I need help on how to filter a very narrowband signal. Signal bandwidth = 0.05 Hz Signal frequency = 5000 Hz Sampling rate = 30000Hz For the filter (lowpass or bandpass FIR): transition bandwidth = 0.02 Hz max passband ripple = 0.15 dB minimal stopband attenuation = 80 dB If I can only get a lowpass filter I will bring the signal to baseband. I do know that I need to use multistage multirate filter. Any help with this simulation will be appreciatted. The simulation is in Matlab. Thanks
Very narrowband filter
Started by ●March 10, 2005
Reply by ●March 10, 20052005-03-10
e09 wrote:> I need help on how to filter a very narrowband signal. > Signal bandwidth = 0.05 Hz > Signal frequency = 5000 Hz > Sampling rate = 30000Hz > For the filter (lowpass or bandpass FIR): > transition bandwidth = 0.02 Hz > max passband ripple = 0.15 dB > minimal stopband attenuation = 80 dB > > If I can only get a lowpass filter > I will bring the signal to baseband. > > I do know that I need to use multistage multirate filter. > Any help with this simulation will be appreciatted. > The simulation is in Matlab. > > ThanksYou don't necessarily need to use a multistage multirate filter. I haven't tried it (no Matlab on my Mac), but for starters, you might load up 'sptool' in Matlab and try to design some IIR bandpass filters for those specifications. You did not mention a phase linearity / group delay variation specification, so IIR warrants consideration, IMO. John
Reply by ●March 10, 20052005-03-10
I am curious how many taps are acceptable and how long your startup transient is allowed to be. Also, is the stability of the signal such that the frequency will not wander in and out of the passband? Does it matter if it does? Do you care that it will ring like crazy if you hit it with something impulsive? Dirk
Reply by ●March 10, 20052005-03-10
"dbell" <dbell@niitek.com> writes:> Do you care that it will ring like crazy if you hit > it with something impulsive?Doesn't the answer to this question have to be either "No" or "I do but there's nothing I can do about it" if the OP does indeed require a narrow bandwidth (e.g., to improve SNR as much as possible) since narrow bandwidth == long time response? -- Randy Yates Sony Ericsson Mobile Communications Research Triangle Park, NC, USA randy.yates@sonyericsson.com, 919-472-1124
Reply by ●March 10, 20052005-03-10
No. The answer could be "I hadn't thought it out and it isn't going to work even if I build it." I can't tell from the OPs post, but considering the possibility might save the OP some time. Dirk Randy Yates wrote:> "dbell" <dbell@niitek.com> writes: > > > Do you care that it will ring like crazy if you hit > > it with something impulsive? > > Doesn't the answer to this question have to be either "No" or "I do > but there's nothing I can do about it" if the OP does indeed requirea> narrow bandwidth (e.g., to improve SNR as much as possible) since > narrow bandwidth == long time response? > -- > Randy Yates > Sony Ericsson Mobile Communications > Research Triangle Park, NC, USA > randy.yates@sonyericsson.com, 919-472-1124
Reply by ●March 11, 20052005-03-11
"dbell" <dirkman@erols.com> writes:> Randy Yates wrote: >> "dbell" <dbell@niitek.com> writes: >> >> > Do you care that it will ring like crazy if you hit >> > it with something impulsive? >> >> Doesn't the answer to this question have to be either "No" or "I do >> but there's nothing I can do about it" if the OP does indeed require a >> narrow bandwidth (e.g., to improve SNR as much as possible) since >> narrow bandwidth == long time response? > > No.Then please tell me how one builds a filter that simultaneously has a very narrow bandwidth and a limited temporal response? -- % Randy Yates % "She's sweet on Wagner-I think she'd die for Beethoven. %% Fuquay-Varina, NC % She love the way Puccini lays down a tune, and %%% 919-577-9882 % Verdi's always creepin' from her room." %%%% <yates@ieee.org> % "Rockaria", *A New World Record*, ELO http://home.earthlink.net/~yatescr
Reply by ●March 11, 20052005-03-11
