Hello Harold,

>         Though I haven't used that chip (my experience is with the 16c74),
you
> should be able to get a maximum of 1 lsb jitter (maybe 240 to 241) when
> the voltage your feeding in is "close to the edge" between values. It
> should be stable when you are not close to the edge.

    Yes, I agree with it.

>         In my view, instability is due to either noise on the input or
noise on
> the reference. There could also be noise due to a high source impedance,
> though I think this would be a slower variation in level due to input
> capacity charge time and variation in input bias current with
> temperature.

    As I'm making tests on a proto-board, I think the problem is noise
*everywhere*. I hope the final product (on PCB) will be more stable.

>         My trick to getting stable A/D readings is to use a ground plane
on the
> board (I do a copper fill of both sides) and use bypass capacitors right
> at the pin that needs bypassing (to the ground plane).

    Do you mean a copper fill *around* the copper tracks? With bypass
capacitors, do you mean a cap between the PIC A/D input and ground? What cap
value?

> I've got a product
> that has 24 slide pots driving 4051 multiplexors driving analog inputs of
> the 16c74. No extra filtering, just pot driving 4051 driving analog
> input. You can set the pot to a position and have an absolutely stable
> A/D reading. If you move it a bit you'll get to a point where it switches
> back and forth between the current number and the next number. The amount
> of noise determines how wide this indeterminate area is. Ideally the A/D
> would settle on one or the other.

    Is it a DMX lighting console? I'll start to develop one soon and also
pretend to make this way.

>         So, again, I haven't used the chip you are, but if the A/D is the
same
> as that in the 16c74, it should be possible to make it stable.

    I think all the Microchip's A/D are equal, besides the bits number.

    Best regards and thank you,

    Brusque

+------------------------+
|     Edson Brusque      |
| Research & Development |
|  www.citronics.com.br  |
+------------------------+