[sdiy] DCO amplitude with filtered PWM

[Declare Update]

ah, sorry. by overcompensating integrator, I do mean saturating the integrator purposely via the DAC that compensates amplitude due to frequency changes.

I am not outputting saw and square simultaneously. I generate the square wave in code as a sub octave, then add that in code to the output DAC feeding the integrator. this makes alternating cycles of the sawtooth either saturate the integrator quickly, or ramp very slowly. the amount of this sub octave added the the DAC output is variable, and controlled by a pot. it is a little more interesting than just mixing in a sub octave square, but not by much.

“PWM” “works” by treating the integrator DAC as the threshold, and ramping the integrator reference DAC. as you can imagine, this doesn’t yield real PWM. I tried this and, while interesting, it was half baked as a PWM attempt. a comparator would certainly
work better for that specific sound. 

this circuit is running at 3.3V, with one DAC channel connected to the negative input of the integrator op amp via a resistor, and another DAC channel driving the positive input directly. this way it can ramp up or down, and reset to any level. 

I built this because I had a thought a while
back that I could simulate detuned sawtooths with one integrator if I could control slope, reset rate, and reset level of a DCO with an MCU. I whipped it up in max to confirm it works, but haven’t gotten around to trying that specifically in the hardware, heh. I’ve found that this much flexibility is not necessary in most cases, though it does let me generate just about any waveform I’m patient enough to work out. 

Cheers,
Chris


Sent from my iPhone

> On Aug 31, 2018, at 12:53 PM, rsdio at audiobanshee.com wrote:
> 
> Would you be willing to explain what you mean by “over compensating the integrator”?
> 
> 
> Does that mean you cannot simultaneously output ramp and square?
> 
> Does your integrator run into saturation purposely?
> 
> Does your integrator charge or discharge during the reset?
> 
> I can almost imagine how you’re getting square by over compensating, but not PWM.
> 
> Brian
> 
> 
>> On Aug 30, 2018, at 8:46 PM, Declare Update <declareupdate at gmail.com> wrote:
>> This core is indeed unipolar, and AC coupled after. In my case, I’m actually creating square and pwm shapes by over compensating the integrator, instead of the usual comparator, since everything is under tight control from the micro. my current scheme has the integrators reference (+ input of the op amp) fed by a DAC as well, so it’s quite flexible. switching to the much slower PWM scheme away from the very fast 12bit DACs on the spendy chips will limit its ability in that regard.
>> 
>> Now I’m interested to see if the Juno 6 pulse width changes with wide bends from the bender or LFO. I’ll have to check it out. 
>> 
>> Including the expo converter is seeming like a better idea with all of this in mind. sticking with the fancy chip and its “dank DACs” is feeling better, too :) These imaginary $0.50 DCOs taunt me so. 
>> 
>> Cheers,
>> Chris
>