10

Introduction to automatic control

10.1 Introduction pedal in an attempt to reduce the error to zero. If

This chapter is devoted to an examination of the head wind is such as to cause a rapid increase
elementary mechanical control systems. The of error, the corrective action will not be the same
discussion will be limited to the class of systems as that for a slow change. Thus we observe that
whose motion can be described by linear the driver’s control action takes account not only
differential equations with constant coefficients. of the magnitude of the error, v,, but also the rate
In practice many control systems have non-linear of change of error, dv,ldt.
elements, but the overall motion can very often Later we shall see that in some control systems
be closely approximated to that of a purely linear a measure of the integral Jv,dt is useful. When a
system. The main features of all control systems human operator is part of the control process, his
can be introduced by discussing specific examples. reaction time introduces a finite delay into the
Let us consider the position control of a system, making it non-linear. Such systems are
machine tool which has only straight-line motion. not discussed further here.
Let the actual position of the tool be defined by x,
and the desired position by x i . The variables x, 10.2 Position-control system
and xi are referred to as the system output and the We can now examine in some detail a particular
system input respectively. elementary position-control system and use it to
The system error, x, , is formally defined by introduce the block-diagram notation by which
control systems are often represented.
x, =xi-x, (10.1) A rotatable radar aerial has an effective
and it is the object of the control system to take moment of inertia 1. The aerial is driven directly
corrective action and reduce this error to zero. by a d.c. motor which produces a torque T, equal
Assume that the tool is initially at rest and that to kl times the motor voltage V ; thus
the system has zero error. If a new position is T,=klV (10.2)
required, the appropriate input is applied, giving
rise to an error in position. The controller then The motor voltage V is effectively the
acts, attempting to reduce the error to zero, and, difference between two voltages V , and vb which
for a linear system, the motion of the tool will be are applied to the two motor terminals so that
described by a linear differential equation. I / = Va-Vb (10.3)
A human operator often forms part of a control
system. As an example of this consider the case of and, of course, if V , and V , were identical the
a man driving a car at a speed which he wishes to motor would have zero output torque. A
remain constant at 100 k d h . This constitutes a potentiometer-and-amplifier system produces the
speed-control system where the desired speed or voltage V , proportional to the desired angular
input, v i , is 100 k d h . The output, v, , is the actual position 6, of the aerial, the constant of
speed of the car, and the error, v,, is the proportionality being k 2 . Thus
difference between input and output. v, = k26i (10.4)
If, for example, the car meets a headwind, the
drop in speed (the error) will be noticed by the A position transducer, attached to the aerial
driver who, among other things, is acting as an whose angular position is 6, (the system output)
error-sensing device, and he will take corrective produces the voltage vb such that
action by adjusting the position of the accelerator Vb = k3 6, (10.5)
158 Introduction to automatic control

If 6, and 6, are equal, then the position error

defined by
6, = 6i - 6, (10.6)
is zero and for this condition it is required that the
voltage V and hence the torque T, be zero. The
voltage V , represents the desired position or
input, and the voltage V,, represents the actual
position or output. The voltage V thus represents 10.3 Block-diagram notation
the error, and we conclude that k2 must equal k3 It is common practice to represent control
and equations 10.3 to 10.6 can be combined to systems in block-diagram form. There are three
give basic elements: an adderlsubtracter, a multiplier,
V = k20e (10.7) and a pick-off point as shown in Figs 10.2, 10.3,
and 10.4.
Equation 10.2 can be written as
T, = kl k2 6, (10.8)
and we see that the motor torque is proportional
to the error.
Equation 10.8 represents the control action of
the system. In order to determine the motion of
the system for a particular input 6 , , we need to
incorporate the dynamics of the aerial itself. (In
mechanical control systems, the object whose
position or speed is being controlled is usually
referred to as the load.)
If the aerial has negligible damping, the only
torque applied to it is that from the motor; thus
T , = Id26,1dt2 (10.9a)
or T,,, =ID28, (10.9b)
where D is the operator ddt.
Eliminating T,,, from equations 10.8 and
10.9(b) ,
k l k2 6, = I D 2 6, (10.10) In Fig. 10.3, the simplest form of the multiplier
For any control system, the relationship E will be a constant, and the most complicated
between input and output is of major importance. form can always be reduced to a ratio of two
From equations 10.6 and 10.10, polynomials in operator D. We can write
kl kZ(0i- 0,) = I D 2 6, ,0
-=E
(ID2+ kl k2) 6, = kl k, ei ( 10.1la) e,
or IO, + kl k2 0, = kl k2 0, (10.1lb) and E is called the transfer operator between 6,
and 6,. Note that if 6, = E61 and 6, = Fez, as
By solving equations 10.11, we can find the shown in Figs 10.5(a) and (b), then 6, = E F & , as
output 0, as a function of time for a given function shown in Fig. 10.5(c).
01.
Note that a purely mechanical analogue of this
system could consist of a flywheel of moment of
inertia I connected to a shaft of torsional stiffness
K = kl k 2 , as shown in Fig. 10.1.

Equations 10.2, 10.6, 10.7, and 10.9(b) are

 10.4 System response 159

represented by the block-diagram elements

shown in Fig. 10.6. Note that there is an
implication of cause and effect: the output of a
block-diagram element is the result of applying
the input(s). In equation 10.9(b), the angular
rotation e, is the result of applying the torque T, .
The equation is thus rewritten as T,(1D2) = e,,
so that Fig. 10.6(d) can be drawn with T, as where w, = ( k , k2/1)”*,as shown in Fig. 10.10.
input.

-
Rather than taking up the required position
e, = ao7 the load oscillates about this position
with circular frequency w,. This performance is
Noting that the output of Fig. 10.6(d) is one of clearly unsatisfactory and it is evident that some
the inputs to Fig. 10.6(b), and connecting the four form of damping must be introduced. The
elements in the appropriate order, we obtain the response would then take the form of either Fig.
system block diagram shown in Fig. 10.7. From 10.11(a) or (b), depending on the amount of
this figure we note that a control system is a damping.
closed-loop system. One of the variables (0,) is
subtracted from a variable (0,) which precedes it;
this is known as negative feedback.

-
One way of providing damping is to attach a
Using the techniques of Fig. 10.5, Fig. 10.7 can damper to the load. If the damper provides a
be reduced to Fig. 10.8. torque Td which opposes the motion of the load
and is proportional to the velocity (viscous
10.4 System response damping), the constant of proportionality being
Returning to equations 10.11 we can determine
7
C , then equation 10.9(a) is replaced by
the response of the system to particular inputs 0,.
Suppose we want the load suddenly to rotate T, - Td = ld20,dt2 (10.13)
through an angle a. at time t = 0 . This T, - CDO, = ID28, (10.14)
corresponds to the step input 0, = 0, t < O ; 0, = ao,
t 2 O shown in Fig. 10.9. It is left as an exercise for +
T = (ID2 CD)O,
the reader to show that the response to this input T,l(ZD2 + CD) = 8,
is given by
The block diagram for the damped load is
e, = a o ( i -coSw,t) (10.12) shown in Fig. 10.12. We note that the effect of
160 Introduction to automatic control

tachogenerator is proportional to its angular

velocity, so that
Vt = k4D0, (10.18)
and the block-diagram form is shown in
adding the damper is to replace I D 2 in the Fig. lo.14.
undamped system by I D 2 + CD. Hence, for the
damped system of [Link].13 (cf. equation
10.11(a) ) 7

