Tim,

wrote in message
oups.com...
Now that PCs with 3.4 GHz processors are available,
plus pretty sophisticated DSP development systems, at
a reasonable cost (several hundreds of dollars), is
there a good homebrew RF spectrum analyzer
available...

There are a couple out there. One with some of the best specs is the one
designed by Tom McDermott that TAPR will offer as a kit and that TenTec will
be manufacturing. More info:
http://www.tapr.org/kits_vna.html?PH...3c5d1a 637bfc .
The price was something in the ballpark of $600, which is nothing compared to
a new HP, Rohde & Schwarz, etc. network analyzer (many such analyzers are
still WELL above $20k...). If you can add another "0" to your price tag,
though, items on eBay do start to look attractive. :-)

One of the design limitations of Tom's analyzer is that the upper end is
100MHz; another is that the dynamic range is something like 60dB.

low cost development board +
high end PC +
GNURadio (or whatever) +
Extensions to GNURadio
--------------------------------
= Pretty (or very) good RF spectrum analyzer

Well, not to put too much of a damper on your hopes here, but I'd have a
couple of concerns: (1) the GNURadio USRP board isn't particularly designed
for "really weak signal" work. Last time I asked (back at Dayton), they
hadn't even measured how much noise was running around the board, but I'd be
really surprised if it's anywhere near as quiet as a good spectrum analyzer.
Secondarily, although there is already a lot of software out there for
GNURadio and Tom's network analyzer board, it's nowhere near as "polished" as
commercial offerings.

I think the USRP board is one of the best new things to come along in amateur
radio in a long time -- it's opening up a wealth of opportunities for
experimenting that would have been much more expensive to get into previously.
Also, GNURadio is (at least in theory) not too tightly tied to the USRP (they
started with some PCI ADC boards, after all), so designing a really quiet
vector or scalar analyzer to plug into its "architecture" would be a pretty
cool undertaking.

Question 2)

In a similar vein, though of course not as expensive,
Does anyone have circuit diagrams for a homebrewed
function generator, working against an accurate
internal reference, and putting out: pure sinusoids,
square waves, saw tooth functions, white noise ....
etc.

This actually is a pretty good application for a USRP, at least if you don't
need super tight control over stuff like phase noise. It's very easy to get
the thing to output any arbitrary waveform you like.

---Joel Kolstad