Jomo Kenyatta University of Agriculture and Technology

BSc Electrical and Electronic Engineering

and

BSc Electronics and Computer Engineering

EEE 2403: CONTROL ENGINEERING I

January 18, 2024

EEE 2403: CONTROL ENGINEERING I Lecture Notes by A. M. Muhia

EEE 2403: CONTROL ENGINEERING I

Prerequisites

Circuit and Network Theory

Ordinary Differential Equations

Purpose

The aim of this course is to enable the student to;

1. understand control systems engineering and the control action.

2. model control systems

3. conceptualize stability of a control system.

Learning Outcomes

At the end of this course, the student should be able to;

1. Perform mathematical modelling of control systems

2. Calculate the response given the input to a control system

3. Determine the stability of a give system using Routh and Hurtwiz criterions

Course Outline

Control systems: Definition; control action, open loop, closed loop, linear and non-linear systems,
servo systems, continuous and discrete types. Principles of feedback: voltage, speed, and position
control. Transfer Function and Mathematical Modelling. Simulation of simple system: block di-
agrams as conceptional tools, Signal Flow diagrams. Time Domain Analysis of linear continuous
system: First order and second order systems. Stability analysis: Hurwitz criterion, Routh Crite-
rion. Root Locus. Frequency Domain Analysis. Bode plots

Teaching Methodology
2 hour lectures and 1 hour tutorial per week, and at least five 3-hour laboratory sessions per
semester organized on a rotational basis.

1
EEE 2403: CONTROL ENGINEERING I Lecture Notes by A. M. Muhia

Instructional Materials/Equipment

1. Control Engineering laboratories;

2. Computer laboratory;

3. Overhead projector;

Prescribed Text Books

1. Distefano J. J., Stubberud A. R. Williams I. J. (1994) Feedback and Control Systems: Theory
and Problems (Schaum’s Outline Series), McGraw-Hill, 2 Ed.

2. Ogata K. (1996) Modern Control Engineering, Prentice Hall, 3rd Ed.

References

1. Kuo B. C. Farid G. (2002) Automatic Control Systems, Wiley, 8th Ed.

2. Gene F. (2005) Feedback Control of Dynamic Systems, Prentice Hall, 5th Ed.

3. Journal of Dynamic Systems, Measurement, and Control

2
EEE 2403: CONTROL ENGINEERING I Lecture Notes by A. M. Muhia

Contents
1 Introduction to Control Systems 1
1.1 Definitions: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Requirements of a good control system . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Classification of Control systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Transfer Function and Mathematical Modelling 7

2.1 Transfer Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Modeling of Electrical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Modelling of Mechanical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Analogous Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3 Block Diagram Representation 30

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2 Canonical form of Feedback Control System . . . . . . . . . . . . . . . . . . . . . . . 31
3.3 Rules for Block Diagram Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4 Signal Flow Graphs 47

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Important of Signal Flow Graphs terms . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.3 Rules of Signal flow Graph Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.4 Mason’s Gain Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5 Time Domain Analysis 62

5.1 Transient Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.2 Steady State Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.3 Analysis of first order System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.4 Analysis of Second order system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.5 Time domain specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

6 Stability Analysis 86
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.2 Hurwitz Stability Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.3 Routh’s Stability Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

3
EEE 2403: CONTROL ENGINEERING I Lecture Notes by A. M. Muhia

7 Root Locus 94
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
7.2 Angle and Magnitude Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.3 Procedure for Drawing Root Locus . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

8 Frequency Domain Analysis 109

8.1 Sinusoidal Response of a Linear System . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.2 Methods used in Frequency Response . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
8.3 Correlation between Time and Frequency Response Specifications . . . . . . . . . . . 112

9 Bode Plots 118

9.1 Bode plots of standard factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
9.2 Frequency Domain Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

4
Introduction to Control Systems Lecture Notes by A. M. Muhia

1 Introduction to Control Systems

1.1 Definitions:

System: An arrangement of physical components connected or related in such a manner as to

form and/or act as an entire unit that attains a certain objective

Control: to mean regulate, direct, or command a system so that the desired objective is achieved

Control system: an arrangement of physical components connected or related in such a manner

as to command, direct, or regulate itself or another system.

Input: the stimulus, excitation or command applied to a control system, typically from an exter-
nal energy source, usually in order to produce a specified response from the control system.

Output: the actual response obtained from a control system. It may or may not be equal to the
specified response implied by the input.
Thus a control system must have

• An input or inputs

• An output or outputs

• An arrangement to achieve this input-output combination

1.2 Requirements of a good control system

Accuracy: accuracy should be very high as any error arising should be detected. Can be improved
using feedback element.

Sensitivity: any control system should be very sensitive to the input signals and insensitive to
any other disturbances Noise: A good system should be able to reduce the effects of noise or
undesired inputs

Stability: stability means bounded inputs and bounded outputs. In the absence of the input,
the output should tend to zero as time increases.

Bandwidth: Operating frequency range decides the bandwidth of any system. For frequency
response of good control system, bandwidth should be large. The required output means maximum
possible output without overshoots and it should be stable for the required input frequency range.

1
Introduction to Control Systems Lecture Notes by A. M. Muhia

Speed: a good control system should have high speed i.e. the output of the system should be as
fast as possible.

Oscillation: for a good control system, oscillations of the output should be constant or sustained
oscillations.

1.3 Classification of Control systems

Control action is that quantity responsible for actuating the system to produce the output. De-
pending on whether such a control action is dependent on the status of the output, control systems
are classified as

1. Open loop systems

2. Closed looped systems

Open loop systems

This is a control system in which the output has no effect on the control action i.e. the control
action is totally independent of the output of the system. The output is neither measured nor fed
back for comparison with the input. To each reference input, there corresponds a fixed operating
condition and as a result the accuracy of the system depends on the calibration. Open loop control
can be used, in practice, if the relationship between the input and output is known and if there are
neither internal nor external disturbances.
A general open loop control system is shown in Figure 1 below.

Figure 1: Open Loop Control System

The reference input r(t) is applied to the controller which generates the actuating signal u(t) to
give the controlled output c(t).
Examples of open loop control systems

1. Automatic washing machine

2. Automatic hand drier

3. Traffic control by means of signal operated on time basis.

4. Bread toaster.

2
Introduction to Control Systems Lecture Notes by A. M. Muhia

Advantages of open loop systems

1. They are simple in construction and design

2. They are easy to maintain

3. Convenient to use when the output is difficult to measure

4. Not much problems of stability

Disadvantages of open loop systems

1. Inaccurate and unreliable

2. Inaccurate results obtained with parameter variations, internal disturbances

3. Recalibration of the controller from time to time is necessary to maintain quality and accuracy

Closed loop control system

This is a control system in which the control action is somehow dependent on the output of the
system. There is a comparison of the state of the output and the reference state. This property is
known as feedback and is the main difference between open loop and closed loop systems.
Feedback is the property of the system which permits the output to be compared with the reference
input so that appropriate control action is formed
The generalized closed loop control system is shown in Figure 2 below

Figure 2: Closed Loop Control System

r (t) - reference input

u (t) - actuating signal
e (t) - error signal
c (t) - controlled output

3
Introduction to Control Systems Lecture Notes by A. M. Muhia

b (t) - feedback signal

Part of the output is fed back to the input. The comparison between the reference input r (t) and
the feedback b (t) gives the error e (t) .
When feedback signal is positive, the system is called positive feedback system and e (t) = r (t) +
b (t).
When the feedback signal is negative, the system is called negative feedback system and e (t) =
r (t) + b (t).
The error is applied to the controller which gives the actuating signal u (t). The action of the
controller will be to drive the controlled output in such a manner so that the error is reduced to
zero i.e. the feedback signal is equal to the input signal.
Examples of closed looped control systems include

1. Automatic electric iron

2. DC motor speed controlled by tacho feedback

3. Sun-seeker solar systems

4. Automatic water level controller.

Advantages of closed loop systems

1. Accuracy is very high since any error arising is corrected.

2. Reduced effects of non-linearities

3. It senses changes in input due to parameter changes, internal disturbances etc and corrects
the same

4. High bandwidth

5. Facilitates automation

Disadvantages of closed loop systems

1. Complicated in design and maintenance costlier

2. System may become unstable

4
Introduction to Control Systems Lecture Notes by A. M. Muhia

Comparison of open loop and closed loop systems

Table 1: Open loop and closed loop systems

Sr. No. Open Loop System Closed Loop System
1. No feedback. Hence feedback Feedback exists. Hence feedback
elements absent elements exists
2. No error detector Error detector present
3. It is inaccurate It is accurate
4. Highly sensitive to parameter Less sensitive to parameter
changes changes
5. Small bandwidth Large bandwidth
6. Stable May become unstable
7. Economical Costly
8. Example: automatic toaster, Examples: Temperature control
hand drier oven, guided missile

Servomechanisms

Out of a number of closed loop systems in the industry, a particular group of applications are
called servomechanisms. Here the controlled output is position, speed, velocity, acceleration etc.
i.e. position or time derivative of position.
A servomechanism is a power amplifying feedback control system in which the controlled variable
is mechanical position or its time derivative, such as velocity, acceleration.
Example: an automobile steering mechanism where the angular position of wheels on the road is
controlled as shown in Figure 3 below

Figure 3: Servomechanism of automobile steering system

Linear and Non-linear Control Systems

Linear systems are systems which obey the principle of superposition and proportionality.
If an input x1 produces output y1 by proportionality theorem an input αx1 produces an output
αy1 . By superposition theorem, an input αx1 + βx2 will produce an output αy1 + βy2 .
Linear systems are systems where mathematical tools like Laplace, Fourier etc. can be used. These
systems can be analysed mathematically or graphically.

5
Introduction to Control Systems Lecture Notes by A. M. Muhia

• Practical systems are not fully linear but exhibit certain non-linearities. The common exam-
ples of non-linearity are saturation, friction, deadzone, backlash etc.

• Non-linear systems exhibit peculiar behaviour

• They do not obey the law of superposition.

• Standard test signals approach cannot be used here

• The stability of linear systems depends on the root location and is independent of the initial
state. In non-linear systems, stability depends on root location as well as the initial condition
and type of input.

• Non-linear systems exhibit self sustained oscillations of fixed frequency and amplitude called
Limit Cycles. Linear systems do not have this feature

• Non linear systems may exhibit hysteresis while linear systems do not show this behaviour

Table 2: Linear and Non Linear Systens

SR. NO Linear Systems Non Linear Systems
1 Obey superposition theorem Do not obey superposition theorem
2 Can be analysed by standard Cannot be analysed by standard
test signals test signals
3 Stability depends only on the Stability depends on the root location,
root location initial condition and type of input
4 Do not exhibit Limit Cycles Exhibit Limit Cycles
5 Do not exhibit Hysteresis / Exhibit Hysteresis /Jump Resonance
Jump Resonance
6 Can be analysed by Laplace, Cannot be analysed by these methods
Fourier and Z-transforms

6
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

2 Transfer Function and Mathematical Modelling

2.1 Transfer Function

The relationship between the input and output of a system is given by the transfer function.

Definitions: The ratio of the Laplace transform of the output (or response) to the Laplace
transform of the input (or excitation) under the assumption of zero initial conditions is defined as
the transfer function of the given system.
For the system shown in Figure below

where
c (t) - output
r (t) - input
g (t)- system
Taking the Laplace transforms

Lc (t) = C (s)

Lr (t) = R (s)

Lg (t) = G (s)

Therefore the transfer function G (s)for Figure above is

Laplace of output C (s)

G (s) = = (1)
Laplace of input R (s)

Example 2.1
Consider a differential equation which gives the relationship between the input x(t) and the output
y(t)
d2 y (t) dy (t) dx (t)
+3 + 2y (t) = 3 − x (t)
dt 2 dt dt
Obtain the transfer function
Solution
Taking Laplace transform and assuming zero initial conditions

s2 Y (s) + 3sY + 2Y (s) = 3sX (s) − X (s)

7
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

And the transfer function is obtained as

Y (s) 3s − 1
=
X (s) 2
s + 3s + 2
Hence the transfer function is defined as the ratio of the Laplace transform of the output to the
Laplace transform of the input under the assumption that the initial conditions are zero

Laplace transform of three basic inputs

The three basic inputs considered in control systems are:

1. Unit step input:

Figure 4: Unit step input

the unit step input is mathematically defined as


1 t≥0
r (t) =
0 t<0

The Laplace transform of the unit step signal is given by

1
R (s) =
s
2. Ramp input:
Ramp signal is mathematically defined as

kt t>0
r (t) =
0 t<0

where k is a constant

8
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Figure 5: Ramp input

The Laplace transform of the ramp signal is given by

k
R (s) =
s2
3. Impulse input:

Figure 6: Unit impulse input

Impulse signal is mathematically defined as


1 t=0
r (t) =
0 t ̸= 0

The Laplace transform of the impulse signal is given by

R (s) = 1

Impulse Response and Transfer Function

Let us assume that the Laplace of an impulse input δ (t) is unity

C(s)
From equation (1) G (s) = then
R(s)

C (s) = G (s) R (s) (2)

9
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Equation (2) gives the relation for finding the response C (s)
Consider an impulse input applied to Fig 6

r (t) = δ (t) = impulse input

Therefore

R (s) = Lr (t) = Lδ (t) = 1

Substituting for R (s) in (2) gives

C (s) = G (s) or c (t) = L−1 C (s) = L−1 G (s) = g (t) (3)

From (3), the impulse response C (s) equals the transfer function G (s)
In other words, if the transfer function is not given but the impulse response is known, then the
transfer function is as good as known.

Definition: The transfer function of a Linear Time Invariant System is defined as the Laplace
transform of the impulse response, with all initial conditions set to zero.

Example 2.2
The impulse response of a system is e−2t . Determine the transfer function of the system.
Solution
The Laplace transform of e−2t is given by

1
L e−2t =
n o
s+2

From the above theory then the transfer function of the system is

1
G (s) =
s+2
Example 2.3
Given the input to the system of example 2.2 as e−3t determine the output of the system.
Solution

1
R (s) = L e−3t =
n o
s+3

1
G (s) =
s+2

10
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

1 1
C (s) = G (s) × R (s) = ×
s+2 s+3
Using partial fractions

1 1
C (s) = −
s+2 s+3
Taking inverse Laplace transform gives

c (t) = e−2t − e−3t

Properties of Transfer Function

1. The transfer function of a system is the Laplace transform of its impulse response for zero
initial conditions

2. Conversely the transfer function can be determined from the system input-output pair taking
the ratio of the Laplace of the output to the Laplace of the input

3. The system differential equation can be obtained from the transfer function by replacing
s-variable with the Linear Differential Operator D defined by D = d
dt
4. The transfer function is independent of the inputs to the system

5. The system poles/zeros can be found from the transfer function (discussed in the next section)

6. Stability can be determined from the characteristic equation (discussed later)

7. The transfer function is defined only for linear time invariant systems. It is not defined for
non-linear systems

Proper and improper transfer functions

A transfer function is said to be proper if the order of the denominator is greater than the order of
the numerator i.e. if the system has more poles than zeros
Example
1 1 s+1
, ,
2
s (s + 2) (s + 3) s + 6s + 4

A transfer function is said to be improper if the order of the numerator is greater than the order
of the denominator i.e. if the system has more zeros than poles

Advantages of Transfer function

1. It is a mathematical model that gives the gain of the given block/system

11
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

2. Integral differential equations are converted to simple algebraic equations

3. Once transfer function is known, any output for any given input, can be known

4. From the knowledge of the transfer function the poles and zeros of a system is found out.
Both poles and poles of a system play a very important role in response of the system.

