4/22/2024

Sinai University
Faculty of Engineering

ENM 4217
Automatic Control

Credit Hours
Lecture Lab/Tutorial Total
2 2 3

Chapter IV
Frequency Response
Analysis

1
 4/22/2024

4. 1 Introduction
The fully discussion of sinusoidal test signal was postponed till this
chapter due to its importance in control engineering.

Now, consider a linear system with a sinusoidal input:

r(t) = A sin(wt)................................................................ (4. 1)

or R(s) = Aw .................................................... (4. 2)

2
(s + w2 )
The steady-state output may be written as:
c(t) = Β sin(wt +f)............................................................. (4. 3)

The magnitude and phase relationship between the sinusoidal

input and the steady-state output of a system is termed the
frequency response.

-The frequency response test on a system or a component is

normally performed by keeping the amplitude "A" fixed and
determining "B" and "f" for a suitable range of frequencies, w.
I.e. it is easy to perform the test.
-Signal generators are readily available for various ranges of
frequencies and amplitudes
- The design and parameter adjustment of the open-loop
transfer function of the system for specified closed-loop
performance is carried out somewhat more easily in
frequency domain than in time domain.
- The effect of noise disturbance and parameter variations are
relatively easy to visualize and assess ()‫يُخمن‬
The frequency response test is not recommended
for systems with very large time constants.

2
 4/22/2024

The frequency response is easily evaluated from the sinusoidal

transfer function which can be simply obtained by replacing "s" with
"jw" in the system transfer function T(s). The transfer function T(jw)
thus obtained, is a complex function of frequency and has both a
magnitude and a phase angle.

4. 2 Correlation Between Time And Frequency Response:

The correlation between

time and frequency response
has an explicit form only for
systems of first- and second-
order.

For the given Figure, the transfer function is given by:

…………………………(4.4)

The transfer function of the system in frequency domain is:

……4.5) 5

where u= . The modulus (or magnitude) and phase

angle can be found from Eq.(4.5) as:
………..……………(4.6)

………………….(4.7)

3
 4/22/2024

The steady-state output of the system for a sinusoidal input

of unit magnitude and variable frequency, ω, is given by:

From Eq. (4. 6) and Eq. (4. 7), it is seen that when
u = 0, Μ = 1 and f=0
u = 1, Μ= and f=

4
 4/22/2024

The resonant frequency:

It is defined as the frequency, where Μ has a peak value. At
this frequency, the slope of the curve equals zero. Let wr be

the resonant frequency and ur = be the normalized resonant

frequency, thus we have:

 -4(1-u
2 u2 r 

dM r)u +8r
= - 1 ×  
= 0 , this gives:
du u=u 2 3
r 
(1-u 2 2 u )  2  2
)+(2

 r r 
1-22 ….(4.8) 9
or u =

and wr=wn 1-22

The resonant Peak:

Substituting into Eq.(4.6), the maximum value of the magnitude
(known as the resonant peak), could be obtained as:

…………………..……………………(4.9)

and the phase angle at resonant frequency, wr, obtained from

Eq.(4.7) is given by:
……………(4.10)

It is obvious that when  approaches zero in Eq.(4.8), wr

approaches wn and Μr approaches infinity in Eq.(4.9). For
0<< , the resonant frequency always has a value less
than ωn and the resonant peak has a value greater than 1.

10

5
 4/22/2024

Cut-off frequency wc
The frequency at which Μ decreases and reaches to . The
response at signal frequencies above cut-off are greatly
attenuated (‫)يحدث ضعف لالستجابة‬.

Bandwidth of the system wb:

It is defined as: the band of frequencies from zero to cut-off
frequency, wc . The bandwidth of a control system indicates its
capacity to satisfactorily response. In general, it may be defined as
the band of frequencies at which the magnitude is above (or equal)
0.707.

