Digital Signal Processing Reference
In-Depth Information
Poles: z ¼ 0:5, jzj ¼ 0:5 < 1; z ¼0:707 j0:707,
jzj¼
q
ð0:707Þ
2
2
þð0:707Þ
¼ 1.
The zero and poles are plotted in
Figure 6.10
.
Since the outmost poles are first order at z ¼0:707 j 0:707
and are on the unit circle, the system is marginally stable.
d. Zeros are z ¼0:5 j 0:5.
Poles: z ¼ 1, jzj ¼ 1; z ¼ 1, jzj ¼ 1; z ¼1, jzj ¼ 1;, z ¼ 0:6
jzj¼0:6 < 1.
The zeros and poles are plotted in
Figure 6.10
. Since the outmost pole is a multiple order (second order) pole at
z ¼ 1 and is on the unit circle, the system is unstable.
6.4
DIGITAL FILTER FREQUENCY RESPONSE
From the Laplace transfer function, we can achieve the analog filter steady-state frequency response
HðjuÞ
by substituting
s ¼ ju
into the transfer function
HðsÞ
. That is,
HðsÞj
s¼ju
¼ HðjuÞ
Then we can study the magnitude frequency response
jHðjuÞj
and phase response
:
HðjuÞ
. Similarly,
in a DSP system, using the mapping in Equation
(6.16)
,
we substitute
z ¼ e
sT
s¼ju
¼ e
juT
into the
z-transfer function
HðzÞ
to acquire the digital frequency response, which is converted into the
magnitude frequency response
Hðe
juT
Þ
and phase response
:
jHðe
juT
Þj
. That is,
HðzÞj
z¼e
juT
¼ Hðe
juT
Þ¼
Hðe
juT
Þ
:
Hðe
juT
Þ
(6.18)
Let us introduce a normalized digital frequency in radians in the digital domain:
U
¼ uT
(6.19)
Then the digital frequency response in Equation
(6.18)
becomes
Hðe
j
U
Þ¼HðzÞ
z¼e
j
U
¼
Hðe
j
U
Þ
:
Hðe
j
U
Þ
(6.20)
The formal derivation for Equation
(6.20)
can be found in Appendix D.
Now we verify the frequency response via the following simple digital filter. Consider a digital
We can determine the system output
yðnÞ
, which consists of the transient response
y
tr
ðnÞ
and the
steady-state response
y
ss
ðnÞ
. We find the z-transform output as
FIGURE 6.11
System transient and steady-state frequency responses.
Search WWH ::
Custom Search