Chapter 5 Frequency Response Method

1. Introduction Concept


2. Frequency Response of the typical Graphics
elements of the linear systems mode
3. Bode diagram of the open loop system
4. Nyquist-criterion

5. System analysis based on the frequency
Analysis
response
6. Frequency response of the closed loop
systems
 5.1 Introduction
Three advantages:
* Frequency response(mathematical modeling) can be obtained
directly by experimental approaches.
* easy to analyze the effects of the system with sinusoidal
voices.
* easy to analyze the stability of the systems with a delay
R
element
5.1.1 frequency response
For a RC circuit:
If : ur  A sin(t   0 ) ur C uc

We have the steady-state response:

1
j C 1
U c ( j )  U r ( j )  U r ( j )
1 jRC  1
R
j C
 5.1 Introduction
Uc( j ) 1
Make: G( j )  
Ur ( j ) jRC  1

then: U c ( j )  G( j )U r ( j )

We have: uc (t )  U cm sin(t   c )

Here: U cm  U c ( j )  G( j )  U r ( j )  c   U c ( j )   G ( j )   U r ( j )
1  tg -1 ( RC  )   0
 A
( RC  ) 2  1
U c ( j ) 1
We call: G( j )  
U r ( j ) jRC  1

Frequency Response(or frequency characteristic) of the electric

circuit.
 5.1 Introduction
Generalize above discussion, we
have:
Definition : frequency response (or characteristic) —the ratio of
the complex vector of the steady-state output versus sinusoid input
for a linear system, that is:
C ( j )
G( j ) 
R( j )
Here: R( j )  the complexvector representationof the sinusoidinput
C( j )  the complexvector representationof the output
G( j )  frequency response(or characteristic)
And we name:
C ( j )
A( )  G ( j )   magnitude response (characteri stic )
R( j )
(amplitude ratio of the steady-state output versus sinusoid input)
 ( )  G( j )  C( j )  R( j )  phase response(characteristic )
(phase difference between steady-state output and sinusoid input )
5.1.2 approaches to get the frequency characteristics
1. Experimental discrimination

Input a sinusoid signal to the control system

Measure the amplitude and phase of the steady-state output

Change frequency
Get the amplitude ratio of the output versus input
Get the phase difference between the output and input

Are the measured data enough？ N

y
Data processing
5.1.2 approaches to get the frequency characteristics
2. Deductive approach
Theorem: If the transfer function is G(s), we have:
G( j )  G( s )
s  j
Proof :
C ( s) M ( s)
assume : G ( s )  
R( s ) ( s  p1 )( s  p2 )    ( s  pn )
A
and r(t)  A sint  R( s ) 
s2   2
Where — pi is assumed to be distinct pole (i=1,2,3…n).

then C ( s )  G( s ) R( s )
M ( s) A
 
( s  p1 )( s  p2 )    ( s  pn ) ( s  j )( s  j )
In partial fraction form:
K1 K2 Kn  A1 A2 
C ( s)        
( s  p1 ) ( s  p2 ) ( s  p 3 )  ( s  j ) ( s  j ) 

M ( s)
Here: Ki   ( s  pi )  R( s )
( s  p1 )( s  p2 )    ( s  p3 ) s   pi

A
A1  G ( s )   ( s  j )
( s  j )( s  j ) s  j
A A G ( j ) j ( G ( j )90o )
 G ( j )  e
2j 2
 A G( j )  j (G( j )90o )
A2  A1  e
2
5.1.2 approaches to get the frequency characteristics
Taking the inverse Laplace transform:
n
c( t )   K i e pi t  A1e jt  A2 e  jt
i 1
n  j (t  G ( j )90o )  j ( t   G ( j )  90 o 
)
e e
  Kie pi t
 A G ( j ) 
 2 

i 1  
n
  K i e pi t  A G ( j ) sin(t  G ( j ))
i 1

For the stable system all poles (-pi) have a negative real parts,
we have the steady-state output signal:
n
lim c( t )  c s ( t )  lim [
t  t 
 K i e pi t  A G ( j ) sin(t  G ( j ))]
i 1
 A G ( j ) sin(t  G ( j ))
 5.1.2 approaches to get the frequency characteristics
the steady-state output:
cs (t )  A G( j ) sin(t  G( j ))
Compare with the sinusoid input r (t )  A sint , we have:
The amplitude ratio of the steady-state output cs(t) versus
sinusoid input r(t):
A G ( j ) C ( j )
 G(jω)   magnitude characteristic
A R( j  )
The phase difference between the steady-state output and
sinusoid input:
[t  G( j )]  t  G( j )  C ( j )  R( j )
 phase characteristic
Then we have :
C ( j )
G( j )   G( s )
R( j ) s  j
 5.1 Introduction
Examples 5.1.1

a unity feedback control system, the open-loop transfer

function: 1
G( s ) 
0.5s  1
If : r ( t )  10 sin(4t  60o )  20cos(4t 45o )
1) Determine the steady-state response c(t) of the system.
2) Determine the steady-state error e(t) of the system.
Solution:
1) Determine the steady-state response c(t) of the system.
The closed-loop transfer function is:
1
 ( s) 
C ( s)

G( s )
 0.5 s  1  1
R( s ) 1  G ( s ) 1  1 0.5 s  2
0.5 s  1
 5.1 Introduction
The frequency characteristic :
1 1
 ( j )  s  j 
0.5 s  2 0.5 j  2
The magnitude and phase The output response:
response :

So we have the steady-state response c(t) :

5
c( t )  2 sin(4t  60o  45o )  5 2cos(4t  45o  45o )
2
5
 2 sin(4t  15o )  5 2 cos 4t
2
 5.1 Introduction
2) Determine the steady-state error e(t) of the system.
The error transfer function is :
0.5 j  1
E ( s ) R( s )  C ( s ) E ( j )  R( j )
  1
C ( s) 0.5 j  2
R( s ) R( s ) R( s )  5
j 0.5  1   10
 1   ( s)  1 
1 E ( j )   R( j )  4
j 0.5  2
0.5 s  2  4  5
 20
0.5 s  1  4

0.5 s  2  j 0.5  1 
 E ( j )      R ( j ) 
The error frequency response:  j 0.5  2   4
(63.4 o  45o )  (4t  60o )  4t  78.4 o

The steady state error e(t) is: (63.4 o  45o )  (4t  45o )  4t  63.4 o

e( t )  2.5 5 sin(4t  78.4o )  5 5 cos(4t  63.4o )

 5.1 Introduction
5.1.3 Graphic expression of the frequency response
Graphic expression —— for intuition
1. Rectangular coordinates plot

  tg 1 (2 )
10 10 10
Example 5.1.2 G( s )   G( j )  
2s  1 j 2  1
1  ( 2 ) 2
G ( j ) G( j )
 G ( j )  G ( j ) 10

0 10 0o
5
0.5 7.07  45o
1 4.47  63.435o 1
0 
o
2 2.4  75.964 0. 5 1 2 3 4 5

3 1.64  80.538o
4 1.24  82.875o
5 0 .995  84.29o - 90o
5.1.3 Graphic expression of the frequency response
2. Polar plot
The polar plot is easily useful for investigating system stability.

Example 5.1.3 K K
G( s )   G( j )  G( s ) 
s(Ts  1) s  j j ( jT  1)
The magnitude and phase response:
;  ( )  G( j )  [90o  tg 1 (T)]
K
A( )  G( j ) 
 1  (T ) 2

Calculate A(ω) and  ( ) for different ω:

Im
KT  
2
Re
 0 1 1 
2T T
  1T
A( )   4 KT KT 0 4KT
5 2 5 -135o
o o 1
 ( )   90  117  135o  180o 2T
 0 -117o
5.1.3 Graphic expression of the frequency response
The shortage of the polar plot and the rectangular coordinates
plot: to synchronously investigate the cases of the lower and
higher frequency band is difficult.
Idea: How to enlarge the lower frequency band and shrink
(shorten) the higher frequency band？
3. Bode diagram(logarithmic plots)
Plot the frequency characteristic in a semilog coordinate:
Magnitude response — Y-coordinate in decibels: 20 log G( j )
X-coordinate in logarithm of ω: logω
Phase response — Y-coordinate in radian: G( j )
X-coordinate in logarithm of ω: logω
First we discuss the Bode diagram in detail with the frequency
response of the typical elements.
5.2 Frequency Response of The Typical Elements
The typical elements of the linear control systems — refer to
Chapter 2.
1. Proportional element
C ( s)
Transfer function: G( s)  K
R( s )
Frequency response:

 G( j )  K  L( )  20 log G( j )  20 log K
G( j )  K  

  ( )  G( j )  0o

Im L( ),  ( )
 ( )  0o
K Re
L( )  20 log K dB
0dB, 0o  (log  )
0.1 1 10 100
Polar plot Bode diagram
5.2 Frequency response of the typical elements
2. Integrating element
C ( s) 1
Transfer function: G ( s )  
R( s ) s
Frequency response:
 1
1  G ( j )   L( )  20 log G ( j )  20 log 
G ( j )   
j  o
  ( )   G ( j )   90
L( ),  ( )
  Im

Re L( ) :  20dB / dec

0dB, 0o  (log  )
 0 0.1 1 10 100  ( )  90o

Polar plot
Bode diagram
5.2 Frequency response of the typical elements
3. Inertial element
C ( s) 1 1
Transfer function: G( s )   G ( j ) 
R( s ) Ts  1 jT  1
0   1 T
1 
G( j )   L( )  20 log 1  (T ) 2    3dB  1 T
1  (T ) 2  20 log(T )   1 T
 
 ( )   tg (T )
1
K 1
1/T: break frequency G ( s )  :
T2 s  1 T2
L( ),  ( )
1
Im  0 20 log K
  0dB, 0o
T
 (log  )
1 Re 0.1 1 10 100
 45 o 20dB/ dec

 90 o

Polar plot Bode diagram

5.2 Frequency response of the typical elements
4. Oscillating element
C ( s) 1
Transfer function: G( s )   0 1
R( s ) T 2 s 2  2Ts  1
1
G( j ) 
(1   2T 2 )  j 2T
2T
G ( j ) 
1
 ( )  tg 1 ( )
1  ( T )  (2T )
2 2 2 2 1   2T 2
0    n (  1 T )

L( )  20 log (1   2T 2 ) 2  ( 2T ) 2    20 log( 2 )   n
 40 log(T )   
 n
maximum value of G( j ) ：
d 2 2
Make: ( G( j ) )  0     r   n 1  2 (0    )
d 2
 r  resonant f requency  M  G( j )  1
r r
M r  resonant p eak 2 1   2
5.2 Frequency response of the typical elements
The polar plot and the Bode diagram:  2  1

