sTABILITY CONCEPTS 139

GHAPTER In this chapler, our major concern in

STABILITY CONCEPTS i) lotest whether the Hystem in stable or not a vesor no answer.
Absolute HLADm
ii) If stable, How slable it is-Relative Stability, Paramneters, such as Damping ralO,
Overshoot, etc. provide indications of the relative stability.
5 Reom the discussions of the Last Chapter,
1. When all the roots of the characteristic equation lie in the L.H.S. of s-plane the
(transient terma) responscs due to initial conditions (or the homogeneous solution)
We havesceninthepreviouschapterthat the time response of a control Ayslem i8 Com will decrease to zero as time approaches infinity (stable condition).
design of a control system is 9. If one or more simple roots are located at the origin (but no roots on the RHS of
only. Basically, theprescribed
bythe pole locations
pole placement in
order to achieve the sppecifications.
Among
most important or the major requirements
he coneernetmany 8-plane) or, one or more pairs of roots on the imaginary axis the responses will he
(a) astep function (6) a ramp function (c) parabolic or increases towards infinity as
> or (d) sustained sinusoidal oscillations, Defined as unstable conditions in
performance specifications, the"stability", from the control 0s the
distinguishIntuitively,
stability'. the term engineer's
two class of systems: 'useful' and 'useless' Asystem 'useless' from tisy
he view point, Linear 8ystem theory.

useful by a DSP engincer! (By

3. If one or more roots found in RHS of s-plane, the responses increase in magnitude
however, can be made toward infinity as t ’a, Again, unstable condition.
Engineer's viewpoint, Digital signal prcessing later in apo
truncations. You will study this under The stability of a linear time invariant system can be stated in a variety of ways.
your care
To get feel of what 'stability' means, refer to the diagrams shown in Fig. 5.4. Strictly speaking, stability is independent of input excitations. However, a well known
definition of stability is as follows:
BIBOStability: "A system is said to be BIBO stable if its output is bounded for any
bounded inpul".
...(5.1)
That is if |rt)| s N<a for t > t
..(5.2)
Then Jct)| s M< for t > to
There are few exceptions to the above definition.
function.
(d Differentiator: For a unit step input the output response is a unit impulse
(a) (b) (c) bounded but the output is not since impulse function is known to have inifinite
Input is
Fig. 5.1. Different conditions of Equilibrium amplitude!
Aball is Integrator: For a unit step input, the output is a ramp function ’ output is not
shown resting on different surface under the influence of gravitationa systems, they are
In Fig. 5.1la) the ball is lying in a very Jarge trough. If the bounded!. However, since a differentiator and integrator are all useful
external disturbing force, it will roll back and forth ball is displaced termed stable system.
relative to its equilibrium point) and (executing overshoots and point:
to the presence of eventually returns back to its equilibriumeitherd 5.1 STABILITY AND IMPULSE RESPONSE
friction). If such a trough
without having a ridge as in Fig. 5.1(6), we can exist, stretched to infinity
on
an input r(t) is given by
stable". However, the ball in Fig. say the system is "Globally Asymplo The output of a linear system for c(t) = r(t) * glt)
...(5.3)
foe
equilibrium position and would never5.1c) be ablewilltorollcomedownbacktheto itshilloriginal disturbedunder:
if it is position * implies covolution operation.
Specifically,
ciarcondition
cumstance!in such asituation is termed as 5.16)d
Where gt) is the system impulse response and
come on top of the between the two just
ridge.
"Unstable". The systemin Fig.
1fit goes described. is stable provided it doesregainitson
It not crossOH c(t) =
...(5.4)

lequi
imitleibdlriyumstable
position. slightly 0
integral is no greater than the integral of
or "Locallysystem is said to beyond
The the ridge it can neverMarginallysta Remembering that "the absolute value of an leads to
be
systems).If it is5.1d) depicts asymptotically stable"condi(a tisituation
surface.
Fig. onally stable or nonlineara
typical ofrestingonaf ne absolute value of the integrand
and", Eq. (5.4)

presence di st urbed, another

it will interesting ..(5.5)
of
system) reachesfrictioitsn. initial move a
Although, the position fixed situation. stops (comesotherstate&
The ball is now to
rest)d'
|C()| s
has distance and
"Non-asymptotically stable". zero value.
Sait

