Cost-Effective Fast Data Acquisition, and the LT5506

[Tim Wescott]

I am working on a radio project in my spare time. The proposed
architecture will use an LT5506 to bring a signal down to baseband,
through some low-pass filters, then into a DSP chip for all the 'real'
processing. I'd like to prototype this thing in pieces before I put it
all onto one tiny board. So I have some questions for the group:

LT5506, 40-500MHz I/Q demodulator with built-in VGA

Have you ever used it?

Do you know if there's an eval board available (Linear doesn't appear to
have anything, but I may not be navigating their site correctly).

Do you know of any alternatives? Something that had built-in low-pass
filters would be tres cool.

Data Acquisition:

I'd like to buy (or build, if I have to) an eval board for the LT5506
and acquire chunks of data to feed to a PC. For this I'd need a data
acquisition device that could sample two channels at no less than 10kHz;
50kHz would be nicer; 16 bits would be nice but I could live with 12
(for 8 bits or less I can use my 'scope). I've found one that's $400 or
so for analog in/USB out -- is that the best I can do? Any suggestions
will be appreciated.

Low-Pass Filters:

My thinking right now is that I'll just come out of the LT5506 into
passive RC filters, oversample the snot out of the signals, and do a
quickie filter-and-decimate before I do the complex signal processing.
Anyone know of a really small low-pass filter I could use that would
have better performance? The LT5506, naturally, has balanced outputs
and its specifications go to hell if you don't use them as such, so any
solution will have to accept a balanced input.

Thanks all.

--
Tim Wescott
Wescott Design Services
Voice: 503-631-7815
Cell: 503-349-8432
http://www.wescottdesign.com

 

[K7ITM]

The data acq part: just get a stereo USB sound port. 16 bits at
22.05ks/s should be easy. I have one that's 24 bits, and it wasn't
anything like $400. You can get 16 bits stereo to 48k or 96k s/s for
much less than $100 these days, even isolated. Nice thing about
delta-sigma converters is that the antialias filtering is relatively
easy. In a design I did some fifteen years ago, I used a five-pole LC
filter implemented with little smt chokes and C0G caps that worked just
fine* and allowed a wide range of output data rates without changing
the filter, but typically you don't need anything like that sort of
rolloff if you're sticking to a narrow range of clock frequencies.

Call your local Linear Technology rep for current eval board
availability.

Cheers,
Tom

* I was expecting to be able to see some distortion from those little
chokes, but distortion from the AAF was well below the 16-bit ADCs'
distortion.

 

[Tim Wescott]

The data acq part: just get a stereo USB sound port. 16 bits at
22.05ks/s should be easy. I have one that's 24 bits, and it wasn't
anything like $400. You can get 16 bits stereo to 48k or 96k s/s for
much less than $100 these days, even isolated. Nice thing about
delta-sigma converters is that the antialias filtering is relatively
easy. In a design I did some fifteen years ago, I used a five-pole LC
filter implemented with little smt chokes and C0G caps that worked just
fine* and allowed a wide range of output data rates without changing
the filter, but typically you don't need anything like that sort of
rolloff if you're sticking to a narrow range of clock frequencies.

Call your local Linear Technology rep for current eval board
availability.

Cheers,
Tom

Thanks. Someone else pointed out the soundcard part off list (d'oh).
Linear Tech doesn't do eval boards for that part, I was hoping maybe
someone knew of someone who does. I think I'm just going to do a
quick-turn eval board myself -- it's one of those itty bitty packages
with the grounding pad in the center, so it pretty much has to be
reflowed to be useful.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

 

[K7ITM]

One way I've considered doing those heat pads on homebrew boards is to
drill a large hole there, say to just clear a 10AWG wire, tin the heat
pad, solder the part down, and finally slip a very short (e.g. 0.09")
length of tinned #10 into the hole, heating it to reflow to the part,
and solder around the the wire to the ground plane. Should work, but
haven't tried it (yet). Such a short, fat wire is pretty low thermal
resistance. You can also add some heat sinking on the back to pull
heat away from the wire stub. Toughest part might be cutting the wire
cleanly perpendicular to its axis. A reasonably clean cut followed by
a bit of filing would do it, and I can imagine better ways if I were
making several.

Cheers,
Tom

 

[Tim Wescott]

One way I've considered doing those heat pads on homebrew boards is to
drill a large hole there, say to just clear a 10AWG wire, tin the heat
pad, solder the part down, and finally slip a very short (e.g. 0.09")
length of tinned #10 into the hole, heating it to reflow to the part,
and solder around the the wire to the ground plane. Should work, but
haven't tried it (yet). Such a short, fat wire is pretty low thermal
resistance. You can also add some heat sinking on the back to pull
heat away from the wire stub. Toughest part might be cutting the wire
cleanly perpendicular to its axis. A reasonably clean cut followed by
a bit of filing would do it, and I can imagine better ways if I were
making several.

Cheers,
Tom

Sounds like you're thinking single-sided board. I was just going to pop
for a quick turn house that delivers two-sided, plated-through holes.
The price is not too bad, all things considered.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/