The Z-Transform

3.1 INTRODUCTION
CHAPTER CONTENTS
We know that transform techniques are important
lntroduction tools in the analysis of signals and linear time-
The Z-transform invariant systems. In this chapter, we shall introduce
One sided and bilateral z-transform the z-transform. We will discuss different properties
of z-transform and its importance in the analysis of
Region of Convergence (ROC)
linear time-invariant (L[I) systems.
Z-transform pf some Common Signals
It may be observed that the Laplace transform is
Properties of Z-Transform
an extension of the continuous-time Fourier transform
Summary of the Properties of because of the fact that the Laplace transform may
z-transform be applied to a broader class of signals than the
The lnverse Z-Transform Fourier transform. Just for instance, there are several
One Sided Z-transform signals for which the Fourier transform does not
s-plane to Z-plane mapping converge but the Laplace transform converges. The
Laplace transform may also be applied to perform
transform analysis of unstable systems.
In fact, the z-transform plays the same role in
the analysis of discrete-time signals and linear time-
invariant (LTI) systems as the Laplace transform
plays in the analysis of continuous-time signals and
linear-time invariant (LTI) systems. This means that
the z-transform is the discrete-time counterpart of the
Laplace transform.

s., THE Z-IRANSFORM

Let us consider a discrete-time linear time-invariant
(LTI) systems with impulse response h(n). If the input
to this discrete-time linear time invariant
(LTI) system is a complex exponential signal of the
form zn, then the response or output y(n) ofthe system
is expressed as
y(n) = H(z) zn ...(3.1)

where H(z) _s h(n) z-" ...(3.2)

158
 Now, for the case z = do with ro real, we have I zl = \ and the summation in equation (3.2)
results in discrete-time Fourier transform of h(n).
More generally, if I z I is not restrict to unity, the summation is then called as the z-transform of
h(n).
In general, the z-transform of a discrete-time signal x(n) is expressed as the power series

X(z) =n=-6! *(n)r-' ...(3.3)

Here, z is a complex variable.

The expression in equation (3.3) is sometimes called the direct z-transform due to the fact that
it transforms the time-domain signal x(n) into its complex-plane representation X(z).
Also, the inverse procedure {i.e.,to obtain x(n) from X(z)} is known as the inverse z-transforrn.
In general, the z-transform ofa discrete-time signal x(n) is expressed as
X(z) = Z{x (n)} ...(3.4)
whereas inverse z-transform is expressed as under:
x(n) = Z-r IyiO)I ...(3.5)
It may be noted that the z-transform is an infrnite power series, it exists only for those values of
z for which this s'eries converges. The region of convergence (ROC) of X(z) is the set of all the values
of z for which X(z) attains a finite value.
In last, evaluating X(z) at the complex number
z =retogives

s
Hence, from above expression, it is clear that if r = 1, ,h;;; z-transform evaluated on the unit
circle produces the discrete-time Fourier transform of the signal x(n).

3.2.1 Advantages- of z-transform

1. Discrete time signals and LTI systems can be completely characterized using Z-transform.
2. The stability of LTI system can be determined using Z-transform.
3. Mathematical calculations are reduced using Z-transform. For example convolution operation
is transformed into simple multiplication operation.
4. By calculating Z-transform of given signal, DFT and FT can be determined.
5. Entire family of digital frlters can be obtained from one proto-type design using Z-transform.
6. The solution of differential equations can be simplifred using Z-transform.

3.3 ONE SIDED AND BILATERAL Z-TRANSFORM

There are two types of z-transform:
1. Single sided z-transform
2. Double sided z-transform (Bilateral)
1. Single sided z-transform
Defrnition: A single sided z-transform of discrete time signal x(n) may be defined as under:

X(z)= !x(n)z-"
n=0
...(3.6)
' Here, "2" is acomplexvariable. In equation (3.6), limits of summation are from 0 to - . Therefore,
while expanding the summation we shall put only positive values of n (from n = 0 to n = -). Hence,
this is single sided or one-sided z-transform.
E
2. Double sided z-transform
Defrnition: A double sided z-transform time signal x(n) may be defined as under:

X(z)=n ! "(n)'-" ...(3.7)

