ch08 (1)
ch08 (1)
0.2 mA
D 8.4 Design the MOS differential amplifier of Fig. 8.5
V to operate at VOV = 0.25 V and to provide a transconduc-
tance gm of 1 mA/V. Specify the W/L ratios and the bias
current. The technology available provides Vt = 0.5 V and
2
V Q Q V
μn Cox = 400 µA/V .
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 631
CHAPTER 8
D 8.6 Design a MOS differential amplifier to operate from transistor is operating. For dc bias calculations you may
±0.6-V supplies and dissipate no more than 0.3 mW in neglect channel-length modulation.
its equilibrium state. Select the value of VOV so that the
value of vid that steers the current from one side of the 1.5 V
pair to the other is 0.2 V. The differential voltage gain
2
Ad is to be 8 V/V. Assume kn = 500 µA/V and neglect
the Early effect. Specify the required values of I, RD ,
PROBLEMS
and W/L. Q6 Q3
vod
vid vid
Q1 Q2
2 2 1.5 V
Figure P8.8
Rs
Specify the required value of R and the W/L ratios for 8.11 For the differential amplifier specified in Prob-
all transistors. Also specify ID and VGS at which each lem 8.2, let vG2 = 0 and vG1 = vid . Find the range of vid needed
632 Chapter 8 Differential and Multistage Amplifiers
PROBLEMS
to steer the bias current from one side of the pair to the other. D 8.18 Design a MOS differential amplifier to operate from
At each end of this range, give the value of the voltage at the ±1-V power supplies and dissipate no more than 1 mW in
common-source terminal, the drain voltages, and vO . the equilibrium state. The differential voltage gain Ad is to
be 10 V/V and the output common-mode dc voltage is to be
8.12 Consider the differential amplifier specified in Prob-
0.2 V. (Note: This is the dc voltage at the drains.) Assume
lem 8.1 with G2 grounded and vG1 = vid . Let vid be adjusted 2
μn Cox = 400 µA/V and neglect the Early effect. Specify I,
to the value that causes iD1 = 0.09 mA and iD2 = 0.07 mA.
RD , and W/L.
CHAPTER 8
8.17 An NMOS differential amplifier is operated at a D 8.22 Figure P8.22 shows a MOS differential amplifer
bias current I of 0.2 mA and has a W/L ratio of 32, with the drain resistors RD implemented using
2
μn Cox = 200 µA/V , VA = 10 V, and RD = 10 k. Find VOV , diode-connected PMOS transistors, Q3 and Q4 . Let Q1 and
gm , ro , and Ad . Q2 be matched, and Q3 and Q4 be matched.
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 633
VDD
CHAPTER 8
at the sources of Q1 and Q2 ? Express these in terms of
the overdrive voltage VOV at which each of Q1 and Q2
operates, and Vt .
(b) For the situation in (a), what current flows in Q3 ? What
Q3 Q4 overdrive voltage VOV3 is Q3 operating at, in terms of VC ,
VOV , and Vt ?
(c) Now consider the case vG1 = +vid /2 and vG2 = −vid /2,
vod
PROBLEMS
where vid is a small signal. Convince yourself that Q3
vid vid
Q1 Q2 now conducts current and operates in the triode region
2 2
with a small vDS . What resistance rDS does it have,
expressed in terms of the overdrive voltage VOV3 at which
it is operating? This is the resistance Rs . Now if all three
transistors have the same W/L, express Rs in terms of VOV ,
I
VOV3 , and gm1,2 .
(d) Find VOV3 and hence VC that result in (i) Rs = 1/gm1,2 ; (ii)
Rs = 0.5/gm1,2 .
Figure P8.22 *8.24 The circuit of Fig. P8.24 shows an effective way of
implementing the resistance Rs needed for the circuit in
(a) Find the differential half-circuit and use it to derive Fig. P8.7. Here Rs is realized as the series equivalent of two
an expression for Ad in terms of gm1,2 , gm3,4 , ro1,2 , MOSFETs Q3 and Q4 that are operated in the triode region,
and ro3,4 . thus, Rs = rDS3 + rDS4 . Assume that Q1 and Q2 are matched
(b) Neglecting the effect of the output resistances ro , find Ad and operate in saturation at an overdrive voltage VOV that
in terms of μn , μp , (W/L)1,2 and (W/L)3,4 . corresponds to a drain bias current of I/2. Also, assume that
(c) If μn = 4μp and all four transistors have the same channel Q3 and Q4 are matched.
length, find (W1,2 /W3,4 ) that results in Ad = 10 V/V.