5. From the characteristic equation, system stability can be inferred.

6. System differential equation can be obtained by replacement of variable ′ s′ by d

dt
7. Transfer function is the characteristic and property of the system itself

8. The value of the transfer function is dependent on the parameters of the system and inde-
pendent of the input applied

Disadvantages of Transfer function

1. Transfer function is valid only for liner time invariant systems

2. It does not take into account the initial conditions. Initial conditions loses its significance

3. It does not give any idea about how the present output is progressing. No idea about the
physical structure of the system is immediately known

Poles and Zeros of a transfer function

The transfer function is given by

C (s)
G (s) =
R (s)
Both C (s) and R (s) are polynomials in s, thus

bm sm + bm−1 sm−1 + · · · + b0
G (s) =
sn + an−1 sn−1 + · · · + an

k (s − b1 ) (s − b2 ) (s − b3 ) . . . (s − bm )
= (4)
(s − a1 ) (s − a2 ) (s − a3 ) . . . (s − an )

k is the system gain factor

Poles

The values of s for which the transfer function magnitude |G (s)| becomes infinite after substitution
in the denominator of the transfer function are called the poles of the transfer function.
For equation (4),s = a1 , a2 , a3 etc. are the poles of G (s). In order to get these poles, equate the
denominator of the transfer function to zero.

12
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

When the poles are not repeated, such poles are called simple poles. If repeated, such poles are
called multiple poles of the order equal to the number of times they are repeated.

Zeros

The values of s for which the transfer function magnitude |G (s)| becomes zero after substitution
in the numerator of the transfer function are called the poles of the transfer function.
For equation (4), the values of zeros are s = b1 , b2 , b3 etc. In order to get these poles, equate the
numerator of the transfer function to zero.
When the zeros are not repeated, such zeros are called simple zeros. If repeated, such zeros are
called multiple zeros of the order equal to the number of times they are repeated.

Pole-zero plot

This is the diagram obtained by locating all poles and zeros of the transfer function in the s-plane.
The s-plane has two axis, real and imaginary axis. Simple poles are shown by X and simple zeros
by a circle O.
Repeated poles are shown by repeated X (equal to the order of multiplicity) while repeated zeros
are shown by overlapping circles or circles side by side (equal to the order of multiplicity).
Example 2.4
Given the transfer function

3 (s + 3) (s + 1.5)3
G (s) =
(s + 5) (s + 7)2
Determine and plot the poles and zeros of the transfer function
Solution
Poles are at s = −5 (simple pole) and s = −7 (repeated pole of 2nd order multiplicity)
Zeros are at s = −3 (simple zero) and s = −1.5 (repeated zero of 3rd order multiplicity)
The pole-zero plot is as shown in Figure 7

2.2 Modeling of Electrical Systems

The basic passive elements of electrical systems are resistance, inductance and capacitance as shown
in Figure 8
For a resistive element, Ohm’s law can be written as

(v1 (t) − v2 (t)) = Ri (t)

(V1 (s) − V2 (s)) = RI (s) (1)

13
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Figure 7: Pole Zero plot

Figure 8: Basic electrical elements

For an inductive element, the relationship between voltage and current is

di
(v1 (t) − v2 (t)) = L
dt

(V1 (s) − V2 (s)) = LsI (s) (2)

For a capacitive element, the electrostatic equation is

Q (t) = C (v1 (t) − v2 (t))

Differentiating both sides yields

dq d
= i (t) = C (v1 (t) − v2 (t))
dt dt

sQ (s) = I (s) = Cs (V1 (s) − V2 (s)) (3)

14
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Laplace transform of Electrical Networks

Element Time-domain expression s - domain expression

R v (t) = i (t) R V (s) = I (s) R
L v (t) = L di V (s) = LsI (s)
dt
C v (t) = C1 R i (t) dt 1 I (s)
V (s) = SC

Example 2.5
Obtain the transfer function for the circuit of Figure 9

Figure 9: Example 2.5

Solution
Applying KVL to this circuit
Z
1
vi (t) = Ri (t) + i (t) dt..........(i)
C
and
Z
1
v0 (t) = i (t) dt.................(ii)
C
Taking Laplace transform of both equations yields

1
Vi (s) = RI (s) + I (s) .......(iii)
sC

1
V0 (s) = I (s) ..............(iv)
sC
From (iv)

I (s) = sCV0 (s) .............(v)

Substituting (v) into (iii)

 
1
Vi (s) = R+ sCV0 (s) .............(vi)
sC

15
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

The transfer function is obtained from (vi) as

V (s) 1
G (s) = 0 =  
Vi (s) 1
sC R + sC

Example 2.6
Determine the transfer function of the network of Figure 10. Hence determine the voltage U2 (t)
when the inputs are

1. U1 (t) = 5

2. U1 (t) = 5t

3. U1 (t) = 5t2

Figure 10: Example 2.6

Solution
The Laplace of the given circuit is as shown in Figure 11

Figure 11: Laplace form of the circuit

Applying KVL to the circuit of Figure 11

 
1
U1 (s) = I (s) + 2 ..................(i)
s

U2 (s) = 2I (s) .................(ii)

From (ii)

16
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

U2 (s)
I (s) = ...............(iii)
2
Substituting (iii) into (i)
 
U2 (s) 1
U1 (s) = +2
2 s

The transfer function is obtained as

U2 (s) s
= ..............(iv)
U1 (s) s + 0.5

1. U1 (t) = 5
∴U1 (s) = 5s

Substituting this value into (iv)

U2 (s) = 5
(s+0.5)

Taking Inverse Laplace Transform

U2 (t) = 5e−0.5t

2. U1 (t) = 5t

∴U1 (s) = 52
s

Substituting this value into (iv)

U2 (s) = 5
s(s+0.5)

By partial fractions

U2 (s) = 10 10
s − (s+0.5)

Taking Inverse Laplace Transform

U2 (t) = 10 − 10e−0.5t

17
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

3. U1 (t) = 5t2

∴U1 (s) = 10
s3

Substituting this value into (iv)

U2 (s) = 2 10
s (s+0.5)

By partial fractions

U2 (s) = 20 40 40
s − s + (s+0.5)

Taking Inverse Laplace Transform

U2 (t) = 20t − 40 + 40e−0.5t

2.3 Modelling of Mechanical Systems

Mechanical systems are of two types:

1. Translational
The motion of the body is along a straight line or curved path

2. Rotational
The motion of the body is about its own axis

Mechanics of translational motion

There are three elements here

1. Mass

Figure 12: Mass

18
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

The physical model of a mass element assumes that mass is concentrated at the center of the
body. When a block of mass M is subjected to a force F, it undergoes a displacement. The
relationship between the force and mass is
F ∝ acceleration

F = M a where a is the acceleration

2
∴F = M dv = M d 2x
dt dt

where v - velocity and x - displacement

2. Spring

Figure 13: Spring

Here the body subjected to force undergoes elastic deformation. The relationship is

F ∝displacement

F = Kx where K is the spring constant

3. Damper

Figure 14: Damper

Viscous friction: it is between moving surfaces separated by a viscous fluid or between a solid
body and a fluid medium. It is proportional to velocity. While friction opposes motion, it is
not always undesirable. Viscous is sometimes introduced intentionally also and is shown by
a dashpot
Here F ∝ velocity

19
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

F = B dx = B d (x1 − x2 )
dt dt

x1 and x2 being the displacement of the two ends of a dashpot e.g. piston

Translational elements

Sr.No. Element Relation Remarks

2
1 F = M d 2x Mass
dt

2 F = B dx = B d (x1 − x2 ) Dumper
dt dt

3 F = K (x1 − x2 ) Spring

Mechanics of Rotational motion

There is a lot of similarity between translational motion and rotational motion as far as modeling
is concerned. Table below is a list of counter parts of these systems

Sr. No. Translational Motion Rotational Motion

1 Mass(M ) Inertia(J)
2 Damper (B) Damper (B)
3 Spring (K) Spring (K)
4 Force (F ) Torque (T )
5 Displacement(x) Angular displacement (θ)
6 Velocity v = dx Angular velocity ω = dθ
dt dt
There are three elements here

1. Inertial J

Figure 15: Inertia

20
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

The relation is
2
T = J d 2θ = J dω
dt dt

Inertia has only one displacement

2. Damper B

Figure 16: Damper

The relation is

T = B dθ = B d (θ2 − θ1 )
dt dt

= B (ω2 − ω1 )

3. Spring K

Figure 17: Spring

Here the relation is

T = K (θ1 − θ2 ) = Kθ

Rotational elements

Sr. No. Element Relation Remarks

2
1 T = J d 2θ Inertia
dt

2 T = B d (θ2 − θ1 ) Damper
dt

3 T = K (θ1 − θ2 ) Spring

21
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

2.4 Analogous Systems

There are two main analogous systems

1. Force-voltage analogy

2. Force-current analogy

Force-Voltage analogy

The force is analogous to voltage. Consider the circuit of Figure 18 below

Figure 18: Series RLC Circuit

Applying KVL
Z
di 1
V (t) = Ri + L + idt
dt C
but
dq
i=
dt
therefore
dq d2 i 1
V (t) = R +L + q
dt dt2 C
Laplace transform gives

1
V (s) = sRQ (s) + s2 LQ (s) + Q (s)
C

1
= s2 LQ (s) + sRQ (s) + Q (s) (1)
C
Comparing this with the Translational system

d2 x dx
F =M +B + Kx
dt2 dt
Laplace gives
F (s) = s2 M X (s) + sBX (s) + KX (s) (2)

Comparing (1) and (2) gives the following analogies

22
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

F → V, M → L, B → R, K → C1 ,X → Q

Comparing with the Rotational System

d2 θ dθ
T =J +B + Kθ
dt2 dt
Laplace transform gives

T (s) = s2 Jθ (s) + sBθ (s) + Kθ (s) (3)

Comparing (1) and (3) gives the following analogies

T → V, J → L, B → R, K → C 1 ,θ → Q

Table 3: Force-voltage analogy

Translational Electrical Rotational
Force F Voltage V Torque T
Mass M Inductance L Inertia J
Damper B Resistance R Damper B
Spring K Elasticity D = C1 Spring K
Displacement x Charge q Displacement θ
dq
Velocity v Current i = Velocity ω
dt

Force-Current Analogy

Here force is analogous to current. Consider the circuit of Figure 19 below

Figure 19: Parallel RLC Circuit

Applying KCL Z
V 1 dv
i= + V dt + C
R L dt
dφ
But V =
dt
therefore

1 dq 1 d2 φ
i= + φ+C
R dt L dt2

23
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Laplace gives

1 1
I (s) = s2 Cφ (s) + sφ (s) + φ (s) (4)
R L
Comparing equations (2) and (4)
F → I, M → C, B → R 1 ,K → 1,X → φ
L
Comparing equations (3) and (4)
T → I, J → C, B → R 1 ,K → 1,θ → φ
L

Table 4: Force-Current Analogy

Translational Electrical Rotational
Force F Currenti Torque T
MassM Capacitance C Inertia J
Spring K 1
Reciprocal of Inductance L Spring K
Damper B Conductance R 1 Damper B
Displacement x Flux linkage φ Displacement θ
dx dφ
Velocity Voltage V = Velocity dθ
dt dt dt

Example 2.7
For the system shown in Fig below , write the system equations

Figure 20: Example 2.7

Solution
Step 1:
There is one mass, hence number of displacement = 1
Take one reference node in addition. Let the displacement be x at M
Step 2:
Mass M has a displacement x. connect between node x and reference Spring constant is connected
to displacement x on one end and reference. Damper B is similarly connected between elements
M and reference Force F is applied to M i.e. displacement x
Step 3:
the node diagram is as shown in Fig below

24
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Step 4:
At x incoming force = F
Outgoing force
d2 x dx
F =M +B + Kx
dt2 dt

F (s) = s2 M X (s) + sBX (s) + KX (s)

 
= s2 M + sB + K X (s)

Example 2.8
Draw the Force-current analogy for the Fig in example 2.7 above and draw the direct analogous
circuit
Solution
In force-current analogy
1 ,K → 1,X → φ
F → I, M → C, B → R L
Therefore
F (s) = s2 M X (s) + sBX (s) + KX (s)

Its analogy is

1 1
I (s) = s2 Cφ (s) + sφ (s) + φ (s)
R L
Using

V (s) = sφ (s)

1 1
I (s) = sCV (s) + V (s) + V (s)
R sL
This is the force current analogous equation. To draw the circuit, replace every element in previous
Figure with its analog value to obtain figure below

Example 2.9
For the mechanical system of Figure 21 below

25
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

1. Draw the mechanical network

2. Write the differential equations of performance

3. Draw the force voltage and force current analogous networks

Figure 21: Example 2.9

Solution
Step 1:
number of displacement = number of masses = number of nodes = 2
Take a reference node in addition. Assume displacements x1 and x2
Step 2:
For Mass
M1 is connected to x1 and M2 to x2
For spring
K1 is connected between x1 and reference
K2 is connected betweenx1 and x2
For damper
B1 is connected between x1 and reference
Step 3: Draw the nodes x1 and x2 and the reference and connect each element to obtain Figure 22
below

Figure 22: Mechanical diagram

26
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Step 4: Writing equations at nodes x1 and x2

At node x1 :
d2 x1 dx
F = M1 + B1 1 + K1 x1 + K2 (x1 + x2 )
dt2 dt

F (s) = s2 M1 X1 + sB1 X1 (s) + K1 X1 (s) + K2 (X1 (s) − X2 (s)) .........(i)

At node x2 :

d2 x1
K2 (x1 + x2 ) = M2
dt2
K2 (X1 (s) − X2 (s)) = s2 M2 X2 (s) .............(ii)

(ii) Force-Current analogy:

F → I, M → C, B → R 1 ,K → 1,X → φ
L
Replace equation (i) and (ii) term by term

1 1 1
I (s) = s2 C1 φ1 (s) + s φ (s) + φ (s) + (φ (s) − φ2 (s))
R1 1 L1 1 L2 1
But V (s) = sφ (s)

1 1 1
I (s) = s2 C1 V1 (s) + V (s) + V (s) + (V (s) − V2 (s)) ........(iii)
R1 1 sL1 1 sL2 1

Equation (ii) becomes

1
(φ (s) − φ2 (s)) = s2 C2 φ2 (s)
L2 1

1
(V (s) − V2 (s)) = s2 C2 V2 (s) ...............(iv)
sL2 1

Equations (iii) and (iv) are F − I analogy equations

Draw the circuit, by replacing every element in previous Figure with its analog value to obtain
Figure 23 below

(iii) Force-Voltage analogy:

F → V, M → L, B → R, K → C 1 ,X → Q

Putting I (s) = sQ (s) into (i) and (ii) gives

1 1
V (s) = sL1 I1 (s) + I (s) + R1 I1 (s) + (I (s) − I2 (s)) ..........(v)
sC1 1 sC2 1

27
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Figure 23: Force-Current analogy

1
(I (s) − I2 (s)) = sL2 I2 (s) ..............(vi)
sC2 1
The F − V circuit is as shown in Figure 24 below

Figure 24: Force-Voltage analogy

Exercise

1. Obtain the differential equations describing the mechanical system shown in Figure 25 and
draw the electric network using force-voltage analogy.

