>I don't understand the problem, yet.
>If the signal causes an interrupt couldn't you just reference the OTHER
>signal to see if there's been any movement (just as Paul described)?
> If the wind was causing pulses, then why couldn't you
>decrement/increment the 'degree-of-turns' register accordingly to keep
>track of subtle movement?

Jitter may be faster than CPU's ability to respond.

We had the exact same problem back in 1987 when using the HEDS-7500 digital
pot by Hewlett Packard 256ppr. This was read by a 2 phase to 4 line dynamic
decoder thus giving us 4 pulse lines to give 1024 ppr (pulse per revolution),
the circuit also gave out an extra line which when inverted could be used to
drive a CPU - in this case a MC68705P3S. But we used the output of the decoder
to generate an instruction which the CPU executed directly - this made the
response time very fast indeed - can't do this at all on a PIC - does not
have the architecture. The decoder chip also handled the very high speed
jitter - provided that it did not go across two edges - then it was
recognised as a valid 2 pulses - ie two edges. Of course we had schmitt
triggers on each input. The HEDS-7500 was a 5v device and had the same
layout, spindle size etc as a normal panel mount potentiometer.

Without the decoder chip and using the conventional techniques here -even
with a flip-flop gave us the jitter problem - especially if the encoder
happened to be sitting right on the edge of a pulse !

Rgds

Mike