Firstly, what is the means of sensing the turbine? If it is something like a reed switch, then switch bounce is an issue. If it is hall effect, optical or suchlike then bounce is not an issue.

Then look at noise as has been suggested. What signal levels are we talking here? Proper shielding and processing of the input signal will help here.

Then look at your micro. Have you tried using a known pulse generator to verify your code.


Then there is the issue of turbulence and other effects of your method of measurement.


Your problem could be one or many of these issues all conspiring to make your life hard!