Outline

• Pole/Zero locations and time response.

Time Response • First order systems.
• Second order systems.
M. Sami Fadali • Dominant poles.
Professor of Electrical Engineering
University of Nevada

1 2

Poles, Zeros, & System Response

System Poles and Zeros
m System Poles and Zeros: poles and zeros of its
K s  z1 s  z m 
 (s  z j )
j 1
transfer function
G(s)  K
s  p1 s  pn  n
Example
 (s  p )
i 1
i

n G(s) Poles at pi zeros at: 1,2

m G(s) Zeros at zi
poles at: 1,3,4

3 4
 Example: System Response Time Response & Poles/Zeros
• The poles of the input give the forced response.
• The poles of the transfer function give the natural
response.
• Real-axis poles: Exponential decay (LHP) or
increase (RHP).
• Complex conjugate poles: oscillatory response.
ି௧ ିଶ௧ ିଷ௧
• Zeros: Affect amplitudes of the different
components and therefore affect the overall
௡௔௧௨௥௔௟ ௙௢௥௖௘ௗ
௥௘௦௣௢௡௦௘ ௥௘௦௣௢௡௦௘
shape of the response.

5 6

4.3 First Order Systems Step Response of 1st Order System

Transfer Function

Differential Equation
Unit Step Response

ି௔௧
7 8
 Time Constant  s Initial Slope
de  t /  e  t / 1
a) Time after which the decaying exponential e t/    
dt  
reaches 37% of its initial value ( e 1 = 0.37) t0 t0

• 3 time constants e 3 0.05 d 1  e  t /   e  t / 1

 
• 4 time constants e 4 0.02 dt
t0
 t0

• 5 time constants e 5 0.01
b) Time after which the rising exponential 1e t/
c) Time Constant
reaches 63% of its final value ( 1e 1 = 0.63) time at which a straight line at the
initial slope intersects the final value.

9 10

Specs. for 1st Order Systems 4.4 Second Order Systems

Rise Time: time to go from 10% to 90% of the final value. Transfer Function
 2n  n2
10%: 1 e  t1 /
 01 G( s)  
.
 e t 2  t1   
0.9
9 s2  2 n s   n2  s   n 2   n2 1   2 
 t 2 /
90%: 1 e  0.9 01.
Differential Equation c  2 n c   2n c   2n r
Tr  t2  t1   ln 9  2.2
Settling Time: time to reach and stay within a specified Unit Step Response C (s)  G (s) R(s)
percentage of the final value.  n2

௦
Definitions:
n = undamped natural frequency rad/s s 2  2 n s  n2 s
 = damping ratio 1 A s  A2
௦
d = damped natural frequency rad/s   2 1
௦
s s  2 n s   n2
11 12
Time Responses of 2nd Order Systems a) Oscillatory Response  = 0
ଶ
௡
ଶ ଶ
௡

ଶ ଶ
௡

• The undamped natural frequency is the

frequency of oscillations when  = 0

13 14

b) Overdamped Response  > 1

Example: Oscillatory n2 n2
G( s)  
s 2  2n s  n2 s  n 2  n2  2  1
 n2
C (s)  G (s) R(s) 
s   n 2   n2  2  1 s
1 K1 K2
  
s s  s1 s  s 2
s1, 2   n   n  2  1   n    2  1  
ଶ
ଵ ଶ ଵ
ଵ ଶ
௦భ ௧ ௦మ ௧
ଵ ଵ
15 16
 Ex. Overdamped System Example: Overdamped
Two real roots & >1
ଶ ଶ ଶ
௡ ௡

ଶ
௡ ௡

Overdamped: two real poles

17 18

c) Underdamped Response  < 1

Partial Fraction Expansion
n2
C ( s)  G ( s) R( s)  1 s   n  d  n d 
s     2

 d2 s C ( s)  
1 A s  A2
n
s s   n 2  d2
  2 1
s s  2 n s  n2 
1 s   n  d  1   2
 

Equate coefficients s s   n 2  d2
n2  s 2  2 n s  n2  A1s 2  A2 s
  
s 2 : 1  A1  0  A1  1 c(t )  1  e nt cos(d t )  sin( t
d )
 1  2 
s : 2 n  A2  0  A2  2 n
19 20
 Time Response
   Example: Underdamped
c(t )  1  e nt cos(d t )  sin(d t )
 1  2


