The only thing that I did, myself, was to put a switch between unregulated power and the 5-volt regulator, as opposed to relying on the AC power switch to the 20 kg transformer on my "old reliable" (>> ~50 ampere) power supply to switch power to the circuit under observation. There is, in this case, a 0.1 ohm resistor between the unregulated supply and the regulator, but, in this context, it has little effect, and I've tried the thing without it. I figured that with switching hundreds of times, "things" could happen, and the resistor would prevent some damage. The addition of a "protection" diode around the 5-volt regulator, was a step, also, to protect things before doing all this "playing" with the supply.

I'm still having problems making the logical connection between all the anecdotes about small system failure and 805x RESET. I'm also having trouble explaining the observed BBRAM corruption that frequently occurred when the decay of Vcc was slow, but which went away when the on-board capacitance was significantly reduced.

Your approach seems interesting enough that I'm likely to give something similar a try, if only to allow for automatic and somewhat more rigorous testing.

RE