Adding more OCXO insulation and ?stability?

[Jebnor]

I'm putting my 'reference*' OCXO in an enclosure.  I'm thinking about packing the enclosure with EPS foam to help with thermal stability.

I'm wondering if OCXOs need a minimum thermal dissipation for stability?  If I insulate it super well, does the heat element -- resistor, transistor,  etc -- bring the temperature over the target range due to minimum 'on' power or hysterisis?  Could temperature oscillations occur if insulated to well?

It's a CTI OCS 5A2B02, 10.000Mhz crystal. Ebay cheapy from a couple years back.

* The best I have, so it's my reference. Take that for whatever that's worth...

[floobydust]

With extra insulation, the OCXO can heat up faster but cool off slower. This increases system gain for heating, lowers it for cooling which I think would tend to more overshoot.  The heater is probably just a proportional controller, guessing it would go unstable if the OCXO was too well insulated and take forever to cool down after. It does need some heat loss.
It seems to be 3W max. and 1.25W typical. We don't know its target temperature or how stable the heat control is, you could look at current draw to see if it's stable.

I think all you really want is to shield it from drafts and sudden room temperature shifts, such as A/C blasts. Is your enclosure metal or plastic, ventilation louvers, PSU in there etc.

I didn't pay attention the the LTZ1000 thread as people there put covers and insulation over the IC, I thought to stop convection air currents or the cooling fan (in HP3458) from throwing the ref temperature around.

[ZigmundRat]

As @floobydust says, OCXO need a certain amount of heat loss, and there is no need to add any insulation- just make sure it’s not got air blowing on it or placed by another heat source.  Add a baffle if you want but it should have free air around it and allow for normal convection.

I don’t know why the thought of insulating an OCXO appears so often. I suppose it’s a reasonable question for the uninitiated, but it’s unnecessary and does not help stability. They are supposed to be hot and dissipate heat.  I’ve been in telecom for >40 years and I’ve never seen an OCXO insulated. None of our products have ever been configured that way and none are now. Don’t get me started on the folks that want to insulate Rb oscillators 

[Jebnor]

I don’t know why the thought of insulating an OCXO appears so often.

For me, it's "insulate to keep hot stuff hot, and cold stuff cold".

In this case, I'll let it breathe  out of the windy air.

Thanks to you both.

[Doctorandus_P]

One advantage of adding extra insulation to an OXCO is that it lowers power consumption for the heating, which could be important for battery operated stuff.

It reminds me of this old EEVblog #1139 @05:10 where Dave tears such a thing apart.

[james_s]

I was playing around with some OCXO's a while back and I was surprised how sensitive they are, putting my hand on top of the can to suck some heat out was enough to cause a significant measurable drift. "Significant" in a relative sense that is, the amount of drift was a heck of a lot lower than you'd see with a non-ovenized oscillator without having to touch it. Definitely a good idea to shield it from drafts but I don't think you want extra insulation.

[floobydust]

One advantage of adding extra insulation to an OXCO is that it lowers power consumption for the heating, which could be important for battery operated stuff.

It reminds me of this old EEVblog #1139 @05:10 where Dave tears such a thing apart.

That OCXO has a 115VAC heater, according to the can   I think it needs the insulation because it's bolted down to the "cold" chassis which acts like a heatsink.

[CaptDon]

I have an adjustable OCXO at 10mhz. I super insulated it and it does well. It stabilizes with no problem. I also found that by shifting the actual 5vdc supply voltage from about 4.9 to 5.1 I can get about +/- .25 hz adjustment range. I have a ten turn pot on the +5 regulator. When making the adjustment you have to tweek and then wait for it to stabilize at the new VCC voltage. I find that if I need to make a tweek I will overshoot the adjustment just a hair and it will stabilize right on target. I made the mechanical adjustment only once. There is a trimmer capacitor located under a removeable screw.