8 Z Transform
Uploaded by
Kavya Banala8 Z Transform
Uploaded by
Kavya BanalaChapter
z-Transform
7
7.1 Introduction
For frequency analysis of a discrete periodic signal we have DTFS where we expressed any signal x[n] in
terms of linear combination of complex exponentials. These complex exponentials are orthogonal to each other.
For frequency analysis of a discrete nonperiodic signal we use Discrete Time Fourier Transform (DTFT). We
will look at DTFT in next chapter. The formulae to calculate DTFT of any signal x[n] is
X (e j ) = x( n) e j n ...(1)
n
The condition to find DTFT of any signal is that the signal x[n] should be absolutely summable, that is
x [ n] < ...(2)
n
For signals which are not absolutely summable we use z-Transform for frequency analysis. The z-Transform
of a signal is defined as,
X (z ) = x [ n] z n ...(3)
n
Here z is available re j. The equation (3) define the bilateral z-transform of a signal x[n]. Another way to
define z-transform is that suppose we have an LTI system with impulse response h[n] and input x[n] = Cn where
C is any constant.
n LTI system
x[n] = C y [n]
h[n]
Fig. 7.1 : LTI system with input Cn
Thus, y[n] = x[n] h[n] = h[k ] x[ n k] h[k] Cn k
n k
y[n] = Cn h[k ] C k ...(4)
k
If we define a new function, H(z) = h[ n] z n ...(5)
n
GATE MASTERS PUBLICATION
298 Signals & Systems
Thus, we can say that, y[n] = C n H(z) ...(6)
z C
Thus output of the system = Input ( H( z) Z C ) .
Study Note
n
• Whenever input to system is Cn for < n< where C is any constant then output is C H (z ) z C .
• Cn is called eigen function for discrete time LTI system and H( z) is called eigen value.
Z C
7.2 Bilateral z-Transform
The bilateral z-transform of any discrete signal x[n] is defined as,
X (z ) = x[ n] z n ...(7)
n
here, z = re j
Thus r is the magnitude of z and is phase of z.
z-Plane
It is convenient to represent complex z as a location an complex plane with distance r from origin and
as the angle from positive real axis. The complex plane also called z-plane here.
Relationship between z-transform and discrete time Fourier transform:
Im{z}
j
re
Re{z}
Fig. 7.2 : The z-plane with point re j shown at a distance
r from origin and angle from real {z} axis
From equation (7) we can see that,
X (z ) = x[ n] z n x[ n] ( r e j ) n
n n
= ( x[n] r n ) e j n
n
Thus we can say that, z-transform of x[n] is DTFT of x[n] r–n and z-transform of any signal x[n] at r = 1
will be equal to it’s DTFT.
That is, z{x[n]} = F{x[n] r–n} ...(8)
and for r = 1, z{x[n] = F{x[n]} ...(9)
Thus we can say that DTFT of any signal x[n] = X( z) z e j but the condition is that z-transform should
GATE MASTERS PUBLICATION
z-Transform 299
be valid for r = 1 that is ROC of z-transform must contain r = 1. We will look at ROC in section 7.3.
Poles and Zeros of z-Transform
The z-transform of any discrete signal X(z) can be written as,
b m z m b m 1 z m 1 ... b1z b0
X (z ) = an z n a n 1 z n 1 ... a1z a0
The values of z where X(z) = 0 are called zeros. That is the roots of numerator are zeros, the values of
z where denominator of X(z) tends of zero or X(z) tends to infinity are called poles. Thus roots of denominator
are called poles.
( z z1 ) ( z z2 ) .... ( z zm )
X (z ) = k
( z p1 ) ( z p2 ) .... ( z pn )
The zeros (z1, z2 ...zm) are values of z where numerator of X(z) goes to zero and X(z) = 0. Thus,
LimX( z) = 0
z zi
and poles (p1, p2,...pn) are values of z where denominator of X(z) goes to zero and X(z) = . Thus,
LimX( z) =
z pi
7.3 Region of Convergence (ROC) for z-Transform
z-Transform of x[n] (X(z)) is DTFT of x[n] r–n, to find DTFT of any signal the condition is that it should
be absolutely summable. Thus X(z) is valid any when x[n]r–n is absolutely summable. That is
x[ n] r n < ...(10)
n
Thus for valid X(z) equation (10) should be satisfied. The region of values of r where equation (10) is
satisfied is called ROC of X(z), region of convergence of z-transform.
Study Note
• ROC define the region of values of r i.e. z for which X(z) is valid or for which X(z) exist.
• DTFT of x[ n] X( z) z e j only when z-transform is valid at r = 1 or ROC contain r = 1.
• In z-plane r = 1 means a unit circle.
Example 7.1
Determine the constraint on r z for each of the following sums to converge:
n 1
1
n 1
1
1 ( 1)n n
(a) 2 z n (b) 2 zn (c) z
n 1 n 1 n0 2
Solution 7.1
( a ) Consider the given summation,
GATE MASTERS PUBLICATION
300 Signals & Systems
n 1 n n
1 1 1 1 1 1 1
2 z n =
2
z
2
2 z n 2 z 2 2 z1
n 1 n0 n0
1 1 1
The summation of the second term on the RHS will converge if z 1 or z . Therefore, the
2 2
1
given summation will converge if z .
2
( b ) Consider the given summation,
n 1 n
1 1 1 1
2 zn = 2 zn (2z)n
n 1 2 n 1 2 n 1
1
Therefore, the given summation will converge if 2 z 1, or z .
2
( c ) Consider the given summation,
1 ( 1)n n 1
1
2 z =
2
z n 2 ( 1)n z n
n0 n0 n0
1 1
=
2
( z1 )n 2 ( z1 )n
n0 n0
1
The first summation will converge if z 1, or z 1. Therefore, the given summation will converge
if z 1.
Example 7.2
Determine the z -transform of the causal signal,
x ( n ) = an u(n )
and depict the ROC and the locations of poles and zeros in the z-plane.
Solution 7.2
By definition, X (z ) = x( n)z n a n u( n)z n
n n
1, n 0
Since, u(n) =
0, n 0
we obtain, X (z ) = a n z n ( az1 )n
n0 n0
1
X (z ) = 1
, for az1 1 or equivalently z a
1 az
GATE MASTERS PUBLICATION
z-Transform 301
z
X (z ) = , z a
za
1
Therefore, an u(n) , z a
1 az1
Example 7.3
Determine the z-transform of the anticausal signal,
x ( n ) = – a n u (– n – 1)
and depict the ROC and the locations of poles and zeros in the z -plane.
Solution 7.3
By definition, X (z ) = x( n) z n a n u( n 1) z n
n n
1, ( n 1) 0 n 1
Since, u(–n – 1) =
0, ( n 1) 0 n 1
1
we obtain, X (z ) = a n z n a n zn ( a 1z)n ( a 1z)n
n n 1 n 1 n 1 0
A change of variables is performed by letting m = n – 1, which also yields n = m + 1, m = 0 as n = 1, and
m = as n = . Therefore,
X(z) = (a 1 z )m 1 (a 1 z ) (a 1 z )m
m 0 m 0
1 1
= (a 1 z ) 1
, for a z 1 or equivalent z a
1 a z
1
= z a
1 az1
z
X (z ) = z a
za
1
Therefore, –an u(–n – 1) , z a
1 az1
7.3.1 Properties of ROC
In this section, we explain properties of the ROC of various classes of signals:
1. The ROC of X(z) consists of a ring in the z-plane ‘centered’ about the origin.
2. If X(z) is rational, then the ROC must not contain any poles.
3. If x(n) is of finite duration and is absolutely summable (i.e. n x( n) ) then the ROC is the
entire z-plane, except possibly z = 0 and/or z = .
4. If x(n) is right-sided and of infinite duration (i.e., x(n) = 0 for all n < N1 for some finite N1), then
the ROC is the region in the z-plane outside of the outermost pole, i.e., outside the circle of radius
equal to the largest magnitude of the pole of X(z). Furthermore, if x(n) is causal (i.e., if it is right-
sided and equal to 0 for n < 0), then the ROC also includes z = .
GATE MASTERS PUBLICATION
302 Signals & Systems
5. If x(n) is left-sided and of infinite duration (i.e., x(n) = 0 for all n > N2 for some finite N2), then the
ROC is the region in the z-plane inside the innermost pole, i.e., inside the circle of radius equal to
the smallest magnitude of the poles X(z) other than any at z = 0 and extending inward to and
possibly including z = 0. In particular, if x(n) is anticausal (i.e., if it is left-sided and equal to 0 for
n > 0), then the ROC also includes z = 0.
6. If x(n) is two-sided and of infinite duration (i.e., the signal is of infinite extent for both n < 0 and
n > 0), then the ROC will consist of a ring in the z-plane.
7. For a finite duration right-sided signal (i.e., x(n) = 0 for n < 0 and n > N1 for some finite N1), the
ROC will be the entire z-plane except z = 0.
8. For a finite duration two-sided signal (i.e., the signal is of finite extent for both n < 0 and n > 0), the
ROC will be the entire z-plane except z = 0 and z = .
9. For a finite duration left-sided signal (i.e., x(n) = 0 for n > 0 and n < –N1 for some finite positive
N1), the ROC will be the entire z-plane except z = .
Study Note
• Defining ROC is very important because expression of z-transform of an u[n] and –an u[–n – 1] is same.
•In z-transform ROC has foramt z > magnitude of pole for right-sided signal, but in Laplace transform ROC has
format s > real part of pole for right-sided signal.
Now lets look at few examples to prove and observe all properties of ROC.
Example 7.4
Find the z-Transform of
( a ) [ n ] ( b ) [ n – 1]
( c ) [ n + 1] ( d ) u[n]
n n
1 1
( e ) 3 n u [ n ] u [ n ] ( f ) 3 n u[ n ] u [ n 1]
2 3
Solution 7.4
( a ) For, x[n] = [n], z-Transform is
X (z ) = x[ n] z n [ n] z n
n n
1, n 0
we know that, [n] =
0, n 0
That is signal is finite duration signal.
X (z ) = 1
We can see that, X(z) will never goto or undefined for any value of z. Thus ROC of X(z) is entire
z-plane.
1, n 1
( b ) For, x[n] = [n 1] =
0, n 1
The z-Transform will be
GATE MASTERS PUBLICATION
z-Transform 303
X (z ) = x [n ] z n [n 1] z n z 1
n n
The signal is finite duration and right sided.
1
Thus, X (z ) =
z
The ROC will be entire z-plane except z = 0 because X(z) has a pole at z = 0 and ROC donot contain
pole. Also we can see that when z = 0 then X(z) is undefined and tends to .
1, n 1
( c ) For, x[n] = [n + 1] =
0, n 1
The z-Transform will be
X (z ) = x[ n] z n [ n 1] z n z1 z
n n
The signal is of finite duration and left-sided. Thus ROC will be entire z-plane except z = because at
z = the X(z) = and thus undefined so ROC is entire z-plane except z = .
1, n 0
( d ) For, x[n] = u[n] =
0, n 0
Thus z-transform will be
n
1
X (z ) = x [n ] z n z n z
n n 0 n 0
1
Thus, X (z ) =
1 z1
Since, the signal is right sided signal and pole of X(z) is at 1, so ROC will be r > 1.
1
X (z ) = , z 1
1 z 1
n
1
( e ) For, x[n] = 3n u[ n] u[ n]
2
n
1
So, x[n] has two signals 3n u[ n] u[ n] .
2
Let, X1(z) is z-transform of 3n u[n].
n
3 1
X 1( z ) = 3n u[n ] z n z 1 3z 1
n 0
Since, signal is right sided and pole of X1(z) is at z = 3, so ROC is z 3.
n
1
Let X2(z) is z-transform of u[ n], thus
2
n n
1 n 1 n 1
X 2( z ) = u[ n] z z 1
n 2 n0 2 1 z 1
2
GATE MASTERS PUBLICATION
304 Signals & Systems
1 1
Since signal is right sided and pole of X2(z) is at z , thus ROC is z .
2 2
1 1 1
Thus z-transform of x[n] is 1
with ROC z ( z 3)
1 3 z 1 2
1 z1
2
ROC is ( z 3) .
n
1
( f ) For, x[n] = 3n u[ n] u[ n 1]
3
n
1
x[n] has two signals 3n u[n]and u[ n 1].
3
n
Let, X1(z) is z-transform of 3 u[n], thus
1
X 1( z ) = , ROC z 3
1 3 z 1
n
1
and let X2(z) is z-transform of u[ n 1], thus
3
n n
1 n 1
X 2( z ) = u[ n 1] z
3
3 z n
n n
= (3z)k {Let k = –n}
k 1
3z
X 2( z ) =
1 3z
1 1
Thus, pole of X2(z) is at z and signal is left sided signal, so ROC is z .
3 3
1
We can see that, x[n] is made up of 2 signals which have ROC of z-transform z 3 and z . Thus
3
1
X1(z) is valid for z 3 and X2(z) is valid for z . Thus no valid ROC for z-transform of x[n].
3
Thus z-transform for x[n] cannot be found.
Study Note
• Thus we can see that for right sided signal which is of finite duration the z-transform will have negative power
of z and the ROC will be entire z-plane except z = 0.
• For left sided finite duration signal the z-transform will have positive power of z and ROC will be entire z-plane
except z = .
Example 7.5
Determine z -transform of signals:
n
(a) a , 0 a 1
n
(b) a , 1 a
GATE MASTERS PUBLICATION
z-Transform 305
(c) 1
Solution 7.5
n
( a ) Here, x[n] = a , 0 a 1
x[n] can be written as, x[n] = an u[n] + a–n u[–n – 1]
x (n) = a n Im{z}
0<a<1 Re{z}
a 1 1/a
n
0
(a) (b)
z-Transform 1
a n u[ n]
1 az1
The pole is at z = a and signal is right sided.
Thus ROC will be z a .
n
1
Similarly, z-transform of a n u[ n 1] u[ n 1] is
a
n
1 z-Transform 1
u[ n 1]
1
a 1 z 1
a
1
Since pole of z-transform is at 1/a and signal is left sided so ROC is z .
a
1 1 1
The z-transform of x[n] will be 1
with ROC z a z .
1 az 1 a
1 z1
a
1
Since, a 1 thus ROC will be a z and
a
a2 1 z
X (z ) =
a 1
(z a) z
a
n
( b ) Here, x[n] = a , a 1
Here, also we can break, a
n = an u[n] + a–n u[–n – 1]
z-Transform 1
a n u[ n] , z a
1 az1
GATE MASTERS PUBLICATION
306 Signals & Systems
n
1 z -Transform 1 1
and a u[n 1] , z
1 az 1 a
The ROC of these two signals will have no intersections when a 1.
Thus when a 1 then we cannot find z-transform of signal x[n].
x(n) = a n
a>1
n
0
( c ) Here, x[n] = 1, we can write,
x[n] = 1 = u[n] + u[–n – 1]
z-Transform 1
u[ n] , z 1
1 z 1
z-Transform 1
and u[ n 1] , z 1
1 z 1
so we can see that, ROC of there two signals donot intersect. Thus ROC of x[n] will be null. Thus for no
value of z z-transform of x[n] will exist.
Thus, x[n] = 1 will have no z-transform.
Study Note
n
• From above example we can see that signal x[n] which is rising both side as a for 1 < a < or any signal
which is constant for – < n < as x[n] = 1 will have no z-transform.
• Similarly we can prove that Cn for – < n < will have no z-transform.
Let us understand why z -transform of some signals cannot be calculated:
Signals which are constant for – < n < , rises both direction an for > a > 1, Cn for – < n <
are those for which z-transform cannot be obtained. The condition for existence of z-transform is that x[n] r–n
should be absolutely summable.
