On Sep 23, 11:33*am, Jim Lux wrote:
Larry Gauthier (K8UT) wrote:
Art,

Funny that you should ask, because the October 2008 QST (just arrived a few
days ago) has a computer-based DIY spectrum analyzer project. Doesn't look
too difficult: *a couple of dozen parts and a software program that's
downloadable from the ARRL web site.

-larry
K8UT
"Art Unwin" wrote in message
....
S A s are usefull for antenna hobbyists. Some twenty years ago there
was a DIY *S A that used a t v tuner in
Ham Radio. Moving forwards some twenty years and most hams have
computors and there are 2.4 Ghz
spectrum analysers out there as well as audio set ups for computors.
So what is available today for hams
in the HF BANDS? There are a lot of older type audio S A s on E bay
but I haven't heard of anybody modifying thoseand with most having
computors it would seem reasonable to think that there were components
available that could performn between audi and 2.4 Ghz. So has anybody
seen any modern type DIY spectrum analysers
around for sale or kits? I have 141t and other types of SA which now
are considered bulky and I would really like something smaller or say
modify something like a audio setup as a project.
Comments

These days, the scanning receiver part isn't the hard part. It's
calibration and user interface.

For instance a Icom PCR1000 makes a fine HF,VHF,UHF spectrum analyzer,
although the resolution BW steps are limited and sweep speeds are limited..

What you get with a "real" spectrum analyzer is a fairly intuitive user
interface (with knobs!), well characterized and consistent RF
performance over a wide variety of operating conditions, and
calibrateability (i.e. stable and repeatable operation)

It's sort of like the difference between a real oscilloscope and a sound
card in a PC with an oscilloscope program. *The real scope probably has
wider applicability, more options arranged in a systematic way, etc.
But for some applications, the soundcard is just fine.

Adding a bit to Jim's good comments... just because spectrum analysis
can be done with fewer and smaller parts than you needed a few years
ago doesn't mean you can get really good performance in a tiny
package. It's still pretty tricky to keep residuals, spurs and
distortion very low while providing the amplitude and frequency
stability that Jim mentions. These days, low phase noise is also a
big benefit, since RBW can be very low.

To me, one of the biggest advances in spectral analysis in the past
few years is the ability to get very decent dynamic range while
digitizing a significant bandwidth, meaning that you can look at a
whole bunch of signals all at the same time, and not miss transient
events because you were scanning a different range when the event
occurred. A side benefit is that a properly done digitizer has
extremely good linearity: you can accurately measure amplitude over a
wide range of signal levels. It's quite possible to digitize the
whole HF range with very respectable dynamic range; the display then
becomes a matter of digital signal processing--and there are chips
capable of keeping up with the high data rate while providing narrow
frequency resolution over large chunks of, or even the whole,
digitized bandwidth. Digitization also makes keeping phase noise low
relatively easy, if you don't have to also do analog frequency
conversion--and even there, if you can do the analog conversion in
only relatively large frequency steps, it's a help.

Cheers,
Tom