(the
Systems ofchanged, its velocity
this type, 1 Some sense. are

138
 CONTROL SYSTEM
140
a bounded signal it follows: STABILITY CONCEPTS
sN<, ie.. nt) is
Now, if |nt)l 141
1 If all the roots of the
require
is 'stable'. characteristic equation have negative real parts, the system
bounded output, we .Tfany characteristic root has positive real part or
Hence if at) is to be a jo axis the system is "unstable"
if there is a repeated root on
the
Te*he condition (1) is satisfied except for the
roots on the j0aKIs, the system is presence of one or more, non
me term Absolute stability with respect"Marginally stable" or "limitedly stable"repeatea
to one or more parameters means that
sP<*, P = system is stable for all values of the parameter or parameters. A system is said tothebe
N
"ronditionally stable" with respect to a parameter, if it is stable only for certain bounded
ranges of values of that parameter.
absolute value curve of the impulse response gt), in
ie.. The area under the the Once the system is identified to be stable question will arise how stable it is. (Degree
(0, ) must be finite. t otability). The Relative stability may be measured by relative settling times of each
Further. by definition characteristic root.

Ge fe*. gt)dt 52 NECESSARY CONDITIONS FOR STABILITY

As explained earlier, for stability all the characteristic roots should have negative real parts.
Anecessary condition can be obtained from the concepts of the Theory of equations. Let the
characteristic equation be
The roots of the characteristic equation are the poles of Gs), the F\s) = as t a,s-l + a,s +...+an-g' +a,-8 + a, = 0 ...(5.13)
function: and when s-takes on these valuesGs) =, Also s = G +j0andoverall
t
e : Without any loss of generality assume a, > 0. (for if a, <0, it can be made positive by
Then Eq. (5.10) becomes multiplying the equation throughout by -1).
From Theory of equations, it follows:

(i) Sumn of the roots, E roots = ...(5.14a)

If one
the jo-aris thenmoreo roots of the
0 and hencecharacteristic
e- s1 = L
equation are in the R.H.S. of s-plant. E ...(5.14b)
Then Eq. (5.11) can be (iü) E products of pairs of roots =
written
as
...(5.14c)
(iü) E products Groups of 3 roots' =
for
...(5.14d)
Rels) > 0 (iv) E products of Groups of 4 roots' =+
l.e.,

Eq. (5.12) ...(5.14e )

"For a system tocont radicts Eq. (5.8), the follows So on till, product of all roots = (-1),
the left half of the be BIBO stable, the rootsstability condition and it automatically 0; (5.14b) means
parts, then Eq. (5.14a) means that a, >co-eficients
s-plane' . i.e., All of the alIlinei
Each
charact characteristic equation must realpars u allroots have negative real All the must
characteristic respornsieriblsetic
on and (5.14e) means a, > 0i.e.,
as root roots shall have Degative L2*O(5.14c) means a. >0so
whichdifferent
will be zero, this means that at least one
are modes ofby the system. In
co-effiient happens to be
for a Specific type of output. c enon-negative. Futber if anv
We making the co-eficient zero. Therefore, the
cancelle d have positive real nart resnonsible in
other
of the
systems-pole-zero numerator calculation
the
system. However, in vast factors
of
due to Gs)
of a system. there ma wll
tocaessary condition for the
svstem to be stable is that "all the
co-effcients of the characteristic
condition stated is sufficient to
frethequencies
of and cancellamajtions)orit.y mput excitation or duet
Such transfer Tunctions do not exhibit llO n be non-npgntine and non vanish". Although the
second order systems, it is not
enough for higher order
transfer functsystem
all13
of Sn stability of First and
ion. stability
exhibit ofp

Accordingly, folsilotwiuatngions,determined by examination

may transfer functions
thetherefore be the us
quite
on stability are
 CONTROL S
142 =0
SYSTEM E STABILITYCONCEPIS 143