1 1.0
n n n
r > 1 r = 0 r < 1
Fig. 7.3 : Plot of r wrt n
GATE MASTERS PUBLICATION
z-Transform 307
For any value of r the signals of above category cannot form x[n] r–n absolutely summable.
Example 7.6
Determine the z-transform and the pole-zero plot for the signal,
n
a , 0 n N 1
x( n )
0 , elsewhere
where a > 0.
Solution 7.6
N 1 N 1
(1 az 1 )N
By definition, X (z ) = x ( n )z n a n z n (az 1 )n
n n 0 n 0 1 az 1
(1 az 1 )N
X (z ) =
1 az 1
1 zN a N
X (z ) =
zN 1 za
since a > 0, the equation zN – aN = 0, or zN = aN, has N roots at
zk = ae j2k/N k = 0, 1, 2,...., N – 1
The zero z0 = a (i.e., root for k = 0) cancles the pole at z = a. Thus,
which has N – 1 zeros and N – 1 poles, located as shown in figure. Since x(n) is of finite duration, the ROC
is the entire z-plane except at z = 0 because of the N – 1 poles are located at the origin.
Im{z}
st
(N – 1) order
pole Unit circle
a 1 Re{z}
Pole-zero pattern for N = 16 and 0 < a < 1
Remember: In a finite duration signal the pole will never be at any finite location. The poles will always be at
origin or at infinity.
7.4 s-To-z-Plane Mapping
Let us consider a continuous time domain signal x(t), the signal is sampled by multiplying the signal
with an impulse train.
GATE MASTERS PUBLICATION
308 Signals & Systems
x(t) x s (t ) x( nT ) (t nT )
n
(t nT )
n
Fig. 7.4 : Sampling of a continous time domain signal
Here T is sampling interval, the Laplace transform of xs(t) is
Xs ( s ) = xs (t ) e st dt n x( nT ) (t nT ) e st dt
which, upon interchanging integration and summation, is
st
Xs ( s ) = x( nT ) (t nT ) e dt x( nT ) e st
n n t nT
= x(t ) e snT x( n) ( esT ) n
n n
By introducing a new variable, z = esT, this equation can be expressed as,
X s (s ) z e sT = x( n) z n X( z)
n
It is cleared from the above discussion that the z-transform can be considered to be the Laplace transform
with a change of variable,
1
z = esT or equivalent, s
ln( z) ...(10)
T
Recall from next chapter that the Laplace variables s was given by
s = + j
where was a constant used to ensure convergence of the integral defining the Laplace transform and
thus the existence of the Laplace transform itself. From the equation,
z = e( + j)t = et ejt ...(11)
so that, the magnitude of z is given by
z = et ...(12)
Now, consider the following three cases:
Case-1: = 0. The imaginary axis (s = j) in the s-plane, = 0 corresponds to z 1 in the z-plane.
That is, the transformation z = esT transforms the imaginary axis (s = j) in the s-plane into a unit circle ( z 1)
in the z-plane.
Case-2: < 0. The left-half of the s-plane, < 0 corresponds to z 1 in the z-plane. That is, the
transformation z = esT maps the left-half of the s-plane into the inside of the unit circle in the z-plane.
Case-3: > 0. The right-half of the s-plane, > 0, corresponds to z 1 in the z-plane. That is, the
transformation z =esT maps the right-half of the s-plane into the outside of the unit circle in the z-plane.
This mapping of the Laplace variable s into the z-plane through z = esT is illustrated in Fig. 7.5.
GATE MASTERS PUBLICATION
z-Transform 309
j Im{z}
s-plane
z-plane
Re{z}
0 1
Fig. 7.5 : Mapping from s-plane to z-plane
Example 7.7
Determine the z -transform of the signal, x ( n ), obtained by sampling the continuous-time
signal,
x ( t ) = e –at u ( t )
every T seconds.
Solution 7.7
The signal, x(n), is obtained by sampling the continuous-time signal,
x(t ) = e–at u(t)
every T seconds therefore,
x(t ) t nT x( nT ) = x(n) = e–anT u(nT)
x(n) = e–anT u(n)
By definition, X (z ) = x( n) z n e anT u( n) z n
n n
anT n
= e z ( eanT z n )
n0 n0
1
X (z ) = aT
for e aT z1 1 z e aT
1 e z 1
7.5 Properties of z-Transform
7.5.1 Linearity
The conversion of a signal from time domain to z-domain is a linear operation.
If, x1(n) X1(z) with ROC = R1
and x2(n) X2(z) with ROC = R2
ax1(n) + bx2(z) aX1(z) + bX2(z) with ROC containing R1 R2 ...(13)
Study Note
If the linear combination is such that there is no pole-zero cancellation, then the ROC will be exactly equal to the
overlap of the individual regions of convergence. But, if the linear combination is such that some zeros are introduced
GATE MASTERS PUBLICATION
310 Signals & Systems
that cancels poles, then the ROC may be larger. A simple example of this occurs when x1(n) and x2(n) are both of
infinite duration, but the linear combination is of finite duration. In this case the ROC of the linear combination
is the entire z-plane with the possible exception of zero and/or infinity.
Example 7.8
Determine the z -transform and the ROC of the signal,
x ( n ) = a n u (n ) – a n u (n – 1)
Solution 7.8
By definition, X (z ) = x( n) z n [a n u( n) a n u( n 1)] z n
n n
= a n u( n) z n a n u( n 1) z n
n n
= ( a z1 )n ( az1 )n
n0 n 1
1 az1
= 1
1 for
az1 1 z a
1 az 1 az
1 az1
X (z ) = 1 ROC is the entire z-plane
1 az1
7.5.2 Time Shifting
If, x(n) X(z) with ROC = R
then, x(n – n0) z n0 X( z) ...(14)
with ROC = R, except for the possible
addition and deletion of z = 0 or z = .
Proof: The z-transform of x(n – n0) is given by
Z[x(n – n0)] = x( n n0 ) z n
n
A change of variables is performed by letting m = n – n0, which also yields n = m + n0, m = – as
n = –, and m = as n = . Therefore,
Z[x(n – n0)] = x( m) z( m n0 ) = z n0 x( m) z m
m m
Z[x(n – n0)] = X( z) z n0
Example 7.9
Using the properties find z-transform of
GATE MASTERS PUBLICATION
z-Transform 311
( a ) [ n + 5]
( b ) 2[3 n + 6] – 5[2n + 5] + 3[ n – 2]
( c ) u[ n] – u[ n – 10]
Solution 7.9
( a ) The z-transform of [n] 1.
The [n + 5] z5, ROC will be entire z-plane except z .
( b ) The z-transform of [n] is 1. Thus,
2[3n + 6] – 5[2n + 5] + 3[n – 2] can be written as, 2[n + 2] + 3[n – 2].
The term 5[2n + 5] will not exist, it will be zero everywhere.
2[n + 2] + 3[n – 2] (2z2 + 3z–2)
Here ROC will be entire z-plane except z = 0 and z = .
1
( c ) The z-transform of u[n] is , so
1 z 1
1 z10 1 z10
u[n] – u[n – 10]
1 z1 1 z1 1 z1
z10 1
The X( z) .
( z 1) z9
Thus it has 9 poles at origin and 9 zeros at circle of radius 1.
ROC of this z-transform will be entire z-plane except origin.
7.5.3 Scaling in z-Domain
If, x(n) X(z) with ROC = R
z0n x( n) X z
then, with ROC = z0 R ...(15)
z0
The notation z0 R implies that the ROC R is scaled by z0 . If R is a z b, then the new ROC is
z0 a z b z0 . Also, if X(z) has a pole (or zero) at z = a, then X(z/z0) has a pole (or zero) at z = z0a. This
indicates that the poles and zeros of X(z) have their radii changed by z0 , and their angles are changed by
{z0}. If z0 has unit magnitude, then the radius is unchanged; if z0 is positive real number, then the angle is
unchanged.
Proof: The z-transform of z0n x( n) is given by
n
z
n
Z[ z0 x( n)] = [z0n x( n)] z n x( n)
n n z0
z
Z[ z0n x( n)] = X
z0
Example 7.10
Find the z-transform of
( a ) e j 0 n ( b ) e j 0 n u[ n]
GATE MASTERS PUBLICATION
312 Signals & Systems
( c ) cos( 0 n ) u [ n ] ( d ) sin( 0 n ) u [ n ]
( e ) a n cos( 0 n ) u [ n ] ( f ) a n sin( 0 n ) u [ n ]
Solution 7.10
(a) x[n] = e j0 n
we can see that magnitude of x[n] will remain constant 1 for – < n < , thus we cannot find
z-transform of this signal. Thus we cannot find z-transform of sin(0n) and cos(0n).
( b ) Here, x2[n] = e j0 n u[n],
the z-transform of step function is
1 1
u[n] 1
1 z 1 1/ z
1
e j0 n u[n] j 0 1
1e z
Similarly
1
e j 0 n u[n]
j 0 1
1e z
In both the above z-transform, poles are at e j0 and e j0 and signals are right sided signal, so
z magnitude of pole z 1.
( c ) Here, x3[n] = cos(0n) u[n]
e j0 n u[ n] e j0 n u[ n]
x3[n] =
2
1 1 1 1 z1 cos( 0 )
Thus, X 3( z ) = j 1
j 1 = 1 2
2 1 e 0 z 1 e 0 z
1 2z cos(0 ) z
and ROC is z 1.
( d ) Here, x4[n] = sin(0n) u[n]
1 j0 n
we can write x4[n] as, x4[n] = [e e j0 n ] u[ n]
2j
1 1 1 z1 sin(0 )
X 4( z ) = j 1
j 1 = 1 2
2 j 1 e z 0 1 e 0 z 1 2 z cos 0 z
and ROC is z 1.
( e ) Here, x5[n] = an cos(0n) u[n]
z
The z-transform, X5(z) = X3
a
1 az1 cos( 0 )
X 5( z ) =
1 2az1 cos(0 ) a 2 z2
and ROC will be z a .
GATE MASTERS PUBLICATION
z-Transform 313
( f ) Here, x6[n] = an sin(0n) u[n]
z
The z-transform, X6(z) = X4
a
az1 sin( 0 )
X 6( z ) =
1 2az1 cos( 0 ) a 2 z2
and ROC will be z a .
Remember: No need to learn how change in signal or z-transform changes ROC, just remember how expression
of z-transform changes and then find ROC using its properties studied in section 7.3.
In this property we have see that,
if, x[n] X(z)
then, an x[n] X z
a
Thus, if y[n] = an x[n] and if X(z) has poles at P1, P2 and zeros at Z1, Z2 then Y(z) will have pole and zeros
at aP1, aP2, and aZ1, aZ2 respectively.
(z Z1 ) (z Z 2 )
That is if, X (z ) =
( z P1 ) ( z P2 )
z z
Z1 Z 2
z
a a
then, Y( z) X =
a z z
( P1 ) P2
a a
j n
2
For example if y[ n] e x[ n] then poles and zeros of Y(z) can be obtained by multiplying ej/2 with poles
and zeros of X(z). That is 90° phase added to poles and zeros.
7.5.4 Time Reversal
If, x(n) X(z) with ROC = R
1 1
then, x(–n) X X( z1 ) with ROC = ...(16)
z R
The time-reversal, or reflection corresponds to replacing z by z–1. Hence if R is of the form a z b,
1 1 1
ROC of the reflected signal is a b or z .
z b a
Proof: The z-transform of x(–n) is given by
Z[(x(–n)] = x( n) z n
n
A change of variables is perform by letting m = –n, which also yields m = as n = –, and m = – as
n = . Therefore,
GATE MASTERS PUBLICATION
314 Signals & Systems
m
1
Z[x(–n)] = x( m) zm x( m) z
m m
1
Z[x(–n)] = X X( z1 )
z
An interesting consequence of the time-reversal property is that, if x(n) is real and even, i.e.,
if x(n) = x(–n)
1
then, X(z) = X( z1 ) X ...(17)
z
In time reversal property we have seen that if z-transform of x(n) is X(z) then,
x[–n] X(z–1)
Thus if a signal x[n] is even that is
x[n] = x[–n]
1
then, X(z) = X(z–1) or X( z) X
z
This is only possible if X(z) has pole at a then it must have pole at 1/a also. Similar condition exist for
zeros also.
Note: Thus if signal is even then its z-transform will have poles and zeros in pairs of (P1, 1/P1), (Z1/1/Z1) respectively.
If a signal x[n] is odd that is
x[n] = –x[–n]
thus, X(z) = –X(z–1)
1
or X (z ) = X
z
Since signal is odd
X(1)= x (n) =0
n
Thus for odd signals the z-Transform will have zero at unit circle. Thus rest of the poles and zeros will
satisfy the condition that if pole is at P1 then there will be pole at 1/P1 and zero at Z1 then zero will be at 1/Z1.
7.5.5 Differentition in the z-Domain
If, x(n) X(z) with ROC = R
dX( z)
then, nx(n) z with ROC = R ...(18)
dz
Multiplication by n in the time domain corresponds to differentiation with respect to z and multiplication
of the result by –z in the z-domain. This operation does not change the location of the poles and hence, the ROC
remains same.
Proof: The z-transform of x(n) is given by
X (z ) = x( n) z n
n
Differentiating both sides of the above equation, we obtain
GATE MASTERS PUBLICATION
z-Transform 315
dX( z)
dz n
= x( n) [ nx( n)] z n 1
dz n dz n
1
= [ nx( n)] z n
z n
dX( z) [ nx( n)] z n Z[ nx( n)]
z =
dz n
Example 7.11
Find the z-transform of g [ n ] = na n u [ n ].
Solution 7.11
1
We know that, z-transform of an u[n] is .
1 az1
1
Thus, an u[n]
1 az1
d 1
na n u[n] z
dz 1 az 1
z( a ) ( 1 z2 ) az1
=
(1 az1 ) (1 az1 )2
az1
Thus, nan u[n]
(1 az1 )2
The poles are at z = a and signal is right sided. Thus ROC is z a .
We can manipulate to get,
a
(n + 1) an + 1 u[n + 1]
(1 az1 )2
1
(n + 1) an u[n + 1]
(1 az1 )2
Since, (n + 1) an u[n + 1] will be zero at n = –1, thus we can say
1
(n + 1) an u[n] ; z a
(1 az1 )2
Example 7.12
Find the z-transform of g [ n ] = n a n .
Solution 7.12
The signal can be written as,
GATE MASTERS PUBLICATION
316 Signals & Systems
n
na , n0
g[n] = n
na , n 0
g[n] = nan u[n] nan u[–n – 1]
az1
The z-transform of nan u[n] ; z a
(1 az1 )2
We can write –na–n u[–n – 1] as –na–n u[–n] because at n = 0 this will be zero.
az1
Thus if x[n] = nan u[n] and X( z)
(1 az1 )
Then, x[–n] X(z–1)
az
Thus, –na–n u[–n]
(1 az )2
1
So, pole is at 1/a and signal is left sided. So ROC is z .
a
z-Transform of g[n] is
1
G(z) = X(z) + X(z–1), a z
a
The condition for existence of G(z) is that a 1.
7.5.6 Time Expansion
The continuous-time concept of time scaling does not directly extend to discrete-time, since the discrete-
time index is defined only for integer values. However, the discrete-time concept of time expansion can be
defined and does play an important role in discrete-time signal and system analysis. Let ‘m’ be a positive
integer, and define the signal,
n
x , if n is a multiple of m
x(m) (n) = m ...(19)
0, if m is not a multiple of m
xm(n) can be obtained from x(n) by placing m – 1 zeros between successive values of the original signal.
Intuitively, we can think of x(m)(n) as a slowed-down version of x(n). Now, if
x(n) X(z) with ROC = R
then, x(m) (n) X(zm) with ROC = R1/m ...(20)
m
That is, R is a z b, then the new ROC is a z b, or a1 / m z b1 / m. Also, if X(z) has a pole
(or zero) at z = a, then X(zm) has a pole (or zero) at z = a1/m.
