10 MHz 0.1 ppb OXCOs on ebay for $3-20

[rhb]

FWIW there are a bunch of bare modules and eval board clone versions of the CTI OSC5A2B0 for crazy cheap.

Best ebay search string is "OCXO 10MHz Frequency Module Crystal Oscillator Constant Temperature" for the $20 plug and play versions and "OSC5A2B0" for the $3 bare modules.  I bought 5 of the former and 10 of the latter.

The two eval board versions that I've tested hold 0.1 ppb with some scrap bath towel to mitigate drafts.  Limit of my GPSDO is around 0.001 ppb best case.  So we are talking very accurate frequency references for $20 delivered.

They arrived set within 1-2 ppm, but a bit of tweaking using a GPSDO and scope got them to 0.1 ppb.  I'm going to mount one in a case withnone,  just insulation and another in a case with active temperature control and then monitor all 3 long term.

[JohnPi]

Any measurements on jitter or Allan variance ? A board like this with a simple DCDC and little filtering or isolation is unlikely to have good short-term characteristics.

[MIS42N]

Are you talking parts per billion or parts per million. I have played with OSC5A2B02 for a few years now, they will reliably settle after 15 minutes warmup to 10MHz within 0.1Hz which is 10 ppb. To get them to settle better than that they have to be powered for days (if used every few days then 2 days is OK. If not used for months can take a week). Also, their sensitivity to the control voltage is about 1ppm/V so to get 1ppb requires the control voltage stable to 1mV or in your instance 0.1ppb requires stable to 100uV, which needs a pretty good voltage regulator (not a run of the mill 7805).

I use them in a cheap GPSDO design, the best I have had reliably is stable to 0.04ppb while being disciplined. https://www.eevblog.com/forum/projects/budget-gpsdo-a-work-in-progress/msg4804529/#msg4804529

Regardless of ppm/ppb etc. they may not be the best OCXO around but certainly great value for money.