ECE 255, MOSFET Amplifiers
26 October 2017
In this lecture, the basic configurations of MOSFET amplifiers will be stud-
ied similar to that of BJT. Previously, it has been shown that with the transistor
DC biased at the appropriate point (Q point or operating point), linear rela-
tions can be derived between the small voltage signal and current signal. We
will continue this analysis with MOSFETs, starting with the common-source
amplifier.
1 Common-Source (CS) Amplifier
The common-source (CS) amplifier for MOSFET is the analogue of the common-
emitter amplifier for BJT. Its popularity arises from its high gain, and that by
cascading a number of them, larger amplification of the signal can be achieved.
1.1 Chararacteristic Parameters of the CS Amplifier
Figure 1(a) shows the small-signal model for the common-source amplifier. Here,
RD is considered part of the amplifier and is the resistance that one measures
between the drain and the ground. The small-signal model can be replaced by
its hybrid-π model as shown in Figure 1(b). Then the current induced in the
output port is i = −gm vgs as indicated by the current source. Thus
vo = −gm vgs RD (1.1)
By inspection, one sees that
Rin = ∞, vi = vsig , vgs = vi (1.2)
Thus the open-circuit voltage gain is
vo
Avo = = −gm RD (1.3)
vi
One can replace a linear circuit driven by a source by its Thévenin equiva-
lence. Then from the equivalent-circuit model in Figure 1(b) and the test-
current method by setting vi = 0, the output resistance, which is the Thévenin
resistor is
Ro = RD (1.4)
Printed on October 26, 2017 at 15 : 23: W.C. Chew and Z.H. Chen.
1
Figure 1: (a) Small-signal model for a common-source amplifier. (b) The hybrid-
π model for the common-source amplifier (Courtesy of Sedra and Smith).
If now, a load resistor, RL is connected to the output across RD , then the
voltage gain proper (also called terminal voltage gain), by the voltage divider
formula, is
RL RD RL
Av = Avo = −gm = −gm (RD k RL ) (1.5)
RL + Ro RL + RD
From the fact that Rin = ∞, then vi = vsig . The overall voltage gain, Gv , is the
same as the voltage gain proper, Av , namely
vo
Gv = = −gm (RD k RL ) (1.6)
vsig
1.2 Final Remarks on CS Amplifier
1. The CS amplifiers has infinite input impedance (draws no current at DC),
and a moderately high output resistance (easier to match), and a high
voltage gain (a desirable feature of an amplifier).
2
Figure 2: A CS amplifier with a source resistance: (top) detail circuit, (bottom)
equivalent circuit T model (Courtesy of Sedra and Smith).
2. Reducing RD reduces the output resistance of a CS amplifier, but unfortu-
nately, the voltage gain is also reduced. Alternate design can be employed
to reduce the output resistance (to be discussed later).
3. A CS amplifier suffers from poor high frequency performance, as most
transistor amplifiers do.
2 Common-Source Amplifier with a Source Re-
sistance
As shown in Figure 2, a T model is used for the equivalent circuit for simplicity.
It is seen that the input resistance of the circuit is infinite because no gate
current flows. As a consequence, vi = vsig . However, because of the existence of
the source resistance, less of the input voltage is divided to vgs , by the voltage-
3
divider formula. Thus
1/gm vi
vgs = vi = (2.1)
1/gm + Rs 1 + gm Rs
It is seen that Rs can be used to make vgs small so that there is less nonlinear
distortion as the small-signal approximations will become better. The output
voltage is generated by the controlled current source yielding
vo = −iRD (2.2)
The current i can be found by
vi gm
i= = vi (2.3)
1/gm + Rs 1 + gm Rs
Thus the open-circuit voltage gain (assume that RD is part of the amplifier) is
vo gm RD RD
Avo = =− =− (2.4)
vi 1 + gm Rs 1/gm + Rs
The above shows that including the source resistance reduces the amplifier gain
by a factor of 1 + gm Rs , but linearity and bandwidth performance (to be shown
later) will improve. This is called negative feedback because when the input volt-
age vi or vgs attempts to increase, the voltage drop across Rs increases reducing
vgs . The source resistance is also called source-degeneration resistance.
Since this is a linear circuit, the Thévenin equivalence of the amplifier looking
in from the right can be easily found. The open-circuit voltage allows us to easily
find the equivalent Thévenin voltage source. The equivalent Thévenin resistor
is Ro which is just RD in this case.
