[sdiy] How DCOs work

[rsdio at audiobanshee.com]

Yes, it's a good video: high-level enough that we can actually finish watching in less than 45 minutes; low-level enough to incite discussion and debate over the terminology used.

I think it's a bit hilarious that he scoffs at those who say that a DCO is an analog oscillator, but then says it's exactly the same as a sub-oscillator! I think that just as much criticism can be made of his claim that a DCO is exactly like a sub-oscillator, as any other attempt to define a DCO.

I still say that a DCO is an analog oscillator because it's core is an op-amp that charges a capacitor. That's almost exactly like a saw-core analog VCO, admittedly with some significant changes. It's not precisely like a saw-core, because any sub-oscillator can be derived directly from the digital pulse train rather than wave-shaping the saw, but that's the only part that isn't derived from a saw core and I assume some DCO circuits may not even include a sub-oscillator.

I would make a point of separating the terminology for master clock and oscillator. Yes, the master clock is a kind of oscillator, but not in the audio sense. Just because it is technically an oscillator circuit doesn't mean that we should think of a polyphonic keyboard as having only a single oscillator. There is a valid point to be made that the frequencies of each voice's oscillators cannot diverge from each other in some ways if they're all derived from the same master clock, but that slight limitation doesn't exactly mean that there is only one oscillator.

Personally, I wouldn't use the terms "analog" and "digital" to distinguish between the two types of master clocks. It seems clearer to discuss that an "analog" clock has a continuous range of possible frequencies, while a "digital" clock only allows discrete selections of frequencies even if there are tens or hundreds of thousands of possible frequencies.

In fact, the terms "analog" and "digital" are a bit too narrow, since they do not always entirely describe the system being discussed. It is important, I think, to consider discrete time versus continuous time, and discrete voltage steps versus continuous voltage. Most DCO circuits are analog in the sense that few have discrete voltages in their waveforms (the exception being the stepped saw of some Korg synths). Many DCO circuits are analog in the sense that they allow continuous frequency tuning, although a significant number of DCO designs are discrete in frequency.

Thanks for sharing the video, Adam, and thanks to everyone else for adding to the discussion!

Brian

p.s. Speaking of imprecise terminology, sometimes I say "square wave" when I really mean bilevel waveforms (that have only two voltages, high and low). A variable pulse width "square" wave is still "square" in my mind, even if it doesn't have a precise 50% duty cycle. However, I totally understand that I'm being lazy, and that a real square has a different harmonic series than other pulse waves.


On Oct 10, 2016, at 6:47 AM, Tom Wiltshire <tom at electricdruid.net> wrote:
> There's a lot of good stuff in that video. The discussion of the difference between having a fixed digital master clock and a high frequency VCO is good.
> 
> One of the reasons people argue over whether DCOs are digital or analog is that no-one seems to agree on how much of the circuit falls under the scope of the term "oscillator". In the video, he argues that the ramp core part is merely another wave shaper, since it would stop without the counter, and that would stop without the master clock, so there's only really one oscillator in the synth - the master clock Which is true in a way, but kind of means there's no such thing as a "DCO", since you've just defined them away. Instead, you've got an oscillator followed by digitally-controlled analog wave shapers.
> 
> But I'm nitpicking. It's a considered and knowledgable discussion of the technology, which is a rare thing with DCOs!!
> 
> 
> On 9 Oct 2016, at 23:22, Adam Inglis <21pointy at tpg.com.au> wrote:
>> I found this video interesting. It explains how DCOs work and the differences between the DCOs found in the Juno 60, 106, Akai AX 80 and the Kawaii SX240...
>> https://www.youtube.com/watch?v=mqY6cVJS9fo
>> 
>> He shows how some DCOs are “more analogue” than others, due to the implementation of the master clock.
>> He then goes through the schematics of said synths highlighting the relevant circuits.
>> At 29:00 he asserts that an inverted sawtooth sounds better in the bass end than a regular sawtooth - I’ve never heard that before.
>> 
>> 
>> Tom of course has a great page on the Juno DCOs
>> http://electricdruid.net/roland-juno-dcos/ 
>> 
>> I guess the most recent application of this technology is the Deep Mind 12
>> 
>> AI
>> 
>