Proof: The z-transform of x(m) (n) is given by,
n n n
Z[x(m)(n)] = x ( m )( n )z x
m
z
n n
A change of variables is performed by letting r = n/m, which also yields r = – as n = – and r = as
GATE MASTERS PUBLICATION
z-Transform 317
n = . Therefore,
Z[x(m)(n)] = x( r ) z mr x( r ) ( zm ) r X( zm )
r r
Example 7.13
Find the z -transform and ROC of the signal,
n a n / 3 , n 0, 3, 6....
g(n) = an / 3
3 0 , elsewhere
where, a 1.
Solution 7.13
n n
Consider the given signal, g(n) = a n / 3 u x x(3) ( n)
3 3
where, x(n) = an u(n), and its z-transform is given by
1
an u(n) , z a
1 az1
Now, using the time-expansion property, we obtain
x(3) (n) X(z3)
n 1 1/3
an / 3 , z a
3
1 az3
Example 7.14
Find the z -transform and ROC of the signal,
n a n , n 0, 10, 20....
g ( n ) = a nu
10 0, elsewhere
where, a 1.
Solution 7.14
n 10 n /10 n n
Consider the given signal, g (n ) = a n u a u x x(10)( n)
10 10 10
where, x(n) = (a10)n u(n)
and its z-transform is given by
1 10
(a10)n u(n) 10 1
, z a
1 a z
Now, using the time-expansion property, we obtain
x(10)(n) X(z10)
n 1 10 /10
a10 n /10 u 10 10
, z a
10
1 a z
GATE MASTERS PUBLICATION
318 Signals & Systems
n 1
an u 10 10
, z a
10 1 a z
In case of time expansion we have see that if X(z) is z-transform of x[n] then,
n m
y[ n] x X (z )
m
Thus if X(z) has poles at P1, P2 and zeros at Z1, Z2 then Y(z) will have poles at P11 / m , P21 / m and zeros at
Z11 / m, Z12 / m .
(z Z1 ) (z Z 2 )
If, X (z ) =
( z P1 ) ( z P2 )
( z m Z 1 ) (z m Z 2 )
then, Y(z) = X ( z m ) m m
( z P1 ) ( z P2 )
7.5.7 Convolution Property
If, x1(n) X1(z) with ROC = R1
and x2(n) X2(z) with ROC = R2
then, x1(n) x2(n) X1(z) X2(z) with ROC containing R1 R2 ...(21)
The z-transform maps the convolution of two signals into the product of their z-transform.
Proof: The z-transform of x1(n) x2(n) is given by
Z[x1(n) x2(n)] = [x1( n) x2 ( n)] z n
n
= x1( m) x2 ( n m) z n
n m
By interchanging the order of summation, we have
x2 ( n m) z n
n
Z[x1(n) x2(n)] = x1( m)
m
Applying the time-shifting property, the bracketed term is X2(z) z–m.
Substituting this into the above equation yields,
Z[x1(n) x2(n)] = x1( m) ( X2 ( z)z m ) X2 ( z) x1( m) z m
m m
Z[x1(n) x2(n)] = X2(z) X1(z) = X1(z) X2(z)
The convolution property states that convolution in the time domain corresponds to multiplication in
the z-domain. Convolution of the two signals, using z-transform, requires the following steps.
1. Compute the z-transforms of the signal to be convolved, i.e.,
X1(z) = Z[x1(n)]
GATE MASTERS PUBLICATION
z-Transform 319
X2(z) = Z[x2(n)]
2. Multiply the two z-transform, i.e.,
X(z) = X1(z) X2(z)
3. Find the inverse z-transform of X(z), i.e.,
x(n) = Z–1 X(z)]
Here also ROC of X1(z) X2(z) will be R1 R2 only when there is no pole zero cancellation. If pole-zero
cancellation take place then ROC can be greater than this range.
Example 7.15
Evaluate the convolution of a unit step function u ( n ) with itself using z-transform.
Solution 7.15
The convolution of a unit step function u(n) with itself can be expressed as
x(n) = u(n) u(n)
Using the convolution property, we obtain
Z[x(n)] = Z[u(n) Z[u(n)]
2
1 1 1
X(z) = 1
1 z 1 z1 1 z1
The inverse z-transform yields equation,
x(n) = (n + 1) u(n)
Example 7.16
If y [ n ] = x 1[ n ] x2[ n ] then find Y ( z ) and it’s ROC if z -transform of x1 [ n ] and x 2 [ n ] are
( z 5)
X1(z ) = , 7 z 9
( z 7) ( z 9)
( z 7)
and X2(z ) = , 4 z 5
( z 4) ( z 5)
Solution 7.16
Here, Y(z) = X1(z) X2(z)
and pole zero cancellation take place thus ROC will not be equal to R1 R2.
1
Y(z) = , 4 z 9
( z 4) ( z 9)
7.5.8 Correlation Property
If, x1(n) X1(z) with ROC = R1
and x2(n) X2(z) with ROC = R2
then, rx1 x2 (n) = x1(n)*x2(n)
= x 1(m) x 2 (m n ) R x1 x 2 ( z )
m
GATE MASTERS PUBLICATION
320 Signals & Systems
1 1
= X1( z) X2 with ROC containing R1 ...(22)
z R2
Proof: We know that, rx1 x2 (n) = x 1( m ) x 2 ( m n ) x 1( n ) x 2 ( n )
m
Taking the z-transform of the above equation and using the convolution property, we obtain
Z[ rx1 x2 ( m)] = Z[x1(m) x2(–m)
1
Rx1 x2 ( z) = X1( z) X2
z
Example 7.17
Determine the autocorrelation sequence of the signal.
x [ n ] = a n u [ n ], a 1
Solution 7.17
Autocorrelation of x[n] is
rxx = x[n] x[–n]
Thus, Rxx = X(z) X(z–1)
1
X (z ) = , z a
1 az1
1 1
and X(z–1) = , z
1 az a
Here X(z) has pole at a and signal as right sided, and X(z–1) has pole at 1/a and signal is left sided. So we
have chosen ROC accordings.
Rxx = X(z) X(z–1)
1 1 1
= , a z
1 az 1 az a
ROC is only valid when a 1, because if a 1 then ROC will be null.
7.5.9 Accumulation Property
If, x[n] X(z) with ROC = R
n
X( z)
then, y[n] = x( k )
1 z1
with ROC = { R z 1}
k
Proof of this property is done using convolution property. Since,
n 1
y[ n] x(k ) = x[ n] u[ n] X( z)
k 1 z 1
Since ROC of X(z) is R and ROC of u[n] is z 1.
1
ROC of X( z) is { R z 1} .
1 z 1
GATE MASTERS PUBLICATION
z-Transform 321
Example 7.18
Find the z -transform of the signal w[ n ] where,
n
w[ n ] = ak, n0
k n
and w[ n ] = 0 for n < 0. Assume that a 1 .
Solution 7.18
From the given data we can say that,
n
w[n] = ak, n0
k n
1 n 0 n
w[n] = a k a k a k a k 1
k n 0 k n 0
n n
= 2 ak 1 2 a k u[k ] [ n]
k 0 k
If an u[n] is x[n] then w[n] is
n
w[n] = 2 x[k ] [ n]
k
1 2
W(z) = 2 X( z) 1
1 1
1 z (1 az ) (1 z1 )
1
7.5.10 First Difference
If, x(n) X(z) with ROC = R
then, x(n) – x(n – 1) (1 – z–1) X(z) with ROC at least = R { z 0} ...(23)
A common use of this property is in situations where evaluation of the z-transform is easier for the first
difference than for the original sequence.
Proof:
Given that, x(n) X(z)
Using the time-shifting property, we get
x(n – 1) X(z) z–1
Now, using the linearity property, we get
x(n) – x(n – 1) X(z) – X(z) z–1
x(n) – x(n – 1) (1 – z–1) X(z)
7.5.11 Conjugation and Conjugate Symmetry
If, x(n) X(z) with ROC = R
then, x(n ) X(z) with ROC = R ...(24)
Proof: By definition, the z-transform of x(n) is
GATE MASTERS PUBLICATION
322 Signals & Systems
*
n
Z[x(n)] =
x (n )z x ( n ) ( z ) n
n n
= [X(z*)]* = X*(z*)
Case-1: If x [ n ] is real i.e. if,
x[n] = x[n]
Thus, X *( z* ) = X ( z )
That is if X(z) has a pole or zero at Z = Z0 then it will have pole at zero at Z Z0 also.
Study Note
Thus when x[n] is real then its z-transform will have complex poles and zeros in conjugate pairs and real poles and
zeros will be single and not in pairs.
Case-2: If x[n] is real and even i.e. if,
x*[n] = x[n] and x[n] = x[–n]
Thus, X*(z*) = X(z) and X(z) = X(z–1)
This means that complex poles and zeros will be in conjugate pairs and also at inverse location that is if
1 1
pole (or zero) at Z Z 0 re j , , then there will be pole (or zero) at Z Z0 re j , at Z e j , at
Z0 r
1 1
Z e j . Thus a complex pole (or zero) will form a pole constellation (or zero constellation).
Z0 r
1 1 1 1
Z0 re j , Z0 re j , e j , e j
Z0 r Z0 r
We can see that if the pole (or zero) is real then if will be at Z = Z0 and Z = 1/Z0 only.
Case-3: If x[n] is real and odd i.e. if,
x[n] = x[n] and x[n] = –x[–n]
Thus, X ( z ) = X(z) and X(z) = –X(z–1)
Thus, if X(z) has a pole (or zero) at Z = Z0 = rej, then there must be a pole (or zero) at Z Z0 re j ,
1 1 1 1
at Z e j , Z e j . Thus here also we get a pole (or zero) constellation.
Z0 r Z0 r
7.6 z-Transform of Causal Periodic Signals
The z-transform of a causal periodic signal can be determined from the knowledge of the z-transform of
its first cycle (period). Consider a causal periodic signal x(n) with period N0 as shown in Fig. 7.5. Let x1(n),
x2(n), x3(n),...., be this signals representing the 1st, 2nd, 3rd,...., cycles of a causal periodic signal x(n). Therefore,
the causal periodic signal x(n) can be written as
x(n) = x1(n) + x2(n) + x3(n) + x2(n) +.....
GATE MASTERS PUBLICATION
z-Transform 323
x(n)
x 1(n) x 2(n) x 3(n)
n
0 N0 – 1 N0 2N0 – 1
Fig. 7.5 : Causal periodic signal
= x1(n) + x1(n – N0) + x1(n – 2N0) + x1(n – 3N0) +....
= x1( n kN0 )
k 0
Assume x1(n) X1(z). Using the time-shifting property, the z-transform of the above equation becomes,
X(z) = X1( z) Z1( z) z N0 X1( z) z2 N0 X1( z) z3N0 ....
= X1( z) [1 z N0 z2 N0 z3N0 ....]
= X1( z) z mN 0 X1( z) ( z N 0 )m
m0 m0
1
X(z) = X1( z) , z N0 1 z 1
1 z N0
X1( z)
= , z 1 ...(25)/
1 z N0
Example 7.19
Consider the signal shown in figure (a). The signal repeats periodically it a period N 0 = 4
for n 0 and is zero for n < 0. Find the z-transform of this signal along with the ROC.
x(n) x 1(n)
2 2
1 1
n n
0 1 2 3 4 5 6 7 8 9 10 11 12 0 1 2 3
(a) (b)
(a) Causal periodic signal (b) its first cycle
Solution 7.19
The given signal x(n) can be expressed as,
x(n) = x1( n kN0 )
k 0
where N0 = 4, and x1(n) is shown in figure (b), which can be expressed as,
x1(n) = (n – 1) + 2(n – 2) + (n – 3)
Taking the z-transform of the above equation, we obtain
GATE MASTERS PUBLICATION
324 Signals & Systems
X1(z) = z–1 + 2z–2 + z–3
= z–1 [1 + 2z–1 + z–2]
We know that the z-transform of a causal periodic signal x(n) is defined as,
X1( z) z1[1 2z1 z2 ]
X (z ) = , z 1
1 z N0 1 z4
Property Sequence Transform ROC
x [n] X(z) R
x 1[n] X1(z) R1
x 2[n] X2(z) R2
Linearity a 1x 1[n] + a2 x2[n] a 1X1(z) + a 2X2(z) RR1 R2
Time shifting x [n – n0 z–n 0 X(z) RR {0 < z < }
z
Multiplication by z n0 z n0 x [n] X R= z R
z0
j 0 n j 0 n –j 0
Multiplication by e e x [n] X(e z) RR
1 1
Time reversal x [–n] X R
z R
dX( z)
Multiplication by n nx [n] z RR
dz
n
x[ n] 1
Accumulation X( z) RR {z > 1}
k 1 z1
Convolution x 1[n] x 2[n] X1[z] X2(z) RR1 R
Table-7.1 : Some properties of the z-transform
7.7 Inverse of z-Transform
There are three methods that are used for evaluation of inverse z-transform:
1. Contour integration methods (or residue method): This method is not used and will not be discussed.
2. Long division method.
3. Partial fraction method.
7.7.1 Contour Integration Method
The inverse z-transform is formally given by,
1
x(n) = X (z ) z n 1 dz
2j c
...(26)
where the integral is a contour integral over a closed path C that encloses the origin and lies within the
region of convergence of X(z). The contour is any circle of radius R in the region of convergence.
7.7.2 Long Division Method
From the formulae of z-transform we have,
GATE MASTERS PUBLICATION
z-Transform 325
X (z ) = x[ n] z n
n
That is if we expand the expression we will get,
X(z) = .... x[–2] z2 + x[–1] z1 + x[0]
+ x[1] z–1 + x[2] z–2 ..... ...(27)
Thus z-transform can be written as power series expansion where signal value x[n] is coefficient of z–n in
expansion. Here we can see that if the signal is right sided that is it exist for n 0 then it will have negative
power of z and if x[n] is left sided signal that is it exist for n < 0 then it will have positive of z. There are three
cases.
Case-1: When ROC of X ( z ) is entire z -plane except z = 0 or z = .
In this case we can directly see that X(z) can be compared with equation (27) and we can find
x[n]. Here x[n] will be non-zero only for finite number of n.
Case-2: When ROC of X(z) is z a .
This means that X(z) has a pole at finite location and since ROC is of the form z a, x[ n] is
right sided signal and X(z) will have negative power of z2 the X(z) will be of the form.
Numerator N( z)
X (z ) = ...(28)
Denominator D( z)
The N(z) will be divided by D(z) to get consecutive negative powers of z.
Case-3: When ROC of X ( z ) is z a .
This means that X(z) has a pole at finite location and since ROC of the form z a, x[ n] is left
sided signal and X(z) will have positive power of z, the X(z) will be of the form,
Numerator N( z)
X (z ) = ...(29)
Denominator D( z)
The N(z) will be divided by D(z) to get consecutive positive power of z.
REMEMBER • The long division method will not given the exact expression of x[n]. We can see that this
method can be helpful in finding value of x[n] at a particular value of n.
• For example we have X(z) and it’s ROC is given and for any right sided signal we have to find
value of x[n] at n = 3 then we will do long division in X(z) and our answer will be coefficient
of z–3. Similarly for any left sided signal we have to find value of x[n] at n = –3 then we have
to find coefficient of z+3.
Example 7.20
Find the inverse z -transform of
1 z 1
X( z )
1
1 z 1
3
when,
1
( a ) ROC: z .
3
GATE MASTERS PUBLICATION
326 Signals & Systems
1
( b ) ROC: z , using a power series expansion.
3
Solution 7.20
1 z1
Given that, X (z ) =
1
1 z1
3
( a ) Since the ROC is z 1/ 3, we express X(z) as a power series in z–1, so that we obtain a right sided
signal. We divided the numerator by the denominator to obtain,
4 1 4 2 4 3
1 z z z ...