cos(AB)=cos(A)cos(B)+sin(A)sin(B) e nt
c(t )  1  cos(d t   ),  sin 1 
sin(A+B)=sin(A)cos(B)+cos(A)sin(B) 1  2

c (t )  1 
e  nt
1  2
 1  2
cos( d t )   sin( d t ) 
e  nt
 1 cos( d t   ),  sin 1 
1  2

e  nt
 1 sin( d t   ),   cos 1    2  
1  2

21 22

Underdamped Poles d) Critically Damped Response

 =1
  cos 1 
 n2
C (s)  G (s)R(s) 
s   n 2 s
1 n 1
  
s s   n 2 s   n

c(t )  1  e nt nt  1, t  0

23 24
 Specifications: 2nd Order Systems
Example: Critically Damped
1.Peak Time Tp : time to first peak (  < 1)
2. Percent Overshoot % OS (  < 1)
௦௧

3.Settling time Ts : time to reach and stay within

2% of final value (5% and 1 % also used)
4. Rise Time Tr : time from 10% to 90% of final
value.

25 26

Specifications: 2nd Order Significance of 2nd Order Criteria

Step Response
• Measures of Speed of Response Tr , Tp , n
Measures of Relative Stability % OS, 
• Mixed Measures Ts
• Tr and Tp increase together
• Use Tp since it has a simpler expression.

27 28
 Time to First Peak Tp Rise Time Tr
Derivative of step response Tr  (1.76ζ 3  0.417ζ 2 + 1.039ζ + 1)/ n
1
_ c sC ( s )  s G ( s )  c  g (t )  impulse response
s
 n2  n2 -1  d 
c  _ -1  2 2
 _  2
 s  2 n s  n  d s   n   d 
2

n
 e  t sin d t   0 at max or min
n

1  2

 
d t  l , l  0,1,2,  Tp  
d n 1   2
29 30

Evaluation of % OS
Effect of Changing 
  
c(t )  1  e nt cos(d t )  sin(d t )
 1  2

 n 
    
cos    
n 1 2 
cmax  c(t ) t T    1  e sin  
p
d  1  2 
  
c 1 


1 2 
%OS  max 100%  e  100 %
1
 lnx  %OS
  x 1
 2  ln 2 x  100
31 32
 Ex. Mass-Spring-Damper
Evaluation of Ts Y ( s) 1
G( s)   2 b
• Difficult since sinusoid not always at peak F ( s ) ms  bs  k
1m
• Obtain a rough estimate based on the exponential  2
s  b m s  k m  m
decay
n2 k k
• Time constant of exponential decay = 1/n G( s)  2
s  2 n s  n2 y
 3
 5% Equate Coefficients
 n n2  k m  n  k m
 4
Ts   2%
 n 2 n  b m    b 2m  k m
 5
  n
1%   b 2 k m 
33 34

Underdamped Time Response s-plane Contours

Constant damping ratio 
  
• Effect of moving pole on s-plane contour  
 1 2 
%OS  e 100 %
on the time response. Constant imaginary part d
• Estimate time response characteristics

from pole locations. Tp 
d
• Used later for design.

4 Constant real part n

Ts 
 n

35 36
 Example
Pole Location & Time Response ଶ
௡
ଶ ଶ ଶ
Constant real part n : ௡ ௡
d changing, ܶ௦ fixed. • For any gain
௡ rad/s for any choice of gain
ସ
– Settling time: ௦ s
఍ఠ೙
Constant imaginary part d :
n changing, ܶ௣ fixed. ௥௔ௗ ଶ
• For ௡ ௦ ఠ೙
• Increasing
Constant damping ratio : (i) decreases (increases OS%)
d is changing & n changing.
(ii) increases ௡ and ௗ (decreases ௣) :
37
faster response 38

Effect of 3rd Pole on Time Response: 3rd Order System

Time Response Step Response
1.4
No Pole

1.2
• Slower response. Pole at -20
1
• Decrease overshoot.
0.8
Amplitude

• Reduced effect for pole farther in LHP. 0.6 13a

G (s) 
• Reduced effect if zero almost cancels 0.4
Pole at -10

s  4 s  13 s  a 
2

a pole. 0.2
Pole at -5
0
0 0.5 1 1.5 2 2.5
Time (sec)

