19EEE04-DCS Unit-1: Introd

Unit-I: Introduction to [

Contents

[1 Introductionl 1

2 Concepts of Digital Control Sys- |

|__ tems! 3

5
4 ‘lypes of Sampling Operations| 6

Quantizing and Quantization Error 6

(6 Sampling Theorem] 7

Syllabus
Basic concepts of digital control systems -
block diagram - analog to digital and digi-
tal to analog conversion - sampling and hold
devices - multiplexing - sampling and sam-
pling theorems.

1 Introduction

These days, systems without digital con-

trollers are outdated. In recent years signif-
icant progress has been made in the analy-
sis and design of discrete-data and digital
control systems. These systems have gained
popularity and importance in industry due
in part to the advances made in digital com-
puters for controls and, more recently, in
microprocessors/micro-controllers and dig-
ital signal processors.
Digital control systems differ from the con-
ventional continuous or analog systems in
that the signals in one or more parts of these
systems are in the form of either pulse trains
or numerical codes. The terms sampled-data

*Manavaalan G, Associate Professor, Department

of Electrical and Electronics Engineering, Coimbat-
ore Institute of Technology, Coimbatore, E-mail:
manavaalan.g@cit.edu.in.

EEE, CIT lo
uction to DCS Manavaalan G

igital Control Systems *

control systems, discrete-data control systems,

and digital control systems have all been used
loosely and interchangeably in the control
systems literature. [I] p. 1]
The application of computer/digital con-
trol has made possible “intelligent” mo-
tion in industrial robots, the optimization
of fuel economy in automobiles, and refine-
ments in the operation of household ap-
pliances and machines such as microwave
ovens and sewing machines, among others.
Strictly speaking, sampled data are pulse-
amplitude modulated signals and are ob-
tained by some means of sampling an ana-
log signal. A pulse-amplitude modulated
signal is often presented in the form of a
pulse train with signal information carried
by the amplitudes of the pulses. Digital data
usually are those signals generated by dig-
ital computers or digital transducers; they
are often in some kind of digitally coded
form. Practical systems found in industry
often contain analog, sampled, as well as
digital data. [2) p. 1]
Among the advantages of digital ap-
proaches for control are the increased flex-
ibility of the control programs and the
decision-making or logic capability of dig-
ital systems, which can be combined with
the dynamic control function to meet other
system requirements. In addition, one hard-
ware design can be used with many differ-
ent software variations on a broad range of
products, thus simplifying and reducing the
design time. |B} p. 1]
The signals used in the description of con-
trol systems are classified as continuous-time
and discrete-time. Continuous-time signals
are defined for all time, whereas discrete-
time signals are defined only at discrete in-
stants of time, usually evenly spaced steps.
The signals for which both time and am-
plitude are discrete, are called digital sig-
nals. Because of the complexity of dealing
with quantized (discrete amplitude) signals,
digital control system design proceeds as if
computer-generated signals were not of dis-

£ 9
19EEE04-DCS Unit-1: Introd

crete amplitude. If necessary, further analy-

sis is then done, to determine if a proposed
level of quantization is acceptable. [4 p. 1]

Types of Signals

x(t}

(a)

x{t}

x{r}

*LTELL
xit)

(d}

0 t

Figure 1: (a) Continuous-time analog singal; (b)

continuous-time quantized signal; (c) sampled-
data signal; (d) digital signal.

1. A continuous-time signal (as shown in

Figure [1] (a)) is a signal defined over a
continuous range of time. The amplitude

EEE, CIT 20
uction to DCS Manavaalan G

may assume a continuous range of val-

ues or may assume only a finite number
of distinct values.
An analog signal defined over a con-
tinuous range of time whose amplitude
can assume a continuous range of val-
ues. We frequently use the terminology
“continuous-time” in lieu of “analog sig-
nal”.

. The process of representing a variable by

a set of distinct values is called quantiza-
tion, and the resulting distinct values are
called quantized values. The quantized
variables changes only by a set of distinct
steps. Continuous-time quantized sig-
nal is shown in FigurefI] (b).

. A discrete-time signal is a signal defined

only at discrete instants of time (i.e., one
in which the independent variable t is
quantized). In a discrete-time signal, if
the amplitude can assume a continuous
range of values, then the signal is called
a sampled-data signal, and is shown in
Figure [I] (c). A sampled-data signal can
be generated by sampling an analog sig-
nal at discrete instants of time. It is an
amplitude-modulated pulse signal.