Im L( ),  ( ) 20 log M r
   0 n  1 / T 20 log( 1 )
Re 2
1 r

0dB, 0o  (log  )
0.1 1 10 100

1  90o
j (   n )
2 40dB/ dec

Polar plot  180o

Bode diagram
1.    r  (  n )  M r 
   0   r   n  unstable system
2. 
  2    0 No resonan ce, Optimal Second - order System
r
2
5.2 Frequency response of the typical elements
5. Differentiating element
Transfer function:
 s differential

G( s)   Ts  1 first  order differential
Ts 2  2Ts  1 second  order differential

Im Im Im

Re Re Re
1 1

differential 1th-order differential 2th-order differential

Polar plot
5.2 Frequency response of the typical elements
Because of the transfer functions of the differentiating elements
are the reciprocal of the transfer functions of Integrating
element, Inertial element and Oscillating element respectively,
that is: inverse
s  1
s
inverse
Ts  1  1
Ts  1
inverse
T 2 s 2  2Ts  1  1
T 2 s 2  2Ts  1

the Bode curves of the differentiating elements are symmetrical

to the logω-axis with the Bode curves of the Integrating
element, Inertial element and Oscillating element respectively.
Then we have the Bode diagram of the differentiating elements:
5.2 Frequency response of the typical elements

L( ),  ( ) L( ),  ( )

 180 o
L( ) :  20dB / dec  40dB / dec
 ( )  90o
 (log  )  90 o
0dB, 0o n  1 T
0.1 1 10 100
0dB, 0o  (log  )
differential 0.1 1 10 100

L( ),  ( )
20 log( 1 )
o 2
 90
o
 20dB / dec 20 log M r
 45
0dB, 0o  (log  ) 2th-order differential
0.1 1 10 100
1th-order differential
5.2 Frequency response of the typical elements
6. Delay element
 e s
C ( s)
Transfer function: G ( s ) 
R( s )

 j  G ( j )  1  L( )  0
G ( j )  e 
 ( )  G ( j )  
Im L( ),  ( )
R=1
0dB, 0o  (log  )
Re 0.1 1 10 100

Polar plot Bode diagram

5.3 Bode diagram of the open loop systems
5.3.1 Plotting methods of the Bode diagram of the open loop
systems
Assume: G ( s )  G1 ( s )  G2 ( s )  G3 ( s )...
here : Gi ( s )  the transf er functio n of the t ypical elements
We have:
 ( )  G( j )  G1 ( j )  G2 ( j )  G3 ( j )  ...

L( )  20 log G  2o log G1  20 log G2  20 log G3  ...

That is, Bode diagram of a open loop system is the superposition
of the Bode diagrams of the typical elements.

10( s  1)
Example 5.3.1 G( s ) H ( s ) 
s 2 (0.01s  1)
5.3 Bode diagram of the open loop systems
G(s)H(s) could be regarded as:
① ②③ ④
10( s  1) 1 1
G ( s) H ( s)   10  (s  1)  
2
s (0.01s  1) s2 0.01s  1
Then we have:
L( ),  ( ) 20dB/dec
②
－40dB/dec
40dB, 90o
①
20dB, 45o
－20dB/dec
0dB, 0o  (log  )
0.1 1 10 100 －20dB/dec
-20dB, -45o ④
-40dB/dec
-40dB, -90o
－40dB/dec
-60dB.-135o
③
-80dB,-180o
5.3.2 Facility method to plot the magnitude response
of the Bode diagram
Summarizing example 5.3.1, we have the facility method to plot
the magnitude response of the Bode diagram:
1) Mark all break frequencies in theω-axis of the Bode diagram.
2) Determine the slope of the L(ω) of the lowest frequency band
(before the first break frequency) according to the number of the
integrating elements:
－20dB/dec for 1 integrating element
－40dB/dec for 2 integrating elements …
3) Continue the L(ω) of the lowest frequency band until to the
first break frequency, afterwards change the the slope of the L(ω)
which should be increased 20dB/dec for the break frequency of
the 1th-order differentiating element .
The slope of the L(ω) should be decreased 20dB/dec for the
break frequency of the Inertial element …
5.3.2 Facility method to plot the magnitude response of
the Bode diagram
Plot the L(ω) of the rest break frequencies by analogy .
Example 5.3.2 10( s  1)
G ( s) 
s(0.1s  1)( 0.012 s 2  0.01s  1)
 20log10  20logω (  1)  ( )  90o  tg 1  tg 1 (0.1 )
 20log10  20log
  0.01
 (1    10)  tg 1
  20log 
 1  ( 0.01 ) 2

L( )   20log10  20log     51.3o
 (10    100)  1
20log  20log(0.1 ) 
    56.5o   10
 20log10  20log    ( )  

 20log  20log(0.1 ) (100   )   174 . 9 o
  100
   179.6o   104
 40log(0.01 )  

The Bode diagram is shown in following figure:

5.3.2 Facility method to plot the magnitude response of
the Bode diagram
L( ),  ( )
G ( s) 
－20dB/dec
10( s  1) 40dB, 90o －20dB/dec
2 2 1.25dB
s(0.1s  1)( 0.01 s  0.01s  1) 20dB, 45o
0dB, 0o  (log  )
0.1 1 10 100
-20dB, -45o
－60dB/dec
-40dB, -90o r
-60dB.-135o
-80dB,-180o
-100dB,-225o
-120dB,-270o

