SHORT NOTES

Control Systems
EEE F242
H.V.D Sir
& Important Laplace Transform Formulae

#
f(t)
#

-
u(t) or 1
I Laplace transform is

t
A used to solve complex
S

n !
equations .

th
gn
+ 1
By taking Laplace transforms
W we can convert differential
Sin wt
S'+w equations to
algebraic
S
polynomials .
This helps
coswt
S + w2
to order
us salve
higher
1
ent S-w
differential equations

- impulse
fit) 1

if input varies
linearly the output also
· Linear
system -> ,

changes linearly

·
Time Invariant
system -> does not
depend on the time at

which
input is
applied

we will
mostly deal with Linear time Invaliant (LTE) systems
o/
system i/P sys
h(t) ~ e
⑧
- -

f(t) impulse x(t) y(t)

response

y(t) x(0) h(t) x(1) h(t-1) +

+

x(x)h(t- es)
: I
=
.......

-
↓ X
input at t 0
shifted only possible
=

->

impulse response in LTI sys

.. y(t =
x(z) h(t -

z) < CONVOLUMON
e
↓ T = 0 ↓ ↳ impulse
/P
%
i/p response
 obtain
On
taking Laplace Transform ,
we

H(s)
H(s) -> transfer
y (S) =

X(s) -

↓ function

2993 Laplace transform

11x1
I
=

of impulse response
([i/py

at initial condition & for LTI

-
Only true z00

Fourier Transform

Jazz) f(t) X (w) 1

e swit
x(t)
=

= =

·
x(w) =

doesn't if not
· Fourier transform at
give a solution Its das

converge ·

eg .
2
jut of
=

I eat , It

analyse diverging signals

we
To ,

1j -a jw)t
(e-
+

add in another component

I'd => Alt
&

solution 18 a
jw)t
get
+

a
-

->
-

When >a we
, -
2

We can redifine the

formula as 10 -
a +

jw)
:
⑧

x (s) =
juct) est dt ->
- >
Laplace Transform
while s =
- +
jw of x(t)

↳ 0 When no
dampening

HIS)= (S-a)(s-b)
·

Transfer function
(S 2)(s -d)
-

here S = a ,
b - zeroes of the
transfer function
S =
c ,
d > Poles
of the
Hansfu function
 Translational Mechanical System

degree of freedom
:
· One

Free
body diagram Laplace Transformed
x(t) >x(t) >x(s)

"
-
E
kx(t) < kx(s) <

esse
M > f(t) f. d M > f(t) frs X(s) <
M
> F(s)

Md <
Ms X(s) <

Opposing
fr
If frSXS) kXCS) FIS)
Mdn(t) fra f(t) MSX(s) +
=

kat=
+

farce to displacement
+

at
inertial mass (F =
mas
due to

1
o/P
415)
Tramsful function
- -

is
:
-

MS2 + frS + h

· Two degues of freedom shortcut

*
*
:

x, (t) xz(t) For a mass

> AV, >
mass
f(t) to that
attatched
- I all elements
↳, k3
both
esse M, Me esse
Ecommon elements of
↳2 masses
-

esse

I fre ↑ fre

For Mi

[M ,
5 + (fr ,
+
fry)s +
(k , +
kz)] X , 13) -

(fr,s +

kn) X21s) =

F(s)

For Mr :

(fVzs +

k2)X , (3)
+

[M25 + (fvz +

f(z)s +
(k +

k3)]X2(5) =
0

Transfer function
:
:.
coeff of X, 15) cauff of XIS)
in M , (s) in M , (5)

Xcls) (fVzs +

(2) -

Xe(s)
cauff of
- caff of XIS]
=

F(S) in M2CS]
in Mc(s)
Rotational Mechanical System
Torque -

Angular Impedance
Displacement
(t) k0(t) ↳
Spring T(t) =

viscous Damper (D)

Tit) D
do DS
=

(5) T(t) 592

dot
Inertia 5
=

torsion

E si)
Throughout
nfacto yac
D

e -
2

in
⑤
For J,:

(5 ,
52 + D, s +
t, (0 , (5) -
k0 , 1s)= + (5)

For J2 :

-
k0 , (5) + (525 +
Des +
k2) 01S) =
0

Oc(s) k
Transfor function
=

T(S)

electricstems
 Block Diagram Reduction
Input output
If MIs) wasn't present
R(S) Els) ((S)
⑱
G(s)
Transfu function
>
⑳

~I ↑) take off
point <(S)
BI
IS q(S) =

< H(s) R(S)

E(s) k(s) = BCS)

feedback wop
=

((S) =
E(S) R(S)

=(R(s) = B(s)) 4/s)

Du to
feedback ,
the new

is
transfor function
:
-
(R)s) = <(S) HCS)) G(s)