Consider the case of the undamped load with a

tachogenerator attached. The tachogenerator is a
relatively small device and applies a negligible
torque to the load so that equations 10.9 are
applicable. Assume that the voltage V, is
+
(ID2 C D + K)Oo = Kei (10.15) subtracted from the voltage V by an operational-
amplifier system so that the voltage V, applied to
where K = kl k2. Dividing by I to obtain the
the motor is
standard form (see equation 9.22) we have
(D*+ 2 c U , , ~+ w,,2)eo = w,,%i (10.16)
v, = v- v, (10.19)
The system block diagram for this case is shown in
where w: = KII and 5 = C/2d(KI).
Fig. 10.15 and we observe that the tachogenerator
For the Same input7 Fig. lo.9>the so1utions to
appears in an inner loop. The equations for the
equation 10.16 are (see also equation 9.33)
component parts of the system are listed below.
e,

[<I
= ao{i-e-~wnr[cosUdt
+
(l/d/l - 12)sin wd t ] }
e, = ei- e, (10.6)
V=k28, (10.7)
= ao{l-e-cwnr[l+wnt]} [=1 (10.17)
= a. { 1- e- cwnr[cosh w e t
v, = v- v, (10.19)
+([/d12-
l)sinhw,t]} [>1 T , = kl V,
wherecod=q,dl-[2andw,=wnd12-l. T, = ID28, (10.9b)
The output 0, does not settle to the required Vt = k4 D 0, (10.18)
value of a. until (theoretically) an infinite time
Eliminating T,, V , , V, V, and 0, we obtain
7
has elapsed. In practice, small amounts of
Coulomb (dry) friction ensure that motion ceases ID28, = kl [k2(@- 0,) - k4D8,]
reasonably quickly.
(ID2+klk4D+klk2)8, = klk2Bi (10.20)
The viscous damper wastes power and cannot
readily be constructed to give a precise amount of The amount of damping in the system can be
damping. There are other methods of introducing altered by regulating the techogenerator voltage
the first-derivative term (CDO,) into the system by a potentiometer circuit. This method of
equation 10.15, and one of these makes use of a introducing damping is known as output velocity
d.c. device known as a tachogenerator, driven by feedback. Another common way of introducing
the load. The voltage Vt produced by the damping is to use proportional-plus-derivative
action (see problem 10.5).

Figure 10.15
 10.5 System errors 161

10.5 System errors 0, = 0. What would be the steady-state error

A system equation relates one of the loop following the application of a constant external
variables to the input(s). It is conventional to torque To to the load?
have the loop variable on the left-hand side of the Equation 10.14 is replaced by
equation and the input(s) on the right. For
example, in Fig. 10.13, e,, V , T, and 0, are loop T, - CDO, + To = ID2O0 (10.24)
variables and 0, is the input; equation 10.15 is an or T, + To = (ID2+ CD)Oo (10.25)
output-input system equation. To obtain the
We could equally well have written -To, since
error-input system equation we can replace 0, in
the direction was unspecified. Putting
this equation by 0,- 0, from equation 10.6, to
T = T,+ To and T/(ZD2+CD) = eo, we can
obtain
draw the system block diagram (Fig. 10.16).
+ +
(ID2 CD K ) ( @- ee)= KO, Notice that the external torque To appears as an
+
(ZD2+ C D K)ee = (ID2+ CD)@ extra input to the system. Combining equations
(10.21) 10.25, 10.8, and 10.6 and putting 0, = 0, we have
If 0, has the constant value a. as shown in k,k20e+ To = ( I D 2 + CD)(-ee)
Fig. 10.9 then all its derivatives are zero and, for
this input, equation 10.21 becomes, for t > O ,
+ +
(ID2 C D kl k2)Oe = - To (10.26)
This system equation is identical in form with
+
(ID2 C D + K ) e e = 0 (1o-22) equation 10.15 with 0, and KOi replaced by 0, and
We already have the solution for eo, equations -TO respectively and SO the solutions can be
10.17. Subtracting these functions of 0, from 0, we written down immediately from equation 10.17.
obtain The steady-state error can be obtained by letting
t-+ 00 and is found to be
e, = croe-conr {coswdt
+ [</V(I - <2)]sinwdt} l<1
= aoe- wnf

r
+ [l/V(J2
{ 1+ w , t }
= age- l w n{coshw,t
l=1

- l ) ] sinhwet}l>1 I (10.23)

where w 2 = KII and 5 = i C / V ( Z K ) . Each of the

above three equations contains the negative
exponential term e-5wnrso that, as t+ 00, ee+O
[eel,-- = [eels, = -To/(klkz)
which is independent of the amount of damping.
(Note that for zero damping the system oscillates
indefinitely with a mean error value of -To/
(kl k2 1).
The complete solution of equation 10.26
consists of (a) the complementary function, which
is the transient part of the solution and dies away
and we say that the final or steady-state error is with time, provided some positive damping is
zero and write present, and (b) the particular integral or
steady-state solution which remains after the
[eelr-m = [eelss = 0 transients have died away. For a constant forcing
We do not need to solve equation 10.22 to find function, the steady-state solution must be a
the steady-state value of 0, since this is merely the constant function.
particular-integral part of the solution, which is Equation 10.26 describes the system for all time
clearly zero. That the steady-state error is not from 0 to 00. In the steady-state, therefore,
always zero can be seen from the following Dee = D20e = 0 and equation 10.26 becomes
example.
Consider the position-control system with k ~ k z [ f ) ~ ] ,-To
,=
viscously damped load which has already been and the steady-state error is
described. Assume that the system is at rest with
[Oelss =-To/(~I~z)
162 introduction to automatic control

Consider once again Fig. 10.13. Assume that

the system is initially at rest then, at time t = 0, it
is required that the load have a constant angular
velocity Ri. The desired position or input is
therefore in Fig. 10.19.
e, = o , t<o
The error-input equation for this system can be
written down directly from equation 10.27, with
and Oj=Rit, t r O K = kl k2 replaced by kl (kZ+ k5/D). Thus
as shown in Fig. 10.17. This is known as a ramp + +
[ID2 C D kl (k2 + ks/D)]ee = CRj
input.
To convert this to a purely differential equation
we simply differentiate with respect to time by
multiplying by D:
+ +
[ID3 CD2 klk2D + kl k5]ee
= DCRi = 0 (10.29)
since CRj is constant. The steady-state error is the
The error-input equation for this particular particular integral of the above equation so that,
input is, from equation 10.21, for the ramp input,
(ID2+ C D + K ) e e [ee 1s=0
= (ID2+ C D ) R i t = CRi (10.27)
The steady-state error is equal to C Q / K and 10.6 Stability of control systems
the error response will be of the same form as The introduction of integral action in the above
equations 10.17. Since 0, = 0, - e,, the output example had the effect of removing the steady-
response can be obtained by subtracting the error state error to a ramp input. It also had the effect
response from the input function. The result is of raising the order of the system. The order is
illustrated in Figs 10.18(a) and (b). defined as the highest power of D on the left-hand
side of a system equation, and in the example it
was raised from two to three.
For any particular control system, the system-
equation loop variable, whichever one is chosen,
will be preceded by the same polynomial in
operator D (see problem 10.2). Thus the
complementary functions (transient responses)
for the loop variables will have different initial
A control system with a residual error is conditions but will otherwise be of the same form.
normally unsatisfactory. Certain steady-state Before the concept of integral action was
errors can be overcome by using a controller introduced in the previous section, all the system
which incorporates integral action. Suppose that, equations were of order two; that is, they were of
in the above example, the voltage V , applied to the form
the motor, instead of being directly proportional [a2D2+alD+ao]x =f(D)y (10.30)
to the error e,, is given by
The transient response, and thus the stability of
v, = k2ee+k,/'eedt (10.28) such a system, depends only on the coefficients
0 ao, a , , and a2. Assuming that ao>O, then,
provided that a,>O and a 2 > 0 , the com-
In D-operator form this is written plementary function will not contain any positive
time exponentials and the system will be stable. If
v,,,= ( k 2 + 3 e e a, = 0 (zero damping) the complementary func-
tion will oscillate indefinitely with constant
and so the block diagram representation of this amplitude and, although not strictly unstable, this
proportional-plus-integral controller is as shown represents unsatisfactory control. Such a system is
 10.6 Stability of control systems 163

described as marginally stable. If either al<O Consider now the array

(negative damping) or a2 <0 (negative mass), the
transient will contain positive exponentials and
the system will be unstable. Figure 10.20
illustrates the various types of stability.

