Control System Engineering-I

Course: B. Tech (EE)

Sem: 5th

Teacher: Dr. B. Rout

Electrical Engineering Department
VSSUT, BURLA

1 Monday, August 21, 2023

 Syllabus
Control System Engineering I ( Course code: UPC)
MODULE-I
Scope of control system Engineering, Various Classification of Control System, Closed
Loop Control Versus Open Loop Control, Mathematical model of physical systems,
transfer function, block diagram algebra, signal flow graph (SFG), Mason’s gain formula.
Feedback Characteristics: Types of feedbacks, effect of degenerative feedback on control
system, regenerative feedback, Application of Control system to engineering and non-
engineering problem.
MODULE-II [10 hrs]
Time domain analysis: Standard test signals: Time response of 1st. order systems to unit
step and unit ramp inputs. Time response of second order systems to unit step input. Time
response specifications. Steady state errors and error constants of different types of
control systems Generalized error series method, Application of MATLAB and its Tool
Box for time response analysis. Effect of poles and zeros on system response.
MODULE-III [8 hrs]
Concepts of stability: Necessary conditions of stability, Hurwitz stability criterion, Routh
stability criterion, application of Routh stability criterion to linear feedback systems,
Relative stability Analysis. Root locus techniques: Root locus concepts, rules for
construction of root loci, determination of roots from root locus, root contours, systems
2with transportation lag, Root locus Monday, August 21, 2023
plots with MATLAB.
MODULE-IV
Frequency domain analysis: Introduction, Polar plots, Bode plots, determination of stability
from Bode plots, Nyquist stability criterion, application of Nyquist stability criterion to
linear feedback systems, Log magnitude versus phase plots, Use of MATLAB for plotting
Bode & Nyquist diagram. Closed loop frequency response: Constant M circles,
constant N circles, use of Nichols chart.
MODULE-V [10 hrs]
Controllers: Proportional, derivative and integral control actions, PD, PI and PID
controllers and their applications to feedback control systems, PID controller gains
tuning by Zeigler- Nichols method. 2-Degree-of-freedom control.
Sensitivity transfer functions (S and T) and their significance: Measure of loop
robustness in terms of the peaks of sensitivity and transfer functions for any PID
compensated system. Ref. Azzo, chapter-14
[6 hrs]
TEXT BOOKS
•K. Ogata, “Modem Control Engineering”, PHI Publishers.
•I.J. Nagrath, M. Gopal, “Control Systems Engineering”, New Age International Publishers.
•Norman S. Nise, “CONTROL SYSTEMS ENGINEERING”, John Wiley & Sons.
Reference Books
•Dorf Bishop, “Modern Control Systems, Pearson International Edition
•John D Azzo, linear Control System Analysis and Design with MATLAB, Marshel
Dekkar Publisher
3 Monday, August 21, 2023
•.
 Module_1 Contents:
Lecture_1 Scope of control system Engineering, Various Classification of Control
System, Closed Loop Control Versus Open Loop Control, Control systems components
Lecture_2 Mathematical model of physical systems, Laplace transfer function,
Properties of T.F
Lecture_3 Transfer function Mechanical system, of translational and rotational
motion, Analogy of Force voltage and Force- current
Lecture_4 Transfer function Electrical system, Transfer function Electronics
system,
Lecture_5 T.F. of electro-mechanical systems, Numerical
Lecture_6 Analogy of Force-voltage and Force-current, Numerical

Lecture_7 Rules for block diagram reduction, system representation in block diagram,
Numerical
Lecture_8 Signal flow graph (SFG), Mason’s gain formula, Numerical

Lecture_9 Comparison between BDR and SFG , Numerical SFG from block
diagram, Types of feedbacks,
Lecture_10 Effect of degenerative feedback on control system, regenerative feedback,
Application of control system in non-engineering field
4 Monday, August 21, 2023
 Module_2
Lecture_11 System Analysis in time domain, Standard test signals-
Characteristics of step, ramp, parabolic , impulse and sinusoidal signal:

Lecture_12 Time response of 1st order systems to unit step and unit ramp inputs,
impulse and parabolic
Lecture_13 Time domain specifications, numerical

Lecture_14 Error constants of different types of control systems, type-0,type-1,type-2,

Numerical

Lecture_15 Generalized error series method, Relation between the static and
dynamic error coefficient Numerical
Lecture_16 Time domain specification numerical, steady state error numerical

Lecture_17 Application of MATLAB and its Tool Box for time response analysis

Lecture_18 Effect of poles and zeros on system response.

5 Monday, August 21, 2023

 Module_3

Lecture_19 Concepts of stability: Necessary conditions of stability

Lecture_20 Hurwitz stability criterion, Routh stability criterion,

Lecture_21 Relative stability Analysis, Application of Routh stability criterion to linear

feedback systems,

Lecture_22 Evan condition ,Root locus techniques. Properties of root-locus

Lecture_23 Construction of root locus: Numerical

Lecture_24 Determination of roots from root locus, root contours.

Lecture_25 Systems with transportation lag

Lecture_26 Plots with MATLAB, Numerical

6 Monday, August 21, 2023

Module_4
Lecture_27 Frequency domain analysis: Introduction, Relation between the time
domain and frequency domain specifications
Lecture_28 Advantages and disadvantages of frequency domain, Frequency response
specifications.
Lecture_29 Bode plot for various factors in the system, determination of stability from
Bode plot
Lecture_30 Polar plot for different types of open loop system, stability and relative
stability from polar plot
Lecture_31 Nyquist stability criterion, application of Nyquist stability criterion to
linear feedback systems.
Lecture_32 Use of MATLAB for plotting Bode & Polar plot

Lecture_33 Constant M circles, constant N circles

Lecture_34 Log magnitude versus phase plots

Lecture_35 Numerical, MATLAB for Nyquist plot

Lecture_36 Closed loop frequency response, Use of Nichols chart

7 Monday, August 21, 2023
 Module_5

Lecture_37 Introduction to Controller, types of controller, Proportional, derivative and

integral control actions, advantages and disadvantages of P,I,D controllers , PID
controller gains tuning by Zeigler- Nichols method, 2-Degree-of-freedom control.

Lecture_38 Sensitivity transfer functions (S and T) and their significance

Lecture_39 Measure of loop robustness in terms of the peaks of sensitivity

Lecture_40 Transfer functions for any PID compensated system.

8 Monday, August 21, 2023

 Module_1

Lecture_1

Scope of control system Engineering

Various Classification of Control System
Advantages of control system
Closed Loop Control Versus Open Loop Control
Control system components

9 Monday, August 21, 2023

 Scope of Control System Engineering
Control systems are an integral part of modern society. The various applications
of automatic control system all around us are:
The space-vehicle system, missile-guidance system, robotics system, motor-
drive system, domestic appliances such as in air conditioning control,
refrigerator control, various modern manufacturing industries in designing the
automobiles and in control of pressure, temperature, humidity, viscosity and
flow rate in process industries.
• The subject of control engineering is interdisciplinary in nature. It embraces all
the disciplines of engineering including Electronics, Computer Science,
Electrical Engineering, Mechanical Engineering, Instrumentation Engineering,
and Chemical Engineering, Production Engineering etc.

• If we are interested to control the position of a mechanical load automatically,

we may use an electrical motor to drive the load and a gearbox to connect the
load to the motor shaft and an electronic amplifier to amplify the control signal.

• The control system engineering may also be extended to non-engineering field

such as: biological system (pancreas regulates blood sugar), economical system,
10 Monday, August 21, 2023
social system, nature’s behavior scenario, political system as well.
What is meant by a system:
• A collection of components or procedures which are included in a specified
boundary and together performs an intended job or objective.
• In control system, A system gives an output(called response) for an input
(called excitation.)
• A system may have many inputs and many outputs called MIMO system
(more than one) Ex: Ship Movement, Motor

What is control system:

•A system (plant) or mechanism which directs the input to the other system or
itself that will provide a desired response based on laws of physical world is
called control system.
• Control system alters the response of a plant or system as desired.

11 Monday, August 21, 2023

Some times the system may consist of combination of few sub systems, but goal is that for
certain input, there must be desired output. Here 4 subsystems are interconnected.

12 Monday, August 21, 2023

 • There are some inputs on which the engineer have direct control over it and
can be used to control the output. These are called as a control inputs.
• The other type of inputs which has no control over it and these will deflect
the outputs from their desired value. These are called as disturbance input.
A ship dynamics has two control inputs such as rudder and engines. The
desired output is position and forward velocity. The disturbances are wind,
waves and current that gives undesired ship motion. (MIMO System)

Ex: MISO System

The block diagram

can be shown as

13 Monday, August 21, 2023

Various kinds of Control Systems
Motor
Input- Electrical Energy (voltage)
Output- Mechanical Energy (torque,
speed)
Air-Conditioner

Input- Electrical Energy (voltage)

Output- Heat Energy ( changes of ambient
temperature)
Touching an moving object

Input- object desired position

Output- hand motion to grapes
object

Ceiling fan with regulator

Input- Voltage
Output- speed
14 Monday, August 21, 2023
 Advantages of Control System

Power amplification
Ex: a radar antenna, positioned by the low-power rotation of a knob at the
input, requires a large amount of power for its output rotation. A control system
can produce the needed power amplification, or power gain.

Remote control
Control systems are also useful in remote or dangerous locations. For example,
a remote-controlled robot arm can be used to pick up material in a
contaminated radioactive environment.

Convenience of input form

Control systems can also be used to provide convenience by changing the form
of the input. For example, in a temperature control system, the input is a
position on a thermostat. The output is a heat. Thus a convenient position of
input yields a desired thermal output.

Compensation for disturbances

15
Control system has the ability to compensate the for disturbances.
Monday, August 21, 2023
 Classification of Control Systems:
Classical Control (1940-60) and Modern control 1960 onwards
Linear and non-linear control system
Open-loop control system and Closed-loop control system:
SISO and MIMO control systems
LTI and LTV Control system
Static and Dynamic Control system
Causal and Non-causal System
Classical Control System deals with frequency domain model,
Mathematical Tools: Laplace Transform
EX: All transfer function model, Bode Plot, Nyquist Plot (Frequency domain model)

Root Locus method (Time domain analysis)

Modern Control deals with time domain model on both continuous and discrete
systems.
Mathematical tools: Vector –Matrices tools, Z-Transform

Ex: State Variable analysis (continuous and discrete) applied on Digital

computers
16 Monday, August 21, 2023
 Linear and Nonlinear systems:
Linear Systems Non-Linear System

• The output does not vary with input

• Output of the system varies linearly with input
• Do not satisfy the homogeneity and
• Satisfies the homogeneity and superposition principle
Superposition principle

Superposition does not hold

Superposition Principle

f(x1+x2)= f(x1) + f(x2)

17 Monday, August 21, 2023

 Linear System

f(x1+x2)= f(x1) + f(x2)

For taking individual sources of V1 and V2 and then add together

We get

18 Monday, August 21, 2023

 Homogeneity: F(ax)= a f(x)

V1 is scaled by constant a and V2 is scaled by constant b

Similarly we can take for inductance also.

19 Monday, August 21, 2023

 Open-loop control system and Closed-loop control system
The control input is not influenced by system output.
It is a control system where its control action only depends on input signal and
does not depend on its output response.
•The output is neither measured nor fed back to the reference input for
comparison.
•In the presence of disturbances, an open loop control system does not perform
the desired output and thus unreliable.
Examples: control of rotating disc, traffic signal without police , washing
machine, bread toaster, ceiling fan regulator etc
EX: control of rotating disc
Block Diagram

20 Monday, August 21, 2023

 Problem in open loop: The controlled variable is sensitive to changes in
disturbance inputs.
For example, if the controller is an electronic amplifier and disturbance 1 is noise, then
any additive amplifier noise at the first summing junction will also drive the process,
corrupting the output with the effect of the noise.
The output of an open-loop system is corrupted not only by signals that add to the
controller's commands but also by disturbances at the output or disturbances in plant itself.
The system cannot correct for these disturbances, either and are simply commanded
by the input.

21 Monday, August 21, 2023

 In an examination, three chapters has to covered in a specific time to get an
grade A. If the professor adds a fourth chapter just before the exam —a
disturbance—the student may not get the desired grade. This extra chapter is a
disturbance for the student.
Other examples of open-loop systems are mechanical systems consisting of
a mass, spring, and damper with a constant force positioning the mass. The
greater the force, the greater the displacement. Again, the system position
will change with a disturbance, such as an additional force, and the system
will not detect or correct for the disturbance.

The controlled variable (output) of a toaster is the color of the toast. The
device is designed with the assumption that the toast will be darker the
longer it is subjected to heat. The toaster does not measure the color of the
toast; it does not correct for the fact that the toast is brown, white, or dark. A
closed-loop toaster oven is more complex and more expensive since it
has to measure both color (through light reflectivity) and humidity inside
the toaster oven.

22 Monday, August 21, 2023

 Closed-loop control system:
It is a control system where its control action depends on both of its input
signal and output response.

Closed loop control of rotating disc

Block diagram representation

23 Monday, August 21, 2023

 The input transducer converts the form of the input to the form used by the
controller. An output transducer, or sensor, measures the output response and
converts it into the form used by the controller. For example, if the controller
uses electrical signals to operate the valves of a temperature control system, the
input position and the output temperature are converted to electrical signals.
Closed-loop systems, then, have the obvious advantage of greater accuracy than open-
loop systems. They are less sensitive to noise, disturbances, and changes in the
environment.

24 Monday, August 21, 2023

 Automobile steering control

Examples: Automobile steering control, automatic electric iron,

missile launcher, speed control of DC motor, steel rolling mills etc.
25 Monday, August 21, 2023
26 Monday, August 21, 2023
27 Monday, August 21, 2023
28 Monday, August 21, 2023
29 Monday, August 21, 2023
30 Monday, August 21, 2023
 Major components of closed loop/feedback system

Controller The error detector with amplifier acts as controller here. The weak
amplifying signal is strengthen by actuator. The manipulating signal from the
actuator is sent to the controlled system and results controlled variable.
Actuator: It is a power device which may be pneumatic motor or valve, hydraulic
motor or an electric motor (servo motor)

Sensor/Transducer: It is a measuring device that converts the output variables to

another suitable variable used to compare with reference variable.
31 Monday, August 21, 2023
 Advantages of Closed loop System

32 Monday, August 21, 2023

 Dis-advantages of Closed loop System

33 Monday, August 21, 2023

 Module_1

Lecture_2

Control system Problems

Mathematical model of physical systems

Laplace transfer function, Properties of

T.F

Numerical

34 Monday, August 21, 2023

 Control Engineering is concerned with basic six problems
 The identification problem: to measure the variables experimentally and convert data
for analysis if the system is very complex.
 The representation problem: to describe the dynamic of the system by an analytical
form or model based approach and knowledge based approach.
The solution problem: to determine the above system model response
The stability problem: general qualitative analysis of the system
The design problem: modification of an existing system or develop a new one
The optimization problem: from a variety of design to choose the best.

Two basic approaches to solve these six problems are conventional and modern approach

Conventional Approach is based on complex function theory. (frequency domain, Bode,

Nyquist, )
Modern Approach is based on the state variable theory, Root Locus Method
Representation Problem (model based approach)
Mathematical Modeling:
Differential equation/Difference equation) ( continuous time system/ discrete time system)
Differential equation and the state variable representation are very popular
mathematical models for describing the dynamics of a continuous control system.
35 Monday, August 21, 2023
The system dynamics are related by its input and output variables with constant
coefficient following the physical laws. Consider a first order system where its
input variable is x(t) and output variable is y(t).
Hence the dynamics of the linear time invariant system is given by its differential
equation as dy (t )
a  by (t )  cx(t )
dt

2
For second order linear time invariant system a y  b dy  cy  ex (t )
d
dt 2 dt
For third order linear time invariant system
d3y d2y dy
a 3  b 2  c  ey  fx(t )
dt dt dt

 Laplace Transfer function ( since differential equations is difficult to represent a

system, this can be transformed to algebraic Equations, frequency domain)
The transfer function model is applicable if the system is linear system
 If the initial conditions, multi input and multi outputs of the system are
36taken into account, then state space model is required which is time
Monday, domain
August 21, 2023
analysis approach.
 What is a Model ?
• Model is a mathematical presentation of a plant or system dynamics.
•Model helps in the analysis (Input-Output) of the system
•It captures the dynamic of the system.
•Dynamics refers to evolution of system variables

For example:
The room temperature when the AC is switched on (how the
temperature behaves or how fast or how slow the desired temperature is
obtained when the AC switch is on.)
 How the speed of a car changes before it reaches the desired speed
How the voltage or current evolves when an inductor is present in an
electrical circuit .

Plant or System Equivalent

Mathematical
Model
37 Monday, August 21, 2023
 Types of Mathematical Modeling

Differential • Dynamics of the system represented in terms of

Equation differential equations
Model •Time domain representation of the system

Transfer • Dynamics represented in terms of a Laplace transform

Function expressions
Model •Frequency domain representation of the system

• State is a set of system variables that describes the system

behavior in conjunction with inputs
State Space
• Dynamics are represented by a set of first order differential
Model
equations using these state variables.
• Time domain representation of the system
•MIMO and LTV is accommodated in this model for
38
analysis. Monday, August 21, 2023
 A linear R-L-C circuit
Example:

di 1
Applying KVL V  Ri  L   idt
dt C
Differential
Equation To know the highest
Model order of the system
di 1
dt C 
V  Ri  L  idt
Transfer
Function
Model By simple writing Laplace
transformation

State Space
Model

39 Monday, August 21, 2023

 The system is very difficult to represent in differential equation form, rather
the differential equation is transformed into an algebraic equation in a complex
variable s. All derivative and integral terms in differential equations are
replaced by algebraic form.

s   function
Any j of complex variable is called complex function. For Ex
G(s)=A+jB since
If a complex function G(s) together with its derivatives exist in a region, it is
said to be analytic in that region.

For ex:

is found to exist everywhere, except at s = – 2. So, it is an analytic function at

all points in the s plane except the point s = – 2.

All the points in the s plane at which the complex function G(s) is found to be
analytic are called ordinary points, whereas the points at which it is not analytic
are called singular points. The terms pole and zero are used to describe two
different types of singular points. The singular points are called poles of the
40 complex function G(s) if the function or it derivatives approach infinity
Monday, at 2023
August 21, these
points.
 The singular points at which the function G(s) equals zero are called zeros. The complex
function G2(s) given by

has zeros at s = – 4, s = – 6, simple poles at s = 0, s = – 1, and a pole of order 2 at s = –

10.
Laplace Transformation

The reverse process of finding f(t) is called inverse Laplace transformation

41 Monday, August 21, 2023

Laplace Transform of common functions

42 Monday, August 21, 2023

Similarly, Laplace Transform of other functions

It is seen that, the exponential function produces a single pole s-plane. For convergence
The real part of the complex s must be greater than alfa
Step Function:

The step function is undefined at t=0

The Laplace transform is given by

The step function whose height is 1 called unit step function.

43Physically, a step function at t=0, corresponds to a constant signal suddenly applied at t=0
 Ramp Function

Laplace
Transform

Sinusoidal

Since,

44 Monday, August 21, 2023

45 Monday, August 21, 2023
 Laplace transform of certain other function

46 Monday, August 21, 2023

 Initial value theorem:

Final value theorem

Transfer function

The block diagram represents the transfer function of the system

47 Monday, August 21, 2023

 General form of Transfer
Function
The transfer function of an nth order single-input, r(t), single output, c(t), linear
time invariant system can be written in the form :
d m r (t ) d m 1r (t ) d n c(t ) d n 1c(t )
bm m
 bm 1 m 1
 ....  b0 r (t )  an n
 an 1 n 1
 ....  a0c(t )
dt dt dt dt

Where, a’s and b’s are system parameters.

The required transfer function of the dynamic system is given by

n m

Thus, the system dynamics can therefore, be represented by an algebraic

equation in s domain by using the concept of transfer function.

48 Monday, August 21, 2023

 Properties of T.F.
Derivative properties

Linearity:

Constant Multiplication

Right shift theorem

Convolution Integral

Limitations:

• It gives no information about the actual structure of the system

• Applicable only to LTI System
• It is a mathematical model and gain of the system
• Impulse response of the system is the transfer function of the system itself
49 Monday, August 21, 2023
 Transfer Function and its relation with impulse Response

In control system, the transfer function is given by C (s)

 G ( s)
R( s )

 C ( s)  R( s) * G( s)
If the input is unit impulse at t=0, then, the output is

C ( s)  G( s) Since, R( s)  1

Taking the Laplace inverse both sides, we get the output time response as

L1C ( s)  L1G ( s)

 c(t )  g (t )

Thus, the inverse Laplace of transfer function of any system is G(s) is called impulse response
of the system itself and is independent of the input function.

50 Monday, August 21, 2023

 Procedure for determining the Transfer function of Control System
The following steps are required

 Formulate the mathematical model or equation of the system adopting physical laws
 Formulate the dynamical equation by algebraic equation b using Laplace transformation
neglecting the initial condition of the system
 Identify the system output and input
 Take the ratio of Laplace transform of output to its Laplace transform of the input
The ratio gives the required transfer function of the particular system
Take Laplace inverse to find total solution.

51 Monday, August 21, 2023

 Numerical:
Find the Laplace transform of the differential equation:
(a) With initial Condition

(b) No initial condition

Solution: (a) Including initial conditions

Taking Laplace transformation

using

(b) No initial condition

Taking Laplace transformation

52 Monday, August 21, 2023

 Solution (b) is more easily solved considering no initial conditions.

Hence the input-output relation is represented as

General form

Example.

53 Monday, August 21, 2023

 Using Laplace Transform:

54 Monday, August 21, 2023

 Module_1

Lecture_3

Modeling of physical Systems

Transfer function Mechanical system (translational and rotational
motion)

55 Monday, August 21, 2023

 Physical Systems:
• Physical systems can be classified into various types:
Mechanical Systems
Electrical Systems
Electronics Systems
Hydraulic Systems
Thermal Systems
Pneumatic System

• Each of the system can be modeled in terms of certain basic elements following
physical laws
• Basic elements of all physical systems can be shown to be analogous

Mechanical Systems: Classification based on type of motion:

 Translational systems having linear motion

Rotational systems having angular motion about its fixed

56axis. Monday, August 21, 2023
 Modeling on mass. Mass is an property of an element that stores the K.E. energy due to
translational motion

When a force is acting on a body of mass M, it causes a displacement x.

ince the momentum P is proportional to velocity V, we get P=MV, where, M is mass which is consta

dP d ( MV ) d 2x
F   M 2  Mx Newton’s second law.
dt dt dt
57 Monday, August 21, 2023
 Modeling on Damper: B

Damper is an element of mechanical system, that generates a force which resists the
translational and rotational motion. Example: friction and despot.

The damping force F is linearly proportional to velocity

Modeling on linear spring K:

The linear spring constant is the property of an element that stores the potential energy
due to
translational motion.

When a force is applied on spring of constant K , it causes a restoring forces of

which is proportional to the displacement.

In mechanical system, mass ,spring constant and damper are the common system parameters whic
can be represented by an equivalent mechanical as well as electrical circuit

The system structure analysis may be done to suit nodal analysis like electrical circuit in mechanica
domain.
58 Monday, August 21, 2023
 Mechanical System: Free body diagram concept

Writing Newton’s law of motion,

FBD

Taking Laplace transformation assuming zero initial conditions, we get

Shows the denominator is the

characteristic equation.
This Equation gives the roots or poles
of the system which decides the transient
59 Monday, August 21, 2023
behavior of the system.
60 Monday, August 21, 2023
 By Nodal Analysis: (Translational Motion), single displacement

Steps:
• Select the non reference node “n” from which displacement occurs and select one reference nod
•All the mechanical elements such as mass, spring constant and dampers are connected between
selected node “n” and reference node.
•Applying Nodal analysis, at point “n” we have Newton’s second law of motion as

Taking Laplace transformation assuming zero initial conditions, we get

61 Monday, August 21, 2023

 Nodal Analysis in mechanical Domain (double displacement):

Steps: 1 Identify the number of non reference nodes

The number of nodes depend on the number of displacement on the system. Here there are two
Displacements x1 and x2

Step-2: Displacement and reference nodes are identified.