Figure 25: Exercise 1

2. Determine the transfer function for the electrical circuit shown in Figure 26 below

3. Write differential equations for the mechanical system shown in Figure 27 below

4. For the circuit shown in Figure 28 below C = 1µF . What values of R1 and R2 will give

28
Transfer Function and Mathematical Modelling Lecture Notes by A. M. Muhia

Figure 26: Exercise 2

Figure 27: Exercise 3

T = 0.6sec and a = 0.1 given the expression for the transfer function as

V0 (s) a (1 + sT )
=
Vi (s) 1 + aST

Figure 28: Exercise 4

29
Block Diagram Representation Lecture Notes by A. M. Muhia

3 Block Diagram Representation

3.1 Introduction

A block diagram is a pictorial representation of the system representing the relationship between
input and output of the system.
Each element of a practical system is represented as a block. The transfer function of that element is
inserted inside the block. The different blocks are interconnected to each other as per the sequence
of operation of the system.
Lines drawn between the blocks indicate the connection of the blocks. An arrow indicates the
direction of the flow of signals from one block to another.

Some important definitions

Block diagram: a pictorial representation of the cause and effect relationship between input and
output of the system

Figure 29: Block diagram

Output: The value of the input is multiplied with the value of block gain to get the output

Output = Gain × Input

R (s)input C (s) Output

Summing point: more than one signal can be added or subtracted at the summing point

Take off point: The point at which the output is taken for feedback purpose.

30
Block Diagram Representation Lecture Notes by A. M. Muhia

Forward path: The direction of flow of signal is from input to output

Feedback path: The direction of flow of signal is from output to input

Advantages of Block Diagram

1. The functional operation of the system can be observed from the block diagram

2. The diagram gives the information about performance of the system

3. Block diagram is used for analysis and design of control systems

4. It is very simple to construct block diagram for big and complicated systems

Disadvantages of Block Diagrams

1. Block diagram for given system are non unique

2. Source of energy in the system is not shown in the diagram

3. In the procedure of reduction of block diagram algebra some important functions may be
omitted or hidden. There is no check for it.

4. The block diagram does not give any information about the physical construction of the
system.

3.2 Canonical form of Feedback Control System

The canonical form consists of

1. One forward path

31
Block Diagram Representation Lecture Notes by A. M. Muhia

2. One feedback path

3. One summing point

4. One take-off point

Figure 30: Canonical form of block diagram

The following terms are used with reference to Figure 30 above

1. R (s) = Laplace transform of input, r (t)

2. C (s) = Laplace transform of controlled output, c (t)

3. E (s) = Laplace transform of error signal, e (t)

4. B (s) =Laplace transform of feedback signal, b (t)

5. G (s) = Forward transfer function

6. H (s) = Feedback transfer function

C(s)
7. = Control ratio = Closed loop transfer function
R(s)
E(s)
8. =Error ratio = Actuating signal ratio
R(s)
B(s)
9. = Primary feedback ratio
R(s)

Derivation of Transfer Function of Canonical Closed Loop System

Referring to Figure 31 below

E (s) = R (s) ± B (s) (1)

B (s) = C (s) H (s) (2)

32
Block Diagram Representation Lecture Notes by A. M. Muhia

Figure 31: Closed loop system

C (s) = E (s) H (s) (3)

Substituting (2) into (1)

E (s) = R (s) ± C (s) H (s)

C (s)
E (s) = = R (s) ± C (s) H (s)
G (s)

C (s) = R (s) G (s) ± C (s) G (s) H (s)

C (s) G (s)
=
R (s) 1 ± G (s) H (s)

Figure 32: Closed loop system transfer function

Positive sign means negative feedback

Negative sign means positive feedback
This can be used as a standard result to eliminate such simple loop in a complicated system
reduction procedure.

3.3 Rules for Block Diagram Reduction

Any complicated system can be converted into canonical form by using block diagram reduction
rules.
The following rules are widely used in block diagram reduction:

33
Block Diagram Representation Lecture Notes by A. M. Muhia

Rule 1: Blocks in Series

Any finite number of blocks in series may be algebraically combined by multiplication.

Figure 33: Blocks in Series

The blocks of Figure 33 (a) can be reduced by an equivalent single block as shown in Figure 33(b).
If there is a summing point or take-off point between the blocks, the blocks cannot be said to be
in series.

Rule 2: Blocks in Parallel

The blocks which are connected in parallel get added algebraically (considering the sign of the
signal).

Figure 34: Blocks in Parallel

Considering the system of Figure 34 above

C (s) = R (s) G1 + R (s) G2 − R (s) G3

= R (s) [G1 + G2 − G3 ]

All the three blocks can be replaced by an equivalent single block as shown in the figure below

This rule cannot be applied directly when a take-off point occurs as shown in the figure below

34
Block Diagram Representation Lecture Notes by A. M. Muhia

For a parallel combination, the direction of flow of signals through the blocks must be the same.
This rule cannot be applied in the system of figure below

Rule 3: Eliminate Feedback Loop

Eliminate feedback loop as follows

Figure 35: Feedback Loop

For negative feedback

C (s) G (s)
=
R (s) 1 + G (s) H (s)
For positive feedback

C (s) G (s)
=
R (s) 1 − G (s) H (s)
Now consider a system shown in Figure 36 below.

Figure 36: Feedback loop

Here G1 andG2 are in series. First the series combination is obtained. Then the feedback loop is
eliminated.

35
Block Diagram Representation Lecture Notes by A. M. Muhia

Rule 4: Associative Law for Summing Point

The order of summing point can be changed if two or more summing points are in series (output
remains the same).

Figure 37: Summing point

For Figure 37(a) above,

y = R (s) − B1

C (s) = y − B2 = R (s) − B1 − B2

For Figure 37(b) above,

y = R (s) − B2

C (s) = y − B1 = R (s) − B2 − B1

Which means output is same. Associative law for summing point is applicable only to summing
points which are directly connected to each other.

Rule 5: Shifting a summing point before a block

In Figure 38(a) below a summing point needs to be shifted before G as shown by the dotted line.

Figure 38: Summing point before a block

Now C (s) = X + GR (s)

36
Block Diagram Representation Lecture Notes by A. M. Muhia

To get the same output, the transformation is shown in Figure 38(b)

 
X
C (s) = G R + = X + GR (s)
G

which is the same output

NB: To shift a summing point before a block in the system, another block having a block function
1/G is added before the summing point as shown in Figure 38(b)

Rule 6: Shifting a summing point after a block

Here our aim is to shift the summing point after G as shown by the dotted line.

Figure 39: Summing point after a block

Here the output of the system is C (s) = (R + X) G

Shifting the summing point without making any change as shown in Figure 39 (b) above , the
output becomes C (s) = RG + X
This output is not the same as the original output. To obtain the same output X is multiplied
by G. This means to shift a summing point after a block, add a block having the same transfer
function at the summing point.

Rule 7: Shifting a take-off point before a block

To shift a take-off point before a block, add a block having the transfer function G in series with
the signal taking-off from the take-off point.
From Figure 40(a) above , C = X = RG
Shifting the take-off point without making any change as shown in Figure 40(b), C = RG andX = R
To satisfy the condition of Figure 40(a), a block having the transfer function G is added in series
with X as shown in Figure 40(c).

37
Block Diagram Representation Lecture Notes by A. M. Muhia

Figure 40: Take-off point before a block

Rule 8: Shifting a take-off point after a block

1 in series with
To shift a take-off point after a block, add a block having the transfer function G
the signal taking-off from the take-off point.

Figure 41: Take-off point after a block

From Figure 41(a), C = RG and X = R

Shifting the take-off point after the block without doing any modifications as shown in Figure 41(b)
yields C = RG and X = RG
1 is added in series with X as
To satisfy the original system, a block with the transfer function G
shown in Figure 41(c).

Rule 9: Shifting a take-off point after a summing point

To shift a take-off point after a summing point, one more summing point is added in series with
take-off point.
Suppose we have to shift a take-off point as shown with the dotted line in Figure 42(a)
Here Z = R and C (s) = R ± Y
Shifting the take-off point after the summing point directly as shown in Figure 42(b) yields C (s) =
R ± Y andZ = R ± Y
We achieve the original value of Z by the operation

Z =R±Y ±Y

38
Block Diagram Representation Lecture Notes by A. M. Muhia

Figure 42: Take-off point after a summing point

This operation is shown in Figure 42(c) below

Rule 10: Shifting a take-off point before a summing point

To shift a take-off point before a summing point, one more summing point is added in series with
take-off point.
Suppose we have to shift a take-off point as shown with the dotted line in Figure 43(a)

Figure 43: Take-off point before a summing point

Here Z = R ± Y and C(s) = R ± Y

Shifting the take-off point after the summing point directly as shown in Figure 43(b) yields C(s) =
R ± Y and Z = R
We achieve the original value of Z by adding a summing point in series with Z as shown in Figure
43(c)

39
Block Diagram Representation Lecture Notes by A. M. Muhia

Procedure to solve block diagram reduction problem

Step 1: Reduce the blocks in series

Step 2: Reduce the blocks in parallel
Step 3: Eliminate feedback path
Step 4: Always try to shift take-off points towards right and summing points towards left.
Step 5: Repeat above procedure till getting simplest form
Step 6: Find out transfer function for the whole system. The transfer function of the overall system
is given by

C (s)
TF =
R (s)

Example 3.1
Find the single block equivalent by block diagram reduction for the given system

Solution
G4 , G3 and G5 are in parallel. Apply rule 2

40
Block Diagram Representation Lecture Notes by A. M. Muhia

Blocks G2 and G3 + G4 + G5 are in cascade. Apply rule 1

Rule 3 can be applied to feedback loop G1 and H1

Now rule 1 can be applied to cascade blocks in the forward path

G1
1+G1 H1 and G2 (G3 + G4 + G5 )

Rule 3 is applied to the resulting feedback loop

41
Block Diagram Representation Lecture Notes by A. M. Muhia

Rule 1 is applied to the resulting block and G6

Example 3.2
Vo (s)
Model the circuit of Figure 61 in block diagram and use it to determine the transfer function
VI (s)

Figure 44: RC Circuit

Solution
Writing the voltage equation
vi (t) = Ri(t) + vo (t) (1)

In Laplace form
VI (s) = RI(s) + Vo (s)

Rearranging
V (s) − Vo (s)
I(s) = I (2)
R
The current through the capacitor is obtained as

d
i(t) = C vo (t) (3)
dt

42
Block Diagram Representation Lecture Notes by A. M. Muhia

From which
1
Vo (s) = I(s) (4)
sC
Using equations 2 and 4, the block diagram of the circuit is drawn as shown in Figure 45

Figure 45: RC Circuit

From the block diagram, the open loop transfer function is given by

1
G(s) =
sRC

The closed loop transfer function is obtained as

1
Vo (s) sRC
=
VI (s) 1
1+
sRC
1
= RC
1
s+
RC
Example 3.3
item Model, in block diagram representation, the electrical circuit of Figure 63, and obtain its
transfer function with VI (s) and Vo (s) as the input and output respectively

Figure 46: RLC Circuit

Solution

43
Block Diagram Representation Lecture Notes by A. M. Muhia

Taking KVL in mesh 1

di (t)
vi (t) = L 1 + vo (t) (5)
dt
In Laplace form
V1 (s) = sLI1 (s) + Vo (s) 0.5 mk

From which
V (s) − Vo (s)
I1 (s) = 1 (6)
sL
In mesh 2 Z
1
i2 (t) − i1 (t)dt + Ri2 (t) = 0 (7)
C
In Laplace form
1
[I (s) − I1 (s)] + RI2 (s) = 0 0.5 mk
sC 2
From which
1
I2 (s) = I (s) (8)
sRC + 1 1
The output is given by
vo (t) = Ri2 (t)

From which
Vo (s) = RI2 (s) (9)

Using Equations 7, 4.4 and 10 we model the circuit in block diagram as follows The open loop

Figure 47: Block Diagram Representation

transfer function is given by

R
G(s) =
s2 RLC + sL
The closed loop transfer function is obtained as

Vo (s) G(s)
=
VI (s) 1 + G(s)

R
=
2
s RLC + sL + R
Substituting for RLC component values

44
Block Diagram Representation Lecture Notes by A. M. Muhia

Vo (s) 1
=
VI (s) 2
s + 2s + 1
Method 2

Figure 48: RLC Circuit

Solution
Taking KVL in mesh 1
di(t)
vi (t) = L + vo (t) (10)
dt
In Laplace form
V1 (s) = sLI(s) + Vo (s)

From which
V (s) − Vo (s)
I(s) = 1 (11)
sL
Taking nodal current
i(t) = i1 (t) + i2 (t) (12)

From which
I2 (s) = I(s) − I1 (s) (13)

The drop across the capacitor is equal to the drop across the resistor
Z
1
i1 (t)dt = Ri2 (t) (14)
C

From which
1
I2 (s) = I (s) (15)
sRC 1
The output is given by
vo (t) = Ri2 (t)

From which
Vo (s) = RI2 (s) (16)

Using Equations 11, 13, 15 and 16 we model the circuit in block diagram as follows Reducing the

45
Block Diagram Representation Lecture Notes by A. M. Muhia

Figure 49: Block Diagram Representation

first feedback
1
G(s) =
sRC + 1
The open loop transfer function is given by

R
G(s) =
s2 RLC + sL

The closed loop transfer function is obtained as

Vo (s) G(s)
=
VI (s) 1 + G(s)

R
=
s2 RLC + sL + R
Substituting for RLC component values

Vo (s) 1
=
VI (s) s2 + 2s + 1

46
Signal Flow Graphs Lecture Notes by A. M. Muhia

4 Signal Flow Graphs

4.1 Introduction

A signal flow graph is a pictorial representation of a system and it displays graphically, the trans-
mission of signals in the system.
Every control system can be represented by a set of linear algebraic equations. These equations are
used to draw the signal flow graph.
Such equations contain dependent and independent variables. These variables are represented by
small circles called nodes
The relationship between nodes is represented by drawing a line between nodes. Such lines are
called branches.
Consider a simple equation
x2 = Ax1
There are two variables x2 and x1 . The signal flow graph for this equation is shown below