=-6
While expanding the summation, we shall put both positive and negative values of 'n'. Thus this
is double-sided z-transform.
What is z-domain?
Observe equations (3.6) and (3.7). Here, we are transforming discrete
time sequences x(n) into X(z). We know that discrete time sequence x(n) is
drawn simply by plotting amplitude versus 'n'. This is known as discrete
domain. In case of z-transform, we are plotting real part of z on X-axis and
imaginary part of z on Y-axis as shown in frgure 3.1. This is known as
z-domain. Fig. 3.1. z-domain
How to denote Z-transfom?
The relationship between x(n) and X(z) is indicated as follows:
x(n) e-Z--+ X(z) ...(3.8)

3.4 REGION OF CONVERGENCE (ROC)* Jt\rru, Hyderbad, sem. Exam.,2007-08)

We know that z-transform is expressed as under:

Z{x(n)} : x(n)z-'
=X(z)=n=-6 ...(3.e)

We may write, complex number z = r d@

Substituting z = r'd', equation (3.9) becomes

X@)l^rr= rer. = X (r elo) = I x(n) r-' e-jon ...(3.10)

The above expression is the discrete-time Fourier transform of the modified discrete-time signal
{x(n) r-n}.
Now,ifr=Lthen lzl=I
In this case, X(z) (equation 3.10) reduces to its Fourier transform.
Hence, the expression in equation (3.10) will converge if {x(n) ,-n; is absolutely summable.
Mathematically, we have
DOYOU KNOW?
i l*,",.-"1 .- ...(3.11) The z-transform is a genera-
lization of the DTFT with similar
Hence, for x(n) to be finite, the magnitude of its advantages.
z-transform, X(z) must also be finite.
* Explain the following terms iD brief with properties :
(il LTI system (iil ROC in z-transform (iii) Stability condition for LTI system.
Hyderabad, Sem. Exam., 2007-08)
 Therefore, the set of values of z in the z-plane for which the magnitude of X(z) is finite, is called
the Region of Convergence (ROC).

This means that convergence of i l*r,,,,

n =0'
-"1 guarantees the convergence ofthe expression
t"
l,(z)

X(z) ) x(n)z-n
= n=0 poleatz=1
Therefore, the condition for z-transformX(z) Zero atz -- O

bo be finite is lz | > 1.
In other words, the Region ofConvergence
(ROC) for X(z) is the area outside the unit circle
in the z-plane. Also, it may be noted that the
region of convergence (ROC) of a rational z-
transform is bounded by the location of its poles. Un1 circte
As an example, the z-transform of the
discrete-time unit-step function u (n) is given as
z tlt;3e'; Pole-zero plot and region of convergence
X(z) = z-7-. This expression has a zero at z = 0 ol the unit-step discrete-time signal u[n]

and a pole at z = L and the region of convergence (ROC) is z

I I
Oliver Heaviside (1850-1925)
> 1 as shown in figure 3.2.

3.4.1 The Region of Convergence (ROC)

for the z-transform : Few Properties
In this section, we shall discuss some important
properties of the region of convergence (ROC) for the z-
transform.
Property 1.:
The region of convergence (ROC) of X(z) consists of a
ring or circle in the z-plane centered about the origin.
Property 2 : Although Laplace published his
The region ofconvergence (ROC) does not contain any transform method to solve
pole. differential equations in 1779, the
Property 3 : method did not catch on until a
The z-transform X(z) converges uniformly if and only if century later. It was rediscovered
the region of convergence (ROC) of the z-transform X(z) independently in a rather awkward
of the given discrete-time signal, includes the unit circle. form by an eccentric British engineer,
The ROC ofX(z) consists of a ring in the z-plane centred Oliver Heaviside (1850-1925), one of
about the origin. This means that the ROC of the z- the tragic figures in the history of
transform of x(n) has values of z for which {x(n) r-"} is science and engineering. Despite his
always absolutely summable i.e. prolific contributions to electrical
engineering, he was severely
t/-/l lx(n)r-"1< - ...(3.12) criticized during his lifetime and was
neglected later to the point that
I

