| Author |
Message |
hebu
Joined: 15 Nov 2004
Posts: 195
|
 11 Aug 2005 3:53 What's different between the two filters? |
|
|
|
|
I have two filter coefficient as below,
1) B =[0.0034 0.0068 0.0034] Numerator
A =[1.171 -2.158 1 ] Denominator
2) B =[0 0.0117 0]
A =[1.16 -2.148 1]
Both coefficients are low pass transfer function, but what's difference between
these two groups?
I am new to Discrete time design, so only I know they can be the low pass
response, the zero at Z=-1 for (1), and the zero at Z=0 for (2). Besides these,
I know nothing about them, please give me some guidance. what's the difference
and characteristic of them?
Thanks,
|
|
| Back to top |
|
 |
emrek
Joined: 26 Jul 2004
Posts: 48
Helped: 4
|
11 Aug 2005 14:47 Re: What's different between the two filters? |
|
|
|
|
| Why do not you try these filters in matlab. You can use fdatool or sptool to analyz these filters even on zero-pole graph.
|
|
| Back to top |
|
|
hebu
Joined: 15 Nov 2004
Posts: 195
|
11 Aug 2005 16:49 Re: What's different between the two filters? |
|
|
|
|
ya, I have tried it, the poles of both filters are all in the unit circle, but the zero
of first filter is at Z=0, and the second is at Z=-1, once the practical implementation makes an inaccruacy coefficient the zero will be outisde of the
cirle. Does this affect the stability? I'm afraid I will miss something.
If there is anything I missed, please remind me.
Thanks for your reply.
|
|
| Back to top |
|
|
khubaibahmed
Joined: 24 Aug 2005
Posts: 31
Helped: 2
|
29 Aug 2005 11:39 Re: What's different between the two filters? |
|
|
|
|
| hebu wrote: |
ya, I have tried it, the poles of both filters are all in the unit circle, but the zero
of first filter is at Z=0, and the second is at Z=-1, once the practical implementation makes an inaccruacy coefficient the zero will be outisde of the
cirle. Does this affect the stability? I'm afraid I will miss something.
If there is anything I missed, please remind me.
Thanks for your reply. |
Zeros does not alter the stability of the system.However for implementation you must take care of floating point and fixed point and scale ur data acordingly so that poles and zeros may not displaced.
|
|
| Back to top |
|
|
zorro
Joined: 06 Sep 2001
Posts: 380
Helped: 47
|
29 Aug 2005 18:56 Re: What's different between the two filters? |
|
|
|
|
Hi hebu,
Both filters are low-pass, the poles are inside the unit circle (not on it) and have nearly the same location.
How had you got these coefficients? Probably they were obtained from the same low-pass continuous-time filter but by different methods: probably the first by bilinear transformation and the second with time invariant impulse response.
Regards
Z
|
|
| Back to top |
|
|
Google
AdSense
|
29 Aug 2005 18:56 Ads |
|
|
|
|
|
|
| Back to top |
|
|
hebu
Joined: 15 Nov 2004
Posts: 195
|
31 Aug 2005 18:27 Re: What's different between the two filters? |
|
|
|
|
I got the first coefficients from matlab filter design tool. I specify a filter type
and got the coefficient. Another one is got when I designed a sc filter by continuous time approach. Design a sc filter by R and c apporach.
And transfer the filter structure and cap ratio to Z transfer function.
|
|
| Back to top |
|
|
