N. T. wrote:
> That's a good point. And you could take not the first previous
> reading, but also, say, N-th previous reading, depending on the
> mechanical properties of the system. This would average readings on
> last N values.

Not a good idea.  Now the timer must not wrap for at least N pulses.

What you are really describing is low pass filtering the series of pulse
periods, although you're not thinking about it that way.  Your average of
the last N periods is a low pass filter, just not a particularly good one.

There is one advantage in applying a box filter in this case, which is that
if you box filter one whole engine revolution of pulses any irregularities
in when the pulses occur within a revolution cancel out.  In this case,
there are two pulses per revolution, so working with the sum of the last tw=
o
pulses might make some sense.  However, I would still compute the period of
each pulse separately, just keep the previous period around too.  Then do a
real low pass filter each pulse on the sum of the last two pulses.  That
should give you a high quality result.

The OP hasn't said what resolution and update rate he needs.  Most likely
something like a simple two pole low pass filter will be just fine.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.