EEEB273 – Electronics Analysis & Design II

Learning Outcome
Able to:
• Describe the mechanism by which a differential-mode
signal and common-mode signal are produced in a
MOSFET diff-amp.
• Describe the dc transfer characteristics of a MOSFET
diff-amp.

Reference: Neamen, Chapter 11

6.1) 6.1)
• Figure 11.21 shows the basic MOSFET • Example 11.10
differential-pair configuration.
• Two identical transistor M1 and M2,
whose sources are connected
together, are biased by a constant
current source IQ which is connected
to a negative supply V_
• The drains of M1 and M2 are
connected through RD to a positive
supply V+
• Assume that M1 and M2 are always
biased in the saturation region. Even
Figure 11.21:
Basic MOSFET
with vG1 = vG2 = 0, the M1 and M2 can
differential-pair be biased in saturation region by the
configuration current source IQ.

6.2) ! 6.2) !
• Use circuit in Figure 11.21. Assume M1 and M2 are • If vd > 0, then vG1 > vG2 and vGS1 > vGS2
matched and neglecting their output resistances:
which implies that iD1 > iD2
iD1 = K n (vGS1 − VTN )
2
(11.60(a))
• Since iD1 + iD2 = IQ (11.62)
iD 2 = K n (vGS 2 − VTN )
2
(11.60(b))

Taking square roots and subtracting

( iD1 − I Q − iD1 ) =(2
K n .vd ) =K v
2 2
n d

iD1 − iD 2 = K n (vGS1 − vGS 2 ) = K n .vd iD1 (I Q − iD1 ) =

1
(
I Q − K n vd2 )
2
where vd = vG1 − vG 2 = vGS1 − vGS 2 (11.61)

is the differential-mode input voltage iD2 1 − I Q iD1 +

1
4
(
I Q − K n vd2 )
2
=0

Lecturer: Dr Jamaludin Bin Omar 6-1

 EEEB273 – Electronics Analysis & Design II

6.2) ! 6.2) !
• Applying the quadratic formula, rearranging terms, • The normalized drain currents are
and noting that iD1 > IQ / 2 and vd > 0, can obtain
iD1 1 Kn Kn 2
= + .vd 1 − vd (11.68)
IQ Kn IQ Kn 2 (11.66) IQ 2 2IQ 2IQ
iD1 = + .vd 1− vd
2 2 2IQ
and iD 2 1 Kn Kn 2 (11.69)
• Using Equation (11.44) can also obtain = − .vd 1 − vd
IQ 2 2IQ 2IQ
IQ Kn IQ Kn 2
iD 2 = − .vd 1 − vd (11.67) These equations describe the dc transfer
2 2 2IQ characteristics for the circuit. They are plotted in
Figure 11.23 as a function of normalized differential
d (
input voltage v / 2 I / K
Q n )

6.2) ! 6.2) !
• Forward transconductance = the slope of the dc
transfer characteristics for the iD1 curve.
• From Figure 11.23, maximum forward
transconductance occurs at vd = 0, so that
diD1
g f (max) = (11.71)
dvd vd = 0
Fig 11.23: Normalized dc transfer characteristics
for MOSFET differential amp • Using (11.48),
• From (11.68) and (11.69), at a specific differential Kn IQ
input voltage, bias current IQ is switched entirely to gm
g f (max) = = (11.72)
one transistor or the other. IQ 2 2
This occurs when vd = (11.70)
max
Kn where gm is each transistor’s transconductance.

6.2) ! 6.3) " #

• The slope of iD2 characteristic curve at vd = 0 is the • Figure 11.24 is AC equivalent circuit of diff-amp
same, except negative. configuration, showing only diff voltage and signal
currents as a function gm. Assume output resistance
Note: looking into the current source is infinite.

• Similar to BJT diff-amp, the differential-mode input • Using this equivalent circuit, the
voltage must be held within a small range of one-sided output voltage at vo2 is
voltages so as to remain linear. g m vd
vo 2 ≡ vo = + RD
2
• However, the vd(max) for the MOSFET diff-amp is
much larger than vd(max) for the BJT diff-amp. • The differential voltage gain is
Why? Because the gain of the MOSFET diff-amp is Kn IQ
much smaller than the gain of the bipolar (BJT) diff- vo g m RD
Figure 11.24: AC Ad = = = .RD
amp. (This will be seen later!) equivalent circuit vd 2 2
for MOSFET diff
amplifier

Lecturer: Dr Jamaludin Bin Omar 6-2

 EEEB273 – Electronics Analysis & Design II

6.4) Small-signal Equivalent Circuit Analysis 6.4) Small-signal Equivalent Circuit Analysis (Cont)
• Can determine basic relationships for differential- • Assume that transistors are matched, with = 0 for
mode gain, common-mode gain, and CMRR from an each transistor, and that constant-current source is
analysis of the small-signal equivalent circuit. represented by a finite output resistance (Ro )

• Two transistors are biased at the same quiescent

current, and gm1 = gm2 gm

• KCL equation at node Vs

Vs
g mV gs1 + g mV gs 2 =
Ro
From the circuit, Vgs1 = V1 - Vs and Vgs2 = V2 - Vs

Vs
Fig 11.25: Small-signal equivalent
Then, g m (V1 + V2 − 2Vs ) = (11.76)
circuit, MOSFET differential amplifier Ro

6.4) Small-signal Equivalent Circuit Analysis (Cont) 6.4) Small-signal Equivalent Circuit Analysis (Cont)
• Solving for Vs V1 + V2 • Based on relationships between input voltages V1
Vs = (11.77) and V2 and differential- and common-mode voltages,
1 as given by Equation (11.29), Equation (11.79) can be
2+
g m Ro written
g m RD g m RD
Vo = Vd − Vcm
• For a one-sided output at the drain of M2 2 1 + 2 g m Ro (11.80)
Vo = Vd 2 = − (g mV gs 2 )RD = − ( g m RD )(V2 + Vs ) • The output voltage, in general form, is
• Substitute (11.77) and rearranging terms yields Vo = Ad Vd + AcmVcm
1 • The transconductance gm of the MOSFET is
V2 1 + − V1
g m Ro g m = 2 K n I DQ = 2 K n I Q
Vo = − g m RD (11.79)
1
2+
g m Ro

6.5) Differential- and Common-mode Gains 6.6) $ % & %

• Therefore, Common-mode rejection ratio (CMRR) is

Differential-mode gain is CMRR =

1
2
[
1 + 2 K n I Q . Ro ] (11.83)
g R R K n IQ
Ad = m D = 2 K n I Q D = .R D (11.82(a))
2 2 2 This demonstrate that the CMRR for the MOSFET
diff-amp is also a strong function of the output
Common-mode gain is resistance of the constant-current source.
− g m RD − 2 K n I Q .R D (11.82(b)) The value of CMRR can be increased by
Acm = =
1 + 2 g m Ro 1 + 2 2 K n I Q . Ro increasing the output resistance of the current
source. The increase can be accomplished by
using a more sophisticated current source circuit,
For an ideal current source, RO = , so Acm = 0. such as the MOSFET cascode current mirror.

Lecturer: Dr Jamaludin Bin Omar 6-3

 EEEB273 – Electronics Analysis & Design II

6.7) Differential- and Common-Mode Input Impedances '( ) # $

• At low frequencies, input impedance of
a MOSFET is essentially infinite.
• This means that both the differential-
and common-mode input resistances of
a MOSFET diff-amp are infinite.
Example 11.11: Compare forward transconductance
of a MOSFET diff pair to that of a bipolar pair.
For MOSFET diff pair, assume Kn = 0.5mA/V2 and IQ = 1mA.
For bipolar diff pair, assume IQ = 1mA.
For MOSFET using (11.72): gm = 1.0mA/V
For BJT using (11.15): gm = 19.2mA/V
gm (BJT) > gm (MOSFET)
gm proportional to Ad Ad (BJT) > Ad (MOSFET)

But Rin (MOSFET) > Rin (BJT)

Design trade-off between gain and input

impedance.

Lecturer: Dr Jamaludin Bin Omar 6-4