Chapter 5

Time Response
Analysis
Test signals
• Impulse
• Step
• Ramp
Test signals
• Sinusoidal.
(frequency response)
 Laplace Transforms

Impulse  (t) 1
step u-1 (t) 1/s
Ramp t u-1 (t) 1/s 2
Order of a

system is ???
Order of a

system is order
of the differential
equation
zeroth order system :

C(s) / R(s) = K
a constant
(algebraic equation)
Examples of zeroth
order systems :
Potentiometer
dc amplifier
dc tachogenerator
First order systems:

C(s) / R(s) = K / (s + a)
 Open loop
1000 / ( 100 + 200 s )
= 10 / ( 1 + 2 s )
Closed loop
1000 / ( 200 + 200 s )
= 5/(1+s)
open loop
vt (s) = (vr (s) )(10 / ( 1+2 s ))
= (25 / s )( 10 / (1 + 2 s)
Vt () = 250 V

closed loop
vt(s) = ( 25 / s)( 5 / ( 1+s ))
Vt () = 125 V.
250

200

150

100

x1
50 x2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

In general K ( 1- e - t / )

K : steady state value of the

function
 : time constant
 smaller, faster response.
Ramp response
of first order
systems
Ramp Response

r (t) = t u-1 (t)

R (s) = 1/ s2
C(s) K

R(s) 1  st
Let K = 1
R C
G
1
G(s) 
1  s
1
R( s)  2
s
 1
c(s) 
s (1  s )
2

1  
 2   
1

s  s 1 s 
 -t/
c(t) = t -  (1  e )
9
8
7
6
5
4
3
2
1
0
0 0. 1 1. 2 2. 3 3. 4 4. 5 5. 6 6. 7 7. 8 9 9
5 5 5 5 5 5 5 5
 
c

t
c 

t
Second order
systems
C(s) n 2
 2
R ( s ) s  2 n s + n 2
For unit step input

1
R (s ) =
s
Applying final value
theorem
C(t) t   =1.0
 Second order
example
- position control.
r + + KT 1 L
Kp KA
Ra s(Js+B)
- -
sKb
Taking
KP = 0.1 V / rad
KA = 10 V / V
KT = Kb = 0.5 Nm/A
or
V/rad/sec.
J = 1 Kg m2

B = 0.5 Nm / rad / sec

 L (s) 1
 2
 r (s) s  s + 1
 n = 1 rad / sec.
 = 0. 5
s 2 +s+1 =
s + 0.5   
2 2
3
2
  j 3
2

- 0.5
3
 -j 2
In general
s2 + 2  n s + n 2

2

 s + n   n 1   2

2
 s1,2   n 
j n 1   2
= -  n  j d
d = n 1 -  2

d damped frequency
  damping factor
 n  undamped
natural frequency
For 0 <  < 1
roots of s are
complex conjugate
For  = 1
roots are
repeating at
s = - n
For  > 1
roots are real
 = 0 : undamped
01:
under damped
  1: critically damped
 > 1: over damped
Unit step response
r (t) = u-1 (t)
1
R (s) = ----
s
 0
2
c( s )  2 n
s ( s  n )
2
 1 s
  2 2
s s  n
 = 0
c ( t ) = 1 - cos nt
(pure oscillation)
 0    1:
n 2
C(s) =
s (s 2 + 2 ns + n 2)
 1 s + 2 n
C(s) = - 2
s s + 2 ns +n 2
 1 s +  n
= - 2
s s + 2 ns +n 2

 n
s + 2 ns +n
2 2
we know
1
 u -1 ( t )
s
s
 cos at
s a
2 2
 a
 sin at
s a
2 2
 s+b
2 2
e -bt
cos at
(s+b) +a
a
2 2
e -bt
sin at
(s+b) +a
 s   n

s  2 n s  n
2 2

( s   n )
( s   n )  d
2 2
  nt
e cosd t
  n

s  2 s  n
2
n
2

 n d
.
d ( s   n )  d
2 2
 n -  n t
.e sin d t
d
  -  n t
 .e sin  d t
1 2
Combine sine cos terms :
1
1- 2

