[time-nuts] OT: Spectrum Analyzer

Magnus Danielson magnus at rubidium.dyndns.org
Sat Oct 31 22:30:14 UTC 2009


Didier Juges wrote:
> Another major difference between the older (sweeping) analyzers and the more modern ones (FFT) is that the sweep type analyzers only look at one particular frequency once per sweep, and for a short time, while the fft types pretty much look at the entire span continuously. That makes the fft type much more capable of identifying low duty cycle transmitters. If the signals you are interested in are CW, it will not make much difference, but that is not always the case.

If looking at signals or artifacts near noise, accumulation over time 
will provide a more consistent view. Shorter time to achieve the same 
monitoring time of several frequencies (as they is being monitored in 
parallel) also means that any instability in longterm cause less 
de-correlation.

I have the precursor to the HP4395A, the HP4195A. Nice unit, but what 
annoys me mostly is the lack of frequency range and that I could do with 
a little more of the accessories. Still, few network analyzers will do
1 mHz steps. I swept the 7 cesium peaks using it and only a mixer as 
part of diagnosis of my cesium beam. Hacking into the 12 MHz level 
worked nicely, and feeding the detector output to the mixer to modulate 
the swept sine so that it emulates a 12 MHz filter while it infact was 
doing just about 9 GHz was kind of sweet.

Oh, someone that has a Linux-tool for LIF floppies?

Cheers,
Magnus



More information about the time-nuts mailing list