I didn't intend to imply that you had to have a very fast processor, but, rather, that one benefit of using an integrator was that it made the duration of the bounces secondary in importance.

As I stated, you'll have one slope, presumably very close to zero, while the relay is open. Then, some time after you turn on the driver, the reed will "make" for the first time, which will allow current to flow into the integrator, raising its voltage. This will continue as the bounces become shorter and shorter, until, finally, there is a steady current flow into the integrator, which will be at the maximal rate at which it charges. So long as the ramp during the bounce is smooth enough, you needn't worry about missing any bounces, since you're not looking for them, but the slope should be "different enough" from the full-rate when the switch is no longer bouncing, that you can clearly recognize it by comparing successive deltas.

BTW, what's a 60386?

RE