Digital Signal Processing Reference
In-Depth Information
Solution
(i) Taking the z-transform of the input sequence, we obtain
X 1 ( z ) = z
z 1 .
Applying the convolution property, the z-transform Y 1 ( z ) of the output response
is given by
2
(1 z 1 )(1 z 1 + z 2 ) .
Taking the partial fraction expansion of Y 1 ( z ) yields
z
( z 1)( z 2 z + 1)
z
Y 1 ( z ) =
H ( z ) X 1 ( z ) =
=
1
1 z 1
1
1 z 1 + z 2 .
Using entries (3) and (12) of Table 13.1 (see Problem 13.5), we obtain
Y 1 ( z ) =
1
1 z 1
←→
u [ k ]
and
π k
3
+ π
6
←→ 1
1 z 1 + z 2 .
Using the linearity property, the output y 1 [ k ]isgivenby
3 sin
u [ k ]
π k
3
+ π
6
y 1 [ k ] =
1
u [ k ] .
3 sin
Note that the output response contains a unit step function and a sinusoidal term
and is, therefore, bounded.
(ii)
Taking the z-transform of the input sequence, we obtain
1
(
3 / 2) z
X 2 ( z ) =
2 .
1 z
1 + z
Applying the convolution property, the z-transform Y 2 ( z ) of the output response
is given by
3
(1 z 1 + z 2 ) 2 .
Using the frequency-differentiation property (see Problem 13.6), it can be
shown that the following is a z-transform pair:
1
2
(
3 / 2) z
z
(
3 / 2) z
Y 2 ( z ) =
H ( z ) X 2 ( z ) =
=
1 z 1 + z 2
1 z 1 + z 2
3
(1 z 1 + z 2 ) 2 .
2
3
π
3 k
π
3 k + π
(
3 / 2) z
←→ =
sin
sin
u [ k ]
6
3
Therefore, the output response is given by
2
3
π
3 k
π
3 k + π
y 2 [ k ] =
sin
sin
u [ k ] .
6
3
Note that the output is a growing sinusoid function because of the k /
3 scaling
factor. Therefore, as k increases, the y 2 [ k ] increases without bound, leading
to an unstable situation.
Search WWH ::




Custom Search