Lecture -3

3. Bode Plots of Transfer Functions

Bode plots of a transfer function (system) can be obtained
using Matlab (experiment). However, to interpret the Bode
plots, we shall take a look at how basic factors of the
transfer function affect the Bode plots.

Example 3.1
2000( s  0.5)
G( s) 
s ( s  10)( s  50)
 Normalization
2( s / 0.5  1)
G( s) 
s ( s /10  1)( s / 50  1)

3-1

Basic factors and their associated corner frequencies:

2

s
 ( s / 0.5  1) (corner freq. at 0.5 rad/s)
1
Ascending
 s /10  1 (corner freq. at 10 rad/s) order of corner
frequencies
1

s / 50  1 (corner freq. at 50 rad/s)

 Bode plots
Magnitude plot is the summation of those of basic factors.
Phase is given by

 ( )  900  tan 1 ( / 0.5)  tan 1 ( /10)  tan 1 ( / 50)

3-2
  (rad/s) 0.1 0.5 1 10 20 50 100 200

 ( )(0 ) -79.4 -48.4 -33.4 -59 -86.7 -124.3 -148 -163

Using Matlab

3 3-3

2
s ( s / 0.5  1)

20log20 10 50
0.1 0.5 1
s / 50  1
0 dB 1
20log4   2 rad/sec. s /10  1

2
  0.5 rad/sec, draw which is -20 dB/dec line passing thru 0 dB at   2 rad/sec.
s
 s 
0.5    10 rad/sec,  1 contributes 20 dB/dec, resulting in 0 dB/dec.
 0.5 
1
 s 
10    50 rad/sec,   1 contributes -20 dB/dec, resulting in -20 dB/dec.
 10 
1

  50 rad/sec,   1 contributes -20 dB/dec, resulting in -40 dB/dec.

s
 50 

3-4
Observations:
The asymptotic magnitude plot is a piecewise straight line with
gradient change only at each corner frequency:

 The –20 dB/dec line at low frequency is due to the integrator.

 If the corner frequency associates with a first order zero, the
gradient will be increased by 20 dB/dec.
 If the corner frequency associates with a first order pole, the
gradient will be reduced by 20 dB/dec.
 The phase starts from 900 due to the integrator. It increases
because of the (minimum phase) zero and decreases afterwards
due to the poles.

3-5

Example 3.2 (Complex poles)

10
G(s) 
s ( s 2  0.4s  4)
Normalization:
2.5 1
G ( s) 
s s 2 / 4  2(0.1) s / 2  1
Basic factors:
 2.5 / s
1
(corner freq. n  2 ,   0.1)
 ( s / 2) 2  2(0.1) s / 2  1

3-6
 2.5
s -20 dB/dec 14dB
20log 2.5
1 2

-60dB/dec

00
2
90 0 
s

1800
1

( s / 2) 2  2(0.1) s / 2  1
3-7

Magnitude plot:
 For   2 rad/sec (Corner Freq.) , draw 2.5/s which is a
– 20dB/dec line, passing through 20log 2.5  8 dB at   1.
 For   2 rad/sec (Corner Freq.) , the pair of complex poles
contributes another –40 dB/dec, thus a line of –60dB/dec.
 At    n  2, make correction based on damping ratio   0.1
1 1
20log  20log  14
2 0.2 dB
0.2 / 2
Phase:  ( )   90 0  tan 1
1  ( / 2) 2
 0.2 1 1.8 2.0 2.2 3.0 10

 ( )(0 ) -91 -98 -134 -180 -224 -257 -268

3-8
20log 2.5
 8 dB

3-9

Example 3.3:
2500( s  10)
G( s) 
s ( s  2)( s 2  30 s  2500)
5(0.1s  1)

s (0.5s  1)[( s / 50) 2  2  0.3( s / 50)  1]
Basic factors:
 5
s
1
 0.5s  1 (corner frequency at 2 rad/s)
 0.1s  1 (corner frequency at 10 rad/s)
1
 (s / 50)2  2  0.3(s / 50)  1 (corner frequency at 50 rad/s
and   0.3 ). 1
Correction at 50 rad/s: 20log  4.4dB
2 3-10
20log(5 / 2)
 8 dB -40dB/dec
2 10 50

-20dB/dec -60dB/dec

You may try to use the Java applet on the website

http://csd.newcastle.edu.au/control//index.html
3-11
to try Bode plots if you don’t have Matlab at home.

1
-20dB/dec
0.5s  1
5 0.1s  1
-40dB/dec
10 50
2 rad/s
s 4.4dB correction
20log(5 / 2)
-20dB/dec
 8 dB
-60dB/dec
1
( s /50)2  0.6( s /50)  1
900

1800

2700

3-12
 Minimum Phase and Non-minimum Phase Systems

• A transfer function is called minimum phase if it has neither

poles nor zeros in the right-half s-plane, including j -axis
except the origin.
• A system with a minimum phase transfer function is
named as a minimum phase system.
• Consider
1  Ts 1  Ts
G1 ( s )  , G2 ( s )  T , T1  0
1  T1s 1  T1s
G1 ( s) is a minimum phase transfer function whereas G2 ( s )
is a non-minimum phase one.

3-13

1  jT 1   2T 2
| G1,2 ( j ) | 
• Note that | G1 ( j ) || G2 ( j ) | 1  jT1 1   2T12
• But their phases are very much different.

G1 ( j )  tan 1 (T )  tan 1 (T1 )

G2 ( j )   tan 1 (T )  tan 1 (T1 )

The net phase change of G1 ( j ) for input frequency between 0

and  is much smaller than that of G2 ( s ) .
• For a minimum phase system, an approximate transfer
function can be determined from only the magnitude plot since
the change of gradient uniquely determines a basic factor, e.g.
an increase of gradient by +20dB/dec implies a stable zero. 3-14
 • An example of non-minimum phase system is the all-pass
network realized with a symmetrical lattice network.

eg., s 1
eg . 1
 s  1  1
2
s 1
 s  1  1
2

-1 1

• Non-minimum phase systems are slow in responding. Hence,

when speed of response is of primary importance, we should
avoid non-minimum phase components. 3-15

Summary
• Bode plots of a transfer function can be derived by
summing bode plots of each basic factor
• Change of gradient of asymptote happens at corner
frequency:

+20 dB/dec  first order zero  s 1

1
-20 dB/dec  first order pole
 s 1
+ 40 dB/dec  second order zeros ( s  1) 2 (i.e.,   1)
or

1
-40 dB/dec second order ploes  s  2
 s  
   2    1
 n   n  
3-16
Exercises:

3.1 Sketch the Bode plots of the following transfer functions:

sa
(a ) G ( s)  ( a  b  0)
sb
1
(b) G ( s )  2
s ( s  1)
s
(c) G ( s) 
( s  1)( s  10)

3-17