. A digital signal is a discrete-time signal

with quantized amplitude. Such a signal
can be represented by a sequence of num-
bers, for example, in the form of binary
numbers.
In practice, many digital signals are ob-
tained by sampling analog signals and
then quantizing them; it is the quantiza-
tion that allows these analog signals to be
read as finite binary words. Figure fl] (d)
depicts a digital signal. Clearly, it is a sig-
nal quantized both in amplitude and in
time.
The use of digital controller requires
quantization of signals both in amplitude
and in time.

Types of Control Systems

Continuous Signal Control System: A sys-

tem in which the signals are continuous-
time signals, and designed and analyzed in

£ 9
19EEE04-DCS Unit-1: Introd

DIOEREIE |

X(t) [ADC
pcccsseasl
K(nT e(nT) 6(z)
5

ro
CONTRI
SYSTEI

co,
X(t)

(t

Figure 2: (a) Digital control with continuous sig

I/Os.

-o- Coe
Qo "Om
ooo o-oo
S/H Laden Decdbesaees
ang Degitat DA
aD compuler converter
converter

4 } 4

Clock

,
L
Figure 3: Block diagram of a digital control syste1
6]

CONTROL
SYSTEM

M(t)
1 H(S) | M(t)

Figure 4: Continuous signal analysis in s-plane

(classical control system). Cont

s-plane. In general it is called as classical

control system as shown in Figure
Digital Control system: A system which
uses digital device for the realization of
compensator/controller is called digital

EEE, CIT 30
uction to DCS Manavaalan G

BT aac |e eta | pe
OL fOH Z
vi CONTINUOUS SIGN

aM(OT) adc | ‘ M(t)

| |

ez) -—) pranr | Yen)

ONTROL
SYSTEM

a M(nT)

»)
nal - I/Os; (b) Digital control with discrete signal

— Hold
*
eircwit
fh A
eruntor
hare
~|
Piant or
procets
4~

ransducer

n showing signals in binary or graphic form.[ p.

control system. It may use continuous /dis-

crete input-output signals. Digital control
system is shown in Figure

2 Concepts of Digital Con-

trol Systems
Discrete-time or digital control systems are
control systems in which one or more vari-
ables can change only at discrete instants of
time. A block diagram of a digital control
system showing signals in binary or graphic
form is given in Figure|3, The controller op-
eration is controlled by the clock. In such

£ 9
19EEE04-DCS Unit-1: Introd

a digital control system, some points of the

system pass signals of varying amplitude
in either continuous time or discrete time,
while other points pass signals in numerical
code, as depicted in the figure.
The output of the plant is a continuous-
time signal. The error signal is converted
into digital form by the sample-and-hold
circuit and the analog-to-digital converter.
The conversion is done at the sampling
time. The digital computer processes the
sequences of numbers by means of an al-
gorithm and produces new sequences of
numbers. At every sampling instant a
coded number must be convert4ed to a
physical control signal, which is usually a
continuous-time or analog signal.
The digital-to-analog converter and the
hold circuit convert the sequence of num-
bers in numerical code into a piece-wise
continuous-time signal. The real-time clock
in the computer synchronizes the events.
The output of the hold circuit, a continuous-
time signal, is fed to the plant, either directly
or through the actuator, to control its dy-
namics.

Sample-and-Hold (S/H)
“Sample-and-hold” is a general term used
for a sample-and-hold amplifier/circuit. It
describes a circuit that receives an analog in-
put signal and holds this signal at a constant
value for a specified period of time. Usually
the signal is electrical, but other forms are
possible, such as optical and mechanical.

Analog-to-Digital Converter (A/D)

Am analog-to-digital converter, also called
an encoder, is a device that converts an ana-
log signal into a digital signal, usually a nu-
merically coded signal. Such a converter
is needed as an interface between an ana-
log component and a digital component. A
sample-and-hold circuit is often an integral
part of commercially available A/D con-
verter. The conversion of an analog signal
into the corresponding digital signal (binary
number) is an approximation, because the
analog signal can take on an infinite num-

EEE, CIT 40
uction to DCS Manavaalan G

Continuous-time
continuous-amplitude Digital
signal words

i Sampler Quantizer Encoder to

Discrete-time Discrete-time
continuous-amplitude _ discrete-amplitude
signal signal

Figure 5: Operations performed by an A/D con-

verter. [4] p. 27]

ber of values, whereas the variety of differ-

ent numbers that can be formed by a finite
set of digits is limited. This approximation
process is called quantization.

