Todays Topics

(Chapter 5: 5.8, 5.9; Chapter 10: 10.2)

• Designing Controllers
– Examples (P, PD and, PID controllers)
• How to design a PID controller
– ZIEGLER-NICHOLS Tuning rules
• Steady-state errors
– Type of a system
– Steady-state errors for step, ramp and
acceleration inputs

1
EXAMPLE CONTROLLER DESIGN PROBLEM:
Suppose we have a simple mass, spring and damper problem.

Input: Force F
Output:
Displacement x

The modeling equation of this system is

Taking the Laplace transform of the modeling equation (1)

Ms X( s ) + bsX( s
2

The transfer function between the displacement X(s) and the input
F(s)
then becomes

Let M=1kg ; b=10 N.s/m ; k=20 N/m ; F(s)=1. Plug these values
into
the above transfer function.
X( s )
=
F( s ) s2 + 10s + 20
The goal of this problem is to show you how each of Kp, Ki and
Kd
contributes to obtain
• Fast rise time
characteristic equation
are (-7.2361, -2.7639).
• Minimum overshoot
What does that mean?
• No steady-state error

2
Let’s first view the open-loop step response. (We apply 1 Newton
forceto see what happens.)

Create a new m-file and

add in the following code:
num=1; den=[1 10 20];
plant=tf(num,den);
step(plant)

X( s ) 1
G (s)= =
F( s )
The DC Gain (Static Gain) of the plant transfer function :
1
K = limG ( s ) = = 0.05

YOU CAN TRY THESE ON

MATLAB ONLINE.
https://www.mathworks.com/products/matlab-online.html

3
 The DC gain of the plant transfer function is 1/20, so 0.05 is the final value of
the output to an unit-step input.

The Static Gain of theplant transfer function :

1
K static = limG ( s ) = = 0.05
s®0

This corresponds to the steady-state error of 0.95, quite large indeed.

Furthermore, the rise time is about 1 second, and the settling time is about 1.5
seconds. Let’s design a controller that will reduce the rise time, reduce the
settling time and eliminates the steady-state error.

Proportional Control:
U(s) 1 Y(s)
_

P Controller Plant

The closed-loop transfer function of the above system with a P controller is:
Kp
Y( s ) s + 10s + 20
2
=
R( s )
1+

From the Table 1, we see that the P controller (Kp) reduces rise time, increases
the overshoot and reduces the steady-state error.

4
Let the proportional gain Kp equals 300. The below plot shows that the P
controller reduced both the rise time and the steady-state error, increased
the overshoot and decreased the settling time by small amount.
P CONTROL

m-file: Kp=300; contr=Kp;

sys_cl=feedback(contr*plant,1); 9
t=0:0.01:2; step(sys_cl,t)

Proportional+Derivative (PD) Control: Now let’s take a look at a PD

control.
U(s) 1 Y(s)
Kp + K

Plant
PD Controller

The closed-loop transfer function of the above system with a PD controller is:
K p + Kd s
Y( s ) + 10s + 20
s2 K p + Kd s
=
R( s ) K + Kd s s2 + ( 10 + Kd )s + ( 20 + K p )
1+ 2 p
s + 10s + 20
From the Table 1, we see that the derivative (D) controller (Kd) reduces both the
overshoot and the settling time.
10

5
Let the proportional gain Kp equals 300 as before and let Kd equals 10.
The below plot shows that Derivative (D) control action reduced both
the overshoot and the settling time, and had small effect on the rise time
and the steady-state error.
P control PD control

m-file: Kp=300; Kd=10; contr=tf([Kd Kp],1); 11

MAK333E System Dynamics and Control Course
sys_cl=feedback(contr*plant,1); t=0:0.01:2; step(sys_cl,t)
Lecture Notes, Dr. E. Altuğ

Proportional+Integral (PI) Control: Now let’s take a look at a PI

control.
U(s) Y(s)

