some help please on more efficient way to detect tones?
in [DSP]
|
Prev: Plastic Logic makes it debut
Next: book for OFDM
From: jacobfenton on 10 Jan 2010 21:05 > >I am sorry but this is where free ride comes to the end. I would be glad >to help you, however this is going to be business. If you are >interested, the contact is at my web site. > >Vladimir Vassilevsky >DSP and Mixed Signal Design Consultant >http://www.abvolt.com > After reading some more, IQ is baseband! Now I am understanding things. I still need to look into our IF requirement, since it can drift quite alot it seems. And how that effects things.
From: Fred Marshall on 11 Jan 2010 13:29 jacobfenton wrote: >> I am sorry but this is where free ride comes to the end. I would be glad > >> to help you, however this is going to be business. If you are >> interested, the contact is at my web site. >> >> Vladimir Vassilevsky >> DSP and Mixed Signal Design Consultant >> http://www.abvolt.com >> > > After reading some more, IQ is baseband! Now I am understanding things. I > still need to look into our IF requirement, since it can drift quite alot > it seems. And how that effects things. Jacob, Yes, IQ is baseband - well .. depending on the sampling frequency and what you want to call "baseband" I suppose. But, for conversation, good enough. Still, I'm a little confused - nothing new in that! :-) If the clock is stable then the IF center frequency is stable is it not? Or, is it tracking the center frequency with the PLL? If the clock is stable then the IF center frequency is stable and you need an IF bandwidth to accommodate Doppler,etc. If the clock is tracking/locked then the center frequency is tracking but the IF bandwidth and the information coming out should be reasonably stable as it's all relative to the carrier, or should be. If you're tracking then I should think that Doppler, etc. would be accounted for and you'd not need to keep the bandwidth wide enough to account for it. This is pretty typical of narrowband space communication systems for example - it keeps the SNR up. Is the latter your situation? I believe in either case the post-demod tones should be stable as you have: - acceptable SNR - are looking at rate of phase changes. Once you've demodulated then the IF bandwidth is immaterial. Only the modulation bandwidth matters (i.e. the band within which the tones exist). Here's how I'm understanding this: You "detect" the signal. You downsample in the IF bandwidth to get baseband. - If the clock is not tracking th carrier then the IF bandwidth needs to take into consideration drift, Doppler and even perhaps the modulation bandwidth. - If the clock is tracking the carrier then the IF bandwidth can be narrower than if you're not. - In no case can the IF bandwidth be narrower than what is caused by the modulation and you may need a bit of bandwidth to accomodate carrier dynamics. This ignores what it takes to "lock" - which may take more bandwidth. All system considerations. Then you detect the modulation signal. At this point it's only the modulation signal bandwidth that matters. And that's why we keep talking about decimation / sample rate reduction because the IF bandwidth and the modulation bandwidth are two different things entirely. Fred
From: jacobfenton on 11 Jan 2010 15:39
>Yes, IQ is baseband - well .. depending on the sampling frequency and >what you want to call "baseband" I suppose. But, for conversation, good >enough. > >Still, I'm a little confused - nothing new in that! :-) > >If the clock is stable then the IF center frequency is stable is it not? >Or, is it tracking the center frequency with the PLL? > >If the clock is stable then the IF center frequency is stable and you >need an IF bandwidth to accommodate Doppler,etc. > >If the clock is tracking/locked then the center frequency is tracking >but the IF bandwidth and the information coming out should be reasonably >stable as it's all relative to the carrier, or should be. >If you're tracking then I should think that Doppler, etc. would be >accounted for and you'd not need to keep the bandwidth wide enough to >account for it. This is pretty typical of narrowband space >communication systems for example - it keeps the SNR up. > >Is the latter your situation? > >I believe in either case the post-demod tones should be stable as you have: >- acceptable SNR >- are looking at rate of phase changes. > >Once you've demodulated then the IF bandwidth is immaterial. Only the >modulation bandwidth matters (i.e. the band within which the tones exist). > >Here's how I'm understanding this: >You "detect" the signal. >You downsample in the IF bandwidth to get baseband. >- If the clock is not tracking th carrier then the IF bandwidth needs to >take into consideration drift, Doppler and even perhaps the modulation >bandwidth. >- If the clock is tracking the carrier then the IF bandwidth can be >narrower than if you're not. >- In no case can the IF bandwidth be narrower than what is caused by the >modulation and you may need a bit of bandwidth to accomodate carrier >dynamics. This ignores what it takes to "lock" - which may take more >bandwidth. All system considerations. > >Then you detect the modulation signal. >At this point it's only the modulation signal bandwidth that matters. >And that's why we keep talking about decimation / sample rate reduction >because the IF bandwidth and the modulation bandwidth are two different >things entirely. > >Fred > > You are correct, currently the system does not lock to the carrier or have any compensation for carrier drift, or local clock drift. This is all accounted for in the IF bandwidth requirement. Our current set up is to have the IF sampling bandwidth big enough to count for all the drift, so we can still recover the information after demodulation. So once demodulated I can decimate for sure, since after demodulation I only care about the information bandwidth, not the IF bandwidth. |