Digital-to-Analog Converter (D/A)

A digital-to-analog converter, also called a

decoder, is a device that converts a digital
signal (numerically coded data) into an ana-
log signal. Such a converter is needed as an
interface between a digital component and
an analog component.

Digital Discrete-time Analog

words signal signal

‘| Zero-order ;
Decoder fala

Figure 6: Operations performed by an D/A con-

verter. [J p. 27]

Plant or Process

A plant is any physical object to be con-

trolled. Examples are a furnace, a chemical
reactor, and a set of machine parts function-
ing together to perform a particular opera-
tion, such as a servo system or a spacecraft.
A process is generally defined as a progres-
sive operation or development marked by a
series of gradual changes that succeed one
another in a relatively fixed way and lead
toward a particular result or end. Examples
are chemical, economic, and biological pro-
cesses.

£ 9
19EEE04-DCS Unit-1: Introd

Transducer

A transducer is a device that converts an in-

put signal into an output signal of another
form, such as a device that converts a pres-
sure signal into a voltage output. The out-
put signal, in general, depends on the past
history of the input.
Transducers may be classified as analog
transducers, sampled-data transducers, or
digital transducers.
An analog transducer is a transducer in which
the input and output signals are continuous
functions of time. The magnitudes of these
signals may be any values within the physi-
cal limitations of the system.
A sampled-data transducer is one in which the
input and output signals occur only at dis-
crete instants of time (usually periodic), but
the magnitudes of the signals are unquan-
tized.
A digital transducer is one in which the input
and output signals occur only at discrete in-
stants of time and the signal magnitudes are
quantized.

3 Sample-and-Hold
In the special case of discrete-time signal
processing with a unit-gain algorithm, and
negligible time delay (i.e., y(k) = x(k)),the
combined action of the A/D converter, the
computer, and the D/A converter can be de-
scribed as a system that samples the analog
signal and produces another analog signal
that is constant over the sampling periods.
Such a system is called a sample-and-hold
(S/H) system. Input-output behavior of an
S/H system is described diagrammatically
in Figure [7]

Xa() valt)

Figure 7: Input-output behavior of a sample-

and-hold system. [4 p. 77]

EEE, CIT 50
uction to DCS Manavaalan G

S/H operations require modeling of the fol-

lowing two processes: (i) extracting the sam-
ples, and (ii) holding the result fixed for one
period.
The impulse modulator effectively extracts
the samples in the form of x(k)d5(t — kT).
The remaining problem is to construct a lin-
ear time-invariant system which will con-
vert this impulse into a pulse of height x(k)
and width T as shown in Figure Bt (a). The
S/H may, therefore, be modeled by Figure}
(b), wherein the ZOH is a system whose re-
sponse to a unit impulse 6(t) is a unit pulse
gno(t) of width T. The Laplace transform of
the impulse response gy9(t) is the transfer
function of the hold operation, namely,

= [aioe
Gro(s) = L{gno(t)}

T
= | e dt
0
1—e st
S
(1)
In a majority of practical digital operations,
T
Xalt) A x*(f) ZOH Yalt)

(2) EnolD)
| 1

t 0 T ft

(a)
T ,
Xs (s) a. ~ a Ghols) = 1-7? | v4
= | Ya(s)

X (s)= ¥ x(Ke™
k=0
(b)

Figure 8: A model of a sample-and-hold opera-

tion. [A p. 77]

S/H functions are performed by a single

S/H device. It consists of a capacitor, an
electronic switch, and operational amplifiers
Figure (| Op-amps are needed for isola-
tion; the capacitor and switch cannot be con-
nected directly to analog circuitry because of
the capacitor’s effect on the driving wave-
form.
Since the voltage between the inverting and
non-inverting inputs of an op-amp is mea-

£ 9
19EEE04-DCS Unit-1: Introd

-—O
Ly 1 Vour
1
Vin 1 Hold
| capacitor
Control 4
aa | logic
Sample/hold & | =>
=
pulse

Figure 9: A sample-and-hold device. [4 p. 78]

sured in micro-volts, we can approximate

this voltage to zero. This implies that the
voltage from the inverting input (—input)
to ground in Figure P|is approximately Vin;
therefore, the output of first op-amp is ap-
proximately Vjy.

