I have no trouble getting one circuit to reset when a stiff power supply is provided. That same circuit, whether using the "classic" 10uF cap-to-Vcc and 8.2K pulldown on reset or using the MAX1232 connected as in the app-note, behaves the same. When I use a Lambda 20-amp 12-volt "5%" SMPS, it fails 100% of the time, despite the "proper" input voltage, and it fails, similarly when I use a 500 mA wall-wart supply to an on-board LM340-5 regulator. However, if I use a Power-One 12-volt 6-amp linear supply to feed it, the thing seems to reset properly every time. I say "seems to" only because I've only tried this about 30 times. It should, of course, be run through 100000 cycles to gain more statistical information, but I'm not ready to do that, yet.

I'm pretty well convinced, that if one provides a proper supply with a very fast (<100 uSec) rise time, to the CPU, one will see much better behavior than what one sees when charging 5000 uF of capacitance from a half-amp regulated supply. I'm just not sure how to do that, yet, nor am I aware of what unforeseen consequences this might inject. I suspect it's a two-edged sword.

RE