ELE 2110A Electronic Circuits

Week 12:
Output Stages, Frequency Response

(2 hours only)
ELE2110A 2008

Lecture 12 - 1

Topics to cover
z

Output Stages

Amplifier Frequency Response

Reading Assignment:
Chap 15.3, 16.1 of Jaeger and Blalock or
Chap 14.1 14.4 of Sedra & Smith
ELE2110A 2008

Lecture 12 - 2

Multistage Amplifiers
Practical amplifiers usually consist of a number of stages connected in
cascade.
z

The first (input) stage is usually required to provide

a high input resistance
a high common-mode rejection for a differential amplifier

Middle stages are to provide

majority of voltage gain
conversion of the signal from differential mode to single-end
mode
shifting of the dc level of the signal

The last (output) stage is to provide

a low output resistance in order to
z
z

avoid loss of gain and

provide the current required by the load (power amplifiers)

ELE2110A 2008

Lecture 12 - 3

Example
z

The input stage (Q1, Q2) is

differential-in and differentialout
biased by current source Q3

(Q4, Q5) is a differential-in

and single-ended-out stage
biased by current source Q6

Q7 provides
additional gain
shifting the dc level of the
signal

ELE2110A 2008

The output stage Q8 is an

emitter follower

Lecture 12 - 4

Output Stages
z

Function of an output stage is

To provide a low output resistance so that it can deliver
the output signal to the load without loss of gain

Requirements of an output stage:

Large input signal range
z

b/c it is the final stage of the amplifier, and usually deals with
relatively large signals.
Small-signal approximations and models either are not
applicable or must be used with care.

Low distortion
High power efficiency

ELE2110A 2008

Lecture 12 - 5

Classification of Output Stages

z

Class A
Class B

Class AB

Class C

Class A: the transistor

conducts for the entire cycle
of the input signal
Class B: the transistor
conducts for only half the
cycle
Class AB: conduction cycle is
greater than 180o and less
than 360o
Used for opamp output stage
and audio power amplifiers

Collector or Drain current waveforms

of different output stages
ELE2110A 2008

Class C: conduction cycle is

less than 180o
Used for radio-frequency
(RF) power amplifications
(mobile phones, radio and
TV)
Lecture 12 - 6

Class-A Amplifier: Source/Emitter Follower

For a source follower biased by an ideal
current source, vGS is fixed and

2 I SS

vO = vI VGS1 = vI VTN +
Kn

Input range:

VSS + VGS vI VDD + VTN

Output range:

VSS vO VDD
The largest output voltage is

vo V
sin t
DD
ELE2110A 2008

(if VSS=VDD)
Lecture 12 - 7

Source Emitter with Load

To maintain class A operation,
is > 0 at all times:
v
i =I
+ o 0
S
SS R
L
vo I SS RL
For largest output amplitude:
We have:

vo V
sin t
DD

VDD sin t I SS RL for all t

The lowest value for the LHS occurs when sin t = -1,

VDD I SS RL
I SS
ELE2110A 2008

VDD

RL
Lecture 12 - 8

Power Efficiency
Average power supplied to the source follower:

VDD sin t
(
)
I
V
V
+
+
0 SS DD SS RL VDD dt

= I SS (VDD + VSS ) = 2 I SSVDD

(if VSS=VDD)

Pav =

1
T

Average power delivered to the load:

2

The largest output

voltage is

vo V
sin t
DD

ELE2110A 2008

VDD

2
VDD
2

=
Pac =
RL
2 RL

Efficiency of amplifier is:

V 2 /( 2 R )
P
L
= ac = DD
P
2I V
av
SS DD
V 2 /( 2 R )
L
DD
= 25 %

2V
/ R V
L DD
DD

I SS

VDD
RL

- Low efficiency
Lecture 12 - 9

Push-Pull Operation: Class B

When a push-pull amplifier is operated in Class B, all of the output current

comes either from the current-sourcing transistor or from the currentsinking device but never from both at the same time.
Source: B. Putzeys, Digital Audios final frontier, IEEE Spectrum, Mar 2003.
ELE2110A 2008

Lecture 12 - 10

Class-B Amplifier
z

A complementary pair of source followers

biased at zero source current
When
V v V
I
TN
TP

ELE2110A 2008

neither transistor conducts

No quiescent (DC) current consumption!

