> What would the consequences be of going to 20K on the midrange 10-bit A/D
> input? I've read the reference manual and the data sheet. If the input is
> bypassed with a 0.1uF or something of that ilk, it seems to be primarily
> limited by the worst case leakage current on the input vs. guaranteed
> accuracy. I'm willing to lose 1 bit or 1.5 bits in the A/D, and high
> temperature operation is not required.

I can see two issues with increasing the A/D source impedance beyond the 10K
specified by Micrcochip.

The first, as you pointed out, is leakage.  That will cause a higher offset
voltage, especially at the high end of the temperature range.  You can't
make the offset voltage go away, but it will be repeatable at the same
temperature for reasonable periods of time.  If you really needed to, you
could calibrate to an external reference, and adjust accordingly in
software.

The second is acquisition time.  A higher impedance will increase the time
constant for charging the internal capacitor the that A/D read the voltage
from.  The simple solution for this is to wait longer between selecting the
channel and starting an A/D conversion.


********************************************************************
Olin Lathrop, embedded systems consultant in Littleton Massachusetts
(978) 742-9014, olin@embedinc.com, http://www.embedinc.com

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads