CONTROL SYSTEMS DESIGN BY

FREQUENCYRESPONSE APPROACH

November, 2018

1 Introduction
The root-locus method was shown to be useful to reshape the transient-response
characteristics of closed-loop control systems. The frequency-response approach,
on the other hand, gives us this information only indirectly. However, as we shall
see, the frequency-response approach is also very useful in designing control sys-
tems. For any design problem, the designer will do well to use both approaches
to the design and choose the compensator that most closely produces the desired
closed-loop response.
In the frequency-response approach, we specify the transient-response per-
formance in an indirect manner.That is, the transient-response performance is
speci…ed in terms of the phase margin, gain margin, resonant peak magnitude
(they give a rough estimate of the system damping); the gain crossover fre-
quency, resonant frequency, bandwidth (they give a rough estimate of the speed
of transient response); and static error constants (they give the steady-state
accuracy). Although the correlation between the transient response and fre-
quency response is indirect, the frequency-domain speci…cations can be easily
met in the Bode diagram approach. After the open loop has been designed, the
closed-loop poles and zeros can be determined. Then, the transient-response
characteristics must be checked to see whether the designed system satis…es the
requirements in the time domain. If it does not, then the compensator must be
modi…ed and the analysis repeated until a satisfactory result is obtained.
Design in the frequency domain is simple and straightforward. The fre-
quency response plot indicates clearly the manner in which the system should
be modi…ed, although the exact quantitative prediction of the transient-response
characteristics cannot be made. The frequency-response approach can be ap-
plied to systems or components whose dynamic characteristics are given in the
form of frequency-response data. Note that because of di¢ culty in deriving
the equations governing certain components, such as pneumatic and hydraulic
components, the dynamic characteristics of such components are usually de-
termined experimentally through frequency-response tests.The experimentally
obtained frequency-response plots can be combined easily with other such plots

1
when the Bode diagram approach is used. Note also that in dealing with highfre-
quency noises we …nd that the frequency-response approach is more convenient
than other approaches.
A Bode diagram of the compensator can be simply added to the original Bode
diagram, and thus plotting the complete Bode diagram is a simple matter.Also, if
the open-loop gain is varied, the magnitude curve is shifted up or down without
changing the slope of the curve, and the phase curve remains the same. For
design purposes, therefore, it is best to work with the Bode diagram.
A common approach to the design based on the Bode diagram is that we …rst
adjust the open-loop gain so that the requirement on the steady-state accuracy
is met.Then the magnitude and phase curves of the uncompensated open loop
(with the open-loop gain just adjusted) are plotted. If the speci…cations on the
phase margin and gain margin are not satis…ed, then a suitable compensator
that will reshape the open-loop transfer function is determined. Finally, if there
are any other requirements to be met, we try to satisfy them, unless some of
them are mutually contradictory.
We might say that, in many practical cases, compensation is essentially a
compromise between steady-state accuracy and relative stability. To have a
high value of the velocity error constant and yet satisfactory relative stability,
we …nd it necessary to reshape the open-loop frequency-response curve. The
gain in the low-frequency region should be large enough, and near the gain
crossover frequency, the slope of the log-magnitude curve in the Bode diagram
should be –20 dB decade.This slope should extend over a su¢ ciently wide fre-
quency band to assure a proper phase margin. For the high-frequency region,
the gain should be attenuated as rapidly as possible to minimize the e¤ects of
noise. Examples of generally desirable and undesirable open-loop and closed-
loop frequency-response curves are shown next.

2 Lead/Lag Compensation
It is possible to consider three di¤erent kinds of frequency compensation: Lead,
Lag and Lead/Lag compensation. Lead compensation essentially yields an ap-
preciable improvement in transient response and a small change in steady-state
accuracy. It may accentuate high-frequency noise e¤ects. Lag compensation,
on the other hand, yields an appreciable improvement in steady-state accuracy
at the expense of increasing the transient-response time. Lag compensation

2
will suppress the e¤ects of high-frequency noise signals. Lag–lead compensa-
tion combines the characteristics of both lead compensation and lag compen-
sation.The use of a lead or lag compensator raises the order of the system by
1.The use of a lag–lead compensator raises the order of the system by 2, which
means that the system becomes more complex and it is more di¢ cult to control
the transient-response behavior.The particular situation determines the type of
compensation to be used.

2.1 Lead compensator

We shall …rst examine the frequency characteristics of the lead compensator.Then
we present a design technique for the lead compensator by use of the Bode di-
agram. Consider a lead compensator having the following transfer function:

The major characteristic of the lead compensator is the positive phase shift
in the intermediate frequencies. The maximum phase shift max occurs at the
frequency ! = ! max , which is the geometric mean of zc and pc .

p 1
! max = zc p c = p
1
max = arcsin( )
1+
The shift in the magnitude curve at intermediate and high frequencies is
undesired but unavoidable. Proper design of the compensator requires placing
the compensator pole and zero appropriately so that the bene…ts of the positive
phase shift are obtained and the magnitude shift is accounted for. The following
paragraphs show how this can be accomplished. Next …gure shows the Bode
plots of magnitude and phase for a typical lead compensator. The values in this
example are Kc = 1; pc = 2:5, and zc = 0:4; so = 0:4=2:5 = 0:16. Changing
the gain merely moves the magnitude curve by 20 log10 (jKc j).

3
4
 The following steps outline the procedure that will be used to design the
phase lead compensator to satisfy steady-state error and phase margin speci…-
cations. Assume Gp (s) is the transfer function of the plant and P Mspec is the
speci…ed phase margin.
1) Determine if the system type needs to be increased in order to satisfy
the steady-state error speci…cation, and if necessary, augment the plant with
the required number of poles at s = 0. Calculate Kc to satisfy the steady-state
error.
G (s)
2) Make the Bode plots of G(s) = Kc sNreqp Nsys
3) Determine the amount of phase shift in G(j!) at the gain crossover fre-
quency and calculate the uncompensated phase margin P Muncomp . The uncom-
pensated phase margin is the vertical distance between 180 and the phase
curve measured at the gain crossover frequency ! 0dB
4) Calculate the values for max and that are required to raise the phase
curve to the value needed to satisfy the phase margin speci…cation;

max = P Mspec P Muncomp + 10

1 sin( max )
= arcsin
1 + sin( max )
5) Determine the value for ! max , which is the frequency such that
jG(j! max )jdB = 10 log( )
6) Compute the lead compensator’s zero and pole.
1
= p
! max
1
zc =
1
pc =

4
Problem 1: Assume Gp (s) = s(s+2) . Design a Lead compensator such that

Kv = 20
P Mspec = 50
0:237s+1
Solution 1: Gc (s) = 10 0:047s+1
20
Problem 2: Assume Gp (s) = s(s+3)(s+10) . Design a Lead compensator
such that
Kv = 10
P Mspec = 30
0:28s+1
Solution 2: Gc (s) = 15 0:08s+1