/V which is a 2.44% decrease. Ex. 10.2 (b) β = 1/ Af ideal = 1/103 β = 0.001 V/V 0.001 =

R1 R1 + R2

R2 ⇒ = 999 R1

Sedra, Smith, Carusone, Gaudet

Vf = βVo = 0.0009 × 10 = 0.009 V Vi = Vs − Vf = 0.01 − 0.009 = 0.001 V (f) 0.8 × 104 1 + 0.8 × 104 × 0.0009

which is a 2.44% decrease.

(e)

Af =

Vo = Af Vs = 1000 × 0.01 = 10 V

= 975.6 V/V

R2 ⇒ = 10.11 R1

A = 0.8 × 100 = 80

R2 = 1110.1 R1

Af =

R1 R1 + R2

(f)

R1 R1 + R2

A = 0.8 × 104

⇒ β = 0.09 V/V 0.09 =

104 1 + 104 β

Ex. 10.3 To constrain the corresponding change in Af to 0.1%, we need an amount-of-feedback of at least 10% = 100 1 + Aβ = 0.1% Thus the largest obtainable closed-loop gain will be 1000 A = = 10 V/V Af = 1 + Aβ 100 Each amplifier in the cascade will have a nominal gain of 10 V/V and a maximum variability of 0.1%; thus the overall voltage gain will be (10)3 = 1000 V/V and the maximum variability will be 0.3%. Ex. 10.4 R1 1 = 0.1 β= = R1 + R2 1+9 Aβ = 104 × 0.1 = 1000 1 + Aβ = 1001 Af =

A 1 + Aβ

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–2

104 = 9.99 V/V 1 + 104 × 0.1

Af =

(d) β =

fHf = fH (1 + Aβ)

Aβ = 4

= 100 × 1001 = 100.1 kHz

R1 20 = = 0.2 V/V R1 + R2 20 + 80

10 × 20 = 7.27 10 + 20 + 80

7.27 = 36.36 V/V 0.2 36.36 A = = 4.4 V/V Af = 1 + Aβ 1 + 36.36 × 0.2

A= Ex. 10.5 (a) Refer to Fig. 10.9(c). β=

R1 R1 + R2

If Aβ were 1, then

(b)

1 1 = = 5 V/V β 0.2

Af

RD

Ex. 10.6 From the solution of Example 10.5, 1 + Aβ = 60.1

Vd R2 Q

V t

Vr

Thus, fHf = (1 + Aβ)fH = 60.1 × 1

R1

= 60.1 kHz

Figure 1 Figure 1 shows the circuit prepared for determining the loop gain Aβ. Observe that we have eliminated the input signal Vs , and opened the loop at the gate of Q where the input impedance is infinite obviating the need for a termination resistance at the right-hand side of the break. Now we need to analyze the circuit to determine Vr Aβ ≡ − Vt First, we write for the gain of the CS amplifier Q, Vd = −gm [RD (R1 + R2 )] Vt

(1)

then we use the voltage-divider rule to find Vr , Vr R1 = Vd R1 + R2

(2)

Combining Eqs. (1) and (2) gives Aβ ≡ −

Vr R1 = gm [RD (R1 + R2 )] Vt R1 + R2

which can be simplified to RD R1 Aβ = gm RD + R1 + R2 Aβ (c) A = β RD (R1 + R2 ) = gm RD + R1 + R2

Sedra, Smith, Carusone, Gaudet

Ex. 10.7 Refer to Fig. E10.7. The 1-mA bias current will split equally between the emitters of Q1 and Q2 , thus IE1 = IE2 = 0.5 mA Transistor Q3 will be operating at an emitter current IE3 = 5 mA determined by the 5-mA current source. Since the dc component of Vs = 0, the negative feedback will force the dc voltage at the output to be approximately zero. The β circuit is shown in Fig. 1 (on next page) together with the determination of β and of the loading effects of the β circuit on the A circuit, β=

1 R1 = 0.1 V/V = R1 + R2 1+9

R11 = R1 R2 = 1 9 = 0.9 k R22 = R1 + R2 = 1 + 9 = 10 k The A circuit is shown in Fig. 2 (on next page). re1 = re2 = re3 = ie =

VT 25 mV = 50 = IE1,2 0.5 mA

VT 25 mV = =5 IE3 5 mA Vi

re1 + re2 +

Rs + R11 β +1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–3

These figures belong to Exercise 10.7.

R2

R2

1

2

R1

Vf 1

Vo

2

R1

b 0

R2

1

Vf R1 Vo R1 R2

2

R1

R2

1

2

R1

R22 R1 R2

R11 R1 R2 Figure 1

RC

20 k Q3

aie

Rs 10 k Q1 Vi

ib3

ie3

Q2

Vo

Rb3

ie

R11

RL 2 k

R22

Ro Ri Figure 2 ie =

Vi 0.05 + 0.05 +

⇒ ib3 = 0.105ie

10 + 0.9 101

Vo = ie3 (R22 RL )

⇒ ie = 4.81Vi

(1)

Rb3 = (β + 1)[re3 + (R22 RL )] = 101[0.005 + (10 2)] = 168.84 k ib3 = αie

RC RC + Rb3

= 0.99 ie

20 20 + 168.84

Sedra, Smith, Carusone, Gaudet

(2)

= (β + 1)ib3 (R22 RL ) = ib3 × 101(10 2) ⇒ Vo = 168.33ib3

(3)

Combining (1)–(3), we obtain A≡

Vo = 85 V/V Vi

β = 0.1 V/V

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–4

Aβ = 8.5

R22 = R1 + R2

1 + Aβ = 9.5

Figure 2 shows the A circuit. We can write

85 = 8.95 V/V 9.5 From the A circuit, we have

Vo = gm (RD R22 )Vi

Ri = Rs + R11 + (β + 1)(re1 + re2 )

A≡

Af =

Thus, Vo = gm [RD (R1 + R2 )] Vi

= 10 + 0.9 + 101 × 0.1 = 21 k Rif = Ri (1 + Aβ)

RD

= 21 × 9.5 = 199.5 k Rin = Rif − Rs = 199.5 − 10 = 189.5 k

Vo

From the A circuit, we have RC Ro = RL R22 re5 + β +1 20 = 2 10 0.005 + 101

R22 Q R11

= 181

Ro

Vi

Ro Rof = 1 + Aβ

Ri

181 = 19.1 = 9.5 Rof = RL Rout

Figure 2 R1 R1 + R2 A Af = 1 + Aβ

β=

19.1 = 2 k Rout ⇒ Rout = 19.2

From A circuit, we have Ex. 10.8 Figure 1 below shows the β circuit together with the determination of β, R11 and R22 . β=

Ri =

1 gm

R1 R1 + R2 R2

R2

1

2

R1

Vf 1

V o

2

R1

b 0

R2

1

R1

Vf R1 Vo R1 R2

2

1

R2

R1

R22 R1 R2

R11 R1 R2

R11 = R1 R2

Sedra, Smith, Carusone, Gaudet

2

Figure 1

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–5

value. Since β = RF , the feedback resistance RF becomes

Ro = RD R22 Rin = Rif = Ri (1 + Aβ)

Rout

RF = 500

1 (1 + Aβ) gm

Rin =

Now, since A is given by

Ro = Rof = 1 + Aβ

Rout =

A=

RD (R1 + R2 ) 1 + Aβ

Rid Rid + RF

= 982.7 mA/V

Ex. 10.9 Af ideal = 10 A/A

Af =

1 Af

ro2 μ + (RF ro2 ) RF + ro2

the value of A becomes 1000 20 100 A= 100 + 0.5 0.5 + (0.520) 0.5 + 20

Comparison with the results of Exercise 10.5 shows that the expressions for A and β are identical. However, Rin and Rout cannot be determined using the method of Exercise 10.5.

β=

1 gm

Thus, Af becomes A 982.7 = 1 + Aβ 1 + 982.7 × 0.5

= 1.996 mA/V

= 0.1 A/A

ideal

A = 1000 A/A 1 + Aβ = 1 + 1000 × 0.1 = 101 Af =

1000 A = = 9.9 A/A 1 + Aβ 101

Rif =

1000 Ri = = 9.9 1 + Aβ 101

Rof = Ro (1 + Aβ) = 100 k × 101

Ex. 10.12 To obtain Af ideal = 100 mA/V, the feedback factor β must be β=

1 = 0.01 V/mA = 10 100

But, β=

RE1 RE2 RE1 + RE2 + RF

For RE1 = RE2 = 100 ,

= 10.1 M Ex. 10.10 Refer to the solution of Example 10.7. If μ is reduced from 1000 V/V to 100 V/V, A will be reduced by the same factor (10) to become

10 =

100 × 100 100 + 100 + RF

⇒ RF = 800

A = 64.9 mA/V

Since the ideal value of (Io /Vs ) is 100 mA/V, the ideal value of the voltage gain is,

and the loop gain will be correspondingly reduced,

Io R C Vo =− = −100 × 0.6 Vs Vs

Aβ = 64.9

= −60 V/V

Thus, the amount of feedback becomes 1 + Aβ = 65.9 and the closed-loop gain Af becomes Af =

64.9 A = = 0.985 mA/V 1 + Aβ 65.9

The input resistance becomes

Ex. 10.13 See figure on next page. Figure 1 shows the circuit prepared for the determination of the loop gain, Aβ ≡ −

Vr Vt

Rin = Rif = (1 + Aβ)Ri

We will trace the signal around the loop as follows:

= 65.9 × 101 = 6.7 M

Ic2 = gm2 Vt

(1)

The output resistance becomes

RC2 Ib3 = Ic2 RC2 + Ri3

(2)

Rout = Rof = (1 + Aβ)Ro

where

= 65.9 × 61 = 4 M Ex. 10.11 Refer to the solution of Example 10.7. To double Af ideal we need to reduce β to half its

Sedra, Smith, Carusone, Gaudet

Ri3 = (β + 1) re3 + [RE2 (RF + (RE1 re1 ))]

(3)

Ie3 = (β + 1)Ib3

(4)

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–6

This figure belongs to Exercise 10.13.

RC3

RC2

RC1

Ic1 rp2

Q1

Vr

Vt

Ib3

Ri3

Ie3

Ie1

re1

Q3

Ic2 Q2

RF If

RE2

RE1

Figure 1 If = Ie3

RE2 RE2 + RF + (RE1 re1 )

(5)

RE1 Ie1 = If RE1 + re1

(6)

Ic1 = αIe1

(7)

Vr = −Ic1 (RC1 rπ 2 )

(8)

gm2 = 40 mA/V, RC2 = 5 k, β = 100, re3 = 6.25 , RE1 = RE2 = 100 , RF = 640 , re1 = 41.7 , α1 = 0.99, RC1 = 9 k, and rπ 2 = 2.5 k Ri3 = 101 0.00625 + [0.1 (0.64 + (0.1 0.0417))]

(9)

Ie3 = 101Ib3

(11)

If = 0.13Ie3

(12)

Ie1 = 0.706If

(13)

Ic1 = 0.99Ie1

(14)

Vr = −1.957Ic1

(15)

Sedra, Smith, Carusone, Gaudet

Rid

Rid mV t

RL

Figure 1 Figure 1 shows the circuit prepared for determining the loop gain

Vr =

(10)

Vr = 249.2 Vt

Rs

V Vr t

Vr Vt

Using the voltage-divider rule, we can write by inspection

Ib3 = 0.347Ic2

Aβ = −

ro

Aβ ≡ −

= 9.42 k

Combining (9)–(15), we obtain

RF

Combining (1)–(8) gives Vr in terms of Vt and hence Aβ ≡ −Vr /Vt . We shall do this numerically using the values in Example 10.8:

Ic2 = 40Vt

Ex. 10.14

−μVt Vr = −μVt

RL [RF + (Rs Rid )] (Rs Rid ) ro + RL [RF + (Rs Rid )] RF + (Rs Rid ) RL (Rs Rid ) ro [RL + RF + (Rs Rid )] + RL [RF + (Rs Rid )]

Thus, Aβ = −

Vr = Vt

μRL (Rid Rs ) ro [RL + RF + (Rid Rs )] + RL [RF + (Rid Rs )]

Q.E.D. Using the numerical values in Example 10.9, we get Aβ = 1035.2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–7

This figure belongs to Exercise 10.15. RF

Vo

Is

Vgs

Rs

gmVgs

ro

Rif

Rout Rof

Rin

(a) If

RF

1

RF 1

2

Vo

2

b

(b)

1 If RF Vo

(c) RF

RF

1

1

2

R11 RF

2

(d)