1


( 1- 2
)
 1
sin ( dt + 
1-  2

where cos  =  .
 c (t) =
-  n t
e
1- sin ( d t +  )
1 2
 e- nt
C(t) 1+
 1- 2
1.0 e-nt
1-
 1-2
Time(t)

Unit step response of 2nd order system for  <1

where (-   n) is the
real part of roots


-n

 d is the imaginary
part of roots
 j n 1 -   d
2


Cos  = 
 n 1- 
2
n

- n

C(t)
Mp Tolerance band
1.0

Time
tr
tp

Time response spceification

Peak overshoot (Mp)

dc
0
dt
 sin (dt + ). cos 

- cos (dt + ).Sin  = 0

 sin (dt) = 0
 dt = 
Time at peak
overshoot ( tp )

tp = /d
 At t = tp ,
 n .t p
e
c(t p )  1  sin ( +  )
1 2
  n . n 1 2
 1 e
 1 2

M p  e
  1 2

e  100%.
As damping factor
increases …

peak overshoot ??
As damping factor
increases …

peak overshoot reduces

 Mp  Mp
0 100 0.6 9.5
0.1 73 0.7 4.6
0.2 53 0.8 1.5
0.3 37 0.9 0.02
0.4 25 1.0 0
0.5 16
100
90
80
70
60
50
40
30
20
10
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1


Settling time (ts)
Assume tolerance
band  5 %
0.05 = e -  n ts
ln (0.05)   nt s
3
or ts =
 n
t s for 2% tolerance is
4
 n
Rise time (tr)
-  n t r
e
c (t r ) = 1 = 1 -
1 2

sin ( d t r +  )
or
sin (dtr + ) = 0
dtr +  = 
or

tr 
d
Steady state
error for step
input = 0
Steady state error
of a second order
system with ramp
input:
E(s) = R(s) - C(s)

1 n 2
= 1-
s2 (s2 + 2 ns + n2)

1 s  2 n s
2
.
s s  2 n s +  n
2 2 2
e(t) = Lt sE(s)
t  s 0

 2
n
For  = 0.5 &
n = 1

ess = 1 unit
 Some more MATLAB
commands .
>> ng = [1 1]; dg =[1 2];
>> printsys (ng,dg)
num/den =
 S1
S2

>> z = roots(ng)
z =
-1
>> p = roots(dg)
p =
-2
>>
>>pzmap(ng,dg)

>>title(‘pole-zero map’)
>>[nt,dt] =
cloop(ng,dg,-1)
>>zc = roots(nt)
zc =
-1
>>pc = roots(dt)
pc = -1.5
>>
>>ng = [1];
dg = [1 1 1];
>>p = roots(dg)
p=
-0.5 + j 0.866
-0.5 - j 0.866
>>t = [0:0.5:10];

>>[y,x,t] = step(ng,dg,t);

>>plot(t,y), grid
>>xlabel(‘Time[sec]’)

>>ylabel(‘output’)
An example :

R + 1 C
K s(s+1)
-
 K
G
s( s  1)
K
C R  2
s sK
nn  K
1
 
2 K
Let K = 100

n = 10 rad / sec

 = 0.05
 Mp 
  1  2

e  100%
= 85 %
ess for ramp input
= 2 /n
= 0.1 / 10
= 0.01
While the steady state
error is acceptable
Mp is not. - Damping
to be increased to say,
0.5
Then 2 n = 1
1
 n = 1
2  0.5
ess increases to 1 - not
acceptable. So we go for
changing the configuration
of the amplifier - derivative
error compensation.
 Error coefficients

R(s) + E(s) C(s)

