CG
CG
• CG Amplifier
• Examples
• Biasing in MOS Amplifier Circuits
ECE315 / ECE515
Common Gate (CG) Amplifier
• CG Amplifier- Input is applied at
the Source and the output is
sensed at the Drain.
• The Gate terminal is used for
establishing appropriate bias
conditions for the transistor.
• Its characteristic can be studied through large-signal behavior as well.
• For large 𝑉𝑖𝑛 i.e. for 𝑉𝑖𝑛 > 𝑉𝑏 − 𝑉𝑇 : 𝑀1 is off and therefore: 𝑉𝑜𝑢𝑡 = 𝑉𝐷𝐷
1 W
I D nCox Vb Vin VT
2
• For lower 𝑉𝑖𝑛 :
2 L
1 W 2
Vout VDD I D RD VDD nCox Vb Vin VT RD
2 L
Vout W V
• Then small signal gain is: nCox Vb Vin VT 1 T RD
Vin L Vin
ECE315 / ECE515
Common Gate (CG) Amplifier (contd.)
Vout W
• Since, 𝜕𝑉𝑇 𝜕𝑉𝑖𝑛 = 𝜕𝑉𝑇 𝜕𝑉𝑆𝐵 = η: nCox Vb Vin VT 1 RD
Vin L
Av g m 1 RD
• Let us look at small-signal model – without channel length modulation and biased
with a constant current source
Vout
KVL in loop-1: KCL at this Av ( gm gmb ) RD gm (1 ) RD
Vin
v1 vin 0 node
V
( g m g mb )V1 out 0
RD
ECE315 / ECE515
CG Amplifier (contd.)
• Input Impedance:
KVL in loop-1: KVL in loop-2:
V1 VX
RD I X ro I X ( gm gmb )VX VX
VX RD ro RD 1
Rin
I X 1 ( gm g mb )ro ( g m g mb )ro ( g m g mb )
VX ro
• Case-I: RD 0 Rin
I X 1 ( g m g mb )ro
• Output Impedance:
Rout 1 ( gm gmb )ro RS ro RD
Example – 1
• Derive the small-signal voltage
gain expression for this amplifier.
Consider the cases when
channel length modulation are
absent and present.
ECE315 / ECE515
Example – 1 (contd.)
• Case-I: λ = 0 both for M1 and M2 → In such a case M2 presents a
degenerating impedance of 1/gm2 to a single stage CS amplifier as shown
below.
1 Vout
gm2 g m1V1
RD
1 g
Vin V1 g m1V1. V1 1 m1
gm2 gm2
1
ro 2 Vout
1
gm2 g m1V1
ro 2 RD
gm2 1 1
Vin V1 g m1V1. ro 2 V1 1 g m1 ro 2
gm2 gm2
RD
Av
Simplification gives: 1 1
ro 2
g m1 g m 2
ECE315 / ECE515
Example – 1 (contd.)
• Case-III: λ = 0 for M2 and λ ≠ 0 for M1 → In such a case the small
signal model looks like:
g m1V1 ro1
1
gm2
ECE315 / ECE515
Example – 2
• Derive expression for the small-signal voltage gain of the following circuit.
(Assume: λ ≠ 0 for both M1 and M2. Neglect body effect)
VX
ro 2 ro1 ro1 ( g m1ro 2 g mb1ro 2 ) Rout
IX
KVL in loop-1 KVL in loop-2
V1 I X ro 2
VX I X ro 2 ( I X g m1V1 g mb1V1 )ro1
ECE315 / ECE515
Biasing using Single Power Supply
• The general form of a • We typically attempt to satisfy three main
single-supply MOSFET bias design goals:
amplifier biasing circuit is:
1) Maximize Gain
Typically, the small-signal voltage gain
of a MOSFET amplifier will be
Av gm
proportional to transconductance 𝑔𝑚 .
Thus, to maximize the amplifier voltage gain, we must
maximize the MOSFET transconductance.
A:
Recall that the transconductance g 2 K V V
depends on the DC excess gate voltage: m GS T
To maximize 𝐴𝑣 , maximize 𝐼𝐷
2) Maximize Voltage Swing
Recall that if the DC drain voltage 𝑉𝐷 is biased
too close to 𝑉𝐷𝐷 , then even a small small-signal vD(t ) VD vd (t ) VDD
drain voltage 𝑣𝑑 (𝑡) can result in a total drain
voltage that is too large, i.e.:
In other words, the MOSFET enters cutoff, and the result is a distorted signal!
