Frequency Domain Analysis of Control System

By the term frequency response, we mean the steady-state response of a system to a sinusoidal
input. Industrial control systems are often designed using frequency response methods. Many
techniques are available in the frequency response methods for the analysis and design of control
systems.
Consider a system with sinusoidal input
( sin r t A t
. !he steady-state output may be
written as,
( sin( c t B t +
. !he magnitude and the phase relationship between the sinusoidal
input and the steady-state output of a system is called frequency response. !he frequency response
test is performed by "eeping the amplitude A fi#ed and determining B and for a suitable range of
frequencies. $henever it is not possible to obtain the transfer function of a system through
analytical techniques, frequency response test can be used to compute its transfer function.
!he design and ad%ustment of open-loop transfer function of a system for specified closed-
loop performance is carried out more easily in frequency domain. &urther, the effects of noise and
parameter variations are relatively easy to visuali'e and assess through frequency response. !he
(yquist criteria is used to e#tract information about the stability and the relative stability of a
system in frequency domain.
Correlation between time and frequency response
!he transfer function of a standard second-order system can be written as,
)
) )
(
( .
( )
n
n n
C s
T s
R s s s

+ +
*ubstituting s by j we obtain,
)
) ) )
+
( .
( ) ( (+ )
n
n n
T j
j j u j u

+ + +
$here,
,
n
u
is the normali'ed signal frequency. &rom the above equation we get,
) ) )
+ )
+
(
(+ ()
( tan -) ,(+ .
T j M
u u
T j u u

+

.
!he steady-state output of the system for a sinusoidal input of unit magnitude and variable
frequency is given by,
+
)
) ) )
+ )
( sin tan
+
(+ ()
u
c t t
u
u u

, +
.
It is seen from the above equation that when,
/, + and /
+
+, and , )
)
, / and
u M
u M
u M

!he magnitude and phase angle characteristics for normali'ed frequency u for certain values of
are shown in figure in the ne#t page.
!he frequency where M has a pea" value is called resonant frequency. 0t this point the slope of the
magnitude curve is 'ero. *etting
/
r
u u
dM
du

we get,
) )
1, )
) ) )
2(+ 3
+
/.
)
(+ ()
r r r
r r
u u u
u u

1 +
]

1 +
]
*olving,
)
+ )
r
u or, resonant frequency
)
+ )
r n
. 4444 44 (/+
!he resonant peak is given by, resonant pea",
)
+
) +
r
M

. 444444 (/)
&or,
+
( /.5/5
)
>
, the resonant frequency does not e#ist and M decreases
monotonically with increasing u.
&or
+
/
)
< <
, the resonant frequency is always less than
n

and the resonant pea" has a

value greater than +.
&rom equation (/+ and (/) it is seen that !he resonant pea"

r
M
of frequency response is
indicative of damping factor and the resonant frequency
r

is indicative of natural frequency for

a given

and hence indicative of settling time.

&or
r
>
, M decreases monotonically. !he frequency at which M has a value of
+
)
is called the
cut-off frequency
c

. !he range of frequencies over which M is equal to or greater than

+
)
is
defined as bandwidth,
b

.
!he bandwidth of a second-order system is given by,
+, )
) ) 2
+ ) ) 2 2
b n

1
+ +
]
4444.(/1
&igure below shows the plot of resonant pea" of frequency
response and the pea" overshoot of step response as a
function of

.
It is seen that the two performance indices are correlated
as both are the functions of the system damping factor

only. &or
+
( /.5/5
)
>
the resonant pea" does not
e#ist and the correlation brea"s down. &or this range of

,
p
M
is hardly perceptible.
&rom equation (/1 it is seen that the bandwidth is
indicative of natural frequency and hence indicative of
settling time, i.e., the speed of response for a given

.
Polar Plot
!he polar plot of a sinusoidal transfer function
( G j
is a
plot of the magnitude of
( G j
versus the phase angle of
( G j on polar coordinates as is varied from 'ero to
infinity. 0n advantage of using polar plot is that it depicts
the frequency response characteristics of a system
over the entire frequency range in a single plot.
!he polar plot of