*8.23 The resistance Rs in the circuit of Fig. P8.7 can be VDD
implemented by using a MOSFET operated in the triode
region, as shown in Fig. P8.23. Here Q3 implements Rs , with RD RD
the value of Rs determined by the voltage VC at the gate of Q3 .
vod
VDD
Q1 Q2 vG2
RD RD vG1
vod
Q3 Q4
Q1 Q2 vG2
vG1 I I
– –
Q3 2 2
–VSS
I I
– VC –
2 2 Figure P8.24
– V SS
(a) With vG1 = vG2 = 0 V, what dc voltages appear at the
Figure P8.23 sources of Q1 and Q2 ? What current flows through Q3
and Q4 ? At what overdrive voltages are Q3 and Q4 oper-
(a) With vG1 = vG2 = 0 V, and assuming that Q1 and Q2 ating? Find an expression for rDS for each of Q3 and Q4
are operating in saturation, what dc voltages appear and hence for Rs in terms of (W/L)1,2 , (W/L)3,4 , and gm1,2 .
634 Chapter 8 Differential and Multistage Amplifiers
(b) Now with vG1 = vid /2 and vG2 = − vid /2, where vid is leave the active mode at vCE ≤ 0.3 V. The collector resistors
PROBLEMS
a small signal, find an expression of the voltage gain RC = 82 k, and the power supplies are ±1.2 V. The bias
Ad ≡ vod /vid in terms of gm1,2 , RD , (W/L)1,2 , and (W/L)3,4 . current I = 20 µA and is supplied with a simple current
source.
*8.25 A design error has resulted in a gross mismatch in the
circuit of Fig. P8.25. Specifically, Q2 has twice the W/L ratio (a) For vB1 = vB2 = VCM = 0 V, find VE , VC1 , and VC2 .
of Q1 . If vid is a small sine-wave signal, find: (b) Find the input common-mode range.
CHAPTER 8
(c) If vB2 = 0, find the value of vB1 that increases the current
(a) ID1 and ID2 .
in Q1 by 10%.
(b) VOV for each of Q1 and Q2 .
(c) The differential gain Ad in terms of RD , I, and VOV .
D 8.28 Design the circuit of Fig. 8.14 to provide a differen-
tial output voltage vO of 1 V when the differential input signal
VDD is 10 mV. A current source of 1 mA and a positive supply
of +5 V are available. What is the largest possible input
common-mode voltage for the circuit to operate as required?
RD RD Assume α 1.
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 635
CHAPTER 8
input resistance Rid . For each circuit, what dc voltage appears 8.35 A particular differential amplifier operates from an
across the bias current source(s) in the quiescent state (i.e., emitter current source I = 0.4 mA. Each of the collector
with vid = 0)? Hence, which of the two circuits will allow a resistances RC = 20 k and a load resistance RL = 40 k
larger negative VCM ? is connected between the two collectors. If the amplifier is
fed in the manner shown in Fig. P8.52 with Rsig = 100 k,
VCC
find the overall voltage gain. Let β = 100.
PROBLEMS
RC RC
amplifier in Fig. P8.36 assuming that β = 100.
vod
5V
vid vid
VCM VCM 25 k
2 2
vo
vi
Re Re 500
VEE
(a) Figure P8.36
VCC
8.37 An npn differential amplifier with I = 0.4 mA, VCC =
RC RC VEE = 2.5 V, and RC = 5 k utilizes BJTs with β = 100 and
vBE = 0.7 V at iC = 1 mA. If vB2 = 0, find VE , VC1 , and VC2
vod obtained with vB1 = +0.5 V, and with vB1 = −0.5 V. Assume
that the current source requires a minimum of 0.3 V for
proper operation.
vid vid 8.38 An npn differential amplifier with I = 0.4 mA, VCC =
VCM VCM
2 2 Re 2 VEE = 2.5 V, and RC = 5 k utilizes BJTs with β = 100 and
vBE = 0.7 V at iC = 1 mA. Assuming that the bias current is
obtained by a simple current source and that all transistors
require a minimum vCE of 0.3 V for operation in the active
I I mode, find the input common-mode range.
