Frequency Response of Transistor Ampliers

for
Electronic Circuits
http://cktse.eie.polyu.edu.hk/eie304
by
Prof. Michael Tse
September 2004
C.K. Tse: Frequency Response of
Transistor Ampliers
2
Introduction
A convenient starting point is the common-emitter
amplier. Our understanding of this amplier (up to
EC1) is that it can provide a fair amount of voltage
gain. We apply a small input at the base and we
expect to get a pretty wide swing at the collector side.
Thats amplication. But, so far, we have assumed
that the signals are not of very high frequency such
that the parasitic capacitances in the transistor do not
have any signicant effects. Remember that the
impedance of a capacitor is inversely proportional to
wC. So, if C is very small, it is practically an open-
circuit at moderately low frequency. But as soon as
we increase the signal frequency, the capacitor
becomes less and less like an open-circuit, and in fact
its impedance begins to drop. Therefore, at high
frequencies, we must take into account the presence
of parasitic capacitances in order to get a fuller
picture of the voltage gain.
R
L
V
BE
+

V
CE
+

I
C
V
CC
input
output
C.K. Tse: Frequency Response of
Transistor Ampliers
3
Review of common-emitter conguration
R
L
v
BE
+

v
CE
+

I
C
V
CC
R
B1
R
B2
v
in
v
o
+

+
+
coupling capacitors
(large enough so that they become short-
circuit at signal frequencies)
~
~
Basic circuit:
R
s
C.K. Tse: Frequency Response of
Transistor Ampliers
4
Review of common-emitter conguration
Small-signal model for low-frequency operation:
g
m
v
be
~
r
p r
o R
L
R
B1
||R
B2
+
v
in

+
v
o

+
v
be

~
R
s
+

Small-signal parameters of BJT:

transconductance i
c
/v
be
current gain i
c
/i
b
input resistance of BJT v
be
/i
b
output resistance of BJT
g
m
=
q
kT
I
c
=
I
c
V
T

I
c
25mV
at room temperature
r

g
m

This is not an ohmic resistor!

r
o
=
V
A
I
c
where V
A
is Early voltage of the BJT
C.K. Tse: Frequency Response of
Transistor Ampliers
5
Review of common-emitter conguration
Small-signal model for low-frequency operation:
g
m
v
BE
~
r
p r
o R
L
R
B1
||R
B2
+
v
in

+
v
o

+
v
BE

~
R
s
+

Gain =
Input resistance = R
in
= R
B1
||R
B2
|| r

Output resistance = R
o
= R
L
||r
o
v
o
v
in
=
R
B1
|| R
B2
R
B1
|| R
B2
( )
+ R
s
|
|
|
|
|
|
g
m
R
L
|| r
o
( ) [ |
g
m
R
L
if R
s
0 and r
o

Parameters of CE amplier:
C.K. Tse: Frequency Response of
Transistor Ampliers
6
What happens when frequency increases?
f (in logscale)

o
=g
m
r

f
T
0
20log
i
c
i
b
(dB)
Lets measure the current gain which is just |i
c
/i
b
|.
arg
i
c
i
b
(deg)
0
45
o
90
o
135
o
180
o
f (in logscale)
starts to drop at f
p
(pole)
and falls to 1 at f
T
.
f
p
f
z
A zero is observed at f
z
.
But this zero gives further negative phase
shift instead of positive phase shift. In
control theory, this kind of zero is called
right-half-plane (RHP) zero.
C.K. Tse: Frequency Response of
Transistor Ampliers
7
Why?
Lets look at the circuit.
i
b
i
c
We omit the biasing circuit, and focus on the short-circuit current gain.
C

g
m
v
be
r
o
r

i
b
i
c
Small-signal model:
We insert a parasitic
capacitor across each
pair of terminals of the
BJT. But since the
output is shorted, we
can omit the one across
C and E in this case.
+
v
be

v
be
r

+
v
be
(1/ jC

)
+
v
be
(1/ jC

)
i
b
= 0
v
be
=
i
b
1
r

+ j C

+ C

( )
Nodal equation of the B node:

B C
E
The current i
c
, from KCL, is
Hence, we get
i
c
= g
m
v
be
v
be
jC

= v
be
g
m
jC

( )
i
c
i
b
= g
m
r

1 j
C

g
m
1+ j C

+ C

( )
r

|
|
|
|
|
|
|
|
|
|
|
|
I
c
I
b
V
EE
+V
CC
C.K. Tse: Frequency Response of
Transistor Ampliers
8
Current gain at high frequencies
From the above analysis, we see that the current gain has a pole and a RHP zero.
This is consistent with the frequency response of |i
c
/i
b
| we saw earlier.
The pole frequency is
The zero frequency is
The low-frequency current gain is

= g
m
r

( j) =
i
c
i
b
= g
m
r

1 j
C

g
m
1+ j C

+ C

( )
r

|
|
|
|
|
|
|
|
|
|
|
|
f
p
=
1
2 C

+ C

( )
r

(in Hz)
f
z
=
g
m
2C

(in Hz)
f
p
f
z
g
m
r

C.K. Tse: Frequency Response of

Transistor Ampliers
9
f
p
f
z
g
m
r

Transition frequency (unity current gain frequency)

We can nd f
T
by setting |(j

)| = 1.
1= g
m
r

1 j
T
C

g
m
1+ j
T
C

+ C

( )
r

1= g
m
r

( )
2
1+
T
2
C

g
m
[
\
|

)
j
2
1+
T
2
C

+ C

( )
2
r

2
[
\
|
|
|
|
|

)
j
j
j
j
j
f
T
transition frequency
1+
T
2
C

+ C

( )
2
r

2
= g
m
r

( )
2
+
T
2
C

g
m
[
\
|

)
j
2

T
2
=
g
m
r

( )
2
1
C

+ C

( )
2
r

g
m
[
\
|

)
j
2

g
m
r

( )
2
C

+ C

( )
2
r

g
m
C

+ C

( )
or f
T
=
g
m
2 C

+ C

( )
C.K. Tse: Frequency Response of
Transistor Ampliers
10
b
o
Principle of gain-bandwidth product
There is an interesting property about (j).
g
m

o
C

+ C

( )
(j)
g
m

1
C

+ C

( )
g
m
1 C

+ C

( )
b
1
0
expressed in dB
2 f
T
Beyond the pole frequency,
the product of b and
bandwidth is constant.
A:

g
m

o
C

+ C

( )
=
g
m
C

+ C

( )
B:

g
m

1
C

+ C

( )
=
g
m
C

+ C

( )
A
B
C C:
1
g
m
1 C

+ C

( )
=
g
m
C

+ C

( )
2 f
T
||
The transition frequency f
T
is a very
important parameter for studying the
frequency response of a BJT
amplier. It tells us how (and r

)
changes as frequency increases.

C.K. Tse: Frequency Response of

Transistor Ampliers
11
How about MOSFET ?
C
gs
C
gd
g
m
v
gs
r
o
i
g
i
d
Small-signal model:
The main difference
between MOSFET and
BJT is that there is no r

in the model of
MOSFET or r

= .
+
v
gs

G D
S
( j) =
i
d
i
g
= g
m
1 j
C
gd
g
m
j C
gs
+ C
gd
( )
|
|
|
|
|
|
|
|
|
|
|
|
We can easily nd the
transfer function of the
current gain as:
2 f
T

(j)
slope = 20dB/dec

T
=
g
m
C
gs
+ C
gd
( )

g
m
C
gs
The transition frequency is given by
since C
gd
is much smaller than C
gs
.
The gain-bandwidth product principle is
valid at any frequency.
C.K. Tse: Frequency Response of
Transistor Ampliers
12
Frequency response of CE amplier
R
L I
c
V
CC
R
B1
R
B2
v
in
v
o
+