When a load resistor RL is added, then the voltage gain proper (also called
terminal voltage gain)
gm (RD k RL ) RD k RL
Av = − =− (2.5)
1 + gm Rs 1/gm + Rs
Because the input resistance is infinite, hence vi = vsig and the overall voltage
gain Gv = Av .
2.1 Summary of the CS Amplifier with Source Resistance
1. The input resistance Rin is infinite.
2. The open-circuit voltage gain, Avo , is reduced by a factor of 1 + gm Rs as
seen in (2.4).
3. For the same nonlinear distortion, the input signal can be increased by a
factor of 1 + gm Rs compared to without Rs .
4. As shall be shown later, the high-frequency response of this design is
improved.
4
Figure 3: (a) Small-signal model for a common-gate amplifier. (b) The T model
equivalent circuit for the common-gate amplifier. Note that the gate current is
always zero in the T model (Courtesy of Sedra and Smith).
In general, the addition of the source resistance Rs gives rise to a “negative”
feedback factor 1 + gm Rs that reduces voltage gain, but improves linearity, and
high-frequency response. Because of the negative-feedback action of Rs , it is
also called the source-degenerate resistance.
3 Common-Gate (CG) Amplifier
The small-signal and a T-model equivalent-circuit common-gate (CG) amplifier
is shown in Figure 3. By inspection, the input resistance Rin is given by
1
Rin = (3.1)
gm
which is typically a few hundred ohms, a low input impedance. The output
voltage is
vi
vo = −iRD , where i = − = −gm vi (3.2)
1/gm
Hence the open-circuit voltage gain is
vo
Avo = = gm RD (3.3)
vi
which is similar to that of the CS amplifier save for a sign change. The output
resistance (or the Thévenin equivalent resistor) of the circuit is
Ro = RD (3.4)
The smaller input impedance is deleterious to the amplifier gain, as by the
voltage divider formula, one gets
vi Rin 1/gm
= = (3.5)
vsig Rin + Rsig 1/gm + Rsig
5
meaning that the vi is attenuated compared to vsig , since Rsig is typically larger
than 1/gm .
When a load resistor RL is connected to the output, the voltage gain proper
(terminal voltage gain) is then
Av = gm RD k RL (3.6)
Thus the overall voltage gain is
1/gm RD k RL
Gv = gm (RD k RL ) = (3.7)
Rsig + 1/gm Rsig + 1/gm
As the input impedance is low, it is good for matching sources with a low input
impedance due the the maximum power theorem, but it draws more current,
implying high power consumption from the signal source.
4 The Source Follower
This is similar to the emitter follower for the BJT, which is used as a voltage
buffer. It is a unit-gain amplifier with a very large input impedance but a
smaller output impedance. Therefore it is good for matching a high-impedance
circuit to a low-impedance circuit.
4.1 Characteristics of a Source Follower
Figure 4 shows the small-signal circuit and a T-model equivalent circuit diagram
for a source follower. The input source is represented by a Thévenin equivalent
voltage vsig and resistor Rsig . A load resistor is connected to the output between
the source and ground.
Since the gate current is zero for this circuit,
Rin = ∞ (4.1)
Using the voltage divider formula, it is seen that voltage gain proper or terminal
voltage gain is
vo RL
Av = = (4.2)
vi RL + 1/gm
For the open-circuit voltage gain, RL = ∞ and
Avo = 1 (4.3)
The output resistance is obtained by replacing the proper part of the amplifier
with a Thévenin equivalence. To this end, with the use of the test-current
method, the value of vi = 0, and thus
Ro = 1/gm (4.4)
6
Figure 4: (a) Common-drain MOSFET amplifier or source follower for small
signals. (b) The T model equivalent circuit for the common-drain or source
follower amplifier. Note that the gate current is always zero in this model
(Courtesy of Sedra and Smith).
7
Because of the infinite input impedance Rin , then vi = vsig , and the overall
voltage gain Gv (also called the total voltage gain) is the same as the voltage
gain proper Av (also called terminal voltage gain)
RL
G v = Av = (4.5)
RL + 1/gm
Since 1/gm is typically small, with large RL , the gain is less than unity, but is
close to unity. Hence, this is a source follower.