+ 2s + 24 Matrix is
systems. Forexample Rs) = s + s The Hurwitz
2 4 0 0
1 3 12
factoredas -j3) = 0 D, =
maybe + j3) (8 - 1 satisfiedby Eq. (5.15a), Eq. 2 4
3) (s - 1
(s +
Although the
necessary
conditionsare
parts.
Therefore, satisfaction of the
positive real stability but there is no guarantee,
(n5e.c1e5b)s arytelatond 1
The various Hurwitz determinants are
3 12
have
complex roots possibilityof system
implies only a to examine the
sufficient
conditions. This \eadk 2 4|
proceed further
CRITERION D, =2 > 0, D, = 1 3 = 2 >0
CONDITIONS-R-H
SUFFICIENT 2 4 0
5.3 the method of investigating
independently publishedcriterion
and E.J. Routh
A. Hurwitz stability of a system. The Routh's is
indeed the the s D, = 1 3 12

conditions of
jtselfbut Routh's
abbreviated
method is
"R-H Criterion"
easy to adopt and the
criterion is known by their Hurwnaimtzes th |0 2 4
= 2(12 - 24) - 1(16 - 0) = - 24- 16 = - 40 < 0
2 4 0 0
5.3.1Hurwitz Criterion 1 12 0
by P(s)= a,s" + a ,sn-1 + a. gh-2 D, =
Let =0. the stabilityequation
characteristic
the For of theitsystem
of the system is necessary and sufficient that the Principal
0 2 4 0

a, = 0 1 3 12
all positive.
of the following determinant are |2 4 0
an-1 = 12|1 12| = - 480 < 0
ln-2 0 2 4
0
ln-3 positive real parts and the system is
D, = 0 Since D, and D, are negative F(s) has roots with
lon-4 unstable.
it is not necessary to evaluate |D,| to see
an-5 If all the co-efficients of F(s) are positive disadvantage
several determinants. This is indeed a
if it is still necessary to evaluate
0
a, overcome in Routh's criterion.
The principal minors are shown with dotted enclosures. They are
called "H.: 5.3.2 Routh's Criterion
Determinants". The Hurwitz determinant D, is given by, (1 s K [ ). we always evaluate second order
determinants only.
This is a much simpler criterion as tabulation are identified by the power of s. The
the
lk-1 This is Tabular method. The rows ofcoefficients of F(s) chosen alternately as below.
the
Frst two rows are entered with
D, = 0 ak-2 4
the rule given below.
ak2 unto s°-th row are filled following
0 The thirdand subsequent rows as a pivot. Multiply the
The previous to the row to be filled is taken
The entry entries in Eq. erirst entry of the row other
of the
ay-, =0, if (5.16) and Eoq. (5.17) are the characteristicequ by the pivot and subtract the product
iagonal entry in the row previous to pivot the pivot. This will be the entry in the current
2,,> n; or 2,, 0. co-eficients of the
t00 diagonal entries. Divide this
difference by
equationExample
is 5.1: Determine the chara:t
stability or otherwise of the System whose
row". This procedure is illustrated below:
As) = s + 2s
+3s + 48 + 12 = 0
 column infirst weunstable, the Columntheof are
the to
Observe column can number If "For
"Routh' The
s " shown be 14
of first always a, " "
Eq. during regarded
Al l zeros,as In The made The
columnshould > and a
(5.18), the 0, of
make then roots theRouth
system th e thprocessdiagonally ot
e without piv
the
elements process
minors
bebe all number element B.A,- a,a,-0A..
ayAzA,
Aa,-B,
We c0-eficient of criterion" to process arrow.
is
positive, ofentries the aray be A,BB;
find
marked same stable of to is
characteristic of of be always
sign (as simplify any continued shown
is togenerating =
ifa, sign, of in given itstated C.
D, D, in
> s"
the changes is row bythe is A,0,
on.andso
Eq. without
0. i.e., first
necessary
computational
in as can the arrow. B,A,-A,B,B, A,
a,d,
(5.16) cquation
a,
column ofBq. below: till element adsB,
-
0s (1
any
positive,
be"Routh' s th e The
and the 5.18) and multapied -A,.
D, D D, D loss
must entries insuf)icient
last other thein =
conare of be array" entry C
instead the
generality, also RHS in of work. or multipliagFirst
diagonal A,A,a,-
same like in column.
with the that divided a,A,
of be of first
a
saying positive this. r ow entriesto Bas =
=
the we the sign. all A,,
column the by becomes the The SYSTEM
CONTROL
entries can s-plane". If
that for entries a
state not, positive missing First
be
stability.
the coresponds
in ai the
in
multipliet
o
n entn e syster the terma
e N ons e
x