Property 4: hardly a textbook today mentions his

If the discrete-time signal x(n) is of finite duration, then name or credits him with
the region of convergence (ROC) will be the entire z- contributions.
plane except z- 0 and.z- *.
'Property 5 :
If discrete-time signal x(n) is a righ ;ided sequence, then the region of convergence (ROC) will
not include infinity (*).
Property 6 i
If the discrete-time signal x(n) is a left-sided sequence, then the region of convergence (ROC)
willnotincludez=0.Butinacaseifx(n)=0foralln>0,thentheROCwillincludez=0.
Property 7:
If the discrete-time signal x(n) is a two-sided sequence, and if the circle I zl = ro is in the ROC,
then the ROC will consist of a ring in the z-plane that includes the circle I , | = .0. This means
that the ROC will include the intersection of the ROC's of the components.
Property 8 _:

If the z-transform X(z) is rational, then the ROC will extend to infinity. This means that the
ROC will be bounded by poles.
Property 9:
If the discrete-time signal x(n) is causal, then the ROC will include z = -.
Property 10:
If the discretq-time signal x(n) is anti-causal, then the ROC will includes z=0

3.4.2 Region of Convergence (ROC) for a two-sided z-transform*

We know that the Z-transform of a discrete-time,signal x(n) is expressed as
Zlx(n)l =X(z)

= | x(n) z-n ...(3.13)

Here z is a complex variable.

The above expression is called a two-sided z-transform. The expression in equation (3.13) may
be written as

Z{x(n)} =X(z)= j x(n)z-o+ ! x(n)z-'

i x(n)z-' =n=_6
n=_€ n=0
Z {x(n)} = X(z)

! x(-n) ," * n=0

= n=1 Z x(n) z-" ...(3.14)

From above exp- sion, it may be observed that the first series is a non-causal sequence which
convergesfor lzl <
The second serie is a causal sequence which converge for lzl > rr.
Hence, the series represented by the expression

Z{x(n)} =X(z)= i x(n)z-n

converges for r, < .

lrl rzn"ora"U ,r-. .r.
Table 3.1. shows finite-duration causal, anti-causal and two-sided signals with their corresponding
ROCs.

Frrr-"7tiu"6ril'r*"n-n"itrr?frp"?G r."y *il"ir"*I"Ifo.Tti

system analysis in the tiansform domain."rr"p"rp".it-io"-."It-o-"g"""ity-"
 Table 3.1. Finite-duration causal, anti-causal and two-sided signals
with their ROCs

S.No. Signale ROC's

1.

2. Anti-Causal

3. TWo-sided

ROC :entire
z-plane
excePt z = 0
Z=a

Table 3.2. shows infinite-duration causal, anti-causal and two-sided signals with their
corresponding ROCs.
 Table 3.2. lnfinite-duration causal, anticausal and
two-sided signals with their ROCs

S.No. Signals ROC's

1.

I Anti-Causal

3.

Table 3.3. Shows some important z-transform pairs and corresponding ROCs

S.No. Discrete-time Z-transfom X(z) ROC

Sigual x(n)
1 6(n) 1 All Z-plane
2 6(n-k) z-k lzl>0,k>0
lzl <-,k<0
1 z
....................-_
3 u (n)
l-z-r - ,-l lrl >1
1 z
4 -u(-n-1) 1-z-' ,-L lrl.l
-=
 S.No. Discrete-time Z-transf,om X(z) ROC
Signal xlnl
z-1 z
5. n u(n)
a-;ry = @-r)2 lrl>L
lz
6. at u(n)
r-;7 = a-^) lrl> laI
lz
7. -anu(-n-1) -l=-
L-az' z-a lrl .lul
az
8. n an u(n) (z - a)2 lrl > lal
az
9. -nanu(-n-1) @-& lrl.lul
z
10. g-an
z -- e-" lrl ,1"*l
z(z +l)
11. n2 u(n)
(z:lF l"l >1
ze-^
t2. ne-u
@-;* lrl > l"-"1
z sin ab
13. sin olon
,2 -zrcos o5 + 1 l"l >1
z (z
- cos ch)
14. cos oon
,2 -22 cos rr:6 + 1 lrl >L
z sin hch
15. sin horon
,2 -zrcoshrog +1 l"l >1
z (z - cos hrrh)
16. cos hrrron
,2 -2zcoshrr5 +1 lrl >1
z e-" sino5
t7. e-e sin oon
,2 - 2z e-a cos coo + e-2u lrl, le-"1

z (z
- e-^ cos rog )
18. e-e cos (Don
,2 -zre-" cos o)o + lzl > le-"1
"-2'
za sin ob
19. a" sin oon
,2 -zra cq + a2 lrl lal
"o"
z(z-acosoh)
20. an cos o)o n
,2 -zracost6 +a2 l"l lal

3.5 Z-TRANSFORM OF SOME COMMON SIGNALS

In this section, we will obtain Z-transform of some standard sequences. Most of the standard sequences
are having infinite duration. However, there are some exceptions like-unit impulse, 6(n).

