[sdiy] Current driving a CD4046

[mark verbos]

Hi Grant,

Quite a few years ago, someone (Paul Maddox maybe) was building a  
wavetable oscillator like the Digisound VCDO. Rather than use a  
CEM3340 he wanted to build his own from normal parts. Since he was  
stepping the wavetable ahead once with each cycle of the oscillator,  
he also needed the VCO to go way up high in frequency.

As I recall, he decided to go with a triangle core based around an  
LM13600. I guess these keep their cool in the high frequencies,  
because they don't have as much reset time messing up their  
linearity, compared to sawtooth cores. Perhaps this will be the  
ticket for you as well. I mean, you only need the square out, right?

Mark






On Nov 21, 2006, at 7:24 PM, Grant Richter wrote:

> Just as a note of interest.
>
> I needed a very wide range VCO for playing the 256 step pages of  
> the Envelooper MARF.
>
> I knew a CD4046 would run 256 kHz at 5 volts which I needed to play  
> the 256 steps in one millisecond.
>
> I hooked pin 9 to +2.5 volts through a voltage divider and ran a  
> regular NPN exponential current source through a 10K to pin 11 (max  
> frequency). It produces a square wave from 15 Hz to 280 kHz with a  
> 100 pF timing cap.
>
> Since I don't need perfect 1 volt per octave tracking (just close  
> enough to half / double rhythms), I didn't test for precise 1 volt  
> per octave pitch performance.
>
> With more work it might work well enough to drive a counter for a  
> wavetable VCO, a Walsh generator or something else that needs an RF  
> clock at 5 volts.