39 40
 Time Response: 1st,2nd, 3rd Order Time Response: 3rd Order System
Step Response

0.8 G(s)=2/(s2+2s+2)
ି఍ఠ೙ ௧
ௗ
Amplitude

0.6

G1(s)=G(s)/(10 s+1)
0.4

0.2
G2(s)=1/(10 s+1)

0
0 10 20 30 40 50
Time (seconds)
41 42

Time Response: 2nd Order System with Zero

Effect of Zero on Time Response
G ( s) 
13 a s  a 
• Faster response. s 2  4 s  13
Step Response

• Increased overshoot.
1

• Reduced effect for zero farther in LHP.

0.8
No zero
• Reduced effect if zero almost cancels a
Amplitude

0.6
pole. Zero at -20

0.4 Zero at -10

0.2
Zero at -5

0
43 0 0.5 1 1.5 2 2.5 44
Time (sec)
Almost Canceling Pole and Zero Nonminimum-phase Systems
n2 s  a    Undershoot in the time response.
G( s) 
s 2  2n s  n2 s  a  ,   1 1.2
Step Response

1 n2 s  a    1
G(s)  2
s s s  2 n s  n2 s  a  0.8

1  a  n2  a  1 As  B 0.6
>> s=tf(‘s’)

Amplitude
   2 
2 
 2
s  a  2 n a  n  s  a s  2 n s  n
2 0.4 >> g=2.6*(-s+5)/(s^2+4*s+13)
a  2 n 2 n a  4 2n2  n2
0.2 >> step(g)
A  1  2 , B  2   0
a  2 n a  n2 a 2  2 n a  n2
n

-0.2
0 0.5 1 1.5 2 2.5 3 3.5
Time (seconds)
Reduced effect if zero
almost cancels a pole. 45 46

Dominant 2nd Order Pair Dominant Pair Example

• For systems with a 2nd order pair and with
zeros or additional poles that 1. Poles s1,2 = 5j5 s3 = 25
(i) are all located far in the LHP, or The underdamped pair is dominant since
(ii) almost cancel, exp(25 t) decays to zero quickly.
then the time response is approximately the • RULE of THUMB factor of 5 is enough.
same as that of the 2nd order pair. 2. Poles s1,2 = 5j5, s3 = 5, Zero z = 5.01
• The 2nd order pair is said to be dominant. The underdamped pair is dominant
RULE of THUMB: factor of 5 is enough. Third pole almost cancels with zero.

47 48
 Saturation
Nonlinearities & the Time Response 0.9
1

No saturation
0.8

0.7

Consider a feedback loop with one of the • Output cannot 0.6

following nonlinearities: exceed a fixed 0.5

value.
0.4
With saturation
1. Saturation 0.3

2. Dead Zone • Clipped output. 0.2

0.1

3. Backlash 0
0 0.5 1 1.5 2 2.5 3 3.5 4

Time s
2
s+2
Step Saturation Transfer Fcn

Scope

2
s+2
Transfer Fcn1
49 50

Dead Zone Backlash

0.8
No deadzone No backlash
0.7
• On reversing 1.8

1.6

• Output is zero 0.6

directions, the output is 1.4

until the input 0.5

unchanged until the 1.2

input exceeds a 1

exceeds a 0.4
0.8
threshold value.
threshold value. 0.3 0.6

0.2 • Reversing directions 0.4

• Reduced output distorts the output. 0.2

With backlash
0.1 With deadzone 0
amplitude. 0
0 1 2 3 4 5 6 7 8 9 10

0 1 2 3 4 5 6 7 8 9 10
Time s
Time s 2 1
2 1
s+2 s
s+2 s
Sine Wave Dead Zone Integrator Sine Wave Transfer Fcn2 Integrator Backlash
Transfer Fcn2
2 sin(4 t) -0.5 to 0.5 2 sin(2 t)
Scope Scope
2 1 2 1
s+2 s
s+2 s
Transfer Fcn1 Integrator1
Transfer Fcn1 Integrator1
51 52