I got one of these compact GPSDOs as well recently (...not that I'ld really need it...
) and had some "toy time" with it today.
First thing I noticed is that the GPS receiver isn't one of the most sensitive ones...while others (tablet, cell phone) work behind the window, I had to place this receiver's antenna under free sky. Well, so be it... At least, the receiver then quickly locks on the satelites. But it takes a considerable time after this for the VCO to get accurate, at least 15 to 20 minutes for the "Error" led to turn off which corresponds to an accuracy of the 10MHz output of 100mHz as the manufacturer states (actually, it's slightly better than this when the light turns off).
To take some measurements, I connected the GPSDO to the "A" input of my HP 53310A Modulation Domain Analyzer and used my Rb standard (LPRO101-based) as an external reference for the MDA.
The performance seems to be quite good, the histogram is symmetric and resembles a Gaussian distribution quite well though it appears to be a little too "peaky". But I wouldn't worry about this, I've seen much worse reference oscillators.
The short-term jitter also isn't too bad at a 6mHzpp and a standard deviation of 900µHz (though I'm not sure how much my Rb reference contributes to this). The 1mHz digit of the average wanders up and down by a count or two over longer times. When I tested my Rb standard against another, identical one, I got two more stable digits of the average, so I guess this really is an effect of the GPSDO.
Finally, I turned the GPSDO by 90° along the longitudinal axis of the enclosure during the measurement, the effect is shown in the third photo. The jump of 15mHz is considerable but the frequency slowly returns back to where it's been before. Also tapping the enclosure results in frequency excursions so the GPSDO should be carefully isolated from any vibration. Otherwise, it seems to perform quite well.
Hope some members might find this information helpful!
Cheers,
Thomas
P.S. Sorry for the crappy screenshots -- had to move the MDA upstairs from the basement in order to get clear sky access for the GPSDO antenna and I didn't want to install the GPIB equipment to take proper digital screencopies as well...
P.P.S. I found in my GPSDO the 10MHz BNC's terminals not to be soldered to the PCB at all (manual jobby -- probably someone simply forgot them). This shows that it's always a good idea to have a peek inside such "el cheapo" gadgets prior to using them.