((S)
R(S)
=

1
G(s)
= G(s)H(s)
=isness

· in series/cascade ·
In parallel
4
C I
R
>
- -
+
7 41 42 43 - 42 >
-

43 7

N
N
R C C
R
- G, 42 43 - - 4, +
G2 -

43 -

shifting of take off point

the block
the block to
after gain
-

f ro m before gain
R C =

R4
R c =

R4 G ·

· 4
-
#
<y =
R
I

[
Y =
R

- vice versa

R R c R4
c R4
=

4
=

4 ⑧

I
I
[ -
G
y R4
=

<
y R4
=
·
shifting the
summing point

after the gain

block to
before the
gain block
-

from
R4 I X
=

2
R
- > 4
R4 I X
=

2
R I
> 4 >
I
->
-

x x I
a X

- vice versa

R4 = XG R4 = x4
=

c
=

c R
R - 4 >
- >4 I
I
->
-

G
a X
a X

Signal Flow Graphs G3

7
>as
R I C 41 42 Gu yu

"
> Gi >
G2 7 Gu > >
x
>
-
=> -8 > ⑧ ⑧ ⑧

43
I

< H1
L
-

H1
<
- 1 ↑

with a node
Each
summing point & take
off point is replaced
elements
Each node is a summation
of incoming

y ,
=
x
-24 ·Ys
=

G242 + 4342 - H:
Yy
G, y, · Yu Guys
·
=

yz =

Mason’s Gain Formula

M
i =
1 - N where N -> no
of forward
ne
paths
↓
Ti i
Ti= gain of forward path it paths from input
without
to

mode
output
i =
1
crossing a

-1- (I
loop gain)
more than once

-
I now touching loops) -
[(3 now
touching loops)
[ (4 now
touching loops)
-
+ . . . . .
 ,
5
Find
transfor
the

-n function
1
1 43
- -
·
⑧
au
⑧ ⑳
-
R C
G2

E < -
H1
-
H2

->
-
N =
2
,
Two forward paths
,
5

-n
in series
R
↳

. an
·
1

-
C
T =
G, G2 Gs Gy ->

IHz
< -
H1 X

1st
gain
forward
of
path sain of and path

3 loops present
= Hi Gu 22 =
-

H243 ; 23 =
- H3
↳
-
;
>45
43
mode
-

↳3 n

↳ & L have I common 1

-
1

·
· ⑧ -
R au C
↳I
↳
hence touching
< H1
En
-

21 L & Les
touching loops
:

a now

:
O
3 non
touching loops

-
=>
1 -

( +

2 +

(3) +

(42 +

2) -
0

i
->
from , if any loop touches it path ,
make it 0

=
1 -
0 =
1
1

2
=
2 -

(2 +

(3) +
(213)

↳ ↳, mode common

Mason's Gain Formula

.. Applying
the transful fine.
= Tit , + 5242 = ousolving you get
L
Time Domain Analysis
First Order System - Time Response
order = no
of poles=1
(whic s =s +jw)
similar
:.
transfer function
will be
toas
will
For
naturally occuring physical system complexe poles
*
any ,

cannot exist
always exist in
pais i.e
only I
complex pole
real
:For a 1st order
system , poles are
always
a/sta
· For a step input if transfer func is
, find response
time domain
Step (u(t))
in
is
-> Input
->

= a
(IS) =

a
* Is in racace domain
=
15
= a
-

35 +
a)

to find in time domain-> inverse response

<(1) =

<"[CCs)Y =

<"[stay= 2"
(5 Ea] -

- at
. c(t) = 1 -
e
-
-

should be on the
* For
any
stable
system ,
the
poles always
left side
of the
yanis
in
-plane

converging response -- stable

system
-
finite output

diverging response -> unstable

system
->
infinite output
due to input
natural response due to
system
21
at
->
-

·
((t) = -
e

Sforced > transient state

+ +a ->> <(H) 1 e" 0 63

Time constant system
=

>
- .

of a
=
=

reaches 63 % output faster than Ya= 100s

If Ha
=
55 -

system
Sheggish
system
 reaches 98 %
·

settling time -> is :

I - output

10 % 90 %
·
Rise time -> Time it takes for system
to
go from to

Tr = 2
-
. 2

a Output

act Only
this
x
Response of a
-> L-e-at
type of response
-

order
first system is possible
-
t >
Step response

Second Order System

· It has 2 polls
induce oscillation & ral
·
complex part of a
pole part causes

in the
damping/decay output

is
· Fol a
purly complex pole ,
the step response oscillatory .

41S) 9 S
eg

W
.
=

=>> Step response :

92 + 9
<(t) =
1- cos3t
-
polys
=

Ij3 .

Fol real pole (response like (adul

purely oscillation
·
a no is
,

2k , t
b
-

a
eg
.