When the switch is closed, the capacitor

rapidly charges to Vin, and Vour is equal to
Vin approximately. When the switch opens,
the capacitor retains its charge; the output
holds at a value of Vn.
If the input voltage changes rapidly while
the switch is closed, the capacitor can fol-
low this voltage because the charging time-
constant is very short. If the switch is sud-
denly opened, the capacitor voltage repre-
sents a sample of the input voltage at the
instant the switch was opened. The capac-
itor then holds this sample until the switch
is again closed and a new sample is taken.

4 Types of Sampling Oper-

ations
A signal whose independent variable time
t is called a discrete-time signal. A sam-
pling operation is basic in transforming a
continuous-time signal into a discrete-time
signal.
There are several different types of sampling
operations of practical importance:

1. Periodic sampling: In this case, the sam-

pling instants are equally spaced, or
th = kT, (k = 0, 1, 2, 3, ...). Periodic
sampling is the most conventional type of
sampling operation.

2. Multiple-order sampling: The pattern of

the t;’s is repeated periodically; that is,

EEE, CIT 60
uction to DCS Manavaalan G

thy — tx is constant for all k.

3. Multiple-rate sampling: In a control system

having multiple loops, the largest time
constant involved in one loop may be
quite different from that in other loops.
Hence, it may be advisable to sample
slowly in a loop involving a large time
constant, while in a loop involving only
small time constants the sampling rate
must be fast. Thus, a digital control sys-
tem may have different sampling periods
in different feedback paths or may have
multiple sampling rates.

4. Random sampling: In this case, the sam-

pling instants are random, or f, is a ran-
dom variable.

5 Quantizing and Quanti-

zation Error
The main functions involved in analog-to-
digital conversion are sampling, amplitude
quantizing, and coding. When the value of
any sample falls between two adjacent “per-
mitted” output states, it must be read as
the permitted output state nearest the actual
value of the signal. The process of represent-
ing a continuous or analog signal by a finite
number of discrete states is called amplitude
quantization.
Quantizing means transforming a continu-
ous or analog signal into a set of discrete
states. Encoding or coding is a process of as-
signing a digital word or code to each dis-
crete state.
The standard number system used for pro-
cessing digital signals is the binary number
system. In this system the code group con-
sists of n pulses each indicating either “ON”
(1) or “OFF” (0). In the case of quantizing,
n “ON-OFF” pulses or bits can represent 2”
amplitude levels or output states.
The quantization level Q is defined as the
range between two adjacent decision points
and is given by

_==FSR
Q

£ 9
19EEE04-DCS Unit-1: Introd

where the FSR is the full-scale range. Note

that the leftmost bit b,,1 of the natural bi-
nary code (by,_1 by—2 ... by bo) has the most
weight (one-half of the full scale) and is
called the most significant bit (MSB). The
rightmost bit bo has the least weight (1/2”
times the full scale) and is called the least
significant bit (LSB).

Quantization Error:
Since the number of bits in the digital word
is finite, A/D conversion results in a fi-
nite resolution. That is, the digital out-
put can assume only a finite number of
levels, and therefore an analog number
must be rounded off to the nearest digital
level. Hence, any A/D conversion involves
quantization error. Such quantization error
varies between 0 and +Q/2, ie.,

oto tx FSR
5 * On = 0< |e(#)| <
This error depends on the fineness of the NIO
quantization level and can be made as small
as desired by making the quantization level
smaller (that is, by increasing the number of
bits n). In practice, there is a maximum for
the number of bits n, and so there is always
some error due to quantization. The uncer-
tainty present in the quantization process is
called quantization noise.

6 Sampling Theorem
Sampling Theorem

Let x(t) be a band-limited signal with

Xa(jw) = 0 for |w| > wm. Then xq(t)
is uniquely determined from its samples
x(k) = Xq(kT) if the sampling frequency
Ws; = af Ws > 2Wm , i.e., the sampling fre-
quency must be at least twice the highest fre-
quency present in the signal.[4) p. 84]
Another statement
If ws, defined as 277/T, where T is the sam-
pling period, is greater than 2w}, or

Ws > 2W4

where w is the highest-frequency compo-

nent present in the continuous-time signal

EEE, CIT 7 0
uction to DCS Manavaalan G

x(t), then the signal x(t) can be recon-

structed completely from the sampled sig-
nal x*(t).[ p. 90]
The theorem implies that if w; > 2w , then,
from the knowledge of the sampled signal,
it is theoretically possible to reconstruct ex-
actly the original continuous-time signal.
We shall use an intuitive graphical approach
to explain the sampling theorem.
We shall assume that a continuous-time sig-
nal x(t) has frequency spectrum X(jw) as
shown in Figure

| Xjolid
§—_ 1

“~w, O Wy iw

Figure 10: Frequency spectrum of a continuous-

time signal x(t). [Zl p. 90]

