[time-nuts] Re: Testing GPSDOs

[Erik Kaashoek]

I'm trying to check the short term stability of a TCXO, for tau below 1 
second
With my frequency counter and timelab the shortest tau usable is 0.02 
seconds but the decrease of accuracy of the counter with this small tau 
starts to limit the measurement. But with Vtune of the TCXO at min or 
max the frequency deviation is well below 0.5e-9. In between Vtune 
produce bigger frequency deviations, don't understand why but if there 
is digital stuff inside the TCXO anything is possible I recently learned.
So I'm trying to use an SA/Modulation analyzer to look at the spectrum 
of overtones  with 1Hz resolution or watch the FM modulation analysis
The FM modulation analysis does not give understandable information. At 
the fundamental of the 10MHz the frequency deviation is reported as 
25Hz, for the harmonics of the 10MHz the frequency deviation increase 
with the harmonic so the 25Hz may be correct.
But the measured modulation frequency is 10kHz so I tried to output the 
demodulated FM into a good PC audio card and use audio FFT SW so see if 
there is anything recognizable but I can not see any peak above the 
noise. Listening to the demodulated FM also does not reveal a lot, just 
a very faint wistle of some tones.
The small span scans on the SA show 50 Hz sidebands, even when the whole 
TCXO is running on battery power so I'm confused.
In short, this requires still a lot of time to learn how to measure and 
to create understanding.
And investing in a timepod is currently not feasible.

Is there any reference material to be found on how to do these short tau 
stability measurements?
The method with the mixers Bob was referring to?
Erik.