Thanks for the suggestions - I am now looking at putting an opto coupler =
in the divider and using a pin from the PIC to enable the divider just =
before taking a sample.


>>> gfiedler@GLOBO.COM.BR 11/14/01 12:29PM >>>
I've never tried it with such an extreme source impedance. You have
basically two error effects: one is that you have to charge the internal
capacity in a certain amount of time (of which the external cap takes
care), the other is the leakage current of up to +-500nA (according to the
general ADC specs, specific part specs may vary).

This leakage current is not specified anymore than its max limits. If you
can live with the max. error the current introduces (which is about +-25
digits with 1M source imp.), this is not a problem.

Otherwise, it depends. If it's a one-off project, you probably can just =
try
it and see what results you get over your temp range. For a production
project, I probably would not trust it to be equal in all parts. Unless
somebody has mass production experience with the value of the leakage
current...

ge

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.