Plant
PI Controller

The closed-loop transfer function of the above system with a PI controller is:
Ki
K +
s
Y( s ) K p s + Ki
= s + 10s + 20 =
2

R( s ) K s3 + 10s2 + ( 20 + K p )s + Ki
K + i
1+ 2 s
s + 10s + 20
From the Table 1, we see that an integral (I) control action (Ki) decreases the rise time,
increases both the overshoot and the settling time, and eliminates the steady-state error.
12

6
Let’s reduce the proportional gain Kp to 30 and let Ki equals to 70. We
have reduced the proportional gain Kp because the integral controller also
reduces the rise time and increases the overshoot as the proportional
controller does (double effect). The below response shows that the
integral controller eliminated the steady-state error.
Closed-Loop Kp=30; Ki=70

Displacement (m)

m-file: Kp=30; Ki=70; contr=tf([Kp Ki],[1 0]); 13

sys_cl=feedback(contr*plant,1); t=0:0.01:2;
MAK333E Systemstep(sys_cl,t)
Dynamics and Control Course
Lecture Notes, Dr. E. Altuğ

Proportional+Integral+Derivative (PID) Control: Now let’s take a

look at a PID controller.
R(s)++ E(s) U(s) Y(s)
K +

Plant
PID Controller

The closed-loop transfer function of the above system with a PID controller is:

K
K + +K s
s
Y( s ) s 2
+ 10s + 20 Kd s2 + K p s + Ki
= =
R( s ) K + K i +K s s3 + ( 10 + Kd )s2 + ( 20 + K p )s +
1+ 2 s K
s + 10s + 20
14

7
 Y( s ) K d s2 + K p s + K i
=
R( s ) s3 + ( 10 + Kd )s2 + ( 20 + K p )s + Ki

After several trial and error runs, the gains Kp=350, Ki=300, and Kd=50
provided the desired response. Now we have obtained the system with no
overshoot, fast rise time and no steady-state error.
Closed-Loop Step Response (PID) Kp=350, Ki=300, and Kd=50

Displacement (m)

m-file: Kp=350; Ki=300; Kd=50; contr=tf([Kd Kp Ki],[1 0]); 15

MAK333E System Dynamics and Control Course
sys_cl=feedback(contr*plant,1); Lecture
t=0:0.01:2; step(sys_cl,t)
Notes, Dr. E. Altuğ

Comparison of
controllers

1
MAK333E System Dynamics and Control Course

8
 Example: Modeling a Cruise Control System

https://instruct.uwo.ca/engin-sc/391b/CTM/examples/cruise/cc.html

https://instruct.uwo.ca/engin-sc/391b/CTM/examples/cruise/ccPID.html

https://www.mathworks.com/products/matlab-online.html

17

GENERAL TIPS FOR DESIGNING A PID CONTROLLER

How to design a PID controller:

1. Obtain an open-loop response and determine what needs to be improved.

2. Add a proportional (P) control to improve the rise time.
3. Add a derivative (D) control to improve the overshoot.
4. Add an integral control to eliminate the steady-state error.
5. Adjust each of Kp, Ki , and Kd until you obtain a desired overall response.
You can always refer to the Table 1 given in these lecture notes to find out
the effect of each term in the PID controller.

IMPORTANT NOTE: You do not need to implement all three controllers

(proportional, derivative and integral) into a single system, if not necessary. For
example, if a PI controller gives a good enough response, then you do not need
to implement derivative (D) controller to the system. Keep the controller as
simple as possible.
18

9
 TUNING RULES FOR PID CONTROLLERS
The figure below shows a PID control of a plant. If a mathematical model of the plant
can be derived, then it is possible to apply various design techniques for determining
parameters of the controller that will meet the transient and steady-state specifications
of the closed-loop system.