characteristicintiniteentriesappears Diticulty
The 1test.While
Special
Cases
5.3.3 rootsthere s' in
row thunstable.
system is numberthe roots
() ()
The FirstGernerally,using inare to
e
Purther
first
Notice The
po Hucritrivwteeitnrzieosn sTACONCE
s BILITY PIS
entries A8difficulty
EpsilonReciprocal
Routh's
and or in
the the element
the the
the
net the
RHS two column s" as that Routh Example the Therefore
well. equation if firsttheof
of
in
test next first row marked
in
Approach test may in of
two R-H
(-20 x4)-(2x 12)(1 sign the (2x Pivot 2 1
Array wi5.t2h: corresponds the
approach:
the breaksposition
row any eriterion
tvpes s-plane. the as
changes sign weon entry 3) Consider
areAs
condition
whatever
tirst polynomial behas 12)-(1
x
-(1x for equivalent. the column
as row of move the 20
overcome (idown.
) a of difieculties again in this s-plane
column works a is certain in the 4) the RHS of of
Replace Repeated
consequence row, zero Tabulation
the from x system F(s) entries
positiveness
inference
in Generally changes First 0) -20 system
as approach:
the most
by asbut difficulties first s
=
12
=
an
of
the roots per the are row is s+
is
zero but following column from there column given of added the
(but ofour otherencountered to Example
2s5.1 of
and 0 drawn (i) statement systemthe is Routh the
element not Replace notwhich arise negative
and row. a for
s' is
2x 3
below. + advantage,
hence equal
all with necessarily) entries sign the 12
3s
Tabulation. Hurwitz
three all causing therefore(indicated -1x0-
2 +
by
cases. s and change,
to s' 4s
state erespect by entries this in as the
positive row + determinants
> approaches: 1/: opposite the discussed the unstable.
12
the 0. is there
th is used ow (from is 12 = Routh's Thus
Determine to to
stability
in breakdown as Negative characteristic
+ 0
F\obtain
2) diliculty the asare (indicated
are 0). 12 Routh'
real below: positive (missing
term) 0 tells s
eriterion
about
wil anon two As
ofthe Fz) roots. nextpivot of we and
the sign hold arises row Cro".
in the characteristic as
move to equation. criterion
negative)
O). hence
system of for 0, making
become If Routh's
Thus from to
Fs)Apply if zero 145
the the the and the
 RHS two
s: Approachs-planeroots As Routh'Approach
s
diffithcin
e ulty
146
The
indicated, Example The RHS The (ii)
of Routh's Following number it
and F(z)
maequat
y ion
5.3: Multiplication
2:the
Tabulation
1
S
2 s-plane,
there0= Reciprocal be'a'
1 Replace
1 1 2: 2: Tabulation F(s)
system in
the are : for
Fz) Consiexampl e whiched, shall by
der aggravat
a
19 19 -2 11 =
the RHS two
is polynomial-put the z6
=
sÑthe should factbeor
is by
zero
unstable. sign
of 2
this 3z5
+
0 2 2
as +illustrates
s5 and
such (s a
2
element z-plane.
changes equation + + 1 charact
below: erist+ic be
tahat facta)or
+
2 3 524
+ s
Test
3 3s
remembered
+ al
>
0.
to
1 3
Therefore in
5 by
is 22 = breaks+First 0 2s the IfaF(-a)obtain
approach:
[ the
+l/z +
equation
>
0 first ... 0....
32z and 5s three < #
0
0, as
there
(Row down entry
+
whilethen s)
(Row
0
column.
multiplied
obtain+
z
+ of
approaches.
+
ot h
makingerwi
3s
the se Mult
(s
=
iply
1
are multiplied +
twoThis
1= the 1=0 factor F(s) theCONTROL
a)
0 row assessment
instead
t
roots
means by (s
is original
+11) by zero +
+3) ENGINEER
SYSTEM
of a) of
F(s)z tnat indicates
remedying
eregarding characterg

