STABILITY

The total response of a linear control system subjected to both inputs and initial conditions is
given by,
Total response = zero-state response + zero-input response
Zero-state response: The zero-state response is due to the input only; all the initial conditions of
the system are zero.
Zero-input response: The zero-input response is due to initial conditions only; all the inputs are
zero.

For a linear time-invariant system to be stable, it should satisfy the following two criteria:
1. When the system is excited by a bounded input, the output is bounded. This concept is
called bounded-input bounded-output stability (BIBO stability).
2. When no input is applied, the output tends towards zero irrespective of the initial
conditions. This stability is called asymptotic stability.

Bounded-Input Bounded-Output (BIBO) Stability:

Let r(t), c(t) and g(t) be the input, output and impulse response of a linear time-invariant system.
They are related in s-domain as,
C(s) = R(s) ∙ G(s)

The convolution integral relating r(t), c(t) and g(t) is


c(t) =  0
r(t   ) g ( )d

Taking the absolute value on both sides of the above equation, we get

c(t )   0
r(t   ) g ( )d
Since the absolute value of the integral is less than or equal to the integral of the absolute value,

c(t )   r(t   ) g ( )d
0

c(t )   g ( ) r(t   ) d
0

The BIBO stability condition is satisfied if for every bounded input ( r(t )  M1  ) , the output
is bounded ( c(t )  M 2  ) .

Therefore, for a bounded input, the bounded output condition is

 
( c(t )  M1 g ( ) d  M 2
0

Thus the notion of BIBO stability is satisfied if the impulse response g(t) is absolutely integrable,
i.e.

0
g ( ) d
is finite. The area under the absolute value curve of the impulse response g(t) evaluated from t=0
to t=  must be finite.

The nature of the impulse response g(t) is dependent on the poles of the transfer function G(s)
which are the roots of the characteristic equation. These roots may be both real and complex
conjugate and may have multiplicity of various orders. If any of the roots of the characteristic
equation is in the right half of the s-plane, then the transient response increases monotonically
with time or oscillates with increasing amplitude. If there are non-repeated roots of the
characteristic equation on the imaginary axis, the response oscillates with increasing amplitude.
If there is a single root at origin, the output is of constant amplitude. If there are repeated roots at
the origin of the s-plane, the output increases without bound.
Conclusion:
1. If all the roots of the characteristic equation have negative real parts, then the impulse

response is bounded and eventually decreases to zero. Therefore 
0
g ( ) d is finite and
the system is bounded-input, bounded output stable.
2. If any root of the characteristic equation has positive real part, g(t) is unbounded and

0
g ( ) d is infinite. The system is therefore unstable.
 3. If the characteristic equation has repeated roots on jɷ-axis, g(t) is unbounded and

0
g ( ) d is infinite. The system is therefore unstable.
4. If one or more non-repeated roots of the characteristic equation are on jɷ-axis, then g(t) is

bounded but 0
g ( ) d is infinite. The system is therefore unstable.

There are a few exceptions to the normal definition of stability. A pole at origin corresponds to
an integrator. A pair of conjugate poles on the imaginary axis corresponds to a sinusoidal
oscillator. Both the integrator and sinusoidal oscillator are useful devices. The response in both
cases is bounded unless the input has a pole matching one of the system poles on the jɷ-axis. So
depending on the final response, such a system may be treated as acceptable or non-acceptable.
These systems are also called marginally or limitedly stable system.

Absolute stability and relative stability:

Absolute stability is a qualitative measure of stability. It refers to the condition of whether the
system is stable or unstable. Once the system is found to be stable, it is of interest to determine
how stable it is, and this degree of stability is a measure of relative stability. Relative stability is
a quantitative measure of how fast transients die out in the system.

With the help of Routh criterion, determine the stability of the systems represented by the
following characteristic equations. For systems found to be unstable, determine the number of
roots of the characteristic equation in the right-half of the s-plane.

Necessary conditions for stability

The necessary (but not sufficient) conditions for stability of a linear time-invariant, single-input
single output system described by the characteristic equation
q(s) = a0 sn+a1 sn-1+…+an-1 s+ an = 0; a0 >0
is that
1. All the coefficients of its characteristic equation be real and have same sign.
2. None of the coefficients should be zero.
So certain conclusions regarding the stability of a system can be drawn by merely inspecting the
coefficients of its characteristic equation in polynomial form. However, these conditions are not
sufficient, because it is quite possible that an equation with all its coefficients non-zero and of
the same sign may not have all the roots in the left-half of the s-plane.
Consider an nth order characteristic equation given as follows:
a0 sn +a1 sn-1 + … + an-1 s + an = 0
Let the roots of this nth order characteristic equation be s = r1, r2, … , rn. These roots are functions
of coefficients a0, a1, a2, …, an-1, an .
Consider a second order polynomial:
a1 a
a0 s 2  a1s  a2  a0 ( s 2  s 2)
a0 a0

