[sdiy] Ten new student projects up!

[René Schmitz]

Hi Aaron and all,

Aaron Lanterman wrote:
> * Scream Machine (was supposed to be a Rene MS-20 with CMOS buffers - 
> the result is technically "broken" but sounds so insanely cool I told 
> them _not_ to fix it, as I would have no idea how to reproduce the crazy 
> sound the got!)

I think I can shed some light on what might be going in there.
The key "mistake" here is the supply to the 4069 which is waaay to soft.
Now what I think that is occuring here is an effect called motorboating, 
that is there is an extra unwanted feedback somewhere with enough 
loopgain and phase shift that an oscillation occurs. (Filter time 
constants, coupling caps and so on.)
Here the feedback is via the supply rails. Something which I often 
encountered in tube circuits, where the supplies are often very soft as 
well. And if you're not careful with decoupling and coupling time 
constants this will lead to audio rate oscillations. (Usually its 
sufficient to make the timeconstants of decoupling caps larger than that 
of the coupling caps, to make this quiet.)
The "remedy" would be to make the supply to the 4069 stiff for audio 
frequencies. I personally wouldn't use a 1k/1k divider for the 7.5V, but 
an emitter follower to buffer the divided supply, but if you'd want to 
stick with that scheme, one could try to add some fairly large 
electrolytics to the supply pins to short out that unwanted feedback to 
GND. And that might be easy enough to try out, without haveing to change 
the circuit much otherwise. Even if just to check out if this is really 
the cause, and changeing it back to the weird behaviour later.

Cheers,
  René

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