3 9 27
1 1
1 z 1 + z–1
3
1 1
1 z
3
4 1
z
3
4 1 4 2
z z
3 9
4 2
z
9
4 2 4 3
z z
9 27
4 3
z
27
4 4 4 3
We can write, therefore, X(z) = 1 z1 z2 z ....
3 9 27
Comparing with, X(z) = x(0) + x(1) z–1 + x(2) z–2 + x(3) z–3 +....
4 4 4
we obtain, x(0) = 1, x(1) , x(2) , x(3) ,....
3 9 27
or equivalently, we may express x(n) as,
4 4 4
x(n) = 0, , , ,.....
3 9 27
( b ) Since the ROC is z 1/ 3, we express X(z) as a power series in z, so that we obtain a right sided signal.
We divided the numerator by the denominator to obtain,
2
–3 – 12z – 36z +.....
1 –1
z1 1 z + 1
3
z–1 – 3
4
4 – 12z
12z
2
12z – 36z
2
36z
We can write, therefore, X(z) = –3 – 12z – 36z2 + .....
= +(–36) z2 + (–12)z + (–3)
Comparing the above equation with,
X(z) = ....+ x(–2)z2 + x(–1)z + x(0)
GATE MASTERS PUBLICATION
z-Transform 327
we obtain, x(–2) = –36, x(–1) = –12, x(0) = –3
or equivalently, we may express x(n) as,
x(n) = {...., 36, 12, 3}
Example 7.21
Determine the inverse z-transform of
X ( z ) = log(1 + a z –1 ), z a
using the power series expansion.
Solution 7.21
According to the logarithmic series expansion,
1 x2 x3
log(1 + x) = n ( x )n x 2
3
....
n 1
1 ( az1 )2 ( az1 )3
Therefore, X(z) log(1 + az–1) = az ....
2 3
1
= n ( az1 )n , z a
n 1
1
= n ( a )n z n , z a
n 1
1
= ( 1)n 1 a n z n , z a
n 1 n
1 n 1 n
X (z ) = n ( 1) a u( n 1) z n , z a
n
Comparing the above equation with,
X (z ) = x( n) z n
n
1
we get, x(n) = ( 1)n 1 a n u( n 1)
n
Example 7.22
Determine the inverse z -transform of
X ( z ) = e z with ROC all z except z
Solution 7.22
According to the power series expansion of ex, we have
xn
ex =
n0 n!
GATE MASTERS PUBLICATION
328 Signals & Systems
zn 0
z n
Therefore, X(z) = ez =
n0 n! n ( n)!
1
= u( n) z n
n ( n )!
Comparing the above equation with,
X (z ) = x( n) z n
n
1
we get, x(n) = u( n)
( n)!
Example 7.23
Find the inverse z -transform of X ( z ) = sin( z ).
Solution 7.23
According to the power series expansion of sin(x), we have
x3 x5 x7
sin(x) = x .....
3! 5! 7!
z3 z5 z7
Therefore, X(z) = sin(z) = z .....
3! 5! 7!
1 7 1 5 1 3
= .... z z z z
7! 5! 3!
Comparing the above equation with
X(z) = ....+x(–7)z2 + x(–5)z5 + x(–3)z3 + z
1 1
we obtain, x(–7) = , x( 5) ,
7! 5!
1
x(–3) = , x( 1) 1
3!
or equivalently, we may express x(n) as,
1 1 1
x(n) = ..., , 0, , 0, , 0, 1, 0
7! 5! 3!
Study Note
While performing long division method on X(z) there are two different ways in which it will be done. For right sided
signal the long division can be performed by writing N(z) and D(z) of X(z) in decreasing power of z and for left sided
signal N(z) D(z) should be written in increasing power of z.
7.7.3 Partial Fraction Expansion Method
If X(z) in rational function, that is
GATE MASTERS PUBLICATION
z-Transform 329
N ( z ) bM z M bM 1 z M 1....b0
X (z ) = ...(30)
D ( z ) aN z N aN 1 z N 1....a0
Here we have to find the inverse z-transform of X(z), ROC of X(z) is given and according to that x[n] will
be found. First of all we will check that X(z) is proper or not, that is (M < N) or not. If M < N then we can
proceed else we will divide N(z) by D(z) and will write X(z) as,
Remainder
X(z) = Quotient ...(31)
D (z )
Remainder N( z)
Quotient of X(z) can be handled easily and we will proceed with as to get result.
D(z) D( z)
Now we have two cases:
Case-1: Roots of D ( z ) are distinct.
Since roots of D(z) are poles of X(z), so this is the case of distinct poles of X(z).
N( z)
Now we have, X (z ) =
D( z)
X( z) N (z )
we will solve for = z D (z )
z
let the poles of D(z) are P1, P2, P3.....PN thus,
X( z) N (z )
= ...(32)
z z ( z P1 ) ( z P2 ) ( z P3 )......( z PN )
we can write above equation as,
X( z) C C1 C2 CN
= 0 ..... ...(33)
z z z P1 ( z P2 ) ( z PN )
The value of Ck can be found by
X (z )
C0 = z X (z ) z 0 ...(34)
z z 0
X ( z ) ( z Pk )
and Ck = ...(35)
z z Pk
z z CN z
Then we can get, X (z ) = C 0 C 1 C2 ....
z P1 z P2 z PN
C1 C2 CN
= C0 1
1
.... ...(36)
1 P1 z 1 P2 z 1 PN z 1
Each pole in expression of X(z) will correspond to two signals. That is pole at z = Pk:
Ck
• can represent Ck(Pk)n u[n] if ROC of X(z) is z Pk in z-plane.
1 Pk z 1
Ck
• can represent –Ck(Pk)n u[–n – 1] if ROC of X(z) is z Pk in z-plane.
1 Pk z 1
GATE MASTERS PUBLICATION
330 Signals & Systems
REMEMBER Each pole in X(z) will correspond to two signals, if ROC of X(z) is z Pk then it will correspond
to right sided signal. Similarly if ROC of X(z) is z Pk then it will correspond to left sided
signal.
Study Note
X( z)
In above analysis we have X(z) in terms of z and we solved for because then only we can get general
z
Ck
expression of that can be easily converted to n-domain.
1 Pk z 1
If in any question the z-transform is
N( z) b0 b1 z1 ....bM z M
X (z ) = ...(37)
D( z) a0 a1 z1 ....a N z N
Then for M < N the function will be proper, and for M N the function will be improper. For proper
function we can directly move to second step but for improper function we have to perform first step.
The first step is convert improper function into proper.
Remainder
X(z) = Quotient
D (z )
Remainder
are we move to second step with . Now we have a case where D(z) has distinct roots and we
D(z)
will solve X(z) only here.
N (z ) N (z )
X (z ) =
D ( z ) (1 P1z ) (1 P2 z 1 ).....(1 PN z 1 )
1
C1 C2 CN
X(z) = 1
1
.... ...(38)
1 P1z 1 P2 z 1 PN z 1
and Ck = X ( z ) (1 Pk z 1 ) ...(39)
z Pk
Study Note
We can solve for inverse z-transform using any method discussed above. We can change rational function X(z)
written with negative power of z into positive power of z and solve similarly we can change rational function X(z)
written in positive power of z into negative power of z and solve.
Example 7.24
Find the inverse z -transform of
1 z 1 z2
X(z) = with ROC 1 z 2
1
1 2 z1 (1 2 z1 ) (1 z1 )
Solution 7.24
GATE MASTERS PUBLICATION
z-Transform 331
1 z1 z2
Given that, X (z ) =
1 1 1 1
1 z (1 2z ) (1 z )
2
For simplification, we eliminate negative powers of z by multiplying both the numerator and denominator of
the above equation by z3. This result is
z3 z2 z
X (z ) =
1
z ( z 2) ( z 1)
2
X( z) z2 z 1
=
z 1
z ( z 2) ( z 1)
2
We use partial fraction expansion to write,
X( z) A1 A A
= 2 3
z 1 z 2 z 1
z
2
1 X( z)
where, A1 = z 1
2 z z 1 / 2
X( z)
A 2 = ( z 2) 2
z z2
X( z)
A 3 = ( z 1) 2
z z 1
X( z) 1 2 2
Therefore, =
z 1 z 2 z 1
z
2
z 2z 2z 1 2 1
X (z ) =
z
1 z 2 z 1 1 1 z1 1 2 z1 1 z1
2 2
The ROC and the locations of the poles are depicted in figure. The ROC, 1 z 2, is a ring in the z-plane.
The pole of the first term is at z = 1/2. The ROC has a radius greater than this pole, so this pole corresponds
to causal (right sided) signal. Therefore,
n
1 1
u( n) 1
2 1 z 1
2
The second term has a pole at z = 2. Here the ROC has a radius less than this pole, so this pole corresponds
to the anticausal (left sided) signal. Therefore,
2
2(2 n ) u( n 1)
1 2 z 1
The third term has a pole at z = 1. Here the ROC has a radius greater than this pole, so this pole corresponds
to causal (right sided) signal. Therefore,
GATE MASTERS PUBLICATION
332 Signals & Systems
2
2u ( n)
1 z1
Im{z}
0 1/2 1 2 Re{z}
Figure : ROC and location of poles
n
1
and hence, we obtain, x(n) = u( n) 2(2)n u( n 1) 2u ( n)
2
Example 7.25
1
Determine the inverse z -transform of X( z) 1 2
.
1 1.5z 0.5 z
Solution 7.25
Here ROC of X(z) is not given, the poles of X(z) are z = 1 and z = 0.5 that is
1
X (z ) = 1
1 0.5z z1 0.5z2
1 2 1
= 1 1
1
(1 0.5z ) (1 z ) 1 z 1 0.5z1
Since ROC is not given and there are two poles in X(z), to find out all possibilities of ROC we draw circle
of z 0.5 and z 1.0 in z-plane. So we can see that whole plane in divided into three regions. For
or
• R1 = z 0.5 both the poles z = 0.5 and z = 1.0 have radius greater than ROC R1 thus both poles give
left sided signal.
2 Inverse
Thus, 1
z-Transform
2 u[ n 1]
1 z
1
(0.5)n u[ n 1]
1 0.5z1
Thus, x[n] = –2 u[–n – 1] + (0.5)n u[–n – 1]
• Similarly for R2 0.5 z 1.0
Pole z = 0.5 has radius less than ROC and Pole z = 1.0 has radius greater than ROC. Thus z= 0.5
correspond to right sided and z = 1.0 correspond to right sided and
2
2u [ n 1]
1 z1
GATE MASTERS PUBLICATION
z-Transform 333
1
and 1
(0.5)n u [ n]
1 0.5z
x[n] = –2u[n – 1] –(0.5)n u[n]
• Similarly for R3 z 1.0 both the poles have radius less than ROC, so both give right sided signals.
2
Thus, 2u [ n]
1 z1
1
(0.5)n u [ n]
1 0.5z1
x[n] = 2u[n] – (0.5)n u[n]
Case-2: When roots of D ( z ) are not of single order.
That is the poles of X(z) are not of single order, then whole solution is same as that of previous case. That
is we will find X(z)/z for rational function X(z) written in positive of z and we will solve for X(z) only when X(z)
is written negative power of z. The only change will be in finding partial fraction. That is if
X (z ) N ( z ) N (z )
= ...(40)
z zD ( z ) ( z P1 ) ( z P2 )....( z Pi )r ....( z PN )
we assume Pi is rth order pole then,
X (z ) C 0 C1 C2 2 1 r CN
= ...... 1 2
3
..... r
.....
z z z P1 z P2 z Pi ( z Pi ) ( z Pi ) ( z Pi ) ( z PN )
we can see that for all single order poles the expression remain same only for non-single order pole we
have some changes. Here,
C 0 = X( z) Z 0 ...(41)
X (z )
Cn = (z Pk ) ...(42)
z z Pk
1 dk X (z )
r – k = k
( z Pi )r ...(43)
k ! dz z z Pi
For converting this X(z) we will follow same procedure that each pole will correspond to a right sided or
left sided signal depending as ROC.
Example 7.26
Determine the causal signal x ( n ) having the z -transform,
1
X( z )
(1 z ) (1 z1 )2
1
Solution 7.26
GATE MASTERS PUBLICATION
334 Signals & Systems
1
Given that, X (z ) =
(1 z ) (1 z1 )2
1
For simplification, we eliminate negative powers of z by multiplying both the numerator and denominator of
the above equation by z3. This results in
z3
X (z ) =
(1 z) ( z 1)2
X( z) z2
=
z ( z 1) ( z 1)2
We use partial fraction expansion to write,
X( z) A1 A A3
= 2
z z 1 z 1 ( z 1)2
X( z) 1
where, A 1 = ( z 1)
z z 1 4
d 2 X( z) 3
A 2 = dz ( z 1) z 4
z 1
X( z) 1
A 3 = ( z 1)2
z z 1 2
X( z) 1/ 4 3 / 4 1/ 2
Therefore, =
z z 1 z 1 ( z 1)2
1 z 3 z 1 z
X (z ) =
4 z 1 4 z 1 2 ( z 1)2
1 z 3 z 1 z1
=
4 1 z1 4 1 z1 2 (1 z1 )2
Taking the inverse z-transform of the above equation yields,
1 3 1
x(n) = ( 1)n u( n) u( n) nu( n)
4 4 2
1 3 1
= ( 1)n n u( n)
4 4 2
Example 7.27
z3 10 z2 4 z 4
Find the inverse z-transform of X( z) with ROC z 1.
2 z2 2 z 4
Solution 7.27
Since the order of the numerator polynomials is greater than the order of the denominator polynomial, the
given rational z-transform is improper. We use long division to express X(z) as the sum of a proper rational
function and a polynomial in z,
GATE MASTERS PUBLICATION
z-Transform 335
0.5z – 4.5
2 2
2z – 2z – 4 z3 – 10z – 4z + 4
2
z3 – z – 2z
–9z2 – 2z + 4
–9z2 + 9z + 18
–11z – 14
11z 14 5.5 z 7
Thus, we may write, X(z) = 0.5 z 4.5 = 0.5z 4.5
2
2z 2z 4 z2 z 2
5.5z 7
X(z) = 0.5z 4.5
( z 1) ( z 2)
Using the partial fraction expansion to expand the rational function, we obtain
0.5 6
X(z) = 0.5z 4.5
z 1 z 2
0.5z1 6z1
= 0.5z 4.5 1
1 z 1 2z1
0.5z1 6z1
Also, because X(z) = 0.5z 4.5 1
1 z 1 2 z1
The inverse z-transform yields, x(n) = 0.5 (n + 1) – 4.5 (n) – 0.5(–1)n – 1 u(–n) + 6(2)n – 1 u(–n)
Study Note
If in an equation X(z) is given and we have to find number of time domain signals x[n] that will correspond to X(z).
Then simply find poles of X(z) and draw circles for all poles z Pk and find out the number of regions z-plane is
divided. The number of region in which z-plane is divided is answer.
7.8 The System Function of Discrete Time LTI Systems
The LTI systems of discrete time domain are expressed in difference equation format,
N N
a k y[ n k ] = bk x[ n k] ...(44)
k 0 k 0
The z-transform of above equation will be
N M
ak Y (z )z k = bk X (z )z k ...(45)
k 0 k 0
M
bk z k
Y( z) k 0
= N ...(46)
X( z) k
ak z
k 0
The ration of z-transform of output and input is called transfer function of the system.
GATE MASTERS PUBLICATION
336 Signals & Systems
7.8.1 Impulse Response, Transfer Function and Step Response
The impulse response h(n) is defined as the output of an LTI system due to a unit impulse signal input
(n) applied at time n = 0. Also,
h(n) H(z) ...(47)
We know that, the output y[n] of a discrete-time LTI system equals the convolution of the input x[n] with
the impulse response h[n],
y[n] = x[n] h[n] ...(48)
Applying the convolution property of the z-transform, we obtain
Y(z) = X(z) H(z) ...(49)
where Y(z), X(z) and H(z) are the z-transform of y[n], x[n], and h[n], respectively.