– –
2 2
8.39 Repeat Exercise 8.7 for an input of –0.3 V.
VEE 8.40 Consider the BJT differential amplifier when fed with
a common-mode voltage VCM as shown in Fig. 8.15(a). As is
(b) often the case, the supply voltage VCC may not be pure dc
Figure P8.33 but might include a ripple component vr of small amplitude
and a frequency of 120 Hz (see Section 3.6). Thus the supply
8.34 Consider a bipolar differential amplifier that, in addi- voltage becomes VCC + vr . Find the ripple component of the
tion to the collector resistances RC , has a load resistance collector voltages, vC1 and vC2 , as well as of the difference
RL connected between the two collectors. What does the output voltage vod ≡ vC2 − vC1 . Comment on the differential
differential gain Ad become? amplifier response to this undesirable power-supply ripple.
636 Chapter 8 Differential and Multistage Amplifiers
PROBLEMS
D 8.41 Consider the differential amplifier of Fig. 8.14 and (b) With no emitter resistances Re , use the large-signal
let the BJT β be very large: model to find iC1 and iC2 when vid = 20 mV.
(c) Now find the value of Re that will result in the same
(a) What is the largest input common-mode signal that can
iC1 and iC2 as in (b) but with vid = 200 mV. Use the
be applied while the BJTs remain comfortably in the
large-signal model.
active region with vCB = 0?
(d) Calculate the effective transconductance Gm as the ratio
(b) If the available power supply VCC is 2.0 V, what value of
of the difference current, (iC1 − iC2 ), to vid in the cases
CHAPTER 8
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 637
Rsig 2
CHAPTER 8
(a) Bearing in mind that for a BJT to remain in the active
mode, vBC should not exceed 0.4 V, show that when
vid has a peak v̂id , the maximum input common-mode
vsig
voltage VCMmax is given by –
2
v̂id v̂
VCMmax = VCC + 0.4 − − Ad VT + id vid
2 2
PROBLEMS
VCM
(b) For the case VCC = 2.5 V and v̂id = 10 mV, use the vsig
–
relationship above to determine VCMmax for the case Ad = 2
50 V/V. Also find the peak output signal v̂od and the
required value of IRC . Now if the power dissipation in
the circuit is to be limited to 1 mW in the quiescent Rsig 2
state (i.e., with vid = 0), find I and RC . (Remember to
include the power drawn from the negative power supply Figure P8.52
−VEE = −2.5 V.)
(c) If VCMmax is to be +1 V, and all other conditions remain
the same, what maximum gain Ad is achievable? 8.53 Find the voltage gain and the input resistance of the
amplifier shown in Fig. P8.53 assuming β = 100.
8.49 For the differential amplifier of Fig. 8.14, let VCC =
+5 V and IRC = 4 V. Find the differential gain Ad . Sketch and
clearly label the waveforms for the total collector voltages 3V
vC1 and vC2 and for (vO ) for the case:
20 k
vB1 = 1 + 0.005 sin(ωt)
vo
vB2 = 1 − 0.005 sin(ωt)
8.50 Consider a bipolar differential amplifier in which vi Q1 Q2
the collector resistors RC are replaced with simple current
sources implemented using pnp transistors. Sketch the circuit
and give its differential half-circuit. If VA = 20 V for all 250 250
Rin
transistors, find the differential voltage gain achieved. 0.2 mA
8.51 A bipolar differential amplifier having resistance Re
inserted in series with each emitter [as in Fig. P8.33(a)] is
biased with a constant current I. When both input terminals Figure P8.53
are grounded, the dc voltage measured across each Re is
found to be 3 VT and that measured across each RC is found
to be 80VT . What differential voltage gain Ad do you expect
8.54 Derive an expression for the small-signal voltage gain
the amplifier to have?
vo /vi of the circuit shown in Fig. P8.54 in two different ways:
8.52 A bipolar differential amplifier with emitter-
(a) as a differential amplifier
degeneration resistances Re and Re is fed with the
(b) as a cascade of a common-collector stage Q1 and a
arrangement shown in Fig. P8.52. Derive an expression for
common-base stage Q2
the overall differential voltage gain Gv ≡ vod /vsig . If Rsig is of
such a value that vid = 0.5vsig , find the gain Gv in terms of Assume that the BJTs are matched and have a current gain
RC , re , Re , and α. Now if β is doubled, by what factor does Gv α, and neglect the Early effect. Verify that both approaches
increase? lead to the same result.
