Module 4 Time response:

Part 2: Steady-State Error

2
 Steady state

• What do we mean by steady-state? This is the limiting value as time

tends to infinity

• In general, a key performance characteristic is the difference

between the steady-state output and the reference/setpoint – this is
denoted as steady-state error/offset

3
 System steady-state gain

Assume the input is a step of magnitude A and the system is G(s).

Hence:

Applying the FVT gives:

G (s) A Assumes
lim s →0 s = lim s →0 G ( s ) A = G (0) A
s that G(0) is
Therefore gain is given as:
finite!
limt → y (t ) G (0) A
= = G (0)
limt → u (t ) A

4
 Find the system gain of the following
A
H (s) =
s +1

5
G (s) =
( s + 1)( s 2 + s + 4)( s + 2)

4
C (s) =
s ( s + 3)

A
K (s) =
s −1
5
Closed-loop transfer functions

R E
C (s) G (s)
Y
-

GC
Y = R
1 + GC
C
U = R
1 + GC
1
E = R
1 + GC 6
 Closed-loop steady-state gain

Using the same definitions as earlier, one can define the closed-loop
steady-gain simply by substituting 𝑠 = 0 into the closed-loop transfer
function
R E
C (s) G (s)
Y
-

GC limt → y (t )  GC 
Y= R  = lim s →0  
1 + GC limt → r (t ) 1 + GC 

G (0)C (0)
Or, in summary: G c (0) =
1 + G (0)C (0)
7
 Use FVT to find the final value of the output
Y of the following loop
s+6 E Y
G (s) =
R
C (s) G (s)
-
( s + 4)( s + 2)
C ( s) = 2
3
R( s) =
s

8
 Use FVT to find the final value of the output
Y of the following loop

s +1
G (s) = R E
C (s) G (s)
Y
s ( s + 3) -

C ( s) = 4
H (s)
10
H (s) =
s + 10
5
R( s) =
s

9
 System steady-state error
R E
C (s) G (s)
Y
Steady-state error/Offset is -
given as the steady-state
value of R-Y.

1 A
For a simple loop diagram we E= R; R( s ) =
can write this value as: 1 + GC s

Applying the final value

1
theorem therefore gives
limt → e(t ) = A
steady-state error as:
1 + G (0)C (0)
10
R E
C (s) G (s)
Y
- Find the steady-state
offset for the following
system

2
G (s) =
( s + 1)( s + 4)
s+3
C ( s) = 6
( s + 10)
1
R( s) =
s
11
 Closed-loop steady-state error analysis

R E
C (s) G (s)
Y
-
For a closed loop of this form, the
steady-state error is given from
the formula.

1
limt → e(t ) = A
1 + G (0)C (0)
Clearly, in general the error is not
zero so the output does not track
the reference in the steady-state
12
 Eliminate steady-state error
R E
C (s) G (s)
Y
The FVT can be written more -
carefully as:

1
limt → e(t ) = lim s →0 A
1 + G ( s )C ( s )
In order for 𝑒(𝑡) to go to zero we need:

1
lim s →0 = 0 → lim s →0 1 + G ( s)C ( s) = 
1 + G ( s )C ( s )
13
 Eliminate steady-state error

• For a simple feedback loop, the steady-state error can be zero if and
only if either of following holds:

lim s →0 G ( s ) =  lim s →0 C ( s) = 

• A transfer function has an infinite gain if and only if it contains an

integrator.

14
Find the steady-state offset for the following
system/controller pair

6
G (s) =
(s + 4)(s + 2)
C (s) = 1, 20, 100

15
Find the steady-state offset for the following
system/controller pair

6
G (s) =
s(s + 4)(s + 2)
C (s) = 1

16
Find the steady-state offset for the following
system/controller pair

6
G (s) =
(s + 4)(s + 2)
20
C (s) =
s

17
 Steady-state error with disturbance
D(s)
Offset is given as the steady-
state value of R-Y. R(s) E(s) U (s) Y(s)
- C (s) G (s)

For a simple loop diagram, by

1 G
inspection.
E= R− D
1 + GC 1 + GC
Remember the final value
theorem gives. limt → e(t ) = lim s →0 se( s )
Finally, assume: 1 
R( s) = ; D( s) =
s s 18
 Steady-state error with disturbance

The FVT gives:

 1  G ( s) 
limt → e(t ) = lim s →0  − 
 1 + G ( s )C ( s ) 1 + G ( s )C ( s ) 

As the earlier part has considered the impact of the reference in detail,
this part looks only at the disturbance.
The total steady-state offset from superposition is:

− G ( s )
ed = lim s →0 = − yd
1 + G ( s )C ( s )
19
 No integrators in the loop

With no integrators in the loop the FVT gives:

 G (s)  G (0)
yd = lim s →0 =
1 + G ( s )C ( s ) 1 + G (0)C (0)
A disturbance will cause steady-state error

20
 D(s)

R(s) E(s) U (s) Y(s)

- C (s) G (s)
Find the steady-state
offset due to the
disturbance

2
G (s) =  G (0)
( s + 1)( s + 4) yd =
1 + G (0)C (0)
s+3
C ( s) = 6
( s + 10)
0.2
D( s) =
s
21
 Integrators in the process

With integrators in G(s) the FVT gives:

lim s →0 G ( s ) = 
And hence

 G (s) 
yd = lim s →0 =
1 + G ( s )C ( s ) C (0)
A disturbance will cause a steady-state error

22
 D(s)

R(s) E(s) U (s) Y(s)

- C (s) G (s) Find the steady-state
offset due to the
disturbance

2
G (s) = 
s ( s + 1)( s + 4) yd =
C (0)
s+3
C ( s) = 6
( s + 10)
0.2
D( s) =
s

23
 Integrators in the control law

With integrators in C(s) the FVT gives:

lim s →0 C ( s) = 
And hence
 G (s)
yd = lim s →0 =0
1 + G ( s )C ( s )
A disturbance will cause a ZERO steady-state error when there is an
integrator in the compensator

24
 D(s)
Find the steady-state
R(s) E(s) U (s) Y(s)
- C (s) G (s) offset due to the
disturbance

2
G (s) =
( s + 1)( s + 4)
s+3
C ( s) = 6
s ( s + 10)
0.2
D( s) =
s

25
 Find the steady-state offset for a unit input
disturbance with the following system/controller
pairs

6
G (s) =
s ( s + 2)( s + 4)
C ( s) = 1

26
 Find the steady-state offset for a unit input
disturbance with the following system/controller
pairs

6
G (s) =
( s + 4)( s + 2)
2
C (s) =
s

27
 Find the steady-state offset for a unit input
disturbance with the following system/controller
pairs

3s + 1
G (s) =
( s − 2)( s + 5)
100
C (s) =
s

28
 Steady State Error for standard inputs

• It is the error between the actual 𝑅(𝑠) 𝐸(𝑠) 𝑌(𝑠)

+ 𝐺(𝑠)
output and the desired output as -
𝑡→∞
𝑒𝑠𝑠 = lim 𝑒(𝑡) = lim 𝑟 𝑡 − 𝑦(𝑡)
𝑡→∞ 𝑡→∞
1
• By Final Value Theorem,
𝑒𝑠𝑠 = lim 𝑒(𝑡) = lim 𝑠𝐸(𝑠)
𝑡→∞ 𝑠→0
𝐺𝑅 𝑅
𝐸 =𝑅−𝑌=𝑅− =
1+𝐺 1+𝐺
𝑠𝑅(𝑠)
𝑒𝑠𝑠 = lim 𝑠𝐸(𝑠) = lim
𝑠→0 𝑠→0 1 + 𝐺(𝑠)

29
 Steady state error for standard inputs

• Unit step input:

1
𝑅 𝑠 =
𝑠
𝑠𝑅(𝑠) 1 1
𝑒𝑠𝑠 = lim = lim =
𝑠→0 1 + 𝐺(𝑠) 𝑠→0 1 + 𝐺(𝑠) 1 + 𝐾𝑝
where 𝐾𝑝 = lim 𝐺(𝑠) is called position error constant
𝑠→0

• Unit ramp (velocity) input:

1
𝑅 𝑠 =
𝑠2
𝑠𝑅(𝑠) 1 1 1
𝑒𝑠𝑠 = lim = lim = lim =
𝑠→0 1 + 𝐺(𝑠) 𝑠→0 𝑠 + 𝑠𝐺(𝑠) 𝑠→0 𝑠𝐺(𝑠) 𝐾𝑣
where 𝐾𝑣 = lim 𝑠𝐺(𝑠) is called velocity error constant
𝑠→0

30
 Steady state error for standard inputs

• Unit parabolic (acceleration) input:

1
𝑅 𝑠 = 3
𝑠
𝑠𝑅(𝑠) 1 1 1
𝑒𝑠𝑠 = lim = lim 2 = lim 2 =
𝑠→0 1 + 𝐺(𝑠) 𝑠→0 𝑠 + 𝑠 2 𝐺(𝑠) 𝑠→0 𝑠 𝐺(𝑠) 𝐾𝑎
where 𝐾𝑎 = lim 𝑠 2 𝐺(𝑠) is called acceleration error constant
𝑠→0
• The error constants 𝐾𝑝 , 𝐾𝑣 and 𝐾𝑎 describe the ability of a system to
reduce or eliminate steady state error
• These values mostly depend on the type of the system
• As the type of the system becomes higher, more steady state errors
are eliminated