(contd t
Pa

therefore
unstable.is
ootS
with
The Note: ThenCheck
Approach firstother As
thewhile
s:-ve (+ve)s:
eTCONCEPTS
ABILITY
learly,
positive = (s
Routh
fromeffected
co-efficients By marked
there s': gÑ
. s": + aos+l columnentries
Tabulation co-efficientselecting
1) 3: on
real are (as"
+ by Multiply
F(s)F(-1) entries
the
inthe 1
3 13 16 13 -1 3 3 2 1 (ao mere tabulation (e-1) 2
(e1) 4 ([2
parts. two + ofs" the = = = first 1)
-
+
sign for a,)s" a,st of thro factor s7 (s1-1 indicating
Fs)
inspection. column
The 1 10 6 4 the sn, + +
changes +
+ Thus, s 2s6 1)
+
by asE’
system modified (a, a,s are (s +F(s) (s
+ 3 - + two are 1 2
0 0 3 7 8 + obtained
Co-efficients
in
a)s + 1), 4g 2 a)
roots all the
0,
isthe + +
equation a,sh- wherever 5 positive. 0....
unstable.First 5s - E-0 lim
0 0 0 0 1 4
+ by 3 of First
+
(a, +.. just of + -4-e? 4e
783 F(s)column
column .. .
is 1 Therefore (e2
+ s+l
(an-2 az)s a,-2 addingpossible = -1)
(Row (Row (Row + =
4 0
8s
$ and inentry
0#
e’0 lim
and t
multiplied
multiplied
divided +an-l
+ pairs s°the
+ RHSthere -2 -4
a,-1)s 4s
therefore (ag remain +
in +2
-1) (e2
of of the
by + + +
multiplication
s
co-efficients 1 s-plane. are
a4) = ’
+4) by by (a,-1 a,un-altered
) two s²
0 lim
F(s) +3) +2) row
sign
= + system
The is -Z,- ve
0 a)s
has starting can negative
changes
+ while 147
two an: be
 equation equation
form
to row DiAlltficulty 2 148
The Example The shoul
TodNotice The (ü) (ü) ()Thients ries
s: and all wil an
overcome
Routh
zero"Auxiliary always that
-a situatConjionsugate Pairs Thercondie tion in
Tabulation 5.4: the give the ofare one
1 1 1 1 1 Routh's row the the
Consider correspond (a) are conjugate indicates row
F(s)
rootdificulty, difficulty
wilequation". roots pairs of
0 1 2 2 3 test be
= +a
depicted of the
is g6+
the replaced pairs forming roots real that
0 0 3
ascontinued".
characteristic
5
3 follows: sÑ that
to is Fig.
The the caused
an in Routh's
+ row 5.2. Fig. a roots
on
3s4 withcaused order just odd Causes
..co-eficientsof 0 0
+
2s'Following the ofpowered evenby Quadrature wiTabul
5.2. the
th ation
+
co-efficients
equation+
the this
above
for (b) 0 imaginaropposi
y te
All Eqn. Aux: From 5s equation the number
difficulty. row difficulty-2 in are
zero example
+ (previous the axis
4dA.
row
co-efficients
ds dAA(s)
3s
+ of is
of
s. of
s-plane signs.
or
zero's"
ds
1 3 the always roots.
= = illustrates
= to
4s s 0 first SYSTEM
CONTROL
+ + of the)
Accordingly,
4s 2s2 derivative even. all
s
+ row the (c)
3
= procedure Solution zero +d
0 of row
the
the of is ol.
Au
show
that Lne
difficulty positive
The 1here s: TheExample Solution
quadr
equataitocn.Thus Mhe RHS ds
4 1

