Frequency response

T.Kaputu

Introduction
What is frequency response?
Why is it useful?
How do I represent frequency response information in a
helpful fashion?
Why is this relevant to feedback loop analysis and design?

Overview
1. Frequency response refers to system behaviour and therefore
is defined for a system.
2. Here the focus is linear systems that can be represented using
a transfer function model.
In simple terms, we are asking how the output of a system
behaves when the system input is a pure sinusoid.

G(S)

u = sin(wt )

y = A sin( wt + )
3

Overview continued
When the input sinusoid had frequency w, the output signal
was:

y = A sin( wt + )

What happens if the frequency is changed?

y C sin( t + )
u = sin( t ) =
Frequency response is a description of how the amplitude and
phase shift of the output, relative to the input, change with
frequency.
4

4
G= 2
s + 3s + 2

Example 1

Overlay the input and output for a system G(s) and a single
frequency (here w = 1rad/s).

Output
amplitude
about 1.2

input
0.5

output

u = sin t
y = 1.2 sin(t 1)

-0.5

Phase shift about

1 second which
-1
corresponds to 1
45
50
55
radian
Time (secs)

60

65

Example 1
Overlay the input and output for a system G(s) and a single
frequency (here = 2/).
4

G=

s 2 + 3s + 2

0.8
0.6

Output
amplitude
about 0.62

0.4
0.2
0
-0.2 Phase

shift about 0.9

second which corresponds
-0.4
to 1.8 radian

u = sin 2t
y = 0.62 sin( 2t 1.8)

-0.6
-0.8

-1
50

51

52

53

56
55
54
Time (secs)

57

58

59

60

Observations
As the frequency of the input changes:
The amplitude of the output changes.
The phase shift of the output changes.
Frequency response is a description of how the amplitude
and phase shift depend upon the frequency of the input, that
is, how do the characteristics of the response, depend upon
the frequency?
7

Definition of amplitude and gain

Some systems include integrators and hence, even with a
sinusoidal input, the output signal is not centred on zero.
Hence, amplitude is half peak to peak.
The following signal, has an amplitude of oscillation of A.

y = B + A sin( wt + )

Gain is the ratio of the

output amplitude to
the input amplitude.
8

Definition of phase shift

Let the output signal be given as

y (t ) = B + A sin( wt + )
Now shift this signal by 1 seconds.

1 = + ( + 1 + )

The phase shift is clearly 1 !

A time shift of 1 is equivalent to a phase shift of 1

Units then balance:
Frequency w is in rad/s and time is seconds.
9

Computing gain and phase from time responses

Find the time between
points of corresponding
phase: t = 1 2 . Then
the phase lag is given as
= (1 2 )

0.25
1

Find the amplitude of

oscillation as half peak
to peak. Compare ratio
of input and output.

y=

0.27 sin(t

[ 0.25 ])

0.27

ENSURE AVOID TRANSIENTS

10

Question
Find the gain and phase for the following system given the
input frequency is 3 rad/s.
2

2 = 1.1

= 0.55

= 0.9

= 3 0.9 = 2.7

1.5

50.8

51.7

1.65

0.55

0.5

-0.5

-1
50

50.5

51.5
51
Time (secs)

52

52.5
11

Bode diagram 2
Simple computation of frequency response

12

Introduction
What is frequency response? Why is it useful?
How do I compute this efficiently?
How do I represent frequency response information in a helpful
fashion?
Why is this relevant to feedback loop analysis and design?
This section looks at the second bullet point of an efficient
method for computing frequency response.

13

Background
The previous section showed how frequency response values
could be inferred from time responses.
1. Gain A(w) is the ratio of output amplitude of oscillation to
that of the input.
2. Phase (w) is the phase difference between the input and
output responses.
However it is tedious and inaccurate to infer A and from time
responses.

u = D sin(wt )

G(s)

y = DA sin( wt + )

14

Efficient computation of frequency response

A more efficient and simple method for determining
() and () is from the system transfer function ()
[assuming this is known].
It can be show that:

A = G ( jw) ; = arg(G ( jw))

That is, substitute = and compute the modulus and
argument of the resulting complex number.
15

Example 1
4
G= 2
s + 3s + 2
The first section gave some time responses and inferred
frequency response from those.
Here, we compare those answers with the analytic approach.
=

4
4
G ( jw) =
=
2
( jw) + 3 jw + 2 3 jw + 2 w2
4

3w
G ( jw) =
; G ( jw) = tan
2
2
2 2
2

w
9w + (2 w )
1

16

Example 1 analytic solutions

Use frequencies of 1 rad/s and 2 rad/s
G ( jw) =

4
9w2 + (2 w2 ) 2

3w
G ( jw) = tan
2 w2
1

4
= 1.26
10

G ( j1) =

3
G ( j1) = tan
= 1.24
1

G ( j 2) =

9 + (2 1) 2

4
36 + (2 4) 2

4
= 0.63
40

6
G ( j 2) = tan
= 1.89rad
2
1

17

4
G= 2
s + 3s + 2

Example 1

Overlay the input and output for a system G(s) and a single
frequency (here w = 1rad/s).
Estimates from section 1

u = sin t
y = 1.2 sin(t 1)

