An independent opinion¶
I happened to notice an old Austron 2110 on a local trading site and bought it for the Austron/Sultzer 1150 OCXO it contained.
After some days the frequency drift have fallen below 1e-9/day which gives me a chance to show some Allan curves again:
The red lines are the HP5065 as I received it, against GPSDO. The lower line has a linear frequency drift removed, but I am not particular confident in that result.
For all the lines the rightmost bit should be considered highly suspect, as the uncertainty is quite large.
The green line is after my powersupply and C-field driver improvements, There is no signs of a frequency drift in the datafile.
The orange lines are the Sulzer 1150 against the HP5065, again with and without linear frequency drift removed. It seems to flatten out at 2e-12.
The blue line is the Sulzer 1150 against the free-running HP-00105 in the HP5065 (ie: not locked to the rubidium atoms), also with linear frequency drift removed.
It puzzles me that the blue line seems to follow the “GPS” hump quite closely. If there is a real effect there, as opposed to a mere coincidence, it’s mostly likely the aircondition.
The Austron/Sulzer 1150 is clearly about a factor of two better than the HP-00150 in the 10-100 tau range, and I’m tempted to plug the 1150 into the HP5065, just to see what happens.
But I have nothing which could actually measure what happens, so that experiment will have to wait.