The apnote on reading resistance using capacitor charge time uses TMRO
(/RTCC) as it has a schmitt input.  
I've used this method in several apps and it works fine, but I'd like
to be able to use TMR0 for something else in a new design.   
Has anyone compared the performance using a normal port pin instead of
TMR0 ? are there any non-obvious 'gotchas', e.g. excessive pin
threshold variations between parts (obviously the method eliminates
most errors but vastly differing thresholds could affect resolution at
the ends of the range. PSU is stable and fairly quiet, I'm measuring
3-40K (thermistor), with a 220n cap.
I'll do a quick test, but any input from others would be welcome.