[sdiy] oscillator jitter / phase noise

[René Schmitz]

Hi Martin and List,

> In some old postings it was said that "old" vco designs
> "could" sound better because there is some random process
> leading to phase noise/jitter.

I think it could be an important part of the entire picture.
But other factors have also be taken into account, like
(since were at it at the moment) the finite reset time, parasitic 
locking and so on.

> This would then lead to very interesting phasing pattern,
> if two or more oscillators beat against each other.
> Obviously a quartz stabilised oscilator, DCO or software
> oscillator will not have this feature.

Unless its a DCO clocked from a clock source with lots of
phase noise. And there is a place to start. Does a DCO start get less 
sterile when clocked from a jittery clock? With a uC that has Xtal and 
RC oscillator options this shouldn't be difficult to compare.

> To me it seems the only explanation, since the actual
> waveform is not very important (after my experiments).
> 
> For the phase noise I could imagine current source noise,
> integrator noise and perhaps most important comparator
> noise. Also noise on the power supply comming from other
> modules could be considered.
> 
> Anyway, has this ever been investigated?
> What do owners of old gear say?

I have often heared from people who prefer the sonic properties of "less 
precise" circuitry over "precision designed" circuitry. In the end drift 
is nothing else than the low frequency corner of phase noise. And can be 
fought against by lower noise opamps, better comparators and so on.

> If not, what would be needed to investigate this?
> Would a very long audio sample be sufficient to make 
> statistics and meaningfull spectral measurements?
> Or is the usuall 20kHz + bandwith too limited? 

One could try a double blind test (a poll perhaps?) to see if people can 
hear a difference. Thats what you're after, perceived sonic properties, 
right?
You would need a reference point to compare any phase noise measurements 
to. Else you have a nice number, but still you don't know if its "warm" 
or "cold". But such polls could generate that.

Cheers,
  René

-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159