All branches in signal flow graphs represent the relationship between the variables
The gain is called the branch gain or branch transmittance

4.2 Important of Signal Flow Graphs terms

Input or source node

A node that has only outgoing branches

Figure 50: Input node

Output or Sink node

A node that has only incoming branches

47
Signal Flow Graphs Lecture Notes by A. M. Muhia

Figure 51: Output node

Chain node

A node that has both incoming and outgoing branches

Figure 52: Chain node

Path

The traversal of branches in the direction of branch arrows such that no node is traversed more
than once

Figure 53: Path

Forward Path

Any path from input node to output node

Figure 54: Forward path

48
Signal Flow Graphs Lecture Notes by A. M. Muhia

x1 − x2 − x3 − x4 - first forward path

x1 − x2 − x4 - second forward path

Feedback path or Feedback Loop

A path which originates and terminates on the same node

Figure 55: Feedback path

x2 − x3 − x2 - feedback path
A special name for a feedback loop that consists of any one node is a self loop

Path gain

The product of all branch gains while going through a forward path

Figure 56: Path gain

For the part x1 − x2 − x3 − x4 the path gain is A ∗ B ∗ C

Touching path

A loop and path are touching when they have a common node
Loops with one or more nodes in common are called touching loops

Figure 57: Touching path

49
Signal Flow Graphs Lecture Notes by A. M. Muhia

Non-touching loops

Loops with no common node between them

Figure 58: Non-touching path

x1 − x4 − x1
x2 − x3 − x2
no common nodes between them

Loop Gain

The product of branch gains forming a loop

Figure 59: Loop gain

For loop x1 − x3 − x5 − x2 the loop gain is L1 , L2

Dummy nodes

A branch having gain 1 can be added at the input as well as the output node. This does not affect
the transfer function of the system.
Dummy nodes cannot be added between chain nodes

Figure 60: Dummy nodes

50
Signal Flow Graphs Lecture Notes by A. M. Muhia

4.3 Rules of Signal flow Graph Algebra

Rule 1: Addition Rule

The value of a variable at a node equals the algebraic summation of signals arriving at the node.
E.g

x4 = 5x1 − 3x2 + 2x3

Rule 2: Transmission Rule

The value of a variable designated as a node in signal flow graphs is transmitted on every branch
that leaves a node

Thus
x2 = 5x1
x3 = 6sx1
x4 = 74x1

Rule 3: Multiplication Rule

When branches are in series , the gains get multiplied

51
Signal Flow Graphs Lecture Notes by A. M. Muhia

Properties of Signal Flow Graphs

1. Applies only to time invariant linear systems

2. The signal flow is only along the direction of arrows

3. The value of a variable at each node is equal to the algebraic sum of all signals entering at
that node

4. Since the system equations can be written differently the signal flow diagram is not unique

5. The gain of SFG is given by Mason’s equation

6. Every block diagram can be represented by SFG but the converse is not true

4.4 Mason’s Gain Equation

Mason’s gain formula states that the overall transfer function is given by

C (s) 1 X
T.F = = pi △i
R (s) △
where
i - number of forward paths
pi - gain of ith forward path
△ - system’s determinant calculated as
△i - value of △ for the part of the graph not touching the ith forward path

△ = 1− (sum of all individual loop gains including self loop gain)

+ (sum of all products of two non-touching loops)
− (sum of all products of three non-touching loops)
+···

Steps

With the notation for p where the first letter stands for serial number and the second number
stands for the number of non-touching loops
e.g
p11 , p22 , p31 - 1st , 2nd , 3rd of single loops
p12 , p22 , p32 - 1st , 2nd , 3rd of loops of two non-touching loops
p13 , p23 , p33 - 1st , 2nd , 3rd of 3 non touching loops
etc then the steps are

52
Signal Flow Graphs Lecture Notes by A. M. Muhia

Step 1: Find out all possible forward paths and their gains
i.e p1 , p2 , p3 . . . pn (n paths are taken)

Step 2: Find out all the single loops and their gains
i.e p11 , p21 , p31 . . . pm1 (m single loops)

Step 3: Find out all the non-touching loops and their gains out of step 2
i.e p12 , p22 , p32 . . . pp2 (p two non-touching loops)

Step 4: Find out all non-touching loops and their gains at > 2
i.e p13 , p23 , p33 . . . pq3 (q three non-touching loops)
Similarly find for other higher numbers of non-touching loops if any

Step 5: The value of △ is given by

   
△ = 1 − (p11 , p21 , p31 . . . pm1 ) + p12 , p22 , p32 . . . pp2 − p13 , p23 , p33 . . . pq3 + · · · · · · (1)

Step 6: Take p1 find △1 by equation (2) excluding all loops of step 2,3, etc that touch p1 path.
Determine △1 p1
Repeat for all forward paths pn

Step 7: Find the transfer function by

(p1 △1 + p2 △2 + · · · )
T.F =
△
Example 4.1

Solution

Step 1: only one forward path R − E − C − C

53
Signal Flow Graphs Lecture Notes by A. M. Muhia

gain = G

p1 = G

Step 2: single loop E − C − E

p11 = ±GH

Step 3:4 no two or three non-touching loops

Step 5:

△ = 1 − (p11 )

= 1 − ±GH

= 1 ± GH

Step 6: There is only one loop from step 2

It touches p1 path

△1 = 1 − (0) = 1

p1 △1 = 6

Step 7:
C p △ G
T = = 1 1 =
R △1 1 ± GH
Example 4.2

Solution

54
Signal Flow Graphs Lecture Notes by A. M. Muhia

Step 1: two forward paths

p1 − G1 G2 G3 G4

p2 − G1 G5

Step 2: four single loops

G1 G2 − H3 p11

G2 G3 − H2 p21

G4 − H1 p31

G5 H1 H4 p41

Step 3: Two non-touching loops

p11 and p31 are two non-touching loops

p12 = p11 ∗ p31

= (−G1 G2 H3 ) (−G4 H1 )

= G1 G2 G4 H1 H4

Step 4: No three non-touching loops

△ = 1 − (p11 + p21 + p31 + p41 ) + p12

= 1 − (−G1 G2 H3 − G2 G3 H2 − G4 H1 + G5 H1 H4 ) + G1 G2 G4 H1 H4

1 + G1 G2 H3 + G2 G3 H2 + G4 H1 − G5 H1 H4 + G1 G2 G4 H1 H4

Step 6: Taking p1 ⇒ G1 G2 G3 G4
loops p11 , p21 , p31 and p41 touch it
△1 = 1 − 0

=1

p1 △1 = G1 G2 G3 G4

Taking p2 ⇒ G1 G5

55
Signal Flow Graphs Lecture Notes by A. M. Muhia

all touch p2 (p11 + p21 + p31 + p41 )

△2 = 1

p2 △2 = G1 G5

Step 7:
p △ + p2 △2
T.F = 1 1
△
G1 G2 G3 G4 + G1 G5
=
1 + G1 G2 H3 + G2 G3 H2 + G4 H1 − G5 H1 H4 + G1 G2 G4 H1 H4

Example 4.3

Solution

Step 1: one forward path

p1 ⇒ G1 G2 G3 G4 G5

Step 2: four single loops

p11 = −G1 H1

p21 = −G1 G2 H2

p31 = −G4 H3

p41 = −G2 G3 G4 H4

Step 3: Two non-touching loops

loops 1 − 3 , 2 − 3 are two non-touching loops

p12 = p11 p31 = G1 G4 H1 H4

p22 = p21 p31 = G1 G2 G4 H2 H3

56
Signal Flow Graphs Lecture Notes by A. M. Muhia

Step 4: no three non-touching loops

Step 5:
∆ = 1 − (p11 + p21 + p31 + p41 ) + (p12 + p22 )

= 1 + G1 H1 + G1 G2 H2 + G4 H3 + G2 G3 G4 H4 + G1 G4 H1 H4 + G1 G2 G4 H2 H3

Step 6: All loops in 2 touch p1

△1 = 1

Step 7:
p △
T.F = 1 1
△
G1 G2 G3 G4 G5
=
G1 H1 + G1 G2 H2 + G4 H3 + G2 G3 G4 H4 + G1 G4 H1 H4 + G1 G2 G4 H2 H3

Example 4.4

Solution

Step 1: 2 forward paths

p1 ⇒ G1 G2 G3 G4 G8

p2 ⇒ G1 G5 G6 G7 G8

Step 2: two single loops

p11 ⇒ −G3 H3

P21 ⇒ −G6 H6

57
Signal Flow Graphs Lecture Notes by A. M. Muhia

Step 3: p11 and p21 are two non-touching loops

p12 = G3 G6 H3 H6

Step 4: no three non-touching loops

Step 5:
△ = 1 − (p11 + p21 ) + p12

= 1 + G3 H3 + G6 H6 + G3 G6 H3 H6

Step 6: taking p1 , p21 does not touch it

△1 = 1 − p21 = 1 + G6 H6

taking p2 , p11 does not touch it

△2 = 1 − p11

1 + G3 H3

p2 △2 = G1 G5 G6 G7 G8 (1 + G3 H3 )

Step 7:
p △ + p2 △2
T.F = 1 1
△
G1 G2 G3 G4 G8 (1 + G6 H6 ) + G1 G5 G6 G7 G8 (1 + G3 H3 )
=
1 + G3 H3 + G6 H6 + G3 G6 H3 H6
Exercise

Example 4.5
Model the circuit of Figure 61 in signal flow graph and use it to determine the transfer function
Vo (s)
VI (s)

58
Signal Flow Graphs Lecture Notes by A. M. Muhia

Figure 61: RC Circuit

Solution
Writing the voltage equation
vi (t) = Ri(t) + vo (t) (2)

In Laplace form
VI (s) = RI(s) + Vo (s)

Rearranging
V (s) − Vo (s)
I(s) = I (3)
R
The current through the capacitor is obtained as

d
i(t) = C vo (t) (4)
dt

From which
1
Vo (s) = I(s) (5)
sC
Using equations 3 and 5, the signal flow graph of the circuit is drawn as shown in Figure 62

Figure 62: Signal Flow Graph

From the signal flow graph, there is only one forward path

1
P1 =
sRC

There is only one loop

1
P11 = −
sRC

59
Signal Flow Graphs Lecture Notes by A. M. Muhia

The determinant is given by

sRC + 1
∆ = 1 − P11 =
sRC
Since the only single path touches the path

∆1 = 1

The transfer function is obtained as

Vo (s) P ∆
= 1 1
VI (s) ∆
1
sRC 1
= =
sRC + 1 sRC + 1
sRC
Example 4.6
Model, Signal Flow Graph, the electrical circuit of Figure 63, and obtain its transfer function with
VI (s) and Vo (s) as the input and output respectively

Figure 63: RLC Circuit

Solution
Taking KVL in mesh 1
di (t)
vi (t) = L 1 + vo (t) (6)
dt
In Laplace form
V1 (s) = sLI1 (s) + Vo (s)

From which
V (s) − Vo (s)
I1 (s) = 1 (7)
sL
In mesh 2 Z
1
i2 (t) − i1 (t)dt + Ri2 (t) = 0 (8)
C
In Laplace form

1 1
[I (s) − I1 (s)] + RI2 (s) = 0F romwhichI2 (s) = I (s) (9)
sC 2 sRC + 1 1

60
Signal Flow Graphs Lecture Notes by A. M. Muhia

The output is given by

vo (t) = Ri2 (t)

From which
Vo (s) = RI2 (s) (10)

Using Equations 7, 4.4 and 10 we model the circuit in signal flow graph as follows There is one

Figure 64: Block Diagram Representation

path
R
P1 =
2
s RLC + sL
There is one loop
R
P11 = −
2
s RLC + sL
The determinant is obtained as

s2 RLC + sL + R
∆ = 1 − P11 =
s2 RLC + sL

The only loop also touches the only path

∆1 = 1

The transfer function is obtained as

Vo (s) P ∆
= 1 1
VI (s) ∆
R
s2 RLC + sL R
= =
2
s RLC + sL + R s2 RLC + sL + R
s2 RLC + sL
Substituting for component values

Vo (s) 1
=
VI (s) 2
s + 2s + 1

61
Time Domain Analysis Lecture Notes by A. M. Muhia

5 Time Domain Analysis

In time domain analysis , time is the independent variable. When a system is given an excitation
(input) , there is a response (output). This response varies with time and is called the time response
Generally the response of any system has two parts

1. Transient response

2. Steady state response

5.1 Transient Response

This is the variation of output response during the time it takes to reach its final value is called
transient period ts
After completion of transient response, the output of a system becomes stable i.e transient response
becomes zero

5.2 Steady State Response

This is the final value achieved by the system output. It starts when the transient response com-
pletely dies out denoted by css (t)
Total response: addition of transient response and steady state response

c (t) = ct (t) + css (t)

Steady state error: The difference between the desired output and actual output of the system.
Illustration

Figure 65: Transient and steady state response

62
Time Domain Analysis Lecture Notes by A. M. Muhia

Steady State error

This is the difference between desired output and actual output denoted by ess and its Laplace
transform as E (s)
Consider the closed loop system below

E (s) = R (s) − B (s) (1)

= R (s) − C (s) H (s)

but
C (s) = E (s) G (s)

⇒ E (s) = R (s) − E (s) G (s) H (s)

E (s) [1 + G (s) H (s)] = R (s)

R (s)
E (s) = (2)
1 + G (s) H (s)

In time domain , error signal is denoted by e (t) and is obtained as

e (t) = L−1 E (s) (3)

Steady state error ess is given by

ess = lim e (t) (4)
t→∞
Applying final value theorem in Laplace transform

lim f (t) = lim sF (s)

t→∞ s→0

⇒ ess = lim e (t)

t→∞

63
Time Domain Analysis Lecture Notes by A. M. Muhia

= lim sE (s)
s→0

R (s)
= lim s (5)
s→0 1 + G (s) H (s)
This shows that the steady state error depends on

1. Types of input - R (s)

2. Types of system - G (s) H (s)

State error Coefficients

Control system has steady state errors for changes in position, velocity, and acceleration
kp - position error constant
kv - velocity error constant
ka - acceleration error constant
Then error constants are called static error constants

Step Input

Consider a step of magnitude A applied as a reference input to the system


A t>0
r (t) =
0 t<0

A
R (s) =
s
From (5)

A 1
ess = lim s
s→0 s 1 + G (s) H (s)

A
= lim
s→0 1 + G (s) H (s)

A
(6)
1 + lims→0 G (s) H (s)
Defining position error constant kp as

kp = lim G (s) H (s) (7)

s→0

64
Time Domain Analysis Lecture Notes by A. M. Muhia

A
ess = (8)
1 + kp
This indicates that the steady state error is controlled by position error coefficient kp and magnitude
of input A

Ramp Input

Consider a ramp input of slope A

r (t) = At

A
R (s) = (9)
s2
From (5)

