DAC King

[Fraser, Colin J]

> -----Original Message-----
> From: Chris MacDonald [mailto:macdonald at evenfall.com]
> Sent: 13 January 2000 08:33
> To: Rene Schmitz
> Cc: synth-diy at mailhost.bpa.nl
> Subject: Re: DAC King
> 
> Rene Schmitz wrote:
> > 
> > Divide 10.5833V by 127 steps = 0.00833V per step. Now its 
> Ok. If LSB error
> > can be neglected. As you say a matter of scaling, and accurracy.
> > 
> 
> I believe 12 bits would suffice for 10 octaves of 
> semitone-step control
> if the DAC had 1/4 LSB relative accuracy.  Using the seven most
> significant bits for semitone selection and holding the 5 least
> signifacant at zero:
> 
> 127 * 1/12V = 10.583 full scale DAC voltage
> 
> 10.583 / 4096 = .002584V LSB step size
> 
> .00258 / 4 = .000645V 1/4 LSB error
> 
> So the error would be +/- 645uV.

In practice I have found 8 bit convertors scaled for 83.333mV per step to be
perfectly acceptable.
I have been told I was just lucky. I have used 8 bit DACs on half a dozen
occasions - I wish I could apply that luck to lottery numbers...
I put an audio example at http://www.octavo.demon.co.uk/colinf/8bit.mp3
This is an analogue synth in one channel and a DCO based synth in the other
playing the same note.
The analogue synth is being driven by the lower 7 bits of an 8 bit DAC
(ZN558) scaled to give 83.33mV per bit.
Doesn't sound too bad (close enough for Rock and Roll).

A 12bit DAC should be more than adequate. I can get a serial input 12 bit
DAC for less than an 8 bit, so I no longer rely on luck.


Colin f