 a0 ( s  r1 )( s  r2 )

 a0 s 2  a0 (r1  r2 )s  a0r1r2

Consider a third order polynomial:

a1 2 a2 a
a0 s 3  a1s 2  a2 s  a3  a0 ( s 3  s  s 3)
a0 a0 a0

 a0 ( s  r1 )( s  r2 )( s  r3 )

 a0 s3  a0 (r1  r2  r3 )s 2  a0 (r1r2  r2r3  r3r1 )s  a0r1r2r3

On extending this expansion to the nth order polynomial, we get

=

If all the roots of a polynomial are real and are in the left half of the s-plane, then all ’s in the
above equation are real and negative. Therefore all polynomial coefficients are positive. If at
least one root is in the right-half of the s-plane, then some of the coefficients will be negative.
Also, it can be observed that if all the roots are in the left-half of the s-plane, no coefficient can
be zero.
For example, consider the characteristic polynomial with all positive coefficients.

The characteristic polynomial can be written as

The coefficients of this polynomial are all positive, but two roots are in the right-half of the s-
plane and so the system is unstable.
Another example is

All the coefficients of this polynomial are positive but two roots are in the right-half of the s-
plane and so the system is unstable.

Routh Stability Criterion

The Routh stability criterion is based on formulating an array(or table) called Routh array (or
Routh table) using the coefficients of the characteristic equation. The first row of the array
consists of first third fifth, etc coefficients and second row consists of second, fourth, sixth, etc,
coefficients all counting from the highest terms. The remaining rows are completed by the
numbers obtained by the indicated operations below for an eight order system.

q(s) = a0 s8 +a1 s7 +a2s6 +a3s5 +a4s4 +a5s3 +a6 s2 +a7 s +a8 = 0

Sufficient Condition for Routh-Hurwitz Stability

The sufficient condition is that all the elements of the first column of the Routh array should
have the same sign. This means that all the elements of the first column of the Routh array
should be either positive or negative.

Routh Array Method:

If all the roots of the characteristic equation exist to the left half of the‘s’ plane, then the control
system is stable. If at least one root of the characteristic equation exists to the right half of
the ‘s’ plane, then the control system is unstable. So, we have to find the roots of the
characteristic equation to know whether the control system is stable or unstable. But, it is
difficult to find the roots of the characteristic equation as order increases.
So, to overcome this problem there we have the Routh array method. In this method, there is no
need to calculate the roots of the characteristic equation. First formulate the Routh table and find
the number of the sign changes in the first column of the Routh table. The number of sign
changes in the first column of the Routh table gives the number of roots of characteristic
equation that exist in the right half of the ‘s’ plane and the control system is unstable.
 Follow this procedure for forming the Routh table:
 Fill the first two rows of the Routh array with the coefficients of the
characteristic polynomial as mentioned in the table below. Start with the coefficient
ofsnand continue up to the coefficient of s0 .
 Fill the remaining rows of the Routh array with the elements as mentioned in the table
below. Continue this process till you get the first column element ofrows0s0isan.
Here,anis the coefficient ofs0in the characteristic polynomial.
Note− If any row elements of the Routh table have some common factor, then you can divide
the row elements with that factor for the simplification will be easy. The following table
shows the Routh array of the nth order characteristic polynomial.
Routh array

s8 a0 a2 a4 a6 a8
s7 a1 a3 a5 a7
s6 a1  a2  a0  a3 a1  a4  a0  a5 a1  a6  a0  a7 a1  a8  a0  0
 b1  b2  b3  a8
a1 a1 a1 a1
s5 b1  a3  a1  b2 b1  a5  a1  b3 b1  a7  a1  a8
 c1  c2  c3 0
b1 b1 b1
s4 c1  b2  b1  c2 c1  b3  b1  c3 c1  a8  b1  0
 d1  d2  a8 0
c1 c1 c1
s3 d1  c2  c1  d 2 d1  c3  c1  a8
 e1  e2 0
d1 d1
s2 e1  d 2  d1  e2 e1  a8  d1  0
 f1  a8
e1 e1
s1 f1  e2  e1  a8
 g1 0
f1
s0 g1  a8  f1  0
 a8
g1
Investigate whether the following systems represented by the characteristic equations are stable
or not.

1. s5  s 4  6s3  12s 2  18s  6  0

s5 1 6 18
s4 12 6 6
 s3 1 6  112 118  1 6 0
 6  12
1 1

s2 6 12  112 6  6  1 0
 14 6
6 6
s1 14 12  (6)  6 0
 14.57
14
s0 14.57  6  14  0
6
14.57

There are two sign changes in the first column of the Routh array. Hence there are two
roots of the characteristic equation in the right-half of the s-plane. Hence the system is
unstable

2. s4  2s3  10s2  8s  3  0

s4 1 10 3
3
s 2 8
2  10  1  8 2  3  1 0
2
s
6 3
2 2

s1 68  2 3 60  20

7 0
6 6
s0 73 60
3
7

All the elements in the first column of the Routh array are positive. So the system is
stable.