+
) )
+ +
( tan
+
+
G j T
j T
T

+
+
is shown in
figure below.
!he polar plot of the transfer function,
+
(
(+
G j
j j T

+
is shown in figure
above.
!he plot is asymptotic to the vertical line
passing through the point (-T, 0).
6olar plots are useful for the stability study of
systems. !he general shapes of the polar plots
of some important transfer functions are given
in figure below.
&rom the plots above, following observations are made,
+. 0ddition of a non'ero pole to the transfer function results in further rotation of the polar plot
through an angle of -7/ as
.
). 0ddition of a pole at the origin to the transfer function rotates the polar plot at 'ero and
infinite frequencies by a further angle of -7/.
Bode Plots
!he transfer function
( G j
is represented by,
(
( (
j
G j G j e


.
!a"ing natural logarithm of both sides,
ln ( ln ( ( G j G j j +
444444444(/2
!he unit of real part is called neper.
*imilarly,
log ( log ( /.212 ( G j G j j +
44444444444444..(/8
!he standard procedure is to plot
)/log ( G j
and phase angle
(
vs. log

. !he unit of
magnitude
)/log ( G j
is decibe. !hese two plots are called Bode pots in honor of 9$ Bode.
Example
+
) )
+ +
( tan
+
+
G j T
j T
T

+
+
.
!he log-magnitude is, ( )
) )
)/log ( +/log + G j T +
.
&or low frequencies ( +, T = , the log magnitude is appro#imated as,
)/log ( +/log+ / G j
db. (/+
&or high frequencies ( +, T ? , the log magnitude is appro#imated as,
)/log ( )/log )/log G j T
. (/)
!he logarithmic plot of equation (/+ is a straight line coincident with the hori'ontal a#is. !he plot
of equation (/) is also a straight line with a slope -)/ db per unit change in
log
. 0 unit change of
log
means
) +
log( , +
or,
) +
+/
.
!his range of frequencies is called a decade. !he slope of the equation (/) is -)/ db,decade.
!he range of frequencies
) +
)
is called an octa!e. *ince -)/log ) : - ; db, the slope -)/
db,decade can also be e#pressed as -)/ db,octave.
&urther at +, T the plot has a value of / db. !he plot is shown in figure below.
!he straight line appro#imation holds good for +, T = and +, T ? . $ith some loss of accuracy
these could be e#tended for frequencies +, T and +, T . !he frequency +, T at which the
two asymptotes meet is called the corner frequency. !he corner frequency divides the plot in low
and high frequency regions.
!he actual log-magnitude plot can be obtained by applying corrections for the errors
introduced by asymptotic appro#imation. !he error at the corner frequency +, T is,
) )
+/log(+ +/log+
+/log(+ + +/log+ 1 db
T + +
+ +
!he error at the corner frequency
+, )T is,
) )
+/log(+ +/log+
+/log(+ +, 2 +/log+ + db
T + +
+ +
.
&or +, T < , the error in log-magnitude
is given by,
) )
+/log(+ )/log T T + + .
!he error caused by the asymptotic plot is shown in figure above.
Simple Rules for Plotting Bode Diagrams
!he open-loop transfer function for a linear system can be written in the form,
( )
)
(+ (+
(
(+ (+ + )
a a
n
a a
n n
" j T j T
G j
j j T j T j j

+ +

1
_ _
1 + + + +

1 , ,
]
L
L L
.
Bode diagram can be s"etched for any general system by simply adding the effects of each pole and
each 'ero in order to determine the angles and intersection points of the asymptotes.
! Factors of t"e form ,(
r
" j
!he log magnitude of this factor is
( )
)/log )/ log )/log
r
"
r "
j

+
and the phase is,
( 7/ r
o
. $ith
log
as abscissa, the plot of above equation is a straight line having a slope
of -#0r db,decade and passing through #0o$ " at + . !his is shown in figure below for r : /,
+, ) and 1.
#! Pole or $ero on t"e real axis
!he pole factor
+,(+ j T +
has e#plained in the previous e#ample. !he phase angle for this
factor is
+
tan T

. 0t corner frequency, the phase angle of this factor is -28. 0t 'ero

frequency it is / and at infinity it is -7/.
!he bode plot for the 'ero factor
(+ j T +
has a slope of <)/ db,decade and a phase angle of
+
tan T

+ . !he db correction is added to the asymptotic plot.

