They're nice, but not necessary. I've been using RC resets on microprocessors and microcontrollers as long as there have been such things, and, frankly, I've never encountered the sorts of difficulties that have been discussed here. If one carefully reads the specifications of the device in question, and they do vary considerably from device to device, one quickly sees what's needed in order to effect a good reset from just the RC circuit. I think what confuses many people is that 805x's have a positivie-going reset, for which we have to thank our friends at Intel.

The manufacturers haven't made this easy. Here's an example of a reset voltage specification and "note."


To me, this says that you might be satisfied with occasional circuit failure. I, personally, wouldn't, though. True enough, these guys sell supervisors and reset circuits, so they may have a vested interest.

Here's one from SIEMENS.


Some of these devices have negative-going reset, and some have positivie-going reset.

The reset circuit uses an RC to tell it that a reset is happening. I believe I've seen as many misapplications of reset circuits as I have of RC reset circuits.

Quite frankly, I think it would work just as well to use a Schmidt-trigger as any of the low-cost reset circuits.

I'm persuaded that, while reset chips can be used quite effectively, they're mostly used as an excuse not to read and interpret the reset spec's of the MCU in question. I haven't bought 'em, so I don't know where they fit in the pricing scheme of things. I suspect they cost more than a couple of Schmidt-gates and an RC circuit, though.

Perhaps it's time this subject got a rigorous treatment here.

RE