ECE315 / ECE515
Biasing using Single Power Supply (contd.)
• To avoid this (to allow 𝑣𝑑 (𝑡) to be as large as possible without MOSFET entering
cutoff), we need to bias our MOSFET such that the DC drain voltage 𝑉𝐷 is as small
as possible.
• Note that the drain voltage is: VD VDD RD ID
• Therefore 𝑉𝐷 is minimized by designing the bias circuit such that the DC drain
current 𝐼𝐷 is as large as possible.
• However, we must also consider the signal distortion that occurs when the
MOSFET enters triode. This of course is avoided if the total drain-to-source
voltage remains greater than the excess gate voltage, i.e.:
vDS (t ) VDS vds (t ) VGS VT
• Thus, to avoid the MOSFET triode mode—and the resulting signal distortion—we
need to bias our MOSFET such that the DC voltage is as large as possible.
To minimize signal distortion, maximize 𝑉𝐷𝑆
ECE315 / ECE515
Biasing using Single Power Supply (contd.)
3) Minimize Sensitivity to changes in 𝑲, 𝑽𝑻
• We find that MOSFETs are sensitive to temperature—specifically, the value of
𝐾 is a function of temperature.
• Likewise, the values of 𝐾 and 𝑉𝑇 are not particularly constant with regard to the
manufacturing process.
• Both of these facts lead to the requirement that our bias design be insensitive to
the values of 𝐾 and 𝑉𝑇 . Specifically, we want to design the bias network such
that the DC bias current does not change values when 𝐾 and/or 𝑉𝑇 does.
• Mathematically, we can express this d ID d ID
and
requirement as minimizing the value: dK dVT
• These derivatives can be minimized by maximizing the value of source resistor 𝑹𝑺 .
Q: So what is current 𝑰𝑳 ?
A: Note that the gate voltage of each MOSFET is the same (i.e., 𝑉𝐺𝑆1 = 𝑉𝐺𝑆2 ),
and if the MOSFETS are the same (i.e., 𝐾1 = 𝐾2 , 𝑉𝑇1 = 𝑉𝑇2 ), and if the second
MOSFET is likewise in saturation.
IL K2 VGS 2 VT 2
2
• If the load resistor becomes larger than 𝑹 + 𝑽𝑻𝟏 𝑰𝒓𝒆𝒇 , the voltage 𝑽𝑫𝑺𝟐 will drop
below the excess gate voltage 𝑽𝑮𝑺𝟐 − 𝑽𝑻𝟐 , and thus the second MOSFET will
enter the triode region. As a result, the drain current will not equal 𝑰𝒓𝒆𝒇 —the
current source will stop working!
ECE315 / ECE515
The MOSFET Current Mirror (contd.)
Although the circuit is sometimes referred to as a current sink, understand that
the circuit is clearly a way of designing a current source.
• Therefore:
VS VG VGS
R2 I
VDD VT
R1 R2 K
ECE315 / ECE515
MOSFET Biasing using Current Mirror (contd.)
• Since we are biasing with a current source, we do not need to worry about drain
current stability—the current source will determine the DC drain current for all
conditions (i.e., 𝐼𝐷 = 𝐼).
• We might conclude, therefore, that we should make DC source voltage VS as
small as possible. After all, this would allow us to maximize the output voltage
swing (i.e., maximize 𝐼𝐷 𝑅𝐷 and 𝑉𝐷𝑆 ).
• Note however, that the source voltage 𝑉𝑆 of the MOSFET is numerically equal to
the drain voltage 𝑉𝐷2 (and thus 𝑉𝐷𝑆2 ) of the second MOSFET of the current
mirror.
Q: So what?!
A: The voltage must be greater than: VGS 2 VT 2 VGS 1 VT 1
VDD Iref R VT 1
in order for the second MOSFET to remain in saturation.
There is a minimum voltage across the current source in order
for the current source to properly operate!
ECE315 / ECE515
MOSFET Biasing using Current Mirror (contd.)
• Thus, to maximize output
VS VGS 1 VT 1
swing, we might wish to set:
(although to be practical, we should make 𝑉𝑆 slightly
greater than this to allow for some design margin).
Q: How do we “set” the DC source voltage 𝑉𝑆 ??
A: By setting the DC gate voltage 𝑉𝐺 !!
• Recall that the DC voltage I
𝑉𝐺𝑆 is determined by the DC VGS VT
current source value I: K