62 Monday, August 21, 2023

 Step-3 Connect mass M between the
node x2 and reference node

Step 4: Connect spring K between the node x1 and x2

Step 5: Connect force F between the node x1 and reference node

Step 6: Apply Newton’s second law at F=K(x1-x2) (1)

node x1
Step 7: Apply Newton’s second law at node x2. At node 2 sum of the forces are
equal to zero.
K ( X 2  X 1 )  MX 2  BX 2 (2)

Equation (1) and (2) are the mathematical model of the mechanical system.

63 The Monday, August 21, 2023

transfer function may be found out as G(s)=X2(s)/F(s) =1/(Ms^2)+Bs
 Mechanical Equivalent network (with gain representation)

Procedures:

 locate the points xa and xb and the reference.

The network elements are then connected between these
points.
The displacements xa and xb are nodes of the mechanical
circuit.
for
At each
node node
at xathe sumf of the
f k forces
K ( xamust
 xb )add to zero(1
) d
As D
for node at xb f k  f M  f B  MD 2 xb  BDxb (2) dt

64 Monday, August 21, 2023

 Since f  f M  f B 
 f  MD 2 xb  BDxb  xb MD 2  BD 
xb 1
 G (3)
f MD 2  BD

Since f  K ( xa  xb )  MD 2 xb  BDxb  Kxa  xb ( MD 2  BD  K )

f
 Kxa  2
(MD 2  BD  K ) From (3)
MD  BD

xa MD 2  BD  K
  2
 G1 (4)
f K ( MD  BD)

Also, since Kxa  xb ( MD 2  BD  K )

xb K
   G2 (5)
xa MD 2  BD  K

xa xb xb
Now multiplying (4) and (5), we get G  G1G2 G  
f xa f

65 Note: From (5) the motion xb resulting from a given Monday, August 21, 2023
motion xa
 From (3) and (4) it is seen that, the motions xa and xb, respectively, resulting from a
given force f.
The signal xa is unaffected by the presence of the block having the transfer function
G2;
Transfer Function—Two Degrees of Freedom

Ex: Find the transfer function, X2(s)/F(s), for the system

viscous friction shown here. All

other frictions are neglected.

Solution: By Nodal analysis:

Here there are two displacements x1 and x2. Hence two number of non reference nodes and one
Reference node.
Connect all the elements as per given figure w.r.t the nodes.
66 Monday, August 21, 2023
 After writing the dynamics at node 1 and 2 applying nodal method, and then by
using
Laplace transformation, we get
At node x1

At node x2

Z X   F 
where

67 Monday, August 21, 2023

 Z X   F 

68 Monday, August 21, 2023

 FBD

If we hold M2 still and move M1 to the right, we see the forces shown in
Figure (a). If we hold M1 still and move M2 to the right, we see the forces
shown in Figure (b).

The total force on M1 is the superposition (c)

The resultant force for mass M1 is

(1)

69 Monday, August 21, 2023

 Similarly, for mass M2

The total force on M2 is the superposition (c)

The resultant force for mass M2

is
(2)

Solving (1) and (2) we get

70 Monday, August 21, 2023

 Rotational system
Also notice that the term associated with the mass is replaced by inertia. The
values of K, D, and J are called spring constant, coefficient of viscous friction,
and moment of inertia, respectively.
In rotational body the torque due to moment of inertia
is proportional to angular acceleration

Similarly including the torque due to damping

and spring stiffness

We get
71 Monday, August 21, 2023
 Considering one degree of freedom and taking initial condition into account, the
behavior of the system is studied as follows.

Taking Laplace transform

For transient response f (t )  0

x(0 )  x0
when

We have

Solving for X(s), we obtain

The roots of numerator is the zeros of the system and the roots of the denominator is the
poles of the system.

The zeros of the system is

K  s  3 x0
Let us consider 2 and Then, X (s) 
M  s  1 s  2 
72 Monday, August 21, 2023
 The pole – zero location is s-plane is

By partial fraction expansion A B

X ( s)  
s 1 s  2

Setting x0  1 we found x0  1  s  1 s  3  s  2  s  3

A 2 B  1
 s  1 s  2  s 1
1
 s  1 s  2  s 2
1

 2 1 
The solution x(t )  L1  X ( s)   L1   t
  2e  e
2t
 s 1 s  2 

It is desired to know the final value or steady state value using final value
theorem
lim x(t )  lim sX ( s)  0
t  s 0

Hence, the final position of the mass is at the equilibrium point x=0
73 Monday, August 21, 2023
 Further the system can be studied considering the damping ratio and the natural frequency

Where,  is the damping ratio and  n is the natural frequency

The roots of the characteristic equation is s1 , s2  n  n  2  1   1

K
n 
Where, M and

When   The
1 roots are real and over damped i.e. transient response will take large
time to settle

  The
1 roots are complex conjugate and under damped i.e. transient response
will damped oscillation

  The
1 roots are repeated and real i.e. critical damped so, the transient response
will damped oscillation
The roots are on imaginary axis and undamped i.e. transient response
  0
will give sustained oscillation
74 Monday, August 21, 2023
 s1 , s2  n  n 1   2  1

The response is under-damped. The plot for the zeros and poles of X(s) are show shown.

1
Where,   cos 

As  varies keeping n constant

The complex conjugate roots follows a

circular path of radiusn

The damping ratio and natural frequency

decides the nature of the transient response
for the second order system.

75 Monday, August 21, 2023

The solution for under damped system is

x(t ) 
x0
1  2
  1   t   
e n t sin n 2

Adjusting the n it varies envelop ent hence, the response x(t)

The larger the value ofn the faster the response converges to zero

This shows, the response x(t) depends on the initial condition of the system. The
transfer function only sees the relation between input and output not the internal
76structure of the system. Monday, August 21, 2023
 Module_1

Lecture_4

Transfer function Electrical system

Transfer function of electronics
system
Numerical

77 Monday, August 21, 2023

 The basic building elements and variables:

Electrical System Elements:

The electrical system can be modeled in terms of certain basic elements following
physical laws
The Basic elements are modeled as followings
V=IR
Resistor R: It is an element which resists the flow of current in an electrical systems.

78
Inductor L: It is an element that stores electrical energy in a magnetic field.

Capacitor C: It is an element that stores electrical energy in an electric

field.

Parallel Circuit: Nodal Analysis

By KCL

79 Monday, August 21, 2023

 Substituting

So, in current source circuit, the basic system variables are current and flux.

Series Circuit By KVL

substituting

So, in voltage source circuit, the basic system variables are voltage and charge.
80 Monday, August 21, 2023
 R-L-C series di 1
circuit
V  Ri  L 
dt C
idt 
Using Laplace
transform

T.F. of the system is

Vc (s) 1 / LC
  2
V (s) s  ( R / L ) s  (1 / LC )

Vc (s) n2 1 R
  2 Where, n  
V (s) s  2n s  n2 LC 2 L/C
81 Monday, August 21, 2023
 R-C series Circuit
The dynamic of the circuit is
1 I (s)
Ri 
C 
idt  V1  RI ( s) 
sC
 V1 ( s)

 1   1 
 V1 ( s)  I ( s)  R    V1 ( s)  I ( s)  R  
 sC   sC 
 1 
The output voltage isV2 ( s)  I ( s)  sC 
 
V2 ( s ) 1 1 1/ 
Taking the ratio V2 to V1, we getG ( s )     1st
V1 ( s) RCs  1  s  1 ( s  1/  )
order
Where,   RC called time constant and
s  1/  is the single pole at real
axis

Ex: Find E0(s)/Ei(s)

82 Monday, August 21, 2023

 The dynamics of the system
is (1)

(2)

where (3)

Taking Laplace Transform, we get

(4)

(5)

(6)
Eliminating I1 from (1) and (2) and
writing Ei (s) in terms of I2, we get
T.F.
(7)

83 The two roots in denominator

Monday, Augustare real
21, 2023
 FIND

Solution
Applying Lalace transform into the system, we
get

For mesh 1
(1)

(2)

84 Monday, August 21, 2023

 Combining eq(1) and (2), we
get

Using Cramer’s rule

85 Monday, August 21, 2023

86 Monday, August 21, 2023
Electronics
System:
OPAMP is the basic building block in electronic circuit for transfer function. The
OPAMP are active element and has high gain when operating in the linear region. This
OPAMP amplifies the signal Used as filter in compensation purposes. It has two input
terminal, the positive terminal is called non-inverting and negative terminal called
inverting terminal. This has also one output terminal .and two terminals for power
supplies.
Ideal OPAMP Characteristics
 The input impedance is high. so, no current flows into either of the input
terminals
The output impedance is low.
 The output voltage is not affected by the load connectedeacross theterminal.
2  e1  0 e2  e1
 The difference between the two input voltages is nearly zero.
e0
A 
e2  e1
Amplification gain

An ideal
OPAMP

87 Monday, August 21, 2023

 Example: ei  e' e'  e0
i1  , i2  ,
R1 R2

Since, negligible current flows into input terminali1  i2

ei  e' e'  e0

R1 R2

Inverting
since e0  A(e2  e' ) '
 e0   A(e  e2 ) opamp

Since e2  0  e0   A(e' )

As A e'  0

ei e e0 R
   0    2
R1 R2 ei R1

88 Monday, August 21, 2023

89 Monday, August 21, 2023
90 Monday, August 21, 2023
 Find for the following
circuits

Example:

Non-inverting amplifier Equivalent circuit

91 Monday, August 21, 2023

 Sol:

92 Monday, August 21, 2023

 Differential Circuit
Integrator circuit

93 Monday, August 21, 2023

 Module_1

Lecture_5

T.F. of electro-mechanical systems

T.F. of Armature control of DC servomotor
T.F. of field control of DC servomotor
Numerical

94 Monday, August 21, 2023

95 Monday, August 21, 2023
96 Monday, August 21, 2023
 Electro-Mechanical System:
This system is the combination of electrical and mechanical system. The electric motor
is coupled with the mechanical load through bearing and gear arrangement. The DC
servomotor is an high power actuating device used to provide high torque in the
application of motor-drive control, robotic control, machine tools etc. The transfer
function of this motor is developed by approximation to an actual motor. Neglecting the
hysteresis and voltage drop across the brush terminals. The T.F. is obtained by applying
the voltage either to armature or field coils. The field needs to be unsaturated.
In armature-controlled D.C. motor,
field is constant

d

dt
97 Monday, August 21, 2023
 The time domain dynamic of the combined
system is
The motor torque Tm   ia

Since field current is constant

(1)
d d
The motor back emf E b   E b  
dt dt

(2)

The electric dynamic (3)

The motor-load torque dynamic

(4)
neglecting the disturbing torque

However, the time domain dynamics are not desirable for system representation,
so, we need the Laplace transform of the above dynamics.
98 Monday, August 21, 2023
 In Laplace
domain
Tm ( s)  Kt I a (s ) (5)

Eb ( s )  Kb s ( s ) (6)

 sLa  Ra  I a (s)  Ea (s)  Eb (s)

Putting the value of Ia from (7) and Tm from
E ( s )  Eb ( s )
 I a ( s)  a (7) (8) in equation (5), we form the block diagram
 sLa  Ra  for closed loop system
 
Tm ( s)  s 2 J  sf  (s ) (8)  (s) Kt
G(s)  
Ea ( s ) s  Ra  sLa   sJ  f   Kt Kb 
Tm ( s)
  s 2 J  sf
 ( s) Kt
 G(s) 
s ( s 2  2n s  n2 )

Since the time constant of armature coil is much less compared to motor inertia, sLa is
neglected.
 ( s) Kt
G(s)  
Ea ( s ) s  Ra  sJ  f   K t Kb  (9)
99 Monday, August 21, 2023
  ( s) Kt / ( Ra f  Kb K t ) Km
   (10)
Ea ( s ) s ( s  1) s( s  1)

Kt Ra J
Km  Motor gain  Motor time constant
where Ra f  Kb Kt Ra f  K b K t
constant
The motor gain Kt and feedback gain Kb are equal during steady state operation. The
 )iamotor is
( Kbthe
power input to and the power delivered
T to shaft is
In the steady state ( Kb )ia  T 

Since T  Kt ia  K b  Kt

So, the transfer function for position control obtained from armature control technique
is of second order system neglecting the armature inductance and it is a closed loop
system.

Field control DC Motor T.F.

or wide speed variation of load, the field control scheme is used in industrial application. During thi
peration, the power maintains constant, since the terminal voltage and armature rated current are
ixed. The field flux is varied by changing the field current with field unsaturated.

100 Monday, August 21, 2023

 The air gap flux is proportional to field current

  K f if

The torque produced by the motor is

Proportional to the field and armature current.
Since, armature current is fixed,
Tm  K 1 ia  K 1 K f i f ia

(11)

As
di f
Field circuit dynamics E f  Rf if  Lf
dt (12)

d 2 d
Motor-load dynamic TL  J  f (13)
dt 2 dt

Neglecting the disturbing torque Tm  T L

Converting into Laplace transform of the eq(11) –(13), we

101 Monday, August 21, 2023
obtain
 (14)
I f ( s)
E f ( s)  R f I f (s)  sL f I f ( s) 
 E f ( s)  I f ( s) R f  sL f   
1
(15)
E f ( s) R f  sL f

 (s)
TL ( s)  Js 2 ( s)  fs ( s) 
 T L ( s )   ( s ) Js 2  fs  
TL ( s )

1
Js 2  fs (16)

Arranging the equation (14) –(16), we obtain the block diagram

The Transfer function of the open loop block is

3rd order system

Writing in pole-zero form

 (s) K t / fR f Lf
Writing in time constant form  G (s)  Where, f 
E f (s) s s   f s   L  Rf
102 Monday, August 21, 2023
 Develop a block diagram of a DC generator (Amplidyne generator) using as a
rotating amplifier, supplying current to resistive load.

An amplidyne is a two stage generator

(rotating amplifier ). The input is applied
To main field winding and the output
voltage is developed across the pair of
the brushes which is short circuited.
This short circuit current produces the
flux for generating the voltage across the
rest pair in amplified signal which is the
Field dynamics
induced voltage E.

Air gap flux= field flux - AR flux

Induced voltage

Voltage causes armature

current

Armature reaction flux is proportional output

to Ia
103 Monday, August 21, 2023
 The cause-effect relation develops a block diagram

By simplifying

104 Monday, August 21, 2023

 Assignments:
Modeling on Thermal System

Modeling on Pneumatic System

Two phase AC servomotor

105 Monday, August 21, 2023

 Module_1

Lecture_6

Analogy of Force-voltage and Force-current,

Numerical

106 Monday, August 21, 2023

 Analogous System
The mechanical translational system or rotational system can be treated as analogous
to electrical system when the differential equations for both cases are compared.

In such systems whose differential equations are of identical form are called analogous systems.

The concept of analogous system is a useful technique for the study of various system
Electrical, mechanical, thermal, pneumatic system etc. If the solution of one of system is obtained,
it can be extended to all other system analogous to it.

The force F (torque T) and voltage V are the analogous variable s in this study.

The R-L-C component of electric circuit is compared with mass-spring-damper elements

of mechanical system

Inductor stores electrical energy Mass stores mechanical energy

107 Monday, August 21, 2023
 Mechanical Rotational system Ex: tachometer, velocity sensor

The rotational motion of a rigid body can be defined as the motion of a body about its fixed axis.
There are three types of torques resists the rotational motion. Inertia torque, damping torque
and stiffness torque.

Shaft inertia (J) stores the K. E. of rotational motion and is the product of MOI and
angular momentum

Damping torque D is available in rotational body due to bearing and viscous friction.

If the rotor shaft of the load or the motor is very large (for high servo-motor), then
the torsional stiffness (K) is taken into consideration like spring constant. If the shaft
is small, the torsional stiffness of a rotating body is neglected.

d 2 d d 2 d For low power

T  J 2 D  K For high power servo TJ 2
D
dt dt dt dt

108 Monday, August 21, 2023

 Analogous System R-L-C series circuit
di 1
V  Ri  L 
dt C
idt  Mechanical
Translational System

dq dq 2 q dx dx 2
V  R L 2  F  D  M 2  Kx
dt dt C dt dt

Rotational System The rotational motion of a d d 2

rigid body is about its axis TD  J 2  K
dt dt

Force-Voltage Analogy

109 Monday, August 21, 2023

 Force-Current Analogy R-L-C
Parallel
dV V 1
I C  
dt R L 
Vdt

d 2
1 d  As V 
d
I C 2   dt
dt R dt L

dx 2
dx d 2 d rotational
F  M 2  D  Kx translational TJ 2
D  K
dt dt dt dt

110 Monday, August 21, 2023

 Analogous between transformer and Gears

111 Monday, August 21, 2023

 Module_1

Lecture_7

Rules for block diagram reduction,

system representation in block
diagram,
Numerical

112 Monday, August 21, 2023

 System Representation Problem, Block-Diagram Reduction:
Block diagram: It is the pictorial representation of the cause-and-effect relationship
between input and output of a physical system. Fig (a) block representation (b) block
with gain of a system

The arrowhead pointing towards the block indicates the input and the arrowhead
leaving away from the block represents the output. The arrows are called signal

The functional operation of the system can be visualized through the block
diagram representing the dynamic of the system

Depending upon the view point of analysis, different blocks can be drawn for a particular system.

Summing Point:It is a circle with cross is the symbol that indicates the summing operation.

The plus or minus sign at each arrowhead indicates whether

that signal is to be added or subtracted.

113 Monday, August 21, 2023

 Branch Point:
A branch point/take off is a point from which the signal leaves concurrently
to other block or summing point.

Block diagram of a closed loop system:

The output signal is fed to the input from branch point

or take off point for comparison

Before feeding the signal to the input, this must be converted to

desired signal through transducer at feedback path.
Unity feedback system when H(s)=1

Closed loop transfer function

For the system the input R(s) and output C(s)
are related as follows.

B ( s)
Open loop gain  E ( s)  G ( s) H ( s)
C ( s)
Feed forward path gain is G ( s)  E ( s) Non-unity feedback system

114 If H ( s)  1 Then open loop gain and feed forward path gain are same
Monday, August 21, 2023
 The output

Actuating error signal

Replacing E(s) from C(s)

C(1GH)  RG

Closed Loop T.F. (1) SISO system

(2)

It is seen that, the output of closed loop system depends on both closed loop T.F. and
nature of input.
In MIMO system, each input can be treated independently at a time neglecting other
inputs. The complete output of the system is obtained taking all inputs following
115 Monday, August 21, 2023
superposition principle.
 MISO system The control input and disturbance inputs are taken into consideration

When D(s)=0 When R(s)=0

C1 G1G2 G1G2 R
  C1  C2 G2 G2 D
R 1  G1G2 H   C2 
1  G1G2 H D 1  G1G2 H 1  G1G2 H
116 Monday, August 21, 2023
 The resulting output is C (s)  C1 ( s)  C2 ( s)

G1G2 R G2 D G2
C ( s)    C ( s)  G1R  D 
1  G1G2 H 1  G1G2 H 1  G1G2 H

As G1G2 H  1 The effect of disturbance C(s)/D(s) is zero. This is the

advantages of closed loop system

Procedures for drawing block diagram

 Write the equation for describing the behavior of individual component
 Take Laplace transform neglecting the initial condition of the set of equations
Represent each Laplace transform equation individually in each block form
 Assemble the elementary blocks to make a complete block that represent the
system
Ex:
The dynamics

RC circuit

117 Monday, August 21, 2023

 Corresponding Laplace transform

The overall block diagram is

Block diagram reduction:

Blocks are connected in series only if the output of one block is not affected by the
next following block.
A number of cascaded blocks are represented by a single block.
A complicated block consisting of many feedback loops can be simplified by step by
step arrangement
118 Monday, August 21, 2023
 Note: On simplification, one should know that, the input and output signal does not
change.
In the complicated loops , one should bring the summing point to the left side and take
off point to the right side.
Solve the inner loop , finally the outer loop
Reduction Rules

Moving a summing point

before the block

Moving a take off point

before the block

Moving a take off point

ahead of the block

Non-unity feedback
to feedback loop

Eliminating a feedback
loop
119 Monday, August 21, 2023
 Reduce the block using the rule

Step_1

120 Monday, August 21, 2023

 Step_2

C ( s) G1G2G3G4 / (1  G1G2 H1 )(1  G3G4 H 2 )


R( s) H3
1  (G1G2G3G4 / (1  G1G2 H1 )(1  G3G4 H 2 ))
G1G4

121 Monday, August 21, 2023

 Find the transfer function for the following block.

Solution:
Step_1, Moving the inner summing point to the left of the block G1 following the rule 1

122 Monday, August 21, 2023

 Find the T.F. for the inner loop containing G1, G2 and H1. Replace this by a single block
Step_2
following rule 5

Step _3 reduce the inner loop in the figure shown in step 2 following rule 5

Step_4 Find the complete T.F. Following rule 5 for the figure shown in step 3

123 Monday, August 21, 2023

Note: The numerator of the closed loop T.F. C/R is the product of their feed forward gain.

Denominator

Ex:
Find Y/R

solution

Step_1

Take off point is shifted to right

124 Monday, August 21, 2023
 Step_2

Step_3

Ex:

125 Monday, August 21, 2023

 Ex : Find C/R for the following

126 Monday, August 21, 2023

 Non-unity feedback to unity feedback system

1st approach

R 
C  ( R  CH )G    C GH
H 

C  GH  1 G
  
R  1  GH  H 1  GH

127 Monday, August 21, 2023

 2nd approach

128 Monday, August 21, 2023

129 Monday, August 21, 2023
 Module_1

Lecture_8

Signal flow graph (SFG), Mason’s

gain formula,
Numerical

130 Monday, August 21, 2023

 Signal Flow Graph (SFG):
Signal flow graph is another form of control system representation which shortens the
representation by replacing the summing point, take off point with simple nodes and
block of gain with transmittance. The two nodes are interconnected by line called branch
with transmittance.
A closed path is a loop that originates and terminates at same node with no node being met twice
along the path.
Two loops are non touching if they do not have common node.
The input and output variables are represented by node as well.
Only applicable to linear system

Ex Consider a straight line equation y  mx

Where x and y are input and output variables. m is the slope or gain. In SFG,
the I/P and O/P variables are represented by nodes and gain by transmittance.
The two nodes are interconnected byyline called branch.
 mx
SFG Source node m Sink node
x y

Ex SFG

131 Monday, August 21, 2023

132 Monday, August 21, 2023
 Ex:
SFG

Ex

SFG

133 Monday, August 21, 2023

 Mason’s gain formula:
Masson’s gain formula is used to determine the overall T.F. (gain) from source node
to sink node of a system. It is very convenient way in comparison to block diagram
reduction. The Masson’s gain formula is given by

C (s)
P
k 1
k k
G (s)  
R (s) 

Where, N=Total number of forward path

Pk=path gain of kth forward path joining from source node to sink
node
Δ= determinant of the graph
=1 - (Σ loop gains of all individual loops) + (Σ sum of the gain products of all
possible combinations two non-touching loops ) - (Σ sum of gain products all
possible combinations of three non-touching loops) + …

Δk= Cofactor of the path Pk

= value of Δ after eliminating all loops that touches kth forward path
=1-(loop gains not touching kth forward path.)
134 Monday, August 21, 2023
 Using Mason’s gain formula:

Step_1 Find the number of forward path and its corresponding path gain.

There is one forward path from source node R(s) to sink node. C(s)

The path gain is P1= G

Step-2 Find the individual loop gain.

135 There is one loop and the loop gain is L1= - GH Monday, August 21, 2023
Step_3 Find the combinations of two non-touching loop and their gain product.

Here no such two non-touching loops. Hence it is zero.

Step-4 Find the cofactor of the determinant

Δk= the value of Δ after eliminating all loops that touches kth forward path

=1-(loop gains not touching kth forward path.)

Here Δ1=1

Step-5 Find the determinant of the SFG Δ

Δ =1 - (Σ loop gains of all individual loops) + (Σ sum of the gain products of all
combinations two non-touching loops ) - (Σ sum of gain products all combinations
of three non-touching loops) + …
Δ = 1- (L1) = 1+GH

Step-6
136 Now the required TF is Monday, August 21, 2023
 Ex: Find SFG using Masson’s gain formula Corresponding SFG

Solution:

Step_1 There are two forward path connecting the input R(s)to output Y(s)

Step_2 There are four individual/self

loops

Step_3
137 Monday, August 21, 2023
 Step_4

Ex:

Sol:

138 Monday, August 21, 2023

 Ex:

Solutions:

Feed forward path gain:

139 Monday, August 21, 2023

 Ex:
Find the over all T.F. by SFG

Solution:

Individual path gain

Three loops
140 Monday, August 21, 2023
 The overall T.F.