Y( z)
Thus, H(z) = ...(50)
X( z)
The z-transform H(z) of h[n] is referred to as the system function (or the transfer function) of the
system.
h[n]
x [n] y [n] = x [n] h[n]
X(z) Y(z) = X(z) H(z)
H(z)
Fig. 7.6 : Impulse response and System function
The step response s(n) is defined as the output of an LTI system due to a unit step input signal, i.e.
x(n) = u(n). The step response of an LTI system with impulse response h(n) is given by
s(n) = h(n) x(n) = h(n) u(n)
= h( m) u( n m)
m
n
s(n ) = h( m )
m
n
Therefore, Z[s(n)] = Z h(m)
m
Using the accumulation property, we obtain
H( z)
S(z ) = ...(51)
1 z 1
n n 1
Conversely, s(n ) = h( m ) h( n ) h( m )
m m
s( n ) = h(n) + s(n – 1)
h(n) = s(n) – s(n – 1)
Z[h(n)] = Z[s(n) – s(n – 1)]
H(z) = [1 – z–1] S(z) ...(52)
GATE MASTERS PUBLICATION
z-Transform 337
REMEMBER • The output of system that is calculated using transfer function H(z) is always output due to
applied input, or forced response.
• The while finding the transfer function H(z) we can assume that all initial conditions are
zero.
Example 7.28
Determine the transfer function and impulse response for the causal LTI system described
by the difference equation,
1
y ( n) y ( n 1) x( n) 2 x( n 1)
2
Solution 7.28
1
Given that, y( n) y( n 1) = x(n) + 2x(n – 1)
2
Taking the z-transform of the difference equation, we obtain
1
Y( z) 1 z1 = X(z) [1 – 2z–1]
2
Hence the system function is
Y( z) 1 2z1 1 2z1
= H( z)
X( z) 1 1 1
1 z1 1 z1 1 z1
2 2 2
The inverse z-transform of the above equation yields,
n n 1
1 1
h(n) = u( n) 2 u( n 1)
2 2
This is the unit impulse response (or unit sample response) of the system.
7.8.2 Characteristics of Discrete-Time LTI System
The characteristics of discrete-time LTI system can be obtained from H(z) and it’s ROC.
1. Causality
For causal system the condition as impulse response is
h[n] = 0 for n < 0 ...(53)
That is h[n] should be a right sided signal, thus ROC of H(z) should be format, ROC z a. Thus
condition for causal system is that H(z) must be a proper function with ROC of form z a.
Study Note
• Here we should have proper function for causality because for example if,
z4
H(z) = , z 1
z 1
GATE MASTERS PUBLICATION
338 Signals & Systems
z3
then, H(z) =
1 z1
thus, h[n] = u[n + 3]
Thus even after ROC is z 1 the system will be non-causal.
• Thus whenever H(z) is improper then even when ROC is of format z 1 the system will be non-causal.
• Thus if any question H(z) is given and we have to find h[n] which is causal, then h[n] should be right sided thus
each pole of H(z) should correspond to right sided signal.
2. Stability
For stable system the condition or impulse response is
h[ n] < ...(54)
n
Since H(z) is z-transform of h[n] then ROC of H(z) is values of r where the condition
h[ n] r n < ...(55)
n
is satisfied. If we compare the above two conditions then we can say that if r = 1 is present in ROC of
H(z) then stability condition will be satisfied. Thus system will be stable only when ROC of H(z) contain
the unit circle.
Study Note
Thus if in any question H(z) is given and we have to find h[n] such that system is stable then we should first of all
draw circles corresponding to each pole (Pk) in z-plane ( z Pk ), and see the option of ROC in z-plane. Then
choose the ROC which contain unit circle. Now write expression of h[n] accordingly.
3. Causal and Stable Systems
For causality and stability we need ROC must contain unit circle and all poles should correspond to
right sided signal. Thus for both the condition to be satisfied together the requirement is that all poles of
H(z) should lie inside unit circle.
Study Note
For causal and stable system first of all the requirement is that all poles of H(z) should remain inside unit circle if
this condition is not satisfied then system cannot be causal and stable. If condition is satisfied then take right sided
signal for each pole and find h(n).
From the above analysis we can see that:
• For anticausal systems i.e. for which h[n] = 0 for n > 0 the ROC of H(z) should be of format z a
and signal should be left sided and H(z) should be proper function.
• For non-causal the system must not follow condition of causal and anticausal system.
• For unstable system ROC of H(z) must not contain unit circle.
GATE MASTERS PUBLICATION
z-Transform 339
Example 7.29
For the following system functions, check whether the corresponding LTI system is causal,
anticausal, or noncausal.
3 4 z1
( a ) H1( z) , z 3
1 3.5 z1 1.5z2
3 4 z 1
( b ) H2 ( z ) , z 0.5
1 3.5 z1 1.5 z2
3 4 z1
( c ) H3 ( z ) , 0.5 z 3
1 3.5z1 1.5 z2
Solution 7.29
( a ) Consider the given system function,
3 4z 1
H1(z) = , z 3
1 3.5z 1 1.5z 2
3 4z1 1 2
= =
(1 0.5z1 ) (1 3z1 ) 1 0.5z1 1 3 z 1
Since the given system function H1(z) is rational and its ROC is the region in the z-plane outside the
outermost pole, the corresponding LTI system is causal. Also it can be verified by its impulse response.
Given that,
1 2
H1(z) = 1
1 0.5z 1 3 z1
Taking the inverse z-transform, we obtain
h1(n) = (0.5)n u(n) + 2(3)n u(n)
Since, h1(n) = 0 for n < 0, this system is causal.
( b ) Consider the given system function,
3 4z1
H2(z) = , z 0.5
1 3.5z1 1.5z2
3 4z1 1 2
= =
(1 0.5z1 ) (1 3z1 ) 1 0.5z1 1 3 z1
Since, the given system function H2(z) is rational and its ROC is the region in the z-plane inside the
innermost pole, the corresponding LTI system is anticausal. Also it can be verified by its impulse
response. Given that,
1 2
H2(z) = 1
1 0.5z 1 3 z 1
Taking the inverse z-transform, we obtain
h2(n) = –(0.5)n u(–n – 1) – 2(3)n u(–n – 1)
Since, h2(n) = 0 for n > 0, this system is anticausal.
( c ) Consider the given system function,
3 4z1
H3(z) = , 0.5 z 3
1 3.5z1 1.5z2
GATE MASTERS PUBLICATION
340 Signals & Systems
3 4z1 1 2
= =
(1 0.5z1 ) (1 3z1 ) 1 0.5z1 1 3 z 1
Since the given system function H3(z) is rational and its ROC is a ring in the z-plane, the corresponding
LTI system is noncausal. In can also be verified by its impulse response. Given that,
1 2
H3(z) = 1
1 0.5z 1 3z1
Taking the inverse z-transform, we obtain
h3(n) = (0.5)n u(n) – 2(3)n u(–n – 1)
Since, h3(t) is two sided, this system is noncausal.
Example 7.30
An LTI system is characterized by the system function,
3 4z1
H(z) =
1 3.5z1 1.5z2
Specify the ROC of H ( z ) and determine h ( n ) for the following conditions:
( a ) The system is causal and unstable.
( b ) The system is noncausal and stable.
( c ) The system is anticausal and unstable.
Solution 7.30
3 4z1 3z2 4z
Given that, H(z) =
1 3.5z1 1.5z2 z2 3.5z 1.5
H( z) 3z 4 3z 4
= 2
z z 3.5z 1.5 ( z 0.5)( z 3)
Using the partial fraction expansion, we obtain
H( z) 1 2
=
z z 0.5 z 3
z 2z 1 2
H(z) =
z 0.5 z 3 1 0.5z1 1 3z1
This system has poles at z = 0.5 and z = 3.
( a ) For this sytem to be causal and unstable, the ROC of H(z) is the region in the z-plane outside the
outermost pole and it must not include the unit circle. Therefore, the ROC is the region, z 3. Since
the ROC, z 3, is the region in the z-plane outside the outermost pole, all the poles correspond to
causal to causal (right sided) signals. Now, consider
1 2
H(z) = 1
, z 3
1 0.5z 1 3 z 1
The inverse z-transform yields,
h(n) = (0.5)n u(n) + 2(3)n u(n)
( b ) For this system to be noncausal and stable, the ROC of H(z) is a ring in the z-plane and it must include
the unit circle. Therefore, the ROC is the region, 0.5 z 3.
GATE MASTERS PUBLICATION
z-Transform 341
The pole of the first term is at 0.5. The ROC has a radius greater than the pole at z = 0.5, so this pole
corresponds to causal (right sided) signal and pole at z = 3.0 correspond to ledt sided signal. Therefore,
we obtain
h(n) = (0.5)n u(n) – 2(3)n u(–n – 1)
( c ) For this system to be anticausal and untable, the ROC of H(z) is the region in the z-plane inside the
innermost pole and it must not include the unit circle. Therefore, the ROC is the region z 0.5. Since
the ROC, z 0.5, is the region in the z-plane inside the innermost pole, all the poles correspond to
anticausal (right sided) signals. Now, consider
1 2
H(z) = 1
, z 0.5
1 0.5z 1 3 z 1
The inverse z-transform yields,
h(n) = (0.5)n u(–n – 1) – 2(3)n u(–n – 1)
Example 7.31
3(1 z1 )
Find the inverse z-transform of X( z ) , assuming that (a) the signal is causal
1 2.5 z1 z2
and (b) the signal has a DTFT, i.e., x ( n ) is absolutely summable.
Solution 7.31
3(1 z1 ) 3( z2 z)
Given that, X (z ) = 1 2
2
1 2.5z z z 2.5z 1
X( z) 3(1 z) 3( z 1)
= 2
z z 2.5z 1 ( z 0.5) ( z 2)
Using the partial fraction expansion, we obtain
X( z) 1 2
=
z z 0.5 z 2
z 2z 1 2
X (z ) =
z 0.5 z 2 1 0.5 z1 1 2 z1
The system has poles at z = 0.5 and z = 2.
( a ) For x(n) to be causal, the ROC of X(z) is the region in the z-plane outside the outermost pole, i.e.,
z 2. Since the ROC, z 2, is the region in the z-plane outside the outermost pole, all the poles
correspond to causal (right sided) signals.
Now, consider
1 2
X (z ) = 1
, z 2
1 0.5z 1 2 z1
The inverse z-transform yields,
x(n) = (0.5)n u(n) + 2(2)n u(n)
( b ) For x(n) to be Fourier transformable, i.e. for x(n) to be absolutely summable, the ROC of X(z) must
include the unit circle. Therefore, its ROC is the region 0.5 z 2. The pole of the first term is at 0.5.
The ROC has a radius greater than the pole at z = 0.5, so this pole corresponds to causal (right sided)
GATE MASTERS PUBLICATION
342 Signals & Systems
signal. Therefore,
1
(0.5)n u(n)
1 0.5z1
The second term has a pole at z = 2. The ROC has a radius less than the pole at z = 2, so this pole
corresponds to the anticausal (left sided) signal. Therefore,
2
–2(2)n u(–n – 1)
1 2 z1
and hence, we obtain
x(n) = (0.5)n u(n) + 2(2)n u(–n – 1)
7.8.3 Response of Discrete Time LTI System to Complex Exponential Signals
We have seen that z-transform of e j0 n , sin0n, cos0n cannot be found because these signals cannot be
made absolutely summable for any value of r, or these signal never satisfy the condition,
x[ n] r n
n
Similarly we can say that, a signal Cn for – < n < have no z-transform (where C is any constant).
Now consider case where Cn is applied as input to discrete LTI system.
n LTI system
C y [n]
h[n]
Fig. 7.7 : Discrete LTI system with Cn as input signal
Here, we cannot apply, Y(z) = H(z) X(z)
because we cannot find X(z). So we apply the general rule formulae of discrete LTI system,
y[n] = h[n] x[n]
That is, y[n] = h[ m] x[ n m]
m
y[n] = h[ m] C( n m)
m
y[n] = Cn h[ m] C m ...(56)
m
We know that, H(z) = h[n] z n ...(57)
n
Thus if we compare equation (56) and (57) we get that,
y[n] = C n H (z ) ...(58)
z C
Thus when input is Cn then output is input multiplied by value of transfer function H (z ) z C .
GATE MASTERS PUBLICATION
z-Transform 343
Note: Cn is called eigen function and H ( z ) z C is called eigen value for discrete time LTI system.
That is when input is e j0 n.
j 0 n LTI system
e e j0 n H( z) ...(59)
H(z) z e j0
Similarly if input is e j0 n then,
–j 0 n LTI system
e
H(z)
e j0 n H( z) ...(60)
z e j 0
If h[n] is real then H(z) = H(z) i.e. if h[n] is real then,
H( z) z e j0 = H( e j0 )
and H( z) z e j0 = H( e j0 )
will have relation, H( e j0 ) = H*( e j0 )
Thus if, H( e j0 ) = H( e j0 ) e j ,
j
that is it has magnitude H( e 0 ) and phase () then,
j j
H( e j0 ) = H ( e 0 ) H( e 0 ) e j
using above equation (60) and (59) and assuming h[n] real we get
e j0 n e j0 n LTI system
cos[0 n]
2 H[z]
e j0 n H( z) z e j0 e j0 n H( z) z e j0
Here the output will be .
2
e j0 n H( e j0 ) e j0 n H( e j0 )
y[n] =
2
e j0 n H( e j0 ) e j e j0 n H( e j0 ) e j
=
2
j
y[n] = H( e 0 ) cos( 0 n )
= H( e j0 ) cos(0 n)
REMEMBER • Similarly we can say that if h[n] is real then
j
1. for sin(0n + ) as input, output will be H( e 0 ) sin(0 n ) .
2. for cos(0n + ) as input, output will be H( e j0 ) cos(0 n ) .
• If it is not specified in question that h[n] is real or not and H(z) is given in question and we
have to find output of e j0 , cos(0 n ), sin(0 n ) then assume h[n] is real.
GATE MASTERS PUBLICATION
344 Signals & Systems
7.8.4 Type of System
There are two types of systems:
1 . Finite Impulse Response (FIR): If impulse response h[n] of any filter or system is non-zero only
for finite values of n then it is called FIR filter. Since FIR filters have h[n] a finite duration signal
thus H(z) the z-transform of h[n]) that is transfer function of system will have no poles at finite
obtain.
2 . Infinite Impulse Response (IIR): These are the systems or filters which have h[n] or impulse
response of infinite duration that is H(z) (the z-transform of h[n]) transfer function of system will
have pole at finite location.
7.9 Unilateral (One-sided) z-Transform
The unilateral z-transform of a signal x(n) is defined as,
Zu[x(n)] = X( z) x( n) z n ...(61)
n0
We shall denote the relationship between x(n) and X(z) as,
Zu
x( n) X( z) ...(62)
The difference in the definitions of the unilateral and bilateral z-transform lies in the lower limit of the
summation. The bilateral z-transform involves a summation from – to +, while the unilateral z-transform
involves a summation from 0 to +. Due to this choice of lower limit, the unilateral z-transform has the following
characteristics:
1. Two signals that differ for n < 0, but that are identical for n 0, will have different bilateral z-transforms,
but identical unilateral z-transform.
2. Any signal x(n) that is identically zero for n < 0 [i.e., the signal x(n) u(n)] has identical bilateral and
unilateral z-transform.
3. The unilateral z-transform is used in analysing causal systems and, particularly, systems specified
by linear constant-coefficient difference equations with non-zero initial conditions (i.e., systems
that are not initially at rest).
