Frequency Response of a LTI System

Lecture 8

We have seen that LTI system response to x(t)=est is H(s)est. We

represent such input-output pair as:
st
st

e H (s)e

Frequency Response

Instead of using a complex frequency, let us set s = j, this yields:

e jt H ( j )e jt
cos t = Re(e jt ) Re[ H ( j )e jt ]

(Lathi 4.8 4.9)

Peter Cheung
Department of Electrical & Electronic Engineering
Imperial College London

It is often better to express H(j) in polar form:

Therefore

E2.5 Signals & Linear Systems

Amplitude
Response
Lecture 8 Slide 1

PYKC 8-Feb-11

Frequency Response Example (1)

Substitute s=j

H ( j ) =

H ( j ) =

2 + 0.01
2

+ 25

Phase
Response
L4.8 p423

E2.5 Signals & Linear Systems

Lecture 8 Slide 2

Frequency Response Example (2)

Find the frequency response of a system with transfer function:

s + 0.1
H ( s) =
s+5
Then find the system response y(t) for input x(t)=cos2t and x(t)=cos(10t-50)

Frequency
Response

cos t H ( j) cos[t + H ( j)]

URL: www.ee.imperial.ac.uk/pcheung/teaching/ee2_signals
E-mail: p.cheung@imperial.ac.uk
PYKC 8-Feb-11

H ( j) = H ( j) e jH ( j )

H ( j ) =

2 + 0.01
2

+ 25

1
( j ) = H ( j ) = tan 1
tan
0.1

j + 0.1
j + 5

1
and H ( j ) = ( j ) = tan 1
tan
0.1
5

for input x(t)=cos2t and x(t)=cos(10t-50)

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 3

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 4

Frequency Response Example (3)

For input x(t)=cos2t, we have:

H ( j 2) =

Frequency Response Example (4)

22 + 0.01
22 + 25

2
1 2


( j 2) = tan 1
tan = 65.3
0.1
5

= 0.372

Therefore

For input x(t)=cos(10t-50), we will use the amplitude and phase response
curves directly:

H ( j10) = 0.894

( j10) = H ( j10) = 26

y(t ) = 0.372cos(2t + 65.3
)

Therefore

y(t ) = 0.894cos(10t 50
+ 26
) = 0.894cos(10t + 24
)

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 5

PYKC 8-Feb-11

Frequency Response of delay of T sec

H(s) of an ideal differentiator is:

H (s) = s and H ( j ) = j = e j / 2

(Time-shifting property)

Therefore

Therefore

H ( j ) =

H ( j) = e jT = 1

and

( j) = T

That is, delaying a signal by T has no effect on

its amplitude.
It results in a linear phase shift (with
frequency), and a gradient of T.
The quantity:

Lecture 8 Slide 6

Frequency Response of an ideal differentiator

H(s) of an ideal T sec delay is:

H (s) = e sT

E2.5 Signals & Linear Systems

and

H ( j ) =

This agrees with:

d
(cos t ) = sin t = cos(t + / 2)
dt

d ( )
= g =T
d

Thats why differentiator is not a nice

component to work with it amplifies high
frequency component (i.e. noise!).

is known as Group Delay.

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 7

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 8

Bode Plot Sketching frequency response
..

without (much) calculation (1)

Frequency Response of an ideal integrator

H(s) of an ideal integrator is:

H ( s) =

s
s

+ 1 + 1
a
a
K ( s + a1 )( s + a2 )
Ka a
1 2
= 1 2
H (s) =
s ( s + b1 )( s 2 + b2 s + b3 )
b1b3 s

s2 b
s + 1 + 2 s + 1
b
b
b
1 2
3

Therefore

and

H ( j ) =

This agrees with:

The POLES are roots of the denominator polynomial. In this case, the
poles of the system are: s=0, s=-b1 and the solutions of the quadratic

( s 2 + b2 s + b3 ) = 0

cos t dt = sin t = cos(t / 2)

Consider a system transfer function:

1
1 j 1 j / 2
and H ( j ) =
=
= e
s
j

H ( j ) =

Thats why integrator is a nice component to

work with it suppresses high frequency
component (i.e. noise!).

which we assume to be complex conjugates values.