Lecture 12 - 11

Class-B Amplifier
z

When VI > VTN,

M1 turns on and acts as an source
follower, vo vI - VTN
M2 off

When VI < VTP,

M2 turns on and acts as an source
follower, vo vI VTP
M1 off

ELE2110A 2008

Power efficiency is high, can be up to

about 80%

Disadv.: Output waveform suffers from a

dead-zone Large distortion

Lecture 12 - 12

Class AB

Class AB exhibits less distortion by allowing the transistors to work

together when the output signal is near zero, in what is called the
crossover region.
Source: B. Putzeys, Digital Audios final frontier, IEEE Spectrum, Mar 2003.
ELE2110A 2008

Lecture 12 - 13

Class-AB Amplifiers

Remove dead zone by biasing transistors into conduction but at a low

quiescent current level
Distortion less than Class-B but worse than Class-A amplifier

z
z

For each transistor, 1800<conduction angle<3600 Class AB amplifier

Power efficiency lower than Class-B but higher than Class-A amplifier
ELE2110A 2008

Lecture 12 - 14

Class-AB Amplifiers
Biasing examples:

DC currents:

K n V

GG V

I =

D
TN

2 2

ELE2110A 2008

I R
I = I exp B B
C S
2V
T

Lecture 12 - 15

Topics to cover
z

Output Stages

Amplifier Frequency Response

ELE2110A 2008

Lecture 12 - 16

Frequency Response of Amplifiers

A typical amplifier:

Block DC and low frequency signals

By-pass high frequency currents
Amplifiers gain is frequency dependent!
ELE2110A 2008

Lecture 12 - 17

Typical Amplifier Transfer Function

Mid-band gain

Lower cut-off frequency

z
z

Upper cut-off frequency

In low frequency side, drop in gain is caused by coupling and

bypass capacitors
In high frequency side, drop in gain is caused by transistors
parasitic capacitors
More on this topic later

In the mid-band range, no capacitors are in effect:

Coupling and bypass capacitors are short circuits
Transistor parasitic capacitors are open circuits

ELE2110A 2008

Lecture 12 - 18

Estimate fL: Short-Circuit Time Constant Method

z

Lower cutoff frequency for a network

with n coupling and bypass
capacitors can be estimated by:

1
L i =1 R C
iS i
f = / 2
L L

RiS = resistance at terminals of ith

capacitor Ci with all other capacitors
replaced by short circuits.
Product RiS Ci is short-circuit time
constant associated with Ci.
ELE2110A 2008

Lecture 12 - 19

SCTC: Example

= 100 and VA=75V

Q-point: (1.73mA, 2.32V)

AC equivalent with finite

coupling capacitances

BJT small signal parameters:

r =

VT
25mV
=
= 1.45k
I B 1.73mA / 100

ELE2110A 2008

VA + VCE 75V + 2.32V

ro =
=
= 44.7k
IC
1.73mA
Lecture 12 - 20

Time Constant Associated with C1

To find the time constant associated
with C1:

(C2 and C3 are short-circuited and set vi = 0)

) = R + ( R r )
R = R + ( R RCE
I
B
I
B in
1S

R1s = 1000 + (7500 || 1450) = 2220

1
1
=
= 225 rad/s
R1s C1 2.22k 2 F
ELE2110A 2008

Lecture 12 - 21

Time Constant Associated with C2

To find the time constant associated
with C2:

(C1 and C3 are short-circuited and set vi = 0)

R = R + ( R RCE
) = R + (R ro ) R + R
C
2S
3
C out
3
3 C

R2 s = 100k + (4.3k || 44.7k) = 104k

1
1
=
= 96.1 rad/s
R2 s C2 104k 0.1F
ELE2110A 2008

Lecture 12 - 22

Time Constant Associated with C3

r + R
th
CC
R = R R out = R
E
E +1
3S
r + (R R )
I B
=R
E
+1

To find the time

constant associated
with C3:

1450V + 1000 || 7500

R3s = 1300k ||
101
= 22.7
1
1
=
= 4410 rad/s
R3 s C2 22.7k 10 F

(C1 and C2 are short-circuited and set vi = 0)

ELE2110A 2008

Lecture 12 - 23

Lower Cutoff Frequency

The lower cutoff frequency is:

L
and

1
= 225 + 96.1+ 4410 = 4730 rad/s

i =1 RiS Ci

L
fL =
= 753 Hz
2

In this example the time constant associated with the bypass capacitor
C3 is dominant.

ELE2110A 2008

Lecture 12 - 24

High Frequency Response

Mid-band gain

Lower cut-off frequency

z

Upper cut-off frequency

At high frequency side, drop in gain is caused by transistors

parasitic capacitors

ELE2110A 2008

Lecture 12 - 25

High Frequency Small Signal Model for BJT

Model for active mode BJT

C: diffusion capacitance of the forward-biased base-emitter junction.
C: depletion capacitance of the reverse-biased base-collector junction.
rx: the resistance of the silicon material of the base region between the
base terminal and the intrinsic base terminal B that is right under the
emitter region.

ELE2110A 2008

Lecture 12 - 26

High-frequency Small Signal Model for

MOSFET

Lov

Model for saturation mode MOSFET

2
C gs = WLCox = the capacitance between the Gate and the conducting channel.
3
C gd = Cov = WLov Cox = the overlap capacitance (very small).

ELE2110A 2008

Lecture 12 - 27

Open-Circuit Time Constant Method to

Determine fH

fH can be estimated by open-circuit time constant method:

m 1
H

RioCi
i =1

= / 2
H

where Rio is resistance at terminals of ith capacitor Ci with all other

capacitors open-circuited.

ELE2110A 2008

Lecture 12 - 28

High Frequency Analysis of C-E

Amplifier

R = R || R = 30k 10k
B 1 2
R = R R = 100k 4.3k
L 3 C

= 100 and VA=75V

Let Q-point be (1.6mA, 3V)

C = 19.9 pF

C = 0.5 pF
rx = 250

ELE2110A 2008

R
B
v =v
th i R + R
I
B

R R
R = I B = 882
th R + R
I
B
Lecture 12 - 29

High Frequency Small Signal Equivalent

Norton source transformation

v
th
is =
R + rx
th
ro = r ( R + rx )
th

ELE2110A 2008

Lecture 12 - 30

Determine Amid

v2 = g m (is r 0 ) RL
v
th
is =
R + rx
th

ro = r ( R + rx )
th

r 0
r
v2
= g m RL
= g m RL
vth
Rth + rx
Rth + rx + r
Amid =

ELE2110A 2008

o RL
Rth + rx + r

100(4120)
= 153
882 + 250 + 1560

Lecture 12 - 31

OCTC: Time Constant Associated with C

To find the time constant
associated with C:

(C is open-circuited and set is = 0)

Ro = ro = r || ( R + rx ) = 1.56k || (882 + 250) = 656

th

C = 19.9 pF
C R 0 = 1.3 10 8
ELE2110A 2008

Lecture 12 - 32

OCTC: Time Constant Associated with C

To find the time constant
associated with C:

(C is open-circuited and set is = 0)

v x = ix r 0 + iL RL = ix r 0 + (ix + g m v) RL

= ro 1+ g m R + L
Ro =

L r
ix

R o = 6561+ 0.064(4120) +

vx

v = i x r 0
4120
= 178k

656

C R 0 = 89.0 10 9
ELE2110A 2008

Lecture 12 - 33

Upper Cutoff Frequency

Upper cutoff frequency:

fH =

ELE2110A 2008

1
1
=
= 9.8106 rad/s
RoC + RoC 1.310 8 + 8910 9

H
= 1.56 MHz
2

Lecture 12 - 34