Normal Board Temperature?

[brouhaha]

I measured the bottom of the board at 65C when it is idle, and 70C when doing a dd from /dev/random.  It's hard to get pinpoint location accuracy with my IR thermometer, but the case temperature of the boost regulator seems to be somewhere over 75C when idle, and I can only imagine what the junction temperature must be.  A friend with a high-end ($25K-ish) FLIR camera has offered to image the board, which would give us fairly precise case temperature information for all of the components, but we won't have a chance to do that for a few weeks.

Another friend with more switching regulator experience looked at the board and suggested that the inductor may be physically too small. If the inductor saturates, that makes the regulator hideously inefficient.

I also suspect that the lack of a bulk capacitor on the input to the boost regulator is likely to make the 3.3V supply noisy. Ideally there should be at least a 4.7 uF capacitor. Is there any particular reason why no bulk capacitor was designed in, and why the input to the boost regulator input is from the 3.3V rail rather than Vbus?

My unit works fine, so none of this should be construed as a complaint, but I think if a few changes were made in a future production run, it would result in improved long-term reliability.