Solvins-gplane The shoWn,As

of
d sTACBOILNITCYEPTS
real Routh's other
there other student
are

part.
are
2 25 -1 1 5.5: of two (s2
the test
two
can
two 27 3 1 1
TabulationConsider
s8
s
Auxiliary
+
and now the
are System + roots
roots sign 27 -1 22 6 3 3 4 + S+ 1 the tel s so
indeed verify s+ can +unstable.
system isthere
that of
row
of changes I= 2
two the unstable.
by 0 27 11 11 11 14
is 4sthe S12 =’ also Equation,
= becomes
0 F(s) 0 because
characteristic +j, solving + gives be
roots ± in
9 20
3s characteristic
=found
s=-1t
(1 the ...co-ef. of 9
+ Sse= V st completely
with the (s -1
N3 + sign
+jN2), Aux.
first 0 9
148 + by 2 2s2
two
positiveequation column.
+ 1 2s
dividing +
equation
11_ + jN2 3 zero
+
equation 2dA. 1 Aux. 3) +1
W3 |W3+1 =0 changes
(1 ds +
real (s F(s)
are -j/2) Hence Eqn. + +
2 and
part. that 20_² sbyAls) = it
2 (Row (Row ds dA A(s)
Difficulty: 2 + =- + is
two + 1)
the = 9s 0.605 0.605 are filled
3 characteristic
roots multiplied 6s= so
divided + = and
0 two
+ 9 solving ±j ±j
solutions
clearly
The + = with
3s rots
responsible by 12s +
0 1.169 1.169
16) by + 11s²
3) 22s
the co-effiofcients
roots resultant F(s)
+
for 9 in 149
the with = the
0 of
 unstable.
System equation two 150
Aliter Two Aliter
1525 -s':2e+12
Two s':(-15e s²: The
Example rootentries
s All
Routh1'sExample
2': 22; z3: z5:
sign + sign
2: 3z4 1: 15¬ on
1 1 6 1 6 5
Multiplication
Fs) F(-1)changes Reciprocal changes - the
3 15 +
2
2 3 1 24e- 5.7: in 5.6:
= =
223 15¬ -12 2 2 imaginary the 6 5 Array 7

-1+1-2+ (s
+ the
in -1 2
+
22
polynomial
2
the
in 144),
0 15 3
Find
the
First 0
Investigate
6 is
1) by + 0
s5 colaxius mn
0 1 1 z First
first number
F(s) a + g4
the
1 column
0
+ the A(s)(co-eff. of 12
ds A(s) +
=factor
column. 0 = s4 dA
1
0 is 5s3
stability+
s 2 + system of the of = =
+ -
method. ..(Row .. 2s 0 6s
2s5 3 Two asE’0. roots gives
+ . . .. (Row
(Difficulty: + array 12
day
+ 7s+ of
+ 15 (Row (Row 2s
is
in ds 1 6
3s4
+ 12
= roots Multiplied
Hence
Multiplied 'Marginally
+ the are =tj
s =
0 5sthe
4g3
# in
Multiplieddivided
1,
3s
+ RHSpositive. 1
syste+m6
0 the =
+ [ 15 0
two by
5s2 Method)
RHS
by 4)
by
roots
by
2[ e)
of or
=stable'.
s-plane
0
whoso
+ limitedly as
However
ENGINER
SYSTENM
CONTROL
18s S-plane. + characteristic
2) 12)
in
+ for
l5 the
RHS the
A(s)
characteri
= equatun
of
s-nlans 0
shun

5.3.s-plane.
4 of
The
Example 2.
Observations
.
Some
sTCONCEPTS
ABIs6LI:TY
sº: s': During
Routh's
following
example.
equation above.s-plane.stage
we
RHS If Hence,
appears, all Rna
the
are zero's) gain
of
2
Whilefirst unless 4 -1
1 1 6 1 2 1
5.8: s-plane, if the 5 1
+4e+48l1 -2 not
-2-1
e
-4e [+ 3[1- () Tabulation looking
Considerinterested element completion
formation
the
thehaving two 3

sign
following purpose
we
5 15 -1 -6 -4 2 4 3
we
into are one changes
may of
isthe is or of
inuse thinterested
the a 0 15 9 18 5
e row only Routh's
of
more
polynomial
situation the in
6
L-Hospital
[-method Routh's
is to zero the
magnitude; asses zero (+
3) (+
6) 0 15
intabulation first

