Poul-Henning wrote:
Charles' design works great from the outside, but doesn't do anything with respect to the thermal energy expended by the encapsulated device themselves, which will cause convection in the inner box.
I have been using the technique for 30+ years, including with many OCXOs (which, obviously, generate significant heat) and have never observed any problems of that nature at the 1e-13 level. I did consider the possibility when I first started doing it, and tested two potential fixes: (i) putting a fan inside the box to homogenize the internal temperature, and (ii) filling the air space inside the box with irregular solid shapes to break up the convection pattern. I tested both methods extensively with instrumented sources, in many variations (fan speeds and orientations, mass and porosity of passive internal shapes), and did not find any difference at the 1e-13 level. I have occasionally used an internal fan just on theoretical grounds, but I have never measured any practical difference.
Thinking about it, this does not seem too surprising -- one would expect any convection to settle into a stable pattern and thus not to cause any temperature changes over time (once it is warm and settled). Whether this explains my results or some other effect predominates (for example, convection may move enough air in the limited space to achieve substantial isothermy), I have confirmed to my satisfaction that it is simply not a factor in practice at the levels we are concerned with.
If you test the "cast aluminum box" method and find that your results do not accord with mine, please publish them and we can discuss what might account for the observed differences and how the method could be improved. Until then, you are just posting speculative musings on the subject based on no data, which does not seem helpful.
And good luck fitting a cubic foot box with a surround of bricks into a 3U rack cabinet, or any other relocatable (much less, semi-portable) enclosure.
Best regards, Charles _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