A
s2
ess = lim s
s→0 1 + G (s) H (s)

A
= lim
s→0 s + sG (s) H (s)

A
=
lims→0 s + lims→0 sG (s) H (s)

A
(10)
lims→0 sG (s) H (s)
Introducing the term velocity error coefficient kv defined by

kv = lim sG (s) H (s)

s→0

A
⇒ ess = (11)
kv
Thus for ramp input , the velocity error coefficient kv will control the steady state error

Parabolic Input

Consider a parabolic reference input of slope A

2

A 2
r (t) = t
2

65
Time Domain Analysis Lecture Notes by A. M. Muhia

A
R (s) =
s3

A
s3
ess = lim s
s→0 1 + G (s) H (s)

A
= lim
s→0 s2 + s2 G (s) H (s)

A
=
lims→0 s2 G (s) H (s)
Defining acceleration error coefficient ka defined by

ka = lim s2 G (s) H (s)

s→0

A
⇒ ess =
ka

5.3 Analysis of first order System

The order of a system is the highest power of s in the denominator of the closed loop transfer
function
Consider a first order system shown below

Figure 66: First order system

1
C (s) Ts
=
R (s) 1 + T1
s
1
=
1 + Ts

Unit Step response of first order system

For unit step input to the first order system

66
Time Domain Analysis Lecture Notes by A. M. Muhia

1
R (s) =
s

C (s) 1
=
R (s) 1 + Ts

1
1 1 T
C (s) = =  
s 1 + Ts s s + T1

In partial fraction form

A B
C (s) = +
s s + T1

1 1
= −
s s+ 1
T
Taking inverse Laplace transform

1 1
c (t) = L−1 −
s s+ 1
T
−1 t
= 1 − exp T

which can be plotted as

The poles of the closed loop transfer function are given by

1 + Ts = 0

1
s=−
T

67
Time Domain Analysis Lecture Notes by A. M. Muhia

5.4 Analysis of Second order system

Definitions

Damping is the tendency of the system to oppose an oscillatory behavior

Damping is measured by a factor called damping ratio denoted by ξ
When ξ is maximum, it provides maximum opposition to oscillatory behavior of the system

Natural Frequency of Oscillation

When ξ is zero , the system oscillates naturally with maximum frequency called natural frequency
of oscillation ωn
Consider the general second order system below

C (s) 2
ωn
=
R (s) 2
s2 + 2ξωn s + ωn

where
ωn - undamped natural frequency
ξ - damping factor
this is the standard form of the closed loop transfer function of a second order system
The characteristic equation of the second order system is given by

s2 + 2ξωn s + ωn
2 =0

The poles of the transfer function are given by

68
Time Domain Analysis Lecture Notes by A. M. Muhia

q
±2ξωn ± 2 − 4ω 2
4ξ 2 ωn n
s=
2
q
= ξωn ± ωn ξ2 − 1

1. When ξ = 0
s = ±jωn ⇒ the roots are purely imaginary, Undamped system

2. When ξ = 1
s = −ξωn = −ωn roots are real and equal, Critically damped system

3. When ξ > 1
q
s = −ξωn ± ωn ξ 2 − 1 roots are real and unequal , Overdamped system

4. When 0 < ξ < 1

q
s = −ξωn ± jωn ξ 2 − 1 roots are complex conjugates , Under damped system

Undamped 2nd order system for unit step input ξ = 0

C (s) 2
ωn
=
R (s) 2
s2 + ωn

For unit step input

1
R (s) =
s

1 ωn 2
C (s) =
2
s s2 + ωn

A Bs
= +
s 2
s2 + ωn
2
ωn
A= =1
2
s2 + ωn s=0
2
ωn
B= = −1
2
s2 + ωn s=0

1 s
C (s) = −
2
s s2 + ωn
" #
−1 1 s
c (t) = L −
2
s s2 + ωn

69
Time Domain Analysis Lecture Notes by A. M. Muhia

= 1 − cos ωn t

The response of undamped second order system for unit step input is completely oscillatory

Figure 67: Undamped second order system response

Under-damped 2nd order system with unit step input 0 < ξ < 1

C (s) 2
ωn
=
R (s) 2
s2 + 2ξωn s + ωn

with R (s) = 1s

1 2
ωn
C (s) =
2
s s2 + 2ξωn s + ωn

A Bs + C
+
s 2 2
s + 2ξωn s + ωn

(A + B) s2 + (2Aξωn + C) s + Aωn
2 = ω2
n

A=1 B = −1 C = −2ξωn

1 s + 2ξωn
C (s) = −
2
s s2 + 2ξωn + ωn

" #
−1 1 s + 2ξωn
c (t) = L −
2
s s2 + 2ξωn + ωn

s + 2ξωn s + 2ξωn
=
2 2
 
s + 2ξωn + ωn (s + ξωn )2 + ωn2 − ξ2 ω2
n

s + 2ξωn
=  
(s + ξωn )2 + ωn
2 1 − ξ2

70
Time Domain Analysis Lecture Notes by A. M. Muhia

s + 2ξωn
=  q 2
(s + 2ξωn )2 + ωn 1 − ξ 2

q
let ωd = ωn 1 − ξ2

s + 2ξωn
⇒
(s + ξωn )2 + ω 2
d
s + ξωn ξωn
= +
2
(s + ξωn ) + ω 2 (s + ξωn )2 + ω 2
d d
1 s + ξωn ξωn
C (s) = − −
2
s (s + ξωn ) + ω 2 (s + ξωn )2 + ω 2
d d
from Laplace transform tables

ω
⇔ exp−at sin ωt
2
(s + a) + ω 2

s+a
⇔ exp−at cos ωt
(s + a)2 + ω 2
 
−ξω n t ξωn −ξω n t
c (t) = 1 − exp cos ωd t − exp sin ωd t
ωd
 
−ξω n t ξωn
c (t) = 1 − exp cos ωd t − sin ωd t
ωd

exp−ξωn t
q 
=1− q 2
1 − ξ cos ωd t + ξ sin ωd t
1 − ξ2

q
q 1−ξ 2
sin θ = 1 − ξ 2 tan θ = cos θ = ξ
ξ

exp−ξωn t  
c (t) = 1 − q sin θ cos ωd t + cos θ sin ωd t
1 − ξ2

71
Time Domain Analysis Lecture Notes by A. M. Muhia

sin (x + y) = sin x cos y + cos x sin y

exp−ξωn t 

c (t) = 1 − q sin θ + ωd t
1 − ξ2

Figure 68: Under-damped second order system response

The response of under damped 2nd order system with unit step input oscillates before settling to
a final value. The oscillations depend on the value of the damping ratio.

Critically damped 2nd order system with unit step input ξ = 1

C (s) 2
ωn
=
R (s) 2
s2 + 2ωn s + ωn

1
R (s) =
s

1 2
ωn
C (s) =
2
s s2 + 2ξωn s + ωn

2
ωn
s (s + ωn )2

A B C
= + +
s (s + ωn ) (s + ωn )2

A=1 B = −1 C = −ωn

1 1 ωn
C (s) = − −
s s + ωn (s + ωn )2

72
Time Domain Analysis Lecture Notes by A. M. Muhia

c (t) = 1 − exp−ωn t −ωn t exp−ωn t

= 1 − exp−ωn t [1 + ωn t]

Figure 69: Critically damped second order system response

The response of critically damped 2nd order system has no oscillations

Over damped 2nd order system with unit step input ξ > 1

C (s) 2
ωn
=
R (s) 2
s2 + 2ξωn s + ωn

1
R (s) =
s
the roots of s2 + 2ξωn s + ωn
2 = 0 are
q
−ξωn ± ωn ξ 2 − 1
q
sa = −ξωn + ωn ξ 2 − 1
q
sb = −ξωn − ωn ξ 2 − 1
let s1 = −sa and s2 = −sb

C (s) ωn 2
=
R (s) (s + s1 ) (s + s2 )

A B C
= + +
s s + s1 s + s2

ωn 2 ωn2
A= =
(s + s1 ) (s + s2 ) s1 s2
s=0

ω2 ω2
⇒A=  n  = n =1
2
q  q
−ξωn + ωn ξ 2 − 1 −ξωn − ωn ξ 2 − 1 ωn

73
Time Domain Analysis Lecture Notes by A. M. Muhia

2
ωn
B=
s (s + s2 )
s=−s1

2
ωn 2
−ωn
= q  q  q  =  q 
ξωn − ωn 2 2 2
ξ − 1 ξωn − ωn ξ − 1 + −ξωn − ωn ξ − 1 2
2 ξ − 1 s1

ωn
C= q 
2 ξ 2 − 1 s2

1 ω2 1 ωn 1
C (s) = − q n  + q 
s s + s1 s + s2
2 ξ 2 − 1 s1 2 ξ 2 − 1 s2

exp−s1 t exp−s2 t
" #
ωn
c (t) = 1 − q −
s1 s2
2 ξ2 − 1
q q
where s1 = ξωn − ωn ξ 2 − 1 and s2 = ξωn + ωn ξ 2 − 1

Figure 70: Over damped second order system response

The response of over damped 2nd order system has no oscillations but it takes longer time for the
response to reach the final steady state value

5.5 Time domain specifications

The transient response of a practical control system often exhibits damped oscillations before
reaching steady state.
A typical damped oscillating response of a system is as shown in Figure 71

Definitions

1. Decay time td - time taken for the response to reach 50% of the final value for the first time

2. Rise time tr - time taken for the response to raise from 0 to 100% for the very first time

74
Time Domain Analysis Lecture Notes by A. M. Muhia

Figure 71: Time domain specifications

3. Peak time tp - time taken for the response to reach the peak value for the very first time
or time taken for the response to reach the peak overshoot mp

4. Peak overshoot mp - ratio of the maximum peak value measured from the final value to
the maximum value
final value = c (∞)


max value = c tp


c tp −c(∞)
mp =
c(∞)
5. Settling time - time taken by the response to reach and stay within a specified error.
It is usually expressed as a % of final value

75
Time Domain Analysis Lecture Notes by A. M. Muhia

Usual tolerance error is 2% or 5% of the final value

Expression for time domain specifications

The response of 2nd order system under damped case is given by

exp−ξωn t

c (t) = 1 − q sin ωd t + θ
1 − ξ2

at t = tr , c (t) = c (tr ) = 1

exp−ξωn tr

⇒1− q sin ωd tr − θ = 1
1 − ξ2

exp−ξωn tr

⇒ q sin ωd tr − θ = 0
1 − ξ2

Since exp−ξωn tr ̸= 0 , then

sin ωd tr + θ = 0

⇒ ωd tr + θ = π

π−θ
tr =
ωd
q
1 − ξ2
θ = tan−1
ξ
q
ωd = ωn 1 − ξ2

q
1−ξ 2
π − tan−1
ξ
tr = q
ωn 1 − ξ 2

Peak time
Differentiating c (t)w.r.t. t and equating to 0

exp−ξωn t

c (t) = 1 − q sin ωd t + θ
1 − ξ2

76
Time Domain Analysis Lecture Notes by A. M. Muhia

q
ωd = ωn 1 − ξ2
q
d ξωn ω n 1 − ξ2
exp−ξωn t sin ωd t + θ − q exp−ξωn t cos ωd t + θ



c (t) = q
dt
1 − ξ2 1 − ξ2

ξωn
exp−ξωn t sin ωd t + θ − ωn exp−ξωn t cos ωd t + θ = 0



⇒q
1 − ξ2
 
ωn
q
exp−ξωn t ξ sin ωd t + θ − 1 − ξ 2 cos ωd t + θ = 0



⇒q
1 − ξ2
q
but cos θ = ξ and sin θ = 1 − ξ 2

ωn
exp−ξωn t cos θ sin ωd t + θ − sin θ cos ωd t + θ = 0



⇒q
1 − ξ2

sin (x − y) = sin x cos y − cos x sin y

ωn
exp−ξωn t sin



⇒q ωd t + θ − θ
1 − ξ2

ωn
=q exp−ξωn t sin ωd t
1 − ξ2

d
c (t) =0
dt tp

ωn exp−ξωn t
⇒ q sin ωd tp = 0
1 − ξ2

sin ωd tp = 0

⇒ ωd t p = π

π
tp =
ωd

π
= q
ωn 1 − ξ 2

77
Time Domain Analysis Lecture Notes by A. M. Muhia

Peak overshoot

% overshoot = %mp

c tp − c (∞)
%mp = ∗ 100
c (∞)
where


c tp = peak response at t = tp
c (∞) = final steady state value
at t = ∞

c (∞) = 1

exp−ξωn t

1− q sin ωd t + θ = 1
1 − ξ2

at t = tp with tp = ωπ
d

exp−ξωn tp

c tp = 1 − q sin ωd tp + θ
1 − ξ2

−ξωn ωπ
d
 

exp π
c tp = 1 − q sin ωd +θ
ωd
1 − ξ2

sin (π + θ) = − sin θ

q
ωd = ωn 1 − ξ2

q
−ξω n π/ω n 1−ξ 2

exp
c tp = 1 − q (− sin θ)
1 − ξ2
q
−ξπ/ 1−ξ 2

exp
c tp = 1 + q sin θ
1 − ξ2

but q
sin θ = 1 − ξ2

78
Time Domain Analysis Lecture Notes by A. M. Muhia

q

−ξπ/ 1−ξ 2
c tp = 1 + exp

c (∞) − c tp
% overshoot = ∗ 100
c (∞)

q
−ξπ/ 1−ξ 2
= exp ∗100

Settling time t (s)

The response of 2nd order system has 2 components

exp−ξωn t
1. decaying eonential component q
1−ξ 2

2. sinusoidal component sin ωd t + θ

The decaying component term reduces oscillations produced by the sinusoidal component. Hence
the settling time is decided by the eonential component.
The settling time can be found by equating the percentage tolerance error to the eonential term
For 2% tolerance
exp−ξωn ts
q = 0.02
1−ξ 2
For least value of ξ
exp−ξωn ts = 0.02

−ξωn ts = ln (0.02)

ln(0.02)
ts =
−ξωn

= 4
ξωn

= 4T for 2% error

For 5% error

exp−ξωn ts = 0.05

−ξωn ts = ln (0.05)

ts = 3
ξωn

79
Time Domain Analysis Lecture Notes by A. M. Muhia

= 3T

In general for a specified % error, ts is given by

ln (% error)
ts =
ξωn

Example 5.1
A control system having unity feedback has an open loop transfer function

20
G (s) =
s (1 + 4s) (1 + s)

Determine

1. Different static error coefficient

2
2. Steady state error if input = r (t) = 2 + 4t + t2

Solution
(1)
kp = lim G (s) H (s)
s→0

20
= lim =∞
s→0 s (1 + 4s) (1 + s)

kr = lim s G (s) H (s)

s→0

20
= lim s = 20
s→0 s (1 + 4s) (1 + s)

ka = lim s2 G (s) H (s)

s→0

20
= lim s2 =0
s→0 s (1 + 4s) (1 + s)
(2)
t2
r (t) = 2 + 4t +
2

2 4 1
R (s) = L {r (t)} = + +
s s2 s3

R (s)
ess = lim s
s→0 1 + G (s) H (s)

