[sdiy] CV/Gate delay project..

Jean-Pierre Desrochers jpdesroc at oricom.ca
Fri Jan 29 23:03:01 CET 2021

Very good idea Tom to use PWM like you did as a DAC mean..


> But now I’m straying into the DAC end, and you were really asking about the ADC end.

I’ll keep the quantizing stuff for another future module..

Well the final idea here is how to cheaply get the ADC/DAC transparent transfer
using a small micro like the PIC16F88 (SPI control).
Surely I’ll need external converters !

But you stated that even using an external ADC, 10 bits would be not enough
and I could hear staircase effects on slowly moving input CV’s ?



De : Tom Wiltshire [mailto:tom at electricdruid.net] 
Envoyé : 29 janvier 2021 16:50
À : Jean-Pierre Desrochers
Cc : Synth-diy at synth-diy.org
Objet : Re: [sdiy] CV/Gate delay project..


If you’re *quantizing* you can get away with less resolution, because you only need to know what *note* it was supposed to be, not the exact pitch. That means in theory you could cover the whole MIDI range with a 7-bit ADC. Of course, in practice, that’s not that realistic, but you can see how a 10-bit ADC could do the job quite reasonably. As you rightly point out, linearity is important, but with 10-bit linearity of even +/-2LSB, you still have more than enough resolution for quantising. Just not for anything else!



Rick Holt and I worked on a quantizer module that used the PIC’s ADC for the input over a several-octave range (sorry, can’t remember exactly any more - five. maybe) and used two PWM channels for the output. Ordinarily PWM is much too coarse and slow for a CV DAC with the filtering that is required, but given two channels, we had one that covered octaves 0-7 (of which we only used some) and with 3-bit PWM, the output frequency was in the 100s-KHz range, which makes the filtering easy. The other channel covered the semitone offset, so was set up as 4-bit, still in the 100s-KHz range. So we had a very cheap, very fast, very linear way of producing accurate Note CVs over many octave. The trimming was similarly pretty simple, since the octave range and the semitone range were independent. Just so long as you don’t need pitch bend or modulation!!


But now I’m straying into the DAC end, and you were really asking about the ADC end.




       Electric Druid
Synth & Stompbox DIY



On 29 Jan 2021, at 21:04, Jean-Pierre Desrochers <jpdesroc at oricom.ca> wrote:


You are right Tom..

I'm looking now for at least 14bits ADC.. 16 bits will probably be.

12 bits DAC's should be fine.

Maxim MX7705EPE+ 16 bits ADC could do (I think).


As a note here:

I saw that  <http://synthesizers.com/> synthesizers.com has a quantizer module Q171
that uses ADC/DAC process on incoming CV’s.

Here are the specs:



Panel Size: Single width 2.125"w x 8.75"h.

Quantization Method: Processor controlled ADC/DAC.

ADC Inputs: 10-bit with precision buffer/scaler.

DAC Outputs: 12-bit with precision buffer/scaler.

Gate Inputs: 0-5V minimum, rising edge.

Gate Outputs: 0-5V, 5ms on, 2ms minimum off.

Power: +15V at 30ma, -15V at 30ma, +5V at 50ma.

 <https://www.synthesizers.com/products/q171/q171data.pdf> https://www.synthesizers.com/products/q171/q171data.pdf

The CV inputs are fed to precision opamps OPA4277
then routed to a micro MEGA644P’s  internal 10 bits ADC.

The CV outputs are from 12bits DAC’s LTC1448.

So… I’ve not tried this module but assumes it works correctly
using a micro’s internal 10bits ADC and an external 12bits DAC.. ??





-----Message d'origine-----
De : Tom Wiltshire [mailto:tom at electricdruid.net] 
Envoyé : 29 janvier 2021 15:37
À : Jean-Pierre Desrochers
Objet : Re: [sdiy] CV/Gate delay project..


I think you’ve answered your own question:


> around 10 steps between each semi-tone.


That’s 10 cents, clearly audible. E.g. Not good enough.





       Electric Druid

Synth & Stompbox DIY





> On 29 Jan 2021, at 16:28, Jean-Pierre Desrochers < <mailto:jpdesroc at oricom.ca> jpdesroc at oricom.ca> wrote:


> Hi list,


> I'm working on a new module for my DOTCOM modular.

> It would memorize incoming CV's and Gate's state for 2msec up to 10sec 

> then spit the stream in steps of 1msec.

> The code for Gate memorizing is done.

> I'm now wondering if using a Microchip PIC 10bits ADC is linear enough 

> to read incoming CV values (0 to 8vdc)...

> There will be an input buffer amplifier that will take care of taking 

> down the 0to8vdc to 0to5vdc for the PIC.

> Incoming CV's 0-8vdc, 12 semitones/octaves.. that makes 1024 / (8 x 12)..

> around 10 steps between each semi-tones.

> My worry is the PIC ADC linearity precision..

> The 10bits memorized values will be output using a 12bits SPI DAC.

> I have a bunch of old PIC16F88 that I want to utilize here.

> Old but still usable for SPI.

> What do you think about PIC's ADC linearity wise ?


> JP


> _______________________________________________

> Synth-diy mailing list

>  <mailto:Synth-diy at synth-diy.org> Synth-diy at synth-diy.org

>  <http://synth-diy.org/mailman/listinfo/synth-diy> http://synth-diy.org/mailman/listinfo/synth-diy

> Selling or trading? Use  <mailto:marketplace at synth-diy.org> marketplace at synth-diy.org


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20210129/68fa7a7a/attachment.htm>

More information about the Synth-diy mailing list