However, if the plant is so complicated that its mathematical model cannot be easily
obtained, then an analytical approach to the design of a PID controller is not possible.
Then we must resort to experimental approaches to the tuning of PID controllers.
The process of selecting the controller parameters to meet given performance
specifications is known as controller tuning. Ziegler and Nichols suggested rules for
tuning PID controllers ( meaning to set values Kp , Ti ,Td ) based on experimental step
responses or based on the value of Kp that results in marginal stability when only
proportional control action is used.
19

Ziegler-Nichols rules are useful when mathematical models of plants

are not known. These rules can, of course, be applied to the design of
systems with known mathematical models.
Such rules suggest a set of values of Kp , Ti and Td that will give a
stable operation of the system. However, the resulting system may
exhibit a large maximum overshoot in the step response, which is
acceptable. In such case we need series of fine tunings until an
acceptable result is obtained.
In fact, the Ziegler-Nichols Tuning Rules give an educated
guess for the parameter values and provide a starting point for
fine tuning, rather than giving the final settings for Kp , Ti and
Td in a single shot.

20

10
 ZIEGLER-NICHOLS RULES FOR TUNING PID CONTROLLERS
Ziegler and Nichols proposed rules for determining values of the proportional gain
Kp , integral gain Ti and derivative time Td based on the transient response
characteristics of a given plant.
There are two methods.

First Method: In the first method, we obtain experimentally the response of the
plant to a unit-step input.

Procedure:
1. Apply unit-step
input
2. Obtain L and T
3. Use Table to
determine
constants

21
MAK333E System Dynamics and Control Course
Lecture Notes, Dr. E. Altuğ

If the plant involves neither integrator (s) nor dominant complex-conjugate poles, then
such a unit-step response curve may look S-shaped as shown in the Figure above.

This method applies if the response to a step input exhibits an S-shaped curve. Such
step response curves may be generated experimentally or from a dynamic simulation of
the plant.
The S-shaped curve may be characterized by two constants, delay time L and time
constant T (or τ). The delay time and time constant are determined by drawing a
tangent line at the inflection point of the S-shaped curve and determining the
intersections of the tangent line with the time axis and line c(t)=K, as shown in the
figure above. 22

11
The transfer function C(s)/U(s) may then be approximated by a first-order
systemwith a transport lag as follows:

U(
Ziegler and Nichols suggested to set the values of Kp , Ti and Td according to the
formula shown in Table 1.

23

Notice that the PID controller tuned by the first method of

Ziegler-Nichols rules gives
1 2
(s+ )
1 1 L
Gc( s ) = K p( 1 + + Td s ) = 1.2 ( 1 + + 0.5Ls ) = 0.6T
L 2Ls

Thus, the PID controller has a pole at the origin (s=0) and
doublezeros at s= -1/L .

24

12
 Example: A plot of the measured signal is shown below. Determine the PID
coefficients using Z-N method.

150 450

25

Second Method: In the second method, we first set Ti=∞ and Td =0. Using the
proportional control action only (see the first Figure below.), increase Kp from 0 to a
critical value Kcr at which the output first exhibits sustained oscillations. (Note: If
the output does not exhibit sustained oscillations for whatever value Kp may take,
then this method does not apply.)

Procedure:
1. Use P control and try
Kp from 0 to Kcr.
2. Kcr is where there is
c(t)
sustained oscillations
3. Determine
corresponding period
4. Use Table to
determine constants
t

Thus, the critical gain Kcr and the corresponding period Pcr are experimentally
determined (see the second Figure above). Ziegler-Nichols suggested that we set the
values of the parameters Kp , Ti and Td according to the formula shown in Table 2.
26

13
Ziegler-Nichols suggested that we set the values of the parameters Kp , Ti
and Td according to the formula shown in Table 2.