R
s
We consider the CE amplier again, but this time, as a voltage amplier.
v
o
v
in
=
R
B1
|| R
B2
R
B1
|| R
B2
( )
+ R
s
|
|
|
|
|
|
g
m
R
L
|| r
o
( ) [ |
g
m
R
L
if R
s
0 and r
o

Recall that the voltage gain ay low frequencies is

C.K. Tse: Frequency Response of
Transistor Ampliers
13
Frequency response of CE amplier
Complete small-signal model:
C

g
m
v
be
r

+
v
be

B C
E
C
ce
+

R
B
R
s
v
in
r
o
R
L
Questions
How many poles?
How many zeros?
Where are the poles and zeros?
Most important is the rst dominant
pole p
1
which limits the gain!
p
1
p
2
z
p
1
p
2
z
?
?
C.K. Tse: Frequency Response of
Transistor Ampliers
14
What are poles and zeros? (Year 1 material)
Theory (if you still remember!)
On the complex number plane, a pole is the point where the output response is even when
the input is nite. Also, a zero is the point where the output response is 0 even when the
input is nite.
All poles must be on LHS of the complex plane. Otherwise, the circuit is unstable.
But zeros can be anywhere.
For circuits with ONLY ONE type of reactive elements (capacitors only), all poles and
zeros are real numbers. Thus, all poles are LHS real numbers. [Some textbooks use the
term negative real poles instead of LHS real poles!]
x x
p
1
p
2
o
z
1
real
imaginary (actually the realistic frequency axis)
j
o
z
2
C.K. Tse: Frequency Response of
Transistor Ampliers
15
What are poles and zeros? (Year 1 material)
In transfer function forms, we write
where A
o
is the low-frequency gain.
x x
p
1
p
2
o
z
1
real
imaginary (actually the realistic frequency axis)
j
o
z
2
F( j) = A
o
1+
j
z
2
[
\
|

)
j
1
j
z
1
[
\
|

)
j
1+
j
p
1
[
\
|

)
j
1+
j
p
2
[
\
|

)
j
Careful! Watch the sign for the RHP zero!
C.K. Tse: Frequency Response of
Transistor Ampliers
16
What are poles and zeros? (Year 1 material)
In practice
We look at the imaginary axis, which is actually the realistic frequency axis.
A pole [or zero] becomes just a corner point on the realistic frequency axis where
the response starts to fall [rise] at a rate of 20 dB/dec.
p
1
p
2
z
1
z
2
20 dB/dec
40 dB/dec
20 dB/dec

We know all poles are LHS

poles, but zeros can be LHS or
RHS.
What is the difference between a
RHS zero and a LHS zero? To
see the difference, we have to
look at the phase.
C.K. Tse: Frequency Response of
Transistor Ampliers
17
What are poles and zeros? (Year 1 material)
Phase shift
A LHS pole comes with negative phase shift.
One LHS pole gives 90.
A LHS zero comes with positive phase shift.
One LHS zero gives +90.
A RHS zero comes with negative phase shift.
One RHS zero gives 90.
p
2
z
1
z
2
20 dB/dec
40 dB/dec
20 dB/dec

p
1
90
180
270
0
NOTE: z
1
is a RHP zero, and z
2
is a LHP zero.
x x
p
1
p
2
o
z
1
real
jw
o
z
2
C.K. Tse: Frequency Response of
Transistor Ampliers
18
Frequency response of CE amplier
Complete small-signal model:
C

g
m
v
be
r

+
v
be

B C
E
C
ce
+

R
B
R
s
v
in
r
o
R
L
For this circuit, there are two independent capacitors, hence the circuit is second order.
Hence, there must be TWO LHS poles. Moreover, there is at least ONE zero due to C

because g
m
v
be
can pull all its current through C

resulting in zero current owing to R

L
.
Lets nd the frequency where this happens!
v
o
v
in
=
R
B1
|| R
B2
R
B1
|| R
B2
( )
+ R
s
|
|
|
|
|
|
g
m
R
L
|| r
o
( ) [ |
1-
j
z
[
\
|