iii) al a0 0 >O
a1 a2 > aOa3 (10.32)
We give below, without proof, the Routh- a3 a2 a1
Hurwitz conditions for stability of a control
(10.46)
164 Introductionto automatic control

x = ReAeJ""' (10.47)
where 0, is the natural circular frequency of the
oscillation and A is the real amplitude.
We will use as an illustration the third-order
system described by equation 10.31. With the
right-hand side set equal to zero we have, for the
complementary function,
[a3D3+ a2D2al D +%]A e'"nt = 0 (10.48)
where we assume that a. , al ,a2 and a3 are all real
and positive.
Now
DA e i o n t = j w ei"d
~ (10.49)
ei0.t = (j w ) 2 ei"d
D ~ A ~ (10.50)
and it follows that
D'Ael"n'= (jwn)'A e i o n f (10.51)
where r is any integer. Y

Equation 10.48 becomes to be fed info the system the force must have a
[a3(jwd3 + a2 (iwJ2 component which is in phase with the velocity
(i.e. the imaginary part of the force must be
+al(jw,)+ao]Ae'""'= 0 (10.52) positive). It follows that if energy is required to
so that keep the system oscillating then the system must
a3 O'wrJ3 + a2 (jwn12 be stable. So we see that
+al(jw,)+ao = 0 (10.53) a1 W > U 3 W 3

sinceAei""#O. or a1>a3w2.
Hence -a3 jw; - a2wn2 al jw, + + a. = 0 Now since w is close to w, we can write
or (-a2wn2 + ao)+ 0, (-a3wn2 + al )j = o a0 = U2W,2 = a 2 ( w + E)'

The real and imaginary parts must separately be where E is a small quantity. So as E+O then
zero, hence w2+ ao/a2.
Hence for a stable system
W,2=- a0 =-.
a1
(10.54) a1 >a3 (ao/a21
a2 03

We conclude that if ala2 = ~ 0 then ~ 3 the

or ala2>a3ao
third-order system will be marginally stable and Note that as previously mentioned all the a's must
will oscillate at a circular frequency 0, given by be positive because if any one a is negative the
equation 10.54. We learned above (inequality output will diverge for zero input.
10.32) that if a1a2>aoa3the system will be stable.
It is clear that if this inequality is reversed the 10.7 Frequency response methods
system will be unstable. An assessment of the behaviour of a closed-loop
A physical reason why this inequality deter- control system can be made from an examination
mines the stability of the system can be found by of the frequency response of the open-loop
considering a small applied sinusoidal forcing system. Graphical methods are often employed in
term, Fe' "', where F is a complex force amplitude this work.
and w is close to w, . The main reasons for using open-loop system
The Argand diagram without the forcing term response methods are
is as shown in Fig. 10.21(a) and that with the (a) the overall open-loop system response can
forcing term is shown in Fig. 10.21(b). For energy be built up quickly using standard response curves
 10.7 Frequency response methods 165

of the component parts of the system; In frequency response methods we are only
(b) in practice most open-loop systems are concerned with the steady-state oscillations
stable which is an advantage if experimental (particular integral) part of the solution and so we
techniques are used! ignore the transient (complementary function)
Consider again a simple position-control sys- response. Since the system is linear, the particular
tem with proportional control driving an inertia integral solution of equation 10.58 must be
load with viscous damping. The block diagram for sinusoidal and at the same frequency w as the
the closed-loop system is shown in Fig. 10.22 input. The steady-state solution is therefore of the
which corresponds to Fig. 10.13 with K = k l k 2 . form
0, = Be I w'.

Substituting for 0, in equation 10.58,

D (1 + r D ) B e I"' = KOAe I"' (10.59)
or, from equation 10.51
jw(l+rjw)BeJ"'= K,AeJw' (10.60)
The forward-pafh Operator (3 (Dl is given by We see that, for sinusoidal inputs of frequency w
00 K to the open loop system, the ratio of output to
G ( D )= - = (10.55) input is
0, ID2 CD +
0, - B el
-
"'
If we disconnect the feedback loop we have the - -- - KO = G(jw) (10.61)
open-loop system of Fig. 10.23 and it can be seen 8 AeJw' jw(l+rjw)
that G ( D ) is ako the open-loop transfer Operator- which corresponds to the transfer operator G ( D )
Here ei is simp1y the input to the open-loop of equation 10.56 with D replaced by j w . G ( j w ) is
system. known as the open-loop transfer function.
We turn our attention now to the closed-loop
system with unity feedback. A unity feedback
system is, by definition, one for which the error
0, = 0, - 0, and therefore, since
e, = G(D) e, (10.62)
eliminating 0, we obtain
[l+G(D)]Oo = G ( D ) e l . (10.63)
(For a system with non-unity feedback see
We can write G ( D ) in standard form as example 10.7.)
KO Suppose now, as was discussed at the end of
G ( D )= (10.56) section 10.6, that the closed-loop system is
D(l+rD) marginally stable, i.e. it oscillates continuously at
where K, = KIC and r = IIC. (Note that r h a s the frequency w, say, for no input. In this case the
dimensions of time.) So, for the open-loop particular integral part of the solution is zero, but
the complementary function, or 'transient' part is
D(i+a)eo==,e (10*57) sinusoidal
We wish to consider the frequency response of 0, = Ce Jwn'.
the open-loop so the input must be sinusoidal and
we can write Substituting into equation 10.63, with the
0 = A e J wr right-hand side set equal to zero we have, for the
complementary function
where A is complex. Equation 10.57 now [l+G(jw,)]CeJ"n'= 0
becomes
therefore
D (1 + r D ) 0, = KOAelw' (10.58)
166 Introduction to automatic control

1 + G(jw,) =0 Therefore
since Ce i o n r 0. + G(jw) =
KO
(10.69)
What we have shown is that, for marginal W d [ l + (OT)’]
stability of the closed-loop system,
A particular real, positive value of w , say wl, can
G(jw,) = -1 (10.64) be found such that I G ( j w ) I = 1 which is
In other words, if there exists a particular value of 1
w (Le. w = w,) which makes the open-loop ~1 =-
T
d[-4 + V($+ KO’ 2)].
transfer function, G ( j w ) , have the (real) value of
- 1, then the closed-loop system will be marginally The phase angle between input and output is
stable and will oscillate continuously at the
frequency w = 0,. Thus we can see that the
open-loop transfer function G ( j w ) can give us
information about the closed-loop performance.
C#J = argG(jw)

(;j)
= argKo+arg 7 +arg
)( :C
1-jTw )
-
(l+’Tjw)

Returning to our example of the open-loop = argKo+arg - +arg

transfer function 10.61 +
1 (WT)’
7T
KO = 0-- -arctan(w~). (10.70)
G ( j w ) = jw (1 + Tjw) 2

we can check if a value of w can be found which Is there a value of w which gives C#J the value -r?
makes G ( j w ) = - 1. In other words, does a value There is only one, which is when w is infinitely
of w exist which simultaneously makes the large. There is therefore no value of w which
amp1itude ratio have unity va1ue and the phase makes G ( j w ) = -1 which shows that the closed-
angle have the value of - 180” or -7~ radians? loop system of Fig. 10.22 can never be marginally
The amplitude ratio is stable. This is a result we already knew, since the
second-order system is always stable.

