>
>
> The bit that doesn't make sense is that an ADC does not spit out a voltag=
e,
> rather you get a binary code word.  To get a voltage, you must have alrea=
dy
> made some assumptions about the ADC reference voltage and transfer
> function,
> and making these assumptions before measuring the actual performance is
> somewhat pointless.
>

Thanks Mike

I see. You are correct. I did assume the reference voltage was 1.023V which
is, as you say pointless (and completely wrong.)

A colleague mentioned that it is only good up to 1V and I corrected them
with 1.023V.  Ooops.

could also measure INL and DNL performance.  However, as someone else
>

INL and DNL ... Googling

So back to the drawing board.  I have several Fluke meters at my disposal.
I will research their resolution and accuracy.

With regards to the code word returned by the ADC, what would be considered
a good approach to determining the value to record as it dithers.  For
example I have applied a fixed voltage and a sample of the returned codes
are :

960
962
958
960
960
961
959
960
960
961
960
959
958
961
959

Should they be averaged, take the mode, median or perhaps the raw data
should be provided along with the average, mode and median?  How many
readings should be taken and how far apart in time.?

I guess it comes down to what will the data be used for.  In this case I
was hoping to provide some data that may prove useful to others.  So to
that end, perhaps I simply need decide and document the setup (temperature,
test equip, number of samples  and the decisions.)

I have much to think about for something I initially considered to be
straight forward.

Cheers

Justin
--=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
.