)
j
1+
j
p
1
[
\
|

)
j
1+
j
p
2
[
\
|

)
j
The low-frequency gain:
C.K. Tse: Frequency Response of
Transistor Ampliers
19
What is a pole? A very simple approach!
To nd the poles, we rst recall that the roll-off is typically a result of an
equivalent low-pass RC lter.
R
eq
C
eq
gain in dB

1
R
eq
C
eq
slope = 20 dB/dec
Basically, if we see a node along the signal path which has
(i) a substantial resistance toward the input side; and
(ii) a substantial capacitance to ground,
then there will be a pole!
C
eq
R
eq
C.K. Tse: Frequency Response of
Transistor Ampliers
20
Finding poles of CE amplier
Complete small-signal model:
C

g
m
v
be
r

+
v
be

B C
E
C
ce
+

R
B
R
s
v
in
r
o
R
L
To nd the poles, we examine the two nodes B and C. Our aim is to nd the equivalent RC
lters at B and C.
First consider node B. We can assume that (before roll-off) the current source g
m
v
be
ows
in the output load R
L
// r
o
such that
So, the voltage across the capacitor C

is
v
CE
= g
m
R
L
|| r
o
( )
v
be
v
BC
= v
be
[1+ g
m
R
L
|| r
o
( )
]
C.K. Tse: Frequency Response of
Transistor Ampliers
21
Finding the rst pole of CE amplier
Consider node B:
C

+
v
be

B
E
+

R
B
R
s
v
in
i
C
= jC

v
be
[1+ g
m
R
L
|| r
o
( )
]
i
C
i
C
i
C
= jC

v
be
i
C
+ i
C
= jC

v
be
[1+ g
m
R
L
|| r
o
( )
] + jC

v
be
= j C

1+ g
m
R
L
|| r
o
( ) [ |
+ C

}
v
be
C
eq
= C

1+ g
m
R
L
|| r
o
( ) [ |
+ C

Equivalent model at node B:

C
r

+
v
be

B
E
+

R
B
R
s
v
in
C

[1+g
m
(R
L
||r
o
)]
Clearly, at node B, we can see that
(i) capacitor C

is pumping current equal to

(ii) capacitor C

is pumping current equal to

They combine to pump current equal totally to
So, we can put one equivalent capacitor to node B:
C.K. Tse: Frequency Response of
Transistor Ampliers
22
First pole is due to Miller
Equivalent model for forward signal ow:
C
r

+
v
be

B
E
+

R
B
R
s
v
in
C

[1+g
m
(R
L
||r
o
)]
Observe that C

has been EXPANDED by a factor of [1+g

m
(R
L
||r
o
)] which is just the dc gain.
This is called Miller effect.
The expanded capacitor at node B to ground is called Miller capacitor:
C
M
= C

[1+g
m
(R
L
||r
o
)]
C
M
C.K. Tse: Frequency Response of
Transistor Ampliers
23
First pole of CE amplier
Equivalent model for forward signal ow:
C
r

+
v
be

B
E
+

R
B
R
s
v
in
C

[1+g
m
(R
L
||r
o
)]
A pole can be found from the equivalent low-pass lter at node B. This is the Miller effect pole.
Clearly, the equivalent R and C are
C
eq
= C

+ C
M
= C

+ C

[1+g
m
(R
L
||r
o
)]
R
eq
= R
s
|| R
B
|| r

This pole is
C
M
p
1
=
1
C
eq
R
eq
=
1
C

+ C

1+ g
m
R
L
|| r
o
( ) [ | }
R
s
|| R
B
|| r

( )
This is a normal LHP pole.
C.K. Tse: Frequency Response of
Transistor Ampliers
24
Finding the second pole of CE amplier
Complete small-signal model:
C

g
m
v
be
r

+
v
be

B C
E
C
ce
+

R
B
R
s
v
in
r
o
R
L
Next consider node C. We can assume that the signal at B looking from C is comparatively
small and that node B is essentially grounded. Therefore, we can approximate that C

goes
to ground at C.
Clearly, the combined capacitance to ground is C

+ C
ce
.
Also, since the equivalent resistance is R
L
|| r
o
, the pole is
g
m
v
be
C
E
C
ce r
o
R
L
C