80
Time Domain Analysis Lecture Notes by A. M. Muhia

 
 
2 4 1  1
= lim s + + 
s→0 s s2 s3 1+ 20
s(1+4s)(1+s)
h i
2 + 4 + 1 (1 + 4s) (1 + s) s2
s s2 s3
= lim s
s→0 s (1 + 4s) (1 + s) + 20

ess = ∞

Example 5.2
The open loop transfer function of a unity feedback control system is given by

25
G (s) =
s (s + 5)

Determine

1. maximum overshoot

2. peak time

3. rise time

4. settling time

Solution

25
G (s) = H (s) = 1
s (s + 5)
closed loop transfer function
C (s) G (s)
=
R (s) 1 + G (s) H (s)

25
=
s2 + 5s + 25
general second order equation

2
ωn
H (s) =
2
s2 + 2ξωn s + ωn

2 = 25
ωn ωn = 5rad/s

5
2ξωn = 5 ξ= = 0.5
2∗5

81
Time Domain Analysis Lecture Notes by A. M. Muhia

q
ωd = ωn 1 − ξ2

√
= 5 1 − 0.25 = 4.33

peak time q
1−ξ 2
π − tan−1
ξ
tr = q
ωn 1 − ξ 2

π − 1.04
= = 0.485sec
4.33
settling time
for 2% error
4 4
ts = = = 1.6sec
ξωn 0.5 ∗ 5
overshoot q
−ξπ/ 1−ξ 2
%mp = exp ∗100

p
= exp−0.5π/ 1−0.52 ∗100 = 16.3%

Example 5.3
A unity feedback system has a forward path transfer function

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

Determine

1. Rise time

2. Peak time

3. Peak overshoot

4. Settling time (2% tolerance)

5. delay time

6. output response to a unit step input

Solution

s+2
G (s) = H (s) = 1
s (s + 1)

82
Time Domain Analysis Lecture Notes by A. M. Muhia

G (s)
C.L.T.F ==
1 + G (s) H (s)

s+2
=
2
s + 2s + 2
compare with general second order system

s2 + 2ξωn s + ωn
2 =0

s2 + 2s + 2 = 0

2 =2
⇒ ωn ωn = 1.414rad/s

1
2ξωn = 2 ξ= = 0.707
1.414

q
ωd = ωn 1 − ξ 2 = 0.999rad/s

q
1 − ξ2
θ = tan−1 = 0.785
ξ

π−θ
tr = = 2.36sec
ωd

π
tp = = 3.14sec
ωd

q
−ξπ/ 1−ξ 2
%mp = exp ∗100 = 4.32%

4
ts = = 4sec
ξωn

exp−ξωn t

c (t) = 1 − q sin ωd t + θ
1 − ξ2

exp−t
=1− sin (0.999t + 0.785)
0.707
Example 5.4

83
Time Domain Analysis Lecture Notes by A. M. Muhia

For the control system below, find the values of k1 and k2 so that mp = 25% and Tp = 4sec .
Assume step input

Solution
k
G (s) = 1
s2

H (s) = 1 + k2 s

C (s) G (s)
=
R (s) 1 + G (s) H (s)

k1
=
s2 + k1 k2 s + k1
comparing the denominator with
s2 + 2ξωn s + ωn
2

2 =k
ωn 1

k1 k2 = 2ξωn

2ξωn 2ξ
k2 = =
k1 ωn

mp = 25%

q
−ξπ/ 1−ξ 2
exp ∗100 = 25

q
−ξπ/ 1−ξ 2
exp = 0.25

84
Time Domain Analysis Lecture Notes by A. M. Muhia

ξπ
q = 1.386
1 − ξ2

ξ2 π2
= 1.92
1 − ξ2

ξ 2 = 0.1628

ξ = 0.4

Tp = 4sec

π π
= = q =4
ωd
ωn 1 − ξ 2

π
=4
ωn ∗ 0.9165

ωn = 0.857

2 = 0.73
k1 = ωn

2ξ
k2 = = 0.93
ωn

85
Stability Analysis Lecture Notes by A. M. Muhia

6 Stability Analysis

6.1 Introduction
A stable system is one that will have a bounded response for all possible bounded inputs A linear
system is stable if and only if all the poles of its transfer function are located on the left of the jω
axis
It is not necessary to determine the actual location of the poles of the transfer function for investi-
gating the stability of a linear system. We only need to find out if the number of poles on the right
of the s-plane is zero or not

Stability Criterion
Simple criterion are used to judge the location the poles of a characteristics equation with respect
to the left or right of the s-plane without actually solving the equation They are;

1. Hurwitz Stability Criterion

2. Routh’s Stability Criterion

6.2 Hurwitz Stability Criterion

The transfer function

C (s) b sm + b1 sm−1 + · · · + bm N (s)

= 0 =
R (s) n
a0 s + a1 sn−1 + · · · + bn D (s)

The necessary and sufficient condition for the polynomial D (s) = a0 sn + a1 sn−1 + · · · + bn = 0
to have all roots on the left half of the s-plane is that the n sub-determinants of the Hurwitz
determinant H should be positive where

a1 a3 a5 − − −
a0 a2 a4 − − −
0 a1 a3 − − −
H= 0 a0 a2 − − −
− − − − − −
− − − − − −
0 0 0 − − an

D1 = |a1 |

a1 a3
D2 =
a0 a2

86
Stability Analysis Lecture Notes by A. M. Muhia

a1 a3 a5
D3 = a0 a2 a4
0 a1 a3

Dn = H

For a stable system, all the sub-determinants should be positive

Example 6.1
Check for stability of the system whose characteristic equation is

s3 + 8s2 + 14s + 24 = 0

Solution

a1 a3 a5
H = a0 a2 a4
0 a1 a3

8 24 0
D3 = 1 14 0 = 2112
0 8 24

D1 = 8 = 8

8 24
D2 = = 88
1 14
D1 , D2 and D3 are positive, hence the system is stable

6.3 Routh’s Stability Criterion

For the transfer function

C (s) b sm + b1 sm−1 + · · · + bm N (s)

= 0 =
R (s) n
a0 s + a1 sn−1 + · · · + bn D (s)

The coefficients of D (s) are arranged in an array called Routh’s array as shown below

87
Stability Analysis Lecture Notes by A. M. Muhia

sn a0 a2 a4 a6
sn−1 a1 a3 a5 a7
sn−2 b1 b2 b3
sn−3 c1 c2 c3
| |
| |
| |
| |
s0 an
The first 2 rows are obtained from the coefficients of D (s).
Elements of the following rows are obtained as below

a a − a0 a3
b1 = 1 2
a1

a 4 − a0 a5
b2 = 1
a1
a a − a0 a7
b3 = 1 6
a1

b a − a1 b2
c1 = 1 3
b1

b a − a1 b3
c2 = 1 5
b1

b a
c3 = 1 7
b1

and so on
Each column will reduce by one as we move down the array. This process is obtained till s0 row is
obtained. The coefficient will be an for this row

Routh’s Criterion

The necessary and sufficient condition for a system to be stable is that all the terms in the first
column of the Routh’s array should have the same sign.
There should be no any sign change in the first column. When there are sign changes in the first
column, then the system is unstable. There are routes on the right hand of the s-plane.

The number of sign changes equals to the number of roots on the right hand side of the s-plane.

88
Stability Analysis Lecture Notes by A. M. Muhia

Example 6.2
Check for stability of the system whose characteristic equation is

D (s) = s4 + 5s3 + 20s2 + 40s + 50 = 0

Solution

s4 1 20 50
s3 5 40 0
s2 12 50 0
s1 230 0 0
12
s0 50 0 0

No sign changes in the first column hence no roots on the right half of the s-plane and therefore
the system is stable.

Example 6.3
Check for stability of the system whose characteristic equation is

D (s) = s3 + s2 + 2s + 24 = 0

Solution

s3 1 2
s2 1 24
s1 −22
s0 24

There are two sign changes in the first column hence two roots on the right half of the s-plane and
therefore the system is unstable.

Special cases

1. If the first element of any row is zero and the rest are non zeros
The next row cannot be formed as division by zero will take place. There are two methods
of this problem

(a) Method 1
It is replaced by a small positive number ϵ. The sign of the element of the first column
is then examined as

Example 6.4

89
Stability Analysis Lecture Notes by A. M. Muhia

Check for stability of the system whose characteristic equation is

D (s) = s5 + s4 + 2s3 + 2s2 + 3s + 5 = 0

Solution

s5 1 2 3
s4 1 2 5
s3 0 (ϵ) −2 0
s2 2ϵ+2 5
ϵ
s1 −4ϵ−4−5ϵ2
2ϵ+2
s0 5

Let ϵ → 0 from the right side

2ϵ + 2 2
lim = 2 + lim = +∞
ϵ→0 ϵ ϵ→0 ϵ
No sign changes

−4ϵ − 4 − 5ϵ2
lim = −2
ϵ→0 2ϵ + 2
Sign changes
There are two sign in the first column hence two roots on the right half of the s-plane
and therefore the system is unstable.
(b) Method 2
1 and complete the Routh’s test for z.
Replace s with Z

Example 6.5
Check for stability of the system whose characteristic equation is

D (s) = s5 + s4 + 2s3 + 2s2 + 3s + 5 = 0

Solution
Replacing s with z1

 5  4  3  2  
1 1 1 1 1
+ + +2 +3 +5=0
Z Z Z Z Z

5Z 5 + 3Z 4 + 2Z 3 + 2Z 2 + Z + 1 = 0

90
Stability Analysis Lecture Notes by A. M. Muhia

Z5 5 2 1
Z4 3 2 1
Z3 − 43 − 23
Z2 1 1
2
Z1 2
Z0 1

There are two sign in the first column hence two roots on the right half of the s-plane
and therefore the system is unstable.

2. When all elements in any one row are zeros

(a) Method 1
Form an auxiliary equation with the help of the coefficients of the row just above the
row of zeros
Take the derivatives of this equation and replace its coefficients in the present row of
zeros. Proceed with the Routh test

Example 6.6
Check for stability of the system whose characteristic equation is

D (s) = s6 + 2s5 + 8s4 + 12s + 20s2 + 16s + 16 = 0

Solution

s6 1 8 20 16
s5 2 12 16 0
s4 2 12 16
s3 0 0 0

s3 row is zero
The auxiliary equation for s4 row is

A (s) = 2s4 + 12s2 + 16

dA
= 8s3 + 24s
dS
Completed array

91
Stability Analysis Lecture Notes by A. M. Muhia

s6 1 8 20 16
s5 2 12 16 0
s4 2 12 16
s3 8 24 0
s2 6 16
s1 2.67
s0 16

No sign change on the first column hence the system may be stable

Solving for auxiliary equation

A (s) = 2s4 + 12s2 + 16 = 0

Let s2 = y

2y 2 + 12y + 16 = 0

y = −4 and y = −2
√
s = ±j2 and s = ±j 2

Non repeated roots on the jω axis give marginally stable system.

(b) Method 2
Form an auxiliary polynomial from the preceding row. The auxiliary polynomial is a
factor of the characteristic polynomial.
Obtain the number of right half plane roots in the auxiliary polynomial
Divide the characteristic polynomial by the auxiliary polynomial and obtain the number
of right half plane roots in the polynomial obtained.
The number of roots of the characteristic polynomial in the right half of the s-plane is
the sum of the two obtained above

Example 6.7
Check for stability of the system whose characteristic equation is

D (s) = s6 + 2s5 + 8s4 + 12s3 + 20s2 + 16s + 16 = 0

Solution
s3 row is zero
The auxiliary equation for s4 row is

92
Stability Analysis Lecture Notes by A. M. Muhia

A (s) = 2s4 + 12s2 + 16 = 0s

√
s = ±j2 and s = ±j 2

Two pairs of roots on the imaginary axis

With D (s) divided by A (s)

Q (s) = s2 + 2s + 2 = 0

Routh array for Q (s)

s2 1 2
s1 2 0
s0 2

No sign change in the first column

All roots of D (s) are on the left half of the s-plane with 2 pairs on the imaginary axis.
The system is therefore marginally stable

93
Root Locus Lecture Notes by A. M. Muhia

7 Root Locus

7.1 Introduction

This is a graphical procedure for sketching the movement in the s-plane of the poles of the transfer
function of the closed loop system by varying the loop gain.
Consider the canonical closed loop system

C (s) G (s)
=
R (s) 1 + G (s) H (s)
The closed loop transfer function poles are given by the characteristic equation

1 + G (s) H (s) = 0

The system gain is varied and therefore the roots of the characteristic equation depends on the
system gain k. k is varied from 0 to ∞
If cell location are joined , the resulting diagram or locus is called the Root locus

Example 7.1
For a unity feedback system with G (s) = ks . Obtain the root locus
Solution

k
G (s) = H (s) = 1
s
closed loop poles are given by

1 + G (s) H (s) = 0

k
1+ =0
s

s+k =0

94
Root Locus Lecture Notes by A. M. Muhia

s = −k

for
k=0 s=0

k=∞ s = −∞

as k is varied from 0 to ∞ , the root locus varies from 0 to ∞

The root locus starts from the poles of the open loop transfer function and ends on the zeros of the
open loop transfer function

Example 7.2
Given that
ks
G (s) =
s+3
determine when the root locus begins and ends
Solution
O.L.T.F pole
s+3=0

s = −3

O.L.T.F zero
ks = 0

s=0

Hence the root locus begins at s = −3 for k = 0 and ends at s = 0 for k = ∞

95
Root Locus Lecture Notes by A. M. Muhia

7.2 Angle and Magnitude Criterion

The closed loop transfer function are given by

1 + G (s) H (s) = 0

G (s) H (s) = −1

= 1∠180° (1)

Since s is a complex number , the above equation has both angle and magnitude and both must
be satisfied as for equation (1)
i.e

|G (s) H (s)| = 1 (2)

angle G (s) H (s) = ±180 (3)

Any point s that lie on the root locus must satisfy equations (2) and (3) which is the magnitude
and angle criterion respectively for k > 0

Example 7.3
Given that

k
G (s) H (s) =
s (s + 2)
check whether points s = −1 + j and s = −2 + j lie on the root locus and determine the value of k

Solution
(1)

s = −1 + j

k
G (s) H (s) =
s=−1+j (−1 + j) (−1 + j + 2)

k
=
(−1 + j) (1 + j)

96
Root Locus Lecture Notes by A. M. Muhia

k k
= =
2
j −1 −2

k∠0
=
2∠180

k
= ∠ − 180
2
angle is −180 , point lies on the root locus
also from magnitude criterion
k
=1
2

⇒k=2

no other gain is possible

(2)

k
G (s) H (s) =
s=−2+j (−2 + j) (−2 + j + 2)

k k
= =
j (−2 + j) −1 − j2

k∠0
=√
5∠63.43

k
√ ∠ − 63.43
5
angle criterion not satisfied hence the point does not lie on the root locus