31
 Type of a system

• Consider following pole-zero form of open loop transfer function of a

system:
𝐾 𝑠 + 𝑧1 𝑠 + 𝑧2 …
𝐺 𝑠 = 𝑛
𝑠 𝑠 + 𝑝1 𝑠 + 𝑝2 …
– Term 𝑠 𝑛 in the denominator denotes the number of poles (n) at
origin
– System with n-poles at origin is called Type-n system
1
– n also indicates the number of integrations ( ) in the system
𝑠
– As 𝑠 → 0, 𝑠 𝑛 dominates in determining the steady state error

32
 Features of steady state errors

• Steady state error is a measure of system accuracy

• In an ideal scenario, the system should match the reference input as
time progresses
• It means the steady state error should be as low as possible and
hence it is an important performance measure
• Steady state error depends on two factors:
– Type of the reference input 𝑅(𝑠)- step, ramp and parabolic
– Type of the system 𝐺(𝑠)
• Steady state errors are calculated only for closed loop stable
system

33
 Steady state errors for different systems

• Type-0 system:
𝐾 𝑠 + 𝑧1 𝑠 + 𝑧2 …
𝐺 𝑠 =
𝑠 + 𝑝1 𝑠 + 𝑝2 …
1 1
➢ 𝑒𝑠𝑠 = lim = (position)
𝑠→0 1+𝐺(𝑠) 1+𝐾𝑝
1 1
➢ 𝑒𝑠𝑠 = lim = = ∞ (velocity)
𝑠→0 𝑠𝐺(𝑠) 0
1 1
➢ 𝑒𝑠𝑠 = lim = = ∞ (acceleration)
𝑠→0 𝑠 2 𝐺(𝑠) 0

• Constant position error, infinite velocity and acceleration errors at

steady state

34
 Steady state errors for different systems

• Type-1 system:
𝐾 𝑠 + 𝑧1 𝑠 + 𝑧2 …
𝐺 𝑠 =
𝑠 𝑠 + 𝑝1 𝑠 + 𝑝2 …
1 1
➢ 𝑒𝑠𝑠 = lim = = 0 (position)
𝑠→0 1+𝐺(𝑠) 1+∞
1 1
➢ 𝑒𝑠𝑠 = lim = (velocity)
𝑠→0 𝑠𝐺(𝑠) 𝐾𝑣
1 1
➢ 𝑒𝑠𝑠 = lim = = ∞ (acceleration)
𝑠→0 𝑠 2 𝐺(𝑠) 0

• Zero position error, a constant velocity error and infinite acceleration

errors at steady state

35
 Steady state errors for different systems

• Type-2 system:
𝐾 𝑠 + 𝑧1 𝑠 + 𝑧2 …
𝐺 𝑠 = 2
𝑠 𝑠 + 𝑝1 𝑠 + 𝑝2 …
1 1
➢ 𝑒𝑠𝑠 = lim = = 0 (position)
𝑠→0 1+𝐺(𝑠) 1+∞
1 1
➢ 𝑒𝑠𝑠 = lim = = 0 (velocity)
𝑠→0 𝑠𝐺(𝑠) ∞
1 1
➢ 𝑒𝑠𝑠 = lim = (acceleration)
𝑠→0 𝑠 2 𝐺(𝑠) 𝐾𝑎

• Zero position error, zero velocity error and constant acceleration

error at steady state

36
 Steady-State Error for
Non-unity Feedback Systems

1 2

3 4

37
 Steady-State Error for
Non-unity Feedback Systems
Problem: Given the control system in figure below, find the value of 𝐾 so
that there is error 0.1 in steady state

38
 Steady-State Error for
Non-unity Feedback Systems
Problem: Find the steady-state error 𝑒 ∞ = 𝑟 ∞ − 𝑐(∞) for a unit step
input given the non-unity feedback system. Repeat for a unit ramp input.
Assume input and output units are the same.

39
 Steady state errors for different systems:
Example
• Find 𝐾 so that the steady-state error to a step input is minimized?

R(s) 1 Y (s)
K
- s+5

2
s+4

40
 Practical: DC Servo Modelling

• Assume that the inductance is so small that the inductor can be

neglected, model the following DC Servo by finding the transfer
function 𝜃ሶ (𝑠)
𝑉(𝑠)

Rotor+Load

• Use the measured data to find the parameters of the model

41
 Virtual laboratories by Anthony Rossister, The University of Sheffield

Practical: DC Servo Modelling

2nd press here to simulate

1st enter 3rd save Graph of

desired result to saved values
voltage graph

31