R22 RF

(e)

Vo

Ii

Rs

R11 Vgs

gmVgs

R22

ro

Ro

Ri (f)

Figure 1

Ex. 10.15 Figure 1(a) above shows the feedback amplifier circuit. The β circuit is shown in Fig. 1(b), and the determination of β is shown in Fig. 1(c), β=− (a) Af

Finally, the A circuit is shown in Fig. 1(f). We can write by inspection Ri = Rs R11 = Rs RF Ro = ro R22 = ro RF

1 RF

Vgs = Ii Ri Vo = −gm Vgs (ro R22 )

ideal

1 = = −RF β

Thus,

(b) The determination of R11 and R22 is illustrated in Figs. 1(d) and (e), respectively:

A≡

R11 = R22 = RF

Af =

Sedra, Smith, Carusone, Gaudet

Vo = −gm (Rs RF )(ro RF ) Ii Vo A = Is 1 + Aβ

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–8

Af =

(c) Rif = Rif =

100 20 1 20 + (1090) + (109020) 5 = −1076.4 A/A = −100

gm (Rs RF )(ro RF ) − 1 + gm (Rs RF )(ro RF )/RF

Q.E.D.

Ri 1 + Aβ

β = −0.1

Rs RF 1 + gm (Rs RF )(ro RF )/RF

Af =

1076.4 A =− 1 + Aβ 1 + 1076.4 × 0.1

1 1 1 gm (ro RF ) = + + Rif Rs RF RF

=−

But,

= −9.91 A/A

1 1 1 = + Rif Rs Rin

Ri = 100 k Rin = Rif =

thus, 1 1 = [1 + gm (ro RF )] Rin RF ⇒ Rin = (d) Rout =

Ri 1 + Aβ

100 108.64 = 921 =

RF 1 + gm (ro RF )

Q.E.D.

Ro = Rof = 1 + Aβ

Rout = Rof = Ro (1 + Aβ) = 929 × 108.64 = 101 M

r o RF 1 + gm (Rs RF )(ro RF )/RF

1 1 1 gm (Rs RF ) = + + Rout ro RF RF ⇒ Rout = ro

1076.4 1076.4 =− 1 + 107.64 108.64

RF 1 + gm (Rs RF )

Q.E.D.

Ex. 10.17 Refer to the solution of Example 1 = −1 10.10. With R2 = 0, β = − 1+0 1 Af ideal = = −1 A/A β

(e) A = −5(1 10)(20 10)

Ri = Rs Rid (R1 + R2 )

A = −30.3 k

⇒ Ri = R1

β=−

1 1 =− = −0.1 mA/V RF 10

Aβ = 3.03

A 30.3 =− = −7.52 k 1 + Aβ 4.03

(Compare to the ideal value of −10 k). Ri = Rs RF = 1 10 = 909 Ro = ro RF = 20 10 = 6.67 k 909 Ri = = 226 Rif = 1 + Aβ 4.03 1 1 Rin = 1 = 291 − Rif Rs Rof =

⇒ Ro = ro2 A = −μ

1 + Aβ = 4.03 Af =

Ro = ro2 + (R1 R2 ) + (gm ro2 )(R1 R2 )

6.67 Ro = = 1.66 k 1 + Aβ 4.03

Rout = Rof = 1.66 k

ro2 Ri 1/gm + (R1 R2 ro2 ) ro2 + (R1 R2 )

⇒ A = −μgm R1 Aβ = μgm R1 Af = =−

A 1 + Aβ

μgm R1 1 + μgm R1

Rin = Rif = Rin =

Ri 1 + Aβ

R1 1 1 + μgm R1 μgm

Rout = Rof = Ro (1 + Aβ) = ro2 (1 + μgm R1 ) ⇒ Rout μ(gm ro2 )R1

Ex. 10.16 Refer to the solution of Example 10.10. ro2 Ri A = −μ 1/gm + (R1 R2 ro2 ) ro2 + (R1 R2 )

Sedra, Smith, Carusone, Gaudet

Ex. 10.18 The feedback shifts the pole by a factor equal to the amount of feedback:

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–9

1 + A0 β = 1 + 105 × 0.01 = 1001 The pole will be shifted to a frequency

complex-frequency variable S = s/104 , and the characteristic equation becomes (S + 1)3 + 103 β = 0

fPf = fP (1 + A0 β) = 100 × 1001 = 100.1 kHz If β is changed to a value that results in a nominal closed-loop gain of 1, then we obtain β=1

(1)

This equation has three roots, a real one and a pair that can be complex conjugate. The real pole can be found from (S + 1)3 = −103 β ⇒ S = −1 − 10β 1/3 = − 1 + 10β 1/3

and 1 + A0 β = 1 + 105 × 1 105 then the pole will be shifted to a frequency fPf = 105 × 100 = 10 MHz

(2)

Dividing the characteristic polynomial in (1) by S + 1 + 10β 1/3 gives a quadratic whose two roots are the remaining poles of the feedback amplifier. After some straightforward but somewhat tedious algebra, we obtain S2 + 10β 1/3 − 2 S + 1 + 100β 2/3 − 10β 1/3

Ex. 10.19 From Eq. (10.40), we see that the poles coincide when

=0

(3)

(ωP1 + ωP2 )2 = 4(1 + A0 β)ωP1 ωP2

The pair of poles can now be obtained as √ S = −1 + 5β 1/3 ± j5 3 β 1/3

(4)

(104 + 106 )2 = 4(1 + 100β) × 104 × 106 ⇒ 1 + 100β = 25.5

Equations (1) and (3) describe the three poles shown in Fig. E10.20.