3. s6  2s5  8s 4  15s3  20s 2  16s  16  0

s6 1 8 20 16
s5 2 15 16
s4 0.5 12 16
s3 -33 -48
s2 11.27 16
s1 -1.16 0
 s0 16

There are four sign changes in the elements of first column of the Routh array. Hence
there are four roots of the characteristic equation in the right-half of the s-plane. Hence
the system is unstable

4. s5  s4  24s3  48s2  25s  5  0

s5 1 24 -25
4
s 1 48 -5
3
s -24 -20

s2 47.16 -5
1
s -22.54
0
s -5

There are three sign changes in the elements of first column of the Routh array. Hence
there are three roots of the characteristic equation in the right-half of the s-plane. Hence
the system is unstable.

5.

The elements in the first column are: 1, 4, 2.5, 2, 3, -76/15, 4 with two sign changes (3 to -76/15
and -76/15 to 4), there are two poles on the RP and the system is not stable.

Routh-Hurwitz Criterion: Special Cases

Special Cases of Routh Array: We may come across two types of situations,
while forming the Routh table. It is difficult to complete the Routh table from
these two situations.
The two special cases are:
 The first element of any row of the Routh’s array is zero.
 All the elements of any row of the Routh’s array are zero.

Special Case 1: First Element of any row of the Routh’s array is zero

If any row of the Routh’s array contains only the first element as zero and at least one of the
remaining elements have non-zero value, then replace the first element with a small
positive integer, ε. And then continue the process of completing the Routh’s table. Now, find
the number of sign changes in the first column of the Routh’s table by substituting ε tends
to zero.

When forming the Routh table, replace the zero with a small number and evaluate the first
column for positive or negative values of .

1. Determine the stability of the closed-loop transfer function

s5 1 3 5
4
s 2 6 3
3
s 0
2
s
s1
s0

s5 1 3 5
s4 2 6 3
3
s 3.5 0
s2 3 0

s1 0 0

s0 3 0 0
 For 0, there will be two sign changes in first column. Hence there are two roots of
the characteristic equation in the right-half of the s-plane. Hence the system is unstable.

2. Determine the stability of the system with the characteristic equation:

s6 1 1 3 5
s5 2 2 4
4
s

s3
s2
s1
s0

s6 1 1 3 5
5
s 2 2 4
4
s 1 5
3
s

s2
5
s1

s0
5

For 0, there will be two sign changes in first column. Hence there are two roots of
the characteristic equation in the right-half of the s-plane. Hence the system is unstable.
 3. Determine the stability of the system with the characteristic equation:

s6 1 2 3
s5 1 2 5
s4 0

s3
s2
s1
s0

s5 1 2 3
s4 1 2 5
s3 -2 0
2
s

s1
5
s0 5
As 0, there will be two sign changes in first column. Hence there are two roots of the
characteristic equation in the right-half of the s-plane. Hence the system is unstable.

Alternative method: Transform the characteristic equation in the s-plane into the z-plane by
replacing s by and apply Routh’s test. So the characteristic equation in the z-plane is :

z5 5 2 1
z4 3 2 1
3
z -1.33 -0.66 0
2
z 0.5 1
1
z 2
0
z 1

There are two sign changes in first column. Hence there are two roots of the characteristic
equation in the right-half of the s-plane. Hence the system is unstable.

Special Case 2: All the Elements of any row of the Routh’s array are zero.

In this case, follow these two steps:

Write the auxilary equation, A(s) of the row, which is just above the row of zeros.
Differentiate the auxiliary equation, A(s) with respect to s. Fill the row of zeros with these
coefficients.
 1. Determine the number of right hand plane poles in the closed-loop transfer
function:

s5 1 6 8
s4 7 42 56
by 7) 1 6 8
3
s 0 0
2
s
s1
s0

A(s) =

s5 1 6 8
s4 1 6 8
s3 2 6
2
s 3 8
1
s
0
0
s 8 0

All the elements in the first column of the Routh array are positive. So the system is
stable.

2.

s5 1 24 -25
s4 2 48 -50
s3 0 0
2
s
s1
s0
 A(s) =

s5 1 24 -25
4
s 2 48 -50
3
s 8 96
2
s 24 -50
1
s 112.7
0
s -50

There is a single change of sign in the first column of the resulting array, indicating that there is
one root with positive real part.
Solving the auxiliary polynomial equation,

yields the roots s2=1 and s2=

s= and s= 5