Dave D., thanks, I have considered the package pin capacitance and it is
more or less ok iff it is very stable. Maybe I will use a discrete JFET
front end which has a lower package capacitance. The main idea is to have
stable readings for up to 1 second (or at least predictable slope). This
includes thermostating the amplifier probably. I've done that before, it
works well. I have done some crude testing using a section of a TL074 and
I need it about 10 to 100 times better than that. The DIL TL074 drops
roughly 1V/sec on package capacitance alone + 5pF external at room
temperature. This is in line with the data sheet. If I'll add a JFET stage
before it I may even get there without a special opamp.

thanks,

Peter

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