27

Notice that the PID controller tuned by the second method of Ziegler-
Nichols rules gives
2
æ 4 ö
÷
1 Pcr ø
+ 0.125Pcr s ) = 0.075Kcr Pcr è
Tis 0.5Pcr s

Thus, the PID controller has a pole at the origin (s=0) and double
zeros at s= -4 / Pcr .
Note: Ziegler-Nichols tuning rules have been widely used to tune PID
controllers in process control systems where the plant dynamics are not
precisely known.
Over many years, such tuning rules proved to be very useful. Ziegler-Nichols
tuning rules can, of course, be applied to plants whose dynamics are known. If
the plant dynamics are known, many analytical and graphical approaches to the
design of PID controller are available, in addition to Ziegler-Nichols tuning
rules. See Example 10-1, page 686, Ogata 4th Edition!
28

14
Example:
We are asked to find PID coefficients for a control system. It has been
observed that as we implemented P controller and increased it to 3,33,
the oscillations have started. The oscillations have a period of 500 mili-
seconds. What are the suitable values for the controllers?

Procedure:
1. Use P control and try
Kp from 0 to Kcr.
2. Kcr is where there is
sustained oscillations
3. Determine

4. Use Table to
determine constants

29

Example: Consider the following system. Apply a Ziegler–Nichols tuning rule for the
determination of the values of parameters and Then obtain a unit-step response curve
and check to see if the designed system exhibits approximately 25% maximum
overshoot.

30

15
31

32

16
33

34

17
 Additional note about ZN metods
Controller Tuning using Computational Method
Remember ZN 2nd method proposes the following PID controller:

1
G (s) = K (1+

(s +
This is in the form: Gc (s) = K
Alternatively we can write a Matlab code that searches suitable K and
a values for the required overshoot and settling time.

See Ogata Section 10.3

35

36

18
37

38

19
 39

CONTROLLER TUNING USING MATLAB

Matlab has additional tools to tune controllers.

1. USING COMMAND LINE

First of all, if you have designed controller (decided

control coefficients), you can check it using the
following:

Controller = pid(Kp,Ki,Kd)

T_pid = feedback(Controller*plant, 1);

step(T_pid) %step response

See the following example.

40

20
 Example: Remember DC motor modeling example

J=0.5;B=0.1;Kt=0.1;Kb=0.5;Ra=1;La=0.5;
s=tf('s');
num=[Kt];
den=[La*J J*Ra+B*La B*Ra+Kt*Kb];
DCmotor=tf(num,den);

We can see open loop response with

step(DCmotor);

Try a PID controller with Kp = 100, Ki = 2, Kd = 2

C=pid(100,2,2)
41

We can determine closed loop transfer function with

unity feedback:

Motor=feedback(C*DCmotor,1);
step(Motor)
Step response looks
stable, but oscillatory and
with overshoot.

Let’s increase Kd to 10.

C=pid(100,2,10)

Motor=feedback(C*DCmotor,1);
step(Motor)
The respone has been improved.
42

21
 Continue CONTROLLER TUNING USING MATLAB
Alternatively we can let Matlab design controllers using the following
command

[C_pid,info] = pidtune(plant,'PID')
If you like, you can tune this controller (change any controller gains) such as
C_pid.Kp=100;

Example: See this on the previous DC motor example

[C_pid,info] = pidtune(DCmotor,'PID')

Matlab designs a PID controller for

us with Kp = 2.81, Ki = 1.29, Kd = 1.29
Motor=feedback(C_pid*DCmotor,1);
step(Motor):

43

2. Using Control System Tuner

Matlab’s Control System Tuner is a great way to tune controllers
automatically.
The Control System Tuner app tunes control systems modeled in
MATLAB® or Simulink® (requires Simulink Control Design™
software). This app lets you tune any control system architecture
to meet your design goals. You can tune multiple fixed-order,
fixed-structure control elements distributed over one or more
feedback loops.

Control System Tuner automatically tunes the controller

parameters to satisfy the must-have requirements (design
constraints) and to best meet the remaining requirements
(objectives). The library of tuning goals lets you capture your
design requirements in a form suitable for fast automated tuning.
Available tuning goals include standard control objectives such
as reference tracking, disturbance rejection, loop shapes, closed-
loop damping, and stability margins. 44

22
 Format: Control System Tuner opens, set to tune this default
architecture. Next, specify the values of the blocks in the
architecture. Click to open the Standard feedback configuration
dialog box.