%! Complex con&ugate poles
!he quadratic factor for a pair of comple# con%ugate poles may be written in normali'ed form as
) )
+ +
=
+ )
+ )
n
n n
u
j u u
j j

+
_ _
+ +

, ,
.
!he log-magnitude of this factor is,
+, )
) ) )
)
) ) )
+
)/log )/log (+ ()
+ )
+/log (+ ()
u u
j u u
u u

1 +
]
+
1 +
]
.
&or + u = ,
)
+
)/log +/log+ /
+ ) j u u

+
.
&or + u ? ,
2
)
+
)/log +/log 2/log
+ )
u u
j u u

+
.
!hus, the log-magnitude curve of the quadratic factor consists of two straight-line asymptotes,
one hori'ontal line at / db for + u = and the other, a line with a slope -2/ db,decade for + u ? .
!he two asymptotes meet on /-db line at u : +. !he asymptotic and the actual plots are shown in
figure right.
!he error between the actual magnitude and the asymptotic appro#imation is as given below>
&or
/ +, u <
the error is
) ) ) )
+/log (+ 2 +/log+ u u 1 + +
]
&or
+ , u <
the error is
) ) ) )
+/log (+ 2 2/log u u u 1 + +
]
!he error versus
u
curves for different values of

are shown in figure below.

!he phase angle of the quadratic factor is
given by,
+
)
)
tan
+
u
u

,
.
!he phase angle plot are shown in figure
above. !he phase angle curve also
depends on

.
'eneral Procedure for Constructing Bode Plots
!he following steps will be used in constructing the bode plot for a given
( G j
.
+. $rite the sinusoidal transfer function in time-constant form.
). Identify the corner frequencies associated with each factor of the transfer function.
1. ?nowing the corner frequencies, draw the asymptotic magnitude plot.
2. &rom the error graphs, determine the corrections to be applied to the asymptotic plot.
8. @raw a smooth curve through the corrected points such that it is asymptotic to the line
segments. !his gives the actual log-magnitude plot.
;. @raw phase angle curve for each factor and add them algebraically to get the phase plot.
Example
) )
;2( ) 2(+ , )
(
( /.8( 1.) ;2 (+ ) (+ /./8 , ;2
s s
G s
s s s s s s s s
+ +

+ + + + + +
!
!he sinusoidal transfer function in time-constant form is,
)
2(+ , )
(
(+ ) (+ /.2
3 3
j
G j
j j j

_ _
+ +

, ,
!
&actor
c
f Aog-magnitude characteristic 6hase angle
characteristic
2, j - *traight line of slope -)/ db,decade,
passing through
)/log 2 +)
db point
at + .
Constant
-7/
+,+ ) j +
+

:
/.8
*traight line of / db for
+
<
,
straight line of slope -)/ db,decade for
+
>
.
/ to -7/
o
,
28
o
at
+

.
+ /.8 j +
)

:
)
*traight line of / db for
)
<
,
straight line of slope <)/ db,decade
for
)
>
.
/ to <7/
o
,
28
o
at
)

.
)
+ /.2
3 3
j
_ _
+

, ,
=
3, /.)
n

1

:
3
*traight line of / db for
1
< ,
straight line of slope -2/ db,decade for
1
>
.
/ to +3/
o
7/
o
at
1

.
!o the asymptotic plot, corrections are to be applied to get the actual plot. !he following list
shows the list of corrections obtained from the error versus log-magnitude curve (plots are given in
the previous pages.
+
+ +
)
) )
1
1 1
&requency Correction
/.8 -1 db
, ) /.)8, ) + -+ db
) <1 db
, ) +, ) 2 <+ db
3, /.) 3 db
, ) 2, ) +; ) db