⇒ β = 0.245 The corresponding value of Q = 0.5. This can also be verified by substituting in Eq. (10.42). A maximally flat response is obtained when √ Q = 1/ 2. Substituting in Eq. (10.42), we obtain

1 (1 + 100β) × 104 × 106 √ = 104 + 106 2 ⇒ β = 0.5

From Eq. (2) we see that the pair of complex poles lie on the jω axis for the value of β that makes the coefficient of S equal to zero, thus βcr =

2 10

3 = 0.008

Ex. 10.21

99R

In this case, the low-frequency closed-loop gain is Af (0) =

R

A0 1 + A0 β

100 = = 1.96 V/V 1 + 100 × 0.5

A (s)

Figure 1 Ex. 10.20 The closed-loop poles are the roots of the characteristic equation 1 + A(s)β = 0 ⎞3 ⎛ ⎜ 1+⎝

Aβ = A(s) × 0.01 =

⎟ s ⎠β = 0 1+ 4 10 10

To simplify matters, we normalize s by the factor 104 , thus obtaining the normalized

Sedra, Smith, Carusone, Gaudet

From Fig. 1, we can easily obtain the loop gain as

=

105

1+

× 0.01 s 2π × 10

1000 s 1+ 2π × 10

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Exercise 10–10

From this single-pole response (low-pass STC response) we can find the unity-gain frequency by inspection as f1 = fP × 1000 = 104 Hz

20 log(1/β) = 80 dB ◦

The phase angle at f1 will be −90 and thus the phase margin is 90◦ . Ex. 10.22 From Eq. (10.48), we obtain | Af (jω1 )| = 1/|1 + e−jθ | 1/β

⇒ β = 10−4 Af =

A 105 = 1 + Aβ 1 + 105 × 10−4

= 9.09 × 103 = 79.2 dB 80 dB

= 1/|1 + cos θ − j sin θ | (a) For PM = 30◦ , θ = 180 − 30 = 150◦ , thus | Af (jω1 )| = 1/|1 + cos 150◦ − j sin 150◦ | 1/β = 1.93 (b) For PM = 60◦ , θ = 180 − 60 = 120◦ , thus | Af (jω1 )| = 1/|1 + cos 120◦ − j sin 120◦ | 1/β =1 (c) For PM = 90◦ , θ = 180 − 90 = 90◦ , thus | Af (jω1 )| = 1/|1 + cos 90◦ − j sin 90◦ | 1/β √ = 1/ 2 = 0.707

Sedra, Smith, Carusone, Gaudet

Ex. 10.23 For a phase margin of 45◦ , φ = −135◦ . Locating the point φ = −135◦ on the phase plot in Fig. 10.38, we find the corresponding point on the gain plot at 80 dB, that is,

Ex. 10.24 To obtain stable performance for closed-loop gains as low as 20 dB (which is 80 dB below A0 , or equivalently 104 below A0 ), we must place the new dominant pole at 1 MHz/104 = 100 Hz. Ex. 10.25 The frequency of the first pole must be lowered from 1 MHz to a new frequency fD =

10 MHz = 1000 Hz 104

that is, by a factor of 1000. Thus, the capacitance at the controlling node must be increased by a factor of 1000.

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–1

Solutions to End-of-Chapter Problems 10.1 Af = 200 =

= 0.001 V

A 1 + Aβ

If A becomes 1000(1 − 0.1) = 900 V/V, then we get

4

10 1 + 104 β

Af =

−3

⇒ β = 4.9 × 10

If A changes to 103 , then we get

1000 = 169.5 5.9

Percentage change in Af =

900 = 0.99889 1 + 900 × 1

Thus, Af changes by =

1000 Af = 1 + 103 × 4.9 × 10−3 =

Vi = Vs − Vo = 1 − 0.999

0.99889 − 0.999 × 100 = −0.011% 0.999

10.4 (a) Figure 1 shows the situation with A set to ∞. We see that, 169.5 − 200 × 100 200

0 0

= −15.3% 10.2 Af ideal = 100 β=

1 Af

Vs

Vo

Vs Vs /R1

R1 Vs /R1

ideal

Af =

1 Aβ A = 1 + Aβ β 1 + Aβ

Af = Af ideal

Figure 1

Af ideal Aβ = 1 + Aβ 1 + (1/Aβ)

Vo =

Vs (R1 + R2 ) R1

Thus, Vo R2 =1+ Af ideal = Vs R1

For Aβ 1,

1 Af Af ideal 1 − Aβ

and,

Thus, the deviation is −

β = 1/ Af ideal =

100 % Aβ

To limit the deviation to 1%, Aβ|minimum = 100 and Aminimum = 104

(b) Af =

A 1 + Aβ

10.3 The direct connection of the output terminal to the inverting input terminal results in Vf = Vo and thus

10 =

A 1 + Aβ

β=1 If A = 1000, then the closed-loop gain will be Af = =

R2

= 0.01

⇒ β = 0.1 −

1000 = 0.999 V/V 1 + 1000 × 1

1 A

(1)

But, β=

A 1 + Aβ

R1 R1 + R2

R1 10 = R1 + R2 10 + R2

⇒ R2 = 10(

1 − 1) β

(2)

= 1 + 1000 × 1 = 1001

Eqns. (1) and (2) can be used to determine the values of β and R2 required in each of the following cases:

or 60 dB

(i)

For Vs = 1 V, we obtain

A = 1000 V/V,

Vo = Af Vs = 0.999 × 1 = 0.999 V

β = 0.1 − 0.001 = 0.099 V/V

Amount of feedback = 1 + Aβ

Sedra, Smith, Carusone, Gaudet

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–2

R2 = 10(

1 − 1) = 91 k 0.099

β=

Vf 0.99 = = 0.198 V/V Vo 5

(ii) 10.7 For each value of A given, we have three different values of β: 0.00, 0.50, and 1.00. To obtain Af , we use

A = 100 V/V β = 0.1 − 0.01 = 0.09 V/V R2 = 101.1 k

Af =

(iii)

A 1 + Aβ

The results obtained are as follows.

A = 20 V/V β = 0.1 − 0.05 = 0.05

Case A (V/V) Af (V/V) Af (V/V) Af (V/V) for for for β = 0.00 β = 0.50 β = 1.00

R2 = 190 k (c) (i)

(a)

1

1

0.667

0.500

(b)

10

10

1.667

0.909

(c)

100

100

1.961

0.990

Thus Af changes by −0.11%.

(d)

1000

1000

1.996

0.999

(ii)

(e)

10,000

10,000

1.9996

0.9999

A = 0.9 × 1000 = 900 V/V Af =

900 = 9.99 V/V 1 + 900 × 0.099

A = 0.9 × 100 = 90 V/V Af =

90 = 9.89 V/V 1 + 90 × 0.09

Thus, Af changes by −1.1%. (iii) A = 0.9 × 20 = 18 V/V Af =

18 = 9.47 V/V 1 + 18 × 0.05

Thus Af changes by −5.3%. We conclude that as A becomes smaller and hence the amount of feedback (1 + Aβ) is lower, the desensitivity of the feedback amplifier to changes in A decreases. In other words, the negative feedback becomes less effective as (1 + Aβ) decreases.