45

Type pidTuner(plant)
You can choose controller type, and adjust transient specs with two sliding
control buttons, and see its affect on step response.

46

23
 Example: See this on the previous DC motor example

Type the following command on Matlab:

pidTuner(DCmotor)

Adjust transient specs with two sliding control buttons, and see
itsaffect on step response.

47

2. Using Control System Designer

Matlab’s Control System Designer is a great way to design controllers.
The Control System Designer app lets you design single-input, single-output (SISO) controllers for
feedback systems modeled in MATLAB® or Simulink® (requires Simulink Control Design™
software).

Using this app, you

can:

• Design controllers using:

• Interactive Bode, root locus, and Nichols graphical editors for adding, modifying, and removing controller
poles, zeros, and gains.

• Automated PID, LQG, or IMC tuning.

• Optimization-based tuning (requires Simulink Design Optimization™ software).

• Automated loop shaping (requires Robust Control Toolbox™ software).

• Tune compensators for single-loop or multiloop control architectures.

• Analyze control system designs using time-domain and frequency-domain responses, such as step responses
and pole-zero maps.

• Design controllers for multimodel control applications. 48

24
 You can also click
edit architecture to
change settings
inside the App.

49

Type controlSystemDesigner(plant)

Same Example:

type

controlSystemDesigner(DCmotor)

50
MAK333E System Dynamics and Control Course

25
Click tuning methods and choose PID for automatic tuning.

51

Alternatively you may manually design your controller.

Need to right-click step and root locus area and specify Design
Requirements.
Move poles and zeros on Root locus and observe its affect on
step response. Click export to export your controller to Matlab.
SEE: https://youtu.be/RPzFLzKkQGs?t=1009

52
MAK333E System Dynamics and Control Course
Lecture Notes, Dr. E. Altuğ

26
 JAVA BASED CONTROL EXERCISES
http://janismac.github.io/ControlChallenges/

If you know Java take the challenge and

try to determine suitable controllers for
different systems.

53
MAK333E System Dynamics and Control Course
Lecture Notes, Dr. E. Altuğ

STEADY-STATE ERRORS IN UNITY-FEEDBACK CONTROL

SYSTEMS
In this section, we will investigate a type of steady-state error that is
caused by the incapability of a system to follow particular types of
inputs.
Any physical control system inherently suffers steady-state error in
response to certain types of inputs.
A system may have no steady-state error to a step input, but the
same system may exhibit nonzero steady-state error to a ramp input.
The only way we may be able to eliminate this error is to modify the
system structure.
Whether a given system will exhibit steady-state error for a given
type of input depends on the type of open-loop transfer function of
the system.
54

27
 CLASSIFICATION OF CONTROL SYSTEMS
Control systems may be classified according to their ability to follow
step inputs, ramp inputs, parabolic inputs and so on.
This is a reasonable classification scheme because actual inputs may
frequently be considered combinations of such inputs.
The magnitudes of the steady-state errors due to these individual
inputs are indicative of the goodness of the system.
Consider the unity-feedback control system with the following open-loop
transfer function G(s):

It involves the term sN in the denominator, representing a pole of

multiplicity N at the origin.
55

K( Tas + 1 )( Tbs + 1 )...( Tms + 1 )

G( s ) =
s N ( T1s + 1 )( T2s + 1 )...( Tps + 1 )
It involves the term sN in the denominator, representing a pole of
multiplicity N at the origin.
The present classification scheme is based on the number of
integrations indicated by the open-loop transfer function.
A system is called type 0 if N=0
type 1 if N=1
type 2 if N=2
. .
. .
. .
Note that this classification is different from that of the order of a
system.
56