There is a resonant peak Mr at:    r   n 1  2 2

1
Mr   1.154  1.25 dB  100 1  2  0.52  70.7
2 1   2
5.3.3 Determine the transfer function in terms of the
Bode diagram
1. The minimum phase system(or transfer function)
Compare following transfer functions:
K (s  1) K (s  1) We have:


G1 ( s )  G2 ( s ) 
(Ts  1) (Ts  1) G1 ( )  G2 ( )  G3 ( )  G4 ( )
T 
K ( s  1) K (s  1)
G3 ( s)  G4(s)  ( ) 2  1
(Ts  1) (Ts  1) K
(T ) 2  1
The magnitude responses are the same.
But the net phase shifts are different when ω vary from zero to
infinite. It can be illustrated as following:
Sketch the polar plot:

G1 ( j )   tg 1 (T )  tg 1 ( ), G2 ( j )   tg 1 (T )  180o  tg 1 ( ) 
 
G3 ( j )   tg 1 (T )  tg 1 ( ), G4 ( j )   180o  tg 1 (T )  tg 1 ( )
 5.3.3 Determine the transfer function in terms of the
Bode diagram
The polar plot:
Im
 
  , K
Im T
 0
  ,  K Im   0, K
K T Re
Re Re

  0,  K 
  , K
T

K (s  1) phase K (s  1) phase G3 ( s) 

K ( s  1) phase
G1 ( s)  G2 ( s) 
Ts  1 shift 0
0
Ts  1 shift －π Ts  1 shift －π
Im
  0,  K It is obvious: the net phase shifts of the
Re G1(jω) is the minimum when ω vary from
zero to infinite.

  , K G1(s) is named: the minimum phase
T
K (s  1) phase transfer function .
G4 ( s ) 
Ts  1 shift π
5.3.3 Determine the transfer function of the minimum phase
systems in terms of the magnitude response
Definition:
A transfer function is called a minimum phase transfer func-
tion if its zeros and poles all lie in the left-hand s-plane.
A transfer function is called a non-minimum phase transfer
function if it has any zero or pole lie in the right-hand s-plane.
Only for the minimum phase systems we can affirmatively deter-
mine the relevant transfer function from the magnitude response of
the Bode diagram .

2. Determine the transfer function from the magnitude response

of the Bode diagram .

Example 5.3.3
5.3.3 Determine the transfer function in terms of the
Bode diagram
we can get the G(s) from L( )
－ 40dB/dec

the Bode diagram : －20dB/dec

 (log  )
0dB, 0o
K ( 0.5s  1)
2 20 200
0.1 1 10 100
G ( s)  40dB/dec
2
s ( 0.005s  1)
－

and :
L( )  2o log K  20 log  2  20 log( 0.5 )  0  K  40
  20
Example 5.3.4 L( )
we can get the G(s) from the 20dB
20dB/dec －20dB/dec
0.5 200
Bode diagram :  (log  )
0dB 10 100
0.1 1
Ks
G ( s) 
(T1s  1)(T2 s  1)
5.3.3 Determine the transfer function in terms of the
Bode diagram
L( )
we can get the G(s) from the 20dB
20dB/dec －20dB/dec
Bode diagram : 0.5 200
 (log  )
0dB 10 100
Ks 0.1 1
G ( s) 
and : (T1s  1)(T2 s  1)
L( )  20 log K  20 log   0.5  0  K 2
L( ω)  20 log 2  20 log   1 / T  20 dB  T1  0.2
1
L( )  20 log 2  20 log   20 log( 0.2 )  20 log(T2 )   200  0  T2  0.05
Example 5.3.5 L( )
－20dB/dec
we can get the G(s) from
8.136 dB
the Bode diagram : 20 dB
 (log  )
2 0dB
K ( 0.01s  1) 0.1 1 10 100
G ( s)  －60dB/dec

s(T 2 s  2Ts  1) －20dB/dec

5.3.3 Determine the transfer function in terms of the
Bode diagram
L( )
－20dB/dec
we can get the G(s) from
8.136 dB
the Bode diagram : 20 dB
0dB  (log  )
K ( 0.01s  1)2 0.1 1 10 100
G ( s)  －60dB/dec

s(T s  2Ts  1)
2 －20dB/dec

1
 10  T  0.1
T
L( )  20 log K  20 log   10  20 dB  K  100 then :
1 100(0.01s  1) 2
20 log  8.136    0.2 G(s) 
2ζ 1   2 s(0.01s  0.04s  1)