4. For a unilateral z-transform, there is a unique inverse transform of X(z); consequently, there is no
need to specify the ROC explicitly (ROC for a unilateral z-transform must always the exterior of a
circle). For this reason, we shall generally ignore any mention of the ROC for unilateral z-transform.
Method of find unilateral z-transform of any signal is
• to use equation (61) directly on the signal to get result.
• or another method is to find unilateral z-transform of x[n] first of all find a new signal x[n] =
x[n] u[n] and then find bilateral z-transform of x[n] to get unilateral z-transform of x[n].
Example 7.32
Find unilateral z-transform of the signal, x ( n ) = a n + 1 u ( n + 1).
Solution 7.32
By definition, Zu[x(n)] = X(z) = x( n) z n a n 1 u( n 1) z n
n0 n0
GATE MASTERS PUBLICATION
z-Transform 345
= a n 1 z n a a n z n = a (az1 )n
n0 n0 n0
a
Zu[x(n)] = X(z) =
1 az1
Example 7.33
Determine the unilateral z -transform of each of the following signals:
( a ) x ( n ) = a n u ( n + 1)
( b ) x ( n ) = ( n + 1) + ( n ) + a n + 3 u ( n + 1)
Solution 7.33
( a ) We have, Zu[x(n)] = X(z) = x( n) z n a n u( n 1) z n
n0 n0
1
= a n z n ( az1 )n =
n0 n0 1 az1
1
Zu[x(n)] = X(z) =
1 az 1
( b ) We have, Zu[x(n)] = X(z) = x( n) z n [( n 1) ( n) a n 3 u( n 1)] z n
n0 n0
= ( n 1) z1 ( n) z n a n 3 u( n 1) z n
n0 n0 n0
= 0 z n a n 3 z n
n0
n0
= 1 a
3
( az1 )n
n0
a3
X (z ) = 1
1 az1
7.9.1 Properties of Unilateral z-Transform
Mostly all the properties of bilateral z-transform are valid in unilateral z-transform except few like Time
reversal and Time shifting if shifting take the signal towards n < 0.
1. Linearity
Zu Zu
If, x1( n) X1( z) and x2 ( n) X2 ( z)
Zu
then, ax1( n) bx2 aX1( z) bX2 ( z) ...(63)
GATE MASTERS PUBLICATION
346 Signals & Systems
2. Scaling in The z-Domain
Zu
If, x( n) X( z)
Zu z
then, z0n x( n) X ...(64)
z0
3. Differentiation in z-Domain
Zu
If, x( n) X( z)
Zu dX( z)
then, nx( n) z ...(64)
dz
4. Time Expansion
Let m be a positive integer, and define the signal,
n
x , if n is a multiple of m
x(m)(n) = m ...(66)
0, if n is not a multiple of m
xm(n) can be obtained from x(n) by placing m – 1 zeros between successive values of the original signal.
Intuitively, we can think of x(m)(n) as a slowed down version of x(n). Now, if
Zu
x( n) X( z)
Zu
then, x( m)( n) X( zm ) ...(67)
5. Conjugation Property
Zu
If, x( n) X( z)
Zu
then, x ( n) X ( z ) ...(68)
6. Convolution Property
If, x1(n) = x2(n) = 0 for all n < 0, and if
Zu
x1( n) X1( z)
Zu
and x2 ( n) X2 ( z)
Zu
then, x1( n) x2 ( n) X1( z) X2 ( z) ...(69)
Study Note
It is important to note that the convolution property for unilateral z-transform applies only if the signals x1(n) and
x2(n) are both zero for n < 0. That is, while we have seen that the bilateral z-transform of x1(n) * x2(n) always
equals to the product of the bilateral z-transform of x1(n) and x2(n), the unilateral z-transform of x1(n) * x2(n) in
general does not equal to the product of the unilateral transforms if either x1(n) or x2(n) is non zero for n < 0.
GATE MASTERS PUBLICATION
z-Transform 347
7. Accumulation Property
If, x(n) = 0 for n < 0 and if
Zu X( z)
then, x(k )
1 z1 ...(70)
k 0
Proof: We know that if x(n) = 0 for n < 0, then
n
x (k ) = x(n) u(n)
k 0
Taking the unilateral z-transform on both sides of the above equation, we obtain
Zu x( k ) = Z [x(n) u(n)]
k 0 u
X( z)
=
1 z1
n
Zu X (z )
Therefore, x ( k )
1 z 1
k 0
8. Time Delay Property or Right Shift of Signal
If x[n] is a signal whose unilateral z-transform is X(z) that is
Zu
x[ n] X( z)
Thus, X (z ) = x[n] z n
n0
Here we assumed that x[n] = 0 for n < 0 and while calculating X(z) the unilateral z-transform of x[n] we
have neglected all values of x[n] for n < 0.
Suppose we give a right shift to x[n] by 1 and then find its unilateral z-transform that is,
Zu[x[n – 1] = x[ n 1] z n = x [1] z 0 x [n 1] z n
n0 n 1
= x [1] ( x [m] z m ) z 1
m 0
= x[–1] + X(z) z–1
Now let us understand what is happening: To find X(z) we have considered x[n] for n 0 only, when x[n]
is shifted by 1 in right side then all value of x[n] shift by 1 and at n = 0 we get x[–1] and when we find
Zu[x[n – 1] then we get a new extra term x[–1] at n = 0 and rest all previous terms shifted by 1 whose Zu
will be X(z) z–1.
General expression will be
Z[x[n – n0]] = x[ n n0 ] z n
n0
By changing variable we get,
GATE MASTERS PUBLICATION
348 Signals & Systems
Z[x[n – n0]] = x[ m] z( m n0 )
m n0
1
= z n0 x[ m] z m x[ m] z m
m n0 m0
1
= z n0 x[ n] zn z n0 X( z) ...(71)
n 1
Zu
Thus, x[ n 2] z2 X( z) x[1] z1 x[2]
Zu
and x[ n 3] z3 X( z) x[1] z2 x[2] z1 x[3]
Pictorially if we look at this then it is easier to understand,
x[n] [..., x [2], x [1], x [0], x [1], x [2]...]
for Zu [x(n)]
Similarly, x[n – 1] [..., x [3], x [2], x [1], x [0], x [1]...]
for Zu [x(n 1)]
so we get whole x[n] shift by 1 right and new term at n = 0.
So, Zu[x[n – 1] = x[–1] z0 + z–1 X(z)
9. Time Advance Property or Left Shift Property
Zu
If, x[ n] X( z)
then, X (z ) = x[n] z n
n0
If, x[n] = [..., x [2], x [1], x [0], x [1], x [2]...]
for Zu [x(n)]
then, x[n + 1] = [..., x [1], x [0], x [1], x [2], x [3]...]
for Zu [x(n 1)]
So, we can see that, Zu[x[n + 1] has one less than term than X(z) because x[0] has gone to n = –1. Thus
we can say that,
Zu[x[n + 1]] = z X (z ) z x (0)
Because whole Because x [0] which is at
signal shift left by 1 n = 1 has to be removed
Similarly, x[n + 2] = [..., x [0], x [1], x [2], x [3], x [4]...]
for Z u [x(n 2)]
This two terms has to be removed,
Zu[x[n + 2] = {z2 X(z) – z2 x[0] – z–1 x[1]}
GATE MASTERS PUBLICATION
z-Transform 349
General formulae is,
n0 1
Zu
x( n n0 ) zn0 X( z) zn0 x( n) z n , n0 0 ...(72)
n0
By definition, we have
Zu[x(n + n0)] = x( n n0 ) z n
n0
A change of variables is performed by letting m = n + n0, which also yields n = m – n0, m = n0 as n =
0, and m = as n = . Therefore,
Zu[x(n + n0)] = x ( m ) z ( m n0 ) z n0 x (m )z m
m n0 m n0
n0 1 n0 1
n m m m
= z 0 x( m) z x( m) z x( m) z
m n0 m 0 m 0
n0 1
n m m
= z 0 x( m) z x( m) z
m 0 m 0
n0 1
n
= z 0 X( z) x( n) z n
n0
n0 1
Zu[x(n + n0)] = zn0 X( z) zn0 x( n) z n
n0
10. First Difference
If x(n) is causal [i.e., x(n) = 0 for n < 0], and if
Zu
x( n) X( z)
Zu
then, x( n) x( n 1) (1 z1 ) X( z) ...(73)
Proof: Given that,
Zu
x( n) X( z)
Using the time-shifting property, we get
Zu
x( n 1) z1 X( z)
Now, using the linearity property, we get
Zu Zu
x( n) x( n 1) X( z) z1 X( z) (1 z1 ) X( z)
11. Initial Value Theorem
If x(n) is causal [i.e., x(n) = 0 for n < 0]
then, x(0) = Lim X( z) ...(74)
z
Proof: By definition, we have
GATE MASTERS PUBLICATION
350 Signals & Systems
X (z ) = x( n) z n
n
Since, x(n) = 0 for n < 0, we obtain
X (z ) = x( n) z n
n0
X(z) = x(0) + x(1) z–1 + x(2) z–2 + ....
Lim X( z) = Lim [ x(0) x(1) z1 x(2) z2 ....] x(0)
z z
Study Note
• Initial value theorem is only applicable for causal sequence.
• In equation (74) X(z) is bilateral z-transform of x[n]. Since sequence is causal bilateral and unilateral z-transform
same.
• For a causal sequence, if x(0) is finite, then Lim X( z) is finite. Consequently, with X(z) expressed as a ratio of
z
polynomials in z, the order of the numerator polynomial cannot be greater than the order of the denominator
polynomial; or, equivalently, the number of finite zeros X(z) cannot be greater than the number of finite poles.
12. Final Value Theorem
The final value theorem states that if Lim x( n) exists and if
n
Zu
x( n) X( z)
then, Lim x( n) = x( ) Lim (1 z1 ) X( n) ...(75)
n z 1
Remember: The limit in equation (75) exists if all the poles of (1 – z–1) X(z) lie inside the unit circle, i.e., all the
poles of (1 – z–1) X(z) have magnitude less than one.
Proof: Let’s take the unilateral z-transform of the signal x(n) – x(n – 1).
Zu[x(n) – x(n – 1)] = [x( n) x( n 1)] z n
n0
N
Zu[x(n)] – Zu[x(n – 1)] = Lim [ x( n) x( n 1)] z n
N n 0
N
X(z) – [z–1 X(z) + x(–1)] = Lim [ x( n) x( n 1)] z n
N n 0
N
Lim [(1 z1 ) X( z) x( 1)] = Lim Lim [x( n) x( n 1)] z n
z 1 z 1 N
n 0
N
Lim (1 z1 ) X( z) x( 1) = Lim [x( n) x( n 1)]
z 1 N n 0
= Lim [ x(0) x( 1) [ x(1) x(0) [x(2) x(1) ....
N
GATE MASTERS PUBLICATION
z-Transform 351
+ [x(N – 2) –x(N – 3)] + [x(N – 1) – x(N – 2) + [x(N) – x(N – 1)
Lim (1 z1 ) X( z) x( 1) = Lim [ x( 1) x( N )]
z 1 N
Lim (1 z1 ) X( z) x( 1) = x( 1) Lim x( N )
z 1 N
Lim(1 z1 ) X( z) = Lim x( N ) x( )
z 1 N
Example 7.34
Find the final value of the signals corresponding to the following z -transforms:
1 z 1
( a ) X 1( z )
1 0.25 z 2
2 z 1
( b ) X2 ( z )
1 1.8 z1 0.8 z2
1
( c ) X3 ( z ) 1
1 2z 3 z2
Solution 7.34
1 z1
( a ) Given that, X 1( z ) =
1 0.25z1
(1 z1 ) (1 z1 )
(1 – z–1) X1(z) =
(1 0.5z1 ) (1 0.5z1 )
It has two poles at z = 0.5 and z = –0.5. Note that both the poles of (1 – z–1) X1(z) lie inside the unit
circle the unit circle, therefore,
x1() = Lim (1 z1 ) X1( z)
z 1
(1 z1 ) (1 z1 ) (0) (2)
= Lim 1 1
0
z 1 (1 0.5z ) (1 0.5z ) (0.5) (1.5)
2z1
( b ) Given that, X 2( z ) =
1 1.8z1 0.8z2
2 z1 (1 z1 ) 2 z1
(1 – z–1) X2(z) = 1 1
(1 z ) (1 1.8z ) 1 0.8z1
It has one pole at z = 0.8, which lies inside the unit circle, therefore,
x2() = Lim (1 z1 ) X2 ( z)
z 1
2 z1 2
= Lim 1
10
z 1 1 0.8z 0.2
1
( c ) Given that, X 3( z ) =
1 2 z 1 3 z 2
(1 z1 ) 1
(1 – z–1) X3(z) =
(1 z1 ) (1 3z1 ) 1 3z1
GATE MASTERS PUBLICATION
352 Signals & Systems
It has one pole at z = –3, which lies outside the unit circle therefore, the final value theorem cannot be
used to find x3().
Example 7.35
Use the initial value theorem to find the initial value of the signals corresponding to the
following z-transforms:
2 z1
( a ) X1( z)
(1 z1 ) (1 0.5 z1 )
1 3 z1
( b ) X2 ( z )
(1 0.1 z1 ) (1 0.6 z1 )
0.5 0.25 z1
( c ) X3 ( z)
1 1.3 z1 0.2 z2 0.1 z3
Solution 7.35
2 z1 2 1/ z
( a ) Given that, X 1( z ) = 1 1
=
(1 z ) (1 0.5z ) 1 0.5
1 1
z z
Using the initial value theorem, we have
2 1/ z 20
x1(0) Lim = Lim
z z 1 0.5 (1 0) (1 0)
1 1
z z
x1(0) = 2
1 3z1 1 3 / z
( b ) Given that, X 2( z ) = 1 1
(1 0.1 z ) (1 0.6 z ) 0.1 0.6
1 1
z z
Using the initial value theorem, we have
1 3 / z 1 0
x2 (0) Lim = Lim = (1 0) (1 0)
z z 0.1 0.6
1 1
z z
x2(0) = 1
( c ) Using the initial value theorem, we have
0.5(0.25 / z) 0.5 0
x 3 (0) = Lim
z
1.3 0.2 0.1 1 0 0 0
1
z z2 z3
x3(0) = 0.5
REMEMBER If in a question we have to find steady state value of output then find Y(z) = H(z) X(z) and then
apply final value theorem on Y(z) and find y(). The value of y() will be valid only when all
poles of (1 – z–1) Y(z) be inside unit circle.
GATE MASTERS PUBLICATION
z-Transform 353
7.10 Solving Difference Equation using Unilateral z-Transform
One of the most common applications of the unilateral z-transform is to solve linear constant-coefficient
difference equations subject to nonzero initial conditions. Such equations are used to model discrete-time LTI
systems. Solving of these equations depends on the time-delay and time advance properties of the unilateral
z-transform. As in the case of the Laplace transform with differential equations, the z-transform converts difference
equations into algebraic equations that are radially solved to find the solution in the z-domain. The procedure is
straight forward and systematic, and we summarizes it in the following three steps:
1. For a given set of initial conditions, take the z-transform of both sides of the difference equation to
obtain an algebraic equation in Y(z).
2. Solve the algebraic equation for Y(z).
3. Take the inverse z-transform to obtain y(n).
Example 7.36
A discrete-time system is characterized by the difference equation:
1 1
y( n)
y ( n 1) x( n ) x( n 1)
2 2
Find y ( n ) for n 0 when x ( n ) = u ( n ) and the initial conditions is given as y (–1) = 1.
Or, determine the step response of the system,
1 1
y( n)
y ( n 1) x( n ) x( n 1)
2 2
when the initial condition is y (–1) = 1.
