Jim,
Are you 100% sure that your AD noise is not caused by the PWM?
I'm asking because I did precisely noise measurements on PWM,
(logic output only, without heavy loads) and I know the PWM  output it's
one
of the greatest agressor for the analog AD sigals in most of the PIC
microcontrollers I work with.
The problem is worse at any other duty cycles than 50%.
You can see this if you have a good spectrum analyzer (even on 8 bit FFT
oscilloscope it's quite visible).

For a light dimmer, 2 bit deviation at 10 bit resolution should be
invisible for most of the
human eyes. The problem is not deviation but the flicker.

Vasile

On Fri, Nov 27, 2015 at 9:38 PM, Jim Ruxton <cinetron@passport.ca> wrote:

> Thanks everyone for some great suggestions to explore. I was expecting
> there would be a "stock" solution to this particular problem other than
> the deadband idea which seems popular. The issue with perception of
> light levels is really mostly on the low end of the brightness scale .
> The worst being when the light is either barely on or completely off. I
> am using LEDs so no forgiveness as in an incandescent bulb.  Perhaps a
> custom lookup table may end up being the simplest approach here
> Jim
>
> On 2015-11-27 01:14 PM, Bob Ammerman wrote:
> > >From personal experience, I know that at the lower end of the brightne=
ss
> > scale a 1/256 change in trigger phase for an AC phase control dimmer is
> very
> > visible. I wouldn't be surprised if 1/1024 is at least noticeable to so=
me
> > people.
> >
> > ~ Bob Ammerman
> > RAm Systems
> >
> >
>
> --
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>
--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.