The ZEROS are roots of the numerator polynomial. In this case, the
zeros of the system are: s =-a1, s=-a2.
L4.9 p430

PYKC 8-Feb-11

Lecture 8 Slide 9

E2.5 Signals & Linear Systems

PYKC 8-Feb-11

Bode Plot Sketching frequency response
..

without (much) calculation (2)

Now let s=j,

Ka a
H ( j ) = 1 2
b1b3

1+
j 1 +

Building blocks for Bode Plots amplitude (1)

j
j
1+
a1
a2

b ( j ) 2
j
1+ j 2 +
b1
b3
b2

Express this as decibel (i.e. 20 log()):

Lecture 8 Slide 10

E2.5 Signals & Linear Systems

zeros at a1 and a2

Pole term

Zero term +20log j

Pole term 20 log 1 +

for
a ,

20log j

j
a

= a/2

poles at 0

= 2a

For s+a

for
a ,

constant term

conjugate poles

poles at b1

Now amplitude response (in dB) is broken into building block components
that are added together.

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 11

For

at = a,
20log 1 + j = 20log( 2) 3dB
PYKC 8-Feb-11

20dB

1
s+a

1 decade
E2.5 Signals & Linear Systems

Lecture 8 Slide 12

Building blocks for Bode Plots amplitude (2)

Now consider the quadratic poles: s 2 + b2 s + b3

Better to express this as:
s 2 + 2 jn s + n2
damping
factor

Building blocks for Bode Plots amplitude (3)

Elsewhere, the exact log amplitude is:

natural
frequency

weak dampling

The log magnitude response is:

strong dampling

40dB

1 decade

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 13

PYKC 8-Feb-11

Bode Plots Example amplitude (1)

Consider this transfer function:

We re-write this as

Step 1: Establish where x-axis crosses the y-axis

Step 2: For each pole and zero term, draw an asymptotic plot.

E2.5 Signals & Linear Systems

Lecture 8 Slide 14

Bode Plots Example amplitude (2)

Since the constant term is 100 = 40dB, x-axis cut the vertical axis at 40.
We need to draw straight lines for zeros at origin and =100.
We need to draw straight line for poles at =2 and =10.

Step 3: Add all the asymptotes.

Step 4: Apply corrections if necessary.

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 15

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 16

Building blocks for Bode Plots Phase (1)

Now consider phase response for the earlier transfer function:

Ka a
H ( j ) = 1 2
b1b3

1+
j 1 +

Building blocks for Bode Plots Phase (2)

Pole term

Pole term

j
j
1+
a1
a2

b ( j ) 2
j
1+ j 2 +
b1
b3
b2

= a /10

= 10a
For s+a

Therefore:

for
a ,

for
a ,

For

Again, we have three type of terms.

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 17

PYKC 8-Feb-11

Building blocks for Bode Plots Phase (3)

2nd order poles: s 2

Phase response is:

PYKC 8-Feb-11

+ 2 jn s + n2

E2.5 Signals & Linear Systems

E2.5 Signals & Linear Systems

Lecture 8 Slide 18

Bode Plots Example phase (1)

Lecture 8 Slide 19

1
s+a

Consider this again:

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 20

Relating this lecture to other courses

You will be applying frequency response in various areas such as filters

and 2nd year control. You have also used frequency response in the 2nd
year analogue electronics course. Here we explore this as a special case
of the general concept of complex frequency, where the real part is zero.
You have come across Bode plots from 2nd year analogue electronics
course. Here we go deeper into where all these rules come from.
We will apply much of what we done so far in the frequency domain to
analyse and design some filters in the next lecture.

PYKC 8-Feb-11

E2.5 Signals & Linear Systems

Lecture 8 Slide 21