Solution 7.36
For x(n) = u(n), we have
1
X (z ) =
1 z1
Now, consider the given difference equation,
1 1 1
y( n) y( n 1) = x( n) x( n 1) u( n) u( n 1)
2 2 2
Taking the unilateral z-transform of the above equation, we obtain
1 1
Zu y( n) y ( n 1) = Zu u( n) u ( n 1)
2 2
1 1 1 / 2 z1
Y( z) [ z1 Y( z) y( 1)] =
2 1 z1 1 z1
1 1
1 z
1 1 2
Y( z) 1 z1 y( 1) =
2 2 1 z1
1
1
1 2
Y(z) =
1 z1 1
1 1
z
2
The inverse z-transform of the above equation yields,
GATE MASTERS PUBLICATION
354 Signals & Systems
n 1 n 1
1 1
y(n) = u( n) u( n) 1 u( n)
2 2 2
Example 7.37
A second-order discrete-time system is characterized by the difference equation,
y (n ) – 3y (n – 1) + 2y (n – 2) = x (n ) – 2x ( n – 1)
Find y ( n ) for n 0 when x ( n ) = u ( n ) and the initial conditions are given as
y (–1) = y (–2) = 1.
Solution 7.37
1
For x(n) = u(n), we have X (z ) =
1 z 1
Now, consider the given difference equation,
y(n) – 3y(n – 1) + 2y(n – 2) = x(n) – 2x(n – 1) = 2u(n) – u(n – 1)
Taking the unilateral z-transform of the above equation, we obtain
Zu[y(n) – 3y(n – 1) – 2y(n – 2)] = Zu[u(n) – 2u(n – 1)]
1 2z1
Y(z) – 3[z–1 Y(z) + y(–1) + 2[z–2 Y(z) + z–1 y(–1) + y(–2)] =
1 z1 1 z1
1 2z1
Y(z) [1 – 3z–1 + 2z–2] –3y(–1) + 2z–1 y(–1) + 2y(–2)] =
1 z1
2 z1
Y(z) [1 – z–1) (1 – 2z–1)] – 3 + 2z–1 + 2 =
1 z1
1 2 z1 1 2 z1
Y(z) =
(1 z1 )2 (1 2z1 ) (1 z1 ) (1 2z1 )
1 1
= 1 2
(1 z ) 1 z 1
The inverse z-transform of the above equation yields,
y(n) = (n + 1) u(n) + u(n) = (n + 2) u(n)
7.10.1 Zero Input Response and Zero State Response
The z-transform method for solving difference equations offers a clear separation between the zero
input response of the system to initial conditions and the zero state response of the system associated with the
input. The zero input response represents the system output when the input is zero, i.e., x(n) = 0, and thus it
is the result of internal system conditions (such as energy storage, initial conditions) alone. In contrast, the zero
state response represents the system response to the external input x(n) when the initial conditions are zero.
The total response is defined as,
Total response = Zero-input response + Zero-state response ...(76)
(natural response) (forced response)
Natural Response and Forced Response: In total response, the response due to the poles of the
system is called natural response of the system. The response due to the poles of the input is called the forced
GATE MASTERS PUBLICATION
z-Transform 355
response of the system.
Example 7.38
Use the unilateral z-transform to determine the output of a system represented by the
difference equation,
5 1
y ( n 2)
y ( n 1) y ( n) 5x ( n 1) x( n)
6 6
in response to the input x ( n ) = u ( n ). Assume that the initial conditions on the system are
y (–1) = 2 and y (–2) = 0. Identify the forced response (zero-state response) y f ( n ) of the
system and the natural response (zero-input response) y n ( n ).
Solution 7.38
For, x(n) = u(n), x(–1) = x(–2) = 0
1
and X (z ) =
1 z1
Now, consider the given difference equation,
5 1
y( n 2) y( n 1) y( n) = 5x(n + 1) – x(n)
6 6
Because the given difference equation is in time advance operator form, the use of the left-shift property
may seem appropriate for its solution. This property require a knowledge of auxiliary conditions y(0),
y(1),....y(N – 1) rather than of the initial conditions y(–1), y(–2),....y(–n), which are generally given. This
difficulty can be overcome by expressing the given difference equation in delay operator form (obtained by
replacing n with (n – 2) and then using the right-shift property. The given difference equation in delay
operator form is
5 1
y( n) y( n 1) y( n 2) = 5x(n – 1) – x(n – 2)
6 6
Taking the unilateral z-transform of the above equation, we obtain
5 1
Y( z) [z1 Y( z) y( 1)] [ z2 Y( z) z1 y( 1) y( 2) ]
6 6
= 5[z–1 X(z) + x(–1)] – [z–2 X(z) + z–1 x(–1) + x(–2)]
5 1
Y( z) [z1 Y( z) 2] [ z2 Y( z) 2 z1 0]
6 6
= 5[z–1 X(z) + 0] – [z–2 X(z) + 0 + 0]
5 1 5 1
Y( z) 1 z1 z2 z1 = 5z–1 X(z) – z2 X(z)
6 6 3 3
5 1 5 1 1
Y( z) 1 z1 z2 = z X( z) [5z1 z2 ]
6 6 3 3
Input terms
Initial condition terms
1/ 3(5 z1 ) z1 (5 z1 )
Y(z) =
5 1 5 1
1 z1 z2 (1 z1 ) 1 z1 z2
6 6 6
6
zero-input component zero-state response
GATE MASTERS PUBLICATION
356 Signals & Systems
1 / 3(5 z 1
) z 1
(5 z 1
)
Y(z) =
1 1 1 1 1 1 1 1 1
1 z 1 z (1 z ) 1 z 1 z
3 2 3 2
Using partial fraction, expansion, we obtain
3 4 / 3 12 18 6
Y(z) = 1
1
1
1
1
1 z
1
1 z1 1 z 1 z1 1 z1
2 3 2 3
The inverse z-transform of this equation yields,
1 n 4 1
n 1
n
1
n
y(n) = 3 u( n) u( n) 12 u( n) 18 u( n) 6 u( n)
2
3 3 2 3
1 n 4 1 n 1
n
1
n
y(n) = 3 u( n) 12 18 6 u( n)
2 3 3 2 3
zero-input response or natural response zero-state response or forced response
y(n) = yzi(n) + yzs(n)
where the zero-input response is
1 n 4 1 n
yzi(n) = 3 u( n)
2 3 3
and the zero-state response is
n n
1 1
yzs(n) = 12 18 6 u( n)
2 3
The poles of the given system are located at p1 = 1/2 and p2 = 1/3. The response y(n) can be separated into
two parts: the natural response yn(n), and the forced response yf(n).
1 n 41
n 1
n
1
n
y(n) = 3 u( n) u( n) 12u ( n) 18 u( n) 6 u( n)
2
3 3 2
3
1 n 41
n
1
n
1
n
= 3 u( n) u( n) 18 u( n) 6 u( n) 12 u( n)
2 33 2 3
n n
1 14 1
y(n) = 15 u( n) 12 u( n)
2 3 3
Forced response = y f (n)
Natural response = y n (n)
7.11 Block Diagram Representation
A linear system can be characterized by its transfer function H(z). Fig. 7.8 (a) shows a block diagram of
GATE MASTERS PUBLICATION
z-Transform 357
a system with a transfer function H(z) and its input and output X(z) and Y(z), respectively. Subsystems may be
interconnected by using cascade (series), parallel, and feedback interconnections, the three elementary types.
7.11.1 Cascade Interconnection
When the systems are connected in series, as depicted in Fig. 7.8 (b), then the transfer function of the
overall system is the product of the individual transfer functions. This result can also be proved by observing
that in Fig. 7.8 (b).
Y( z) W( z) Y( z)
H(z) = H1( z) H2 ( z)
X( z) X( z) W( z)
We can extend this result to any number of systems connected in cascaded. It follows from this discussion
that the systems connected in cascade can be interchanged without affecting the overall transfer function. This
commutation property of LTI systems follows from the commutative and associative property of convolution.
X(z) H(z) Y(z)
(a)
W(z) Y(z)
X(z) H1(z) H2(z) = X(z) H(z) = H1(z) H2(z) Y(z)
(b)
X(z) H1(z)
H1(z)
X(z) = X(z) H(z) = H1(z) + H2(z) Y(z)
Y(z)
H2(z)
X(z) H2(z)
(c)
+ E(z)
X(z) H1(z) Y(z)
–
H1( z)
= X(z) H1( z) Y(z)
1 H1( z) H2 ( z)
H2(z)
(d)
Fig. 7.8 : Elementary connection of block diagram
7.11.2 Parallel Interconnection
When the systems are connected in parallel, as depicted in Fig. 7.8 (c), then the transfer function of the
overall system is the sum of the individual transfer functions. This result can also be proved by observing that in
Fig. 7.8 (c).
Y(z) = X(z) H1(z) + X(z) H2(z)
Y(z) = X(z) [H1(z) + H2(z)]
GATE MASTERS PUBLICATION
358 Signals & Systems
Y( z)
= H(z) = H1(z) + H2(z)
X( z)
We can extend this result to any number of systems connected in parallel.
7.11.3 Feedback Interconnection
When the output is fed back to the input, as shown in Fig. 7.8 (d), the overall transfer function H(z) =
Y(z)/X(z) can be computer as follows. The input to the adder are X(z) and –H2(z) Y(z). Therefore, E(z), the
output of the adder, is
E(z) = X(z) – H2(z) Y(z)
but, Y(z) = H1(z) E(z)
= H1(z) [X(z) – H2(z) Y(z)]
Y(z) = H1(z) X(z) – H1(z) H2(z) Y(z)
H(z) [1 + H1(z) H2(z)] = H1(z) X(z)
Y( z) H1( z)
= H( z) ...(76)
X( z) 1 H1( z) H2 ( z)
Therefore the feedback loop can be replaced by a single block with the transfer function shown in
equation (76). The transfer function H1(z) is the forward path transfer function, and H2(z) is the feedback path
transfer function.
7.12 System Realization
We now develop a systematic method for realization (or implementation) of an arbitrary Nth order
transfer function. The most general Nth order transfer function is given by
b0 zN b1 zN 1 ... bN 1 z bN
H(z) =
zN a1 zN 1 ... a N 1 z a N
7.12.1 Direct Form 1 Realization
Consider for simplicity a third-order system characterized by a transfer function:
b0 z3 b1 z2 ... b2 z b3 b0 (b1 / z ) (b2 / z 2 ) (b3 / z 3 )
H(z) = =
z3 a1 z2 a2 z a3 a a a
1 1 22 33
z z z
We can express H(z) as,
b1 b2 b3 1
H(z) = b0 2 3
z (1 ( a1 / z) ( a2 / z ) ( a3 / z )
z
z 2 3
H1( z ) H2 ( z )
We can realize H(z) as a cascade of transfer function H1(z) followed by H2(z), as depicted in Fig. 7.9 (a),
where
W( z) b b b
H1(z) = b0 1 22 33 ...(77)
X( z) z z z
GATE MASTERS PUBLICATION
z-Transform 359
b1 b2 b3
W(z) = b0 2 3 X( z) ...(78)
z z z
Y( z) 1
and H2(z) = ...(79)
W( z) 1 ( a1 / z) ( a 2 / z2 ) ( a3 / z3 )
a1 a2 a3
Y(z) = W( z) Y( z) ...(80)
z z2 z3
We shall first realize H1(z) given in equation (71). Equation (72) shows that the output W(z) can be
synthesized by adding by adding the input b0X(z) to b1X(z)/z, b2(X(z)/z2), and b3(X(z)/(z3). Because the transfer
function of a unit delay is z–1 = 1/z, the signals X(z)/z, X(z)/z2, and X(z)/z3 can be obtained by a successive delay
of the input x(n). The left-half section of Fig. 7.9 (b) shows the realization of H1(z).
W(z)
X(z) H1(z) H2(z) Y(z)
(a)
b0 W(z)
X(z) Y(z)
z–1 z–1
b1 –a 1
z–1 z–1
b2 –a 2
z–1 z–1
b3 –a 3
(b)
Fig. 7.9 : (a) Realization of a transfer function in two steps
and (b) direct form-I realization of a third-order discrete-time LTI system
We next consider the realization of H2(z) given by equation (79). Equation (80) shows that the output
Y(z) can be synthesized by subtracting a1(Y(z)/z), a2(Y(z)/z2), and a3(Y(z)/z3) from W(z). To obtain signals
(Y(z)/z2), and (Y(z)/z2), and (Y(z)/z3), we assume that the already have the desired output Y(z). Successive
delay of Y(z) yields the needed signals Y(z)/z, Y(z)/z2 and Y(z)/z3. the right-half section of Fig. 7.9 (b) shows the
realization of H2(z). We can generalize this procedure, known as the direct form I realization, for any value of N.
This procedure requires 2N delay units to realize and Nth-order transfer function. This realization is noncanonic
since it employs six delay units to implement a third-order transfer function. A realization is canonic if the
number of delay units used in the realization is equal to the order of the transfer function realized. Thus canonic
realization has no redundant delay units.
7.12.2 Direct Form-2 Realization (Canonic Form Realization)
In the direct form I, we realize H(z) by implementing H1(z) followed by H2(z), as shown in Fig. 7.9 (a).
GATE MASTERS PUBLICATION
360 Signals & Systems
We can also realize H(z), as shown in Fig. 7.10 (a), where H2(z) is followed by H1(z). This procedure is known
as the direct form II realization.
The direct form I realization [Fig. 7.9 (b)] implements zeros first [the left-half section represented by
H1(z)] followed by realization of poles [the right-half section represented by H2(z)] of H(z). In contrast, the
direct form II realization implements poles first followed by zeros.
Again consider for simplicity a third-order system characterized by a transfer function:
b0 z3 b1 z2 b2 z b3
H(z) =
z3 a1 z2 a2 z a3
b0 (b1 / z ) (b2 / z 2 ) (b3 / z 3 )
H(z) =
a a a
1 1 22 33
z z z
We can express H(z) as,
1 b1 b2 b3
H(z) = 2 3 b0 z 2 3
(1 ( a1 / z) ( a2 / z ) ( a3 / z )
z z
H2 ( z ) H1( z )
We can realize H(z) as a cascade of transfer function H2(z) followed by H1(z), as depicted in Fig. 7.10 (a),
where
V( z ) 1
H2(z) = ...(81)
X( z) (1 ( a1 / z) ( a2 / z2 ) ( a3 / z3 )
a1 a2 a3
V(z) = X( z) V( z) ...(82)
z z2 z3
Y( z) b b b
and H1(z) = b0 1 22 33 ...(83)
V( z) z z z
b1 b2 b3
Y(z) = b0 2 3 V( z) ...(84)
z z z
The left-half section of Fig. 7.10 (b) shows the realization of H2(z) and the right-half section shows the
realization of H1(z).
We observe that in Fig. 7.10 (b) the signal variables at nodes 1 and 1 are the same, and hence, the two
top delay units can be shared. Likewise, the signal variables at nodes 2 and 2 are the same, which permits the
sharing of the two middle delay units. Following the same argument, we can share the delay units leading to the
final structure shown in Fig. 7.10 (c)
This implementation halves the number of delay units to N, and thus more efficient in hardware utilization.
This is the direct form II realization. This realization is canonic since it employs N delay units to implement a
Nth-order transfer function.
GATE MASTERS PUBLICATION
z-Transform 361
V(z) b0
X(z) Y(z)
z–1 z–1
–a 1 b1
1 1
W(z)
X(z) H2(z) H1(z) Y(z)
z–1 z–1
(a)
–a 2 b2
2 2
z–1 z–1
–a 3 3 3 b3
(b)
V(z) b0
X(z) Y(z)
z–1
–a 1 b1
z–1
–a 2 b2
z–1
–a 3 b3
(c)
Fig. 7.10 : Direct form II realization of a third-order LTI system
Example 7.39
Let x ( n ) be an absolutely summable signal with rational z -transform X ( z ). If X ( z ) is known to
have a pole at z = 1/2, could x ( n ) be
( a ) a finite-duration signal? ( b ) a left-sided signal?
