Hi

I've implemented a precision edge-triggered frequency measurement device. I do so using 2 modes, pulse-counting and 1/time-period. The 2nd mode is more critical to my operation. The input is sine-wave-ish with an amplitude of 0.75 V and around 2 volts of DC offset.

I am converting it into a square wave using a simple RC + LM351 zero-detector circuit. I later realized (AFTER the PCB was made :D) that it should have been a schmitt trigger, to increase the noise margin.

So I used software to do the same: every pulse edge I disable the interrupt, run a little delay (0.1 msec) and check if the interrupt pin is still high. If it is, then I don't consider that pulse in my calculations. This should filter out pulses of frequencies higher than 10 KHz (right??)
Sort of a reverse comb filter?

My question is as to whether this method is a good way of protecting my algorithm from errors due to short, sharp spurious pulses, which is the main kind of noise i might face.

Is it comparable to the schmitt-tigger method?

Are there any advantages/disadvantages to this method? Aliasing is not a problem

kundi