Equivalent model at node C:

p
2
=
1
C

+ C
ce
( )
R
L
|| r
o
( )
This is a normal LHP pole.
C.K. Tse: Frequency Response of
Transistor Ampliers
25
Finding the zero of CE amplier
Complete small-signal model:
C

g
m
v
be
r

+
v
be

B C
E
C
ce
+

R
B
R
s
v
in
r
o
R
L
To nd the zero, we can simply try to nd the frequency (on complex plane) where the
response is 0. This is quite easy! Almost by inspection, we have, when output is 0,
So, the zero is +g
m
/C, which is a RHP zero.
v
bc
= v
be

g
m
v
be
sC
= v
be
s =
g
m
C
C.K. Tse: Frequency Response of
Transistor Ampliers
26
Complete transfer function of CE amplier
Equivalent model for forward signal ow:
C
r

+
v
be

B
+

R
B
R
s
v
in
C

[1+g
m
(R
L
||r
o
)]
The complete transfer function for the gain is
C
M
g
m
v
be
E
C

+
C
ce
r
o
R
L
v
o
v
in
=
R
B1
|| R
B2
R
B1
|| R
B2
( )
+ R
s
|
|
|
|
|
|
g
m
R
L
|| r
o
( ) [ |
1
j
z
[
\
|

)
j
1+
j
p
1
[
\
|

)
j
1+
j
p
2
[
\
|

)
j
z =
g
m
C
p
2
=
1
C

+ C
ce
( )
R
L
|| r
o
( )
p
1
=
1
C
eq
R
eq
=
1
C

+ C

1+ g
m
R
L
|| r
o
( ) [ | }
R
s
|| R
B
|| r

( )
note the -ve sign (its RHP)
C.K. Tse: Frequency Response of
Transistor Ampliers
27
Complete frequency response of CE amplier
Gain in dB

p
1
p
2
z
90
180
270
The RHP zero gives a
further 90 phase shift!
C.K. Tse: Frequency Response of
Transistor Ampliers
28
Locations of poles and the RHP zero
g
m
(mS)
f
p
1
p
2
z
0.1
10
1M 10M 100M
C
gd
(pF)
f
p
1
p
2
z
0.1
10
1M 10M 100M
1
100
These are not Bode plots!
They actually show the
positions of the poles and
the RHP zero for different
values of g
m
and C
gd
.
Note the pole splitting
when g
m
is greater than a
certain value. This actually
improves the stability, as we
will see when we discuss
feedback later.
pole splitting
C.K. Tse: Frequency Response of
Transistor Ampliers
29
Miller effect gives the dominant pole
The rst pole (dominant pole) denes the roll-off frequency and limits the
bandwidth of the amplier.
The main cause, as we have seen, is Miller effect!
In general, we can extend Miller effect to any situation where a capacitor
appears across the input and output nodes of an amplier.
v
i
v
o
= Av
i
A
C
v
i
v
o
= Av
i
A
(1+A)C
Forward signal model:
C.K. Tse: Frequency Response of
Transistor Ampliers
30
Discussion
Is there any way to beat Miller?
Direction 1: kill C
A capacitor ying over a signal input and a swinging amplied output. We can
kill Miller if we can either stop the input or the output from moving! The
question is how to do it, with the signal amplication still maintained.
Direction 2: kill R
How about R
s
, which is clearly an evil that causes the roll off? Can we make it
as small as possible? What kind of amplier should be used to buffer the input
so that R
s
can be smaller?
The process for making Miller disappear is called broadbanding an amplier. I
will tell you more about beating Miller in the nal-year elective High Frequency
Circuit Design.