638 Chapter 8 Differential and Multistage Amplifiers
PROBLEMS
CHAPTER 8
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 639
CHAPTER 8
8.60 For the differential amplifier shown in Fig. P8.2, let Q1 Note that this equation indicates that RD can be deliberately
2
and Q2 have kp (W/L) = 4 mA/V , and assume that the bias varied to compensate for the initial variability in gm and RD ,
current
has an output resistance of 30 k. Find VOV ,
source that is, to minimize Acm .
gm , Ad , Acm , and the CMRR (in dB) obtained with the In a MOS differential amplifier for which RD = 5 k and
output taken differentially. The drain resistances are known RSS = 25 k, the common-mode gain is measured and found
to have a mismatch of 2%. to be 0.002 V/V. Find the percentage change required in one
of the two drain resistors so as to reduce Acm to zero (or close
D *8.61 The differential amplifier in Fig. P8.61 uti-
PROBLEMS
to zero).
lizes a resistor RSS to establish a 0.02-mA dc bias current.
Note that this amplifier uses a single 2-V supply and thus the D 8.63 A MOS differential amplifier utilizing a simple
dc common-mode voltage VCM cannot be zero. Transistors Q1 current source to provide the bias current I has a CMRR of
2
and Q2 have kn W/L = 1 mA/V , Vt = 0.4 V, and λ = 0. 60 dB. If the CMRR must be raised to 100 dB by adding
a cascode transistor to the current source, what must the
(a) Find the required value of VCM .
intrinsic gain A0 of the cascode transistor be? If the cascode
(b) Find the value of RD that results in a differential gain Ad
transistor is operated at VOV = 0.2 V, what must its VA be? If
of 15 V/V.
for the specific technology utilized VA = 5 V/µm, specify the
(c) Determine the dc voltage at the drains.
channel length L of the cascode transistor.
(d) Determine the single-ended-output common-mode gain
VD1 /VCM . (Hint: You need to take 1/gm into account.) 8.64 In a bipolar differential-amplifier circuit, the bias cur-
(e) Use the common-mode gain found in (d) to find the rent generator consists of a simple common-emitter transistor
change in VCM that results in Q1 and Q2 entering the triode operating at 200 µA. For this transistor, and those used in the
region. differential pair, VA = 20 V and β = 50. What common-mode
input resistance would result? Assume RC ro .
*8.69 In a particular BJT differential amplifier, a production 8.75 One approach to “offset correction” involves the
error results in one of the transistors having an emitter–base adjustment of the values of RC1 and RC2 so as to reduce the
junction area that is twice that of the other. With the inputs differential output voltage to zero when both input terminals
grounded, how will the emitter bias current split between the are grounded. This offset-nulling process can be accom-
two transistors? If the output resistance of the current source plished by utilizing a potentiometer in the collector circuit,
is 400 k and the resistance in each collector (RC ) is 15 k, as shown in Fig. P8.75. We wish to find the potentiometer
find the common-mode gain obtained when the output is setting, represented by the fraction x of its value connected in
taken differentially. Assume α 1. (Hint: The CM signal series with RC1 , that is required for nulling the output offset
current vicm /REE will split between Q1 and Q2 in the same voltage that results from:
ratio as the bias current I does.) VCC
(x) (1 x)
Section 8.4: DC Offset
1k
D 8.70 An NMOS differential amplifier for which the RC1 RC2
MOSFETs have a transconductance parameter kn and whose 5k 5k
drain resistances RD have a mismatch RD is biased with a
current I.
Q1 Q2
(a) Find expressions for Ad and VOS in terms of kn , RD ,
RD /RD , and I. Use these expressions to relate VOS and
Ad .
2
(b) If kn = 4 mA/V , RD = 10 k, and RD /RD = 0.02, find
1 mA
the maximum gain realized if VOS is to be limited to 1
mV, 2 mV, 3 mV, 4 mV, and 5 mV. For each case, give
the value of the required bias current I. Note the trade-off
Figure P8.75
between gain and offset voltage.
D 8.71 An NMOS amplifier, whose designed operating (a) RC1 being 4% higher than nominal and RC2 4% lower than
point is at VOV = 0.3 V, is suspected to have a variability nominal
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 641
CHAPTER 8
(b) Q1 having an area 5% larger than nominal, while Q2 has find the input voltage that would restore current balance to
area 5% smaller than nominal. the differential pair. Repeat using large-signal analysis and
compare results.
8.76 A differential amplifier for which the total emitter bias
current is 400 µA uses transistors for which β is specified D 8.84 A large fraction of mass-produced differential-
to lie between 80 and 200. What is the largest possible input amplifier modules using 20-k collector resistors is found
bias current? The smallest possible input bias current? The to have an input offset voltage ranging from +2 mV to
largest possible input offset current? –2 mV. By what amount must one collector resistor be
PROBLEMS
adjusted to reduce the input offset to zero? If an adjustment
D 8.77 An NMOS differential pair is to be used in an
mechanism is devised that raises one collector resistance
amplifier whose drain resistors are 10 k ± 1%. For the pair,
2 while correspondingly lowering the other, what resistance
kn W/L = 4 mA/V . Decide whether to use a bias current I of
change is needed? If a potentiometer connected as shown
160 µA or 360 µA. Contrast the differential gain and input
in Fig. P8.75 is used, what value of potentiometer resistance
offset voltage for the two possibilities.
(specified to 1 significant digit) is needed? Assume that the
8.78 An NMOS differential pair operating at a bias current offset is entirely due to the finite tolerance of RC .
2
I of 100 µA uses transistors for which kn = 200 µA/V and
W/L = 10. Find the three components of input offset voltage Section 8.5: The Differential Amplifier with
under the conditions that RD /RD = 4%, (W/L)/(W/L) = a Current-Mirror Load
4%, and Vt = 5 mV. In the worst case, what might the
total offset be? For the usual case of the three effects being 8.85 A current-mirror-loaded NMOS differential amplifier
independent, what is the offset likely to be? is fabricated in a technology for which |VA | = 5 V/µm.
All the transistors have L = 0.5 µm. If the differential-pair
8.79 A differential amplifier uses two transistors whose β transistors are operated at VOV = 0.25 V, what open-circuit
values are β 1 and β 2 . If everything else is matched, show that
differential gain is realized?
the input offset voltage is approximately VT 1/β1 − 1/β2 .
Evaluate VOS for β 1 = 50 and β 2 = 100. (Hint: The bias 8.86 The differential amplifier of Fig. 8.31(a) is biased with
current I will still split equally between the two emitters.) I = 200 µA. All transistors have L = 0.5 µm, and Q1 and
Q2 have W/L = 20. The circuit is fabricated in a process for
8.80 Two possible differential amplifier designs are consid- 2
which μn Cox = 400 µA/V and |VA | = 6 V/µm. Find gm1,2 ,
ered, one using BJTs and the other MOSFETs. In both cases, ro2 , ro4 , and Ad .
the collector (drain) resistors are maintained within ±2% of
nominal value. The MOSFETs are operated at VOV = 200 D 8.87 In a current-mirror-loaded differential amplifier of
mV. What input offset voltage results in each case? What the form shown in Fig. 8.31(a), all transistors
are charac-
terized by k W/L = 4 mA/V , and VA = 5 V. Find the bias
2
does the MOS VOS become if the devices are increased in
width by a factor of 4 while the bias current is kept constant? current I for which the gain vo /vid = 20 V/V.
*8.81 A differential amplifier uses two transistors having VA 8.88 A current-mirror-loaded NMOS differential amplifier
values of 100 V and 200 V. If everything else is matched, operates with a bias current I of 200 µA. The NMOS tran-
find the resulting input offset voltage. Assume that the two are operated at VOV = 0.2 V and the PMOS devices at
sistors
transistors are intended to be biased at a VCE of about 10 V. VOV = 0.3 V. The Early voltages are 20 V for the NMOS and
12 V for the PMOS transistors. Find Gm , Ro , and Ad . For what
**8.82 A differential amplifier is fed in a balanced or
value of load resistance is the gain reduced by a factor of 2?
push–pull manner, and the source resistance in series with
each base is Rs . Show that a mismatch Rs between the 8.89 The differential amplifier in Fig. 8.35(a) is operated
values of the two source resistances gives rise to an input with I = 500 µA, with devices for which VA = 10 V and
offset voltage of approximately (I/2β)Rs / [1 + (gm Rs )/β]. β = 100. What differential input resistance, output resistance,
short-circuit transconductance, and open-circuit voltage gain
*8.83 In a particular BJT differential amplifier, a production
would you expect? What will the voltage gain be if the input
error results in one of the transistors having an emitter–base
resistance of the subsequent stage is equal to Rid of this stage?
junction area twice that of the other. With both inputs
grounded, find the current in each of the two transistors and 8.90 Figure P8.90 shows a differential cascode amplifier
hence the dc offset voltage at the output, assuming that the with an active load formed by a Wilson current mirror.
collector resistances are equal. Use small-signal analysis to Utilizing the expressions derived in Chapter 7 for the output
642 Chapter 8 Differential and Multistage Amplifiers
PROBLEMS
resistance of a bipolar cascode and the output resistance (d) If I is delivered by a simple NMOS current source
of the Wilson mirror, and assuming all transistors to be operated at the same VOV and having the same channel
identical, show that the differential voltage gain Ad is given length as the other four transistors, determine the CMRR
approximately by obtained.
1
Ad = βgm ro 8.92 The MOS differential amplifier of Fig. 8.31(a) is
3
biased with a simple current mirror delivering I = 200 µA.
Evaluate Ad for the case of β = 100 and VA = 20 V.
CHAPTER 8
Q1 Q2
vd
I
I
RSS
I
Q8 Q7
VEE 5V
Figure P8.90
Q5 Q6
D 8.91 Design the current-mirror-loaded differential
MOS amplifier of Fig. 8.31(a) to obtain a differential gain
of 60 V/V. The technology available provides μn Cox =
2
4μp Cox = 500 µA/V , Vt = 0.4 V, and VA = 15 V/µm and
operates from ±0.8-V supplies. Use a bias current I = 200
Figure P8.94
µA and operate all devices at VOV = 0.15 V. Assume all
transistors have the same channel length.
(a) Find the W/L ratios of the four transistors. *8.95 (a) Sketch the circuit of a current-mirror-loaded MOS
(b) Specify the channel length required of all transistors. differential amplifier in which the input transistors are
(c) If VICM = 0, what is the allowable range of vO ? cascoded and a cascode current mirror is used for the load.
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 643
CHAPTER 8
vg3 vid 4
ro Q3 Q4 ro
i13 i12 i2 i1
1
vo (g r ) v
2 m o id
PROBLEMS
i11 i3
i9 i6 i4
(vid 2) Q1 ro ro Q2 (vid 2)
i8 i10 i7 i5
vs vid 4
Figure P8.96
(b) Show that if all transistors are operated at an overdrive given by
voltage VOV and have equal Early voltages VA , the gain is (W/L)A
VOS1 = VOV /2
given by (W/L)A
2 where VOV is the overdrive voltage at which Q1 and Q2 are
Ad = 2 VA /VOV operating.
(b) Repeat for a mismatch (W/L)M in the W/L ratios of the
Evaluate the gain for VOV = 0.20 V and VA = 10 V.
mirror transistor Q3 and Q4 to show that the corresponding
8.96 Figure P8.96 shows the current-mirror-loaded MOS VOS is given by
differential amplifier prepared for small-signal analysis. We (W/L)M
have pulled ro out of each transistor; thus, the current in VOS2 = VOV /2
(W/L)M
the drain of each transistor will be gm vgs . We have also
simplified matters by indicating approximate values for where VOV is the overdrive voltage at which Q1 and Q2 are
some of the node voltages. For instance, the output voltage operating.
vo = 12 gm ro vid , which we have derived in the text. The (c)
For a circuit in which all transistors are operated at
voltage at the common sources has been found to be roughly VOV = 0.2 V and all W/L ratios are accurate to within ±1%
+vid /4, which is very far from the virtual ground we might of nominal, find the worst-case total offset voltage VOS .
expect. Also, the voltage at the drain of Q1 is approximately
D 8.98 Consider the bias design of the Wilson-loaded
−vid /4, confirming our contention that the voltage there is
cascode differential amplifier shown in Fig. P8.90.
vastly different from the output voltage, hence the lack of
balance in the circuit and the unavailability of a differential (a) What is the largest signal voltage possible at the output
half-circuit. Find the currents labeled i1 to i13 in terms of without Q7 saturating? Assume that the CB junction
(gm vid ). Determine their values in the sequence of their conducts when the voltage across it exceeds 0.4 V.
numbering and assume gm r o 1. Assume that all transistors (b) What should the dc bias voltage established at the output
are operating at the same VOV . Write the current values on (by an arrangement not shown) be in order to allow for
the circuit diagram and reflect on the results. positive output signal swing of 1.5 V?
(c) What should the value of VBIAS be in order to allow for a
8.97 This problem investigates the effect of transis-
negative output signal swing of 1.5 V?
tor mismatches on the input offset voltage of the
(d) What is the upper limit on the input common-mode
current-mirror-loaded MOS differential amplifier of
voltage VICM ?
Fig. 8.31(a). For this purpose, ground both input terminals
and short-circuit the output node to ground. 8.99 In addition to the random offset voltages that result
(a) If the amplifying transistors Q1 and Q2 exhibit a W/L from the mismatches inevitably present in the differen-
mismatch of (W/L)A , find the resulting short-circuit output tial amplifier, the current-mirror-loaded bipolar differential
current and hence show that the corresponding VOS is amplifier suffers from a systematic offset voltage. This is due
644 Chapter 8 Differential and Multistage Amplifiers
PROBLEMS
to the error in the current transfer ratio of the current-mirror 8.101 Consider the current-mirror-loaded MOS differential
load caused by the finite β of the pnp transistors that make amplifier of Fig. 8.31(a) in two cases:
up the mirror. The situation is illustrated in Fig. P8.99. Show
(a) Current source I is implemented with a simple current
that the resulting input offset voltage is given by
mirror.
VOS = −2VT /βP (b) Current source I is implemented with the modified
Wilson current mirror shown in Fig. P8.101.
CHAPTER 8
and evaluate VOS for βP = 50. Also, if the simple current Recalling that for the simple mirror RSS = ro Q and for
S
mirror is replaced by the pnp version of the Wilson mirror the Wilson mirror RSS gm7 ro7 ro5 , and assuming that all
transistors have the same VA and k W/L, show that for
of Fig. 8.37(a), find the expected systematic input offset
I
I
Figure P8.99 RSS
Q Q
Figure P8.101
v
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 645
VDD 0.9 V
CHAPTER 8
Q3 Q4
IREF 200 A
Q6
PROBLEMS
A
Q1 Q2 B vo
Q5
Q8 Q7
VSS 0.9 V
Figure P8.104
D 8.103 A current-mirror-loaded MOS differential ampli- 8.105 Consider the amplifier of Fig. 8.37, whose parameters
fier has a differential voltage gain Ad of 40 V/V. The output are specified in Example 8.6. If a manufacturing error results
resistance of the current-mirror load is equal to the output in the W/L ratio of Q7 being 24/0.4, find the current that Q7
resistance of the amplifier at 40 k. If the current-mirror gain will now conduct. Thus find the systematic offset voltage that
is 0.98 A/A and its input resistance Rim is 500 , find the will appear at the output. (Use the results of Example 8.6.)
required value of the output resistance RSS of the bias current Assuming that the open-loop gain will remain approximately
source to obtain a CMRR of 80 dB. unchanged from the value found in Example 8.6, find the
corresponding value of input offset voltage, VOS .
Section 8.6: Multistage Amplifiers
D 8.104 The two-stage CMOS op amp in Fig. P8.104 is *8.106 Figure P8.106 shows a bipolar op-amp circuit that
resembles the CMOS op amp of Fig. 8.37. Here, the
fabricated in a 0.18-µm technology having kn = 4kp =
2
400 µA/V , Vtn = −Vtp = 0.4 V. input differential pair Q1 –Q2 is loaded in a current mir-
ror formed by Q3 and Q4 . The second stage is formed
(a) With A and B grounded, perform a dc design that will by the current-source-loaded common-emitter transistor
result in each of Q1 , Q2 , Q3 , and Q4 conducting a drain Q5 . Unlike the CMOS circuit, here there is an output
current of 100 µA and each of Q6 and Q7 a current of stage formed by the emitter follower Q6 . The func-
200 µA. Design so that all transistors operate at 0.2-V tion of capacitor CC will be explained later,
in Chap-
overdrive voltages. Specify the W/L ratio required for ter 11. All transistors have β = 100, VBE = 0.7 V, and
each MOSFET. Present your results in tabular form. ro = ∞.
What is the dc voltage at the output (ideally)?
(b) Find the input common-mode range. (a) For inputs grounded and output held at 0 V (by negative
(c) Find the allowable range of the output voltage. feedback, not shown) find the emitter currents of all
(d) With vA = vid /2 and vB = −vid /2, find the voltage gain transistors.
vo /vid . Assume an Early voltage of 6 V. (b) Calculate the gain of the amplifier with RL = 1 k.
646 Chapter 8 Differential and Multistage Amplifiers
5V
PROBLEMS
0.4 mA
0.5 mA
Q6
CHAPTER 8
Q1 Q2 CC
vo
Q5
1 mA RL
Q3 Q4
5V
Figure P8.106
8.107 A BJT differential amplifier, biased to have re = 50 (a) Find the dc bias current in each of the three transistors.
and utilizing two 50- emitter resistors and 5-k loads, Also find the dc voltage at the output. Assume VBE =
drives a second differential stage biased to have re = 25 . 0.7 V, β = 100, and neglect the Early effect.