141 Monday, August 21, 2023

 EX: Find SFG from the block

Ex:

SFG

142 Monday, August 21, 2023

 Solution
there are two forward paths between the input R(s) and the output Y(s). The forward
path gains are
P1 = G4(s) for the path R(s)E(s)X2(s)Y(s)
P2 = G1(s)G2(s)G3(s) for the path R(s)E(s)X4(s)X3(s)X1(s)Y(s)
there are five individual loops. The loop gains are

L1 = –G1(s)G2(s)H2(s) for the loop E(s)X4(s)X3(s)E(s)

L2 = – G1(s)G2(s)G3(s) for the loop E(s)X4(s)X3(s)X1(s)Y(s)E(s)
L3 = – G4 for the loop E(s)X2(s)Y(s)E(s)
L4 = – G2(s)G3(s)H1(s) for the loop X4(s)X3(s)X1(s)Y(s)X4(s)
L5 = G4(s)H1(s)G2(s)H2(s) for the loop E(s)X2(s)Y(s)X4(s)X3(s)E(s)

there are no non-touching loops in this signal flow graph. The determinant Δ for the
graph is given by

Δ = 1 – (L1 + L2 + L3 + L4 + L5)
= 1 + G1(s)G2(s)H2(s) + G2(s)G3(s)H1(s) + G1(s)G2(s)G3(s) + G4(s) -G4(s)H1(s)G2(s)H2(s)

143 Monday, August 21, 2023

 The value of Δ1 is computed in the same way as Δ by removing the loops that touch the
first forward path P1. In this example, since path P1 touches all the five loops, Δ1 is
found as Δ1 = 1

Proceeding in the same way, we find Δ2 = 1, The T.F.

C (s) P11  P2  2
G (s)  
R(s) 

C(s) G4  G1 G2G3
G(s)  
R(s) 1  G1 G2 H2  G2G3H1  G1G2G3  G4  G2 G4 H1H2

144 Monday, August 21, 2023

 Module_1

Lecture_9

Comparison between BDR and SFG ,

Numerical on SFG from block
diagram
Types of feedbacks

145 Monday, August 21, 2023

 Comparison between block diagram and SFG
Block Diagram Reduction Signal Flow Graph

Applicable to LTI Applicable to LTI

Each element is represented by block Each variable is represented by

node
Self loop does not exist Self loop exists

It is time consuming method Requires less time

Block diagram is required at each and At each step SFG is not required
every step
Feedback path is present Feedback loop is present

T.F. is shown in block T.F. shown on branch

146 Monday, August 21, 2023

Procedures for getting SFG from block diagram
• Replace all the summing points and take off points by nodes
•If the branch connecting a summing point and take off point has unit gain,
then both summing point and take off points are represented by a single node
• If there are more take off points from the same signal, then all points
represented by a single node
• If there is no block between two summing points, then those summing points
are considered as one node.

147 Monday, August 21, 2023

 SFG from block diagram model

Step-1: All variables and signals are replaced by nodes.

Step-2: Connect all nodes according to their signal flow.
Step-3: Each of gains is replaced by transmittances of the branches connected between
two nodes of the forward paths.
Step-4: The feedback path is replaced by transmittances of -1 multiplied with its gain.

148 Monday, August 21, 2023

 Ex:

149 Monday, August 21, 2023

 Ex:

150 Monday, August 21, 2023

 Ex:

151 Monday, August 21, 2023

 We can also find

152 Monday, August 21, 2023

 Types of feedback on control system

In open loop control system, there is no provision for supervision to correct the behavior
of system output due to lack of proper system components.

The open loop system is highly sensitive to disturbances and changes in parameters of
the plant.

The closed loop known as negative feedback system is used for self correcting
mechanism.

The most important property of negative feedback system is to reduce the sensitivity if
There are twoparameter
any system types of feed back occurs.
variation control system

Negative feedback system (degenerative)

Positive feedback system (regenerative)

153 Monday, August 21, 2023

Degenerative feedback control system: It is a control system where the feedback signal
opposes the input signal.

C ( s) G (s)

R(s) 1  G (s) H (s)

when G(s)H(s)  1 C (s)  finite

Regenerative feedback control system: It is a control system where the feedback

signal supports or adds the input signal.

C ( s) G ( s)

R(s) 1  G (s) H ( s)

when G(s)H(s) 1 C(s) 

154 Monday, August 21, 2023

However, the regenerative is sometimes used for increasing the loop gain of feedback system.

The positive feedback is necessarily required for the system in the inner followed by the negative
in the outer loop.
G1 G1G2 G3
1  G1 H1 1  G1 H1

The closed loop T.F. is C ( s) G1G2G3 / (1  G1 H1 ) G1G2 G3

 
R( s) 1  G1G2G3 H / (1  G1H1 ) 1  G1 H1  G1G2G3 H

1
If G1H1  1 the inner loop gain becomes very high and the closed loop T.F. H (s )
becomes
Thus, due to the high inner loop gain, the closed loop T.F. becomes insensitive to G(s), i.e.
Forward path gain
155 Monday, August 21, 2023
 Module_1

Lecture_10

Effect of degenerative feedback on control system,

regenerative feedback,
Application of control system in non-engineering
field

156 Monday, August 21, 2023

  Effect of feedback on parameter variation

open loop

The parameter changes due to environmental and aging factor

Let G( s) be the change in G ( sdue

) to parameter variation and C ( s ) be the

Corresponding change in output.

then C (s)  G(s) R(s) becomes C ( s)  C ( s)  G ( s)  G( s )  R(s)

 G( s) R( s)  G ( s) R( s)

157 Monday, August 21, 2023

  C ( s)  C ( s)  C ( s)  G( s) R( s)
 C (s)  G( s) R( s) (1)

Hence, it is observed that, for a small variation of gain parameter, the output is
proportionally varied.
parameter variation on closed loop system C ( s) G ( s)

R( s ) 1  G ( s ) H ( s)

G ( s)
 C (s)  R( s)
1  G ( s) H ( s)

Let there is a small change in outputC ( s) G

due to change in( s )
G ( s)  G ( s )
So, C ( s )  C ( s)  R( s )
1  G ( s)  G ( s)  H ( s)

G ( s)  G ( s)
C ( s )  C (s)  R( s)
1  G ( s) H ( s)  G ( s) H ( s)

As G( s) H (s)  G( s) H ( s)
158 Monday, August 21, 2023
  G ( s)  G ( s ) 
We have C ( s)  C (s)    R(s)
 1  G ( s ) H ( s ) 

 G (s)   G ( s) 
C ( s)  C ( s)    R ( s )    R( s )
 1  G ( s ) H ( s )   1  G ( s ) H ( s ) 

 G ( s) 
C ( s)    R( s ) (2)
 1  G ( s ) H ( s ) 

1
Comparing (1) and (2), we observe that, the output variation in closed loop is1  G ( s) H (s)

times that of open loop system. Usually G(s)H(s) is very high which means, feed
back reduces the change in output due to effect of parameter variation.

159 Monday, August 21, 2023

  Effect of feed back on sensitivity
Sensitivity:
The performance of an open loop control will change due to any change in its
parameter values.

The parameters changes because of environmental conditions or because of

ageing.

The resistance of coils change with temperature, the gain of transistors change
with time, and the friction of a bearing is dependent upon the level of
lubrication.

The performance of the open loop system will also change because of
unwanted 'noise‘ signals which will always be affecting any physical system.

We then say that the performance of the system is 'sensitive' to parameter

variations and to unwanted noise signals.

One of the prime reasons for using feedback is that it reduces the sensitivity of
160the system to both parameter variations and external noises orMonday,
disturbances
August 21, 2023
Sensitivity of T.F. T(s) on G(s ) due to degenerative Feedback :

The effect of a parameter change on system performance can be expressed in terms of a

sensitivity function.

T / T T G
 SGT   
G / G G T

In an open-loop system, C (s)  G(s) R(s)

When G(s) changes to G (s)  G (s) Then, C (s)  G(s) R(s)

T / T G G
 SGT    1 In open loop T  G
G / G G G

In an closed-loop system, G G(s)

T  C (s)  R(s)
1  GH 1  G (s) H(s)

T G   G   G 
SGT    G /  
G T G 1  GH   1  GH 
161 Monday, August 21, 2023
 Closed loop   G 
SGT    1  GH 
G  1  GH 

1  GH   GH 1
 2
 1  GH   SGT 
1  GH  1  GH 
1
Since the GH is very high, the sensitivity for a closed loop system is reduced by a factor
1  GH 

This implies that, the closed loop system is less sensitive to G variation w.r.t. open
loop system
Sensitivity on feedback gain H

Open loop: In open loop system, H=0

T T H
So, sensitivity H H T  0
S 

T H   G  H GH
Closed loop S HT    S HT    1
  1  GH
H T H  1  GH  G / (1  GH )

The closed loop system is more sensitive variation in feedback path

162 Monday, August 21, 2023
parameter
 Negative sign implies that, T decreases with increase in H.

R2
Ex: K 
Find the transfer function T of this inverting amplifier where 1
R1

Find the sensitivity due to change in A and due to change in feedback gain K

Given the gain A  10 4

R2
K   0 .1
R1

163 Monday, August 21, 2023

 The equivalent circuit of the amplifier is
V0  A Vi  KV0 

 V0  AKV0  AVi

 V0 1  AK   AVi

V0 A
T  
Vi 1  AK
The sensitivity due to change in A is
V0 A 104
T A 1 1 T    10
S TA     0.001 Vi 1  AK 1  104  0.01
A T 1  AK 1  104  0.1

The sensitivity due to change in feed back gain K is

T K KA 103
S KT     1
K T 1  AK 1  104  0.1

164 Monday, August 21, 2023

 Ex:
Calculate the sensitivity of the system transfer function w.r.t.
(a) Forward path gain
60
(b) Feedback path gain s  s  10 

given   1.6 rad/sec

0.8

Sol:
a)
T G
The sensitivity of the T.F. w.r.t. forward path gain is given bySGT 
G T
G
since T  1  GH

1 1 s 2  10s
SGT    2
1  GH 1  60
* 0.8 s  10s  48
s ( s  10)

s 2  10s  2  j10 (1.6)2  j10(1.6)

Substituting s  j SGT  2
 2

s  10s  48   j10  48 (1.6)2  j10*1.6  48

165 Monday, August 21, 2023

 2.56  j16
SGT 
45.44  j16
(2.56) 2  162
The magnitude of the sensitivity is SGT   0.336
2 2
(45.44)  16

(b) The sensitivity of the system T.F. w.r.t. feedback

gain is
T H GH  48 
S HT     2 
H T 1  GH  s  10s  48 

substituting s  j  48 
S HT    2 
   j10  48 

48
S HT    0.996  1
2 2
(45.44)  16

Negative sign indicates that, T decreases with increase in H. Hence, feedback reduces the
sensitivity
166 Monday, August 21, 2023
  Effect of feed back on stability Consider first order system. The T.F. is

V0 ( s ) 1 1
 
Vi ( s) 1  RCS 1   s

Where  is the time constant

The general representation of first order system in block diagram is

1
K 1 s

V0 ( s ) K
G
Where K is the amplifier gain. The open loop T.F. in time constant form is( s )  
Vi ( s) 1   s
K / 1
In pole-zero form G(s)  Open loop pole is at s  
s  1/  

1
s

167 Monday, August 21, 2023
 Closed loop system

K
'
1  s Where K is the sensor gain

K'
Now,
Closed loop pole
K /1  s K /
CLT .F .  T ( s )  ' 
1  KK ' KK  1  KK ' 
s 1 s 
 1  s   

1
s

Open loop pole

Hence, the increasing the system gain K, the stability is improved in closed loop as
compared to open loop system.
168 Monday, August 21, 2023
 Effect of Feedback on bandwidth: (the range of frequencies over which the system
Responds satisfactorily )
K' / K
K'  G ( s)   G( s) 
Open loop G (s)  s  1 s  1
s 

Closed loop (unity feed back) G(s) K / (1  s ) K

T (s)   T (s)  
1  G ( s) 1  K / (1  s ) s  1  K

K / (1  s ) K
 T ( s)  
1  K / (1  s ) s  1  K

K / (1  K ) K / (1  K ) 
 T ( s)   where  CL 
s / (1  K )  1 s CL  1 K 1

1
B.W. is the frequency at which the DC gain drops to
2
K K 1
For open loop G ( s)   
s  1 j  1 2

K / (1  K ) K / (1  K ) 1
For closed loop T ( s )   
s CL  1 j CL  1 2
169 Band width of anAugust
Monday, audio21,
amplifier
2023
 K K / (1  K ) 1 1 1
   
j  1 j CL  1 j  1 (1  K )( j CL  1) 2

 j  1  (1  K )( j CL  1)  2   2 2  1  (1  K )( 2 CL

2
 1)  2

and

Also  bCL  (1  K )bOL

If the gain G1 is very high, disturbance in forward path is less affected

170 Monday, August 21, 2023
  Effect of feedback on disturbance:

Disturbance in forward path:

Td ( s )
 C ( s) 
G1 H

171 Monday, August 21, 2023

 Properties Open Loop System Closed Loop System Comments

Transfer Open loop gain is higher

Function because G(s) is normally
& Gain high
Parameter Variation in output with
variations parameter variation is
lower in closed loop
system
Sensitivity Closed loop is much less
sensitive to G while open
loop is 100%
But closed loop is highly
sensitive with H variation

Stability For 1st order system, Pole is at For closed loop system,
the pole is at the pole shifted more
stable position by
adjusting K
Band Band width increases in
Width closed loop system
Disturbanc The disturbance does Noise or disturbance is Closed loop system are
172 Monday, August 21, 2023
e not arise considerable reduced able to control
disturbance
 Module_2

Lecture_11

System Analysis in time domain, Standard test signals-

Characteristics of step, ramp, parabolic , impulse and
sinusoidal signal:

173 Monday, August 21, 2023

 Time Domain Response
The manner in which the dynamic system responds to an input, expressed as a function
of time
is called time domain response.
It is possible to compute the time domain response of a system if the mathematical
model of
the system is known.

The time domain response has two components,

 Transient response/natural response: This response will decay (for stable
system) exponentially or oscillatory to zero as time increases and is independent of
the input quantity.
The oscillatory or exponential nature are due to system’s storage energy device. This
device does not allow the response instantaneously.
 Steady state response/forced response: This is the response of a system after
transient part decays to zero. However, this response depends both on system dynamics
and the input quantity. If the response of the system at steady state does not exactly
agree with given input, then steady state error will occur. This error is the indication of
accuracy.
So the total response is given by
174
The dynamic response along with steady state response decides theMonday, systemAugust
behavior
21, 2023
(absolute stability, relative stability and steady state error of the system)
 The difference between the input xi (in this case ramp) and output x0 during transient
period is
called transient error and the difference between the input and out during steady state
period is
Called steady state error.

The transient and steady state response are studied with certain standard signal called test
Typical
signal. test signal: The common test signal are step, ramp, acceleration, impulse and
sinusoidal. The mathematical and experimental analysis of a control system can be
175 Monday, August 21, 2023
carried out easily.
 Which test signal is required ?
If the system is subjected to sudden disturbances, a step input is the good choice.
If the inputs to a system is gradually changing with time , then ramp function is the right
choice.
If the input to a system is more gradually increased with time than ramp, then parabolic
function is used.
If the inputs is subjected to shock input, then impulse function is the best one.
Impulse signal: The impulse input is useful when we consider the convolution integral
For the output y(t) in terms of input r(t) which is written for the system shown

Impulse response

Since the integral has a value only at   0 , therefore

is the impulse response for the system.

176 Monday, August 21, 2023
 When the system is subjected to a large magnitude , narrow width pulse. An
impulse t  0
is a pulse with width The strength of impulse is its area A

The impulse function whose area is unity is called

unit impulse function.

Also the derivative of unit step signal is the unit impulse.

dr (t )
 (t ) 
dt

 dr (t )  1
R( s)  L( (t ))  L    s  1
 dt  s

Unit step function: Displacement input signal

177 Monday, August 21, 2023

 Ramp function:

Also called velocity input signal. It is a step function of velocity. Integration of step is ramp functi

A is the slope of the line.

Parabolic Function:

This is the acceleration input and integration of ramp is the parabolic

function.

The derivative of unit ramp is also called parabolic function.

178 Monday, August 21, 2023

 Sinusoidal signal Test

This is one of the test signal used

testing of frequency domain
Response. This gives a steady state
response
Information of the system. The steady
state response also a sinusoidal
responserwith some phase angle.
(t )  sin(t )

R ( s )  L ( r ( t ))  L sin ( w t ) 
s2   2

Application of step signal to a system

Consider a simple system given by the transfer function in which only one pole and one
zero is present. Let us apply the step input to this system. The system results the transient
and steady state response.

The pole-zero
Location for given
system G(s)

179 Monday, August 21, 2023

 When step input is applied, the output or response becomes

Where,

Thus,

Summary:
 The pole at step function generates the
forced response i.e. the pole at origin generates
a step function at output.
 A pole of the system (transfer function G(s)) The zero of the system implies
generates the natural response, i.e. pole at -5 the magnitude of the forced
e 5t
from origin of s-plane generates and natural response.
 The pole at -5 is on real axis of s-plane gives
exponential response. The pole farther away
180 from the origin gives faster response. Monday, August 21, 2023
 Module_2

Lecture_12

Time response of 1st order systems to unit impulse, step and unit
ramp inputs, Parabolic input

181 Monday, August 21, 2023

 Types and order of the system:
Types of the system: The number of poles of open loop T.F. G(s)H(s) at origin denotes
type of the system. Generally H(s) = 1 or unity feedback is considered.

Order of the system: The highest power of s in the denominator of G(s)H(s) is called
as order of the G(s)H(s).
K ( s  2)
Let G (s) H (s) 
s( s  1)( s  5)

Here, one pole is at origin (type-1) and the highest power of s in denominator is 3. So the
transfer function G(s)H(s) is 3rd order system.
Note: Usually the type is the property of open loop T.F. G(s)H(s) while the order is the
property of closed loop T.F. i.e. number of closed loop poles on the system.
G (s)
T (s) 
1  G ( s) H ( s )

The order of system means, the order of characteristic equation 1+G(s)H(s)

Time response of 1st order system with impulse input or

Impulse response of 1st order system

Consider an impulse of area A applied to a 1st order

182system without zero. Monday, August 21, 2023
 The output is

So, where a=1/T, pole of the system

The inverse transform becomes

At time t=0, the output is AK/T and

decreases exponentially as t increases.
Initial slope of the output x0 (t) corresponds to the
speed of response. If the slope is high,
system will be faster.
dx0 (t ) AK
The initial slope is at t=0  2
dt T

Note: Since, the output only gives the transient

response, there is no steady state response,
hence no steady state error
e (  )  xi (  )  x 0 (  )


AK Since r ( ) is not defined at infinity
2
T
183 e (  ) is not Monday,
definedAugust 21, 2023
 Impulse Response to 2nd Order System

184 Monday, August 21, 2023

185 Monday, August 21, 2023
 Step response of 1st order system:
Consider a first order system without zero .

where R(s)= 1/s is step input

Open loop pole location
The step response is Of G(s)

The time response is

Where, the input pole at the origin generates the forced response c f (t )  1 and the system pole
at -a generates the natural response cn (t )  eat  et / 

1
Where  time constant
a

t   so
The parameter a only decides the system transient response. When e at  0.37

So, c(t )  1  0.37  0.63

At time constant t  the transient response decreases to 37% and hence, the system response

rises to 63% of its final value.

186 Monday, August 21, 2023
 1
The initial slope at t=0 is 

dc ( t ) d
dt

dt

1  e  t /   1

At t  4 c(t) 1e4  98%

When the output reaches within 2% of

the defined output , the system reached its
steady state condition.
4
At t  4 
a
Settling time:The time required to settles at desired valuet s  4

The time required to reach from 0.1 to 0.9 of the final value . It is found the differenc
Rise time:
2.3 0.11 2.22
of time from c(t) =0.9 to c(t)=0.1. tr   
a a a
Hence,  e ( t )  1  (1  e  t / )  e  t /
The error e ( t )  r (t )  c ( t )

The steady state error is e ss  tlim e ( t )  0 Applying final value theorem



187 Hence, the steady state error is zero for a first order system with step input.
Monday, August 21, 2023
188 Monday, August 21, 2023
 Ramp response of 1st order system

Consider the ramp input with slope Q

applied to a 1st order system.
The output X0(t)

(1)

 1
Multi plying both sides by s 2  s   We get
 T

(2)

Equating the coefficients on both sides (2)

(3) From (3)

(4) From (5)

(5) Substituting in (4)

Hence From (3)

189 Monday, August 21, 2023
 Putting A, B, C in (1)

Taking inverse transform

If the slope and gain Q=K=1, then (6)

The 1st term in equation (6) represents the input, the 2nd term represents the steady state error
And 3rd represents the transient component.

The error is found as e(t )  xi (t )  x0 (t )

 e(t )  t  t  T  Tet /T

 e(t )  T (1  e t /T )

Hence, the steady state error ess  tlim


e(t )  T

Thus, reducing the time constant (T), reduces not only the steady state error, but
190
also improves its speed of response. Monday, August 21, 2023
 Parabolic response of 1st order system

With the same 1st order system, the parabolic input is given and the resulting output
and its steady state error are found as
1
Input R(s) 
s3
The time response of the output is
t2
c(t )  T 2  Tt   T 2 e  t /T
2
t2
The error is e(t )  r (t )  c(t )  t  T  Tt   T 2 et /T
2 2
2

The steady state error ess  lim e(t )  

t 
is
Summary:
 Step input gives the desired information about the speed of transient response
The speed of response is inversely proportional to the time constant of the system
The ramp and parabolic input do not give any additional information regarding the speed
of response
 The steady state error is not defined for impulse input
191
 Monday, August 21, 2023
So, it is more appropriate to study behavior of any system by unit step input
 Module_2

Lecture_13

Time response of 2nd order systems to unit step, Time domain

specifications

192 Monday, August 21, 2023

 Time domain response of second order system
In the first order system, the system parameter (poles) changes the speed of response.
However , we will see in second order system that how the system parameters can
change the form of the response.
In this system, the characteristics or the transient behavior is known from its
Consider
component a standard
values. form of second order system considering only pole. By proper
n
selecting the two parameters such as natural frequency  damping ratio ,
and we
get the closed loop poles. This closed loop pole results various shape or form of the
transient response. n2
The open-loop T.F. of a standard proto type second order system is G(s) 
given by s(s  2n )
The closed loop T.F. of the system for unity feedback
system i.e H(s) 1

CLTF
C(s) G(s) n2
T (s)   
R(s) 1 G(s)H (s) s2  2n s  n2

The solution of the characteristic equation results the closed loop poles i.e.

193 s2  2n s  n2  0 Monday, August 21, 2023

The roots of the characteristic equation are given by s1 , s2  n  jn 1   2

Keeping the natural frequency constant and varying the damping ratio , various forms of
transient response due to closed loop poles are developed.

For example if
s1 , s2  n  jn 1   2
poles are complex conjugate and lie in
0 1
left half of s-plane and the transient response is under damped.

poles are negative and equal s1 , s2  n

 1
the transient response is critically damped.

poles are negative real and unequal. s1 , s2  n  n  2  1

 1 The transient response is over damped
Without oscillation.

 0 poles are on imaginary axis and complex conjugate and

s1 , s2   jn
the transient response is undamped or oscillatory.

194 Monday, August 21, 2023

195 Monday, August 21, 2023
 Step response of second order system

n2 n2
The output is C (s)   C ( s) 

s s 2  2n s  n2  
s s 2  2n s  (n )2  n2  (n )2 
n2 n2
 C ( s) 

s ( s  n )2  n2 (1   2 )   C (s) 

s ( s  n ) 2
 d2  Since d  n 1   2

Taking the Laplace inverse, we get

A Bs  C
By partial fraction C ( s)  
s ( s  n )2  d2

Solving the equations, we get A=1, B=-1 C  2n

1 s  2n 1 s  n n d 

Now C (s)       
s ( s  n )2  d2  s ( s  n )2  d2 d ( s   )2  d2 

Since,

196 Monday, August 21, 2023

 This results that, the transient oscillation of the response is damped oscillation with
damped frequency d and varies with 

The error signal for this system is

The error thus exhibits the damped oscillation

At steady state t  No error exist between input and output.