G(s)
-
We know
C (s) = R (s) . G (s) / ( 1+G (s) )
and
E (s) G (s) = C (s)
 E (s) = R (s) 1 + G (s) 
Steady state error (ess )
= Lt s E (s)
s0
step input
R (s) = 1 /s
(constant position)
1 1
ess  Lt s .
s0 s 1  G( s)
Let

Lt G( s)  K p
s 0
(position error coeff.)
Then

1
ess 
1 K p
Ramp input

R (s) = 1/s2

(Constant velocity)
Steady state error
in position
 1 1
 Lt s 2 .
s 0 s 1  G(s)
1
Lt  1
s 0 sG ( s ) Kv
K v = velocity error coefficient
parabolic input
R (s) = 1/s 3
(constant acceleration)
Steady state error
in position
 1 1
 Lt s. 3
s 0 s 1  G (s)
1
 Lt 2  1
s 0 s G ( s ) Ka
Ka = acceleration error
coefficient
 ( s  z1 )( s  z2 )(....)
G ( s)  n
s ( s  p1 )( s  p2 )(....)
n  type of the system
Type ‘0’ system

K p  Lt G ( s )  finite value
s 0

1
 e ss   finite value
1 K p
Type ‘0’ system

K v  Lt sG (s)  0
s 0
 e ss  
Type ‘0’ system

K a  Lt s G (s)  0
2
s 0
 e ss  
Type ‘1’ system

K p  L t G (s)  
s 0
 e ss  0
Type ‘1’ system
K v  L t s G (s)
s 0

 finite
 e ss  1 K v  finite
Type ‘1’ system

K a  Lt s G (s)  0
2
s 0
 e ss  
Type ‘2’ system

K p  L t G (s)  
s 0
 e ss  0
Type ‘2’ system

K v  L t s G (s)  
s0
 e ss  0
Type ‘2’ system

K a  Lt s G (s)  finite
2
s 0
 e ss  1 K a  finite
input
u-1(t) t2
Type tu-1(t) u 1 ( t)
2
1
0 1  Kp  
1
1 0 
Kv
1
2 0 0 Ka
When feedback is non unity,
K p  Lt G ( s ) H ( s )
s 0
K v  Lt s G ( s ) H ( s )
s 0

K a  Lt s G ( s ) H ( s )
2
s 0
 Derivative error
Compensation

R + 1 C
K s(s+1)
-
r +
Kp KA
-
L
+ 1 L
KT/Ra s(Js+B)
-
sKb
with
K = 100
n = 10 rad/sec
= 0.05
Mp = 85% (not acceptable)

t s (2%)  4 n  8 sec.

ess for ramp input
(type ‘1’ system)

= 0.01
Objective : increase the
damping factor
from 0.05 to 0.5.
Change the amplifier
transfer function
to K + s KD
 K  sK D
 G( s) 
s(s  1)
G( s ) K  s KD
 2
G( s )  1 s  s  K  s K D
n = K ( no change)
= 10 rad/sec
2 n = 1 + KD
 for  = 0.5,
KD = 9
essfor ramp input
remains the same
at 0.01
4
t s ( 2%)   0.8 sec.
 n
Derivative output Compensation

R + + 1 C
K s(s+1)
- -
sKt

Introduce another feed back loop

 K
G ( s) 
s(s  1  K t )
K
 G ( s) (1  G ( s)) 
s  s (1  K t )  K
2
Specifications
= 0.5

ess for ramp input

= 0.01
K v  Lt sG( s)
s 0

K
  100 - - - - - (1)
1 Kt
2   n  1  Kt
or K  1  K t    (2)
From (1) & (2),
K  100
or K  10000
K t  99.
Comparing both methods…

derivative output
•transducer required
•gain reduced due to
feedback
 Integral Control

+  KI  1
KA  
 s s ( s  1)
-
 K As  K I
G( s) 
s (s  1)
2
type 2 
ess for step input = 0
ess for ramp input = 0
ess for parabolic
input is finite
 However system
becomes of 3 rd order
- stability problem -
will be discussed in
later chapters