For the non-minimum phase system we must combine the

magnitude response and phase response together to determine
the transfer function.
5.3.3 Determine the transfer function in terms of the
Bode diagram
L( ),  ( )
Example 5.3.6
－20dB/dec
10(0.1s  1) 10(0.1s  1) 0dB, 0o  (log  )
G1  G2  0.1 1 10 100
( s  1) ( s  1) －90o
－180o
10( 0,1s  1) 10(0.1s  1)
G3  G4 ( s ) 
( s  1) ( s  1)
All satisfy the magnitude response
10(0.1s  1)
But only G4 ( s ) 
( s  1)
satisfy the phase response sim ultaneously.
10(0.1s  1)
So, we have : G( s) 
(s - 1)
5.4 The Nyquist-criterion
A method to investigate the stability of a system in terms of the
open-loop frequency response.
5.4.1 The argument principle(Cauchy’s theorem)
Assume: G ( s ) H ( s )  K 1 ( s  z1 )( s  z 2 )...( s  z m ) nm
( s  p1 )( s  p2 )...( s  pn )
here : z i  open-loop zeros; p j  open-loop poles .
Make : K 1 ( s  z1 )( s  z 2 )...( s  z m )
F ( s)  1  G( s) H ( s)  1 
( s  p1 )( s  p2 )...( s  pn )
( s  p1 )( s  p2 )...( s  pn )  K 1 ( s  z1 )( s  z 2 )...( s  z m )

( s  p1 )( s  p2 )...( s  pn )
K F ( s  s1 )( s  s 2 )...( s  sn )
 si  zeros of the F ( s )
( s  p1 )( s  p2 )...( s  pn )
Note: si→ the zeros of the F(s), also the roots of the 1+G(s)H(s)=0
5.4.1 The argument principle
Now we consider the net phase shift F (s) if s travels 360o
along a closed path Γ of the s-plane in the clockwise direction
shown in Fig.5.4.1 .

 ( s  si )  ( s  p j )

n n
Because : F ( s ) 
i 1 j 1 S-plane
If the zeros si is enclosed by Γ , then : Im Γ
( s  si )  2 s  si
s
si
If the zeros si is not enclosed by Γ , then : Re

( s  si )  0 si
Similarly we have:
p j is enclosed by Γ : ( s  p j )  2 Fig. 5.4.1

p j is not enc losed by Γ : ( s  p j )  0

5.4.1 The argument principle
If Z zeros and P poles are enclosed by Γ , then:

 ( s  si )  ( s  p j )

n n
F ( s ) 
i 1 j 1
 Z  ( 2 )  P  ( 2 )  ( P  Z )  2

It is obvious that path Γ can not pass through any zeros si or

poles pj . Then we have the argument principle:
If a closed path Γ in the s-plane encircles Z zeros and P poles of
F(s) and does not pass through any poles or zeros of F(s) , when s
travels along the contour Γ in the clockwise direction, the corres-
ponding F(s) locus mapped in the F(s)-plane will encircle the
origin of the F(s) plane N = P-Z times in the counterclockwise
direction, that is:
N=P－Z
5.4.1 The argument principle
here: N —— number of the F(s) locus encircling the origin of the
F(s)-plane in the counterclockwise direction.
P —— number of the zeros of the F(s) encircled by the path
Γ in the s-plane.
Z —— number of the poles of the F(s) encircled by the path
Γ in the s-plane.
5.4.2 Nyquist criterion

Im
If we choose the closed path Γ so S-plane
that the Γ encircles the entire right
hand of the s-plane but not pass Re
through any zeros or poles of F(s)
shown in Fig.5.4.2 .
The path Γ is called the Nyquist-path. Fig. 5.4.2
5.4.2 Nyquist criterion Im
 S-plane
When s travels along the the Nyquist-path:
F ( s )  1  G( s ) H ( s )  1  G( s ) H ( s )  0
s  j Re

F ( s)  1  G( j ) H ( j )  G( j ) H ( j )  1  F ( s )
s   j
Because the origin of the F(s)-plane is Fig. 5.4.2
equivalent to the point (-1, j0) of the G(jω)H(jω)-plane, we have
another statement of the argument principle:
When ω vary from - (or 0) →+  , G(jω)H(jω) Locus mapped
in the G(jω)H(jω)-plane will encircle the point (-1, j0) in the
counterclockwise direction:
N  P  Z [or N  (P  Z)/ 2 for  from 0   ]
here: P — the number of the poles of G(s)H(s) in the right hand
of the s-plane.
Z — the number of the zeros of F(s) in the right hand of the
s-plane.
5.4.2 Nyquist-criterion
If the systems are stable, should be Z = 0, then we have:
The sufficient and necessary condition of the stability of the
linear systems is : When ω vary from - (or 0) →+  , the
G(jω)H(jω) Locus mapped in the G(jω)H(jω)-plane will encircle
the point (-1, j0) as P (or P/2) times in the counterclockwise
direction. ——Nyquist criterion
Here: P — the number of the poles of G(s)H(s) in the right hand
of the s-plane.
Discussion :
i) If the open loop systems are stable, that is P = 0, then:
for the stable open-loop systems, The sufficient and necessary
condition of the stability of the closed-loop systems is :
When ω vary from - (or 0) →+  , the G(jω)H(jω) locus
mapped in the G(jω)H(jω)-plane will not encircle the point (-
1, j0).
5.4.2 Nyquist-criterion
ii) Because that the G(jω)H(jω) locus encircles the point (-1, j0)
means that the G(jω)H(jω) locus traverse the left real axis of
the point (-1, j0) , we make:
G(jω)H(jω) Locus traverses the left real axis of the point (-1, j0)
in the counterclockwise direction —―positive traversing‖.
G(jω)H(jω) Locus traverses the left real axis of the point (-1, j0)
in the clockwise direction —―negative traversing‖.
Then we have another statement of the Nyquist criterion:
The sufficient and necessary condition of the stability of the
linear systems is : When ω vary from - (or 0) →+  , the
number of the net ―positive traversing‖ is P (or P/2).
Here: the net “positive traversing” —— the difference between the
number of the “positive traversing” and the number of the “negative
traversing” .
5.4.2 Nyquist-criterion
Example 5.4.1
The polar plots of the open loop systems are shown in
Fig.5.4.3, determine whether the systems are stable.
Im Im
stable stable
(-1, j0)
(-1, j0)
  Re   Re
 0
(1) P=2 (2) P=0
 0  0
Im Im
unstable unstable
(-1, j0)
  Re (-1, j0)
  Re
 0
(3) P=2
Fig.5.4.3 (4) P=0
5.4.2 Nyquist-criterion
Note: the system with the poles (or zeros) at the imaginary axis
10
Example 5.4.2 G ( s ) H ( s ) 
s( s  1)( 0.5s  1)
There is a pole s = 0 at the origin in this system, but the Nyquist
path can not pass through any poles of G(s)H(s).
Idea: We consider a semicircular Radius
detour around the pole (s = 0) repre- s  j Im
r 
sented by setting s  e j (  0)   0  0
at the s = 0 point we have: Re
  j 90 o   1 1 j 90 o －2 －1
  0  s  εe  G( j0 ) H ( j0 )   e Radius
o  
εe  j 90  0  0
j 0o 1 1 j 0o
  0  s  εe  G ( j 0) H ( j 0)   e s   j s  j
o 
εe j 0
Fig. 5.4.4
o 1 1 o
  0  s  εe j 90  G ( j 0 ) H ( j 0 )   e  j 90
j 90 o 
εe
5.4.2 Nyquist-criterion
It is obvious that there is a phase saltation of the G(jω)H(jω) at
ω=0, and the magnitude of the G(jω)H(jω) is infinite at ω=0.
Radius
s  j Im   0
r  Im