197 Monday, August 21, 2023
 I
f

198 Monday, August 21, 2023

If the damping is   0 the response becomes undamped oscillation with natural frequency
n

The output will be

The damped frequency is given by d  n 1   2 is always less than undamped natural

frequency n

When increases beyond unity, the transient response becomes overdamped.

When   1 Critically damped, the two poles are equal s1 , s2  n

For step input

  1Over damped, two poles are negative real and unequal. s1 , s2  n  n  2  1

199 Monday, August 21, 2023

 Thus the two response is decaying exponentially. The faster exponential term is neglected
considering slower one.

Neglecting the one of the faster pole, the solution is

200 Monday, August 21, 2023

 Time domain specifications
The standard time domain specifications are available for the second order system with
step input as

201 Monday, August 21, 2023

 With delay Input

202 Monday, August 21, 2023

 1. Delay Time (td): The delay time is the time required for the response to reach half
the final steady state value the very first time
2. Rise Time (tr): The time required for the response to rise from 0% to 100% (for
underdamped system) of its final value. For overdamped system, tr is 10% to 90%
of its final value
3. Peak Time (tp): The peak time is the time required for the response to reach the
first peak of the overshoot
4. Maximum (percent) overshoot, Mp: The maximum overshoot is the maximum
peak value of the response curve measured from unity. It is expressed as

c(t p )  c()
Maximum percent overshoot =  100%
c ( )

The amount of maximum overshoot indicates the relative stability of the system

Calculation of time domain specifications:

Delay time: As per definition
c(td )  0.5

e t
n d

0.5  1  sin(d td   ) 1  2
1 2 Where    tan 1


203 Monday, August 21, 2023
 ~ 1  0.7
td 
n

Rise time (tr) As per definition for underdamped system

e  t
n r

1 1 sin(d tr   )
1 2

e  t
n r

0 sin(d tr   )  e t sin(d tr   )  0

n r

1  2

but e t  0 n r so sin(d tr   )  0

 sin(d tr   )  sin(n ) Where n=1 for first intersection of response

with input (i.e 1st overshoot)
 d tr      d tr    
 
 tr 
204 d Monday, August 21, 2023
 Peak time (tp)

At t=tp, the output c(tp) will attain maximum value. So the slope at the maximum
point is zero.
e
  t
n p

dc(t p ) As c(t p )  1  sin(d t p   )

0 1  2
dt

dc(t p ) n  t e

  t n p

 e sin(d t p   ) 
n p
d cos d t p     0
dt 1 2
1  2

 n sin(d t p   )  n 1   2 cos d t p   

 cos   sin(d t p   )  sin   cos d t p   

Where sin   1   2
cos   
 tan(d t p   )  tan  Monday, August 21, 2023
205
 n
 d t p  n  tp 
d

Note:

206 Monday, August 21, 2023

 Peak overshoot Mp:

As per definition,

M p  c (t ) max  1  M p  c (t p )  1

e
 n t p

M p  1 sin(d t p   )  1
1  2


 n

e n 1 2

 Mp   sin(d  )
1 2 d


1 2
e
 Mp  sin  since sin       sin 
1  2



1 2

207
 %M p  e  100% Mp is independentMonday,
of natural frequency
August 21, 2023
 Settling time (Ts):

The settling time is determined when the response is within the  2 % tolerance of its final vlaue

This occurs when

The speed of decay of the transient response

depends on time constant 1
  n

For a given  n the settling time is inversely

proportional to damping ratio  .

208 Monday, August 21, 2023

 Impulse response to second order system;
For the unit impulse, the output of second order system is

For under damping case,

dc(t )
since  (t ) 
dt

Impulse response

209 Monday, August 21, 2023

 Lecture_14

Steady state Error, Error constants of different types

of control systems, type-0,type-1,type-2, Numerical

210 Monday, August 21, 2023

 Steady State Error
The difference between the actual output and desired output is known as error. The t  
error at
is called steady state error

Error E(s)  R ( s )  B ( s )
 R (s)  C (s)H (s)
 R ( s )  E ( s )G ( s )
1
 E(s)  R (s)
1  G (s) H(s)

Steady state error is

211 Monday, August 21, 2023

 Steady state error depends on two factors:

Type and magnitude R(s)

Type of Open loop T.F. G(s)H(s)

Static position error constant:

Usually H(s)=1 unity feedback system
A
Step Input R(s) 
s
sR ( s ) A A
e ss  lim sE ( s )  lim  
s0 s0 1  G (s) 1  lim G (s) 1  K p
s 0

Where, K p  lim G (s) Positional error coefficient or

s 0 static position error constant
A
e ss 
212 1 K p Monday, August 21, 2023
 While for type 1 or higher Kp is infinite.

213 Monday, August 21, 2023

 Static velocity error constant:
Ramp input

Type-1 to
ramp input

The steady state error is

214 Monday, August 21, 2023

215 Monday, August 21, 2023
This analysis indicates that, a type 0 system is incapable of following a ramp input in the
steady State.
The type 1 with unity feedback can follow the ramp input with a finite error.
The type 2 or higher system can follow a ramp input with zero error at steady state.
Static acceleration error constant:

For acceleration input

216 Monday, August 21, 2023

217 Monday, August 21, 2023
 Note:
Both type 0 and type 1 systems are incapable of following input in the steady state.

The type 2 with unity feedback follows a parabolic input with a finite error signal.

The type 3 or higher system follows the parabolic input with zero steady state error.

218 Monday, August 21, 2023

 Conclusion
Type 0 system
A
K p  lim G (s) H(s)  K e ss 
Step input s 0 1 K

1
ramp input K v  lim sG (s) H(s)  0 ess   
s0 Kv

1
Parabolic input K a  lim s 2 G (s) H(s)  0 e ss   
s0 Ka

219 Monday, August 21, 2023

 Type 1 System

A
K p  lim G (s) H(s)   e ss   0
Step input 1 
s 0

A
K v  lim sG (s) H(s)  K e ss 
ramp input K
s 0

1
e ss   
ramp input K a  lim s 2 G (s) H(s)  0 Ka
s0

Type 2 System
A
Step input K p  lim G (s) H(s)   e ss   0
s0 1 
ramp input K v  lim sG (s) H(s)   1
s 0 e ss   0
K
Parabolic
K a  lim s 2 G (s) H(s)  K A
input s0 e ss 
Monday, August
220 K 21, 2023
 Steady-state error and error constant for different types of input are summarized as
follows
Type Step input Ramp input Parabolic input

Kp e ss Kv e ss Ka e ss
Type 0 A
K 0  0 
1 K
Type 1 A
 0 K 
0
K
Type 2  A
 0 0 K K

Note:
 If the order of the input is same as type of system, then the error is finite.
If order of input is greater than the type of system, then the steady state error is infinite.
If type of system is greater than order of input , then error is zero

The steady state error for impulse response is zero for all type of system.
221 Monday, August 21, 2023
222 Monday, August 21, 2023
 Advantages of static error coefficient method:
Simple calculation

Limitations:
Applicable only to stable system
Applicable only to three standard input signals
Cannot give exact value of error. It gives only mathematical value i.e. 0
or infinite

223 Monday, August 21, 2023

 Mod_2

Lecture_15

Generalized error series method, Relation

between static and dynamic error coefficient,
Numerical

224 Monday, August 21, 2023

 Generalized Error coefficient Method/Dynamic error coefficient method
In type 0, 1,and 2 systems, only one of the static error constants kp, kv and ka is non-zero
and finite and the other two are either zero or infinity. The dynamic error coefficient
method is more generalized approach to estimate the steady state error. This method is
used to find the at any instant of time.
1
Consider the error in closed loop system E ( s )  R (s) (1)
1  G (s)H (s)

1
 F1 ( s ) F2 ( s ) Where, F1 ( s )  F2 (s)  R ( s )
1  G (s) H (s )

Using convolution integral

t
e (t )  
0
F1 ( ) F2 ( t   ) d 

t
e (t )   0
F1 ( ) R ( t   ) d  (2)

By Taylor series
 2' ''  3
R (t   )  R (t )   R (t )  R (t )  R '''  . . . (3)
225 2 ! 3 !Monday, August 21, 2023
 Substituting (3) in (2) we get,
t t t
 2 ''
e(t )   F1 ( ) R(t )d    R (t ) F1 ( )d   R (t ) F1 ( )d  ...
'
Neglecting higher order terms
0 0 0
2!

  
2
ess  lim e(t )  R(t )  F1  d  R (t )   F1 ( )d  R (t )  F1 ( )d
' ''
t 
0 0 0
2!

define Calculation of k0

k0   F1 ( )d
0


k1     F1 ( )d  k0e s  F1 ( s)
0

Taking limit as s  0 both sides

k0  lim F1 ( s)
s 0

ess  k0 R(t )  k1R ' (t )  k2 R '' (t )

Where
1
Where k0, k1 and k2 are known as dynamic error coefficient. F1 ( s) 
G(s) H ( s)
226 Monday, August 21, 2023
 Calculation of k1 d
 k1e s  F1 ( s ) Where
ds 
since k0 e s  F1 ( s) k1    F1 ( )d
d d 0
 k0 e s  F1 ( s)
ds ds Taking limit
d
 k0   e s   F1 ( s) k1  lim
d
F1 ( s)
ds s 0 ds

Substituting k0
similarly
d
 k0   e  s   F1 ( s )
ds
Putting k0
d2
k2  lim 2 F1 ( s)
s 0 ds

d
 F ( )d ( )e
 s
1  F1 ( s)
0
ds

d

 s
 F1 ( ) d e  F1 ( s )
0
ds
227 Monday, August 21, 2023
 Ex:
For the system shown, find the steady state error using dynamic error coefficient
method r (t )  4  3t  4t 2
for the input given by

Sol:
60
G( s) 
( s  3)( s  4)
where
2
r (t )  4  3t  4t

ess (t )  k0 r (t )  k1r ' (t )  k 2r '' (t ) 1

F1 ( s ) 
60
1
( s  3)( s  4)

s 2  7 s  12
F1 ( s )  2
228 s Monday,
 7 s August
72 21, 2023
 s 2  7 s  12 12
k0  lim 2   0.1667
s 0 s  7 s  72 72
d  s 2  7 s  12 
k1  lim  2   0.0810
s 0 ds s  7 s  72
 

d 2  s 2  7 s  12 
k2  lim 2  2   7.39*10 3
s 0 ds
 s  7 s  72 

ess (t )  0.1667(4  3t  4t 2 )  0.0810(3  8t )  7.39*103 *8

ess (t )  0.9  1.14t  0.66t 2

229 Monday, August 21, 2023

 Relation between static and dynamic error coefficient
KN ( s )
Since the open loop transfer function for unity feedback system is expressedGas
( s) 
D( s)
E (s) 1
As error is 
R(s) 1  G (s)
E ( s) D( s)
 
R( s) D( s )  KN ( s )

Dividing the numerator and denominator and arranging the result in ascending
order of E ( s)
 k0  k1s  k2 s 2  ...
S , we obtain R(s)

E (s)  (k0  k1s  k2 s 2  ...) R(s)

Taking the Laplace inverse , we get

e(t )  k0 r (t )  k1r ' (t )  k2r '' (t )  ... Alternative way to write the error series.

For type 0 system, the error is given by E (s)  (k0  k1s  k2 s 2  ...) R(s)
1
For step input, R( s) 
s
230 Monday, August 21, 2023
 The steady state error is given ess  lim sE ( s )  k0
s 0
by
But the steady state error of a type 0 system due to unit step input is found during static error
coefficient is 1
ess  Where Kp is the positional error coefficient
1 K p

Comparing the two equations , we obtain

1
k0  For type 0 system
1 K p

Similarly for type 1 system, the error equation is

E ( s )   k1s  k 2 s 2  ... R( s )
1
For a unit ramp function R ( s ) 
s2
ess  lim sE ( s)  k1 Type-1
The steady state error is s 0
1
1 k1  k0  0
e  So, Kv
Since , we obtained in static error coefficient method, ss K
231 v Monday, August 21, 2023
 For type -2 system, the error is E ( s)  (k2 s 2  k3 s 3  ...) R( s)
1
For a unit parabolic input, R ( s) 
s3
The steady state error is

ess  lim sE ( s )  k2
s 0
1
But it is found that, the acceleration error constant is ess 
Ka

So, for type -2 system,

ess  lim sE ( s )  k2
s 0

1
k2  k1  0 k0  0
Ka

232 Monday, August 21, 2023

 Hence the other way of representation of steady state error for composite input is

R1 R R
ess   2 3
1  K p Kv K a

R1, R2 and R3 are the magnitude of step, ramp and parabolic input

ess  0.005
Ex Assuming r(t)=(0.1t ) and it is desired that

Find the range of K for error to be within the specified limit for the given system

Solution: 1 H (s)  1 r (t )  0.1t

G ( s) 
s( s  1)

K
For the ramp input, type 1 system, K v  lim sG ( s) H ( s ) K v  lim s K
s 0 s 0 s ( s  1)
233 Monday, August 21, 2023
 A 0.1 0.1
Steady state error is ess    0.005   K  20
Kv K K

So for any value of K, greater than 20 , ess will be less than 0.005. Hence, the range
of K is
20  K  

200
Ex: For a unity feedback system G ( s)  s ( s  8) and r (t )  2t

Determine the steady state error if it is desired to reduce this existing error by 5%.
Find the new value of the gain of the system

Solu: 2 200
For the given input R( s )   Kv  lim sG ( s ) H ( s )  lim s
s2 s 0 s 0 s ( s  8)
A 2
 ess    0.08
K v 25

Now, the error to be reduced by 5% of the existing value, with new of G(s)of gain K2 instead of 2

 5 
ess1  ess   ess   ess   0.05*0.08   0.0796
 100 
234 Monday, August 21, 2023
 A
 0.0796  Kv  lim sG( s)
Kv s 0

K2
Where G( s)  K2
s( s  8) So, Kv 
8

Now, A 2 16
0.0796   
Kv K2 / 8 K2 K 2  201 New gain

235 Monday, August 21, 2023

 Mod_2

Lecture_16

time domain specification numerical, steady state error on non-

unity feed back system, numerical

236 Monday, August 21, 2023

 Numerical on time domain specifications

Ex: A second order servo-system is governed by the following differential

equation.
d 20 d0
2
 6  200  100i
dt dt

Where 0 and i are the output and input respectively. For this system find the

(i) Peak overshoot (ii) rise time (iii ) peak time (iv) delay time (v) settling time (5%)

Solution:
d 20 d0 0 ( s) C (s ) 100
 6  200  100i    2 Given Closed loop system
dt 2 dt i ( s ) R( s) s  6s  20

The characteristic equation s 2  6s  20  0

2 2
Comparing this with standard form s  2n s  n  0

We get, n2  20 n  4.47 rad/sec 2n  6    0.67

237 Monday, August 21, 2023
 (i) Peak overshoot  /( 1 2 )
%M p  e *100  5.57%

  co s  1 
(ii) Rise time t r  2
 0 .6 9 4 sec
n 1 


(iii) Peak time (tp) t p  2
 0 .9 4 6 sec
n 1 

1  0 .7 
(iv) Delay time (td) t d   0 .3 2 8 sec
n

3
(v) Settling (ts) ts   1 .0 0 0 sec
 n

238 Monday, August 21, 2023

   0 .5 for the following
Ex: Determine the value of K such that the damping ratio
system
Find (i) tr (ii) tp (iii) Mp (iv) ts

Sol: Take off point to be shifted to the right of integrator

239 Monday, August 21, 2023

 2
The characteristic equation gives s  (0.8  16 K ) s  16  0

2 2
Comparing with standard equation s  2n s  n  0

2n  0.8  16 K  K  0.2

So, n 2  16  n  4 rad/sec
For   0.5

1   60
  cos    60 180  0.605 sec
(i) tr   
n 1   2 4 1  0.52 4 1  0.52
  4
tp    0.906 sec (iv) ts   2 sec
(ii) 2 2 n
n 1   4 1  0.5

(iii)  /( 1 2 )

240 %M p  e  16.3% Monday, August 21, 2023
 Ex:

Sol:

241 Monday, August 21, 2023

 Ex:

Sol:

242 Monday, August 21, 2023

 steady state error on non-unity feed back system
Consider a non-unity feedback system. To get the steady state error, this non-unity must be
converted into unity feedback system.

243 Monday, August 21, 2023

 Ex: Given the non-unity feedback system and step input function is applied.

Find the steady state error

244 Monday, August 21, 2023

 We fine the equivalent T.F. Ge

245 Monday, August 21, 2023

 Ex:
K
A second order servomechanism with unity feedback, has the open-loop T.F.G(s) 
s(s  4)
Find the gain K so that, the steady state error shall not exceed 0.4 degree when the
input shaft is rotated at 3 rpm

Solution: 
 2    rad/sec
Input speed is 3 rpm    3   10
 60 

t 
Hence the input r (t )  t  R(s) 
10 10 s 2
K
since G( s)  Velocity error constant K
s( s  4) K v  lim sG ( s) 
s 0 4
A  /10 4
Steady state error is ess   
Kv Kv 10 K


Given steady state limit is 0.40  ess  0.4
180
4 0.4
Hence,   K  180
246 10 K 180 Monday, August 21, 2023
 Lecture_17

Application of MATLAB and its Tool Box for time response

analysis

247 Monday, August 21, 2023

 Basics of MATLAB:

Some command

248 Monday, August 21, 2023

 Online help:

249 Monday, August 21, 2023

 Relational operator

line style option

250 Monday, August 21, 2023

 Basics of MATLAB:
The MATLAB is an interactive programme for engineering and scientific calculation
which includes base programme and variety of tool boxes.

The tool boxes are collection of special files, called M-files, which extends the functionality of
the base programme.

A variety of objects such as statement, matrices, graphics and scripts are allowed to interact with
the programme.

Statement has the form as shown

A typical statement is

251 Monday, August 21, 2023

 I f the statement is followed by a semicolon (;)
the output is not displayed.

When the variable name and (=) are omitted from an expression,
the result is assigned by a generic variable by ans

The user guide has complete list of available

trigonometric and elementary math functions.

252 Monday, August 21, 2023

With decreasing power of s

253 Monday, August 21, 2023

 MATLAB Application in control system:

Closed loop system

Feedback function

254 Monday, August 21, 2023

 Application of Step function:

255 Monday, August 21, 2023

256 Monday, August 21, 2023
257 Monday, August 21, 2023
 MATLAB in Time domain specifications:

For step
input

258 Monday, August 21, 2023

259 Monday, August 21, 2023
260 Monday, August 21, 2023
261 Monday, August 21, 2023
 Model conversion

Ex:

262 Monday, August 21, 2023

263 Monday, August 21, 2023
 Eigen values and eigen vectors

264 Monday, August 21, 2023

 Laplace and inverse Lapalce transform
Using MATLAB

Ex

265 Monday, August 21, 2023

266 Monday, August 21, 2023
 n  
 tp   tr 
d d

267 Monday, August 21, 2023

268 Monday, August 21, 2023
269 Monday, August 21, 2023
270 Monday, August 21, 2023
271 Monday, August 21, 2023
 an
d

272 Monday, August 21, 2023

273 Monday, August 21, 2023
274 Monday, August 21, 2023
 Mod_2

Lecture_18

Effect of poles and zeros on system response.

275 Monday, August 21, 2023

 The time domain specification were found for a second order, two complex poles and no
zeros
For more than two poles and presence of zeros can not be accounted for the finding of
time domain specifications.

The system for such are considered to be approximate as second order system by
locating the dominant poles.

276 Monday, August 21, 2023

  (1)

Larger is imaginary line i.e. less peak time, the more is overshoot.
So, the damping ratio is more. Hence, rise time
decreases and the settling time is less.

(2)

Settling time is inversely proportional to real part of the pole

The vertical lines represents the real part of pole

corresponding to its settling time.
277 Monday, August 21, 2023
 Effect of step response by changing the position of poles
at different location
of s-plane.
Fig-a: Moving the pole
on vertical axis. Settling time
is same with different natural
frequency.

Fig-b: Moving the pole

on horizontal axis.
The natural frequency is same.
As the pole moves to the left,
The response is very fast
and damps more quickly.

Fig-c: Moving the pole

along the constant radial line.
The percentage overshoot
remains same. 


The farther the pole,  %M p  e 1 2

 100%
fast is the response.
278 Monday, August 21, 2023
 Consideration of only one additional pole with complex pole:

The time domain output is

279 Monday, August 21, 2023

 Case_II

The pure exponential will die out much more rapidly than the second order under
damped step response. Thus , the total response will be the response duet to second
order response.
Case_I The exponential decay is significant and the system
can not be represented by second order system. The third pole has the effect to the overall
the response.

Case _III Thus , the total response will be the response duet to second order
280 Monday, August 21, 2023
response.
 If the real pole is 5 times father than the dominant pole, the system response is
represented by second order system.

Consider a second order system. A zero is added to the left half of s-plane in the transfer function.

C (s) ( s  a)
T (s)  
R( s) ( s  b)( s  c)

Adding a zero to the response C(s), we have

281 Monday, August 21, 2023
 “a” is very small, then the derivative term sC (s) contributes more to the response and has greater
ffect. For step response, the derivative is typically positive at the start of the response.

or small value of “a” , there is a more overshoot in the second order system because the derivative
erm will contribute more positive to the first overshoot of the response.

Where a= -3, -5, -10

282 Monday, August 21, 2023

 If “a” is positive small and is placed right side of s-plane, then the response
will initially follow derivative action sC(s) in opposite direction.

A system which exhibits this type of response is called non-minimum phase

system.

Example: A motor cycle initially steers left when commanded to the right.

283 Monday, August 21, 2023

284 Monday, August 21, 2023
285 Monday, August 21, 2023
286 Monday, August 21, 2023
287 Monday, August 21, 2023
 Minimum Phase TF

Non-Minimum Phase
TF

G1*G2= All-Pass Phase TF

288 Monday, August 21, 2023

Ex: Nonmimum phase electronic system, where step input is given

(1)

(2)

(3)

Putting the value of (1) in (3) we get (4)

(5)

Using (4) and (5) in (2)

After simplifying and rearranging (6)

289 Monday, August 21, 2023
 For step input
where

By partial fraction expansion, with zero

290 Monday, August 21, 2023

 Module_3

Lecture_19

Concepts of stability: Necessary conditions of stability

291 Monday, August 21, 2023

 Stability Criterion
Concept of stability: As we know that, the response of the system is given by

Using these concepts, we present the following definitions of stability, instability,

and
marginal stability:
A linear, time-invariant system (LTI) is stable if the natural response approaches
zero as time approaches infinity.

A linear, time-invariant system is unstable if the natural response grows without

bound as time approaches infinity.
A linear, time-invariant system is marginally stable if the natural response neither
decays nor grows but remains constant or oscillates as time approaches infinity.

Thus, the definition of stability implies that only the forced response remains as
the natural response approaches zero.
292 Monday, August 21, 2023
 Using the natural response:
1. A system is stable if the natural response approaches zero as time approaches
infinity.
2. A system is unstable if the natural response approaches infinity as time
approaches infinity.
3. A system is marginally stable if the natural response neither decays nor grows
but remains constant or oscillates.

Using the total response (BIBO):

1. A system is stable if every bounded input yields a bounded output.
2. A system is unstable if any bounded input yields an unbounded output.
hree circular cone resting on its base on a horizontal surface(equilibrium positions ) . Slightly tilted
rom their position gives the information of stability as follows.

293 Monday, August 21, 2023

 stable

stable

unstable

unstable

294 Monday, August 21, 2023

The stability of the system is checked adopting the forcing response by impulse response approach

since

With initial conditions assumed zero, the output of the system is given by

For open loop system, C ( s)  G ( s) R( s)

L1 (C ( s))  L1 (G ( s) R( s ))

the forcing function output c(t) due to input r(t) is known as convolution integral, and is
given by

where h(t) is the impulse response. When the impulse response g(t) is bounded and
satisfies the additional condition

then the integral will be finite for a bounded input |r

(t)|.
295 Monday, August 21, 2023
 i.e. For stable, the response of the system is based on the bounded input. i.e. Less
than some finite value for all time. t
c(t )   h( ) r (t   ) d
0
t
c(t )   M h( ) d
0

Since, the output c(t) can not exceed to input r(t) for stable
system

t
c(t )
 h( ) d    i.e. the impulse response is finite
0
M

296 Monday, August 21, 2023

 Pi is the poles of the transfer
function
297 Monday, August 21, 2023
 How do we determine if a system is stable?
If the closed-loop system poles are in the left half of the plane and hence have a
negative real part, the system is stable. That is, stable systems have closed-loop
transfer functions with poles only in the left half-plane.