7.3 Procedure for Drawing Root Locus

1. The root locus is always symmetrical about the real axis

2. Number of loci
let n = number of O.L.T.F poles
m = number of O.L.T.F zeros

(a) if m < n
number of loci = number of poles = n
each loci will start from an O.LT.F pole and end at O.L.T.F zero
(n − m) loci will end at infinity

97
Root Locus Lecture Notes by A. M. Muhia

(b) if m > n
number of loci = number of zeros = m
each loci will start from an O.LT.F pole and end at O.L.T.F zero
(m − n)loci will end at these zeros from ∞

Example 7.4
for
k (s + 1)
G (s) =
s (s + 2) (s + 3)

n=3 m=1

n−m=2

loci ends at infinity

one loci ends at s = −1 (zero at −1)

3. Real axis Loci

Determined by open loop transfer function poles and zeros. Any point on the real axis is a
part of the loci iff the number of poles and zeros to its right is odd.

k (s + 1)
G (s) H (s) =
s (s + 2) (s + 3)
the poles and zeros are plotted and the loci analyzed as follows

(a) All points between s = 0 and s = −1 not (inclusive) have one pole at s = 0 (odd). Hence
loci is present in the region

98
Root Locus Lecture Notes by A. M. Muhia

(b) between s = −1 and s = −2 (not inclusive) , there are two (one pole at s = 0 , zero at
s = −1) to the right .Hence loci absent
(c) loci present between s = −2 and s = −3 (not inclusive), 3 elements (2 poles s = 0, −2
and one zero s = −1) to the right
(d) no loci beyond s = −3. four elements (3 poles, one zero to the right)

4. Angle of asymptotes
A line which the locus touches at infinity. The loci of these branches that **** to infinity
move along the asymptotes
Number of asymptotes = n − m
angle of asymptotes is given by
(2k + 1)
β= 180
n−m
for k = 0, 1, 2, . . . . . . n − m − 1
from
k (s + 1)
G (s) H (s) =
s (s + 2) (s + 3)
number of asymptotes = n − m = 3 − 1 = 2

(2k + 1)
β= 180 k = 0, 1
2

β = 90° , 270°

3 loci started at s = 0, −2, −3

one loci end at s = −1
The loci starting at s = −2 and s = −3 do not have any zero at real negative axis and
therefore they must break the real axis and move out
The point they break away is called break away point, they will be guided by asymptotes
which they should touch at k = ∞

5. center of asymptotes (centroid)

The asymptotes touch the real axis at a point called centroid σc where
P P
real parts of poles of OLT F − real parts of zeros of OLT F
σc =
n−m
The centroid σc and angle of asymptotes , fixes the location of asymptotes
From previous example
poles at s = −2, −3 and 0

99
Root Locus Lecture Notes by A. M. Muhia

zero at s = −1 P P
− (0 + 2 + 3) − −1
σc =
3−1

−5 + 1
= = −2
2

6. Breakaway point
The point at which the root locus comes out of the real axis.
Obtained from
dk
=0
ds
7. Angle of arrival/departure
The root locus arrives at a complex zero at k = ∞. Similarly, the root locus departs from a
complex pole at k = 0 . These angles are known as angle of arrival and departure respectively
angle of departure from poles ϕD is given by

ϕD = 180° + arg 6 H

where arg 6 H → angle of G (s) H (s) excluding the pole when angle is calculated
angle of arrival ϕA is given by
ϕA = 180° − arg 6 H

8. jω crossover or intersection of imaginary axis

we make use of Routh criterion

100
Root Locus Lecture Notes by A. M. Muhia

9. Sketch the root loci

10. Comment on stability of the system

Example 7.5
Draw the root locus of the system

k
G (s) H (s) =
s (s + 3) (s + 6)
obtain the value of k when ξ = 0.6 from the root locus
Determine the value of k for marginal stability and critical damping

Solution

1. number of loci

n=3

m=0

number of loci is 3
loci ending at infinity is 3
There are three loci all proceeding to end at infinity

2.

3. Real axis loci

(a) present between −3 < σ < 0

(b) absent between −6 < σ < 3

101
Root Locus Lecture Notes by A. M. Muhia

(c) present between −∞ < σ < −6

three loci starting at s = 0 , s = −3 , s = −6 all proceed to infinity along 3 asymptotes
between s = 0 and s = −3 , two loci start towards each other. They must break away
between 0 and −3

4. number of asymptotes and their angles

n−m=3

(2k + 1)
β= 180 k = 0, 1, 2
n−m

(2k + 1)
180
3

= 60° , 180° , 300°

5. centroid
P P
real parts of poles of OLT F − real parts of zeros of OLT F
σc =
n−m

(0 + −3 + −6) − 0
=
3−0

= −3

6. Breakaway point

 
k = − s3 + 9s2 + 18s

dk
=0
ds

 
0 = − s3 + 9s2 + 18s

s2 + 6s + 6 = 0

s = −4.73 s = −1.27

102
Root Locus Lecture Notes by A. M. Muhia

taking s = −1.27 (real axis loci present below)

s=0 and s = −3

therefore h i
k = − (−1.27)3 + 9 (−1.27)2 + 18 (1.27)

= 10.39

critical damping k = 10.39

7. angle of arrival
Not needed : no complex poles at zeros are present

8. Intersection of imaginary axis

Characteristic equation
s (s + 3) (s + 6) + k = 0

s3 + 9s2 + 18s + k = 0

Routh array
s3 1 18
s2 9 k
s1 162−k
9
s0 k
for stability column 1 should be position

162 − k
>0
9

k < 162 k>0

⇒ 0 < k < 162

103
Root Locus Lecture Notes by A. M. Muhia

jω axis cross over s = ±j4.24

Example 7.6
The open loop transfer function of a system is

k
G (s) H (s) =
s (s + 2 + j2) (s + 2 − j2)
determine the complete root locus and comment on the stability of the closed loop system
Solution

1. number of loci

n=3

104
Root Locus Lecture Notes by A. M. Muhia

m=0

number of loci is 3

2. Real axis loci

present between 0 < σ < −∞

3. number of asymptotes and their angles

n−m=3

(2k + 1)
β= 180 k = 0, 1, 2
n−m

(2k + 1)
180
3

= 60° , 180° , 300°

4. centroid
P P
real parts of poles of OLT F − real parts of zeros of OLT F
σc =
n−m

(−2 − 2) − 0
= = −1.33
3
5. Breakaway point

1 + G (s) H (s) = 0

s3 + 4s2 + 8s + k = 0

105
Root Locus Lecture Notes by A. M. Muhia

k = −s3 − 4s2 − 8s

dk
= −3s2 − 8s − 8 = 0
ds

3s2 + 8s + 8 = 0

s = −1.33 ± j0.94

6. Angle of departure

k
σH =
s (s + 2 + j2) (s + 2 − j2)

k
σH ′ =
s (s + 2 + j2) s=−2+j2

k
=
(−2 + j2) (−2 + j2 + 2 + j2)

k
=
(−2 + j2) j4

k∠0
√
8∠135 4∠90

arg σH ′ = −225°

ϕD = 180° + arg σH ′ = −45°

ϕD = −45 at s = 2 + j2 and
ϕD = 45° at s = −2 − j2

7. jω crossover
1 + G (s) H (s) = 0

s3 + 4s2 + 8s + k = 0

106
Root Locus Lecture Notes by A. M. Muhia

s3 1 8
s2 4 k
s1 32−k
4
s0 k

32 − k = 0 kmax = 32 k=0

for stability 0 < k < 32

4s2 + k = 0

4s2 + 32 = 0

s = ±j2.82

system unstable for k > 32

system marginally stable for k = 32
system stable at 0 < k < 32

107
Root Locus Lecture Notes by A. M. Muhia

108
Frequency Domain Analysis Lecture Notes by A. M. Muhia

8 Frequency Domain Analysis

Frequency response is the steady state response of a system to a sinusoidal input.

8.1 Sinusoidal Response of a Linear System

Consider a system as shown in Figure below where

r(t)- input function(sinusoidal)
y (t)- output function
g (t)- Transfer function

c (t) = g (t) r (t)

Taking Laplace transform

C (s) = G (s) R (s)

and

C (s) N (s) N (s)

G (s) = = =
R (s) D (s) (s + a) (s + b) (s + c) + · · ·

N (s)
C (s) = R (s) (1)
(s + a) (s + b) (s + c) + · · ·

Consider r (t) = A sin ωt

Thus

Aω
R (s) =
s + ω2
2

N (s) Aω
C (s) = (2)
(s + a) (s + b) (s + c) + · · · s + ω 2
2

eanding equation (2) in partial fraction form

A1 A2 A3 B1 B2
C (s) = + + + ··· + +
s+a s+b s+c s − jω s + jω

The inverse Laplace becomes

109
Frequency Domain Analysis Lecture Notes by A. M. Muhia

c(t) = A1 exp−at +A2 exp−bt +A3 exp−ct + · · · + B1 expjωt +B2 expjωt (3)

The terms with coefficients A1 to An contribute to decaying eonents and therefore the transient
response. The terms with coefficients B1 and B2 contribute the steady state response and thus

css (t) = B1 expjωt +B2 expjωt (4)

Thus

Aω
C (s) = G (s)
s + ω2
2

The coefficients B1 and B2 are calculated as

 
Aω
B1 = G (s) (s − jω) |s=jω
s2 + ω 2

A
= G (jω)
2j
 
Aω
B2 = G (s) (s + jω) |s=jω
s2 + ω 2

A
= G (−jω)
2j

G (jω) is a complex quantity written as

G (jω) = |G (jω)| expjϕ

Also

G (−jω) = |G (−jω)| exp−jϕ

Putting the values ofB1 and B2 into equation (4)

A A
css (t) = |G (jω)| expjϕ expjωt + |G (−jω)| exp−jϕ expjωt
2j 2j

expj(ωt+ϕ) − exp−j(ωt+ϕ)
" #
= A |G (jω)|
2j

= A |G (jω)| sin (ωt + ϕ) (5)

110
Frequency Domain Analysis Lecture Notes by A. M. Muhia

therefore
css (jω) = |C (jω)|

where

|C (jω)| = A |G (jω)|

= |G (jω)| |R (jω)|

Conclusions

1. The sinusoidal output has same frequency as input

2. The output is sinusoidal and has a phase difference of ϕ and a different magnitude

3. The system merely changes the amplitude and phase of the output with respect to the input

8.2 Methods used in Frequency Response

Frequency response characteristics have two plots

1. Magnitude function

2. Phase function

For both plots, angular frequency ω is the variable

Performance Specifications

Consider a closed loop system of Figure below

The transfer function for the system is given by

C (s) G (s)
M (s) = =
R (s) 1 + G (s) H (s)

C (jω) G (jω)
M (jω) = =
R (jω) 1 + G (jω) H (jω)

1. Response Peak Mr
This is the maximum value of M (jω) as ω is varied It indicates relative stability of the closed
loop system. A larger value of Mr corresponds to large maximum overshoot in a step response

111
Frequency Domain Analysis Lecture Notes by A. M. Muhia

2. Resonance frequency ωr
This is the frequency at which peak response Mr occurs

3. Phase Margin ϕP M
It is a measure of relative stability. Let ϕ1 be the phase angle of a system at unity gain. The
phase margin is given by

ϕpM = 180 + ϕ1

The frequency at which the gain is unity, ω1 is called crossover frequency.

4. Gain Margin
The frequency ω2 where the phase angle of transfer function is −180° is called phase crossover
frequency.
The gain at this frequency is given by|G (jω2 )| |H (jω2 )|
The gain margin is given by

1
Gain M argin =
|G (jω2 )| |H (jω2 )|

The phase margin indicates how much the system angle can be increased to cause the system
to become unstable from a stable condition
The gain margin indicates how much gain can be increased to cause system instability

8.3 Correlation between Time and Frequency Response Specifications

For a canonical second order system

2
ωn
G (s) =
s2 + 2ξωn s + ωn

With usual notation

2
ωn
G (jω) =
(jω)2 + 2ξωn (jω) + ωn

2
Dividing the numerator and denominator by ωn

1
G (jω) =  2  (1)
2ξω

1 − ωωn + j ωn

a) Resonant peak
Let u = ωωn

112
Frequency Domain Analysis Lecture Notes by A. M. Muhia

1
G (jω) = h i
1 − u2 + j2ξu

therefore

1
|G (jω)| = r 2 =M (2)
1−u2 + (2ξu) 2

2ξu
∠G (jω) = tan−1 (3)
1 − u2
For resonant frequency ωr

d
|G (jω)| = 0
du

   
− 1 2 1 − u2 (−2u) + 8ξ 2 u
d 2
|G (jω)| = 3/2 =0
du  2
1−u 2 + (2ξu) 2

−4u + 4u3 + 8ξ 2 u = 0

u2 + 2ξ 2 − 1 = 0

q
u= 1 − 2ξ 2

therefore
ω
q
= 1 − 2ξ 2
ωn
At ω = ωr
q
ωr = ωn 1 − 2ξ 2

1
Mr = r 2
1 − u2r + (2ξur )2

1
= q (4)
2ξ 1 − ξ 2

Mr is a function of the damping factor ξonly

113
Frequency Domain Analysis Lecture Notes by A. M. Muhia

Similarly

2ξu
ϕr = tan−1
1 − u2
√
1 − 2ξ
= tan−1
ξ

Conclusion

1. As ξapproaches zero, Mr approaches infinity and ωr approaches ωn

2. For 0 < ξ < 0.707, Mr is greater than unity and ωr is less than ωn

3. For ξ > 0.707, there is no resonant peak and the maximum value Mr of is unity and ωr = 0

b) Cut off frequency ωc and Bandwidth (BW)

Cut off frequency ωc is the point where M has a value of 0.707.
Usually control systems are low pass filters, hence M = 1 for zero frequency, bandwidth becomes
equal to ωc
Let uc = ωωc
n
Substituting into equation (7)

1
0.707 = r 2
1 − u2c + (2ξuc )2

 2
1 − u2c + (2ξuc )2 = 2

  q
2 2
uc = 1 − 2ξ ± 4ξ 4 − 4ξ 2 + 2

  q 1
2
BW = ωc = ωn 1 − 2ξ 2 ± 4ξ 4 − 4ξ 2 + 2 (5)

Conclusions
From (11) BW is a function of ωn and ξ

1. BW is directly proportional to ωn

2. BW and Mr are directly proportional for 0 < ξ < 0.707

3. As time increases, BW decreases

114
Frequency Domain Analysis Lecture Notes by A. M. Muhia

Example
For a second order system with unity feedback

200
G (s) =
s (s + 8)

Determine the frequency domain specifications

Solution

200
M (s) =
2
s + 8s + 200

2
ωn
s2 + 2ξωn s + ωn

ωn = 14.14 and ξ = 0.283

1
Mr = q = 1.842
2ξ 1 − ξ 2
q
ωr = ωn 1 − 2ξ 2 = 12.96rad/s

  q 1/2
BW = ωc = ωn 1 − 2ξ 2 ± 4ξ 4 − 4ξ 2 + 2 = 30.25rad/s

Example
Consider a closed loop unity feedback system with

k
G (s) =
s (s + a)
Determine the values of k and a that will satisfy the conditions Mr = 1.04 and ωr = 11.55rad/s.
Determine also the settling time and bandwidth.
Solution

