This is a topic that somewhat interests me, so hope you don't mind if I chime in!

And Murray knows much more about all this than I do, so please correct me if I'm wrong anywhere!

What you're describing is the ADC and DAC - the thing that takes the digitally stored patch and converts it to/from an analog control signal. The ADC and DAC can be any number of bits - 8bit,10bit,12bit,16bit, etc, this essentially determines how many "steps" there will be in the digital<->analog conversion.

The jupiter 8's DAC was actually upgraded from 12 to 14bit in it's lifecycle. However, this made very little practical difference, technically a patch's settings might be stored slightly more accurately but in practice the difference is just too minor to notice and the change likely had more to do with parts availability and supply.

Essentially, you're turning that slider or knob which is a 0-10v control signal into either 4096 (12bit) vs 16384 (14bit) "slices". The value stored will be some number within those ranges - if the knob is at max or 10v, the ADC would theroetically convert that into the maximum value: 4096 or 16384, if the knob is at minimum, the value would be 0. If the knob's in the middle - you'd likewise expect a 5v voltage, the ADC to generate a value around 2048 or 8192 - on a 12bit or 14bit jupiter 8 accordingly.

Now - this is just a high level technical overview and the actual implementations likely vary, they might be 16bit, or 0-5v, or -5 to +5v, and the way the values are stored digitally will vary based on the coders discretion i'm sure.

The jupiter 8 and prophet 5 do have hardware env generators, similar to those other synth's Murray mentioned, so in their case, the DAC is providing them a control signal, and they generate the envelope, as opposed to the computer generating the entire envelope directly, as was done in some later synths. Likewise with the VCO - you can almost just imagine it's storing/recalling the values you're feeding to the VCO.

-Jordan

On Wed, May 02, 2018 at 01:04:48PM -0400, Chas D 350ypvs@... [kiwitechnics] wrote:
> Thank you Murray for explaining the complex workings of the JX-3P (and
> other DCO synths) so clearly and precisely!
>
> Out of interest, how do synths such as the Jupiter 8, Prophet 5 and
> Oberheim OBXa store their patches considering they are VCO rather than a
> DCO synths? I vaguely recall that the Jupiter 8, or the earlier models at
> least, were described as "12 Bit" so I assume that despite the VCO sound
> generation they are still using some form of digital circuitry for patch
> storage and tuning?
>
> On 2 May 2018 at 03:32, Murray Hodge yahoo@... [kiwitechnics] <
> kiwitechnics@yahoogroups.com > wrote:
>
> >
> >
> > Almost all synths that have the ability to save a patch use a MUX to
> > control all the voltage controlled items in them and this part of the
> > hardware is completely digital. These controls include filter cutoff,
> > Resonance amount, VCA Levels etc etc. The big difference between 'digital'
> > & 'analogue' synths is the way the note frequency is generated. The JX-3P
> > for example uses digital dividers which output a square wave and the
> > frequency (divide number) is written into them each time the note changes..
> > This then needs to be converted into a Saw wave using hardware. The SH-101
> > for example uses an 'analogue' oscillator chip (CEM3340) and the frequency
> > is still controlled via the MUX and the wave forms are done within the 3340
> > chip.
> >
> > Unless the synth has a hardware ENV generator (for example the Korg
> > PolySix, Korg Poly61 use a SSM2056 ENV chip) all envs are generated in
> > software and output via the MUX to the various output parts that need
> > control. Calculating the various control outputs is a very complex part of
> > the code as there are many sources that need to be mixed together to make
> > the final output for each control. For example the VCF cutoff may have LFO
> > (various sources normal or plus), ENV level (various sources normal or
> > inverted), Keyboard note (including bender & portamento) and note level,
> > matrix output, control knob level and so on. Do this for each voice and
> > each control and you can begin to see that the cpu is pretty busy. The
> > JX-3P has 29 MUX control outputs and each of these has to be recalculated
> > in real time and this takes up a lot of the processing time. When it is not
> > doing this the cpu also has to scan the keyboard, scan the switches, scan
> > the bender, send and receive midi and display all the leds.
> >
> > In short the signal path from the DCO output to the synth output is
> > analogue but everything else and all the controls are digital. Also if the
> > DAC has a fault all the MUX outputs will also be faulty. If any of the 29
> > outputs are working then the DAC is working.
> >
> > MH
> >
> >
> > On 2/05/2018 2:28 PM, Chas D 350ypvs@... [kiwitechnics] wrote:
> >
> >
> > Florian,
> >
> > We appear to be getting terminology confused with each other.
> >
> > The JX-3P uses digital CONTROL, but its actual sound generation is
> > analogue (oscillators, amplifier, filter, chorus). I believe the envelopes
> > are software generated. Many people incorrectly state that the 3P is a
> > digital synthesizer, when it's actually analogue with digital control
> > circuits.
> >
> > A Yamaha DX7, Casio CZ and Roland D50 are fully digital, and use "digital
> > oscillators", entirely different to "digitally CONTROLLED oscillators" as
> > per the 3P. Many times I see the 3P incorrectly described as having digital
> > oscillators or being described as a digital synth. A digital oscillator
> > requires a digital to analogue converter to convert digital data to
> > analogue audio, much the same way that a CD player uses D/A converters to
> > convert digital information on the disc to analogue audio where it can be
> > heard with the human ear.
> >
> > This is what I took the OP's post to imply when he talked about D/A
> > converters being faulty, because if it were a fully digital synth, then
> > that would explain the loss of all audio. With the 3P, that wouldn't apply
> > because of the 3P's analogue signal path. If the OP was talking about the
> > 3P's digital control circuits with the digitally controlled oscillators,
> > then my original comment wouldn't apply.
> >
> > I guess half the problem with many people's misunderstanding of the 3P,
> > along with other DCO synths, is that they take a hybrid approach by mixing
> > digital technology with analogue sound production. The lines get even more
> > blurred when you bring in the likes of the Korg DW series, which flips
> > things over by using digital waveforms for fully digital sound generation,
> > and then using analogue filters and amps!
> >
> > Hopefully the OP's problem was solved simply by changing the internal key
> > velocity parameter!
> >
> > On Tue, May 1, 2018, 5:39 AM Florian Anwander fanwander@...
> > [kiwitechnics] <kiwitechnics@yahoogroups.com> wrote:
> >
> >>
> >>
> >> Hello Chas
> >>
> >> On 01.05.18 01:43 , Chas D 350ypvs@... [kiwitechnics] wrote:
> >>
> >> You can rule out a digital to analogue converter
> >>
> >> Though I think, that the OP has different problem: That is not true. The
> >> D/A-converter provides everything for the sound control. For example the
> >> envelope for the VCA. If the D/A is broken, then there is no envelope or
> >> gate for the VCA and so there would be no sound.
> >>
> >> Florian
> >>
> >
> >
> >