3.5.1 Z-transform of unit impulse 6(n) (RGTU, Bhopal, Sem. Exam., 2007-08)
Unit impulse 6(n) has been shown in figure 3.3. If,is given by,
 srcNALs AND SY+MS

6(n),= 1 Onlyatn=0
,= Q Otherwise
According to definition of ,-transform we have,

! *(n)z-'
x(4)=n=-€ ...(3.15)

Here. x(n) = 6(n)

Fi9.3.3. Unit impulse 6(n)

Therefore, x(z)u ) 6(n)z-" ...(3.16)

Since 6(n) is preseilt only at r.\- 0 we can directly write as under:

X(z) = ilO) ro
But 6(0) = 1 and zo = I
Therefore, X(z) = 1 ...(3.17)
Region of convergence: In equat\n (3.17), there is no'z'term. Hence, ROC is entire z-plane.
This means that z can have any value.
Thus, the z-transform pair i\given by,
6(n) e-Z--+

3.5.2. Z-transform of Delayed Unit lmpulse, 6(n - k)

Here, 6 (n - k) is a delayed unit impulse. It indicates that 5(n) is delayed by'k" samples. It has
been shown in figure 3.4.
It is given by,
6(n-k) = 1 onlyatn =kandk> 0
= Q otherwise
According to definition of z-transform, we have,

) x(n)z-"
X(z)=n=-6 ...(3.18)
Fig. 3.4. Delayed
unit impulse
Here, x(n) = 6(n-k),I>-O-

Therefore, v/-\-
zr\Lt - t2 &n-k)z-'
n=-e

But, 6(n - k) = 1 only at n = k. Ttrus, we obtain,

X(z) = l.z-k
Hence, X(z) = z-k

Region of convergence: Here, X(z) = ,-o = Since, k is positive, (k > 0) for any value of 'z'
i.
(except z = 0) we will get finite valueofX(z). Thus, ROC is entire z-plane except z = 0. This is because
if we substitute z = 0 then X(z) = -. Thus, the-transform pair is given by
6(n - k) +
t-- Z=*

3.5.3. Z-transform of advanced unit impulse,6(n + k), k > 0

Here, 6 (n + k), k > Q is an-advanced unit impulse. It indicates that 6(n) is advanced by '+ k'
samples. It has been shown in figure 3.5.
 It is given by,
6(n+k) = 1 onlyatn =-k, k> 0
-0 otherwise
According to definition of z-transform, we have,

X(z)=
"i-*tt')'-"
Here, x(n)=6(n+k) Fig. 3.5. Advanced unit impulse

Theerfore, X(z) = + k) z-'

But 6(n + k) lonlyatn=k.

"i-""
=
Thus, we get,
X(z) = l.zk
Therefore, X{z) = zk
Region of aonvergence: Here, k > 0. Hence, we will get finite values of X(z) for all 'z' except
z=-.Since,forz=-,weget,X(z)=(-)k--.Therefore,ROCisentirez-planeexceptz=-.
Thus, z-transform pair is given by
6(n+k) < z > z*k

3.5.4. Z-transform of unit step, u(n)

We know that u(n) is unit step as shown in figure 3.6. It is given by,
u(n) = 1 forn>0
-0 otherwise Fig. 3.6. Unit step
According to the definition of z-transform we havil,

i *(n)'-"
X(z)=n=-- ...(3.19)

Here, x(n) = u(n). Since, u(n) is present from n = 0 to 1 = e, we shall change the limits of
summation.

Therefore, X(z)= ) u(n)z-" ...(3.20)

n=0
But, the magnitude of u(n) is always 1.

Therefore, X(z)=
n=0
Zt.r-" = n=0
){r-')" ..G.21.)