1=: Ic(Jw)l= ljw(l+Tjw)

KO
1 The polar, or Argand, diagram of the
open-loop frequency response is known as a
Nyquist diagram (after H. Nyquist’s work in the
early 1930s). A sketch of the Nyquist diagram for
1 the transfer function of equation 10.61 is shown in
(10.65)
= l K o 1 1 ; l l ~ l Fig. 10.24 where the arrow shows the direction of
KOis real so that increasing frequency. It can be seen that G ( j w )
never has the critical value of - 1. The plotting of
Nyquist diagrams and a logarithmic form of
1K0~=Ko. (1o.66) frequency response are discussed later in this
chapter.
Further

and
l i l =-l-l1 1
w i = t / - j / =;
1
(10.67)

IG1 1 1 1
1 1 - Tjw
= 1(1+ Tjw)(1- Tjw) = 2;:(+-:

-
-
I 1 - Tjw 1 - d [ 1 + (wT)’]
1 + (137)’ - 1+ (WT)*
1
- (10.68)
d [ 1 -k (WT)’]
 10.7 Frequency response methods 167

Assume now that the proportional controller of Sketches of the Nyquist diagrams for the
the above example is replaced by a controller with open-loop response of the above system with
proportional-plus-integral action. The open-loop proportional-plus-integral are shown in Fig.
transfer operator of equation 10.55 is thus 10.25. It will be observed that the plot for T = T~
replaced by passes through the critical point G (jw) = - 1. The
frequency response of equation 10.73 is discussed
G ( D )=
+
K Ki/D
(10.71) again in example 10.5.
ZD2 + C D
which we can write in the form
KO(1 + 7oD) (10.72)
G ( D )=
D~(I+~D)
where KO= Ki/C, T~ = KIKi and 7 = IIC.
For sinusoidal inputs of circular frequency w we
can replace D by j w as before to obtain the
open-loop transfer function G(jo) = K + -Ki
1W
KO(1 + 70jw) (10.73) I( jo)’+ Cjo
G(jw) =
(jw)’ (1 + 7jw) - +
KO( 1 ?do)
-
It is easy to show (see example 10.5) that the
amplitude ratio is Figure 10.25
+
(jw)*( 1 Tjw)

Substituting for G ( D )from equation 10.72 into

~ o d / [ 1 (w70>21
+
v[
I G (jw) 1 = w2 1+ (w7)21 (10.74) the closed-loop inputloutput system equation
10.63 we obtain
and the phase angle is [7D3+ D 2 + K07, D + K O ]0,
4 = argG(jw) = arctan(w7,) = [KOTOD+&]0i
-7r- arctan(w7) (10.75)
If the system is marginally stable. we have an
To check for marginal stability of the closed- equation of the form 10.48 for the complementary
loop system we look for the possibility of a value function, where a. = K O , al = K o ~ oa2
, = 1 and
of w which simultaneously gives 1 G(jw) I the a3 = 7. From equation 10.54, we find
value of unity and 4 the value of -7r. We note
that, from equation 10.75,4 will have the value of KO---K07o
_
- - 7 ~ if T = T~ for any value of w . If IG(jw)I = 1 1 7
then, from equation 10.74 with 7 = T~ or T ~ =
7
Kolw2 = 1
which confirms the result found from considera-
so, from equation 10.72 tion of the open-loop frequency response.
w2 = KO= Ki/C = KII,
Bode diagrams
We have shown that, for the open-loop frequency The overall open-loop amplitude ratio is the
response, the amplitude ratio will be unity and product of the amplitude ratios of the component
simultaneously the phase lag will be - 7 ~radians at parts, and the overall phase angle is the sum of
the excitation frequency w = d ( K / I ) provided the phase angles of the component parts (see, for
that Ki/C = KII or CK = IKi. This is equivalent examples, equations 10.65 and 10.70).
to saying that the closed-loop system will be When graphical techniques are employed it is
marginally stable provided that convenient to plot the logarithm of the amplitude
ratio I GI (logarithms to the base 10 are always
CK = IKi used). Traditionally, although not essentially, the
and, if this is the case, the system will oscillate logarithm of the amplitude is multiplied by 20 to
continuously at the frequency d ( K / I ) . give the ratio in the form of decibels (dB). When
168 introduction to automatic control

the frequency response information is presented

in the form of two graphs, one log I G ( j w ) 1 or
20 log I G ( j w ) I plotted against logw and the other
the phase angle 4 plotted against logw, the graphs
are known as Bode diagrams after H. W. Bode
who presented his work in the 1940s.
It is useful to build up a number of standard
Bode diagrams of simple functions since know-
ledge of these enables (a) the rapid sketching of
the overall frequency response plots and (b) the
reduction of experimental results into component
parts to assist with analysis. Below we use the
notation E ( D ) for the transfer operator of a Figure ,o.27 '
component part and G (D) for the transfer
operator of the OPen-IooP. E ( b J )and G (io)are +
(iii) E3(D) = 1/(1 7 D ) ,the first order lag
the corresponding frequency response transfer
functions. The amplitude ratio I E3 ( j w ) I = [ 1+ ( w ~ ) ~ ] - ~ ' ~
from equation 10.68 and the phase angle is
(i) E , ( D ) = K , a constant 4 = arg (1 + 7 j w ) - l = -arctan (07) from equation
The amplitude ratio IEl(iw)I of the frequency 10.70. With regard to the overall shape of the
response is simply K and the phase angle 4 is zero Bode diagrams for this function we note that at
at all frequencies so that the Bode diagrams for low frequencies (small w ) , I E3 ( j w ) I + 1 and
this function are as shown in Fig. 10.26. 4-0 whereas at high frequencies (large w ) ,
lE3(jw)1+(w7)-' and 4 + - ~ / 2 . So at low
frequencies log I E ( j w ) I+ -logo- log7 or
-logw+log(1/7) which is a straight line of slope
-1 on the graph of logIE(jo)/ plotted against
logw (or if decibels are used the slope is -20
dB/decade or -6 db/octave).
The log amplitude ratio and phase graphs are
each therefore asymptotic to straight lines at
both low and high frequencies. At the particular
frequency w = 1/7, known as the break
or corner frequency, I E , ( j w ) I = 2-"* and
loglE,(jw)l= -0.1505 (and 2010gIE3(jw)1=-3
dB) also 4 = arctan(-1) = -7d4 radians or -45".
The Bode diagrams for this function are shown in
(ii) E,(D) = 1/D, the integral operator Fig. 10.28.
The amplitude ratio 1 E 2 ( j w )I = l / o from
equation 10.67. The log of the amplitude ratio
is log(l/w) = -logw (or, in decibel form,
-20 log w ) . The phase angle 4 = arg ( - j / w ) = -~ / 2
at all values of the frequency w . The Bode
diagrams for this function are shown in Fig. 10.27.
Each graph is a straight line with the log(ampli-
tude ratio) graph having a slope of -1. (If
decibels are used for this graph the gradient is
-20. A tenfold increase in frequency is known as
a decade and log10 = 1 so that this slope is often
described as -20dB/decade. A doubling in
frequency is known as an octave and
log2 = 0.3010 so the slope can also be described
as -6 dB/octave since 20 x 0.3010 = 6.02 = 6).
 10.7 Frequency response methods 169

As an aid to sketching the phase graph of the logw axis for any other value of [Link] same
Fig. 10.28 it should be noted that the gradient of applies to the phase graph c$ = arg[E3(jw)]. The
the graph at the break frequency w = 1/r is proof of this phenomenon is left to the reader.
-(ln10)/2 or -1.151 radians per decade =
-66"/decade (the proof of this is left for problem +
(iv) E,(D) = 1 TO,the first order lead
10.25). It is left to the reader to show that the log
Accurate values of amplitude ratios in decibel amplitude ratios and the phase curves are those
form and phase angles for the function E3( j w ) are for the transfer operator E3(D) = (1 + TO)-'
listed in the table below and plotted to scale in rotated about the logw axis, as shown in
Fig. 10.29 where a logarithmic scale is used for Fig. 10.30.
the frequency axis. In practice log graph paper is
normally used for Bode diagrams.