Poles in the right half-plane (rhp) yield either pure exponentially increasing or
exponentially increasing sinusoidal natural responses. These natural responses
approach infinity as time approaches infinity.
Thus, unstable systems have closed loop transfer functions with at least one pole
in the right half-plane and/or poles of multiplicity greater than 1 on the imaginary
axis.
The marginally stable systems have closed-loop transfer functions with only
imaginary axis poles of multiplicity 1. When the pole is at origin, it is said to be
limitedly stable.
Whether the linear system is stable or unstable does not depend on the input or
driving
function of the system. The poles of the input do not affect the property of the
stability but
they contribute only the steady state response.

The
298 control system should not have closed loop poles on the imaginary axis
Monday, August 21, since
2023 in
practical
 It is to be noted that, all closed loop poles lie in the left half of s-plane does not guarantee
satisfactory transient characteristics. If the dominant complex conjugate closed loop pole
lies close
to origin or jw axis, the transient response may exhibit excessive oscillations or may be
very slow.

To guarantee fast, well damped, transient response characteristics, it is necessary that, the
closed loop
The
poles lie in a particular region in the complex closed
plane, loop
such polesisare
region found from
bounded by thethe
shaded
area. characteristic equation of the closed loop system

299 Monday, August 21, 2023

 By factorization,

For R(s) as step input

The magnitude of the relative residues depends on the location of the closed loop pole in left half
of s-plane.
300 Monday, August 21, 2023
 a pole is located far away from the origin in left half of s-plane, then the residue for this pole is sma
The transient for such a remote pole is small and last for a short time.

If there is a closed loop zero closed to a closed loop pole, then the residue at this pole is small.

A pair of closely located closed loop pole and zero will be effectively cancel each other.

The characteristic equation for second order system is given by

he higher order system can be approximated to second order system for the transient response analy
301 Monday, August 21, 2023
 Location of Roots and their corresponding responses

302 Monday, August 21, 2023

303 Monday, August 21, 2023
304 Monday, August 21, 2023
305 Monday, August 21, 2023
 Module_3

Lecture_20

Testing for stability, Hurwitz stability criterion,

Routh stability criterion

306 Monday, August 21, 2023

 Testing for stability: (Necessary condition for stability)
• All the coefficients of the characteristic equation in s be real and positive.
•None of the coefficients should be zero.
some of the coefficients are zero or negative it can be concluded that, the system is unstable.

n n 1 n2
Let the characteristic equation is F ( s)  a0 s  a1s  a2 s  ....... an 1s  an  0

Where all coefficients are real and positive and

Hurwitz Criterion: (Necessary and Sufficient conditions)

The Hurwitz criterion is in terms determinants which is developed from the coefficients from the
Characteristics equation.
n n 1 n2
Let the characteristic equation is F ( s )  a0 s  a1 s  a 2 s  ....... an1s  an  0

For the stability of the system, it is necessary that the n determinant s formed from the coefficients
of the characteristic equation be positive, where these
determinants are
taken as the principal minors of the following arrangement called Hurwitz determinant
307 Monday, August 21, 2023
 Method I F ( s)  a0 s n  a1s n1  a2 s n2  ....... an 1s  an  0
 a1 a3 a5 . . . . a2 n 1 
a a2 a4 . . . . a2 n  2 
 0
0 a1 a3 . . . . a2 n  3 
 
0 a0 a2 . . . . a2 n  4 
The coefficients are arranged as H 
0 0 a1 . . . . a2 n  5 
 
0 0 a0 . . . . a2 n  6 
. . . . . . . . 
 
 0 0 0 . . . . an 

Here, the element of the determinant decreases by two along the row by one down the
column. For stability, the following conditions must satisfy
a1 a3 a1 a3 a5 . . . . a2 n 1
1  a1  0 2  0
a0 a2 a0 a2 a4 . . . . a2 n  2
0 a1 a3 . . . . a 2 n 3
0 a0 a2 . . . . a2 n  4
. .
a1 a3 a5 n  0 0 a1 . . . . a2 n 5  0
 3  a0 a2 a4  0 0 0 a0 . . . . a2 n  6
0 a1 a3 . . . . . . . .
0 0 0 . . . . an

308 Monday, August 21, 2023

 3 2
Example: F (s)  s  s  s  4  0

Check the stability using Hurwitz determinant

Solution: a1 a3 a5 1 4 0
H  a0 a2 a4  1 1 0 
0 a1 a3 3*3
0 1 4  3*3

Sub-determinant D1  1  0;

1 4 
D2     1  4  3  0;
1 1 

1 4 0
D3  1 1 0   1( 4  0)  4( 4  0)  0(1  0)  4  16   12  0
0 1 4 

Sub-determinant D2 and D3 are negative. Hence, the system is unstable

309 Monday, August 21, 2023
 Example:
Find stability of the following system given by and

using Hurwitz stability criterion

Solution:

The characteristic equation is F ( s)  a0 s 2  a1s  a2  0

Here, a0  1 a1  1 a2  K a3  0 a4  0 a5  0
a1 a3 a5 1 0 0
Hurwitz determinant is H  a0 a2 a4  1 K 0 
0 a1 a3 3*3
0 1 0  3*3
1 0 0
D1  1  0 1 0  D3  1 K 0  1(0)  0(0)  0(1)  0
The sub-determinant D2     K 0  0
1 K  0 1 0

310 Monday, August 21, 2023

 Example_2:
CLTF

The characteristic equation is F ( s)  a0 s 3  a1s 2  a2 s  a3  0

F ( s )  s 3  6 s 2  8s  K  0
a0  1 a1  6 a2  8 a3  K a4  0 a5  0

Sufficient condition for stability through Hurwitz determinant is

a1 a3 a5 6 K 0
H  a0 a2 a4  1 8 0  D1  6  0 6 K
D2   48  K   0
0 a1 a3 0 6 K  3*3 1 8
3*3

6 K 0
D3  1 8 0  6(8 K  0)  K ( K  0)  0  48 K  K 2  K ( 48  K )
0 6 K

So, the system will be always stable forK  48

311 Monday, August 21, 2023
Drawback of Hurwitz determinant :

 This method is time consuming for higher order system

Roots located at right hand s-plane is not determined
Difficult to predict marginal stability

312 Monday, August 21, 2023

Routh’s /Routh-Hurwitz Table: (Necessary and Sufficient Conditions)

We first create the Routh table in array form. Begin by labeling the rows with powers of
s from the highest power of the denominator of the closed-loop transfer function to s°.
Next start with the coefficient of the highest power of s in the denominator and list,
horizontally in the first row, every other coefficient. In the second row, list horizontally,
starting with the next highest power of s, every coefficient that was skipped in the first
row. F ( s)  a s n  a s n 1  a s n 2  ....... a s  a  0
0 1 2 n 1 n

313 Monday, August 21, 2023

Similarly, the coefficient of any row is obtained by considering the previous two rows.

314 Monday, August 21, 2023

 Example:

Solution:

Note: Most often we use Routh-Hurwitz replacing Routh array also.

315 Monday, August 21, 2023

 Example:

316 Monday, August 21, 2023

 Example:

The complete array is

317 Monday, August 21, 2023

 Example: Let us now shift one of the poles at origin

So the previous system becomes

The characteristic polynomial becomes

Note: If the last element of first column contains zero, one pole is at origin.
318 Monday, August 21, 2023
 While preparing the table, we may encounter four different cases which require some
modifications of the procedure for computing the elements of the array. The distinct
special cases are:
(1) a zero in the first column, but other elements in the row containing the zero
are nonzero;
(2) Row containing all zero element
(3) as in case (2) plus repeated roots on the jω-axis.

In presence of the cases mentioned above, we will face difficulties in completing the
table. We shall consider examples to illustrate the procedure for overcoming the
difficulties in.
Case 1: A zero in the first column.

a first column in any row is zero, but the remaining terms are not zero, or there is no remaining
rms, then the zero term is replaced by a small positive number and the rest array are evaluated.
Method-1
(i) Replace 0 by ε (very small number) and complete the array with ε.

(ii) Examine the sign change by taking

319 Monday, August 21, 2023
 Ex:

320 Monday, August 21, 2023

321 Monday, August 21, 2023
 Ex:

Solution:

322 Monday, August 21, 2023

323 Hence, the system is unstable Monday, August 21, 2023
 Case-2 All elements in a row is zero
A second special case exists when all the elements in a row of the Routh array are
zero. This is caused when the polynomial contains singularities that are symmetrical
located about it’s real and imaginary axes in s-plane namely:

(i) One pairs of roots on the jw-axis

If all the elements of a particular row are zero, then further row can not be created. So, the
zero elements are replaced by the derivatives of the auxiliary polynomials, formed from the
just previous row.

The polynomial whose coefficients are the entries in the row above the row of zeros is called an
Auxiliary polynomial and the roots of this polynomial give symmetrical locations of poles on
Imaginary axis.

324 Monday, August 21, 2023

 Solution :

325 Monday, August 21, 2023

 A( s )  s 4  5s 2  5  0 s1, 2   j1.1756 and s3, 4   j1.902

326 i.e. roots are not repeated on imaginary axis Monday, August 21, 2023
 (i) Repeated roots on imaginary axis
D(s)  s 6  3s 5  4s 4  6s 3  5s 2  3s  2  0
Routh-Hurwitz Table
AE just above row of all zeros
ROZ1
AE1  2s 4  4s 2  2

d AE1 
 8s 3  8s
ds

ROZ2
AE2  2s 2  2

d AE2 
 4s
No sign change in 1st column ds
ROZ are two i.e. there are two pairs of imaginary roots
 Total no. of poles are 6. Out of 6, 4 poles are symmetrical and repeated and
2 are on negative half of s-plane
No poles are on right half of s-plane
 If poles are repeated on imaginary axis, then the system is unstable
ROZ always comes on odd power of s
327 Monday, August 21, 2023
 AE1  2s 4  4s 2  2  0

s 4  2s 2  1  0

s   2 * s
2 2 2 2
 
*1  1  s 2  1  0
2


 s2 1  0 2

 s1, 2   j1 and s3, 4   j1

When the poles are repeated on imaginary axis, then , the system is unstable.

Rest two roots are found on the negative of real axis, which is justified as

Divide the auxiliary equation to the original

equation
s 6  3s 5  4s 4  6 s 3  5s 2  3s  2
4 2
 s 2  3s  2
328 2s  4s  2 Monday, August 21, 2023
 Now check the routh array from the result

s 2  3s  2

System is unstable

No sign changes in first column. So all the two roots are lying in the left half s-
plane
s 2  3s  2  0

s1, 2  1,2

Assignment:

329 Monday, August 21, 2023

 (ii) One or more real roots with opposite sign
D( s )  s 4  s 3  s  1  0 Plot the poles and check the stability

Solution:

AE just above row of all zeros

AE  s 2  1
d AE 
 2s
ds

Note:
 Total poles=4
No. of sign changes 1 means no. of pole is 1 at right s-plane
RHS pole=1
The auxiliary equation provides the symmetrical 2-pole location on real axis
Rest 2-poles are on real axis of negative half s-plane
330 Monday, August 21, 2023
 (iii) One or more pairs of complex roots with their mirror images
About the jw-axis, together forming quadrates in the s-plane

D( s)  s 6  s 5  7 s 4  6s 3  31s 2  25s  25

A( s)  s 4  6s 2  25

d  AE 
 4s 3  12s
ds

Now, two sign changes in 1st column. Hence, two poles are on right side of s-plane.
The system is unstable. To check the auxiliary poles, we have to do

Factorize the auxiliary polynomials as A( s)  s 4  6s 2  25  0

331 Monday, August 21, 2023

Monday, August 21, 2023 332
Roots forming a quadrate s3,4  1  j 2 s1 , 2   1  j 2
 s  1  j 2 s  1  j 2 s  1  j 2 s  1  j 2 
 s2  5  2s s2  5  2s
  
2 2
 s2  5
   2 s 
2
 s2  5
   4s2
2 2
s 5   6 s 2  10 s 2
A( s)  s 4  6s 2  25
333 Monday, August 21, 2023
 Ex:
Find the range of K of the following system such that, the system will be stable ,
unstable and
marginal stable.

Solution:

Since K is assumed positive, we see that all elements in the first column are always
positive except the s1 row. This entry can be positive, zero, or negative, depending upon
the value of K. If K < 1386, all terms in the first column will be positive, and since there
334are no sign changes, the system will have three poles in the left half-plane
Monday,and be 21,
August stable.
2023
 If K > 1386, the s1 term in the first column is negative. There are two sign changes,
indicating that the system has two right-half-plane poles and one left half- plane pole,
which makes the system unstable.
If K = 1386, we have an entire row of zeros, which could signify jw poles. Returning to
the s2 row and replacing K with 1386, we form the even polynomial

(1)

(2)

(2) we obtain the Routh-Hurwitz

Table for K=1386

Since there are no sign changes from the even polynomial (s2 row) down to the
bottom of the table, the remaining root is in the left half-plane. Therefore the system
335 is marginally stable. Monday, August 21, 2023
 Alternate Method
The procedure to be carried out when a row of zeros occurs is (a) form the auxiliary
Polynomial (b) divide the original polynomial by the auxiliary polynomial and (c) test the remainin
Polynomial by means of the Routh array.
The Routh-Hurwitz criterion is often used in limited applications to factor polynomials
containing even factors.

336 Monday, August 21, 2023

337 Monday, August 21, 2023
 Module_3
Lecture_21

Advantages and disadvantages of Routh-Hurwitz , Relative

stability Analysis, Application of Routh stability criterion to
linear feedback systems,

338 Monday, August 21, 2023

 Advantages and disadvantages of Routh-Hurwitz
Advantages of Routh-Hurwitz stability

(i) Stability can be judged without solving the characteristic equation

(ii) Less calculation time
(iii) The number of roots in RHP can be found in case of unstable condition
(iv) Range of value of K for system stability can be calculated
(v) Intersection point with the jw-axis can be calculated
(vi) Frequency of oscillation at steady-state is calculated

Dis-advantages of Routh-Hurwitz stability

(i) It is valid for only real coefficient of the characteristic equation

(ii) Unable to give exact locations of closed-loop poles
(iii) Does not suggest methods for stabilizing an unstable system
(iv) Applicable only to the linear system
(v) Higher order system is not easily factorized.
339 Monday, August 21, 2023
 Relative Stability Analysis:
Routh_Hurwitz stability criterion provides information whether all roots lie in the left-
hand side of the s-plane or not. Sometimes, the engineer is interested to know
information about relative stability, specially in presence of parameter variation in the
process due to environmental changes or change in operating conditions, which will
affect one or more of the coefficients in the characteristic equation.
The relative stability may be found by shifting the imaginary axis to the left by some
constant
amount σ
This can be easily done by replacing s by s + σ in the characteristic equation and
applying the Routh-Hurwitz criteria.

In this case, the number of changes of sign in the first column of the array developed for
the polynomial in s corresponds to the number of roots that are located to the right of
the vertical line s = – σ.

Thus, this test reveals the number of roots that lie to the right of the vertical line s = – σ.

So, a large value of σ on the negative axis indicates that the system has high relative
stability and is less prone to be unstable due to small parameter variation.

340 Monday, August 21, 2023

 Introducing the constant real value , the concept of relative stability can be viewed as one
in which the characteristic equation is tested to find out whether or not any roots lie to the
right of an imaginary axis which passes throughσthe point s = - , where,

Example: How many roots of the characteristics equation

Lie to the right of s=-1 in the s-plane.

341 Monday, August 21, 2023

 Solution:

A system with such a denominator polynomial is therefore a stable system because there
are no changes of sign in the first column of the Routh array.

The imaginary axis is now shifted from s=0 to s= - 1 by making the substitution s = r -
1, the new polynomial equation is

and although it is now apparent by the negative sign in the polynomial that at least one
root of this polynomial lies to the right of the point r = 0, the Routh array will be formed
for completeness.

342 Monday, August 21, 2023

From the Routh array it is apparent that as only one change of sign occurs in the first
column, so only one root of the polynomial lies to the right of the r = 0 (s = - 1) point.
Also, from the previous Routh array it was shown that no roots exist to the right of the
s = 0 point, thus the root in question must lie between the s = - 1 and s = 0 points.
Hence , no root of the characteristic equation lie to the right of s= - 1 in the s-plane.

Application of Routh-Hurwitz Stability Criterion to Linear feedback

system

Routh-array is
The characteristic equation is

We require for stability

343 Monday, August 21, 2023

 Module_3

Lecture_22

Root locus techniques: Evan

condition, Properties of root locus

344 Monday, August 21, 2023

 Root Locus concept on 1st order system
Routh-stability criterion was only giving the range of K for which the system is stable or not,
but root locus gives, the location of poles for particular value of K on the root-locus.
When any parameter of the system like, pole, zero or gain changes, the closed loop
pole of the system changes. This change in closed loop pole migrates in the s-plane.
The locus of the migration is called root locus.

 Root locus method is a graphical representation of the closed –loop poles ort he roots
of the characteristic equation when the gain of the system varies from zero to infinity
 It is a time-domain analysis for stability study
 It provides a measure of sensitivity of roots to the variation in the parameters being considered.
 The loci always starts at open loop poles and terminates at open loop zeros

345 Monday, August 21, 2023

 K
Open loop pole TF G ( s) H ( s) 
Ts
The open loop pole s=0
The characteristic equation gives the closed loop pole. The characteristic equation is

K
1  G ( s) H ( s)  1   Ts  K  0
Ts
K
s
T
So, by varying the value of gain K from zero to infinity, various values of the closed loop
Poles are obtained.
346 Monday, August 21, 2023
 Root-Locus diagram

It is observed that, as K is increased, the system is more responsive. Practically, K has the
limitation to avoid the system malfunctioning

347 Monday, August 21, 2023

348 Monday, August 21, 2023
349 Monday, August 21, 2023
350 Monday, August 21, 2023
351 Monday, August 21, 2023
 Root Locus of a feedback system:

Any value of s that is root of the characteristic equation satisfy the both the magnitude as well
as angle condition as proposed by Evan in 1948.

However, the angle condition is more significant than the magnitude condition. After finding the an
We go for K calculation.

352 Monday, August 21, 2023

 Ex:

This does not satisfy the angle criterion. Hence, the point s=-1+j4 is not on the root locus for the
Given system. After knowing the angle condition, the magnitude condition is applied know
the value of gain K.

353 Monday, August 21, 2023

 Root locus techniques: (Time domain design tool)
The root locus technique is used to analyze and design the effect of loop gain (K) upon
the system's transient performance and stability.
By varying only the system gain or amplifier gain keeping all other parameters constant,
the closed-loop poles migrates in s-plane and the locus of the migration points is called
root locus. The system behavior is associated with the closed loop poles.
This plot give absolute stability and relative stability information.

Example: The open loop transfer function of a unity feedback system

is,

Determine the location of its closed loop poles for increasing values of K from 0 to .
Also determine the limiting value of K beyond which the closed loop system with unity
feedback becomes unstable.
Solution:
All the three poles of the plant, S1 = - 1, S2 = - 2 and S3 = - 3, are negative. Hence, the
plant, i.e., the uncontrolled system (without feedback) is stable. To find the closed loop
poles we need to do feedback system and the effect of system gain K on the closed loop
354
poles. Monday, August 21, 2023
 The transfer function of the closed loop system is,

The characteristic equation of the closed loop

system is,
So,

In order to obtain the roots or the poles of the closed loop system, we must factorize the
above closed loop characteristic equation for different values of K.

( 1 ) K = 0: The closed loop characteristic equation becomes the plant characteristic

equation. Hence, the closed loop poles will be the same as the open loop poles. That is,
S1 = - 1, S2 = - 2, S3 == - 3.

(2) K = 6: The characteristic equation is

Thai is,

(3) K=24: Thus,

355 Monday, August 21, 2023
 (4) K=60:

That is,

The locus of the closed loop poles for increasing values of K is

shown
Such a diagram, showing the movement of the roots of the closed loop characteristic
equation, is also called a root locus diagram. This figure shows that for values of gain K
greater than 60, the closed loop poles will enter the r.h.s. of the , s -plane. The system
will then become unstable.
An unstable system cannot perform any useful
function. It is dangerous and may damage to the
plant. Therefore, it is necessary that the feedback
control system be so designed that it does not
become unstable under any condition of
operation.

356 Monday, August 21, 2023

 Evan Conditions (Magnitude and Angle conditions) :
For constructing root locus Evan conditions must be satisfied.

The closed loop transfer function given by Equation

The roots of the characteristic equation is made by

A pole, “s” exists when the characteristic polynomial in the denominator becomes
zero.
Alternately, a value of “s” (a complex number) is a closed-loop pole if

The complex of may be represented by polar form as magnitude and angle

form

where
Angle condition:

For any value of “s” a complex number , if the angle is an odd multiple of 180°, that
357 value of “s” is a system pole for some particular value of K. Monday, August 21, 2023
 Magnitude condition
Since the angle criterion is satisfied, all that remains is to satisfy the magnitude
criterion, Scalar
value

Example:

The open loop pole-zero location of the system is

The closed-loop transfer function, T(s), is

If point “s” is a closed-loop system pole for some value of gain K, then “s” must
satisfy the magnitude and angle conditions.

Consider the point s= -2 + j 3 in s-plane. If this point is a closed-loop pole for some
value of gain K, then the angles of the zeros minus the angles of the poles must equal
an odd multiple of 180°.
358 Monday, August 21, 2023
 The closed-loop system is

The location of the closed loop pint s= -2+j3 is located in s-plane to find the angle criterion.
From the figure angle condition at the point
s= -2 + j 3 in s-plane
3
1  tan 1    56.300  2  tan 1 3  71.570 3  900
2

 4  1800  tan 1 3  1800  71.570  108.430

Therefore, - 2 +j 3 is not a point on the root locus, or

alternatively, - 2 +j3 is not a closed-loop pole for any
gain. So, magnitude condition also not satisfied.
We now proceed to evaluate that value of gain.

K is not on the RL path

359 Monday, August 21, 2023

 Example: K
G( s) 

s s 2  7 s  12 

Solution:
K K
G( s)  

s s 2  7 s  12  ss  3s  4

K
G (1  j1) 
 1  j1 1  j1  3 1  j1  4

K
G (1  j1) 
 1  j12  j13  j1

360 Monday, August 21, 2023

 Angle condition:

1
1  (1  j1)  180 0  tan 1    135 0
1

1
2  (2  j1)  tan 1    26.56 0
2

1
3  (3  j1)  tan 1    18.430
3

G(1  j1)  1  2  3  1350  26.560  18.430  1800

Thus, angle condition is satisfied

Then, we go calculation of gain K

361 Monday, August 21, 2023

 Magnitude Condition
G( s)  1  j 0 G( s)  1 G(1  j1)  1

K
G (1  j1) 
 1  j12  j13  j1

K
G (1  j1)  1
(1  j1) (2  j1) (3  j1)

K K
G (1  j1)  1 G (1  j1)  1  K  100  10
2 5 10 100

Alternate Method (graphical method)

If zero in OLTF is not given, take vector length of zero as unity

2 5 10
K  100  10
1
362 Monday, August 21, 2023
 Example : K ( s  2)
G(s) 
( s  1)(s  3)

The characteristic
equation
( s  1)(s  3)  K ( s  2)  0

For K = 0 the two loci will therefore start at the open-loop poles, i.e. s = - 1 and -3. If
we are just considering K > 0, then the two loci will end at s= - 2 and s=

Consider the point s  1.4

The angle criterion is satisfied because the angle to that point from both the zero at s = - 2
and the pole at 0s 0= -3 is , whereas the angle from the poleat s= - 1 is .
0
Clockwise direction =0
Anti-clockwise direction =  
We then have G(s)  00    00  

and from the magnitude criterion

363 Monday, August 21, 2023

 in which 0.6 is the magnitude to the point at s = - 1.4 from the zero at s = - 2, and 0.4 is
the magnitude from the pole at s = - 1 whereas 1.6 is the magnitude from the pole at s =
- 3.
Hence, that a value of K = 32/3 will result in a root on the locus at s = -
1.4
Properties of Root locus plot: The following properties to be followed for construction of R-L
The root locus starts at open loop poles of G(s)H(s) at K = 0

Characteristics equation of a closed-loop system

For K=0,

Open-loop poles are also closed-loop poles for K=0. A root-locus starts from every open-
loop pole.
364 Monday, August 21, 2023
 The root locus ends at open loop zeros of G(s)H(s) at

Characteristics equation of a closed-loop system

Root-locus ends at an open-loop zero or at infinity.

