It's been awhile since I cracked out the DSP so I am not sure of the
specifics (it turns out that normal precision with a well-selected reference
is plenty for my current project) but by oversampling and low-pass filtering
(an average over n samples is the simplest and crudest variety) that you can
effectively gain a couple of extra bits of resolution, IF the noise in your
system is of the appropriate type.  Sometimes extra noise (dithering) of a
known characteristic (Noise with a triangular PDF of and amplitude around 1
LSB is one choice) is even added right before the A/D.  This is most
beneficial if your original signal changes very slowly wrt sample frequency
and so the noise is strongly correlated with signal level.  Adding
intentional noise causes the signal to cross more A/D level boundaries,
thereby decorrelating the noise and making the process more effective.  4x
oversampling and filtering can gain an effective 1 bit of resolution.  Check
out comp.dsp for articles on "oversampling dithering resolution" etc  -
that's where I last read about the idea.

Nick


----- Original Message -----
From: "John Ward" <john@AGENT-J.COM>
To: <PICLIST@MITVMA.MIT.EDU>
Sent: Thursday, January 15, 2004 3:28 PM
Subject: Re: [PICLIST] [PIC:] more precision out of a 10 bit A/D


> I hear you and thats a great idea, however, i really need the extra 2 bits
of
> precision as the values i'm reading are repeatable and stable as is. The
> point is that with a 10 bit output, the values are a bit rough and i need
to
> see smaller changes
>
> J
> On Thursday 15 January 2004 23:11, you wrote:
> > > HI All, Hope you all have a very "entertaining " 2004 ;)
> > >
> > > Does anyone have an idea of how i can increase the sensitivity of
> > > a 10 bit A/D
> > > on a 12f675 ?
> > >
> > > In the old days with the 80c515/535 , we used to change the
> > > reference voltage
> > > after some basic measurements and then "squeeze" out a few more bits
of
> > > precision.
> >
> >         What about just super sampling? Say something like an external
> > sample and hold crircuit followed by multiple samples with your 10bit
A/D?
> > If your signal is not changing quickly you can even skip the external
> > sample and hold. TTYL
> >
> >
> > ----------------------------------
> > Herbert's PIC Stuff:
> > http://repatch.dyndns.org:8383/pic_stuff/
> >
> > --
> > http://www.piclist.com#nomail Going offline? Don't AutoReply us!
> > email listserv@mitvma.mit.edu with SET PICList DIGEST in the body
>
> --
> http://www.piclist.com#nomail Going offline? Don't AutoReply us!
> email listserv@mitvma.mit.edu with SET PICList DIGEST in the body
>

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body