( c ) a right-sided signal? ( d ) a two-sided signal?
Solution 7.39
( a ) No. we know that for a finite-length signal, the ROC is the entire z-plane. Therefore, there can be no
poles in the finite z-plane for a finite-length signal.
( b ) No. since the signal is absolutely summable, the ROC must include the unit circle. Since it is given that
the signal has a pole at z = 1/2, the ROC can never be the region z 1/ 2. Therefore, the signal cannot
be left-sided.
( c ) Yes. since the signal is absolutely summable, the ROC must include the unit circle. Since it is given
GATE MASTERS PUBLICATION
362 Signals & Systems
that the signal has a pole at z = 1/2, a valid ROC would be the region z 1/ 2. Therefore, the signal
could be right-sided.
( d ) Yes. since the signal is absolutely summable, the ROC must include the unit circle. Clearly, we may
define the ROC which is a ring in the z-plane add includes the unit circle. Therefore, the signal could
be two-sided.
Example 7.41
Suppose we are given the following five facts about a particular LTI system with impulse
response h ( n ) and z -transform H ( z ):
1 . h(n) is real.
2 . h(n ) is right-sided.
3. Lim H( z) 1
z
4 . H ( z ) has two zeros.
5 . H ( z ) has one of its poles at a non-real location on the circle defined by z 3 / 4.
Answer the following two questions:
( a ) Is this system causal?
( b ) Is this system stable.
Solution 7.41
( a ) Since Lim H( z) 1 , H(z) has no poles in infinity. Furthermore, since h(n) is given to be right-sided,
z
h(n) has to be causal.
( b ) Since Lim H( z) 1 ,and h(n) is causal, the numerator and denominator polynomials of H(z) have the
z
same order. Since H(z) has two zeros, we may conclude that it also has two poles. Since h(n) is real,
the poles must occur in conjugate pairs. Also, it is given that one of the poles lies on the circle defined
by z 3 / 4. Therefore, the other pole also lies on the same circle. Clearly, the ROC for H(z) is z 3 / 4,
and will include the unit circle. Therefore, we may conclude that the system is stable.
Example 7.41
A causal signal x ( n ) has the z -transform,
X ( z ) = sin( z –1 ) (1 + 3 z –2 + 2 z –4 )
Find: x (11).
Solution 7.41
Given that, X(z) = sin(z–1) (1 + 3z–2 + 2z–4)
z3 z5 z7 z9 z11
n
= z1 2 4
(1 3 z 2 z ) = x( n) z
3! 5! 7! 9! 11! n0
Since x(11) is simply the coefficient in front of z–11 in this power series expansion of X(z), therefore
1 3 2
x(11) =
11! 9! 7!
GATE MASTERS PUBLICATION
z-Transform 363
Example 7.42
Find the inverse z -transform of
2 z 2 3 z 4
X( z) , z 0
z2 4 z 3
Solution 7.42
We see that, X(z) can be written as,
X(z) = (2z–1 + z–3 + 3z–5) X1(z)
z
where, X 1( z ) =
z2 4 z 3
Thus, if x1[n] X1(z)
then by the linearity property and the time-shifting property, we get
x[n] = 2x1[n – 1] + x1[n – 3] + 3x1[n – 5]
X1( z) 1 1 c c
Now, = 2 1 2
z z 4z 3 ( z 1) ( z 3) z 1 z 3
1 1
where, c1 =
z 3 z 1 2
1 1
c2 =
z 1 z 3 2
1 z 1 z
Then, X 1( z ) =
2 z 1 2 z 3
Since the ROC of X1(z) is z 0, z 0 x1[n] is a right sided sequence, and we get
1
x1[n] = [( 1)n ( 3)n u[ n]
2
Thus, we get
x[n] = [(–1)n – 1 –(–3)n – 1] u[n – 1]
1 3
[( 1)n 3 ( 3)n 3 ] u[ n 3] [( 1)n 5 ] ( 3)n 5 ] u[ n 5]
2 2
Example 7.43
n
1
The output y [n ] of a discrete-time LTI system is found to be 2 u[ n] when the input x [ n] is
3
u [ n ].
( a ) Find the impulse response h[n] of the system.
n
1
( b ) Find the output y [ n ] when the input x[ n ] is u[ n].
2
Solution 7.43
z
(a) x[n] = u[ n] X( z) ; z 1
z 1
GATE MASTERS PUBLICATION
364 Signals & Systems
n
1 2z 1
y[n] = 2 u[ n] Y( z) , z
3 1 3
z
3
Hence, the system function H(z) is
Y( z) 2( z 1) 1
H(z) = ; z
X( z) 1 3
z
3
Using partial-function expansion, we have
H( z) 2( z 1) c1 c
= 2
z 1 z z 1
z z
3 3
2( z 1)
where, c1 = 6
1
z
3 z 1 / 2
2( z 1)
c2 = 4
z z 1 / 3
z 1
Thus, H(z) = 6 4 , z
1 3
z
3
Taking the inverse z-transform of H(z), we obtain
n
1
h[n] = 6 [ n] 4 u[ n]
3
n
1 z 1
(b) x[n] = u[ n] X( z) ; z
2 1 2
z
2
Then, Y(z) = X(z) H(z)
2 z ( z 1) 1
= , z
1 1 2
z z
2 3
Again by partial fraction expansion we have,
Y( z) 2 z ( z 1) c c
= 1 2
z 1 1 1 1
z z z z z 3
2 3
2( z 1)
where, c1 = 6
1
z
3 z 1 / 2
2( z 1)
c2 = 8
1
z
3 z 1 / 3
GATE MASTERS PUBLICATION
z-Transform 365
z z 1
Thus, Y(z) = 6 8 ; z
1 1 2
z z
2 3
Taking the inverse z-transform of Y(z) we obtain,
n n
y(n) = 6 1 8 1 u[ n]
2 3
Example 7.46
Find the steady state response of the discrete time system described by
0.1 z
H( z)
z 0.5
When excited by x [ n ] = cos(0.2 n + 10°) with discrete frequency = 0.2.
Solution 7.46
For the given input, x[n] = cos 0.2n
j
Response, y[n] = H (e ) cos(1n )
0.1 e j 0.2
H( e j ) =
2
e j 0.2 0.5
0.111.3 0.111.3
= 0.098 j 0.0196 0.5 0.402 j 0.0196
0.111.3
= 0.2514.1
0.402.8
Therefore, y[n] = 0.25 cos[0.2n + 10° + 14.1°]
= 0.25 cos[0.2n + 24.1°]
GATE MASTERS PUBLICATION
366 Signals & Systems
1 8
(c) X ( z ) , z
64 2 9
1 z
81
1 Subjective Practice Problems
6 . Let,
1
y[n] = u[ n]
9
n
1 . Let, x[n] = (–1)n u[n] + an u[–n – n0] Determine two distinct signals such that each has
Determine the constraints on the complex number z-transform X(z) which satisfies both of the following
a and the integer n0, given that the ROC of X(z) is conditions:
1 z 2. [ X ( z ) X ( z )]
1. Y (z 2 )
2
2 . Consider the signal, 2. X(z) has only one pole and only one zero in the
z-plane.
1 n
cos , n 0 7 . Consider the following system functions for stable
x[n] = 3 4
LTI systems. Without utilizing the inverse
0, n 0 z-transform, determine in each case whether or not
Determine the poles and ROC for X(z). the corresponding system is causal.
4 1 1 2
3 . Suppose that the algebraic expression for the z z
1
(a) 3 2
z-transform of x[n] is 1 1
z1 1 z1 1 z1
1 2 2 3
z
1
X( z) 4
1 2 5 1 3 2 1
z
1 z 1 z z 2
4 4 4 (b)
1 3
How many different regions of convergence could z2 z
2 16
correspond to X(z)?
z 1
4 . Let x[n] be a signal whose rational z-transform X(z) (c)
4 1 2 2 3
contains a pole at z = 1/2. Given that, z z z
3 2 3
n
1
x1[n] = x[ n] 8 . Consider a causal LTI system whose input x[n] and
4
output y[n] are related through the block diagram
is absolutely summable and
representation shown in figure.
n
1
x2[n] = x[ n] x[n] y [n]
8
is not absolutely summable, determine whether x[n]
is left sided, right sided, or two sided.
z–1
5 . Determine, for each of the following z-transforms,
whether the corresponding signal has an
2/3 –6
approximately lowpass, bandpass, or highpass
characteristic:
z–1
z1 8
(a) X( z) , z
8 9
1 z1 –1/9 8
9
8 1 (a) Determine a difference equation relating y[n]
1 z 8
(b) X( z) 9 , z and x[n].
16 1 64 2 9 (b) Is this system stable?
1 z z
9 91
GATE MASTERS PUBLICATION
z-Transform 367
9 . Determine the unilateral z-transform of each of the 5 6
following signals, and specify the corresponding 1 1
z z
regions of convergence. 4 4
(c) 6
(d) 5
1 1
n z z
1 2 2
```(a) x1[ n] u[ n 5]
4
14. The following is known about a discrete-time LTI
(b) x2[n] = [n + 3] +[n] + 2n u[–n]
system with input x[n] and output y[n]:
n
1 1. If x[n] = (–2)n for all n, then y[n]= 0 for all n.
(c) x3 [ n]
2 2. If x[n] = (1/2)n u[n] for all n, then y[n] for all n
is of the form
10. A right sided sequence x[n] has z-transform,
n
1
3z 10
z 7
5z 2
4z 1
1 y[n] = [ n] a u[ n],
X( z) 4
10 7 3
z 5z z
where a is a constant.
Determine x[n] for n < 0. (a) Determine the value of the constant a.
11. We are given the following five facts about a discrete (b) Determine the response y[n] if the input x[n] is
time signal x[n] with z-transform X(z): x[n] = 1, for all n
1. x[n] is real and right sided. 15. When the input to an LTI system is
2. X(z) has exactly two poles. n
3. X(z) has two zeros at the origin. 1
x( n) u( n) (2)n u( n 1)
3
1 j / 3
4. X(z) has a pole at z e . and the corresponding output is
2
n n
8 1 2
5. X(1) . y( n) 5 u( n) 5 u( n)
3 3
3
Determine X(z) and specify its region of convergence. (a) Find the system function H(z) of the system. Plot
the poles and zeros of H(z) and indicate the ROC.
12. A causal LTI system is described by the difference
(b) Find the impulse response h(n) of the system.
equation,
(c) Write a difference equation that is satisfied by
y[n] = y[n – 1] + y[n – 2] + x[n – 1]
the given input and output.
(a) Find the system function H(z) = Y(z)/X(z) for
(d) Is the system stable? Is it causal?
this system. Plot the poles and zeros of H(z) and
indicate the region of convergence. 2
(b) Find the unit sample response of the system. 16. Determine the inverse transform of X( z) ez with
(c) You should have found the system to be unstable.
ROC all z except z .
Find a stable (noncausal) unit sample response
that satisfies the difference equation.
17. Find the z-transform of the digital signal obtained by
13. Determine which of the following z-transforms could sampling the analog signal x(t) = e–4t sin(4t) u(t) at
be the transfer function of a discrete-time linear intervals of T = 0.1 seconds.
system that is not necessarily stable, but for which
18. Using long division, determine the inverse z-
the unit sample response is zero for n < 0. State
transform of
your reasons clearly.
1 2z1
(1 z1 )2 ( z 1)2 X( z)
(a) (b) 1 2 z1 z2
1 1
1 z1 z if (a) x(n) is causal and (b) x(n) is anticausal.
2 2
GATE MASTERS PUBLICATION
368 Signals & Systems
n
20. (a ) When input 1 is
x( n) u( n) u( n) 21. Computer the unit step response s(n) of the system
2
with impulse response
applied to a linear causal time invariant system, the
output is 3n , n0
n n h( n) 2 n
1 1 , n 0
y( n) 6 u( n) 6 u( n) 5
4 3
Find the transfer function of the system.
(b) What is the difference equation representation
of the system?
GATE MASTERS PUBLICATION
You might also like
- 12-Revision of Z-Transform With Numericals-09-11-2023No ratings yet12-Revision of Z-Transform With Numericals-09-11-202351 pages
- Unit 2 - Digital Signal Processing - WWW - Rgpvnotes.inNo ratings yetUnit 2 - Digital Signal Processing - WWW - Rgpvnotes.in13 pages
- Z - Transform: Presented By-: Mohammad RameezNo ratings yetZ - Transform: Presented By-: Mohammad Rameez42 pages
- N U A N X: Definition of The Z-TransformNo ratings yetN U A N X: Definition of The Z-Transform4 pages
- Chapter Three:Z Transform and Its Application To LTI SystemsNo ratings yetChapter Three:Z Transform and Its Application To LTI Systems52 pages
- Jatiya Kabi Kazi Nazrul Islam University: Assignment OnNo ratings yetJatiya Kabi Kazi Nazrul Islam University: Assignment On27 pages
- The Z-Transform: Electrical & Electronics Engineering Department Eastern Visayas State UniversityNo ratings yetThe Z-Transform: Electrical & Electronics Engineering Department Eastern Visayas State University39 pages
- Module 5 - Z Transform & Inverse Transform - Lecture Notes PDFNo ratings yetModule 5 - Z Transform & Inverse Transform - Lecture Notes PDF104 pages
- Biosignals & Biosystems: Block 2. The Z-TransformNo ratings yetBiosignals & Biosystems: Block 2. The Z-Transform69 pages
- Z - Transform and Its Application To The Analysis of Lti SystemNo ratings yetZ - Transform and Its Application To The Analysis of Lti System41 pages
- SSC Mts Ex: Studymaterialfor GenralawarenessNo ratings yetSSC Mts Ex: Studymaterialfor Genralawareness5 pages
- SSC Mts Ex: Studymaterialfor GenralawarenessNo ratings yetSSC Mts Ex: Studymaterialfor Genralawareness3 pages
- SSC Mts Ex: Studymaterialfor GenralawarenessNo ratings yetSSC Mts Ex: Studymaterialfor Genralawareness8 pages
- 1070uf 06-12-24 Ee Ec SignalsystemssolutionsNo ratings yet1070uf 06-12-24 Ee Ec Signalsystemssolutions19 pages
- Semiclassical Methods For Hawking Radiation From A Vaidya Black HoleNo ratings yetSemiclassical Methods For Hawking Radiation From A Vaidya Black Hole12 pages
- Partial Differential Equations For Scientists andNo ratings yetPartial Differential Equations For Scientists and7 pages
- Takashi Okamoto - Introduction To WormholesNo ratings yetTakashi Okamoto - Introduction To Wormholes14 pages
- CENG 131: Temperature-Entropy (TS) Diagram of NitrogenNo ratings yetCENG 131: Temperature-Entropy (TS) Diagram of Nitrogen6 pages
- The Use of Control Valve Sizing Equations With Simulation Based Process DataNo ratings yetThe Use of Control Valve Sizing Equations With Simulation Based Process Data14 pages
- Differential Equation and Area Under CurveNo ratings yetDifferential Equation and Area Under Curve26 pages
- (Graduate Studies in Mathematics) Jennifer Schultens - Introduction To 3-Manifolds. 151-American Mathematical Society (2014)100% (2)(Graduate Studies in Mathematics) Jennifer Schultens - Introduction To 3-Manifolds. 151-American Mathematical Society (2014)300 pages
- Modern Physics II Homework #3: Professor Richard DelzeneroNo ratings yetModern Physics II Homework #3: Professor Richard Delzenero5 pages
- Complex Variables & Transforms (20A54302) : Lecture NotesNo ratings yetComplex Variables & Transforms (20A54302) : Lecture Notes263 pages