F(s)
and (difficultknowing elements
Table
the
rule for column.
proceed.
=determining
OR stability will (+
2)
0 1 2 4
s° to if
1). the always and/or at
+determine we Hence
2s This may The any
number we
0 1 + one
stage
...... situation system needshow two
3s have the the or
(+ + of more
not the a roots
number
sign row
2) 6s to is roots
is unstable proceed
+ write system of
as elements
negative (not
illustrated
3s F(s)
[ of in
’ comprising
+ anroots the
beyond is are
4s 0,
asunstable.
auxliary
although RHS in
+ inper
1 in RHS 151
= the the (1) this
0 of of
 The 5.3.5
in investigation
systems.
determining RHS factolriaHowever
auxi rysmeansymmet, ryourically 152
Tabulation
Routhis The For Example R-H The
s': s: s:
what
Application
This
of As
[’ divides there bid Taking
characteristic
s4 criterion s-plane. Routh to
+ equatapplionicataireon
K260 26 1 1 8s3 values 5.ranges
9: can
be is number
there
of 0,
Array
the dettwoerminleocated the
+
Therefore limit
K 36s²
equation
of
K
The done
frequently
open of R-H
of
are
s': charact
for
eristicusing roots ofthe roots as
this co-ef icieL-ntHsospiin tal sign in e0,
+ a by of sign
K 10 80 36 80s will paramneter RHS the
1
+ loop forcingCriterion changes 48-0
1 2
the
Gs)+
K
=
is
=system
the Gs) =
transfer
the
used investigator
s-plane polynomial F(sequat
) = of
ion right rule the the wil
(Row
.. 0.... K
0
0 s for
first
to to in of ent
hand
s-prlyane.result
or
s
(s+ determineLinear
be functionstable
roots. th
first
e
1 siscompl
+1 F(se)tely rosuggest
4 2
w.
s s
as in
multiplied (sstable? 4) column
8)
(+-
+ (s K
operation.
must = s-plane. thNotis ice the
4) + Feedback columnbe
1
s+2s This that
would irst
andgives the that
(s 4s of
entries the
+ a

26)
by
4s
+ 20) unityconditions
to
cautious
entries Alternativelv
+
2s the sign entyou ry
wrongly
A(s) cannot of
+
feedback Systems
be
20) E-method in
indicating two resul*
+ of ENGINERSYSTEM
CONTROL
+
positive. 4s chanis zero!roW s'
entry
=this
indicate
for using +1 ² let
K :+ divide
system stability 1 £
0= Procedure the = four
’0
the
0.
positive.
ini,
This
is roots Indeei
and sign
row
given of wile

ha Ii in
ty
by
h
L.e., l.e.,
relative of the s'
analysis
stability
the
Relative
system.
row
RHS Since Borcing The Let us Thus
sTCONCEPTS
ABIFor LITY
s-plane.
is