28
As the type number is increased, accuracy is improved; however,
increasing the type number aggravates the stability problem. A
compromise between steady-state accuracy and relative stability is
always necessary.
We will see that, if G(s) is written so that each term in the
numerator
and denominator, except the term sN , approaches unity as s
approaches 0, then the open-loop gain K is directly related to
thesteady-state error.
Let us investigate and fill the Table below, the steady-state errors for
type0, type 1, and type 2 systems when they are subjected to various
inputs.
STEADY-STATE Step Input Ramp Input Acceleration
ERRORS r(t) =1 r(t) =t Input r(t) =t2/2
Type 0 system ? ? ?
Type 1 system ? ? ?
Type 2 system ? ? ?

57

Steady-State Errors: Consider the system shown in the Figure below.

E(s) C(s)
+ G(s)

The closed-loop transfer function is C( s ) G( s )

=
R( s ) 1 + G( s )

The transfer function between the error signal e(t) and the input signal
r(t) is
E( s ) C( s )
= 1- =
R( s ) R( s ) 1 + G( s )

where the error e(t) is the difference between the input signal and the
output signal.
58

29
The Final-Value Theorem provides a convenient way to find the
steady-state performance of a stable system. Since E(s) is

R( s ) R( s )
The steady-state error is
sR( s )
e = lime( t ) = limsE( s ) = lim
The static error constants defined in the following are figures of
merit of control systems. The higher the constants, the smaller the
steady- state error.
In a given system, the output may be the position, velocity, pressure or
temperature, or the like. The physical form of the output, however, is
immaterial to the present analysis. Therefore, in what follows, we will
call the output ‘position’, the rate of change of the output ‘velocity’ and
so on.
59

Static Position Error Constant Kposition: The steady-state error of the

system for a unit-step input is

sR( s ) 1
ess = lim sE( s ) = lim = lim s =
s®0 s®0 1 + G( s ) s®0 1 + G( s ) s 1 + G( 0
)
The static position error constant Kposition is defined by
K position = limG( s ) = G( 0 )
s®0
Thus, the steady-state error in terms of the static position error constant Kposition
is given by

1 + K position

60

30
For a type 0 system,

p
s®0

For a type 1 or higher system,

K( Tas + 1 )( Tb s + 1 )...
p = lim = ¥, for N ³ 1
1 2

Hence for a type 0 system, the static position error constant Kp is

finite, for a type 1 or higher system Kp is infinite!

61

For a unit-step input, the steady-state error ess may be summarized as

follows:

for type 0 sys

1+ K

From the foregoing analysis, it is seen that the response of a feedback

control system to a step input involves a steady-state error (offset) if
there is no integration in the feedforward path!

62

31
q If small errors for step inputs can be tolerated, then a type 0 system
may be permissible, provided that the gain K is sufficiently large. If the
gain K is too large, however, it is difficult to obtain reasonable relative
stability.
q If zero steady-state error for a step input is desired, the type of the
system must be one or higher!

63

Static Velocity Error Constant Kv: The steady-state error of the system
with a unit-ramp input is given by
1
= lime( t ) = limsE( s ) = lim s = lim
1 + G( s ) s2 s®0 sG( s )
The static velocity error constant Kv is defined by

Kv = limsG( s )
s®0

Thus, the steady-state error in terms of the static velocity error

constant Kv is given by

1
ess =
K
64

32
Thus, the steady-state error in terms of the static velocity error constant
Kv is given by
1
ss =

The term velocity error is used here to express the steady-state error
for a ramp input. The dimension of the velocity error is the same as
the system error. That is, velocity error is not an error in velocity,
but it is an error in position due to a ramp input!

For a type 0 system, sK( T s + 1 )( T s + 1 )...

Kv = lim =0
( T1s + 1 )( T2s + 1 )...
For a type 1 or higher system,
sK( Tas + 1 )( Tbs + 1
= lim

65

For a type 2 or higher system,

sK( Tas + 1 )( Tb s + 1 )...