5V = 2500 V/V 2 mV 5V = 50 V/V Af = 100 mV Amount of feedback ≡ 1 + Aβ 10.8 A =

=

A 2500 = = 50 Af 50

or 34 dB Aβ = 49 β=

49 = 0.0196 V/V 2500

10.9 Anominal = 1000 Alow = 500

10.5 Af ideal = 10

Ahigh = 1500

1 β= Af

If we apply negative feedback with a feedback factor β, then

= 0.1

ideal

20 log(1 + Aβ) = 40 dB

Af , nominal =

⇒ 1 + Aβ = 100 ⇒ A = 990 990 A = Af = 1 + Aβ 1 + 990 × 0.1 = 9.9

10.6 A =

1000 1 + 1000β

Af , low =

500 1 + 500β

Af , high =

1500 1 + 1500β

It is required that Vo 5V = 500 V/V = Vi 10 mV

Vf = Vs − Vi = 1 − 0.01 = 0.99 V

Sedra, Smith, Carusone, Gaudet

Af , low ≥ 0.99Af , nominal

(1)

and Af , high ≤ 1.01Af , nominal

Microelectronic Circuits, 8th International Edition

(2)

© Oxford University Press 2021

Chapter 10–3

If we satisfy condition (1) with equality, we can determine the required value of β. We must then check that condition (2) is satisfied. Thus, 1000 500 = 0.99 × 1 + 500β 1 + 1000β ⇒ β = 0.098

1000 1 + 1000 × 0.098

99 = 0.099 V/V 1000 For A = 500 V/V, we have β=

500 1 + Aβ

⇒ Densensitivity factor ≡ 1 + Aβ = 50 49 = 0.098 V/V 500 If the A = 1000 amplifiers have a gain uncertainty of ±10%, the gain uncertainty of the closed-loop amplifiers will be

β=

= 10.101 Af , low =

500 = 10 1 + 500 × 0.098

Af , high =

1500 = 10.135 1 + 1500 × 0.098

Thus, the low value of the closed-loop gain is 0.101 below nominal or −1%, and the high value is 0.034 above nominal or 0.34%. Thus, our amplifier meets specification and the nominal value of closed-loop gain is 10.1. This is the highest possible closed-loop gain that can be obtained while meeting specification. Now, if three closed-loop amplifiers are placed in cascade, the overall gain obtained will be Nominal Gain = (10.1)3 = 1030 Lowest Gain = 10 = 1000 3

Highest Gain = (10.135)3 = 1041 Thus, the lowest gain will be approximately 3% below nominal, and the highest gain will be 1% above nominal.

10.10 From Eq. (11.10), we have 1 dAf /Af = dA/A 1 + Aβ Since −40 dB is 0.01, we have 0.01 =

Aβ = 99

Af = 10 =

For this value of β, we obtain Af , nominal =

⇒ Densensitivity factor ≡ 1 + Aβ = 100

1 1 + Aβ

±10% = ±0.1% 100 If we require a gain uncertainty of ±0.1% using the A = 500 amplifiers, then =

±0.1% =

50 ⇒ Gain uncertainty = ±5%

10.12 Af = 10 V/V 1 + Aβ =

A 100 ⇒ A = 1000 V/V 100 − 1 = 0.099 V/V β= 1000

10.13 The open-loop gain varies from A to 10A with temperature and time. Correspondingly, Af varies from (25 − 1%) i.e. 24.75 V/V to (25 + 1%) or 25.25 V/V. Substituting these quantities into the formula for the closed-loop gain A 1 + Aβ

Af =

For 1 dAf /Af = dA/A 5

we obtain

1 + Aβ = 5

24.75 =

A 1 + Aβ

(1)

25.25 =

10A 1 + 10Aβ

(2)

Dividing Eq. (2) by Eq. (1), we obtain

⇒ Aβ = 4

1.02 = 10 10.11 For A = 1000 V/V, we have Af = 10 =

±10% = 100 ±0.1%

10 =

⇒ Aβ = 99

we have

Gain uncertainty of A = 500 amplifiers

1000 1 + Aβ

Sedra, Smith, Carusone, Gaudet

1 + Aβ 1 + 10Aβ

1.02 + 10.2Aβ = 10 + 10Aβ ⇒ Aβ = 44.9

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–4

Substituting in (1) yields

A0 = 1000 V/V

A = 24.75 × 45.9 = 1136

Af 0 =

and β=

A0 1 + A0 β

⇒ 1 + A0 β =

44.9 = 0.0395 1136

A0 1000 = = 100 Af 0 10

99 = 0.099 V/V 1000 10 MHz = 0.1 MHz = 100 kHz f3dB = 100 Thus, the open-loop amplifier 3-dB frequency must be at least 100 kHz.

β= 10.14 A0 = 105 V/V f3dB = 100 Hz (a) Af 0 = 103 =

105 1 + A0 β

10.16 The closed-loop gain for the first (high-gain) segment is

⇒ 1 + A0 β = 100 or 40 dB

1000 1 + 1000β

3 β = 9.9 × 10−4 0.001 V/V

Af 1 =

fH = (1 + A0 β)f3dB = 100 × 100

and that for the second segment is

= 10 kHz

Af 2 =

(b) Af 0 = 102 =

105 1 + A0 β

1 + A0 β = 1000

(1)

100 1 + 100β

(2)

We require Af 1 = 1.1 Af 2

β = 9.99 × 10−3 0.01 V/V

Thus, dividing Eq. (1) by Eq. (2) yields

fH = 1000 × 100

1.1 = 10

= 100 kHz

1 + 100β 1 + 1000β

1.1 + 1100β = 10 + 1000β

(c)

⇒ β = 0.089

105 Af 0 = 10 = 1 + A0 β 1 + Aβ = 104 −2

β = 9.999 × 10

0.1 V/V

Af 1 =

1000 = 11.1 V/V 1 + 1000 × 0.089

Af 2 =

100 = 10.1 V/V 1 + 100 × 0.089

= 1 MHz

The first segment ends at |v O | = 10 mV × 1000 = 10 V. This corresponds to

(d)

vS =

fH = 104 × 100

Af 0 = 1 =

105 1 + A0 β

1 + Aβ = 105 β = 9.9999 × 10−1 1 V/V fH = 10 × 100 5

10 V 10 = 0.9 V = Af 1 11.1

The second segment ends at |v O | = 10 + 0.05 × 100 = 15 V. This corresponds to v S = 0.9 +

15 − 10 Af 2

10.15

5 = 1.4 V 10.1 Thus, the transfer characteristic of the feedback amplifier can be described as follows:

Af 0 = 10 V/V

For |v S | ≤ 0.9 V, v O /v S = 11.1 V/V

fH ≥ 10 MHz

For 0.9 V ≤ |v S | ≤ 1.4 V, v O /v S = 10.1 V/V

= 10 MHz

Sedra, Smith, Carusone, Gaudet

= 0.9 +

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–5

This figure belongs to Problem 10.16.