 A point on the real axis lies on the root locus if the sum of open loop poles and zeros on the
Real axis to the right hand side of this point is odd.

Ex: K ( s  1)(s  4)
G(s) 
s ( s  3)(s  5)

365 Monday, August 21, 2023

 STEPS FOR SOLVING ROOT LOCUS DIAGRAM

366 Monday, August 21, 2023

  Step1: Number of Root Locus Branches

K
Example: G ( s)  No. of open loop poles P3
s ( s  1)( s  2)

So, since PZ

Z 0 B P3
No. of open loop zeros

Step-2: Plot the open loop poles and zeros

367 Monday, August 21, 2023

  Step-3: Find the existence of real axis loci

 step-4: Number of Asymptotes and their angle

A PZ
AZ P

A0

Note: Asymptotes are the guidelines for plotting the root locus from the centroid

For K
G(s)  A PZ 3
s( s  1)( s  2)
368 Monday, August 21, 2023
  Step-4: Angle of Asymptotes:

the root locus branches are approximated by asymptotic lines with asymptotes angles
given by
A 
2q  11800
PZ

where q = 0, 1, 2, ... (P– Z – 1); P is the number of finite poles of G(s)H(s), and Z is the
number of finite zeros of G(s)H(s).

q=0,1,2 as P-Z-1=2

However, the angle of asymptotes are drawn from the

centroid point of root locus

A1 
2q  1180 0

180 0
 60 0
PZ 30 q=0

A2 
2q  11800

3 *1800
 1800 q=1
PZ 30

A3 
2q  1180 0

5 *1800
 300 0 q=P-Z-1=3-1=2
PZ 30
369 Monday, August 21, 2023
 K
 Step-5 Centroid of RL G(s) 
s( s  1)( s  2)

centroid( ) 
 Re(P)   Re(Z ) 
1  2
 1
PZ 3
Note: Center may or may not be a part of the root locus

90% of RL having centroid on RL

370 Monday, August 21, 2023

  Step-6: Break Away Point (BAP)/Break In Points (BIP)
The root locus between two open loop poles move towards each other as the gain factor K
is increased till they are coincident. At the coincident the value of K is maximum. Any
further increased in K break the root locus into two parts.
The break away/break in points are found by writing the characteristic equation and
there form solving the value of s by

Break away point appears between two adjacent poles and break in point occurs between two adjac
zeros.

Breakaway points on the real axis Breakin points on the real axis
371 Monday, August 21, 2023
Finding of BAP (two adjacent poles must be the RL branch)
1  G ( s) H ( s )  0
K
1 0
s( s  1)(s  2)

s 3  3s 2  2s  K  0

K   s 3  3s 2  2s
dK
 3s 2  6s  2
ds
For getting valid point of BAP on RL

dK
 0  3s 2  6s  2  0  3s 2  6s  2  0
ds
 6  62  4 * 3 * 2 12
s1, 2   1   0.43,1.57
2*3 6

However, s1  0.43 Valid BAP

s2  1.57 Invalid BAP
372 Monday, August 21, 2023
  Intersection points with imaginary axis (jw).

The value of K and the point at which the root locus branch crosses the imaginary axis is
determined by applying Routh criterion to the characteristic equation. The roots at the
intersection point are imaginary. This is called K marginal value.
The characteristic equation is s(s  1)(s  2)  K  0

s 3  3s 2  2s  K  0

373 Monday, August 21, 2023

 Also, for 3rd order system only s 3  3s 2  2s  K  0

For oscillatory system

Internal product = External product

Hence, K 6

374 Monday, August 21, 2023

 When the closed loop poles lie on the imaginary axis, the system gives sustained
oscillation.
Step-8: Angle of Departure or arrival

Not exist as there is no complex poles and zeros

375 Monday, August 21, 2023
 Step by Step Procedure to draw the Root Locus
diagram:
Draw the root-locus of the feedback system whose open-loop transfer function is given by

Step 1: Determine the number of open-loop poles and zeros

Number of open-loop poles n=3

Number of open-loop zeros m=0
Open-loop poles: s=0, s=-2 and s=-4

Step 2: Mark open-loop poles and zeros on the s-plane

Step 3: Determine parts of the root-locus on the real

axis

376 Monday, August 21, 2023

 Step 4: Determine breakaway and break-in point

Characteristic equation,

Breakaway point as  b  0.85  b  3.15

But σb = –3.15 is not on the root-locus and therefore not a breakaway or break-in point.
 b point
Gain at the breakaway 0.85

377 Monday, August 21, 2023

Step 5: Draw asymptotes of the root-locus

A 
2q  11800
PZ

centroid( ) 
 Re(P)   Re(Z )
PZ

Steps 6 & 7: Since there are no complex open-loop poles or zeros, angle of departure
and arrival need not be computed
378 Monday, August 21, 2023
 Step 8: Determine points on the root-locus crossing imaginary axis

K
1  GH  1  0  s 3  6s 2  8s  K  0
s( s  2)(s  4)

A system with such a denominator is therefore stable as long as

K>0
Hence when K =48 roots will exist on the imaginary axis, and these are the points we
are looking for. Substituting K = 48 into the characteristic polynomial equation, results
in
6s 2  48  0  6s 2  48 s    8   j 2.83 rad/sec
379 Monday, August 21, 2023
 s   8   j 2.83

For K=48

380 Monday, August 21, 2023

 Equating both real and imaginary parts of last equation, we get,

and

Hence , we get

Substituting in ,we get in

K=48
381 Monday, August 21, 2023
 Draw the root-locus of the feedback system whose open-loop transfer function
is given by

Step 1: Determine the number of open-loop poles and zeros

Number of open-loop poles n=3
Number of open-loop zeros m=0
Open-loop poles: s=0, s=0 and s=-1

Step 2: Mark open-loop poles and zeros on the s-plane

382 Monday, August 21, 2023

 Step 3: Determine parts of the root-locus on the real axis
Test points on the positive real axis

Step 4: Determine breakaway and break-in point

383 Monday, August 21, 2023

 Step 5: Draw asymptotes of the root-locus
Angle of asymptotes:

384 Monday, August 21, 2023

 Steps 6 & 7: Since there are no complex open-loop poles or zeros, angle of departure
and arrival need not be computed.

Step 8: Determine points on the root-locus crossing imaginary axis

385 Monday, August 21, 2023

  Angles of departure from complex poles and angles of arrival at complex
zeros
The root locus begins to departs from at an angle called departure angle when there is
complex pole in s-plane. Similarly, The root locus arrives to at an angle called arrival
angle when there is complex zeros in s-plane.

Angle of departure  d  1800   P   Z

Angle of Arrival
 A  1800   P   Z

386 Monday, August 21, 2023

 BP4

387 Monday, August 21, 2023

388 Monday, August 21, 2023
389 Monday, August 21, 2023
390 Monday, August 21, 2023
  0.5  j 0.865

Calculation of angle of departure froms  0.5  j 0.865

391 Monday, August 21, 2023

 Angle of departure

d  1800  argG(s) H ( s)  1800  2400  600

392 Monday, August 21, 2023

393 Monday, August 21, 2023
394 Monday, August 21, 2023
395 Monday, August 21, 2023
396 Monday, August 21, 2023
 Assignments

Ex

Ex

Ex

397 Monday, August 21, 2023

398 Monday, August 21, 2023
 Assignment

399 Monday, August 21, 2023

 Assignment

 1  G(s) H (s)  0

s 3  10s 2 Ks  K
 3  0
s  10s 2 s 3  10s 2

Assignment K ( s  1)
 1 0
s 2 ( s  10)

400 Monday, August 21, 2023

 Example

401 Monday, August 21, 2023

402 Monday, August 21, 2023
403 Monday, August 21, 2023
404 Monday, August 21, 2023
405 Monday, August 21, 2023
 Angle of departure Root Locus of

406 Monday, August 21, 2023

 Example:

With Zeros

Solution:

Since B=3, as P>Z

407 Monday, August 21, 2023

408 Monday, August 21, 2023
409 Monday, August 21, 2023
410 Monday, August 21, 2023
 Example:

411 Monday, August 21, 2023

 Sketch the root locus

Solution:

412 Monday, August 21, 2023

413 Monday, August 21, 2023
414 Monday, August 21, 2023
415 Monday, August 21, 2023
416 Monday, August 21, 2023
 Angle of Arrival :

Solution:

417 Monday, August 21, 2023

  A  1800   P   Z

418 Monday, August 21, 2023

419 Monday, August 21, 2023
420 Monday, August 21, 2023
 Example:
G ( s) H ( s) 
K ( s  2)( s  4) 2
For, s  2s  2 The complex conjugate poles are
s 2  2s  2
s1  1  j1
Solution: s2  1  j1

Departure angle is at
s1  1  j1

and s2  1  j1

d  1800  

where    p  z

421 Monday, August 21, 2023

  p1  900
 z1  tan 1 (1)  450
1
 z 2  tan 1    18.430
3

   p1  z1  z 2
  900  450  18.430  26.430

d  900  450  1800  26.430  153.430

anticlockwise
422 Monday, August 21, 2023
 K ( s  2)(s  4)
G ( s) H ( s) 
s 2  2s  2

No. of open loop poles=2

The Characteristic equation is
No. of open loop zeros=2
K ( s  2)( s  4)
1  G ( s) H ( s)  1  0
No. of RL branches = 2 s 2  2s  2

No. of Asymptotes A= P-Z=0 K ( s  2)( s  4)

2
 1
s  2s  2
Asymptotes angle =0
 ( s 2  2s  2)
K
Centroid=0 s 2  6s  8

Break In Point exists

423 Monday, August 21, 2023

 To know BIP dK
0
ds

 ( s 2  2s  2)
K
s 2  6s  8

dK   
 s 2  6s  8 2s  2  s 2  2s  2 2s  6
 

0
ds 
2
s  6s  8 2
 

s 2
 6s  82s  2  s 2  2s  22s  6

s 2
 3s  1  0

3 9 4 3 5
s1, 2    0.38,2.61
2 2

Where, -2.61 is valid BIP point as it lies on RL branch

424 Monday, August 21, 2023
-2.61 is the valid BIP point which lies on the root locus.

The break in point lies in between the adjacent zeros

425 Monday, August 21, 2023

426 Monday, August 21, 2023
427 Monday, August 21, 2023
428 Monday, August 21, 2023
 Module_3

Lecture_24

GATE problem and solution, +ve feedback system

solution, complementary RL, Root Contour

429 Monday, August 21, 2023

 Ex: Whether s=-3, s=-5, s=-7, s=-9 lies on root locus or not of
K ( s  2)(s  6)
G( s) H ( s)  K 0
( s  4)(s  8)

Solution: Check Evan Conditions

G( s) H ( s )  1
Magnitude Condition: G( s) H ( s)  1

Angle Condition Condition: G( s) H ( s)  1800

K (3  2)(3  6) K (3  2)(3  6) K (1)(3)

Magnitude Condition: G( s) H ( s) s 3     ve
(3  4)(3  8) (3  4)(3  8) 1* 5
satisfying
K (1) * 3 G( s) H ( s)  1
As G( s) H ( s) s  3  1  1
1* 5

5
K 
3

S0, s=-3 lies on root locus

430 Monday, August 21, 2023

 Check s=-5
K (5  2)(5  6) K (3)(1)
G ( s) H ( s ) s  5    ve
(5  4)(5  8) (1)(3)

+ve value comes at 0 or 360 degree. Hence, not 180 degree

So, s=-5 is not on root locus

Check s=-7
K (7  2)(7  6) K (5)(1)
G ( s) H ( s) s  7    ve
(7  4)(7  8) (3)(1)

Hence, s=-7 lies on root locus

K (5) * (1) 3
G( s) H ( s) s  7  1  1 K 
(3) *1 5

431 Monday, August 21, 2023

 Check s=-9 G ( s ) H ( s ) s  9 
K (9  2)(9  6) K (7)(3)
  ve
(9  4)(9  8) (5)(1)

Hence, s=-9 is not on root locus

The existence of RL branch

432 Monday, August 21, 2023

Difference between +ve (CRL) and –ve (RL) feedback system

For +ve feedback system, Complementary root locus is required

433 Monday, August 21, 2023

434 Monday, August 21, 2023
 Example: K ( s  1)
G( s) H ( s)  Draw the root locus plot for +ve feedback system
( s  2)(s  3)

 G(s) H (s)  1
Note: For negative feedback system, 1  G( s) H ( s)  0

G( s) H ( s)  1800

1  G( s) H ( s)  0  G( s) H ( s)  1
For positive feedback system,

G (s) H (s)  00

435 Monday, August 21, 2023

 For positive feedback system,
K ( s  1)
1 0 s 2  5s  6
s  2(s  3) K 
s 1

dK
0
ds

2 44
s1, 2   1  1.414  2.414,0.014
2

No. of Asymptotes A=P-Z=2-1=1

Angle of Asymptotes

2q *1800
A   00
PZ

As q=P-Z-1=0

436 Monday, August 21, 2023

 RL exists when the sum of poles and zeros are even to the right of any point in root
locus
for +ve feedback system.

Virtual zero is considered to be a nothing for counting the RL branch. i.e it is a

0, so even
437 Monday, August 21, 2023
 Example:

Plot the RL for a –ve feedback system

Solution: Root-Locus plot

Open loop pole- zero location

438 Monday, August 21, 2023

439 Monday, August 21, 2023
 Example:
Plot the root locus for negative feedback system

s 2  2s  25  0 Open loop pole P = s1, 2  1  j 4.9 Open loop zero Z=1

(2q  1) *1800
A=P-Z=2-1=1 A   1800
PZ

Root-Locus

440 Monday, August 21, 2023

 Departure angle:

d  1800  
  z   p
z  (1800  tan 1 4.9)  1010
 p  900
  z   p  1010  900  110

d  1800  110  1690

441 Monday, August 21, 2023

 Example: G(s) H (s) 
K ( s  1)
0
s  2(s  3)
Plot the root locus for positive feedback system

A  P  Z  2 1  1

2q *180 0
A   00
PZ

Since q=P-Z-1=0 Centroid=0

Virtual zero is considered to be a nothing for

counting the RL branch. i.e it is a 0, so even

442 Monday, August 21, 2023

 Example K
G ( s) H ( s) 
( s  1)s  2( s  3)

Solution
RL branch B=P=3

Asymptotes A=P-Z=3

2q *1800 q=0,1,2,…P-Z-1
Asymptote angle A 
PZ

1  2  3
Centroid   2
3

443 Monday, August 21, 2023

 Asymptote angle

2q *1800 q=0,1,2,…P-Z-1
A 
PZ

00 q=0
A1   00
3

2 *1800 q=1
A2   1200
3

2 *1800
A3   2400 q=2
3

444 Monday, August 21, 2023

 Break Away Point

The characteristic equation is

K
1 0  ( s  1)(s  2)(s  3)  K  0
( s  1)(s  2)(s  3)

K  ( s  1)(s  2)(s  3)

K  s 3  6s 2  11s  6

For BAP, dK 3s 2  12s  11  0

0
ds

 12  144  4 *11 * 3  6  3
s1, 2    2.57,1.43
6 3

However, -2.57 is the valid BAP

445 Monday, August 21, 2023
 Root Locus for K
G( s) H ( s) 
( s  1)s  2( s  3)

446 Monday, August 21, 2023

 System with Delay

so, one has to do the Pade’s approximation

447 Monday, August 21, 2023

 K ( s  2) s( s  2)
For T=1 1  G ( s)  0 1 0 K
s ( s  2) ( s  2)

dK
 s 2  4s  4  0 s1, 2  2  2 2  4.82,0.828
ds

s1  4.82 BIP

s2  0.828 BAP

448 Monday, August 21, 2023

449 Monday, August 21, 2023
 Complementary Root Locus K (1  s)
G( s) 
s( s  3)

 K ( s  1)
G ( s)  Proper form
s( s  3)

When K is negative, this signifies the complementary root locus

And is varying 0K 

CRL, the sum of P+Z to the right of any point on CRL is even

450 Monday, August 21, 2023

 Solution P2 Z 1  K ( s  1)
G( s) 
B2 PZ s( s  3)

A  P  Z  2 1  1

2q *1800 q  0,1,2...( P  Z  1)
A 
PZ

A  00 As q0

Centroid


Re( P)   Re( Z )  0  3 1
  4
PZ 2 1

451 Monday, August 21, 2023

 K ( s  1) K ( s  1)
BAP/BIP 1  G ( s)  0 1 0 1
s ( s  3) s( s  3)

s ( s  3)
K
( s  1)

dK
 s 2  2s  3  0
ds

s1, 2 
2 (2) 2
 4(3)   2  4  3,1
2 2

s1  3 BIP

s2  1 BAP

452 Monday, August 21, 2023

453 Monday, August 21, 2023
454 Monday, August 21, 2023
 Root contour
The parameter other than gain K is varied to get the root locus. The other parameter may
be poles and zeros. Here, the pole is varied keeping gain K as constant.

Procedures:
K
G ( s) H ( s) 
s(s   )

K  s 2  s  K  0
1  G( s) H ( s)  1   s(s   )  K  0
s(s   )

Rearranging the equation  s 2  K  s  0

Dividing s2  K In the above equation, we get

455 Monday, August 21, 2023
 s
1 0
s2  K

Now, we have a new characteristic equation after arranging the original ch. Equation as

1  G ' ( s) H ' (s)  0 where

Since, G ' ( s) H ' ( s )  1

s
 2
 1
s K

 (s 2  K )
  Gain of the root contour
s

456 Monday, August 21, 2023

 Since

The open loop poles are s2  K  0  s1, 2   j K

 (s 2  K ) d d   ( s 2  K ) 
BAP Sinc      0
s ds ds  s 
e

 s * 2s  (s 2  K ) 
 2
  0  2s 2  s 2  K
 s 

 s2  K

 s1, 2   K

Here, s K Valid point since K is a pole here.

457 Monday, August 21, 2023
 Intersection at Imaginary axis

The characteristic equation is 1  G ( s ) H ( s )  0  s 2  s  K  0

Routh Array

When  0 the auxiliary equation is s2  K  0  s1, 2   j K

458 Monday, August 21, 2023

 Root Contour for open loop pole K is fixed, the gain is varying from
0  

459 Monday, August 21, 2023

 Root Locus for Transportation Lag
System

460 Monday, August 21, 2023

 Root Locus for Transportation Lag System
Ke  s
G (s) 
( s  1)

Solution
 1 
1  s 
2 
es  
 1 
1  s 
 2 

 es 
2  s 
2  s 

 K s  2 
G( s) 
( s  1)s  2 

461 Monday, August 21, 2023

 2q *1800
A  P  Z 1 A   00
PZ

BAP
( s  1)s  2  K ( s  2)
1  G ( s)  0
( s  1)s  2

( s  1)( s  2) ( s  1)( s  2) s 2  3s  2
K ( s  2)  ( s  1)(s  2) K  
( s  2) ( s  2) ( s  2)

dK ( s  2) * (2 s  3)  ( s  1)(s  2)
0  2
0
ds ( s  2)

( s  2) * (2 s  3)  ( s  1)(s  2)
 2
0
( s  2)

 s 2  4s  8  0  s1, 2  5.464,1.4640

s1  1.4640 Valid BAP

s2  5.464 Valid BIP

462 Monday, August 21, 2023
 ( s  1)s  2  K ( s  2)
1  G ( s)   s 2  3s  2  Ks  2 K  0
( s  1)s  2

 s 2  s(3  K )  2 K  2  0

Intersection with Imaginary axis

When, K=3, the auxiliary equation becomes s2  2 *3  2  0  s2  8  0

s1, 2   j 2 2
463 Monday, August 21, 2023
464 Monday, August 21, 2023
 Module_3

Lecture_25

Plots with MATLAB, Numerical

465 Monday, August 21, 2023

 Using MATLAB, construct the Root Locus

466 Monday, August 21, 2023

467 Monday, August 21, 2023
468 Monday, August 21, 2023
469 Monday, August 21, 2023
470 Monday, August 21, 2023
471 Monday, August 21, 2023
472 Monday, August 21, 2023
473 Monday, August 21, 2023
474 Monday, August 21, 2023
475 Monday, August 21, 2023
476 Monday, August 21, 2023
477 Monday, August 21, 2023
478 Monday, August 21, 2023
479 Monday, August 21, 2023
480 Monday, August 21, 2023
 Plot of Root Contour

The system involves two variables, parameter a and K.

The closed loop system is

481 Monday, August 21, 2023

482 Monday, August 21, 2023
483 Monday, August 21, 2023
 Module _4

Lecture_26

Frequency Domain Analysis

484 Monday, August 21, 2023

485 Monday, August 21, 2023
 Whenever, the system response is plotted w.r.t time, then we called it as time domain response.

Whenever, the system response is plotted w.r.t frequency, then we called it as time domain response

Note: Routh Criterion and Root Locus are the time domain approaches.
Laplace transform tool is used to analyze the system behavior.

However, Fourier transform is used to analyze the system in frequency domain

486 Monday, August 21, 2023
 Why Frequency domain ?

Certain parameters of the system are there which can not be analyzed through time domain such as
Peak resonant, resonant frequency, bandwidth, phase margin and gain margin. These are frequency
dependent parameters of the system. However, there is a co-relations between the time domain and
Frequency domain parameters which can be studied.

There are certain devices which deals with only frequency domain parameters such as Filter. It pass
Certain signal in terms of frequency.

The noise, signal are dealt with only on frequency domain.

When we go the frequency domain from time domain, we require Fourier Transform

Fourier Transform is the mapping between time domain and frequency

domain.

When we talk about Fourier Transform, we replace the s in transfer function by s=jw

487 Monday, August 21, 2023

 Let us take a real signal say f (t )  e 2t u (t )

Its transfer function is 1

F (s) 
2s
1
The Fourier transform is F ( jw) 
2  jw

To find the magnitude response and phase of the system, we find the Fourier Transform as

1
The magnitude F ( jw)  This is called magnitude spectrum Even function
2
4 w

 w This is called angle spectrum Odd function

The angle is F ( jw)   tan 1  
2

For real signal it is always the magnitude spectrum is even and phase spectrum
is odd function

488 Monday, August 21, 2023

 Magnitude Plot Phase plot

In frequency domain, we generally applied the sinusoidal as input to LTI system because,
sinusoidal is a eigen function i.e. the output obtained is a similar type of input by
Only changing in amplitude of the magnitude and phase angle.

The frequency does not change

489 Monday, August 21, 2023

 Prove:
In LTI system, the output is obtained by convolution of

y (t )  x(t ) * g (t )  Y ( s)  X ( s) * G( s)

By Fourier transform Y ( jw)  X ( jw) * G( jw)

 X ( jw) X ( jw) * G( jw) G( jw)

 X ( jw) G ( jw)  X ( jw)  G ( jw) 

So, Y ( jw)  X ( jw) G( jw) Multiply the amplitude

Y ( jw)  ( X ( jw)  G( jw)) Add the angle

490 Monday, August 21, 2023
 Hence , the output sinusoidal signal may be

A sin w0t  B sin(w0t   )

Where, B  A G ( w0 )   G ( w0 )

Example:

Calculate the steady state output if input x(t)= sin 2t and transfer function of the system is
1
G( s) 
s2
Solution:
1 1 1
G( jw)  G( jw)     450
jw  2 j2  2 2 2

1
Steady state output y (t )  sin(2t  450 )
2 2
491 Monday, August 21, 2023
 Advantages of Frequency Response
 It is more difficult to design a time domain response of a system for higher order system
The transfer function of a higher system can be identified by computing the frequency response
of the system for a wide range of frequency.
 The stability of a non-linear system can be analyzed by the frequency response method
 The frequency response can be applied to a system when the input is not deterministic.
The stability, relative stability of the system can be analyzed without evaluating the characteristic
Equation
 It is very convenient to measure the sensitivity to noise and parameter variation in frequency
Response.

Dis-advantages of Frequency Response

 Frequency response is not recommended for the system with very large time constant.
 It can only be applied to a linear system, however if applied to non-linear system, the result
obtained is not exact.
 This method is outdated when compared with the methods developed for digital computer
and modeling.

492 Monday, August 21, 2023

 Frequency Response Analysis

493 Monday, August 21, 2023

 The response may be leading or lagging depends on the system
components.
494 Monday, August 21, 2023
 The following plots are used in frequency response.