Assessment of closed-loop stability

We know that if, at a particular value of
excitation frequency in the open-loop frequency
response, the amplitude ratio is unity and
simultaneously the phase angle is -180" (i.e.
G ( j w ) = -1) the loop, when closed by unity
feedback, will be marginally stable. The closeness
of the open-loop Nyquist plot to the critical point,
G ( j w ) = -1, is a measure of the closed-loop
stability.
Take, as an example, a control system with an
open-loop transfer function of the form
Figure 10.29 K
G ( D )= (10.76)
D (1 + 7 1 D ) (1 + 7 2 D)
w/(rads-') 20ioglE3(.b)l'dB +'degrees
The Nyquist diagrams of G ( j w ) for three
1/(107) -0.04 -5.71 particular values of K are sketched in Figs
1/(4~) -0.26 -14.04 10.31(a), (b) and (c).
1/(2 7) -0.97 -26.57 In Fig. 10.31(b) a value of K has been chosen
1/ r -3.01 -45.00 which makes the curve pass through the critical
2/r -6.99 -63.43
4/7 - 12.30 -75.96
1o/r -20.04 -84.29

Once scales for the graphs have been chosen,

all of the graphs of log I E3( j w ) I will have the same
shape, independent of the value of 7.A template
can be made of the curve which has been drawn
for a particular value of 7,the break point being
w = 1/7, then the template can be moved along
170 Introduction to automatic control

point (marginally stable closed-loop) while in Figs determining, from the open-loop response,
10.31(a) and (c) smaller and larger values of K whether or not the closed-loop system is stable
than that used in (b) are used respectively. and will suffice for all systems described in this
The closed-loop equation, from equation 10.63, chapter. There are a group of transfer functions
is where this rule does not apply, but they are
outside the scope of this book and for these the
reader is referred to more advanced texts on
D ) ( 1 5+- 2 0 ) ]e0
[I + D ( 1 + ~ l K frequency response methods).
The closeness of the open-loop frequency
-
K response curve to the critical point is an indication
D (1 + T~ D ) ( 1 + 72 D ) e, of the performance of the closed-loop system and
a measure of the closeness can be obtained from
hence the phase margin +,, and the gain margin g ,
[7172D3 + ( 7 1 + T ~ ) DD~++K ] 0, = KO, which are defined in the open-loop Nyquist
diagram of Fig. 10.32.
The phase margin is given by $,, = 180"+ 4
when the amplitude ratio is unity. It is the angle
between the negative real axis and the G ( j w )
vector when I G ( j w ) I = 1.
The gain margin is the reciprocal of the
amplitude ratio when the phase angle 4 = -180".
If the gain K i n , for example, equation 10.76 is
multiplied by an amount equal to the gain margin,
the open-loop will then pass through the critical
point and the corresponding closed-loop will be
marginally stable.
Typical values for satisfactory closed-loop
performance are: 4, should not be less than
about 45" and g , should not be less than about 2
or 6 dB.

Comparing this with the standard form of

equation 10.31 and applying inequality 10.32 we
find, for a stable system,
K<(7* +72)/(5-172)
It follows that if K = ( T ~+ ~ ~ ) / ( 7 ~the7 2closed-
)
loop system will be marginally stable and if
K > (71 + 7 2 ) / ( ~T1 ~ the
) closed-loop system will be
unstable. Figure 10.31(b) corresponds, as men-
tioned previously, to a marginally stable system
while Figs 10.31(a) and (c) correspond to stable
and unstable closed-loop systems respectively.
If one considers walking along the curve of Fig.
10.31(a) (stable closed-loop) in the direction of The phase margin and gain margin can of
increasing frequency it will be observed that the course be found from Bode diagrams and these
critical point G ( j w ) = -1 falls to the left of the are illustrated in Fig. 10.33.
curve. Similarly for Fig. 10.31(c) (unstable
closed-loop) the critical point falls to the right.
This idea can be used as a rule of thumb for
Discussion examples
Example 10.1
Figure 10.34 shows a hydraulic relay with
feedback used in a control system. Oil under
pressure is supplied at P via the spool valve to the xla = x 2 / b (ii)
power ram and can exhaust to the drain at either and equations (i) and (ii) are represented by the
Q or R. The value contains a sliding sleeve and
block diagram of Fig. 10.36.
the displacements of the spool and sleeve

Eliminating x from equations (i) and (ii),

Dx2 = k(xl - ax2/b)A
(D + kaA/b)x2= kAxl
(bDl(kaA)+ 1)x2 = ( b / a ) x l
Figure 10.34 or (TD+ 1 ) x 2 = Kxl (iii)
where T = b/(kaA) and K = bla.
measured from the centralised position are x 1 and The transfer operator is
x respectively. Neglecting compressibility and
inertias, the volumetric flow rate q through the x2
--- - K
(iv)
valve can be taken to be proportional to the x1 T D + ~
effective valve opening; that is, which can be represented by the block diagram of
q = k(x1 - x ) Fin. 10.37.
The power ram, whose displacement is x 2 , has an
effective area A , and the sleeve is connected to
the ram by the slotted lever EFG, which pivots Figure 10.37
about F. The complementary function (c.f.) of equation
Draw a block diagram for the relay and show +
(iii) is the solution of (TD 1)xZ = 0, which is of
that the transfer operator between x2 and x 1 is of the form x2 = X 2 e m r ;hence
the form mX2emr+ =0
XZemr
---
x2- K and, dividing by X2emr,
x1 T D + ~
rn = -117
Find the values of the gain K and the first-order
The c.f. is therefore
time-constant T and determine the response of the
relay to a step change in x1 of magnitude XI. x2 = X2e-r'T
Solution The velocity v2 of the ram downwards The particular integral is
is equal to the product of the flow q and the area
A: ~2 = KX1
and so the complete solution is
~2 =D x=
~ qA = k(xl - x ) A 6)
From the geometry of the feedback link x2 = X 2 e p r ' T + K X I (VI

(Fig. 10.35), At t = 0, x2 = 0, so that

172 Introduction to automatic control

0 = X,(l)+KX, system which drives the power ram R whose

Xz = -KX1 vertically upward displacement is w. The velocity
I.i, of the ram is basically kl times the displacement
and substituting for X 2 in equation (v) gives y, but the pneumatic system introduces a
x2 = K X l (1 - e-r'T) first-order lag of time-constant T. The ram adjusts
valve V, and the flow qi into the tank from a
constant-head supply is k2 times the ram
displacement w. A dashpot of damping constant
C and a spring of stiffness S connect end N to the
ram and to ground respectively. The mass of all
the links can be neglected.
a) Draw a block diagram for the system and
show that it contains proportional-plus-integral
action.
The response is sketched in Fig. 10.38. The b) Show that there will be no steady-state error
transfer operator (iv) thus represents Propor- in level following a sudden change in the desired
tional control (the constant of proportionality level hi or demand qo.
being K ) with a first-order lag of time-constant T. C) Determine the necessary and sufficient
conditions for system stability.
Example 10.2
Figure 10.39 shows a level-control rig. The tank T Sofurion Considerink? sma1l disp1acements Of