61.22, we Routh Example This

e the stability,
(overshoot.
stability: cannot
the determinethe
Thereforethere First sº: s: array
s s system
5.10: system
(K s' roW Z: s°
: : meansfirst
assign row column -(5901K" is s4
are 0.083 row: 15 (99K10) 20K 1
will column
+
Fig. settlingThis 5910 980 20K characteristic l sº
: values 20K be
the 2 a
5.3. sign-j0.1)
acomplex 5901K 99K
20K (99K10) - row: row
K+ entries stable
the for
term s 260K<
0< entries
Shifting time given - of + :
z-plane changes 5901 100
10
> 980K 260
- 0 K 5s K
concerned
etc.) system (K -<0 980K to thatequation >0
of
value >
+
+ must
- 0 be (10
the The in 0.83 ’ 100) K
positive, correspond range >0
+ be
orgin S-plane
R-H with isthe for 100
K
’K>
K> 0
K)s
+
unstable K+j of
a positive.
’
first 15 K 5
10+
in criterion the and 0.1)
< 99 10 + K<
the 0 we feedback
15
column. for = to 260
s-plane attributes for <
0.10101010... have a =
This
Re K 0 stable 0
can all = 15 control
Accordingly
0.0101010....;
values
be
used of system.
transient
of
for
K.
a 2
preliminary rootsentry
behaviour
are in
the 153
in
 than2'. are
negative negativthean-2. largest basis, thmodi
eir fiedobtain i.e.,
+o transient 154
Solution:
(i) (ii)Calculate
(i)
Example the
in
Since Routh
Array is Solution:Example real
a
with origin n Ifwe If
Thesubstitute the
The The The LHS all time-constant canwe
equat on. fieredspect of
parts imodi this response real
characteristic closedcosed the 5.12: of
the
z We 5.11: find z-plane to simply part
entries
z-plane. more
ß range
A 22
+
substitute Check the
equat iz-onplane.
If is decays thof
10s
+ loop loop
polessystemvalues of
unit
in + associatmiednimnegat
um ivestable, the implies faster moset
feedbackTherefore the 20: 3 z': 2z if in
equation
+ Gs) =
z2: 2
+ Fz) s the than this the Make point S a
(21 are
is of First 2
3 =2-
= = roots diwitshtance means this -o =shi2= ftin g it dominant
K)
+
is
more stable s(s+3)
K+7) (s control column 0 (z 2 of the
-¡. z- plane. with 0 0
t
is
s
+ negative
which
for (s
+
K
roots
all
0
2)°
- the
system. Thus, of that
substitution
respect 2=S=ofsmaller. root
13K Gls) 13)system
ofare + equation the theUsing ¡ -Þ the 1s
th e
0 0
(2
8 shifting
roots to origin. -
= = than positive
has given - most R-H S-pin
lane O,
0 1
2)2 F(s) thecriterion the then
G(s)
+ -1. the from
dominant
= vertical original Thus
equation, this + g and the
open 28(2 imaginary
= + constant
timesYSTEM
0
means
-
8s analyse origin CONTROL
loop roots of
characterisite
2) + axis
F(s) + 28s of
transfer all
35 + axis on the s-plane
are the = 35 and
heth a
stablity
0 =0 trig
indeed roots system
function are
henet

m. of
M
ty N

Solving
ainedspecifically,Pe
Lne be
Can uaonmethod 0therStability
hand, n0
yes
or
THE5.4
sTACONCEBILITYPTS
theexample.
e got ofand analysis For Routh The
haracteristic
open concept directly the answer. METHOD ROOT negativeHence
LOCUS Substitute Thus Forci ng
regarded helps
loop closed the
Tabulation the s': sº: 13Ks: s': ²: 10s3:Routh1
Consider of we Fitrhste
from loop in usingIt than First
row: z' row: 20
range
poles Root assessing does conclude zß s 2103K
- 13K - tabulation
2103K
tbasically
he 2': 2: z:1 =
R-H -1 column + z 10 10
equation: are alocusdampingthe poles not 72Fz) of >
for - 40-5K 7 for
- column 0
second
at Root criterion 40-5K12K 12K 12
+ 1 for K
arestability lend that
7 entries 7 this (4 = and >0 is
s technique as (z
locus. known, - +
=order
0 ratio, a itself the 12 equation K)z obtain - which ’ K>0entries
.:. 13K 21 +
K
and s technique will >0 > to 1 3K
control can settlinghence wellasfor only 1<K< 8 closed 0 + + the
or ’ be 12K12- + K
4 (12K the < to
t he 10 210
= 5K positive,
K>1 is closed be
modified
-
system be
time ideaantheindesign. as tell loop
< - (2
- or positive: 0
a. whether 12)
The well control poles
40 1² loop K
and of or we = +
equation <
tim e 70
closed understood
with
other its system the
The
of
K< need
0 0 (21 system is
G(s) domain.
time Root the 8 +
loopP system K
time
response given (z stable
= design. locus
poles s(s+ a) after domain From -
K is 1)
going Technique, stable system + <K< is0
are isthe It 13K=
automatically
andspecifications is
obtained through root a or are
H(s) graphical not-an 0
on more 155
=
locus, 70.
by 1. a the