Analytic solution
0.5

u = sin t
y = 1.26 sin(t 1.24)

-0.5

-1

45

50

55
Time (secs)

60

65

18

4
G= 2
s + 3s + 2

Example 1

Overlay the input and output for a system G(s) and a single
frequency (here w = 2rad/s).

u = sin 2t
y = 0.62 sin( 2t 1.8)

Estimates from section 1

1
0.8

Analytic solution

0.6
0.4
0.2
0

u = sin 2t
y = 0.63 sin( 2t 1.89)

-0.2
-0.4
-0.6
-0.8
-1
50

51

52

53

54
55
56
Time (secs)

57

58

59

60

19

Example 2
Find the frequency response (gain and phase definitions) for ().

2s + 1
G= 3 2
s + s + 3s + 2

2 jw + 1
=
j (3w w3 ) + 2 w2

2 jw + 1
G ( jw) =
( jw) 3 + ( jw) 2 + 3 jw + 2

G ( jw) =

4w2 + 1
(3w w ) + (2 w )
3 2

2 2

3
2
w
(
3
w

w
)
1
1
G ( jw) = tan
tan
1
2 w2

NOT NICE

Be careful to take
solution in the
correct quadrant.

YUCK

20

General formulae for gain and phase

() = |()| =
() = () =

Re ()
tan1

+ Im ()

Im ()
Re ()

We will not tend to use this means of finding the

frequency response because it is not efficient and does
not use the easy rules that come from
multiplying/dividing complex numbers.
21

Conclusions

System output amplitude and phase depend on the

frequency of the input signal.
In general for a transfer function G(s), then:
= |( )|; = arg(())

However, a crude computation of these formulae leads to

very cumbersome expressions and poor insight.

The following section will show how the parameters () and

() can be determined efficiently and with methods that
give easy and useful insight.
22

Bode diagram 3
Using factors to compute frequency response

23

Introduction
What is frequency response? Why is it useful?
How do I compute this efficiently?
How do I represent frequency response information in a helpful
fashion?
Why is this relevant to feedback loop analysis and design? This
section looks at the second bullet point of an efficient method
for computing frequency response.
24

Frequency response formulae to avoid, in general.

An efficient method for determining the gain and phase
() and () is from the system transfer function G(s).

However, the following expressions tend to be cumbersome to

use.
() = |()| =
() = () =

Re ()
tan1

+ Im ()

Im ()
Re ()

This section introduces an alternative based on factorisation.

25

Frequency response for simple factor

Consider simple factors with LHP roots (numerator or
denominator) such as () = ( + ) () = 1/( + ).
() = +

|( )| =

1
() =
+
|()| =

2 ; ()

tan1

; () =

tan1

You may find

it useful to
convince
yourself that
the phases
given here
are indeed in
the correct
quadrant
given
, > 0.
26

Multiplication and division of complex numbers

Students are reminded of the following rules for complex
numbers.
1. Modulus of the product is the product of the moduli.
2. Phase (or argument) of the product is the sum of the phases.
||||

|| = ||||||;
=
;
||||||

||2 ||
2
=
3

||3 ||

= + + ;

2
3 = 2 + 3

27

Frequency response of two factors

Use rules of complex numbers to find the frequency response for
the following.
+
=
+

+
+
=
=
+
+

2 + 2
2 + 2

= + +
+

tan1

1
tan

28

Example
Given () = 2(0.5 + )/(2 + ), find the gain and phase for a
frequency of 4.
2( + 0.5)
+2

=2

=4

2 + 0.52

tan1

2 + 22

=2

1
tan
0.5
2

= tan1 8 tan1 2

16.25
20

29

Example
Which is the amplitude and phase of the output?
4
() =
; = sin2
+ 6)( + 4
= (2 + )
4
4
=2
=
=
40 20
2 + 62 2 + 42
= tan
=

tan1

1
tan
6
4

1
1
1
tan
3
2

30

Example
Write the gain and phase expressions
+2
() =
+ 3)( + 4

2 + 22

2 + 9 2 + 16

tan1

1
1
tan
tan
2
3
4
31

Example
Write the gain and phase expressions
+ 1)( + 2
() =
+ 3)( + 4

2 + 12

2 + 4

2 + 9 2 + 16

tan1

+ tan1

1
1
tan
tan
2
3
4
32

Example
Write the gain and phase expressions
4
() =
( + 4)

= 900

2 + 16

= 90

tan1

4
33

Conclusions
1.

2.
3.

Shown how one can use the gain and phase of simple
factors to build the overall gain and phase for a transfer
function.
= |( )|; = arg(())

Uses simple rules of complex numbers.

