Le Tue, 18 Sep 2007 20:20:58 -0700, John Larkin a écrit:

On Wed, 19 Sep 2007 00:17:53 GMT, Joerg
wrote:

John Larkin wrote:

I was playing with our new spectrum analyzer and decided to just hang
a 12" cliplead off the input to see what was floating around.

See pics. The biggest signal was about -26 dBm, an FM station I guess.
Just about every slot on the FM band is full.


Looks like a luxury analyzer. Aeroflex? I thought you could buy those
only with a GSA account or for government contract work.


It's the new 3 GHz Aeroflex; anybody can buy it. With all the options
(tracking gen, preamp, high resolution) it was $13K or something. It has
a few firmware bugs, but I think they're sincere about fixing them.
Except for a few glitches, it's a lovely box, with a beautiful display.

I was disappointed in the harmonic distortion, -65 dB typically, but it
looks like that's actually not real bad compared to other units. We have
some RF signal generators that have THD numbers like 1.5%!

To test some amps we're doing, I guess we'll have to buy a few sets of
matched lowpass/highpass filters, so we can pump in a clean sinewave,
then and pull out the fundamental before we feed it to the spectrum
analyzer. What a nuisance. Maybe we should find/make a couple of tunable
notch filters.

How does one measure the THD of an amplifier, down around -70 dbC, in
the 1-30 MHz range maybe?

Why don't spectrum analyzers do log sweeps?


Some do, some don't.
I've both, one of which can nicely do linear and log sweeps.

-65dB isn't that bad, depending at which level it is specified.
To measure lower (but not extreme) THD without notching out the
fundamental, you can lower the input level (2nd order SA distorsion will
lower at twice the rate, 3rd at 3 times the rate, ...), and regain the
dynamic range by lowering the IF bandwidth.

You can easily check how low a level you have to go with a BP filtered
gen.

--
Thanks,
Fred.