has a horizontal cross-syfi-,:! area A , and liquid link GHJ, from similar triangles (Fig. 10.40),
flows into the tank at a rate qi and out at a rate qo.
The height h, of liquid in the tank is sensed by the
float F, which displaces end G of link GHJ, where
G H = HJ = l l . The desired height hi can be
adjusted by altering the vertical position of
end J. Link LMN is connected as shown; (hi-xyfl = (h,+x)/f,
LM = MN = f 2 . Point M, whose vertically down- x = ;(hi - ho ) = '2he (i)
ward displacement is y , is connected to the
where x is the downward displacement of H (and
flapper U of a pneumatic flapper-nozzle valve
L) and he is the error in level. Similarly, denoting
the upward displacement of N by z ,
y = t(x-z) (ii)
The motion of the ram R is given by
Dw = [kl/(l +T D ) ] ~ (iii)
The downward force acting on N due to the
spring is Sz, and the upward force due to the
dashpot is C ( w - 2 ) . Since the links are light,
the net force must be zero,
i.e. S z - C ( + - 2 ) = 0
+
(CD S ) z = CDw (iv)
The flow into the tank is
4, = k2W (VI
and the net inflow is equal to the area A times the
rate of change of height h,:
41 - 40 = ADho ( 4
Equations (i) to (vi) are represented in the
Figure 10.41

block diagram of Fig. 10.41. The transfer oper- which leads to

ator for the inner loop can be obtained by
eliminating y and z from equations (ii), (iii), and +
A (C T S ) (kl C2/2 - .S2) > k l k2C 3714 (ix)
(iv). The result is We note that this inequality cannot be satisfied
unless condition (viii) is satisfied.
W -
- -
Ik (CD + S)
x D(TD + l)(CD + S ) + Ik1 CD Example 10.3
In a simple angular-position control system the
Ik1 ( C + S/D)
- driving torque on the load is k times the error 0,.
TCD~+ (TS + c )D + (S + t k , C ) The load is a flywheel of moment of inertia I
which indicates proportional-plus-integral action whose motion is opposed by a dry friction torque
with a second-order lag. Eliminating x , w , and qi which can be assumed to have a constant
and replacing h, by hi - h e , we find magnitude Tf . Viscous damping is negligible.
Numerical values are k = 0.2 N d r a d ,
[ A d D 4+ A ( T S +C)D3+ A(S + bk1 C)D2 I = 1x lO-3 kg m2, and Tf = 0.015 N m.
+aklk2CD+4klk2S]he Initially the system is at rest with zero error and
=AD2[KD2+(7S+C)D then the double step input 0, shown in Fig. 10.42
+ (S + Ik1 C ) ]hi is applied. Find (a) the final position and (b) the
time taken for all motion to cease.
+ +
+ D[7CD2 (TS C)D
+ (S + Ikl C ) ]qo (vii)
Note that, for step changes in desired level hi or in
demand qo, the right-hand side of the system
equation (vii) is zero and therefore the steady-
state or particular-integral value of the error he is
zero.
The system equation is of the fourth-order form
Solution If we choose to denote 0, (and hence
(a4D4+a3D3+u2D2+alD+ao)he
= ... Oi and 0,) as positive in the anticlockwise
The first condition for stability is that all the sense, then the positive direction Of its deriva-
coefficients a. , al , a 2 , a 3 , u4 be positive, which is tives, bo and e,, is also anticlockwise. If the
satisfied. The next condition is flywheel happens to be rotating anticlockwise
(b,>O), then the friction torque Tf will be
u 1 a2 > (13% clockwise and vice versa (see Figs 10.43(a) and
( k1 kz C/4)(S + k 1 C/2)A (b) ).
>A(TS+C)(klk2S/4)
which reduces to
k l c2/2>
7S2 (viii)
The final condition for a fourth-order system is
al(a2a3)-u~a32-u12a4>0
174 Introduction to automatic control

The two equations of motion are overcome the static friction torque Tf. So,
whenever 6, = 0, motion will not continue unless
ke, - T, = le, (for 6, >0)
ke, + Tf = le, (for 6, <0)
(i) I I
0.2 I 0, >0.015 or 0, I >0.075. This corresponds
(ii) to the range AB, 0.9255 0,s 1.075, when 0, = 1
Replacing 0, by 0, - e,, we obtain and to the range CD, 1.425r8,11.575, when
0, = 1.5 (see Fig. 10.44).
+
le, ke, = kei - Tf (6, >0) Immediately after .the input 0, is applied,
le, + ke, = kei + Tf (6,<0) 0, = 0, = 0 and so the trajectory starts at point
Substituting numerical values and dividing by (a). The initial error is 0, - 0, = 1 - 0 = 1 and the
k = 0.2, initial driving torque ke, = (0.2)(1) = +0.2 N m,
causing the load velocity 6, to be positive. The
(5 x i0-’)eO + e,
=0.925 (6,>0)
+
(5 x io-’)e0 e,
= 1.075 (6,<0)
(5 x i0-’)eO + e,
I ei= 1 (iii)

(iv)
appropriate centre for the trajectory is 0, = 0.925
(point A). We could have arrived at the same
conclusion by noting that the trajectories always
follow a clockwise pattern so that, when motion
commences, the velocity 6, will be positive and
= 1.425 (bo>O)
+
(5 x 10-’)8, e,
= 1.575 (6,<0) I 0, = 1.5 (VI

(vi)
Each equation is a second-order linear differ-
entia1 equation with constant coefficients, no
the centre of the arc at point A.
The input 0, = 1 for 0.625 s. The angle swept by
the radii which generate the trajectory is equal to
o,t so that, while 0, = 1, the total angle of
rotation will be [1/(5 x lO-3)]’/2 x 0.625 radians or
506.4 degrees; that is, one revolution plus 146.4
(viscous) damping term e,, and a constant forcing degrees. After point (b), 6, becomes negative so
term. Thus the equations represent simple we shift the centre to point B and so on until point
harmonic motion and we know (section 9.8) that (c) is reached. 0, then takes on the value of 1.5 rad
if we draw a phase-plane plot of 6,/w, against 0, and so the appropriate centres become points C
the result will be a series of circular arcs with and D. The trajectory continues and then ceases
centres on the 0, axis corresponding to the at point (e), due to insufficient driving torque.
particular integral or ‘equilibrium’ position for The final position is 0, = 1.48 rad, and the
the relevant equation. Since the equations have trajectory makes 2.5 revolutions so that the
been arranged to have a unity coefficient for e,, total time is 2.5(27~/0,) = 2.5(27~)/(5X lO-3)-’/2
the right-hand sides of the equations are the = 0.89s.
particular integrals. It should be noted that this technique can be
Each time the velocity 6, = 0, we must check used for any input function 0, if it is approximated
whether there is sufficient driving torque k0, to by a series of small steps as shown in Fig. 10.45.