4(5)= => c(t) = 1 =

2
9 + as + b -
Bkzt
-
e

both real
poles k,, + 7
=

-
·
If 415) =

5 + as +
b

CCH-1-cost
0 (15) ss -> poles=Ijfb +

1fa
=
=
=

Wn=is t Wi =

b
excillations
frequency of
:

... natural

F
West q
- +

S
:G (s)
=
= =>

wit ↳ real part of pole

 real part
a
·

Damping
Ratio :

Es= of pole -
a =
2 WnEs
2W
Wn

Wi
: G(s) =

S+ 2EgWns+Wi

polls S,
EgWnI We -1
=

Ratio (Es) Pole location Step Response

Damping
Oscillation
0 I jWn Undamped

0 =
Eg = 1 Complex Undurdamped Oscillation

2 real poles at
1 critically damped
same
point
1 Oscillation
E =
I real Overdamped

·
For an overdamped & critically damped system ,
the
response
is like 1st order
system :it behaves like

1
st ordel
& 2 2
:
Rise time Tr =
.

a 'can be replaced
EgWn
with real palt of

is t
Time
settling poll' Wn' the

!
> We
sel" will be close to

accurate
=2 1 tolanc
%

*
Underdemuped system
mean ene

e5wnt
:
Step response

Mr
<(t) 2 Sin Welt +O
-
=

1 -

>

.
Wat We 1-5 ;
O=
tan")
 Rise time Tr= is
Fron
· :
=

for undudamped
it is 0 % -> 100 %

Time for man overshoot Tos-

:

Es
Peak Time Tp
we-si
-
=

·% overshoot=e
1 -

As the
pole location
changes from 1+ 2 +
3: (horizontally)

coordinate remains the same

3 2 I y
. ⑧ ----- Win 1-E
:Peak Time
"
remains the same

Ol
increases
En coordinate magnitude
-

⑧ ⑧
time decreases
setting
· overshoot decreases /: tano dechases)

Practice what polis vertically

-
diagonally
when +
happens have

in
you add
the
· When a zero
transfer func
Wn(s+a)
4291
-
=

152 + 24Wns+wn

It simplifies to
Scales the
->
e response
c'(s) = Sc(x) +
ac(s)
-
↓
in the
derivative -> affects only
transient stati

Hence
-
,
adding a zo
only scales the response
its nature
- does not
change
 Third Order System
second order
:

· When
you
add a pole to a
system complex paurs
poles↑

as in
be exist
can
only -

added to real

(S) Wi axis
El
q
I

RIS) wi) (S +AS

⑧

(s +

, wns
25
+

=>
a(s) =

RCs) +
RIS)
*
I
·
a

(S A)
⑧
+

ke

- At
=> :CH =

( ) +
e
-
> damping
x
If the new pole is at a distance greater than 5a from
& the
the original poles ,
it can be
ignored system can be

treated as a 2nd order system

Type of System & Steady State Error

4(5) = (S a)(s b)
+
+

=
ab(1 + +,
5)(1 +

Tes)
=
↓ (1 +
Sti)(1 + Stz) . . .

(S + 2) (s +
d) cd(1 +
Ts) (1+ Tus) 5 (1 +

ST3) (1+ STul ...

at
Type O
system
-> n = o -> 0 poles origin

Type I

system ->n= 1 -
I pole at origin

SECs)
-

ecor= elt) It
Steady state whil elt) Wit) ((t)
If =
-
· =

t -
es
S -
0 input output

· RIS) (2S)
G
=
>

-- a
=

=
state eva E(s) RIS) s
Steady
-
=

:
ess= see
· For a Type O
system RCS) =

Als
e > a Step

91s) = lest s
A
input

ess =

s =

SAS =ass" r

R(S) =

Er
- a
samp
input ess=
et Sx =
S
S- O
1 +

>
parabolic
a(s) =

Es
i

ess =

It a
=

Steady State Error (for Unit

negative feedback loop)

Input Type O
Type 1 Type 2

A
Step C -
1 +
k

A
samp 0 ⑨
-

A
parabolic 0 0
-

k
 Sensitivity
does
CT/
How T
change
o Ite
T
Sa -
= on
changing I is the

SF/z sensitivity

2 4
= =

· - -

S =

I = 1 : + =

4 = :

4
T =

=
=

=4
- X
-

S Get
sensitive
a te
C
=

feedback
systems an prefund

closed loop polis same

open loop &
are
4/s+
&

(4 C :
=1)
=

R C open loop poly--a

X Ysta
(T s+)
.

closed loop polis = -a -

=

<
↓
men towards left
I
For positive feedback T -> poll--ath mal
=
-

S + a -
k towards right
(may class origin)

Hence
negative feelback loops all more stable than positive
-
fudback
-
used in oscillators
used in
amplifies