Re: [Discuss-gnuradio] cable modem tuner now integrated

[Darrin John Eden]

Awesome! I can't wait to be in a position to acquire said hardware. 

The "find interesting signals" is such a classy problem. I suppose it runs the Information Theory gamut in a big way.

I wonder if serving a normative subset of the data a node collects might enable those of us without handy access to The Hardware to do our part and explore interesting signals? The address@hidden folks have a lot of us working on it now, but who wants to be tied to just one antenna?

Darrin

On Monday, November 26, 2001, at 12:25 AM, Eric Blossom wrote:

> Hi Folks,
>
> I've now got a very nice cable modem tuner module (eval board) hooked
> into GNU Radio.  From a hardware point of view it was a zero glue
> operation.  Just a couple of cables and rigging up a power supply.
>
> FYI, it's a Microtune 4937DI5 3x7702 module.  Tunes from 50M to 860M
> and has about a 6MHz wide IF centered at 5.75M.  Very convenient for
> our 20M sample A/D board.  Cables right up.  You control it by bit
> banging it over the parallel port.  It has a I2C interface plus a few
> additional bits.
>
> I've attached some FFT graphs from various parts of the spectrum.
> Lots of good stuff to work with.  The center of the IF is indicated at
> 5.75MHz.  The displayed frequency axis values don't currently take
> into account where the module is tuned.
>
> Here's what they are:
>
> fm92.png
> Lots of broadcast FM stations.  Should be able to cook up a
> cool demo that demods more than one station at a time without
> retuning the RF front end.  This will show something that you
> can't do with a regular radio.  Existing code demods the
> channel that is tuned to 5.75M
>
> ch35-597.25M.png
> Local TV channel 35.  Should be able to extract the audio.
>
> 461.2M.png
> Not too exciting in the snap shot, but in real time there are
> all kinds of signals coming and going.  Some are conventional
> push-to-talk FM, but some appear to be burst transmissions.
>
> narrow-band-891M.png
> Narrow band FM.  Note close channel spacing.  Some are FM, some
> are digital.
>
> spread-spectrum-879M.png
> Right hand side of graph is spread spectrum lump.
> Conventional narrow band on left hand side.
>
> One thing I've noticed out of watching the FFT for a while is we need
> software that "finds interesting signals".  
>
> I plan on reworking the spectrum display so that there are a few
> markers available that can be positioned to show signals of interest,
> select this one to demod, etc.  Annotations into a data base might be
> nice too.
>
> Eric
λφ(decentric.org)だ