Figure 10.44
 torque TLso that
T,-T~~,-(~~,+T~)=o
To = ( C + Tb)W, + TL
= (0.2) + 0.35)200
+50= 160Nm
a) When the speed is varying, the equation of
motion for the load with the brake removed and
the governor disconnected (Le. 4 remaining
Example 10.4 constant at O O ) , allowing for the time lag 7 , is
A four-cylinder petrol engine is connected TD/(1+ TD)- CO, = IDw,
directly to a load. The total effective moment of
inertia is I and the damping constant is C. A To- Tb O, = (1+ 7D)(ID + C) W ,
governor increases the throttle angle 4 of the [ h D 2 + ( I + c T ) D + ( c + Tb)]W,= To
carburettor by N degrees for each rads of speed Substituting numerical values,
change. For steady-state operation over a
particular operating range, the torque TD driving [0.45D2+ 3.03D + 0.551o, = 160
the ‘rankshaft increases by Ta for each degree The final value of 0, is the particular integral of
increase in throttle angle and decreases by Tb for the above equation, which is
each rads increase in the engine speed 0,.
During speed variation, the carburettor intro- w, = 160/0.55 = 290.9 rads
duces a first-order lag of time-constant 7. b) When the governor is connected, it will
Numerica1 va1ues are I = 3 kg m2, c = o.2 N m operate in such a way as to increase the torque if
per rad/s, N = lo, Ta = 4 N m 7 Tb = 0-35 N m, the speed drops and vice versa. We can assume
and T = 0.15 s. that, before the load was removed, the error in
A brake app1ies a torque T L Of 50 N m to the speed o, was zero; i.e. wi,the desired speed, is
load which is being driven by the engine at a equal to 200 rads. The extra torque Tg provided
‘Onstant ’peed Of 200 rads. The brake is then by the governor is found from the two equations
removed. Assuming that the system is linear,
obtain the differential equation which describes 4 = No,
the engine speed (a) if the governor is Tg = Ta4
disconnected and (b) if the governor is connected.
Find the final speed for each case and draw a so that
block diagram of the system for case (b). Tg = NTaw,
and TD in part (a) is replaced by
TD= T o - T b o , + T g
Substituting q - w, for o,, the resulting
equation is
[ITD~+(Z+CT)D+(C+Tb+NTa)]wo
= To+ NTao,
[0.45D2+ 3.03D + 40.551 o, = 8160
Solution Let us measure the throttle angle 4 The SteadY-state value of 00is
from its original position when the engine speed is w, = 8160140.55 = 201.2 rads
steady at 200 rads. Figure 10.46 shows the engine
torque TD for 4 = 0 and steady-speed operation A block diagram of the control system is shown
and is of the form in Fig. 10.47.
TD = T o - TbW,
The engine torque TDfor this case is simply equal
to the damping torque Coo plus the braking
176 Introduction to automatic control

Figure 10.47
Example 10.5 Assuming next that TO<T (i.e. 1 / ~ 0 > 1 / ~the
)
Consider the control system with proportional- break points of Eb and Ed are interchanged
plus-integral action with the forward path transfer resulting in the Bode diagrams of Fig. 10.49.
operator Finally, if ro = 7, the numerator and denomina-
tor terms (1 + T ~ ~ oand ) (1 + +) in equation
K+Ki/D - Ko(l+~oD) 10.77 cancel so that the transfer function reduces
G ( D )= -
ID2 + CD D2(1 + TD) to
as in equations 10.71 and 10.2 and obtain sketches KO
for T ~ > T , T~ = T and T ~ < T of the overall G(jw) =-
open-loop frequency response in Bode and (jwI2
Nyquist form. Again letting KO= 1, the Bode diagrams for this
system are simply the straight line graphs shown
Solution The open-loop frequency response in Fig. 1o.5o.
transfer function For any positive value of KOnot equal to unity
KO (1 + TOjW) an additional component of log KO would be
G(jw) = (1o-77) added to the amplitude ratio plots, so that the
(jO)2(1+ T j U )
overall frequency response amplitude ratio curves
can be broken down into its individual compo- obtained above would simply be moved upwards
nents by an amount equal to logKO. If the above Bode
diagrams are re-plotted in Nyquist form, the
Ea(iW) = KO curves of Fig. 10.25 result.
+
E,, ( j w ) = 1 Toj w
&(io) = l/(jw)2
+
Ed (io)= 1/( 1 T j O ) . Example 10.6
The forward path transfer function of a control
The Bode diagrams for E,, Eb and Ed have been
system is given by
discussed above, see Figs 10.26, 10.30 and 10.28.
+
C(D 5)
E = -
c tw)k) -

so, combining two sets of Bode diagrams as in

G ( D )=
D (D+ 2)(D + 3)
From the open-loop frequency response deter-
mine
Fig. 10.27 the amplitude part for E c ( j w ) is a
straight line of gradient -2 passing through the a) the phase margin &, if C = 10,
origin and the phase angle is constant at -180". b) the value of C if &, = 45", and
The break frequencies for Eb and Ed are at c) the gain margin for each case.
w = l/To and w = 1 / respectively.
~
Solution We can rewrite the transfer operator in
Let us assume first that T ~ > T (Le. TO< UT) standard fom as
and initially, for simplicity, that KO = 1 and only
the straight line approximations are required on KOU+ 71D)
the amplitude ratio graph. The component parts G ( D )=
D(1+72D)(1+730)
of the transfer function are shown with the dashed
lines in Fig. 10.48 and the overall open-loop where KO= 5C/6, r1 = 1/5, T~ = 1/2 and T~ = 1/3.
response with the full lines. The Bode diagrams for each of the five
Figure 10.48

Figure 10.49
178 Introduction to automatic control

The amplitude ratio is seen to be unity

somewhere between w = 3 and w = 4 rads. By
trial and error we find that, at w = 3.34 rads,
IG ( j w ) I = 1.O001and the phase angle is -163.8".
The phase margin is therefore
& = 180- 163.8 = 16.2".
(This would be too small in a practical system; the
closed loop response would be too oscillatory).
b) If the phase margin c&, is to be 45" we need
to find the value of w at which
4 = - 180"+ 45" = - 135". From the above table
we see this occurs somewhere between w = 1 and
w = 2 rads. By trial and error we find that at
w = 1.427 rads the phase angle C$ = -135.02".

'I
Using the value of C = 10 as in a) above we find
that the corresponding value of G ( j w ) 1 is 4.464.
For a phase margin of 45" this value of I G ( j w ) 1
-
component parts of C ( j w ) can be drawn, and the should be unity so C needs to be reduced by a
overall open-loop frequency response can be factor of 4.464. This gives C = 1014.464 = 2.24.
obtained by combining the components as in
example 10.5. The phase margin and the gain Example 10.7
margin can then be determined as in Fig. 10.33. Figure 10.51(a) shows the block diagram for a
This method is left as an exercise to the reader. system where the feedback is operated on by a
Alternatively we can work directly from the transfer operator H ( D ) .
overall amplitude ratio and phase functions, as
follows.
The overall open-loop frequency response
function is
KO (1 + T l b )
G(jw) =
j w (1 +
72jw)( 1 T+) +
The amplitude ratio is Determine (a) the closed-loop transfer oper-
ator and (b) equivalent unity feedback system.
Kod[1 + (T1w)21 410.78) Solution
' G ( j w ) I =0d1+(T20)2d[1+(T3W)
a) The difference between the desired input and
and the phase angle is the signal which is now fed back is
4 = arctan(T1w)- ~/2-arctan(72w) e: = e, - H ( D ) e,
- arctan (T~ w ) (10.79) therefore the output
a) To find the phase margin we need to establish
the value of w at which 1 G ( j w ) I = 1. Substituting e, e, - G ( D ) H ( D )e,
= G ( D )e: = G (D)
the numerical values into equations 10.78 and thus
10.79 and by trying a few values of w , we obtain

