Martin J. Maney wrote:

A very thoughtful solution. (I was waiting for you to respond
to this thread, Martin).



Another solution you might consider is a voltage to frequency
converter, or more properly a current to frequency converter
(e.g. Analog Devices AD652 or National's LM331). Your approximately
1 meg ohm resistor can be configured with these devices such that
the voltage across it is constant. So as resistance changes, the
current changes. The change in current causes a proportionate change
in output frequency. You can use your PIC to count frequency pulses
over a fixed interval.

There are two great features of V/F converters:
1) You can trade digital resolution for conversion time. I.e. if you
want to wait for a second or two, it's possible to get as high
as ~19 bits of resolution (AD652).

2) You can chose the gating interval such that low frequency line
noises like 50 and 60 hz may rejected (I've exceeded 100db of both
normal and common mode rejection of 60 Hz by chosing a gating interval
of 33.33333milliseconds i.e. two cycles of 60Hz)



One more thing I would like to add to Martin's solution. Be aware that
if you are going to amplify the signal before sending it to the A/D, you
are going to have to subtract the 2.5 V "common mode" voltage. There are
several ways of doing this: a) a bridge configuration, b) a resistor
between the voltage divider node and a negative voltage reference.

Scott