Feedback and Stability

 Obtain the transfer function of the ideal feedback

system.
 Analyze the four ideal feedback circuit
configurations.
 series–shunt feedback amplifiers.
 shunt–series feedback amplifiers.
 series-series feedback amplifiers.
 shunt–shunt feedback amplifiers.
 Advantages of Negative Feedback

1. Gain Sensitivity – variations in gain is

reduced.
2. Bandwidth Extension – larger than that of
basic amplified.
3. Noise Sensitivity – reduce noise

4. Reduction of non-linear distortion

Disadvantages of Negative Feedback

1. Circuit Gain – reduced compared to that of

basic amplifier.
2. Stability – possibility that feedback circuit
will become unstable and oscillate at high
frequencies.
 Configuration of Feedback Amplifier

A 1
Af  Af 
1  A 
βA (Loop Gain) is usually positive, can be complex and in some cases negative!
β is usually passive and is almost completely independent of the real amplifier!
β appears to dominate the gain of the closed loop amplifier, hence we
have a stable amplifier!
 Gain Versus Frequency

A( s)
Af (s) 
1  A( s)
Ao
Let A( s) 
s
1
H
A f ( s)   Ao   1
Hence
 1  Ao  s
1
H (1  Ao)

Classwork IV: Derive the Af(s) equation

Gain Bandwidth Product

GB   Ao   H (1  Ao)  AoH
 1  Ao 
Gain Versus Frequency
Gain sensitivity: Reduction of Nonlinear
Distortion
Assume open
loop gain
changes due to
temperature or
other reasons

Closed loop gain is lower but stable

Basic Feedback Configurations
Voltage Amplifier Current Amplifier

Trans-conductance Amplifier Trans-impedance Amplifier

 Ideal Series-Shunt Feedback
Voltage amplifier

=AvVe

CL voltage gain
Av
Avf 
1   v Av
1. Sample Vo back to
input
2. FB Zin is infinite, no
loading caused
3. Vcvs
Voltage gain is now independent of transistor parameters
 Ideal Series-Shunt Feedback
Voltage amplifier

Ve  Vi  V fb
Vi  Ve  V fb  Ve   vVo  Ve   ( AvVe )
Vi
Ve 
1  vAv
Ve Vi
Ii  
Ri Ri (1  vAv)
Vi
Rif   Ri (1   v Av )
Ii
 Equivalent Circuit:
Series-Shunt Feedback Circuit
V  V fb  V   vVx  0

Vx  AvV Vx  Av (  vVx ) Vx (1   v Av )
Ix   
Ro Ro Ro

Vx Ro
Rof  
Ix 1   v Av
Summary of Feedback Amplifier
Functions
 Op-Amp Voltage Amplifier
Series-Shunt Feedback Circuit
Av
Ideal Avf 
1   v Av

Derive Vo/Vi, Ri and Ro, letting Ro=0 and Ri big:-

 R1 
Vf b  
R R 
Vo
 1 2 
Vo  AvV
AvV
Vi  V  V fb  V 
R
1 2
R1
 
 
Av
Vi  V 1  
 R 
 1 2 
 R1 
 
Vi Vi  Av 
Rif    Ri 1    Ri(1   vAv)
Ii V / Ri  R
1 2 

 R1 

 Loop Stability
A
Af 
1  A
A
Af 
1  T ( s)
T ( j )  T ( j ) 

If loop gain is unity when phase is 180 degrees, then T(jω)= -1

This leads to closed loop gain going to infinity

So.. An output exist even if input is zero

Circuit is oscillating!
 Loop Stability
A
Af 
1  A
A
Af 
1  T ( s)
T ( j )  T ( j ) 

T ( j )  1
  180
Stable

T ( j )  1
  180
Unstable
 Bode Plot:
Single-Stage Common Emitter Amplifier
A
Af 
1  T ( s)
T ( j )  T ( j )  C1 includes Cπ and Cmiller

 RC  1 
Ai  g m R   
R 
 C L 
R 1  sR 1
C

Aio f
Ai   
2   tan 
1

 f  f1 
1   
 f1 
Always
Stable!
 Bode Plot:
Two-Stage Common Emitter Work out the
Amplifier current gain of
this circuit!

Potentially
Aio  1  f  f  unstable!
Ai  2 2
   tan    tan 1  
 f  f    f1   f 2 
1    1   
 f1   f2 
 Bode Plot:
Three-Stage System

Aio
T( f ) 
 f  f  f 
1  j 1  j 1  j 
 f 1  f 2  f3 
Bode Plot: Three-Stage System

Likely
unstable at
some
frequency !

Aio  1  f 1  f  1  f 

Ai    tan    tan    tan  
2 2

2
  f1   f2   f 3 
 f  f  f 
1    1    1   
 f1   f2   f3 
 Bode Plot:
Phase and Gain Margins
Bode Plot:
Phase and
Gain
Margins
 Problem-Solving Technique:
Frequency Compensation

1. Insert dominant pole or move existing pole to

dominant pole position to stabilize circuit.
a. Assume dominant pole frequency is small
b. Determine frequency of resulting loop gain
function to achieve required phase margin
2. Set T(f) = 1 to find required dominant pole
frequency.
3. Use one of a number of techniques (e.g.
Miller compensation) to obtain required
dominant pole frequency.
 Bode Plot:
Frequency Compensation
 Bode Plot:
Miller Compensation Capacitor