;:
the following table e, =
G ( D )ei
1+ G ( D ) H ( D )
w/(rad s-')
IW4l ddwrees so that the transfer operator is given by
-123.7
146.9 !%= G ( D )
1.27 -1 60.3 e, 1+ G ( D ) H ( D )
0.72 - 167.3
b) Figure 10.51(b) shows the equivalent unity
 Problems 179

feedback system such that the closed-loop [ZDZ+(C+K2)D+K1K3]0,= KIK30i

transfer operator is the same as (a) 10.4 A voltage V is produced which is K1 times the
error in a position-control system. The load is a
i.e. G(D) - G'(D) flywheel of moment of inertia I , and the damping
+
1 G(D)H(D) 1 G'(D) + torque at the load is equal to C times the angular
velocity of the load. The moment of inertia of the rotor
or G + G G ' = G ' + G H G ' of the motor which drives the load is I, and the torque
G = G'(l+ GH-G) developed between the rotor and the stator is given by
G T, = K2 V. Obtain the system equation for the output
finally G' = 0, and also determine the damping factor for each of
(l+GH-G) the following cases: (a) the motor is directly connected
to the load; (b) as (a) with an external torque QL
Note that the open-loop transfer 'perator for (a) applied to the load; (c) a gearbox is placed between the
is GH and if G ( j w ) H (j w ) equals -1 then the motor and the load such that OM = n q .
closed-loop system is marginally stable. Also for
10.5 The amount of damping in a position-control
G ' (io) equa1s -'
(b) the open-loop transfer operator is Gt and if
then this system is margina11y system is increased by using proportional-plus-
&rjvatjve action such that the driving tqrque TD
stab1e. A little a1gebra 'On shows that applied to the load is given by TD = kl 0, + k2 e, where
G ' ( j w ) = -1 implies that G ( j w ) H ( b )= -1, So e, is the error. The moment of inertia of the load is I
that all the analysis carried out in this chapter and the viscous damping constant is C. If the damping
considering unity feedback applies equally well to ratio of the system is 4, show that k,Z = (C+ k2)'.
systems with non-unity feedback Operators using 10.6 Figure 10.54 shows a hydraulic power ram B fed
the appropriate open-loop transfer operator. by a spool valve V. It can be assumed that the ram
velocity is proportional to the spool displacement
Problems measured from the centralised position, the constant of
proportionality being k. The slotted link PQR is
10.1 For the system represented by the block diagram connected to the spool and ram as shown.
of Fig. 10.52, obtain the transfer operator for x l y .

10.2 For the control system of Fig. 10.53, obtain the

system equation for each of the loop variables.

Show that the transfer operator for the arrangement

is of the form x l y = A / ( l + TD) and write down ex-
pressions for the gain A and first-order time-constant T .
10.3 A motor used in a position-control system has its 10.7 The hydraulic relay of problem 10.6 is modified
input voltage V , , its output torque T , , and its angular by the addition of a spring of stiffness S and a damper
velocity omrelated by the equation of damping constant C, as shown in Fig. 10.55. Show
that the modified arrangement gives proportional-plus-
T, = K1 V , - K~w,
integral action with a first-order lag of time-constant T
The motor is connected directly to a load of moment of by obtaining the transfer operator in the form
inertia I whose motion is opposed by a viscous damping
torque equal to C times the angular velocity of the load. x (1 + ~ / ( T ~ D ) )
-=A
If the motor voltage V , is given by K3e,, where 0, is Y (1+ TD)
the position error, show that the output-input system
equation is
180 Introduction to automatic control

10.14 See problem 9.21. In a position-control system,

the driving torque on the load is 0.2 N d r a d of error.
The load is a flywheel of moment of inertia 5 x lO-4
kg m2 whose motion is opposed by a dry-friction torque
such that the torque required to initiate motion is
0.022Nm but once motion has started the resisting
torque is 0.015 N m. Viscous damping is negligible.
Initially the system is at rest and then a step input of 1
radian is applied. Find (a) the time taken for all motion
to cease and (b) the steady-state error.
10.15 In an angular-position control system the load
consists of a flywheel of moment of inertia IL and the
driving torque is K times the position error. Damping
of the load is brought about by a viscous Lanchester
damper in the form of a second flywheel of moment of
Evaluate the constant A , the lag constant T, and the inertia I D mounted coaxially with the first and con-
integral time constant T ~ . nected to it by a viscous damper. The torque transmit-
ted through the damper is C times the relative angular
10.8 Consider the level-control system of example
velocity of the flywheels.
10.2 with the spring removed and the dashpot replaced
by a rigid link. The system is steady, supplying a a) Show that the system is stable.
constant demand Q,. Show that if the demand is
increased by lo%, the level drops by 0.2 Q o / k 2 . b) Determine the steady-state errors following inputs
of the form (i) A u ( t ) , (ii) Atu(t), and (iii) A t 2 u ( t )
10.9 The load of a position-control System is an where A is constant and u ( t ) = 0 for t < 0 , u ( t ) = 1 for
undamped flywheel of moment of inertia I. The driving t z 0.
torque on the load may be assumed to be KOtimes the
motor input voltage I/. A three-term controller and l0.l6 Figure ''-56 shows a flywheel A which is driven
amplifier whose output is V have a combined transfer by a motor M having an Output torque Q . Flywhee1 A
operator drives flywheel B by viscous action, the torque trans-
mitted being C times the relative angular velocity. One
Vl0, = (K, + K2D+ K3ID) end of a spring of torsional stiffness S is attached to B,
where 0, is the position emor, and D the operator d d t . the Other end being [Link] moments Of inertia Of A
and B are In and IB respectively; the inertia of M is
a) Show that the maximum value of K3 for stability is negligible. -
KO K1 K2lI.
b) Show that the steady-state position error for each
of the following inputs is zero: (i) step input, (ii) ramp
input, and (iii) acceleration input.
10.10 A simple position-control system has a vis-
cously damped load. The moment of inertia of the load
is 4 kg m2 and the damping constant is 8 N m per rads.
The driving torque applied to the load is K times the a) Derive a differential equation relating Q to the
position error and the system has a damping ratio of angular position 0.4 of A.
unity. (a) Find the value of K . (b) If the system is b) If A is the load in a position-control system and Q
initially at rest and then at t = 0 the input shaft is is K times the error, obtain the fourth-order output-
rotated at 0.4 rads, find the steady-state position error. input system equation and show that the system is
10.11 For the previous problem, show that the posi- always stable.
tion of the load is given by 10.17 In a speed-control system, the driving torque
e~= 0.4t-0.8[1 - e-'(1 + i t ) ] where t is the time in T,, which is applied to the load increases by 0.01 N m
seconds. Find when the maximum acceleration of the for each rads of errOr w,. The load consists of a
load occurs and determine its value. flywheel of moment of inertia 0.5 kgm2 with viscous
10.12 Derive all of equations 10.17. damping amounting to 0.04 N m per r a d s of load
speed.
10.13 See example 10.3. Rewrite equations (iii) to
(vi) in terms of 0, instead of 0,. Draw the phase-plane a) If the load is running at a constant speed of 150
plot of 0,/w, against 0, and hence show that the final r a d s with no error, determine the equation relating TD
error 0, is 0.02 rad. to we and find the time-constant of the system.
the displacement x of the trolley such that
D2x = ( A + BD) 6 where A and B are positive con- Frequency (w)/ Amplitude Phase lag/
et-ntc (rad/-') ratio degrees
0.6 17.0 -
1 10 80
2 5.5 69
3 4.0 60
4 3.2 50
5 2.7 45
10 2.2 29
20 2.1 16
100 2.0 5

mics of the pendulum are represented by Frequency (w)/ Amplitude Phase lag/
(g - &D2)6 = D2x and that the control will be success- (rad/s-') ratio degrees
ful provided that A >g.
Initially the control is switched off and the pendulum 1 5.0 96
held at an angle 6 = el. At time t = 0 the pendulum is 2 2.5 100
released and simultaneously the control is brought into 4 1.1 110
action. Show that, in the steady-state, the trolley has a 8 0.5 130
constant velocity to the right and determine this 20 0.1 155
velocity. 50 0.02 170
100 - 175
10.21 Obtain accurate Bode plots of the transfer
182 Introduction to automatic control

a) What is the appropriate value of K? 10.26 The transfer function of a first-order lag is of
the form E ( j w ) = (1 + dw)-'. Show that, at the break
b) What is the gain margin? frequency w = UT,the slope of the phase-frequency
c) What is the damping ratio of the closed-loop plot is -66"/decade.
system?