12μ 1 + 12μβ

For |v S | ≥ 1.4 V, v O = ±15 V

100.5 =

The transfer characteristic is shown in the figure above.

Dividing (4) by (3) yields

10.17 Let the gain of the ideal (nonvarying) driver amplifier be denoted μ. Then, the open-loop gain A will vary from 2μ to 12μ. Correspondingly, the closed-loop gain will vary from 95 V/V to 105 V/V. Substituting these quantities into the closed-loop gain expression, we obtain 2μ (1) 95 = 1 + 2μβ 105 =

12μ 1 + 12μβ

(2)

6(1 + 2μβ) 1 + 12μβ

⇒ μβ = 41.58 Substituting into (3) provides 99.5(1 + 2 × 41.58) 2 = 4187 V/V

μ=

which is about a factor of 10 higher than the gain required in the less constrained case. The value of β required is 41.58 = 9.93 × 10−3 V/V 4187 Repeating for Af = 10 V/V (a factor of 10 lower than the original case): β=

Dividing Eq. (2) by Eq. (1) yields 1.105 =

1.01 =

(4)

6(1 + 2μβ) 1 + 12μβ

1.105 + 1.105 × 12μβ = 6 + 12μβ ⇒ μβ = 3.885

(a) For ±5% maximum variability, Eqs. (1) and (2) become

Substituting in Eq. (1) yields

9.5 =

95(1 + 2 × 3.885) = 416.6 V/V 2 3.885 β= = 9.33 × 10−3 V/V 416.6 If Af is to be held to within ±0.5%, Eqs. (1) and (2) are modified to

μ=

99.5 =

2μ 1 + 2μβ

Sedra, Smith, Carusone, Gaudet

(3)

2μ 1 + 2μβ

10.5 =

(5)

12μ 1 + 12μβ

(6)

Dividing (6) by (5) yields 1.105 =

6(1 + 2μβ) 1 + 12μβ

⇒ μβ = 3.885

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–6

which is identical to the first case considered, and 9.5(1 + 2 × 3.885) μ= = 41.66 V/V 2 which is a factor of 10 lower than the value required when the gain required was 100. The feedback factor β is 3.885 β= = 9.33 × 10−2 V/V 41.66 which is a factor of 10 higher than the case with Af = 100. (b) Finally, for the case Af = 10 ± 0.5% we can write by analogy μβ = 41.58 μ = 418.7 V/V β = 9.93 × 10−2 V/V

and a high value of Af , high =

Thus, the cascade of two stages will have a range of Lowest gain = 9.9572 = 99.14 V/V Highest gain = 10.0232 = 100.46 V/V which is −0.86% to +0.46% of the nominal 100 V/V gain, well within the required ±1%. 10.19 If the nominal open-loop gain is A, then we require that as A drops to (A/2) the closed-loop gain drops from 10 to a minimum of 9.8. Substituting these values in the expression for the closed-loop gain, we obtain 10 =

10.18 If we use one stage, the amount of feedback required is 1 + Aβ =

1000 A = 10 = Af 100

Thus the closed-loop amplifier will have a variability of ±30% = ±3% 10 which does not meet specifications. Next, we try using two stages. For a nominal gain of 100, each stage will be required to have a nominal gain of 10. Thus, for each stage the amount of feedback required will be

Variability of Af =

1000 = 100 1 + Aβ = 10 Thus, the closed-loop gain of each stage will have a variability of ±30% = ±0.3% 100 and the cascade of two stages will thus show a variability of ±0.6%, well within the required ±1%. Thus two stages will suffice. =

We next investigate the design in more detail. Each stage will have a nominal gain of 10 and thus 1000 = 100 1 + Aβ = 10 ⇒ Aβ = 99 ⇒ β = 0.099 Since A ranges from 700 V/V to 1300 V/V, the gain of each stage will range from Af , low =

700 = 9.957 V/V 1 + 700 × 0.099

Sedra, Smith, Carusone, Gaudet

1300 = 10.023 V/V 1 + 1300 × 0.099

A 1 + Aβ

(1)

A/2 1 1 + Aβ 2 Dividing Eq. (1) by Eq. (2) yields 1 2 1 + Aβ 2 1.02 = 1 + Aβ

9.8 =

1.02 = =1+

(2)

2 + Aβ 1 + Aβ 1 1 + Aβ

1 = 50 0.02 Substituting in Eq. (1) gives

⇒ 1 + Aβ =

A = 10 × 50 = 500 V/V and 50 − 1 = 0.098 V/V 500 If β is accurate to within ±1%, to ensure that the minimum closed-loop gain realized is 9.8 V/V, we have A/2 9.8 = 1 1 + A × 0.098 × 1.01 2 ⇒ A = 653.4 V/V β=

10.20 Af =

A 1 + Aβ

(1)

0.1A 1 + 0.1Aβ

(2)

100 = 99 =

A 1 + Aβ

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–7

Dividing Eq. (1) by Eq. (2) gives 10(1 + 0.1Aβ) 1.01 = 1 + Aβ =

10 + Aβ 1 + Aβ

=1+ ⇒

10.22 Let’s first try N = 2. The closed-loop gain of each stage must be √ Af = 1000 = 31.6 V/V

9 1 + Aβ

9 = 0.01 1 + Aβ

Thus, the amount-of-feedback in each stage must be 1000 A = 31.6 = 1 + Aβ = Af 31.6

1 + Aβ = 900 Aβ = 899 Substituting (1 + Aβ) = 900 into Eq. (1) yields A = 100 × 900 = 90, 000 V/V The value of β is β=

Thus, both the midband gain and the 3-dB frequency are lowered by the amount of feedback, (1 + AM β).

