[dsp] Analog front end design

[scott at opentrac.org]

I've been using the dsPIC filter designer lite to design filters so far.
I'm not designing for a PIC, but the program's not really PIC-specific.  It
generates working C code that I can feed real data and then plot the results
in Excel.  I'm still a bit lost when it comes to actually choosing the
parameters.  I know the passband I want, but I'm not sure where to set the
stopband, for example.  The ripple parameters I find don't really matter if
I'm constraining it to a limited number of coefficients - I just take the
best it can give me.  Not always sure what's 'good enough' though.

It does do a good job of showing me the results of my design.  I can at
least keep tweaking things until it looks like what I want.

I'm not so worried about clock recovery and such in the digital domain.  I
can handle digital, I've done stuff like that before - decoding data off an
XR2211, for example.  But I'm lacking a lot of the math background I need
for the DSP portions and it's a steep learning curve.

Anyway, is what I'm describing with the autocorrelation (?) the same thing
you're talking about?

Thanks,

Scott

-----Original Message-----
From: David Willmore [mailto:willmore at optonline.net]
Sent: Monday, April 25, 2005 6:43 PM
To: scott at opentrac.org
Cc: willmore at optonline.net; 'TAPR DSP Mailing List'
Subject: Re: [dsp] Analog front end design


> Generally, wouldn't you want the sample rate to be a multiple of the baud
> rate to avoid ISI when you're filtering?  Or maybe I'm just designing my
> filters wrong.  Seems like it'd make clock recovery simpler, too.

I can see a lot of ways to do it.  If you rely on your sample rate being a
multiple of the data rate, you leave yourself open to a lot of errors if
those rates ever differ.  I see it as better to just admit that they'll
never be firmly related and deal with the consequences.

For clock recovery, just lock a DPLL to the square of the data signal?

> Anyway, I'm trying something along the lines of what you're suggesting, I
> think.  Right now it's just a test program on the PC.  It's multiplying
each
> sample in the input waveform by a sample 90 degrees apart at 1700 hz.
> Haven't got the LPF implemented properly yet.  I set it up to output a
> digital value based on a couple of thresholds and I can plot the result in
> Excel and see the bits clearly.  It's not going to handle much noise that
> way, though.

The LPF is important. :)  All you really need is a second order IIR filter.
The DSPGuide should be able to help you design a simple one of those.


> Can you suggest a good book that would cover this?  Preferably something
> accessible to a relative newbie.  =]

Hmm, Sklar's book is good.  I'll try to find my copy and get you a
reference to it.  I think there's a second edition of it out.

Cheers,
David