4046

[Don Tillman]

   Date: Thu, 21 Nov 1996 18:21:40 GMT
   From: chordman at concentric.net (Zippy)

   Till:
   >The 4046 is so incredibly cool and bizarre, that trying to make it
   >operate like a traditional VCO doesn't seem reasonable.  I mean, it's
   >not exponential and it's only got a square wave output -- that's a
   >lousy VCO.

   Well, I agree that it's a cool and bizarre part, but I wouldn't agree
   that it is necessarily a 'lousy' VCO.  Perhaps it doesn't suit your
   particular needs, but just because it isn't exponential, doesn't make
   it a 'lousy' VCO either.  A lousy VCO (IMHO) is one that won't stay in
   tune or that doesn't track the CV through a usable range (among other
   things).  Limited in some senses, yes.  Lousy? not.

I meant lousy in a "traditional VCO" context, just like I said.  With
only a square wave out (can't waveshape that much) and 1% typical
linearity, it's nothing to write home about.

Besides the readily available 566 chip does two waveforms (triangle
and square, sawtooth available with four extra parts) with better
linearity for a few pennies less.  Cost was the original topic, right?

   Don, many of us cannot devote a large amount of time to synth
   building.  I am one of those people.  I have already built VCOs and
   several wave shapers, so the addition of an additional VCO square wave
   output isn't really that bad.  In fact, my guess is that it would
   thicken my sound, certainly not detract from it.  I pretty much resent
   being told to 'get off my duff' and to 'get creative'.  Who's to say
   that using this part in this manner *isn't* creative?

Slapping in an additional VCO doesn't sound especially creative to
me. 

But let's say you have a synth at home.  With linear VCOs.  And let's
say you want to fatten the sound with an additional VCO.  Okay great.
Wouldn't it make sense to copy one of the existing VCOs, so at least
any nonlinearities and temperature dependancies have a chance of
tracking?

   Perhaps you can share with us what you have done with it...  That
   would be considered constructive by all here on DIY.

I personally haven't done much with them, but here are some ideas:

First know that this is a phase-locked loop chip, with two (three on
the Philips version!) very different phase detectors with very
different operating characteristics.  

So the most obvious use would be to track other oscillators.

Or other sound sources.  

Add noise to the sound source for deliberate mistracking.  

Screw with the loop characteristics for undershoot and overshoot.  

Add a bias to the loop for phase shift.  

Add a divider chain to lock on harmonics, subharmonics, and ratios.

Use the xor gate to lock onto harmonics within a specific frequency
range (something like Hammond foldback).  

Combine the phase detector outputs together.  

Run a whole string of 4046's one after each other and let their loop
responses add.

My point that with all this cool stuff available, it seems silly to
just use the chip as a square-wave only VCO module.

  -- Don