  0  0
  
Re G ( s) H ( s) 
Re
Radius
10 (-1, j0)
  
  0  0 s( s  1)( 0.5s  1)  0

s   j s  j   0
Fig. 5.4.4 Fig.5.4.5
In terms of above discussion , we can plot the system’s polar
plot shown as Fig.5.4.5.
The closed loop system is unstable.
Example 5.4.3
10 10
G ( s) H ( s)  
s( s  1)( s  4) s( s  1)( s  j 2)( s  j 2)
2

Determine the stability of the system applying Nyquist criterion.

Solution Im

Similar to the Example   j2

5.4.2, the system’s polar plot
Re
is shown as Fig.5.4.6 .
(-1, j0)
The closed loop system     0
is unstable.   j2
  0
5.4.3 Application of the Fig.5.4.6

Nyquist criterion in the Bode diagram

5.4.3 Application of the Nyquist criterion in the Bode diagram
G(jω)H(jω) locus traverses the left real axis of the point (-1, j0)
in G(jω)H(jω)-plane → L(ω)≥0dB and υ(ω) =－180o in Bode
diagram (as that mentioned in 5.4.2).
We have the Nyquist criterion in the Bode diagram :
The sufficient and necessary condition of the stability of the
linear closed loop systems is : When ω vary from 0→+  , the
number of the net ―positive traversing‖ is P/2.
Here: the net ―positive traversing‖ —the difference between
the number of the ―positive traversing‖ and the number of the
―negative traversing‖ in all L(ω)≥0dB ranges of the open-loop
system’s Bode diagram.
―positive traversing‖ — φ(ω) traverses the ―-180o line‖ from
below to above in the open-loop system’s Bode diagram;
―negative traversing‖ — φ(ω) traverses the ―-180o line‖ from
above to below.
5.4.3 Application of the Nyquist criterion in the Bode diagram
Example 5.4.4
The Bode diagram of a open-loop stable system is shown in
Fig.5.4.7, determine whether the closed loop system is stable.
Solution
Because the open-loop system is L( )
－20
－40
stable, P = 0 . －90o
－60
In terms of the Nyquist
－40 φ(ω)
criterion in the Bode diagram: －20
0dB, －180o  (log  )
The number of the net
－40
―positive traversing‖ is
－60
0 ( = P/2 = 0 ). －270o
L(ω)
Fig.5.4.7
The closed loop system is stable .
5.4.4 Nyquist criterion and the relative stability
(Relative stability of the control systems)
In frequency domain, the relative stability could be described
by the ―gain margin‖ and the ―phase margin‖.
1. Gain margin Kg
1
Kg  K g ( dB )   20 log G ( j ) H ( j )
G ( j ) H ( j )   g
  g

g : G ( j ) H ( j )    1800  Phase-cros sover freq uency

g
2. Phase margin γc
 c  G ( j ) H ( j )    ( 1800 )  G ( jc ) H ( jc )  1800
c
ωc : G ( j ) H ( j )  1  Gain-crossover frequ ency
 c
3. Geometrical and physical meanings of the Kg and γc
5.4.4 Nyquist criterion and the relative stability
The geometrical meanings is shown in Fig. 5.4.8.
Im
The physical signification : 1/Kg

Kg— amount of the open-loop gain in

decibels that can be allowed to increase
－1 Re
before the closed-loop system reaches to
be unstable. γc
For the minimum phase system:
Kg>1
the closed loop system is stable . stable
unstable
γc —amount of the phase shift Critical
Fig. 5.4.8
of G(jω)H(jω) to be allowed before stability

the closed-loop system reaches to be unstable.