11

11

- Example: Determine the resonance peak, Mr, and resonant frequency,

wr, for the system
C(s) 5
T(s) = =
R(s) s2 +2s+5
C(jw) 5 5
T(jw) = = =
R(jw) (jw)2 +2(jw)+5 5-w2 +2(jw)

M= 5 = 5 = 5
(5-w2 )2+4w2 (25-10w2+w4 )+4w2 (25-6w2+w4 )
= 5(25 - 6w2 + w4 )-0.5

dM = 0  (-12w + 4w 3 ) = 0 i.e. w 2 = 3  w = 1.7321

dw r r r r

Substituting into M equation; yields:

Mr=5/sqrt(25-6*3+9)=5/4=1.25
12

12

6
 4/22/2024

4.3 Polar Plots

The sinusoidal transfer function G(jw) is a complex
function and is given by:
……….................... (4.14)
.............................. (4.15)
Phase-angle, f, is always measured positively in
counter clockwise direction (‫)عكس اتجاه عقارب الساعة‬.
Polar plot is the locus of the tip of the phas or G(jw) as
the input frequency, ω, is varied from 0 to 2, and the
corresponding values of magnitude, Μ, and phase-
angle, f.

13

13

For example, consider a simple RC filter shown in

figure, we have:

, where t = RC.

…….(4.16)

M= .................................. (4.17)

14

7
 4/22/2024

The polar plot of G(jw) of Eq.(4.17) is drawn in the Figure,

(when w = 0, Μ =1 and f= 0). When and f = -
45°.

15

15

b) a second-order system of transfer function

................ (4.18)

It should be noted that

for a second-order
system, for all damping
ratios , when ω=ωn,
the phase lag is 90° and
the magnitude is

16

16

8
 4/22/2024

For example a third-order system:

then by substituting different values of frequencies, the following table

can be obtained :

17

17

The plots is asymptotic to the vertical line passing

through the point (-t,0).

Polar plot for G(s) =

18

18

9
 4/22/2024

The major advantage of the polar plot lies in stability

study of systems which was developed by N. Nyquist in
1932. Because of his work, the polar plots are commonly
referred to as Nyquist plots.

Inspection and comparison of all last few figures, the

following observations would be made:
(i) Addition of a nonzero pole to a transfer function results in further
rotation of the polar plot through an angle of -90° as w→.
(ii) Addition of a pole at origin to a transfer function rotates the polar
plot at zero and infinite frequencies by a further angle of -90°.
(iii) The effect of addition of a zero to a transfer function is to rotate
the high frequency portion of the polar plot by 90° in counter-
clockwise direction.

19

19

4.4 Bode Plots:

A logarithmic plot which consists of two
graphs, one for the logarithm of |G(jw)|
and the other for phase angle of G(jw).
Both plots are plotted against
frequency in logarithmic scale.

20

20

10
 4/22/2024

21

22

11
 4/22/2024

𝟒 (𝟏+𝟎.𝟓 𝑺)
- 𝑮 𝒔 𝑯 𝒔 =
𝒔𝟐 𝟏+𝟎.𝟐 𝑺 (𝟏+𝟎.𝟎𝟐 𝑺)
i. The transfer function is given into time-constant form.
ii. The sinusoidal form is obtained by replacing (s) by jw as:
𝟒 (𝟏+𝟎.𝟓 𝒋𝒘)
𝑮 𝒋𝒘 𝑯 𝒋𝒘 =
−𝒘𝟐 𝟏+𝟎.𝟐 𝒋𝒘 (𝟏+𝟎.𝟎𝟐 𝒋𝒘)
iii. The factors of this transfer function in order of their occurrence as frequency
increases are:

(1) constant gain, Κ = 4,

The constant gain is M= 20 log(4)  12 db, Slope=0 db/decade , Ф1 =0

(2) 1st order (simple) zero at s = -2; corner frequency w1= 2, Slope=20
db/decade , Ф2=-tan-10.5w
The asymptotic approximation of the magnitude of the zero at w1= 2 has a slope of
+20 db/decade just after the corner frequency at w1= 2 and zero decibel before it.

23

(3) double pole at origin, Slope=-40 db/decade , Phase angle, ф3=-180o.

The magnitude of the pole at the origin extends from zero to infinity frequency and
has a slope of -40 db/decade intersecting the 0-db line at w = 1.

(4) 1st order simple pole at s = -5; corner frequency w2 = 5. Slope= -20 ,
ф4=-tan-10.2w

The asymptotic approximation of the magnitude of the simple pole at w2= 5 has a
slope of -20 db/decade just after the corner frequency at w2= 5 and zero decibel
before it.

(5) 1st order simple pole at s = -50; corner frequency w3 = 50. Slope=-20 ,
ф5=-tan-10.1w
0.02 w
The asymptotic approximation of the magnitude of the simple pole at w3 = 10 has a
slope of -20 db/decade just after the corner frequency at w3 = 10 and zero decibel
before it.

24

12
 4/22/2024

The constant gain is M= 20 log(4)  12 db, Slope=0 ) 1st order (simple) zero at s = -2; corner frequency w1= 2, Slope=20
(3) double pole at origin, Slope=-40 db/decade (4) 1st order simple pole at s = -5; corner frequency w2 = 5. Slope= -20
(5) 1st order simple pole at s = -50; corner frequency w3 = 50. Slope=-20
0.1 1 10 100 1000 10000 105

80

60

40 40

20 20
1, slope 0

0 0

-20 -20

-40 -40

-60 -60

-80

25

At each corner frequency and half and double

w Error Net w Net error Phase
ω1 = 2 (db) (ø)
½ ω1= 1 2 3 1 1
2 ω1 = 4 1 1 2 3
4 1 2.5 -1
5 -3 4 1
2.5 -1 5 -3
10 -1 10 -1
50 -3 25 1-
25 -1 50 -3
100 -1 100 -1

26

13
 4/22/2024

27

0.1 1 10 100 1000 10000 105

-45

-90

-135

-180

-225
-270

28

14
 4/22/2024

20
20log Magnitude

Asymptotes
0 -20 db/decade

-20
Exact Curve

-40
-1 0 1 2
10 1 1 10 2 10 10
2t t t w

-20
phase Deg

-40

-60

-80
-1 0 1 2
10 10 10 10
w

29

29

The transfer function of a system is represented by:

........................................................(4.22)
Taking the normal logarithm (base 10) on both sides, yields:

= log |G( jw)| + 0.434 j f(w) ...............................(4.24)

The standard procedure is to plot 20 log and phase angle

f(w) versus the frequency, w, which is, usually, drawn on a logarithmic

scale. In this representation, the unit of magnitude, 20 log |G(jw)|, is
decibel, abbreviated as db.

As an example, consider the RC filter with transfer function as given in

Eq.(4.17),

The logarithmic gain (or magnitude) is:

................ (4.25)
30

30

15
 4/22/2024

For small frequencies, , the log-magnitude is approx. as

20 log |G(jw)| = - 10 log(1) = 0 db.................................... (4.26)

For large frequencies, w>> , the log-magnitude is approx. as

20 log |G( jw)| = -20 log ωt ................................................. (4.27)
=-20 log ω - 20 log t ................................... (4.28)
The plot of Eq.(4.28) is a straight line with a slope = -20 db per unit change in
log w. A unit change in log ω means:
ω
log ω - log ω = log( 2 )=1 or w2 = 10 w1
2 1 ω
1
This range of frequencies is called a decade. Thus the slope of Eq..(4.28) is -
20 db/decade. The range of frequencies w2 = 2 w1 is called an octave. At one
octave below, f=-26.57, and at one octave above, f=-63.43

The following plot can then be constructed:

31

31

20
20log Magnitude

Asymptotes
0 -20 db/decade

-20
Exact Curve

-40
-1 0 1 2
10 1 1 10 2 10 10
2t t t w

-20
phase Deg

-40

-60

-80
-1 0 1 2
10 10 10 10
w

32

32

16
 4/22/2024

The frequency w= 1t , at which the two asymptotes meet, is called

the corner frequency or the break frequency. This break
frequency divides the plot in two regions; a low frequency region
and a high frequency region.

33

33

The actual plot can be obtained by applying corrections for the errors
introduced by asymptotic approximation. During the period 0 < w 
, the error can be evaluated by subtracting Eq.(4.26) from Eq.(4.2.5).
Therefore, the error at the corner frequency w = is given by:

-10 log (1 + 1) + 10 log(1)  -3 db................................ (4.29)

The error at frequency of one octave below the corner frequency, at

w= , is given by: 1
2t
-10 log(1 + ) 1+ 10 log(1)  -1.0 db ................................(4.30)
4
Similarly for the frequency range  w < , the error in log-magnitude
is given by the difference between Eq.(4.25) and Eq.(4.27) as:

..................................... (4.31)

34

17
 4/22/2024

The error at the corner frequency is found to be:

-10 log(1+1) + 20 log(1)  - 3 db ............................... (4.32)
and the error at frequency one octave above corner frequency,
at w = 2
t , is:
-10 log(1+ 4) + 20 log(2)  -1 db ............................... (4.33)
In practice, a sufficiently accurate log-magnitude plot is obtained
by correcting the asymptotic plot by -3 db at the corner frequency and
by -1 db one octave below and one octave above the corner frequency.
Then a smooth curve is drawn through these three points approaching
the low and high frequency asymptotes.