We know that, standard expression of geometric series is given by

1
I A" =Ao +A+A2 +AB + ... = 1 +A+A2+A3
n=0
+ ... = 1_a if lAl <1 ...(3.22)

Let A = z-1. Thus, equation (3.21) becomes,

11 .1
X(z) = 1_A if lAl = .,,_r- if lz-11 < t
Multiplying numerator and denominator by z, we obtain
zz
X(z) =
;7; if lZ-11 a.7 = j if lz-11 < 1
 In equation (3.22), the condition if I A | < 1 indicates that if this condition is satisfred then only
the geometric series is convergent otherwise not. In the given example, we get this condition as,
lr-r | . 1. This, means that this is the condition of ROC.
Thus ROC: lr-'l .

This means rh" | ; <1

Therfore, lll < lrl 1

Hence, ROC is lz | > 1
This ROC is plotted as shown in figure 3.7.
Thus, ROC is exterior part of circle having radius 1.
Therefore, the z-transform pair is given by,
' u(n) . z ,'z-lrr
-L itl rlrt Fi9.3.7. ROC cl unit step
Fig.3.7.

Why to draw circle only?

The given ROC is l"l > l.In case of circle having radius 1, every point outside the periphery of
circle satisfres this condition (lrl , 1). For other diagrams like square or triangle, this condition
cannot be satisfied.

Trnportant Note: The unit step is an infinite sequence and as shown in frgure 3.7, the range of
u(n) is from n = 0 to rI = @ . This means that sequence is present only for positive values of n.
Hence, it is a causal sequenoe.

3.5.5. Z-transform of Unit Ramp

(Anna UniversiQl, Chennai, Sem. Exam., 2008-09)
We know that unit ramp sequence is denoted by r(n). It has
been shown in figure 3.8.
Its magnitude is as under:
dtn= 0, r(n) = 0
atn= 1, r(n) = 1
atn= 2, r(n) = 2
atn= 3 r(n) = 3 ... Fig.3.8. Unit ramp
Thus, it is expressed as under:
r(n)=n forn>0
-0
otherwise ...(3.23.)
Since this sequence is again causal sequence, we can write, r(n) = nu(n). ...8.24)

Important Note: Whenever any sequence is multiplied by u(n) then the magnitude ofthat sequence
is not changed. But since u(n) is present, only for positive values of n (n > 0); the result of
'sequence.
multiplication is a causal Therefore, multiplication of any sequence with u(n) always
results in causal sequence.
According to the definition of z-transform, we have,

f
X(z)=n=-6 x(n)z-o ...$.25)
But, here, x(n) = r(n) = nu (n)
 Therefore, X(z)= ! nu(n)z-' ...(3.26)
n=-€
Upto now, we have discussed that the multiplication by u(n) results in a causal sequence. Hence,
the limits of summation will be from n = 0 to n = *.'We know that since the magnitude of u(n) is 1; it
will not affect the calculation. It is just used to change the limits. Hence, now onwards, we shall not
write u(n) in the calculations.

Therefore, X(z)=
n=0
Ir(r-')"
Irr*" = n=0 ...(3.27)

Now, let us use standard summation expression,

i^'e" = r!r=
n=0 (l-Af if lAl < 1

Let A = z-1. Thus, equation (3.23) becomes,

z-l
X(z) = O_Fp if lz-11 < 1
To convert this equation into positive powers of 'z' ,lel us multiply numerator and denominator
by 12-
q1
z'-z '
Therefore, X(z)= iflz-11 <1
Fo_Fp
z
or X(z) =
]n_r;r+A if lz-11 < 1

z
or X(z)=
? -2r+l if lz-11 < 1

z
Hence,' . X(z) = A _L)2 if I z-t <1 ;

Region of convergence: Here, the condition lr-tl .l indicates the ROC.

Irl
Therefore, ROC is
l;l , t= 7. lrl
orROCis lzl >1
Therefore, ROC is exterior part of circle having radius 1. This ROC is same as shown in frgure
3.7. Therefore, the z-transform pair will be

nu(n) - r+ if lzl >1

3.5.6. Z-transform of Right Hand Exponential Sequence

As the name indicates, it is an exponential sequence present at the right hand that means only
for positive values of 'n'. Hence, it is a causal exponential sequence. Thus, it may be expressed as
under:
x(n) = crn u(n) = ctt for n 2 0
-0 for n<0 ...(3.28)
This exponential sequence has been shown in frgure 3.9.
According to definition of z-transforrn, we have