Potting calibrated components

Yes, there is a micro and I do use software calibration already, the problem was drift. Funny enough, baking the unit for many hours at 100 deg C fixed the problem, so I assume it was an interaction between the potting compound and the PCB.

We used to temperature-cycle precision potted modules, after curing, for at least 5x full cycles of the operating temperature range. My theory was it relieved the static stress on components that had been created by the shrinkage of the potting compound when cured.

I also found that potting in small sections (of about 5 cu-inch per section) reduced the total stress. It also gave a chance of tweaking the calibration after early part-pots.

Room temperature potting compounds were the worst for shrinkage (and resultant stress), even to the extent of pulling electrolytic capacitors apart. We settled on Emerson and Cummins 2651MM, which needed to be heated to about 70C to start the cross-linking and send it off.

Nope, not at all! Rather, alt.sleeping_on_the_couch or perhaps alt.law.divorce-court.

Is there a micro involved? If so, use software calibration. If no micro, then you need to find alternative methods of calibration after potting.