For the minimum phase system: γc>0 the closed loop system is
stable .
5.4.4 Nyquist criterion and the relative stability
Attention :
For the linear systems:
The changes of the open-loop gain only alter the magnitude of
G(jω)H(jω).
The changes of the time constants of G(s)H(s) only alter the
phase angle of G(jω)H(jω).
Example 5.4.5
The open loop transfer function of a control system is:
e s
K
G ( s) H ( s) 
s(0.1s  1)
(1) Determine Kg and γc when K =1 and τ =1.
(2) Determine the maximum K and τ based on K = 1 and τ = 1.
5.4.4 Nyquist criterion and the relative stability
e s
K
Solution G ( s) H ( s) 
s(0.1s  1)
(1) Determine Kg and γc ( K =1 and τ =1)
In terms of: G ( j ) H ( j )    1800 G( j ) H ( j ) 1
g  c

  tg 10.1 g   g     g  1.43;  c 1  (0.1 c )  K  c  1
2
2  1 K 1
1  1  (0.1 )2
Kg    1.44  3.17 (dB)
G ( j g ) H ( j g ) K  g 1.43
K 1

 c  G ( jc ) H ( jc )  1800   900  tg 10.1c  c  c 1  270

 1
(2) Because K g  1.44  maximum K  1.44
K 1
Because  c  1  270  c  270  maximum   0.47
 5.4.4 Nyquist criterion and the relative stability
Example 5.4.6
The G(jω)H(jω) polar plot of a system is shown in Fig.5.4.9.
(1) Determine Kg
(2) Determine the stable range of the open loop gain.
Solution 2 Im
(1) Determine Kg 1.5
(- Re
1
Kg   1.25  1.94( dB) 1, j0)
0.8 0.8
G(jω)H(jω)
(2) Determine the stable range of
the open loop gain.
Fig.5.4.9
Assume : G( jω) H ( jω)  K0GH ( jω)
and the critical stable value of the open loop gain is Kc .
In terms o f :
2 Im
K 0GH ( j )  0.8 and
  g 1.5

5 (- Re
K c1GH ( j )  1  K c1  K 0 1, j0)
  g 4 0.8
K 0GH ( j )  1.5 and G(jω)H(jω)
  g
2
K c 2GH ( j )  1  Kc 2  Fig.5.4.9
  g 3
K 0GH ( j )  2 and
  g
1
K c 3GH ( j )  1  K c 3  K0
  g 2
Then we have the stable range of the open loop gain K :
2 5 1
K 0  K  K 0 or K  K0
3 4 2
 5.5 System analysis based on the frequency response
5.5.1 Performance specifications in the frequency domain
1. For the closed loop systems
The general frequency response of a closed loop systems is
shown in Fig. 5.5.1
(1) Resonance frequency ωr:
A(ω)
G ( j )
Assume : A( )   ( j )  Mr
1  G ( j ) H ( j )
d A(0)
 r satisfy : A( ) 0 0.707A(0)
d   r
ω
(2) Resonance peak Mr : 0
ωr ωb

M r  A( ) Fig. 5.5.1

  r
(3) Bandwidth ωb:
2
 b satisfy : A( )
 b
 A(0)
2
2. For the open loop systems
(1) Gain crossover frequency ωc:
c satisfy : G ( jω) H ( jω) 1
ωωc
For the unity feedback systems, ωc≈ ωb , because:
G ( j ) 1 G(j  )  1
 ( j )  
1  G ( j ) G ( j ) G(j  )  1
(2) Gain margin Kg:
1
Kg  ; K g ( dB)   20 log G ( j ) H ( j )
G ( j ) H ( j )   g
  g

Here g satisfies : G ( j ) H ( j )    1800

g
(3) Phase margin γc:  c  G ( j ) H ( j )    ( 1800 )
c
Here c satisfies : G ( j ) H ( j ) 1
 c
 5.5 System analysis based on the frequency response
Generally Kg and γc could be concerned with the resonance
peak Mr : Kg and γc ↑ —— Mr ↓.
ωc could be concerned with the resonance frequency ωr and
bandwidth ωb : ωc↑ —— ωr and ωb↓.
5.5.2 Relationship of the performance specifications between the
frequency and time domain
The relationship between the frequency response and the time
response of a system can be expressed by following formula:
1 
C ( t )  L C ( j ) 
1 jt
 C ( j )e d
2  
G ( j )
here : C ( j )   ( j ) R( j )  R( j )
1  G ( j ) H ( j )
But it is difficult to apply the formula .
5.5.2 Relationship of the performance specifications between the
frequency and time domain
(1) Bandwidth ωb(or Resonance frequency ωr)  Rise time tr
Generally ωb(or ωr )↑—— tr ↓ because of the “time scale”
theorem:
1 
In terms of c ( t )  L C ( j ) 
1
 C ( j ) e jt
d
2 
If :    , C ( j )  C ( j )
1  j t /   1
d    c ( t /  )
1
Then : c ( t )   C ( j )e
2    
That is : βω  t So ωb(or ωr )↑—— tr ↓
β
alike : ωc↑—— tr ↓ because of ωc≈ ωb .
For the large ωb , there are more high-frequency portions in
c(t), which make the time response to be faster.
5.5.2 Relationship of the performance specifications between the
frequency and time domain
(2) Resonance peak Mr  overshoot σp%
Normally Mr ↑ —σp% ↑ because of the large unbalance of the
frequency signals passing to c(t) .
Kg and γc ↓ —σp% ↑is alike because of Kg and γc ↓—Mr ↑.

