AW: PolyModular

[Rene Schmitz]

At 19:44 05.05.99 +0200, you wrote:

>It's really not limiting, IMO. Asking directly: *Where* would the limits be
>?

I'm using my modular thru a (sort of) MIDI->CV-interface, that way I get a
voltage proportinal to pitch already. Of course one could do a pitch to
frequency lookup table inside the interface, but then I'd have to use wider
DACs etc. (This has been discussed several times on DIY)

>What I suggested was a VCO in the style of the Korg MS series. This has
>*both*, linear and expo inputs. You can use the expo inputs for anything you
>like,
>including KOV. 

But then the advantage is gone !? 

>I just say let's learn from Korg and connect the KOV to the
>linear
>input, and scale it by switching resistors for footage selection, and you
>avoid
>a lot of drift problems. It is still an expo (V/Hz) VCO  for any other
>modulation
>you choose.
>So, where are the limits ??

To be honest, I don't exactly know if I got this right. As I see it, you
use a 
KOV which is proportonal to frequency, then have a divider chain (div by 2,
4, 8). This gets into the reference current input of the expo convertor.
Now with that the expo inputs scales the V/Hz characteristic of the linear
input. You're right this is no limit. The problem I have with it is that
you would need to make a polyphonic keyboard with linear output to make use
of the advantages. 

Doing the maths on this shows that you'd get rid of the kT/q term if the
exponential input is 0V. (I=Ikbd * exp (0)) If the expo input would be used
to offset the frequency, then you'd get I=Ikbd * exp (q*Ube/(k*T)), the
right factor being dependant on temperature. This is of course the case
without compensation. I have the feeling that if one makes use of the exp
inputs for setting pitch one would need to compensate such a VCO. 

Another problem that I see is that of voltage offsets. This wasn't a
problem in the Korg synths, because the offsets were common to all VCOs.
Say if you have an offset that detunes VCO from 500 Hz to 502 Hz, and you
have set the second VCO one octave below, it would get half that offset
voltage, resulting in a detune of 1 Hz. So the relative pitches are
preserved. Indeed very clever.

Now, in a PolySynth, you'd usually have different offsets. (unless we have
trimmers for all opamps in the KOV chain) Using quad packages would here be
probably out of question. 

Just my random thoughts.

 René




             ,   : (uzs159 at uni-bonn.de)
|") [" |\ | ["   : (http://www.uni-bonn.de/~uzs159)
|"\ [_ | \| [_   : (http://members.xoom.com/Rene_Schmitz)