[sdiy] digital control of CV

Tom Wiltshire tom at electricdruid.net
Mon Sep 3 21:06:40 CEST 2007


On 3 Sep 2007, at 18:32, jure zitnik wrote:

> i'd be needing 16-24 CV channels for stuff like envelopes, filter
> settings, vco settings... i realize i'd need a more accurate DAC for
> VCO pitch control, but how much bitdepth really?

I mentioned the Prophet 5 used a 14-bit DAC for the VCO. 16-bit would  
be ample.
If the top seven bits represented MIDI note numbers from 0-127,  
there'd be 9 bits per semitone,
or 512 steps. This is about a fifth of a cent, considerably less than  
is audible.

> also, how would i go about making switches (like VCO waveform)
> digitally controlled?

Analogue switch ICs like the 4053 or 4051 are the way this is  
typically done, and it seems to work fine.

> no matter if i decide for muxing or multiple DACs, i'd still like to
> know how fast are those chips, would a serial multichannel (like 8
> channels) DAC be maybe too slow for stuff like fast envelopes?

Depends if you're intending to simply _control_ the envelope with CVs  
from the microprocessor, or whether you're going to _generate_ the  
envelope with the microprocessor.

For control, the rate needn't be that high, since no-one can twiddle  
a front panel control extremely rapidly. I've written code that  
samples potentiometer positions using a DAC at about 6KHz, and this  
is still probably well in excess of what is required.

For generation of envelopes, it's a whole different story. For that,  
I've used sample rates of 20 or 25KHz. This  produces good results.

> what about demuxing, how many channels would i be able to get on a  
> single
> line (12 bits for instance)...

Sorry, can't help with this one.


More information about the Synth-diy mailing list