The phase angle of this network, ....... (4.34)

Since the phase angle is given by inverse tangent function,

Eq.(4.34), the phase characteristic is symmetric about the
inflection point = - 45°.

35

35

4.4.1 Construction of Bode plot for quadratic

factor with complex poles
The factor may be written in normalized form as
.............................................. (4.36)

where u = ww is the normalized frequency. The log-magnitude

n
is given by:

.............................. (4.37)
For u <<1, the log-magnitude is given by:
20 log |G(jw)|  -10 log(1) = 0 ...................................(4.38)
and for u >>1, the log-magnitude is:
.............................. (4.39)
36

36

18
 4/22/2024

Therefore, the log-magnitude curve of the quadratic factor

under consideration, consists of two straight-line
asymptotes, one horizontal line at 0 db for u << 1 and the
other, a line with slope -40 db/decade for u >>1. These two
asymptotes intersect on 0-db line at u = 1, i.e. at w = wn
which represents the corner frequency of the plot.
The error between
the actual magnitude
and the asymptotic
approximation is
shown in error plot
and is given by:

a) for 0 < u  1, the

error is:

........................ (4.40)
37

37

b) for 1 < u  ∞, the error is:

..................................(4.41)
It might be noticed that the error is a function of  and u.
The phase angle of the quadratic factor is given by:
.........................................................(4.42)
which is also a function of both u and  · The phase
angle is zero at u = 0, -90° at u= 1, and -180° at u = . The
plots for zeros factor are typical the same as those for poles
but with opposite signs.
4.4.2 General procedure for constructing Bode plots
Draw the Bode plots for the transfer function:
1st., test the 2nd. order gain for
complex roots. If so, then find
wn and . i.e.; b -4ac=3.2*3.2-4*1*64<0, then compare with 2nd
2

order general form (s2+2wns+wn2) yields; wn2=64 and wn=8,

also 2wn=3.2 leads to =0.2
38

38

19
 4/22/2024

1. Rewrite the sinusoidal transfer function in the time-

constant form by replacing (s) with jw as:
............................. (4.43)

2. Identify the corner frequencies associated with each factor

of the transfer function in order of their occurrence as
frequency increases .
(i) constant gain, Κ = 4,
(ii) pole at origin,
(iii) pole at s = -0.5; corner frequency w1= 0.5,
(iv) zero at s = -2; corner frequency w2 = 2,
(v) pair of complex poles with  = 0.2, wn = 8; corner
frequency w3= 8.

39

39

3. Draw the asymptotic magnitude plot. This plot consists of

straight-line segments with line slope changing at each
corner frequency by + 20 db/decade for a zero and - 20
db/decade for a pole. For complex conjugate zero or pole
the slope changes by ± 40 db/decade.
(i) The constant gain is 20 log(4)  12 db in a horizontal line.
(ii) The magnitude of the pole at the origin extends from zero
to infinity frequency and has a slope of -20 db/decade
intersecting the 0-db line at w = 1 and f=-90.
(iii) The asymptotic approximation of the magnitude of the
pole at w1= 0.5 has a slope of -20 db/decade just after the
corner frequency at w1= 0.5 and zero decibel before it.
(iv) The asymptotic magnitude for zero at w2 = 2 has a slope
of +20 db/decade just after the break frequency at w2 = 2
(ν) The asymptotic approximation for the pair of complex
poles at w3= 8 has a slope of -40 db/decade due to the
quadratic form and f=-tan-1(2u/(1-u2)).
40