Randy, You don't. My point was that there is an alternate answer to those you proposed that might apply, that is (rephrased) "I do care and it makes the solution not usable." I asked the original questions because I have seen people try this very type of design to find out that they could not use it because of what was actually a fundamental property of their design. Since the OP was not specific about the application, I thought I would raise some issues before he implemented the 'solution'. Dirk Randy Yates wrote:> "dbell" <dirkman@erols.com> writes: > > Randy Yates wrote: > >> "dbell" <dbell@niitek.com> writes: > >> > >> > Do you care that it will ring like crazy if you hit > >> > it with something impulsive? > >> > >> Doesn't the answer to this question have to be either "No" or "Ido> >> but there's nothing I can do about it" if the OP does indeedrequire a> >> narrow bandwidth (e.g., to improve SNR as much as possible) since > >> narrow bandwidth == long time response? > > > > No. > > Then please tell me how one builds a filter that simultaneously has a > very narrow bandwidth and a limited temporal response? > -- > % Randy Yates % "She's sweet on Wagner-I thinkshe'd die for Beethoven.> %% Fuquay-Varina, NC % She love the way Puccini lays downa tune, and> %%% 919-577-9882 % Verdi's always creepin' from herroom."> %%%% <yates@ieee.org> % "Rockaria", *A New World Record*,ELO> http://home.earthlink.net/~yatescr
Reply by ●March 11, 20052005-03-11
Randy Yates wrote:> "dbell" <dirkman@erols.com> writes: > >>Randy Yates wrote: >> >>>"dbell" <dbell@niitek.com> writes: >>> >>> >>>>Do you care that it will ring like crazy if you hit >>>>it with something impulsive? >>> >>>Doesn't the answer to this question have to be either "No" or "I do >>>but there's nothing I can do about it" if the OP does indeed require a >>>narrow bandwidth (e.g., to improve SNR as much as possible) since >>>narrow bandwidth == long time response? >> >>No. > > > Then please tell me how one builds a filter that simultaneously has a > very narrow bandwidth and a limited temporal response?It can't be done. Dirk knows it can't be done. He very kindly wants to the be sure the OP knows it can't be done. I lent a physicist an adjustable power supply one time. It had a current limiter too; he specifically wanted one like that. He brought it back half an hour later, saying that it needed to be fixed. I showed him that the voltage could be set between zero and 20 volts, and the current could be clamped between a few microamps and half an amp. He said that what he needed to do was to set them simultaneously. I took him a long time to understand why that couldn't be done without adjusting the load. My neighbor once told me that the starter on his car didn't work. Asked how he knew, he said it was plain: the car wouldn't start. "It goes rrr rrr rrr, but it won't start." That one was easy. I told him "starter" was a misnomer; it's really a cranker, and the rrr rrr rrr sound proved that both battery and cranker were fine. Then I adjusted his points. Moral: never take for granted that the other guy understands the problem or the constraints on a solution. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●March 11, 20052005-03-11
"e09" <ewong9@juno.com> wrote in message news:1110490865.518669.77780@o13g2000cwo.googlegroups.com...>I need help on how to filter a very narrowband signal. > Signal bandwidth = 0.05 Hz > Signal frequency = 5000 Hz > Sampling rate = 30000Hz > For the filter (lowpass or bandpass FIR): > transition bandwidth = 0.02 Hz > max passband ripple = 0.15 dB > minimal stopband attenuation = 80 dB > > If I can only get a lowpass filter > I will bring the signal to baseband. > > I do know that I need to use multistage multirate filter. > Any help with this simulation will be appreciatted. > The simulation is in Matlab. >What the other folks were alluding to is this (I think): Are you aware that the settling time for such a filter would be around 20 seconds (1/0.05Hz) ? Is that acceptable to you?\ Fred
Reply by ●March 11, 20052005-03-11
"dbell" <dirkman@erols.com> writes:> Randy, > > You don't. My point was that there is an alternate answer to those you > proposed that might apply, that is (rephrased) "I do care and it makes > the solution not usable." I asked the original questions because I > have seen people try this very type of design to find out that they > could not use it because of what was actually a fundamental property of > their design. Since the OP was not specific about the application, I > thought I would raise some issues before he implemented the 'solution'.I see now your intention - you were definitely asking a good question. However, even though it's academic, logically your response to my question was incorrect because I added the clause "if the OP does indeed require a narrow bandwidth." The case you bring up means in effect that a narrow bandwidth can't be a requirement. -- Randy Yates Sony Ericsson Mobile Communications Research Triangle Park, NC, USA randy.yates@sonyericsson.com, 919-472-1124
