Agilent 53131A Calibration

[Laval]

Hello everyone.

I have received my OCXO module for my Agilent 53131A Frequency counter and I installed it. Now, I want to calibrate the unit against my rubidium standard but there is one thing I'm not sure about. How long do I have to let the unit powered on for the OCXO to be settled and ready for calibration ? Someone on YouTube was saying 24 hours which seems like a lot to me. I didn't find any specifics for my time base model (see picture). Anyone have an idea ?

[KE5FX]

It depends on how long the OCXO has been sitting around unpowered.  Right now it'll be undergoing an accelerated aging process of sorts, in which it is likely to be drifting significantly more than usual.  There could also be some noticeable frequency jumps during this period.  Calibration will largely be a waste of time until it settles down.

24 hours minimum, I'd say.  Better to wait at least a week before assuming it will meet its manufacturer's stability specifications.  Some of them will take longer than that to stabilize, but can still turn out to be good OCXOs in the long run.

[Laval]

I see, I don't think that thing is brand new but better safe then sorry. I'll have the thing powered on for a week. Thank you for the answer.

[SCSKITS]

The OCXO should be close to spec after about 20 minutes to an hour. It will pass the calibration at that point.
For best long term results, the calibration should be re-run after the counter is ON for an extended period.
The OCXO on the board is NOS, but has been on the shelf for a long time, so will get better over time.
Note, with the counter plugged in (even with the display OFF in standby), the OCXO is powered and will get better over time.

ed

[SCSKITS]

You can also apply the rubidium 10MHz output to the input of the counter and watch the measured frequency approach 10MHz as the OCXO warms up.
Once it gets closer than about 0.5Hz the calibration should pass. Calibration can take up to about 5 minutes.
This quick calibration lets you get confirmation that all is well with the counter to OCXO board communication.
As I said previously, calibration should be repeated after the counter has been plugged in for an extended period.

ed

[KE5FX]

It depends on how long the OCXO has been sitting around unpowered.  Right now it'll be undergoing an accelerated aging process of sorts, in which it is likely to be drifting significantly more than usual.  There could also be some noticeable frequency jumps during this period.  Calibration will largely be a waste of time until it settles down.

24 hours minimum, I'd say.  Better to wait at least a week before assuming it will meet its manufacturer's stability specifications.  Some of them will take longer than that to stabilize, but can still turn out to be good OCXOs in the long run.

Example -- a 10811-60111 OCXO that I purchased in an eBay lot a couple of years ago had been sitting on the shelf unpowered until I needed to replace a bad 10811-60109 in a 5061A Cs standard.   I thought it made a pretty good demonstration of what can be expected when an older OCXO is first put into service. 

After an initial 48-hour burn-in period (not shown), this is how it behaved over the next 10 days:



Drift falls below the 5E-10/day spec within a few days and ultimately ends up close to 1.5E-10/day, but it's still not quite done settling down.  There would be little point in calibrating it during the first week of operation after long-term storage.



This shows how the OCXO's ADEV evolves over time, one trace per day, from lightest=earliest to darkest=most recent.  For the first 24-48 hours it drifted so much that the ADEV was meaningless.  As the first plot suggests, it's still slightly jumpy towards the end of the run, but the overall trend is improving.