You're fortunate in that your employer only requires you to do the easy and fun part of the job. I would therefore guess that you leave the engineering, which is mostly documentation, to someone else.


That's anecdotal evidence, perhaps, but in no sense is it proof. It's just like your case where the phone calls stopped. That may have simply proven that they were tired of complaining. I have to rely on what I have observed, particularly since nobody else has even bothered to examine the conditions of such failures.


That's true. One thing I believe I've pointed out before, is that some of these problems go away completely on 805x versions that use negative-going RESET. I'm persuaded that the problem with RESET is entirely due to the fact that the RESET signal is tied to a changing Vcc rather than to a fixed GND, and that the raw supply often enough can't force a rise time on Vcc that is adequate to start everything appropriately. I'd even guess that the occurrence of RESET malfunctions, whether they involve flash corruption as this one does, or not, will increase with an increasing rise and fall time of Vcc during the power-up and power-down transients. Because of the peculiarities I've observed during power-down, I'm more suspicious of that than of power-up.