899 = 9.989 × 10−3 V/V 90, 000

If A were increased tenfold, i.e, A = 900, 000, we obtain 900, 000 = 100.1 V/V Af = 1 + 8990 If A becomes infinite, we get

The 3-dB frequency of each stage is f3dB |stage = (1 + Aβ)fH = 31.6 × 20 = 632 kHz Thus, the 3-dB frequency of the cascade amplifier is f3dB |cascade = 632 21/2 − 1 = 406.8 kHz which is less than the required 1 MHz. Next, we try N = 3. The closed-loop gain of each stage is Af = (1000)1/3 = 10 V/V

A Af = 1 + Aβ 1 1 = 1 β +β A 1 = 100.11 V/V = 9.989 × 10−3 =

and thus each stage will have an amount-of-feedback 1000 = 100 1 + Aβ = 10 which results in a stage 3-dB frequency of f3dB |stage = (1 + Aβ)fH = 100 × 20 = 2000 kHz = 2 MHz

s 10.21 A = AM s + ωL Af = =

The 3-dB frequency of the cascade amplifier will be f3dB |cascade = 2 21/3 − 1

A 1 + Aβ

= 1.02 MHz

AM s/(s + ωL ) 1 + AM βs/(s + ωL )

which exceeds the required value of 1 MHz. Thus, we need three identical stages, each with a closed-loop gain of 10 V/V, an amount-of-feedback of 100, and a loop gain

AM s = s + ωL + sAM β =

AM s s(1 + AM β) + ωL

=

s AM 1 + AM β s + ωL /(1 + AM β)

Aβ = 99 Thus, β = 0.099 V/V

Thus, AM 1 + AM β ωL ωLf = 1 + AM β

AMf =

Sedra, Smith, Carusone, Gaudet

10.23 Refer to Fig. 1 (next page). For v I = +0.7 V, we have v O = 0 and

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–8

V

(b) From Example 10.4, we obtain RL [R2 + R1 (Rid + Rs )] Aβ = μ RL [R2 + R1 (Rid + Rs )] + ro

vS

ve

m 100

×

vO

vI

10 [90 + 10 (100 + 100)] Aβ = 1000 10 [90 + 10 (100 + 100)] + 1 ×

V

= 43.11 43.11 Aβ = = 431.1 V/V A= β 0.1

vO

Af =

Slope 0.99 V/V vS

7 mV

9.9 = =

Slope 0.99 V/V

A 1 + A × 0.1

Thus μ must be increased by the factor 1010 = 2.343 to become 431.1 μ = 2343 V/V

vI +0.7 = = +7 mV μ 100

Similarly, for v I = −0.7 V, we obtain v O = 0 and −0.7 vI = = −7 mV ve = μ 100 Thus, the limits of the deadband are now ±7 mV. Outside the deadband, the gain of the feedback amplifier, that is, v O /v S , can be determined by noting that the open-loop gain A ≡ v O /v e = 100 V/V and the feedback factor β = 1, thus

=

A 1 + Aβ

⇒ A = 1010 V/V

Figure 2

Af ≡

431.1 A = = 9.77 V/V 1 + Aβ 1 + 43.11

(c) To obtain Af = 9.9 V/V, we use

7 mV

ve =

10 (100 + 100) 100 × [10 (100 + 100)] + 90 100 + 100

= 1000 × 0.9009 × 0.0957 × 0.5

Figure 1

0

R1 (Rid + Rs ) Rid × [R1 (Rid + Rs )] + R2 Rid + Rs

vO A = vS 1 + Aβ

100 = 0.99 V/V 1 + 100 × 1

The transfer characteristic is depicted in Fig. 2.

10.25 (a) Setting the open–loop gain to ∞, the voltage Vgs1 reduces to zero and the current in the source of Q1 (Id1 ) becomes zero. This isolates the feedback network which is the voltage divider (RF , RS1 ). Thus, RS1 β= RS1 + RF and the ideal value of the closed-loop gain is RF 1 Af ideal = = 1 + β RS1 RF 10 = 1 + 0.1 ⇒ RF = 0.9 k (b)

10.24 Refer to Fig. 10.13. (a) The ideal closed-loop gain is given by Af =

1 R2 R1 + R2 =1+ = β R1 R1

R2 10 ⇒ R2 = 90 k 10 = 1 +

Sedra, Smith, Carusone, Gaudet

Figure 1 on the next page shows the circuit for determining the loop gain. Observe that we have broken the loop at the gate of Q2 where the input resistance is infinite, obviating the need for adding a termination resistance. Also, observe that as usual we have set Vs = 0. To determine the loop gain Aβ ≡ −

Vr Vt

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–9

A 1 + Aβ

Af =

RD2

RD1 Vd2 Id1

Vr Vt

Id3 Q3

Q2

Q1

Id3

Id1

RF

313.3 = 9.7 V/V 1 + 31.33

Thus, Af is 0.3 V/V lower than the ideal value of 10 V/V, a difference of −3%. The circuit could be adjusted to make Af exactly 10 by changing β through varying RF . Specifically, 313.3 1 + 313.3β

10 =

1/gm1 If

RS1

=

RS2

⇒ β = 0.0968 But, β=

Figure 1

RS1 RS1 + RF

0.0968 = we write the following equations: Vd2 = −gm2 RD2 Vt Id3 =

1 + RS2 gm3

(1) Vd2

(2) 1

RF + RS1

gm1

RS2

1 RF + RS1

+ RS2 gm1

If = Id3

Id1 = If

RS1 RS1 +

(3)

(4)

1 gm1

Vr = Id1 RD1

(5)

0.1 0.1 + RF

⇒ RF = 933 (an increase of 33 ). 10.26 All MOSFETs are operating at ID = 100 µA = 0.1 mA and |VOV | = 0.2 V, thus 2 × 0.1 = 1 mA/V 0.2 All devices have |VA | 10 ro = = = 100 k ID 0.1 gm1,2 =

(a)

Substituting the numerical values in (1)–(5), we obtain Vd2 = −4 × 10Vt = −40Vt Id3 =

Q3

(6)

Vd2 1 1 + 0.1 0.9 + 0.1

4 4

Id3 = 2.935Vd2

Q4 Vo

Rs Q1

Q2

(7)

