[sdiy] vcos

[Tom Wiltshire]

On 8 Apr 2009, at 18:52, Simon Brouwer wrote:
>
> Paul Anderson schreef:
>> On Wed, Apr 8, 2009 at 9:15 AM, cs e <modulemania at gmail.com> wrote:
>>> Would examining an oscillator's spectrum on a spectrum analyzer or
>>> some software equivalent not be useful?  Everybody's talking about
>>> oscilloscopes.
>>>
>> A spectrum analyzer is extremely useful, though only for a sine wave.
>> With a triangle wave or square wave you'll see many other components.
>
> Isn't that the point? To see those other components and compare their
> magnitudes to those in the triangles and squares of other vco's?

Yes, but no spectrum analyzer I've ever seen would provide enough  
detail or resolution to identify the subtleties we're talking about.  
Probably that's just because I'm always on a budget! Maybe decent  
(euphemism for "expensive") tools would be able to do the job.

Still, it might be interesting even with a basic analyzer. I think  
this kind of discussion needs to be much more based on measurable  
facts than they usually are, and this is one way to provide some  
actual evidence, instead of the typical "VCO type Q is much better  
because it has/doesn't have jitter/leading edge artifacts/distortion"  
with no proof offered.

My own slant on this is that (like Seb) I think digital has reached  
the point (increased accuracy and sample rates) where we can  
reasonably expect to model more or less whatever analogue weirdness  
it is that makes a great analogue synth great. However, in order to  
do that, we first need to study and understand that technology and  
it's faults/features in some considerable detail. I don't think this  
information is widely available, although some research must have  
been done.

T.