40

20
 4/22/2024

4. From the error plot or calculations, determine the

corrections to be applied to the asymptotic plot, then
construct the following table:
Individual Frequency corrections net net corrections
frequency
Angle f
ω1 = 0.5 from Eq.(4.29) = -3 db 0.25 -1 db -110
½ ω1= 0.25 from Eq.(4.30) = -1 db 0.5 -3 db -122
2 ω1 = 1 from Eq.(4.33) = -1 db 1 -1+1=0 db -130
ω2 = 2 2 +3 db -127
from Eq.(4.29) = +3 db
½ ω2= 1 from Eq.(4.30) =+1 db 4 +1+2=+3 db -124
2 ω2= 4 from Eq.(4.33) = +1 db 8 +8 db -191
ω3= 8, =0.2 from Eq.(4.40) = +8 db 16 +2 db -260
½ ω3 = 4 from Eq.(4.40) = +2 db
32 -267
2 ω3= 16 100 -269
from Eq.(4.41) = +2 db
41

41

2

-20 db/decade

constant gain


pole at origin
 +20 db/decade
20log Magnitude

−

− 
-40 db/decade
− 
-20 db/decade

− 2
−   2
   
w

42

42

21
 4/22/2024


-20 db/decade
 -20 db/decade

2
-40 db/decade

20log Magnitude

-40 db/decade -60 db/decade



− 

− 2 -20 db/decade

-60 db/decade
− 

− 
−   2
  4.0  
0.25 w 16
0.5 8.0
2.0
43

43

5. Draw a smooth curve

through the corrected
points such that it is
asymptotic to the line
segments. This gives
the actual log-
6. The phase characteristic magnitude plot.
may be obtained by
adding the phase due to
each individual factor as
follows:
(i) The phase of the constant
gain is, of course, zero
degrees.
(ii) The phase of the pole at
the origin is a constant of
-90°.
44

44

22
 4/22/2024

(iii) The phase angles of the factor (1+ jtw)1are determined

from Eq.(4.34) as: f=± tan-1(tw) ± 45° at w = 1 t
after locating these points, a smooth 1
± 26.6° at w = 2t
curves are to be passed through them
to produce the phase plots of factors: ±5.7° at w = 1
(1+ jtw)1 t
± 63.4° at w = 2 t
(iv) The phase angles of the factor :

2
-1 ± 84.3° at w = 10t
   is
 w − w  
  
1+ j2 

w  w  
   

 n  n
 
 

are determined from Eq. (4.42) as: 0° at ω = 0, -90° at w= 8, -

180° at w→ .
The total angle=-90-tan-1(2w)+tan-1(0.5w)-tan-1(2*0.2w/(64-w2))

45

45

 (1+ j0.5w)


K
-1
 w  w 

2
− (1+ j2w) -1
1+j0.4 + j  
 j1w 
8  8 

Phase Phi

−

−
Resultant Phase
−2

−2
−   2
   
w
46

46

23
 4/22/2024

The following figure shows the Bode plots for a quadratic log
of Eq.(4.37) and Eq.(4.42) for different values of damping
ratio .

Bode plots for

G(s)=

47

47

Example:
Draw Bode and Polar plots for G(s)=e-ts
G(jw)=e-tjw=1 -tw radians, which can be represented as:

48

48

24
 4/22/2024

4.5 Log-Magnitudes versus Phase angles diagram

An alternative approach
to graphically portraying the
frequency response is to plot
the logarithmic magnitude in
db versus the phase angle
for a range of frequencies.

Since this information is

equivalent to that portrayed by
the Bode diagram, it is
normally easier to draw the
Bode plots first, then the
information should be
5(1+ j ω)
transferred to the coordinates
of the log magnitude versus G(jw)= 10
 2
 jω  
phase diagram.  ω  
 ω
jω 1+ j  1+ j0.6 +   
 
49  2  50  50  
 

49

The corner frequencies

are so adjusted that the
db value at the corner
frequency on the
asymptotic plot differs
from actual log-
magnitude plot by an
amount which is in close
agreement with the db
correction of the known
studied kind of factors.
3. In the low frequency range, the plot is dominated by a
K
factor of the form. In most practical systems, “r”
(jw)r
equals 0, 1, or 2. The value of r is determined as follows:
(a) if the low frequency asymptote is horizontal line, r=0, at x
db, it indicates that the transfer function represents a type-
50
0 system with a gain Κ given by 20 log k=x.
50

25