M (s) = 2 k
s +as+k

√ a
ωn = k and ξ = √
2 k

1 2k
Mr = q = p
2ξ 1 − ξ2 a 4k − a2

115
Frequency Domain Analysis Lecture Notes by A. M. Muhia

Substituting forMr = 1.04

p
1.04a 4k − a2 = 2k

4k 2 − 4.32a2 k + 1.08a4 = 0

k = 0.686a2 or k = 0.393a2

From ωr = 11.55rad/s condition

q
ωr = ωn 1 − 2ξ 2

1p
= 2k − a2
2

2k − a2 = 266.7

For k = 0.686a2

1.372a2 − a2 = 266.7

a = 26.77 and k = 491.61

For k = 0.393a2

0.786a2 − a2 = 266.7

This does not give real value of a

a = 26.77 and k = 491.61

√
ωn = k = 22.17

and
a
ξ = √ = 0.604
2 k

4
ts = = 0.298
ξωn

116
Frequency Domain Analysis Lecture Notes by A. M. Muhia

  q 1/2
BW = ωc = ωn 1 − 2ξ 2 ± 4ξ 4 − 4ξ 2 + 2 = 28.15rad/s

117
Bode Plots Lecture Notes by A. M. Muhia

9 Bode Plots
The frequency response of any linear system is indicated by plotting two curves or graphs

• Graph 1- The magnitude (M) of the transfer function

The magnitude of the open loop transfer function is plotted in decibels on Y-axis. On the
X-axis frequency is plotted on a logarithmic scale.
The logarithm of the magnitude of the transfer function G(jω) is expressed in decibels as

M = |G(jω)| = 20log|G(jω)dB

• Graph 2- The phase angle (ϕ) as function of frequency (ω)

The phase angle of the open loop transfer function is plotted on the Y-axis. On the X-axis,
same as in graph 1, frequency is plotted on a logarithmic scale

These characteristics are called Bode plots. Both the graphs are plotted in one semilog paper with
same X-axis scale.
Two units used to express frequency ratios are the octave and decade.
f2
An octave is a frequency band from f1 to f2 where = 2. The number of octaves in the frequency
f1
range from f1 to f2 is
log(f2 /f1 ) f
= 3.32log 2
log2 f1
There is an increase of 1 decade from f1 to f2 when f 2/f 1 = 10. The number of decades in the
frequency range from f1 to f2 is

f
log 2 decades
f1

As a number doubles, the decibel value increases by 6 dB. As a number increases by a factor of 10
the decibel value increases by 20 dB.

Advantages of Bode Plot

1. Both low and high frequency characteristics of a transfer function can be shown in one diagram

2. Gain margin and phase margin can be obtained with minimum computational efforts from
the plots

3. The can be plotted with ease using asymptotic approximations

4. It indicates clearly relative stability of the system

5. Data for constructing polar plots and Nyquist plots of complex Transfer function can be
obtained from bode plots

118
Bode Plots Lecture Notes by A. M. Muhia

6. Frequency domain specifications can be obtained easily from bode plots

7. Stability of open loop transfer function can be obtained using bode plots

8. Transfer function of the system can be obtained from bode plots

The frequency transfer function can be written in a generalized form as

N (s) b sm + b1 sm−1 + ... + bm

G(jω) = = 0
D(s) a0 sn + a1 sn−1 + ... + bn

This must be factorized to general type

K1 sg (s + z1 )(s + z2 )...
G(jω) =
sk (s + p1 )(s + p2 )...

NB: either sg or sk is present, not both

Ksg (1 + s/z1 )(1 + s/z2 )...

=
sk (1 + s/p1 )(1 + s/p2 )...
Note:
K1 z1 z2 ...
K=
p1 p2 ...
Putting s = jω
K(jω)g (1 + jω/z1 )(1 + jω/z2 )...
G(jω) =
(jω)k (1 + jω/p1 )(1 + jω/p2 )...
where K is the gain constant or Bode gain.
In general a transfer function can be considered to be a combination of the following terms

• Constant term

• Poles and Zeros at the origin

• Poles and Zeros not at the origin

• Complex Poles and Zeros

9.1 Bode plots of standard factors

Bode gain K or Constant factor

For the magnitude in dB

20logG(jω) = 20logK = K2 dB

is plotted on the Y-axis

For the phase, K is either positive or negative. Thus for positive value of K, the bode plot is a

119
Bode Plots Lecture Notes by A. M. Muhia

straight horizontal line at 0o for all ω and the phase angle is −180o for all ω for negative value of
K
Both the plots are as shown in Figure 72

Figure 72: Magnitude and Phase Plots for Constant factor

1
Poles at the origin of order k or integral factor
(jω)k
Consider
1
G(jω) = = (jω)−k
(jω)k
The magnitude in dB is obtained as

20logG(jω) = 20log(jω)−k

= −20klog(jω) dB

The quantity plotted on the X-axis is log(jω) and on Y-axis in this case is

Y = −20klog(jω) = −kX

This corresponds to a graph of Y = M X where the slope is −20k and is a straight line passing
through the origin given by (X = 0, Y = 0). Since X = log(jω) = 0 at ω = 1
Note:

• ω = 0 is not defined as log(jω) = −∞

• If k = 1, the straight line passes through (0 dB, ω = 1) at slope = −20 db/dec

• If k = 2, the straight line passes through (0 dB, ω = 1) at slope = −40 db/dec and so on

120
Bode Plots Lecture Notes by A. M. Muhia

For the phase

• For k = 1 argG(jω) = −90o

• For k = 2 argG(jω) = −180o

• For k = 3 argG(jω) = −270o

• For k = k argG(jω) = −90k o

Both the magnitude and phase plots are as shown in Figure 73

Figure 73: Magnitude and Phase Plots for Integral factor

Zeros at the origin(jω)g of order g or derivative factor

Consider
G(jω) = (jω)g

The Magnitude in dB is obtained as

20logG(jω) = 20log(jω)g

= 20glog(jω)

This is a plot at a slope of 20g dB/dec

If g = 1, the straight line passes through (0 dB, ω = 1) at slope= 20db/dec
If g = 2, the straight line passes through (0 dB, ω = 1) at slope= 40db/dec and so on

121
Bode Plots Lecture Notes by A. M. Muhia

The Phase is obtained as

For g = 1 argG(jω) = 90o
For g = 2 argG(jω) = 180o
For g = 3 argG(jω) = 270o
For g = 4 argG(jω) = 9g0o

Both the magnitude and phase plots are as shown in Figure 74

Figure 74: Magnitude and Phase Plots for Derivative factor

1
Finite poles at or first order integral factor
(1 + jω/p1 )

Consider
1
G(jω) = ω
(1 + j )
p1
The quantity plotted on the Y-axis is

1 ω −1
Y = 20logG(jω) = 20log ω = 20log(1 + j p )
(1 + j ) 1
p1

ω
= −20log(1 + j )
p1
For ω ≪ p1 , the equation becomes

Y = −20log(1) = 0 dB

For ω ≫ p1 then equation becomes

ω
Y = −20log( ) dB
p1

122
Bode Plots Lecture Notes by A. M. Muhia

ω ω ω
If we change X-axis to log( ) instead of logω then the graph of −20log( ) versus log( ) is a
p1 p1 p1
straight line of slope −20 dB/dec. Then this case can be approximated into two parts.

ω ω
For ≪ 1,the slope= 0 dB and for ≫ 1, the slope= −20 dB. Also the change in slope of
p1 p1
0 dB to −20 dB is assumed to occur at ω = p1 . This can be summarized as follows
ω
• The graph has a slope of 0 dB for ≪1
p1
ω
• The graph has a slope of −20 dB for ≫1
p1
ω
• The change of slope occurs at = 1 i.e. ω = p1
p1
Error in approximation
The actual magnitude curve is
1
|G(jω)| = v
u 2
t1 + ω
u
p21

The actual magnitude is given by

1
|G(jω)| = √
1+1
Therefore 20log|G(jω)| = −3 dB

The phase is obtained as

ω
argG(jω) = −tan−1
p1
As ω becomes larger and larger the phase angle approaches −90o from 0o
Example
Plot the Magnitude and Phase plots for

1
G(jω) = ω
1+
50

Solution
Magnitude plot
|G(jω)| = 0 dB f or ω < 50

= −20 dB/dec f or ω > 50

Phase plot
ω
argG(jω) = −tan−1
50
for ω = 0 argG(jω) = 0o
as ω → ∞ argG(jω) → −90o

123
Bode Plots Lecture Notes by A. M. Muhia

Figure 75: Magnitude and Phase Plots for First order integral factor

Finite zeros at (1 + jω/z1 ) or first order derivative factor

For the Magnitude plot

The analysis has the same treatment as that of first order poles
ω
• The graph has a slope of 0 for ≪1
z1
ω
• The graph has a slope of 20 dB for ≫1
z1
ω
• The change of slope occurs at = 1 i.e. ω = z1
z1
For Phase plot
ω
argG(jω) = tan−1
z1
for ω = 0 argG(jω) = 0o
as ω → ∞ argG(jω) → 90o
Error in approximation
At ω = z1 s
ω2
|G(jω)| = 1+
z1
In dB this gives 3 dB at ω = z1
ω
The magnitude and phase plots for (1 + j ) is as shown in Figure 76
z1

Second order poles

Consider
ωn
G(s) =
2 2
s + 2ζωn s + ωn
Then
2
ωn 1
G(jω) = = ω 2 ω
(jω)2 + j2ζωn ω + ω 2 (1 − ( ) + j2ζ
ωn ωn

124
Bode Plots Lecture Notes by A. M. Muhia

Figure 76: Magnitude and Phase Plots for First order derivative factor

For the magnitude plot

1
|G(jω)| = v"
2 #2 
ω 2
u  
t 1− ω
u
+ 2ζ
ωn ωn

As G(s) is a second order factor approximate, Bode magnitude plot is drawn as

0 dB line for ω ≪ ωn
−40 dB/dec line for ω ≫ ωn
Also do necessary error correction at ω = ωn depending on the value of ζ
The phase angle is given by

ω
2ζ
ωn
argG(jω) = −tan−1
ω 2
 
1−
ωn
Error in Approximation
1
At ω = ωn the value of |G(jω)| =
2ζ

1
∴ |G(jω| in dB = 20log
2ζ

This error depends on ζ

Second order zeros

ω 2
 
ω
G(jω) = 1 − + j2z
ωn ωn

The magnitude is approximated as

0 dB line for ω ≪ ωn
40 dB/dec line for ω ≫ ωn

125
Bode Plots Lecture Notes by A. M. Muhia

The nature of the error has just reversed as compared to earlier case.

The phase is approximated as

ω
2ζ
ωn
argG(jω) = tan−1
ω 2
 
1−
ωn

The phase plots are functions of both ω and ζ.

9.2 Frequency Domain Specifications

Gain Margin (GM)

It is a measure of relative stability and is the reciprocal of the open-loop transfer function, evaluated
at the phase crossover frequency ωpc , the frequency at which the phase is −180o .
The gain margin in dB is the distance on the Bode magnitude plot from the amplitude at the phase
crossover frequency up to the 0 dB point.

Phase Margin (ϕP M )

As a measure of relative stability, it is defined as 180o plus the phase angle ϕ1 of the open loop
transfer function at unity gain. That is

ϕP M = 180 + arg|G(jωgc )|degrees

Where G(jωgc ) = 1 and ωgc is the gain crossover frequency which is the point on the magnitude
plot where the magnitude is unity.

Relative Stability

The relative stability of a system is indicated by the gain and phase margin. The parameters are
easily determined from the Bode plots.
In most cases, positive gain and phase margins will ensure stability of the system.
For stable systems ωgc < ωpc i.e. GM and P M are positive
For unstable systems ωgc > ωpc i.e. GM and P M are negative
When GM is +∞, then the system is absolute stable.
When both GM and P M are zero, then the system is marginally stable
Example

126
Bode Plots Lecture Notes by A. M. Muhia

Figure 77: Gain Margin and Phase Margin

For the system with open loop transfer function

10
G(s) =
s(s + 1)(s + 5)

Determine the stability of the system by plotting the Bode plot of the system

Solution
Writing G(s) in standard form
2
G(s) = s
s(1 + s)(1 + )
5
Substituting for s = jω
2
G(jω) =
jω
jω(1 + jω)(1 + )
5
In the given transfer function, the following four factors are present

1. Constant term k = 2
The magnitude is obtained as
20 log 2 = 6.02 dB
The Phase is 0o
1
2. Pole at the origin
jω

127
Bode Plots Lecture Notes by A. M. Muhia

The magnitude plot is a straight line of slope −20 db/dec crossing the 0 dB line at ω = 1
The phase is −90o
1
3. First order pole The magnitude plot consists of
1 + jω
a line of slope 0 dB/dec for ω < 1 and a line of slope −20 dB/dec for ω > 1
The magnitude plot terminates at ω = 5
The phase is given by −tan−1 ω
1
4. First order pole ω The magnitude plot consists of
1+j
5
a line of slope 0 dB/dec for ω < 5 and a line of slope −20 dB/dec for ω > 5
The magnitude plot terminates at ω = ∞
ω
The phase is given by −tan−1
5

Preparing a table for the magnitude plot

Factor Resultant slope Start point (ω) End point (ω)
k=2 straight line at 6.02 dB 0.1 ∞
1
−20 dB/dec 0.1 1
jω
1
−40 db/dec 1 5
1 + jω
1
ω −60 dB/dec 5 ∞
1+j
5
The resultant phase is obtained as

ω
ϕ(ω) = −90 − tan−1 (ω) − tan−1 ( )
5

Preparing the phase angle table

ω
ω −900 −tan−1 ω −tan−1 ϕ(ω)
5
0.01 −900 −0.57 −0.11 −90.68
0.05 −900 −2.86 −0.57 −93.43
0.1 −900 −5.71 −1.15 −96.86
0.5 −900 −26.56 −5.71 −122.27
1 −900 −45.00 −11.31 −146.31
10 −900 −84.29 −63.43 −237.72
100 −900 −89.42 −87.14 −266.56
500 −900 −89.89 −89.42 −269.31
1000 −900 −89.94 −89.71 −269.65
The bode plot is as shown in Figure 78
From the Bode plot Gain margin = 12 dB

128
Bode Plots Lecture Notes by A. M. Muhia

Figure 78: Bode Plot

Phase margin = 15o

Gain crossover frequency ωgc = 1.4 rad sec

Phase crossover frequency ωpc = 2.7 rad sec

ωgc < ωpc and therefore the system is stable

129