+
+
&requency (et Correction
/.)8 -+ db
/.8 -1 db
+ / db
) <1 db
2 1 db
3 3 db
+; <) db
+
+
!he phase angle curve may be drawn using the following procedure.
+. &or the factor ( ) ,
r
" j , draw a straight line of -%0r.
). !he phase angles of the factor
+
(+ j T
t
+ are
a. 28 at :+,! t
o
b. );.; at :+,)! t
o
c. 8.5 at :+,+/! t
o
d. ;1.2 at :),! t
o
e. 32.1 at :+/,! t
o
1. !he phase angles for the quadratic factor are
a.
n
7/ at :
o
b. 0 few points of phase angles are read off from the normali'ed Bode plot for the
particular

.
Experimental Determination of (ransfer Functions
!he system transfer function within a certain degree of accuracy can be obtained from bode plots by
fitting an asymptotic log-magnitude plot to the e#perimental data using the procedure outlined
below.
+. Bse the e#perimental data to plot the e#act log-magnitude and phase angle versus frequency
curves on a semilog graph sheet.
). @raw the asymptotes on the log-magnitude curve "eeping slopes as a multiple of
)/ db,decade. t 0d%ust the corner frequencies "eeping in mind the correction factors.
1. If the slope changes by )/ db,decade & at
+
:
, a factor
+
+,(+ ,
&
j + will e#ist in the
transfer function.
2. If the slope changes by )/ db,decade & + at
)
:
, a factor
)
(+ ,
&
j + will e#ist in the
transfer function.
8. If the slope changes by 2/ db,decade at
1
:
, either a double pole or a pair of comple#-
con%ugate pole will e#ist in the transfer function. If the error between the asymptotic and
actual curve is about ; db , then a factor of the form
)
1
+,(+ , j + is present and if the
error is positive, then a quadratic factor of the form
)
1 1
+,-+ ) ( , ( , . j j + + is
present. !he value of

will be obtained from the error graph.

;. In the low frequency range, the plot is determined by a factor of the form, ,(
r
" j .
a. &or type-/ system, the low frequency asymptote is a hori'ontal line at # db such that
)/log . " '
from where we can get " .
b. If the low frequency asymptote has a slope of )/ db,decade. , it will indicate a
factor of the form,
, " j
. !he frequency at which asymptote intersects the /-db line
numerically represents the value of " also the asymptote has a gain of )/ log" at
:+.
c. If the low frequency asymptote has a slope of 2/ db,decade , then the transfer
function has a factor of the form
)
,( " j . !he frequency at which asymptote
intersects the /-db line is numerically equal to
"
also the asymptote has a gain of
)/ log" at :+.
!he log-magnitude curve of type-/, type-+ and type-) systems are shown in figure below.

0fter obtaining the transfer function from the log-magnitude curve, the phase angle curve is
constructed and is then compared with the one obtained e#perimentally. If the two curves are in fair
agreement and the curves at very high frequencies tend to 7/ ( , q p
o
then the transfer function is
of minimum phase type. If the computed phase angle is +3/
o
less negative than the one obtained
e#perimentally, then the transfer function is of nonminimum phase type and one of the 'eros of the
transfer function lies in the right-half s-plane.
!he e#perimental log-magnitude and phase-angle curves of a system are shown in the ne#t page.
&irst draw the asymptotes on the e#perimental curve. &rom the slope changes and the corner
frequencies, the transfer function may be written as,
)
)
8 + +
( (+ ,+/
(+ , ) + ) ( , 8/ ( , 8/
8(+ ,+/

(+ , )+ ) ( , 8/ ( , 8/
G j j
j j j j
j
j j j j

+
+ + +
+

+ + +
0t the corner frequency of 8/ the error between the actual and the appro#imate plot is <2 db.
Corresponding

from the error graph is obtained as /.1.

!he phase angle at very high frequencies is -)5/ which is equal to 7/ (2 + )5/
o o
. !herefore
the Bode plot represents a minimum-phase transfer function.

)og*magnitude +ersus P"ase Plots
It is a plot of log-magnitude versus phase angle with frequency as the running parameter. It is
obtained from the Bode plots by reading the values of db and at different frequencies. !he
relative stability of the closed-loop control systems can be determined quic"ly using this plot.