As a search of this site on RESET and RC RESET will probably show, there's been a lot of trouble associated with this seemingly "little" feature.

What's at issue is the fact that the part is a fully synchronous device, i.e. nothing will function at all until the oscillator starts. Then, for a while the oscillator is not entirely stabile, so its behavior is somewhat unpredictable.

Intel discovered this early on, and specified a VERY large (10 uF) capacitor and a relatively large (8.2 K) resistor as the RC combination for its parts. This makes for a very slow (a week or ten days, at least) reset. Read the reset spec's and what it is most likely to tell you is that it wants the reset to be positive for at least two cycles once the oscillator is stabile. I don't know how you're supposed to reckon with that. However, some 805x's have to drive the reset pin as well as taking it as input. That complicates the business of selecting a reset IC if you even want to use one. Some reset IC's generate an output that's valid only if they're "kicked" (as in a watchdog) once in a while, and at least one of them suggests using the ALE line to do that. Unfortunately, until the oscillator is stabile, you can't rely on ALE to be stabile either, so it's likely to generate a few extra pulses during the power-on reset, not that it necessarily matters.

8032 has no internal memory, so the flash-corruption issue doesn't plague them. However, since the later 805x's had the pulldown resistor built into the MCU, they required only the cap to Vcc. These didn't produce a RESET output, though, and, in fact, I'm not aware of any MCU's that don't have internal watchdogs that have this feature.

You'd best read the spec's on your MCU carefully in order to make certain that whatever reset IC, should you choose to use one, and I'm told it's advisable with flash-based MCU's, to ensure that it doesn't clash with the MCU on reset. Two devices driving the same signal seldom is pretty unless they're intended to work in that mode.

RE