To show the validity of this sampling the-

orem, we need to find the frequency spec-
trum of the sampled signal x*(t). The
Laplace transform of x*(t) is given as

X*(s) = Lx*(t)

- z L X(s + jwek), with ZIC

- (2)
=5, Y* X(s-+ jewsk) + + MOH) with ZIC
— 3)

To obtain the frequency spectrum, we sub-

stitute jw for s in Equation (Q). (To dis-
cussing frequency spectra, we need not be

£ 9
19EEE04-DCS Unit-1: Introd

| X*{ food |

Figure 11: Frequency spectrum |X « (jw)| of a sa

Ws > 2wy. [0 p. 91]

|X*(jod th

é A -
/ i\ i
i if \ , ey
i f> voy. a
t_ j ju [| pez] 4

Figure 12: Frequency spectrum |X « (jw)| of a sa

ws < 2wy. [ p. 91]

concerned with the value of x(0+).) Thus,

1 cO

= T, dx (jw + jwsk)

4 ExGle a)
={ + 2X(j(w)) (4)
+ ZX(j(w+s))
+
Equation () gives the frequency spectrum
of the sampled signal x* (tf). We use that the
frequency spectrum of the impulse sampled
signal is reproduced an infinite number of
times and is attenuated by the factor 1/T.
Thus, the process of impulse modulation of
the continuous-time signal produces a series
of side-bands.
Since X*(s) is periodic with period 271/ws,
X*(s) = X*(s + jwsk), k =0,1,2,...
if a function X(s) has a pole ats = 5},
then X*(s) has poles at s = s1 + jwsk

EEE, CIT 80
uction to DCS Manavaalan G

w, > 2u,

y 11
Lt i i i ae
w, t), 3 w, 2u, w
é 2

umpled-time signal x « (t) for sampling frequency

w, <2w,

_-

Gd

umpled-time signal x « (t) for sampling frequency

(k = 0,1,2,...). Figures [11] and [12] show

plots of the frequency spectra X* (jw) versus
w for two values of the sampling frequency
Ws > 2w, and ws; < 2w4, respectively.
Each plot of |X*(jw)| versus w consists
of |X(jw)|/T repeated every w, = 27/T
rad/s. In the frequency spectrum of
|X*(jw)| the component | X(jw)|/T is called
the primary component, and the other compo-
nents, | X(j(w + wsk))|/T, are called comple-
mentary components.
If ws > 2w 1, no two components of | X* (jw) |
will overlap, and the sampled frequency
spectrum will be repeated every ws rad/s.
If ws < 2w4, the original shape of | X(jw)| no
longer appears in the plot of |X*(jw)| versus
w because of superposition of the spectra.
Therefore, we see that the continuous-time
signal x(t) can be reconstructed from the
impulse-sampled signal x*(t) by filtering if
and only if ws > 20.

£ 9
19EEE04-DCS Unit-1: Introd

It is noted that although the requirement on

the minimum sampling frequency is speci-
fied by the sampling theorem as w, > 2w),
where w is the highest-frequency compo-
nent present in the signal, practical con-
siderations on the stability of the closed-
loop system and other design considera-
tions may make it necessary to sample at a
frequency much higher than this theoretical
minimum. (Frequency, ws is chosen to be
10w to 20w 1.) [ p- 92]

The superimposition of the high-frequency

behavior onto the low frequency is known
as frequency folding or aliasing. Refer Fig-
ure (12)

References
[1] Benjamin C. Kuo. Digital Control Sys-
tems. Oxford University Press, India,
2nd edition edition, 2016.

[2] Katsuhiko Ogata. Discrete-Time Control

Systems. Pearson education, Inc., 2nd
edition edition, 2003.

[3] Gene F. Franklin, J. David Powell, and

Michael Workman. Digital Control of Dy-
namic Systems. Pearson education, Inc.,
3rd edition edition, 2003.

[4] M Gopal. Digital Control and State Vari-

able Methods - Conventional and Intelligent
Control Systems. Tata McGraw Hill Ed-
ucation Private Limited, 4th edition edi-
tion, 2003.

EEE, CIT 90
uction to DCS Manavaalan G