Odp: [sdiy] DAC for MIDI2CV

[Roman]

I'd have to look at my printed AD catalogue to check the specs
(ADI site seems not to have AD1856 anymore...), but anyway:
audio DACs in general have poor gain error, offset error and drift.
Something you really need small in MIDI2CV. Besides they use
internal Vref, which doesn't have to be precise.
As for INL and THD, fair estimate is this:
INL = 10^(THD/20), or THD = 20log(INL)
INL in full scale ratio of course.

I don't know a DAC to recommend. Those chips you mentioned
seem to be the cheapest option. If you can find them.

I wanted to do the same thing, i.e. 16-bit dac for CV and
pitchbend, but gave up the idea, mainly thanks to members
of this list. It may be cheaper and better to have separate
DACs for every task, than have 16-bit sucker doing
everything.
We had a flame war about DACs few weeks back. Check
the archives.

Roman


----- Original Message -----
From: Seb Francis <seb at is-uk.com>
To: synth diy <synth-diy at dropmix.xs4all.nl>
Sent: Friday, October 04, 2002 9:08 PM
Subject: [sdiy] DAC for MIDI2CV


> Hi Again,
>
> Thanks everyone for your input on my VCA tempco question.  The VCA is now
working beautifully, and passes the "hairdryer" test using a single-opamp and
tempco potential divider input stage ;-)
>
>
> I am back to the design of my MIDI2CV converter, and the thing that is causing
me the most trouble is finding the right DAC for not too much money.  I want at
least 14bit +-1LSB, but 16bit would also be fine and they seem much more common.
I want to add things like pitch bend, and have LFOs in software, so high
resolution is important.
>
> But, I'm having real trouble in UK finding a reasonably priced one (e.g. <£20)
which has low integral linearity error.  Having spent a lot of time looking
through datasheets, it seems accurate DACs designed for control applications are
generally expensive (it also needs to settle reasonably fast because it's
running 16 CV outs).
>
> However, I have seem quite a few so called "PCM Audio" DACs (e.g. BB PCM56, AD
AD1856N), and these _seem_ cheap and well spec'ed in comparison with control
DACs.  But none of the datasheets quote the integral linearity error - they give
only the differential linearity error and THD.  Am I to assume that the integral
linearity error is quite high on these types of DAC, and they are not suitable
for pitch control applications?  I'm getting a headache from reading all these
datasheets!
>
> If someone can point me to a suitable 14 or 16 bit serial input DAC (available
in UK) that would be perfect, but failing that a general answer about the
suitability of "audio" DACs would be much appreciated.
>
> Thanks in advance,
> Seb
>
>