All BJTs have β = 100. What is the voltage gain of the first (b) Find the input resistance and the output resistance.
stage? Also find the input resistance of the first stage, and (c) Use the current-gain method to evaluate the voltage gain
the current gain from the input of the first stage to the vo /vi .
collectors of the second stage.
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem
Problems 647
Table P8.110
CHAPTER 8
Transistor Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8
360 µA/V , μp Cox = 90 µA/V , VA = 6 V for all devices,
2 2
8.114 In the multistage amplifier of Fig. 8.38, emitter
PROBLEMS
VDD = VSS = 0.9 V. Determine the width of Q6 , W, that will resistors are to be introduced: 100 in the emitter lead
ensure that the op amp will not have a systematic
offset of each of the first-stage transistors and 25 for each of
voltage. Then, for all devices evaluate ID , VOV , VGS , gm , the second-stage transistors. What is the effect on input
and ro . Provide your results in a table similar to Table 8.1. resistance, the voltage gain of the first stage, and the overall
Also find A1 , A2 , the open-loop voltage gain, the input voltage gain? Use the bias values found in Example 8.7.
common-mode range, and the output voltage range. Neglect
D 8.115 Consider the circuit of Fig. 8.38 and its output
the effect of VA on the bias currents.
resistance. Which resistor has the most effect on the output
D *8.111 In a particular design of the CMOS op amp of resistance? What should this resistor be changed to if the
Fig. 8.37 the designer wishes to investigate the effects of output resistance is to be reduced by a factor of 2? What
increasing the W/L ratio of both Q1 and Q2 by a factor will the amplifier gain become after this change? What
of 4. Assuming that all other parameters are kept unchanged, other change can you make to restore the amplifier gain to
refer to Example 8.6 to help you answer the following approximately its prior value?
questions:
**8.116 The MOS differential amplifier shown in
(a) What change results in VOV and in gm of Q1 Fig. P8.116 utilizes three current mirrors for signal
and Q2 ? transmission: Q4−Q6 has a transmission factor of 2 [i.e.,
(b) What change results in the voltage gain of the input (W/L)6 /(W/L)4 = 2], Q3−Q5 has a transmission factor of 1,
stage? In the overall voltage gain? and Q7−Q8 has a transmission factor of 2. All transistors
are
(c) What is the effect on the input offset voltages? (You sized to operate at the same overdrive voltage, VOV . All
might wish to refer to Section 8.4). transistors have the same Early voltage VA .
8.112 Consider the input stage of the CMOS op amp in (a) Provide in tabular form the values of ID , gm , and ro of
Fig. 8.37 with both inputs grounded. Assume that the two each of the eight transistors in terms of I, VOV , and VA .
sides of the input stage are perfectly matched except that the
threshold voltages of Q3 and Q4 have a mismatch Vt . Show VDD
that a current gm3 Vt appears at the output of the first stage.
What is the corresponding input offset voltage? Q5 Q3 Q4 Q6
D 8.113 The two-stage op amp in Figure P8.104 is fab-
ricated in a 65-nm technology having kn = 5.4 × kp =
2
540 µA/V and Vtn = −Vtp = 0.35 V. The amplifier is Q1 Q2
operated with VDD = +1.2 V and VSS = 0 V. vo
(a) Perform a dc design that will cause each of Q1 , Q2 , Q3 ,
and Q4 to conduct a drain current of 200 µA and each
I
of Q6 and Q7 to conduct a current of 400 µA. Design so
that all transistors operate at 0.15-V overdrive voltages.
Specify the W/L ratio required for each MOSFET. Q7 Q8
Present all results in a table.
(b) Find the input common-mode range. VDD
(c) Find the allowable range of the output voltage.
(d) With vA = vid /2 and vB = −vid /2, find the voltage gain Figure P8.116
vo /vid . Assume an Early voltage of 2.4 V.
648 Chapter 8 Differential and Multistage Amplifiers
PROBLEMS
(b) Show that the differential voltage gain Ad is given by transfer ratio is given by
Ad = 2gm1 ro6 ro8 = VA /VOV 1
Ai Ai (ideal) 1 −
gm ro
(c) Show that the CM gain is given by
ro6 ro8 1 where gm and ro are the parameters of the input transistor
Acm of the mirror.)
RSS gm7 ro7
CHAPTER 8
problems with green numbers are considered essential; * = difficult problem; ** = more difficult; *** = very challenging
= simulation; D = design problem