Formulae used here assume LHP factors only; RHP factors
are covered in the next section.

NOTE: Where a transfer function has a quadratic factor with

complex roots it is more difficult to produce a simple rule
so students should treat those with care (or use a
computer algebra package).

34

Bode diagram 4
Frequency response of RHP factors

35

Introduction
What is frequency response? Why is it useful?
How do I compute this efficiently?
How do I represent frequency response information in a helpful
fashion?
Why is this relevant to feedback loop analysis and design?
This section continues the focus on efficient (and insightful)
methods for computing frequency response.
36

Background
The previous section showed that an efficient method for
determining frequency response gain and phase for a system
transfer function () is to deduce the gain and phase or each
factor and then use rules of complex numbers.
1. Modulus of the product is the product of the moduli.
2. Phase (or argument) of the product is the sum of the phases.
() = + ; |( )| =

+ 2 ; ()

tan1

1
1

1
() =
; |( )| =
; () = tan
2
2
+

+
HOWEVER, Simple formulae are for LHP factors only!

37

Frequency response for simple numerator factor with a

RHP root.
Consider a simple factor with a RHP root, () = ( ).
Gain is straightforward, but YOU MUST SKETCH THE complex
number before attempting the phase to ensure you are in the
correct quadrant.
() = |( )| =

() =

tan1

0.1

2 + 2

0
-0.1

1 0.5

-0.2
-0.3
-0.4
-0.5

-0.2

0.2

0.4

0.6

0.8

381.2

Frequency response for simple numerator factor with

a RHP root.
Consider a simple factor with a RHP root, () = ( ).
Gain is straightforward, but YOU MUST SKETCH THE complex
number before attempting the phase to ensure you are in the
correct quadrant.
() = |( )| =
2.5

1
() = 180 tan

2 + 2

-0.5+2j

1.5

0.5

-0.5

-0.5

-0.4

-0.3

-0.2

-0.1

0.1

39

Frequency response for simple denominator factor with a

RHP root.
Consider a simple factor with a RHP root, () = 1/( ).
Gain is straightforward, but YOU MUST SKETCH THE complex
number before attempting the phase to ensure you are in the
correct quadrant.
1
1
=
( ) =

2 + 2
2.5

Calculate the phase of the

factor and add a minus sign as
in denominator

-0.5+2j

1.5

0.5

() = 180

tan1

-0.5

-0.5

-0.4

-0.3

-0.2

-0.1

0.1

40

Summary 1
The gain/phase for RHP factors are summarised, but the advice is,
ALWAYS SKETCH, the corresponding complex number to ensure
you get the phase in the correct quadrant.
2 + 2
() =

1
() = tan

|( )| =

|( )| =
1
2 + 2

() =

() = tan1

41

Summary 1
The gain/phase for RHP factors are summarised, but the advice is,
ALWAYS SKETCH, the corresponding complex number to ensure
you get the phase in the correct quadrant.

G ( jw) =
a jw

G ( jw) =
a jw

G ( jw
=
)

w2 + a 2

w
G ( jw) =
tan
a
1
G ( jw) =
2
2
w +a
1 w
G ( jw) = tan
a
1

42

Summary 2
The gain/phase for RHP factors are summarised, but the advice is,
ALWAYS SKETCH, the corresponding complex number to ensure
you get the phase in the correct quadrant.

G ( jw) = jw a

1
G ( jw) =
jw a

G ( jw) = w2 + a 2
w
G ( jw) = 180 tan
a
1

G ( jw) =

2
2
+
w
a

1 w
G ( jw) = tan
180
a
43

What is the gain and phase of G(s)?

5
G (s) =
(s 2 3)
= 180

tan1

2 3

= (180 tan1
=

2 + 2 3

2 3

)
2 3
2

44

Example
What is the asymptotic output signal for the following?
1
() =
; = sin
+1

= ( + )

1 = tan1

tan1

tan1

= 2 tan1 = 90

2 + 1

2 + 1

=1

=1

45

Example
Write the gain and phase expressions
6( + 2)
() =
3)( + 4

3 = 180 tan
=

tan1

3 = 3

1
1
tan
180 tan
2
4
3

6 2 + 4

2 + 9. 2 + 16

46

Conclusions
1.

Shown how one can determine the gain and phase of simple
factors with RHP roots.

2.

Advised that students always sketch the complex number

associated to the RHP root to ensure no silly errors arise in
allocating the correct phase.

3.

Given some examples of finding the frequency response of

transfer function models with multiple factors, some of
which may include RHP roots.
47

Exercise
Find the frequency response expressions for the following
transfer functions.
0.2

1 + 0.4

5
( + 1)
+2
2 ( + 1)

0.4( + 1)
( + 4)( + 2)
8(1 )
( + 4)( + 2)
+5
+2

3( + 6)
( + 10)( + 1)( + 3)( + 4)

2( + 2)
( + 10)( + 1)( 4)

2
48