[sdiy] DCO

[Magnus Danielson]

karl dalen wrote:
>> If it has an integrator and a digital counter that resets
>> the integrator, then it's a DCO. Otherwise not. This seems
>> to be the de facto definition.
> 
> Among who'm? DCO is/was just a marketing term not a god
> description of a particular technique! Not even reasonable!
> 
> Particularly if the reference oscillator are just sitting
> there on a specific frequency!;)
> 
> Hehe, one could say a Matrix6 is VCO, well the DAC modulates
> the tuning of the LC oscillator!
> 
> I'd say its a mental minefield! :)

I agree. Within the context of musical synthesizers, terms like "DCO" 
has by some been assigned a certain meaning. In general engineering 
speech it only denotes an (analogue) oscillator under some form of 
digital control, and the meaning shift widely as to what that means, 
which is context dependent. It can mean a crystal oscillator with a 
built-in DAC, it can mean coarse tuning of LC-oscillator by switching in 
varous amounts of capacitors (used in fullcustom or ASIC designs) or it 
can even include capacitance modulation (remember - on/off) to dither in 
  sigma/delta fashion the correct capacitance to in average have the 
correct capacitance and thus in average have the correct frequency.

For instance, you could use a DAC to steer the control current of CCO 
stage, and consider it a DCO.

So, in the end, there is no good general definition of what a DCO is, 
except that it is an oscillator which somehow is under digital control.

As a side note, early wave-table synthesis did not use 
NCO/phase-accumulator/DDS technology. They rather evolved from more and 
more elaborate waveshaping.

Another side-note is that even with digital control of octave, note and 
waveshape does not prohibit use of analog modulation, as the top 
oscillator may still be a VCO. Elka Synthex and Aelita uses this topology.

So, I may accept "Roland DCO" as a compromize term, in lack of a propper 
name. Analogue Waveshaped NCO should be much more descriptive.

Cheers,
Magnus