DAC resolution

[Fraser, Colin J]

> -----Original Message-----
> From: Paul Schreiber [mailto:synth1 at airmail.net]
> Sent: 09 March 1999 12:40
> To: Fraser, Colin J; jamos at technotoys.com; synth-diy at mailhost.bpa.nl
> Subject: DAC resolution
> 
> Sigh.
> 
> The *quick* version is:
> 
> a) if you are controlling decent VCOs, say 4-6 at a time
> b) you have "musical" ears and can hear 0.2% pitch deviation
> (this was documented 25 ears ago in JASA)

I have often used my cv controlled monosynths in unison with dco based
polysynths without any noticable beating.

> If an 8-bit DAC is spanning 1 volt, the dac has 1/2LSB accuracy.

Any quantised continuous signal has an average error of 1/2 LSB.
The accuracy of the resistors in a laser trimmed R-2R network must be
significantly less, otherwise I would not have the successful, tried and
tested results that I get with 8 bit convertors generating cvs.

> If you are controlling a monosynth and playing fast techno 
> and not beating
> VCOs in a slow string pad, there is a good possibility you 
> will not NOTICE
> the error. You may get a "good" DAC-08 that is like 1/8th LSB 
> error by luck.
> But the proper engineering solution is not to use one.

Have you tried this in practice or are you just reading the data sheets ?


Colin f