Jon Nicoll wrote:
>
> Hello All
>
>         I want to incorporate a simple resistor-based DAC within a
> PIC application. I have in mind using an R-2R ladder, either directly
> on a PICs port pins, or (for other reasons) via a latch, and then to
> an op-amp buffer etc.
>
> I realise that there are limitations to this approach, but certainly in
> terms of speed, it should be adequate for my application. Has anyone any
> experience in improving the accuracy of this basic arrangment? I am
> thinking
> in terms such as:
>
>         - Voh and Vil levels and variability for the PIC/latch (logic family?)
>         - loading of the port pins
>         - resistor tolerances
>         - modifications of the simple R-2R ladder, alternate weightings etc.
>
>         Any suggestions gratefully received.


You can always trade analog speed for processor bandwidth and harware
complexity. If the processors isn't too busy and low speed is all that
is needed then one resistor and a cap delta modulation is fine. It will
work
well to at least the high end of audio frequencies. ( It can be
used it to generate DTMF tones within the CCITT specs for example)

The bigest problem with R-2R is the output impediance of a port and the
switching offsets.

Walter Banks