5 Summary Table and Comparisons
The following concluding points are in order for the MOSFET and BJT ampli-
fiers.
1. MOS amplifiers have high input impedance (except for CG amplifiers).
This is an advantage over BJT amplifiers.
2. BJT’s have higher transconductance gm than MOSFET’s giving BJT am-
plifiers higher gains.
3. Discrete-circuit amplifiers, e.g., circuits assembled on printed-circuit board
(PCB), BJT’s are prevalent because of their longer history and wider
commercial availability.
4. Because of easier fabrication, integrated circuit (IC) amplifiers are domi-
nated by MOSFET’s.
5. The CS and CE configurations are best suited for gain amplifiers because
of their larger than unity gain. A cascade of them can be used to increase
the gain.
6. The addition of Rs in a CS amplifier improves the linearity of the circuit
and better high frequency performance.
Table 7.4 from Sedra and Smith summarizes the characteristics of the MOS-
FET amplifiers.
6 Discrete-Circuit Amplifiers
Due to history and tradition, most discrete-circuit amplifiers are BJT’s. Also,
capacitive coupling is often used in discrete amplifier design. They make the
design simpler as capacitors are DC blockers, and can be approximated as a
short circuit for AC signals.
8
6.1 A Common-Source (CS) Amplifier
The circuit to be analyzed here is shown in Figiure 5(a). The bias point which
is a DC operating point, is determined by Figure 5(b) where all capacitors are
open circuited. The small signal model, which is for AC signal, is shown in
Figure 5(c) where all capacitors are short circuited.
It is noted that the MOSFET source (S) terminal is grounded for the AC
signal because of the large coupling capacitor CS , and hence, it is also called the
signal ground or AC ground. Hence, CS is also called the bypass capacitor
as its impedance is much smaller than that of RS . The presence of RS is to
stabilize the biasing point. Looking at Figure 5(b), if RS is not there, since VGS
is small, and hence, VG has to be small. Thus, all of the fluctuation of VG will
appear across VGS . However, with RS present, any fluctuation in VG will be
shared by VGS and the voltage drop across RS .
Then CC1 is another coupling capacitor, which will be acting approxi-
mately like a short circuit to AC signals, but is a DC blocker. The second
coupling capacitor CC2 is also acting like a short circuit to the AC signal, or the
small signal. These give the rational for the small signal model in Figure 5(c).
Using the small signal model for AC signals, and the hybrid-π model for
MOSFET as shown in Figure 5(c), it is seen that
Rin = RG1 k RG2 (6.1)
9
Figure 5: (a) Common-source MOSFET amplifier with the biasing circuit in
place. (b) The biasing circuit at DC, where the capacitors are open circuited.
(c) Small signal equivalent circuit model for AC signals where the capacitors
are assumed to be short circuited (Courtesy of Sedra and Smith).
Here, Rin can be kept high by making RG1 and RG2 high, usually in the
megaohm range. It is seen that the voltage gain proper (terminal voltage gain)
is
Av = −gm (RD k RL k ro ) (6.2)
and the overall voltage gain
Rin
Gv = − gm (RD k RL k ro ) (6.3)
Rin + Rsig
6.2 A Common-Emitter Amplifier
This is the most commonly used configuration of the BJT amplifiers, as shown in
Figure 6(a) with the coupling capacitors CC1 and CC2 , and the bypass capacitor
10
Figure 6: (a) Common-emitter BJT amplifier with the biasing circuit in place.
(b) The biasing circuit at DC, where the capacitors are open circuited. (c) Small
signal equivalent circuit model for AC signals where the capacitors are assumed
to be short circuited (Courtesy of Sedra and Smith).
CE in place. These capacitors, to simplify the analysis, are assumed to be open
circuited for DC or the bias-point analysis, but are short circuited for the AC
small signal analysis. Again, as in the MOSFET case, RE is there to stabilize
the bias point of the base voltage.
The equivalent small signal model is shown in Figure 6(b). From it, it is
seen that
Rin = RB1 k RB2 k rπ (6.4)
In the above RB1 and RB2 should be kept large, around tens to hundreds of
kilo-ohms, to maintain high input impedance.
The overall voltage gain Gv is then given by
Rin
Gv = − gm (RC k RL k ro ) (6.5)
Rin + Rsig
11