Steve:
First comment, your 2.5V reference is below the book minimum.  But then, I
do have one production product that uses 2.5, and I don't have a problem
with it.

To protect pins, you can clamp a pin to a rail, with diodes, but I am never
too confident, as the diodes typically have a higher bias than the max volts
on the pin.  So your zener may clamp OK at 3.3 or whatever you use, but
going negative, it will be at 0.7, which is too far low.  AND you cannot
forget that a generic zener has a prety good leakage, and will modify your
readings if you have any series resistance in the line.

The other approach that I have seen, and I think it relies on the family of
parts in use, is to add a series resistor in the line, allow the voltage to
climb over the rail, and limit current with the value that keeps the current
below absolute max, which relates to the value that causes migration in the
chip.  But based upon family, this method may cause latchup.  I don't know
enough about all of the families to know which ones latch up.

But, Motorola makes a part that comes in 8 pins, and acts to clamp the
line.  Cant remember a whole lot about it, but it's part number is TCF6000.

In other words, I haven't got a real good answer.  Gluck.

Chris Eddy

stephen mcalonan wrote:

> Hello everyone!
> I am using the 5 channel A/D on the PIC16C73A running at 5V.  I am using
> an external voltage reference of 2.5V on pin RA3/AN3/Vref.  My question
> is: Should I use a zener diode, etc. to protect the port from voltages
> from 2.5V to Vdd+.6 or will this just give me a full scale reading?  Are
> there any other thought on protecting A/D inputs?  Thanks in advance.
> Steve McAlonan
>
> ______________________________________________________
> Get Your Private, Free Email at http://www.hotmail.com