 Polar/Nyquist plot
 Bode plot or Logarithm Plot
 Nicholos Method/Magnitude-Phase angle plot

Frequency domain specifications

The performance measure of a feedback control system is also described in terms of

frequency-domain parameters.

The common frequency-domain specifications are

• peak frequency response, Mp,
• resonant frequency ωr , Closed loop
stability
• bandwidth specification

• phase margin and

• gain margin Relative stability
specifications

495 Monday, August 21, 2023

 Frequency domain specifications
Resonant Peak Mr
This is the maximum value of M(jw). This gives the indication of the relative stability of a stable
closed loop system.

Consider a closed loop system of a second order system

G( s) wn2
M ( s)  T ( s)   M ( s)  2
1  G( s) H (s) s  2wn s  wn2

wn2
Put s=jw M ( jw) 
( jw) 2  j 2wn w  wn2
496 Monday, August 21, 2023
 wn2
M ( jw)  2
( wn  w2 )  j 2wn w

Dividing both numerator and denominator by wn2

1
M ( jw)  1
  w 2   M ( jw) 
1      j 2 w
  wn  
 
1  u 2  j 2u
wn
 

1
M ( jw) 
1  u   2u 
2 2 2

1
M ( jw) 
1  u    2u 
2 2 2
Normalizing the frequency u 
w
wn

497 Monday, August 21, 2023

 1  2u 
M ( jw)  M  and M ( jw)     tan 1  2 
1  u   2u 
2 2 2
 1  u 

1  2u 
M    tan 1
2 
1  u   4 u
2 2 2 2
1 u 

498 Monday, August 21, 2023

The magnitude decreases from 1 to 0 monotonically or attains a maximum value and then,
decreases to zero.
499 Monday, August 21, 2023
 In decible , the magnitude is plotted as M dB  20 log10 M ( jw)
K
Where, the magnitude M including DC gain K M ( jw) 
1  u   4 u
2 2 2 2

K
M dB  20 log10
In dB 1  u   4 u
2 2 2 2

M dB  20 log10 K  20 log10 1  u    4 u
2 2 2 2

1
2

It goes to negative when the second term is more than the first term. The second figure is
When K=1
500 Monday, August 21, 2023
 Frequency domain Specifications (parameters of frequency response )

The frequency domain specifications are obtained from the magnitude curve when the
magnitude is expressed in p.u quantities
1
M
Since,
1  u    2u 
2 2 2

Resonant frequency Wr

The resonant frequency occurs when the slope

Of the magnitude response becomes zero.

 
d  1 
  0
 
1
du 
 1  u  
2 2
 2u 
2 2 

501 Monday, August 21, 2023
  2(1  ur2 ) * (2ur )  8 2ur  0  8 2ur  4ur (1  ur2 )  2 2  (1  ur2 )

 ur2  1  2 2  ur  1  2 2

wr The resonant frequency

AS ur   wr  wn 1  2 2
wn

2
The frequency wr is always positive and it is possible if and only if1  2  0

1
     0.707
2

502 Monday, August 21, 2023

 The resonant peak occurs when 0    0.707

The resonant peak does not occur when   0.707

Time domain response

Frequency domain response

for  

Mr  wr 

503 Monday, August 21, 2023

 Calculation of Mr
To get resonant peak Mr put the value of u=ur in the magnitude equation

1 1
M M 
1  u   2u 
1  u   2u  
2 2 2 1
2 2 2 2

Since, ur  1  2 2 
ur2  1  2 2 
1
Mr  1 1

  
 1  1  2 2 2  2 1  2 2 2 

 2  Mr 
4  4
4 2
(1  2 ) 2

1
2

1 1
 Mr   Mr 
4  4 
1 1
4 2 4 2 2 4 2
 4  8  4

504
 Mr 

2 1   2  Valid when   0.707
Monday, August 21, 2023
 For  0 Mr  

For   0.707 The peak resonant is equal to DC value

  0.707
Mr  

505 Monday, August 21, 2023

 ras
Similarly the phase angle at resonant frequency ur can be calculated

 2u 
1
Since,    tan  2 
1 u 

Putting the resonant frequency value u  ur  1  2 2

 2 1  2 2 
r   tan 
1 
 1  1  2 2 
 

 1  2 2 
r   tan  1 
  
 

506 Monday, August 21, 2023

 Co-relation between time domain and frequency domain

507 Monday, August 21, 2023

 Bandwidth: Wc
he range of frequency between zero and cut-off frequency is called as bandwidth frequency and
he cut off frequency wc is defined as the frequency at which the magnitude of the system is reduced
o 3 dB of the system at zero frequency .

Assuming the magnitude of the system at zero frequency is 1, the cut off frequency is

508 Monday, August 21, 2023

 Since,

509 Monday, August 21, 2023

 The expression for BW is same as the cut-off frequency

Cut-off rate :
The slope of the magnitude curve obtained near the cut-off frequency is called the cut-off rate.
The cut-off rate indicates the ability of the system to distinguish the signal from noise.

510 Monday, August 21, 2023

 Gain crossover frequency: wgc
The frequency at which the gain of the system is unity is called the gain crossover frequency
wgc
Phase crossover frequency: w pc
The frequency at which the phase angle of the system is -180 degree is called the phase
w pc
crossover frequency

Study of Relative Stability (GM,PM)

511 Monday, August 21, 2023

Gain Margin: Gain margin is defined as the margin in gain allowable by which gain can be increased
ll the system reaches on the verge of instability.

athematically, it is the reciprocal of the magnitude of G(jw) H(jw) at phase cross over frequency W p

1
GM 
G ( jw) H ( jw) w  w pc

Wpc the phase angle at which -180 deg is obtained.

Thus, the gain margin is a factor by which the open loop gain G(jw)H(jw) can be multiplied to make
The product unity.

In dB
1
GM  20 log
G ( jw) H ( jw) w w pc

GM  20 log G( jw) H ( jw) w w

pc

512 Monday, August 21, 2023

Phase Margin (PM): The amount of additional phase lag which can be introduced in the system
Till the system reaches on the verge of instability is called phase margin.

PM  G ( jw) H ( jw) 0
 (180 )    1800
w  wgc

wgc  wpc
PM is +ve

GM is +ve
wgc  wpc
Closed Loop system
Is stable

513 Monday, August 21, 2023

 (a) and (b) stable; wgc  wpc

514 Monday, August 21, 2023

 (c) and (d) unstable.
wgc  w pc

515 Monday, August 21, 2023

Checking of closed loop Stability

516 Monday, August 21, 2023

 Example: 2
G(s) H (s) 
s(1  0.5s)(1  0.05s)
Find wpc, wgc, PM, GM

Solution: According to definition wpc is the frequency at which the phase angle
of OLTF becomes -180 degree

2
  G ( s) H ( s) 
s(1  0.5s)(1  0.05s)

 1800  900  tan 1 0.5w  tan 1 0.05w

 tan 1 0.5w  tan 1 0.05w  900

Multiplying tan both sides we get,

 
 tan tan 1 0.5w  tan 1 0.05w  tan 900
517 Monday, August 21, 2023
 As
0.5w  0.05w
   
tan tan 1 A  tan 1 B 
A B
1  0.5w * 0.05w 1  AB

0.5w  0.05w
 
1  0.5w * 0.05w

 1  0.025w2  0

 wpc  6.32 rad/sec

GM 1
: GM 
G ( jw) H ( jw) w w pc

2
G ( jw) H ( jw) 
jw(1  j 0.5w)(1  j 0.05w) w pc

518 Monday, August 21, 2023

 2
G ( jw) H ( jw) 
j 6.32(1  j 0.5 * 6.32)(1  j 0.05 * 6.32) 6.32

 G( jw) H ( jw) 6.32  0.0918

1
Since, GM  20 log
G ( jw) H ( jw) w w pc

GM  20 log G( jw) H ( jw) w w

pc

GM  20 log10 0.0918  20.74dB

519 Monday, August 21, 2023

 Gain cross over frequency:
At the gain cross over frequency the magnitude of OLTF is unity

G( jw) H ( jw)  1

2
G ( jw) H ( jw)  1
2 2
w 1  0.25w 1  0025w

2
1
2 2
w 1  0.25w 1  0025w

wgc  2rad / sec

520 Monday, August 21, 2023
 To get PM   G( s) H ( s) w wgc

  900  tan 1 0.05  tan 1 0.05 w  wgc  2

    140.710
PM  1800  

PM  1800  140.710

PM  39.290

As wpc  6.32 wgc  2

since wgc  w pc The closed loop system id stable

521 Monday, August 21, 2023

 Numerical on Frequency domain specifications

522 Monday, August 21, 2023

 Example:

523 Monday, August 21, 2023

524 Monday, August 21, 2023
 Example:

525 Monday, August 21, 2023

 # Assignment

# Assignment

# Assignment

526 Monday, August 21, 2023

 Graphical Representation of frequency response
Determining the frequency response of the system by using tabular method is more difficult if
The system poles and zeros increases. So, graphical method is essential for this problem.

There are different types of graphical method used for frequency response such as

Bode plot (Asymptotic plot or Logarithmic

Plot)
Polar plot
Nyquist plot
Constant M and N circle
Nichols Chart

527 Monday, August 21, 2023

 Bode Plot:
It is one of the popular method for determining the stability of the system when the system is
Subjected to sinusoidal input. The stability of the closed loop system is determined based on the
Frequency response of the open loop transfer function G(s)H(s) of the system.

The Bode plot consists of two plots

(1) Magnitude plot (To plot the logarithmic magnitude or gain of the loop transfer function in dB
Verses frequency w i.e. M  20 log10 G( jw) H ( jw) in dB verses w

  G( jw) H ( jw)

(2) Phase angle plot : To plot the phase angle of loop transfer function
verses w
Different frequency domain specifications are obtained from Bode plot Such as gain
cross over frequency, phase cross over frequency, phase margin and gain margin etc.
The closed loop stability is determined from the PM and GM from the bode plot
itself.
The bode plot is only possible for minimum phase transfer function
Bode plot is done on semilog graph paper. At origin the low frequency value can not be equal zero
As

The plot is done dB versus log w. But the x-axis is given by w

528 Monday, August 21, 2023
 Bode-Plot:
This plot is only valid for minimum phase system. The bode plot requires the open
loop transfer function in time constant form. The common Bode plots are

Put s= jw

529 Monday, August 21, 2023

530 Monday, August 21, 2023
 Bode plot for constant gain Kb M  20log Kb

When Kb>0, M=+ve dB

When Kb<0, M=-ve dB

When Kb=1, M=0dB

The varying of constant gain Kb does not affect phase angle. So the phase angle is zero

531 Monday, August 21, 2023

532 Monday, August 21, 2023
533 Monday, August 21, 2023
 When the frequency is increased to double, then the slope is expressed as dB/octave

For ex: At w=0.5 M=-20log0.5=6dB

At w=1 M=-20log1=0dB
At w=2 M= -20 log2=-6dB

So, when the frequency is changed from w1=1rad/sec to w2=2w1=2rad/sec, the magnitude
Changed from 0dB to -6dB. So, the slope is -6dB/octave

When the frequency is changed from w1=1 rad/sec to 10rad/sec, then the slope is
changed to 20 dB/dec
For w=0.1, M=20 dB
w=1, M=0dB
w=10, M=-20dB
w=100, M=- 40 dB

So, 6dB/Octave=20 dB/dec

534 Monday, August 21, 2023

 Pole at origin 1 1 j
G(s)   
s jw w

1
M  G ( jw) 
w

In dB M  20 log10 w

The magnitude plot is a straight line passing through 0 dB at w=1 with slope -20dB/dec

For w=0.1, M=20 dB

w=1, M=0dB
w=10, M=-20dB
w=100, M=- 40 dB

In bode plot, the frequency ratio are expressed in term of decades or octaves. An octave is a
frequency band from w1 to 2w1 and decade is a frequency band of w1 to 10 w1 where w1 is
the frequency of any value.
535 Monday, August 21, 2023
 Similarly for

The bode plot passing through at zero dB at 1 with slope -40dB/dec

Similarly for n-poles at

origin

With slop -20n dB/dec

536 Monday, August 21, 2023
 Zero at origin

Therefore, the plot is a straight line passing through 0dB at w=1 with a slope of
20dB/decade
The phase plot is a horizontal line at 90 degree

537 Monday, August 21, 2023

538 Monday, August 21, 2023
 1
Simple Pole (1st order pole) G ( jw) 
Asymptotic Plot
1  jwT1

1 1
G  20 log  G  20 log(1  w T1 )2 2 2 2
 G  10 log(1  w2T1 )
2 2
1 w T 1

Plotting procedures
wT1  1 Low frequency range w2T12 is neglected as compared
Case-1
to 1
G  10 log(1  w2T12 )  G  10 log1  G  0dB Slope is
0dB/Decade

Case-2 wT1  1 high frequency range 1 is neglected

G  10 log(1  w2T12 )  G  10 log w2T12  G  20 log wT1

The slope is found as 1

wT1  1 w
T1
1
 G  20 log wT1  G  20 log T1  G  20 log1  G  0dB 0dB/Decade
T1

When wT1=10  G  20 log10  G  20dB - 20dB/Decade

539 Monday, August 21, 2023
 So, the magnitude plot is a straight line of 0dB magnitude when wT1<<1 and again
it is a straight line when wT1>>1. The two straight lines meet at a point (w=1/T1)
which is the corner frequency. This corner frequency divides the plot into two regions
(Low frequency and high frequency region)

540 Monday, August 21, 2023

 1
The exact plot can be plotted as follows G ( jw)  2
1  jwT1  G  10 log(1  w2T1 )

For w T1 =1 The error can be found as Error = exact plot - asymptotic

plot 2
Error  10 log(1  w2T )  (10 log w2T 2 )
1 1

2 1 2
Exact value is G  10 log(1  w2T1 )  G  10 log(1  T )
2 1
T1

 G  10 log(1  1)  G  10 log(2)  G  3dB

Asymptotic value is G  10 log(1  w2T12 )  G  10 log1  0dB

2
Error  10 log(1  w2T1 )  (10 log1)  Error  3dB

For w T1 < 1
Error = exact - asymptotic
Let w = 1/2T1
2
Error  10 log(1  w2T1 )  (10 log1)
1 2 5
Error  10 log(1  T1 )  (10 log 1) Error  10 log   (10 log 1)
4T12 4

Error  0.969dB  1dB

541 Monday, August 21, 2023
 For w > 1/T1 Let w=2/T1

Error = Exact – asymptotic

2
Error  10 log(1  w2T1 )  (10 log w2T12 )

4 2 4 2
Error  10 log(1  2
T1 )  ( 10 log T )
2 1
T1 T1

Error  10 log(5)  (10 log 4)

Error  0.969dB  1dB

542 Monday, August 21, 2023

 1
The phase angle for simple pole G ( jw) 
1  jwT1
is

Assumed to be zero frequency

At corner frequency

1 
   tan 1  T1    tan 1 1  450
 T1 

For wT1 >>1

543 Monday, August 21, 2023

 1st order zero G ( jw)  1  jwT1  G( jw)  1  w2T12


20 log G ( jw)  20 log 1  w T 2
1 
1
2 2 
 10 log 1  w2T12 

Asymptotic Plot For wT1 <<1 G  10 log1  0dB

For wT1 >>1 G  10 log w2T12  20 log wT

Let wT1 = 1 i.e. w = 1/T1 G  10 log w2T12  20 log wT1  G  0dB

When w =
10/T1
G  10 log w2T12  20 log wT1

10
G  20 log T1 G  20 log10 G  20dB
T1
544 Monday, August 21, 2023
 Error  10 log(1  w2T 2 )  (10 log1)

545 Monday, August 21, 2023

 Error Calculation:

546 Monday, August 21, 2023

547 Monday, August 21, 2023
 Asymptotic Phase angle:

548 Monday, August 21, 2023

 Numerator Quadratic factor

2
Put s=jw, then divide den. by wn

549 Monday, August 21, 2023

 Since,

 2 2
 2
 w   w 
 20 log10 G( jw)  10 log10 1      4 2   
  wn    wn  
  
Approximation Plot

550 Monday, August 21, 2023

 2
 w
20 log10 G ( jw)  20 log10  
 wn 
 w
 40 log10    40 log10 w  40 log10 wn
 wn 
When w = 0.1 wn G=
40dB
w = wn G=0dB
w=10 wn G=-
The
40dBlow frequency asymptote
is straight line with slope 0
dB/dec and high frequency
asymptote is a straight line
with slope -40 db/dec meet
at w = wn. This point w=wn
is called corner frequency.

However, the asymptotes

derived are independent of
damping ratio zeta
551 Monday, August 21, 2023
 However, near corner frequency w=wn a resonant peak is usually is expected and the
magnitude of the resonant peak depends on zeta.

The error introduced by the asymptotic is as follows

 w
2
 w 
2

20 log10 G ( jw)  10 log10 1    4   
2

 wn   wn  
At w=wn

20 log10 G ( jw)  20 log10 2

552 Monday, August 21, 2023

 The error is
negative
The exact plot is
below the
asymptotic plot

The error is
positive
The exact plot is
above the
asymptotic plot

No peak
resonant occurs

553 Monday, August 21, 2023

554 Monday, August 21, 2023
555 Monday, August 21, 2023
556 Monday, August 21, 2023
557 Monday, August 21, 2023
 Draw the bode plot for the OLTF
20
G (s) H (s) 
s2
Solution
: 20 10
G (s) H ( s)  G ( s ) H ( s ) 
s2 1 s / 2

10
G ( jw) H ( jw) 
1  jw / 2
Corner frequency w=2 rad/sec, DC gain K=10
In this problem there is a only only simple pole at corner frequency is 2

Now two asymptotic plot is drawn, one for DC gain K and other for simple pole

20 log K=20 log 10=20 dB Its magnitude plot is a straight line with slope 0 dB/dec

Since there is a simple pole at corner frequency w=2 rad/sec, the slope will be -20
dB/dec
The magnitude plot will start at 20 dB line upto 2 rad/sec, then the slope begins -
558 20 dB/dec at w=2 Monday, August 21, 2023
 Magnitude plot

559 Monday, August 21, 2023

560 Monday, August 21, 2023
 Example:

Solution:

561 Monday, August 21, 2023

562 Monday, August 21, 2023
 Magnitude plot of 10
G ( jw) H ( jw) 
jw(1  j 0.5w)(1  j 0.1w)

Wgc =3.9 rad/sec

563 Monday, August 21, 2023

 Since, wgc < wpc, the system is stable
564 Monday, August 21, 2023
 wpc =4.5 rad/sec

565 Monday, August 21, 2023

 Example:

Solution
:

566 Monday, August 21, 2023

 (iv) The change in slope at different corner frequencies is shown in
table

567 Monday, August 21, 2023

568 Monday, August 21, 2023
569 Monday, August 21, 2023
 Since, wgc< wpc The closed loop system is stable

570 Monday, August 21, 2023

 Example:

10( s  2) Pole-zero form

G(s) 
s ( s  3)( s  5)

Solution:

571 Monday, August 21, 2023

572 Monday, August 21, 2023
 Step 3 At each corner frequency the slope is changed.

1 st corner frequency is c=2 rad/sec . So magnitude plot will be starting from any
frequency less than wc=2. Let it is wc=0.1

The magnitude plot will change at next corner frequency and the procedures
repeats

573 Monday, August 21, 2023

574 Monday, August 21, 2023
 K K
Step 4 since G( s) H (s)  20 log10 G ( s ) H ( s )  20 log10  
s s
20 log10 G ( jw) H ( jw)  20 log10 K  20 log10 w w 0.1

Step 5 Angle Calculation

575 Monday, August 21, 2023

576 Monday, August 21, 2023
 Example: 104 (1  s )
G(s)  Draw the Bode Plot
GATE-2006 s (10  s)(100  s) 2

Solution: 104 (1  jw)

G ( jw) 
jw(10  jw)(100  jw) 2

577 Monday, August 21, 2023

578 Monday, August 21, 2023
579 Monday, August 21, 2023
580 Monday, August 21, 2023
581 Monday, August 21, 2023
582 Monday, August 21, 2023
 Example:

583 Monday, August 21, 2023

 K
G ( s)  Since, there are two poles at
s2 origin

To find the value of K 32  20 log K  40 log1

 32  20 log K
32
log10 K   1.6
20
 K  101.6  39.8
584 Monday, August 21, 2023
585 Monday, August 21, 2023
586 Monday, August 21, 2023
587 Monday, August 21, 2023
588 Monday, August 21, 2023
589 Monday, August 21, 2023
590 Monday, August 21, 2023
591 Monday, August 21, 2023
592 Monday, August 21, 2023
593 Monday, August 21, 2023
 Disadvantages of Bode Plot:

Only valid for Minimum phase Transfer Function

594 Monday, August 21, 2023

 Types of system for Bode Plot:
All Pass
System:
A system which passes all frequencies without any attenuation is known as all pass
system or all pass filter.

Such system is characterized by a transfer function with a zero in mirror image position
to imaginary axis for every pole on left side.

G(s) 
1  Ts 
The TF
is (1  Ts )

595 Monday, August 21, 2023

 Minimum Phase Transfer
Function:
The system with all poles and zeros in the left half of s-plane is known as minimum
phase transfer function.
For stable system, it is necessary to keep all the poles in left half of s-palne, but there is
no restriction on zeros of the transfer function.

Consider a system

596 Monday, August 21, 2023

The transfer function G(s) is the product of two transfer function.

Where G1(s) has all the poles and zeros in the left half s-plane and Ga(s) is an all pass
system.
Since the contribution of magnitude of Ga (s) is unity, the magnitude plot of G(s)
and G1(s) are identical. But the all pass system contributes the phase angle of 180
degree when frequency changes from 0 to infinity. This all pass system adds to G(s)
more angle than G1(s) without affecting to the system magnitude. This makes the
transfer function G(s) more lagging angle.
This transfer function G(s) is called as non-minimum phase transfer function.

The larger phase angle contributes the deterioration or damage to the system as
well. Hence, the large phase angle is to be avoided.

This transfer function G1(s) is called as minimum phase transfer function.

597 Monday, August 21, 2023

 Phase angle plot of, Minimum , Non-Minimum phase and All pass
system

Example:
Non-Minimum phase
TF

598 Monday, August 21, 2023

599 Monday, August 21, 2023
600 Monday, August 21, 2023
601 Monday, August 21, 2023
 POLAR PLOT

602 Monday, August 21, 2023

 Polar plot of a TF G(jw) is a plot of magnitude of G(jw) versus the phase angle of G(jw)
on polar co-ordinate as w varies from 0 to infinity.

G ( s)  G( jw)  G( jw) G( jw)  M

The polar plot gives the information of closed loop stability from the OLTF .

The polar plot is not essential. Only the rough sketch.

The following points are to be followed for polar plot.

603 Monday, August 21, 2023

604 Monday, August 21, 2023
605 Monday, August 21, 2023
 Type 0 1st order system:

Solution:

606 Monday, August 21, 2023

 Type-1 2nd order
system
Solution: 1
G ( jw) 

w 1  w2T12 
  900  tan 1 wT1

607 Monday, August 21, 2023

 Type-0 2nd order system

1
Solution: M
1  w T 1  w T 
2
1
2 2
2
2    tan 1 wT1  tan 1 wT2

608 Monday, August 21, 2023

 To know the intersection points:

609 Monday, August 21, 2023

610 Monday, August 21, 2023
611 Monday, August 21, 2023
612 Monday, August 21, 2023
 Type 0 3rd order System

613 Monday, August 21, 2023

 Type 0 4th order System 1
G ( s) 
1  sT1 1  sT2 1  sT3 1  sT4 

Type 2 4th order system

614 Monday, August 21, 2023

 Zero contributes zero degree
and pole contributes -90 degree

615 Monday, August 21, 2023

616 Monday, August 21, 2023
 Example:
Solution:

(1)
617 Monday, August 21, 2023
618 Monday, August 21, 2023
619 Monday, August 21, 2023
620 Monday, August 21, 2023
621 Monday, August 21, 2023
622 Monday, August 21, 2023
623 Monday, August 21, 2023
624 Monday, August 21, 2023
625 Monday, August 21, 2023
626 Monday, August 21, 2023
627 Monday, August 21, 2023
628 Monday, August 21, 2023
629 Monday, August 21, 2023
 NYQUIST PLOT

630 Monday, August 21, 2023

 Nyquist Plot
 Polar plot is valid on frequency from zero to + infinity while Nyquist plot is valid
from minus infinity to plus infinity
Sometimes calculation of gain margin is not possible, this is when the polar plot
does not cross (-ve) real axis. So, stability analysis is difficult.
This can be possible by Nyquist Plot

Nyquist plot tells that, how many number of closed loop poles or zeros of characteristic
equation lies on the right half plane.

631 Monday, August 21, 2023

Nyquist Stability Criterion
The following points are to be understood

 The relationship between the poles of 1+G(s)H(s) and poles of G(s)H(s)

The relationship between the zeros of 1+G(s)H(s) and the poles of the closed
loop TF
Mapping
Mapping Contour
Principle of Arguments
632 Monday, August 21, 2023
 Consider a negative feed system.

(1)

(2)

(3)

633 Monday, August 21, 2023

 Mapping

Mapping Contour:

634 Monday, August 21, 2023

635 Monday, August 21, 2023
636 Monday, August 21, 2023
637 Monday, August 21, 2023
Principle of Arguments:

According to this principle, if the s-plane contains Z zeros and P poles of s-contour within
it, then, the mapped F(s)plane contour encircles the origin N=P-Z times in in
anticlockwise direction when P>Z. So N= -ve. If Z >P, N= +ve .
If P=Z, N=0 no encircle. Here, P= no. of open loop poles or poles of F(s) lying rt. Half s-
plane and Z= zeros of q(s).

This mapping is called Nyquist daigram, or Nyquist plot of G(s)H(s)

638 Monday, August 21, 2023

 Application:

Fig (a) is stable and Fig (b) is unstable.

639 Monday, August 21, 2023
640 Monday, August 21, 2023
 A general construction procedure is outlined for continuous systems in the following
steps.

Step 1: Check p( s ) for poles on the jw-axis and at the origin. Since, no pole at
origin, the Nuquist path will be
Nyquist contour or path

Step 1: Draw the polar plot for the path ad in p(s) plane.
Step 2 : Draw the polar plot for the path fa, that is the mirror image of ad.
Step 3: Plot the image of path def. This path at infinity usually plots into a point in the
p(s)-plane.
641 Step 4: Connect all curves drawn in the previous steps Monday, August 21, 2023
Step: Find the Nyquist plot
 The origin poles or imaginary axes poles are to be detoured

642 Monday, August 21, 2023

 Ex: Nyquist plot for type 0 system

S-plane
Since no pole at origin, the Nyquist counter will be

For the path ad, the polar plot is drawn where the frequency is varying from zero to infinity.

Where the magnitude is zero and the angle varies from

643 Monday, August 21, 2023
For the path fa, the polar plot is drawn which is the mirror image of path ad.

The complete Nyquist plot is shown on p(s) plane

p(s) plane
Here, P=0, N=0
So, N=Z-P=0
Hence, closed loop
system is stable.

Ex:

Since no pole at the origin, the Nyquist contour is

644 Monday, August 21, 2023

645 Monday, August 21, 2023
 Here, P=0, N=0

Since N=Z-P

Z=0

So, no zeros of p(s) in right half of s- plane.

646
Hence, closed loopMonday,
system is stable.
August 21, 2023
 EX:
Open loop system is unstable.

647 Monday, August 21, 2023

 Since P=1, the encirclement about -1+j0 is clockwise , So N=1

Hence N=Z-P Z= N+P Z=1+1=2 Which is not equal to zero

Hence, the closed loop system is unstable

648 Monday, August 21, 2023

 Ex: Open loop system is unstable

649 Monday, August 21, 2023

650 Monday, August 21, 2023
651 Monday, August 21, 2023
652 Monday, August 21, 2023
 Effect of Poles at the Origin

653 Monday, August 21, 2023

654 Monday, August 21, 2023
655 Monday, August 21, 2023
656 Monday, August 21, 2023
 Ex:

657 Monday, August 21, 2023

658 Monday, August 21, 2023
659 Monday, August 21, 2023
 Ex:

the entire Nyquist Path is broken up as before

(a) Semicircle abc (b) imaginary axis cd (c)

semicircle deg and (d) imaginary axis ga

Let us now evaluate the mappings for these segments

of Nyquist Path .

660 Monday, August 21, 2023

661 Monday, August 21, 2023
 This operation can be performed using this MATLAB command:
nyquist
[real, imag, omega] = nyquist(NUM, DEN, omega);

Ex:

662 Monday, August 21, 2023

663 Monday, August 21, 2023
664 Monday, August 21, 2023
 Here, P=0
N=0
Since N=Z-P
Z=N+P=0

No encirclement about -1+j0 Hence, closed loop stable.

Monday, August 21, 2023

665
666 Monday, August 21, 2023
667 Monday, August 21, 2023
668 Monday, August 21, 2023
669 Monday, August 21, 2023
670 Monday, August 21, 2023
671 Monday, August 21, 2023
672 Monday, August 21, 2023
673 Monday, August 21, 2023
 NYQUIST’S STABILITY CRITERION APPLIED TO SYSTEM HAVING DEAD
TIME

674 Monday, August 21, 2023

675 Monday, August 21, 2023
676 Monday, August 21, 2023
677 Monday, August 21, 2023
 Relative Stability:

678 Monday, August 21, 2023

679 Monday, August 21, 2023
680 Monday, August 21, 2023
 M- Circle (Constant Magnitude of CLTF loci)
The closed loop frequency response specifications are studied in this M and N circles.
Such specifications are Mr, Wr and band width Wb.
 M-Circle determines the magnitude of CLTF from the information of OLTF
Applicable on feedback system only
Gives the information of resonant peak (Mr), resonant frequency (wr)and
bandwidth (wb) of CLTF

Mr: This is the maximum value of M, the magnitude of the closed loop system. A large resonant pe
Corresponds to large overshoot in transient response.

Wr: This is the frequency at which the resonance peak Mr occurs. This frequency is related to
the frequency of oscillations in step response and is indicative of speed of transient response.

Wr: Bandwidth: It measures the the ability of a feedback system to reproduce the input signal
and measures the noise rejection characteristics. The closed loop system filters out the signal
Whose frequencies are greater than cut off (-3db). This range of frequency is the BW frequency.

681 Monday, August 21, 2023

682 Monday, August 21, 2023
683 Monday, August 21, 2023
 Equation of a circle

Where center is

684 Monday, August 21, 2023

685 Monday, August 21, 2023
 Constant M-Circle

686 Monday, August 21, 2023

687 Monday, August 21, 2023
688 Monday, August 21, 2023
The polar plot of G(s)H(s) is done on the M-circle. The largest of M-circle for which the M-circle
is tangent to the polar plot is known as Mr and the corresponding frequency is known as Wr.

N-Circles or Constant Phase angle Loci

689 Monday, August 21, 2023

690 Monday, August 21, 2023
691 Monday, August 21, 2023
 the equation of a circle with N as a parameter. It has its center at

nserting a given value of into the equation of circle. Various N-circles are obtained
y considering various N. All N-circles pass through origin and -1+j0 point regardless of the value N

The tangent of angles in the first and third quadrant is positive. Therefore, the yq
coordinate is the same for an angle in the first quadrant and for the negative of its
supplement, which is in the third quadrant. As a result, the constant contour is
only an arc of the circle.
After constructing the M and N circles, draw the G(jw)H(jw)plot which passes through M and N
circle. Note down the intersection points on M and N circles. Plot the closed frequency response cur
Monday, August 21, 2023
To 692
know the Mr and Wr.
693 Monday, August 21, 2023
694 Monday, August 21, 2023
695 Monday, August 21, 2023
 Nichols Chart

The resonant peak, resonant frequency and bandwidth of closed loop system are known from
Nichols Chart

696 Monday, August 21, 2023

697 Monday, August 21, 2023
 Procedure to plot Nichols Chart
Step 1 Draw the bode plot selecting the frequency such that the phase angle lies between 0 to 180
Step 2 The Bode plot follows the Table-1

698 Monday, August 21, 2023

 Bode Plot

db
Phase angle

frequency
699 Monday, August 21, 2023
 Step 3 Draw the Nichols plot from bode plot, take each unit of magnitude as 4db on Y axis

Step 4 Corresponding to this magnitude at 0db,4 db 8db etc and -4db,-8db…..Find the corresponding
Frequencies from X-axes for the same Bode plot. Then Find the phase angle corresponding to these
Frequencies.

700 Monday, August 21, 2023

ep 5 Table 2 is drawn getting the various values of phase angle from the corresponding g frequencie

701 Monday, August 21, 2023

702 Monday, August 21, 2023
 Available Nichols Chart

The thick line represents the magnitude in db and dotted represents the phase angle
703 Monday, August 21, 2023
Step 6 : Draw the Nichols Chart from the Table 2 from the supplied Nichols Chart

For ex: For 20db with

-98 degree phase angle
gives w=0.2rad/sec
and so on.

Join all the frequency

Points.

704 Monday, August 21, 2023

Step-7 From different values of magnitude of supplied Nichols chart which intersecting the our
G(s)H(s) plot, find the various frequencies and the phase angle as well. This is shown in Table 3.

705 Monday, August 21, 2023

Step 8 Draw the magnitude and phase plot versus frequencies from Table 3 taking another
semilog graph paper . This is the required plot that determines the Mr, Wr and Wb

Magnitude Plot

706 Monday, August 21, 2023

 Both Magnitude and
Phase Angle Vs frequency plot

degree

So we get, Resonant peak Mr= 4db rad/sec

Resonat frequency Wr= 1.68 rad/sec

Band width Wb= 2.5 rad/sec

707 Monday, August 21, 2023
 Module 5
Lecture_37

Introduction to Controller, types of classical controller

708 Monday, August 21, 2023

 Classical Controller
Control system designed may not able to meet the design goals such as time domain
specifications and frequency domain specifications.

Approach is made to force the system to meet the design goal using the controllers.

A controller is the combination of both error detector and amplifier with adjustable gain.

The classical controller used in industrial applications are proportional controller,

derivative controller , integral controller and their combinations as well.

The most common use of classical controller is PID controller.

The adjustable gain is done through PID controllers which is placed ahead of the
amplifier.

Almost all industries use PID and modification of PID as controller in applications.

Analog PID mostly available hydraulic, pneumatic, electric and electronics or

combination of both.
709 Monday, August 21, 2023
 The controller parameters (gain, time constant) of PID are tuned in such a manner that,
the transient and steady state error are minimized in closed loop system. The are called
performance specifications.

For plant model simple----- analytical approach required for gain calculation

For plant model complex------ Experimental approach is done

However, the PID control action has been brought from basic three control action such as P, I and D
Control action.
710 Monday, August 21, 2023
 Proportional Control
The proportional controller in true sense, an amplifier with an adjustable gain.

In a proportional control scheme, the controller output is proportional to the actuating signal.

The output of the controller and error are related by

m(t )  K p e(t ) M ( s)
  Kp
E ( s)

K p  Lt G( s) H ( s)  
s 0
Consider a second order system
wn
K v  Lt sG( s) H ( s) 
s 0 2

The steady state error is given

by
1 2
ess  
K v wn

711 Monday, August 21, 2023

 1 2
ess  
K v wn

1 2
 ess  
K v  wn

owever, the gain Kv can not be increased beyond the certain limit which brings instability to the sys

Since, the SS error depends on the increase of natural frequency, this creates the unstable of the syst
2
 ess 
wn 

Thus, a sluggish over damped system response can be made faster by increasing the
forward path gain of the system. This increase in gain reduces the steady state error but
the overshoot is increases and gives more oscillations and approaches towards
instability
712 Monday, August 21, 2023
 Integral Control

In integral action , the control signal or the output of the controller is proportional to the
integral of the actuating signal. Thus,
t
ec (t )  K I  ea (t )dt
0

713 Monday, August 21, 2023

 d
ec (t )  K I ea (t )
dt
• Since the rate of change of controller output ec is proportional to the actuating signal.
• For a constant actuating, control signal increases linearly,
• For a zero actuating signal , the control signal remains stationary or without change.

Hence, the integral control is sometimes called as a reset signal.

Effect of Integral action:

Consider a plant With integral action as

714 Monday, August 21, 2023

715 Monday, August 21, 2023
716 Monday, August 21, 2023
 If the system shown in Figure is asymptotically stable, and the inputs r and d
(disturbances)
are constant, it follows that all signals in the closed loop will tend to constant values.

In particular the integrator output v(t) tends to a constant value. It follows that the
integrator input e(t ) tends to zero.

Since we have arranged that this input is the tracking error it follows that e(t) = r −
y(t) tends to zero and hence y(t) tracks r as t → ∞.

So, the integrator input is forced to become tracking error to zero.

717 Monday, August 21, 2023
 P-I Controller:

718 Monday, August 21, 2023

719 Monday, August 21, 2023
720 Monday, August 21, 2023
 Take a second order
wn2 Type 1, 2nd
system G(s) 
s( s  2wn ) order system

When the PI controller is added,

G( s ) 
K p  K I / s wn
2


K I  sK P wn2 Type 2, 3rd order
s( s  2wn ) s 2 ( s  2wn ) system

Hence, adding controller, the system order increases. The transient response gets
affectedif Ki is not designed properly.

721 Monday, August 21, 2023

 K I w n2
Ka 
2
1 2
e ss  
Ka K I w n2

722 Monday, August 21, 2023

 1 2
e ss  
PI controller Ka K I w n2

It increases the order of the system

 It increases the type of the system
Steady state error reduces tremendously for same type of input but not affecting the
transient response
Decreases the bandwidth
Increases the rise time
Acts as a low pass filter

723 Monday, August 21, 2023

 Derivative Control
In derivative control action, the output of the controller or control signal is the derivative
of the actuating signal

724 Monday, August 21, 2023

 Effect of derivative action:
Consider a second order system with derivative controller

When the actuating signal is constant, the derivative controller does not produce any
control signal. Hence, the steady-state error is effective only during ne transient
period of the output response. Because of this reason, a derivative control is usually
not used alone. It is always used along with other types of controllers.

If the actuating signal is time varying, the derivative controller produces a proper control
signal that
is proportional to the rate of change of the actuating signal. This controls the error from
becoming to large.
If the plant transfer function does not contain an integrator then the addition of a
derivative controller improves the damping ratio. In the absence of the controller, the
characteristic equation of the system
725 Monday, August 21, 2023
 The addition of derivative controller with the plant, improves the damping ratio.

Without derivative action,

The characteristic equation of the given system is,

K
1  G( s)  0  1 0
s  a s  b 

Compared with s 2  2wn s  wn2  0

ab
1 
2 ab  K

With Derivative controller

726 Monday, August 21, 2023

 K * sK D K * sK D
The characteristic equation is 1 0 G ( s ) 
s  a s  b  s  a s  b 
 s  a s  b   K * sK D  0

 s 2  sb  sa  ab  K * sK D  0

Which can be compared with s 2  2wn s  wn2  0

The damping a  b  KK D
ratio
2 
2 ab

727 Monday, August 21, 2023

Thus, inclusion of derivative action

 improves the damping ratio

Reduces the maximum overshoot
Reduces the settling time
Increases the rise time
It acts as lead compensator
Improves the GM, PM and stability
Increases the band width
 SS error remains unchanged (with P-D controller)

Thus, the derivative controller in the forward path improves the transient performance
parameters.
728 Monday, August 21, 2023
 Derivative Feedback Control
The derivative feedback control is also known PD control. For derivative feedback
control the actuating signal is obtained as the difference between the proportional error
signal and derivative (rate ) of the output signal. Therefore, the actuating signal for
derivative feedback control action is

For a second order unity feedback control system block diagram incorporating
derivative
feedback control is shown

729 Monday, August 21, 2023

 Derivative feedback (tachometer feedback) control

The characteristic equation for the overall transfer function

730 Monday, August 21, 2023

 The damping ratio for the above characteristic equation is

The damping ratio is increased by using derivative feedback control and therefore, the
maximum overshoot is decreased. However, the rise time is increased.

The forward path transfer function is

731 Monday, August 21, 2023

 Substituting G(s)and H(s)

732 Monday, August 21, 2023

 It is noted that using derivative feedback control the steady state error is increased
and as the damping ratio is increased the maximum overshoot is decreased.

In both types of control actions derivative error as well as derivative feedback, the
damping ratio is increased which results in decreasing maximum overshoot but the time
response expression is different.

In case of derivative control the rise time is decreased and the steady state error
remains unchanged whereas in derivative feedback control the rise time is decreased
and the steady state error is increased.
733 Monday, August 21, 2023
734 Monday, August 21, 2023
 Ex: The overall transfer function of a control system is given by

It is desired that the damping ratio be 0.8. Determine the derivative rate feedback
constant , and compare rise time, peak time, maximum overshoot and steady state error for
unit ramp input without and with derivative feedback control.

Solution: The characteristic equation is

Comparing with the second order characteristic equation

therefore, the damping ratio for the system without derivative feedback control is

735 Monday, August 21, 2023

 The damping ratio with derivative feedback control is given by

736 Monday, August 21, 2023

 (b) With derivative feedback control
The overall transfer function of the system using derivative feedback control is given
by

737 Monday, August 21, 2023

738 Monday, August 21, 2023
 P-D Controller:
Feed-forward P-D

739 Monday, August 21, 2023

740 Monday, August 21, 2023
741 Monday, August 21, 2023
 However, this can be shown that, the steady state error will not change with and
without using PD controller in the second order system.

Thus, P-D control action has such effects

 improves the damping ratio

Reduces the maximum overshoot
Reduces the settling time
Decreases the rise time
It acts as lead compensator
Improves the GM, PM and stability
Increases the band width
 SS error remains unchanged

742 Monday, August 21, 2023

 PID Control action
P-I improves the steady state response, whereas P-D improves the
transient responses, thus the combination of these two controllers
improve the overall time domain performances by proper tuning the
controller gains called parameters of the controller. This can be achieved
by one of the classical method called Zeigler-Nichols tuning controller
method.

743 Monday, August 21, 2023

744 Monday, August 21, 2023
745 Monday, August 21, 2023
746 Monday, August 21, 2023
747 Monday, August 21, 2023
 Structure of PID Controller

748 Monday, August 21, 2023

 Zeigler-Nicholas Tuning Of PID- An experimental approach (Step response
method )

Z-N approach is used when mathematical model of the plant is not

known.
The Ziegler-Nichols step response method is an experimental open-loop tuning method
and is applicable to open-loop stable plants.

This method first characterizes the plant by two parameters K and L obtained from its
step response. K and L can be determined graphically from a measurement of the step
response of the plant.

First, the point on the step response curve with the maximum slope is determined and the
tangent is drawn.

The intersection of the tangent with the vertical axis gives K, while the intersection of the
tangent with the horizontal axis gives L.

Once K and L are determined, the PID controller parameters are then found out
following the table.
749 Monday, August 21, 2023
 First method of Z-N When Plant has no integrator and dominant conjugate
By obtaining step response poles

Step input S-
shaped

750 Monday, August 21, 2023

 Based on step response

751 Monday, August 21, 2023

 The Ziegler-Nichols 2nd method By evaluating critical gain and critical
period
Dynamics of plant is not be known

The Ziegler-Nichols frequency response method is a closed-loop tuning method. This

method first determines the point where the Nyquist curve of the plant G(s) intersects
the negative real axis.

It can be obtained experimentally in the following

way:
• Turn the integral and differential actions off i.e. and
• set the controller to be in the proportional mode only and close the loop
• Slowly increase the proportional gain kp until a periodic oscillation in the output is
observed. The value of Kp is called then critical gain Kcr.
• Find the period of oscillation Pcr. Knowing the Kcr and Pcr the gain of PID controller
are found.
752 Monday, August 21, 2023
 response PID controller T.F.

Z-N 2nd method table

753 Monday, August 21, 2023

 When the plant has an integrator, the Zeigler tuning rule (neither 1 st and 2nd ) may not be
applied in some cases.

Example:
H(s)=1

Because of the integrator, 1st method is not applied. Since the step response will not give a S shape
Rather the response increases with time.

Also, the 2nd method is not valid since, the closed loop system does not give sustained
oscillations at any value of Kp

Hence, the solution for this is analyzed through Routh-Array formulation.

The characteristic equation is

754 Monday, August 21, 2023

 When the plant has an integrator, the Zeigler tuning rule (2 nd method ) may be applied
in some cases.

Example:

755 Monday, August 21, 2023

756 Monday, August 21, 2023
 From the 2nd method table, Kp, Td and Ti are to be found as:

757 Monday, August 21, 2023

 The PID controller has one pole at origin and two no’s of zeros which shows that
PID controller itself is a improper function. So the structure of PID controller is
to be modified by changing its TF to bring into proper function .

758 Monday, August 21, 2023

 Summary:
P-I Control Action:
reduces the steady state error.
Increases the rise time.
Acts as low pass filter
decreases the bandwidth.

P-D Control Action:

It increases the damping ratio
Reduces the peak overshoot
Reduces the settling time
Improves the transient responses without affecting the steady state response.
Increases the B.W.
Improves the stability

P-I_D Control Action:

PID control action: Both steady state error and undesired transient occurrences are
reduced
759 Monday, August 21, 2023
 Lecture_38

Sensitivity transfer functions (S and T) and their

significance, Open loop and closed loop with unity and
non-unity feed back system

760 Monday, August 21, 2023

SENSITIVITY
The environmental conditions to which a control system is subjected affect the accuracy
and stability of the system.

The performance characteristics of most components are affected by their environment

and by aging. Thus, any change in the component characteristics causes a change in the
transfer function and therefore in the controlled quantity. The effect of a parameter
change on system performance can be expressed in terms of a sensitivity function.

This sensitivity function is a measure of the sensitivity of the system’s response

to a system parameter variation and is given by

761 Monday, August 21, 2023

 To illustrate the effect of changes in the transfer function, four cases are considered for
which the input signal r(t) and its transform R(s) are fixed. Although the response Y(s) is
used in these four cases, the results are the same when M(s) is the control ratio.
762 Monday, August 21, 2023
 Case 1: Open-Loop System

A change in the transfer function G(s) therefore causes a proportional change in the
transform of the output Yo(s).
763 Monday, August 21, 2023
 Case 2: Closed-Loop Unity-Feedback System

764 Monday, August 21, 2023

 Comparing with open loop with closed loop , it is seen that the effect of changes of
G(s) upon the transform of the output of the closed-loop control is reduced by the
factor 1/(1+G(s)) compared to the open-loop control. This is an important reason why
feedback systems are used.

Case 3: Closed-Loop Non-unity-Feedback System

[Feedback Function H(s) is Fixed and G(s) is Variable]

765 Monday, August 21, 2023

 Comparing with unity feedback, it is seen that, the non-unity feedback closed-loop
variation is reduced by the factor 1/(1+G(s)H(s)). Since, the term 1/(1+G(s)H(s)) is
larger than the term 1/(1+G(s)). , then there is an advantage to using a non-unity-
feedback system. Further, H(s) may be introduced both to provide an improvement in
system performance and to reduce the effect of parameter variations within G(s).

766 Monday, August 21, 2023

 Case 4: Closed-Loop Non-unity-Feedback System
[Feedback Function H(s) is Variable and G(s) is Fixed]

Multiplying and dividing the above equation by H(s) and also dividing by equation

results in

767 Monday, August 21, 2023

 It is seen that a variation in the feedback function has approximately a direct effect
upon the output, the same as for the open-loop case.

Summary: In other words, by use of feedback compensation the performance of

the system can be made to depend more on the feedback term than on the forward
term.

768 Monday, August 21, 2023

SENSITIVITY ANALYSIS
An important aspect in the design of a control system is the insensitivity of the system
outputs to items such as: sensor noise, parameter uncertainty, cross coupling effects, and
external system disturbances.

The analysis in this section is based upon the control system shown in Figure

Where T(s)= Y(s)/R(s) and where F(s) represents a pre-filter. The plant is described by
P(s) and may include some parameter uncertainties.

In this system G(s) represents a compensator. The pre-filter and compensator are
designed to minimize the effect of the parameter uncertainties. The goal of the design is
to satisfy the desired figures of merit (FOM). In this text it is assumed that F(s)=1. The
effect of these items on system performance can be expressed in terms of the sensitivity
function which is defined by
769 Monday, August 21, 2023
770 Monday, August 21, 2023
771 Monday, August 21, 2023
 Based on the magnitude characteristic of Lo for low- and high frequency ranges,
then:

Then, becomes

Then

772 Monday, August 21, 2023

 THANK
YOU

773 Monday, August 21, 2023