K v = lim = ¥, for N ³ 2
s®0 s N ( T s + 1 )( T s + 1 )...
1 2

For a unit-ramp input, the steady-state error ess may be summarized as

follows:

ess =
Kv
1
ess = = , for type 1 systems
Kv

ess =
Kv

66

33
 e
K
1 1
e = = , for type
K K
1
ess =

q The foregoing analysis indicates that a type 0 system is incapable of

following a ramp input in the steady-state.
q The type 1 system with unity feedback can follow the ramp-input
with a finite error.
In steady-state operation, the output velocity is exactly the same as the
input velocity, but there is a positional error. This error is proportional to
the velocity of the input and is inversely proportional to the gain K.
67

The Figure below shows an example of the response of a type 1 system

with unity feedback to a ramp input.

The type 2 or higher system can follow a ramp input with zero error at
steady state.

68

34
 ess =
a
Note that the acceleration error, the steady-state error due to a
parabolic input, is an error in position.
The values of Ka are obtained as follows:

For a type 0 system,

s 2 K( Tas + 1 )( Tbs + 1 )...
Ka = lim =0
s®0 ( T1s + 1 )( T2s + 1 )...
For a type 1 system,

Ka = lim s K( Tas + 1 )( Tbs + 1 )... = 0

2

s®0 s( T1s + 1 )( T2s + 1 )...

70

35
For a type 2 system,
s 2 K( Tas + 1 )( Tbs + 1 )...
Ka = lim
s®0 s ( T1s + 1 )( T2s + 1 )...

For a type 3 or higher system,

s 2 K( Tas + 1 )( Tbs + 1 )...
Ka = lim N = ¥, for N ³ 3
s®0 s ( T1s + 1 )( T2s + 1 )...

Thus the steady-state error for the unit-parabolic input (acceleration

input) is:

ess

for type3 or higher systems

71

ess = ¥, for type 0 and type1systems

1
e = , for type 2 systems

ess = 0, for type3 or higher systems

Note that both type 0 and type 1 systems are incapable of following a
parabolic input (acceleration input) in the steady-state. The type 2 system
with unity feedback can follow a parabolic input with a finite error
signal. The figure below shows an example of the response of a type 2
system with unity feedback to a parabolic input.
The type 3 or higher
system with unity feedback
follows a parabolic input
with a zero error at
steady-state.

72

36
Summary: The Table below summarizes the steady-state errors for type
0, type 1, and type 2 systems when they are subjected to various inputs.

K = limsG( s ) K = lims2G( s )
s®0 s®0
STEADY-STATE Step Input Ramp Input Acceleration
ERRORS r(t) =1 r(t) =t Input r(t) =t2/2
Type 0 system 1/(1+Kp) ∞ ∞

Type 1 system 0 1/Kv ∞

Type 2 system 0 0 1/Ka

The finite values for steady-state errors appear on the diagonal line.
Above the diagonal, the steady-state errors are infinity, below the
73
diagonal, they are zero.

q Remember that the terms position error, velocity error, and

acceleration error mean steady-state deviations in the output position.
q A finite velocity error implies that after transients have died out the input
and output move at the same velocity but have a finite position difference.
q The error constants Kp , Kv and Ka describe the ability of a unity-
feedback system to reduce or eliminate steady-state error. Therefore,
they are indicative of the steady-state performance!
qIt is generally desirable to increase the error constants, while
maintaining the transient response within an acceptable range.
qIt is noted that to improve the steady-state performance we can increase
the type of the system by adding an integrator or integrators to the
feedforward path. This, however, introduces an additional stability problem.
qThe design of a satisfactory system with more than two integrators in
series in the feedforward path is generally not easy.
74

37
Example:

Figure 5-58

75

z =

76

38
 B 2z
ess = = , z = , w =
K w 2 KJ

77

78

39
79

80

40
81

82

41
42