Some experiential formulas:

Overshoot  p %  0.16  0.4( M r  1)  (1.1  M r  1.8)
1
and Mr 
sin  c
For mostdesignproblem, an optimumvalueof Mr : 1.1  Mr  1.5

Settling time ts 
k
c
 1
, k  2  1.5
  1
 1  2.5


 1  350   c  900 
 sin  c   sin  c 
5.5.2 Relationship of the performance specifications between the
frequency and time domain
(3) A(0) → Steady state error ess
G ( j )
A( 0 )  A( )  0 
1  G ( j ) H ( j )  0
1
assume : G ( s)  KG   G0 ( s) H ( s)  K H  H0 ( s)
v
s
 1 For the unity feedback system, H(s)  1 :
 K v 1
then : A(0)   H  1 v 1

1 
v0 A( 0)   KG v0
 KG K H 1  KG

So for the unity feedback systems: A(0)  1, A(0)   ess 

5.5.2 Relationship of the performance specifications between
the frequency and time domain
(4) Reproductive bandwidth ωM → accuracy of Reproducing r(t)
A(ω)
Reproductive bandwidth ωM : Mr

G ( j ) △
A( )   A( 0)   A(0)
1  G ( j ) H ( j )   0.707A(0)
M
ω
 : allowed reproducin g error 0
ωM ωr ωb
Fig. 5.5.2

for a given ωM , △↓—higher accuracy of reproducing r(t) .

for a given △, ωM ↑ —higher accuracy of reproducing r(t) .
Demonstration
assume : E ( j )  R( j )  C ( j )   e ( j ) R( j )
and :  e ( j )  
5.5.2 Relationship of the performance specifications between
the frequency and time domain
For the frequency spectrum of r(t) shown in Fig.5.5.3 .
1  M jt
e(t )    e ( j ) R ( j ) e d R( j )
2 
M
  M jt
  R( j )e d    r ( t ) ω
2  0
ωM
M
That is : e ( t )    r ( t )    e( t )  Fig. 5.5.3

5.5.3 Relationship of the performance specifications between the

frequency and the time domain: for the typical 2th-order system
For the typical 2th-order system:
 n2  n2
G( s)    ( s) 
s( s  2 n ) s 2  2 n s   n 2
5.5.3 Relationship of the performance specifications between the
frequency and the time domain: for the typical 2th-order system
We have:

 b   n (1  2 2 )  2  4 2  4 4

2 
 r   n 1  2 (0   
2
) 
2    ,  n   p % , t s , t r ...
1 
Mr  
2 1   2

c  n 1  4 4  2 2

2 
 c  tg 1 
   ,  n   p % , t s , t r ...
1  4 4  2 2 
Kg   
5.5 System analysis based on the frequency response
5.5.4 ―three frequency band‖ theorem
The performance analysis of the closed loop systems according
to the open loop frequency response.
1. For the low frequency band
the low frequency band is mainly concerned with the control
accuracy of the systems.
The more negative the slope of L(ω) is , the higher the control
accuracy of the systems. The bigger the magnitude of L(ω) is, the
smaller the steady state error ess is.
2. For the middle frequency band
The middle frequency band is mainly concerned with the
transient performance of the systems.
ωc↑—tr ↓; Kg and γc ↓—σp% ↑
 5.5.4 ―three frequency band‖ theorem

The slope of L(ω) in the middle frequency band should be the

–20dB/dec and with a certain width .
3. For the high frequency band
The high frequency band is mainly concerned with the ability of
the systems restraining the high frequency noise.
The smaller the magnitude of L(ω) is, the stronger the ability of
the systems restraining the high frequency noise is.
L( )  20 log G( j ) H ( j )
Example 5.5.1
－40
Compare the performances
between the system Ⅰand 0dB －20 ω
system Ⅱ
－40

Fig. 5.5.4
Ⅰ Ⅱ
 5.5 System analysis based on the frequency response
Solution : L( )  20 log G( j ) H ( j )
essⅠ> essⅡ
－40
σpⅠ% =σpⅡ%
trⅠ > trⅡ －20 ω
0dB
The ability of the system Ⅰ
－40
restraining the high frequency Fig. 5.5.4
noise is stronger than system Ⅱ Ⅰ Ⅱ
L(ω)
Example 5.5.2 －20dB/dec
For the minimum phase system, the open loop
magnitude response shown as the Fig. 5.5.5. 1 ω
Determine the system’s parameter to make the 0.1 －40dB/dec
system being the optimal second-order system Fig. 5.5.5
and the steady-state error ess< 0.1.
Solution :
 5.6 Frequency response of the closed loop systems
5.6 Frequency response of the closed loop systems
How to obtain the closed loop frequency response in terms of
the open loop frequency response.
5.6.1 The constant M circles: How to obtain the magnitude
frequency response of the closed loop systems in terms of the
open loop frequency response…… (refer to P495)
5.6.2 The constant N circles: How to obtain the phase frequency
characteristic of the closed loop systems in terms of the open
loop frequency response…… (refer to P496)
5.6.3 The Nichols chart: How to obtain the closed loop frequency
response in terms of the open loop frequency response……
(refer to P496)
Chapter 5 Frequency Response Methods