Vt Vr

0.1

1 + 0.1 0.9 + 0.1

4

If = Id3

If = 0.0933Id3 Id1 = If

0.1 +

1 4

= 0.286If

Figure 1

(9)

Vr = 10Id1

(10)

Combining (6)–(10) gives

Figure 1 shows the circuit prepared for determining the loop gain Aβ. Vo = −gm1,2 [ro2 ro4 (R1 + R2 )]Vt

Vr = −31.33Vt

Vr =

⇒ Aβ = 31.33

Thus,

A=

31.33 Aβ = = 313.3 V/V β 0.1

Sedra, Smith, Carusone, Gaudet

R2

(8)

0.1

R1

(1)

R2 Vo = β Vo R1 + R2

Aβ ≡ −

Vr = gm1,2 [ro2 ro4 (R1 + R2 )]β Vt

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–10

= 1(100 100 1000)β

10.28 Refer to Fig. 10.9.

= 47.62β

(a) β =

Thus,

R2 1 Af ideal = = 1 + β R1

A = 47.62 V/V A (b) Af = 1 + Aβ 5=

R1 R1 + R2

R2 1 ⇒ R2 = 4 k

5=1+

47.62 1 + 47.62β

(b) From Example 10.3, we have

⇒ β = 0.179 V/V

Aβ = (gm1 RD1 )(gm2 RD2 )

R2 = 0.179 R1 + R2 ⇒ R2 = 179 k R1 = 821 k 10.27 Because the op amp has an infinite input resistance and a zero output resistance, this circuit is a direct implementation of the ideal feedback structure and thus

1 × 1 + gm1 R1

R1 1 RD2 + R2 + R1

gm1 = 1 1 × (4 × 10)(4 × 10) 1 + 4 × 1 10 + 4 + (1 0.25)

A = 1000 V/V

= 22.54 22.54 Aβ = = 112.7 V/V A= β 0.2

and

Af =

β=

R1 R1 + R2

=

The ideal closed-loop gain is 1 R2 Af ideal = = 1 + β R1 R2 10 ⇒ R2 = 90 k 10 = 1 +

where gm = 4 mA/V

10 = 0.1 V/V 10 + 90

RD = 10 k R1 + R2 = 1 M (the potentiometer resistance)

Aβ = 1000 × 0.1 = 100 Af =

Thus,

A 1 + Aβ

To obtain Af that is exactly 10, we use

⇒ Aβ = 99 β = 0.099 0.099 =

R1 R1 + R2

0.099 =

10 10 + R2

⇒ R2 = 91 k

Sedra, Smith, Carusone, Gaudet

10 × 1000 = 39.6 V/V 10 + 1000 A Af = 1 + Aβ A=4×

1000 = 9.9 V/V = 1 + 100 1000 10 = 1 + Aβ

112.7 = 4.79 V/V 1 + 22.54

10.29 Refer to Fig. 10.9(c) and to the expressions for β, Aβ, and A given in the answer section of Exercise 10.5. RD (R1 + R2 ) A = gm RD + R1 + R2

Thus,

β=

A 1 + Aβ

5=

39.6 1 + 39.6β

⇒ β = 0.175 V/V 0.175 =

R1 R1 + R2

⇒ R1 = 0.175 × 1000 = 175 k 10.30 (a) Setting the open-loop gain to ∞, results in Vbe1 = 0, Ie1 = 0 and Ib1 = 0, which isolates the feedback network as the voltage divider (R1 , R2 ), thus

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–11

β=

R1 R1 + R2

If the loop gain is large, the closed-loop gain approaches the ideal value R2 1 Af ideal = = 1 + β R1 =1+

10 = 11 V/V 1

(b) Figure 3 0.1 mA

8.4 V

0.1 mA Q1

Q2

0

0.003 mA 0.7 V R2

0.3 mA 7.7 V

10 k 0.7 mA Rs 100 k

1 mA

R1 1 k

Vt re1 + Rs

Ie1 =

(1)

Ic1 = α1 Ie1

(2) Ve2 = −(β2 + 1)Ic1 RL [R2 + (R1 Rib )] (3) Vr = Ve2

R1 Rib (R1 Rib ) + R2

(4)

Combining (1) to (4), we can determine Aβ as

Figure 1 The dc analysis is shown in Fig. 1, from which we see that IE1 0.1 mA

Aβ ≡ −

(β2 + 1) RL [R2 + (R1 Rib )] = α1 re1 + Rs

IE2 0.3 mA

×

VE2 = +7.7 V (c) Setting Vs = 0 and eliminating dc sources, the feedback amplifier circuit simplifies to that shown in Fig. 2.

Vr Vt

R1 Rib (R1 Rib ) + R2

Substituting α1 (β2 + 1) = α(β + 1) = β = 100 re1 =

VT 25 mV = 250 = IE1 0.1 mA

Rs = 100 RL = 1 k

Q2 B1

Q1

R1 = 1 k R2 = 10 k

R2

Rib = 101(0.25 + 0.1) = 35.35 k

Rs

R1

RL

we obtain Aβ =

Rib

100 1 [10 + (1 35.35)] 0.25 + 0.1

×

Figure 2

1 35.35 (1 35.35) + 10

= 23.2 Now, breaking the feedback loop at the base of Q1 while terminating the right-hand side of the circuit (behind the break) in the resistance Rib ,

(d) A =

Rib = (β1 + 1)(re1 + Rs )

Af =

results in the circuit in Fig. 3 which we can use to determine the loop gain Aβ as follows:

=

Sedra, Smith, Carusone, Gaudet

23.2 Aβ = = 255.2 V/V β (1/11)

A 1 + Aβ

255.2 = 10.55 V/V 1 + 23.2

Microelectronic Circuits, 8th International Edition

© Oxford University Press 2021

Chapter 10–12

10.31 (a) The feedback circuit consists of the voltage divider (RF , RE ). Thus, β=

RE RE + RF

IC2 2 mA = = 80 mA/V VT 0.025 V

rπ 2 =

β2 100 = 1.25 k = gm2 80

RE = 0.05 k

and, RF 1 Af ideal = = 1 + β RE

re1 =

VT 25 mV = 25 = 0.025 k IE1 1 mA

β3 = 100

Thus,

RC2 = 1 k

RF 0.05 ⇒ RF = 1.2 k 25 = 1 +

RF = 1.2 k we obtain

(b) Figure 1 on next page shows the feedback amplifier circuit prepared for determining t