[sdiy] Varactor-based VCF

[Aaron Lanterman]

One of my students, Brian, had heard about varactors in his other classes. 
All semester we've been talking about replacing resistors with OTAs or 
Vactrols, and Brian thought it would be interesting to try replacing the 
capacitors. I figured there was a reason varactors weren't popular in this 
application, but I thought it would be an interesting experiment anyway 
just to see what would happen.

Alas, we haven't seem to come up with anything that filters in any 
remotely variable way.

Varactors work by backwards-biasing them; the capacitance allegedly 
changes with the bias. To get interesing C sweeps, you need to swing the 
bias between something like 5 to 14 or therabout volts. One issue is to 
get audio range cutoffs, you need a pretty high resistor (Brian's been 
trying 1 Mohm and 2 Mohm)

I suggested two ideas for him to try, and he's been banging his head on 
both. I'm afraid I may have led him astray, and could use y'alls help.

Brian sent me screenshots of  MultiSim schematics he made of the two 
ideas. The first is 
here:

http://users.ece.gatech.edu:80/~lanterma/Small_Varactor_Circuit.bmp

My idea was to run +CV through one diode set, and then -CV through 
another, to get a Q-point of zero volts inbetween them to feed our signal 
to; and then put the input through the resistor in an R-C structure, 
hoping that the C's at the CV points are then at "AC" ground. (To get 
higher capacitance, he's using a couple varacs in parallel; he bought the 
very last 5 varactors they had at Ack Electronics).

(The 1 kohm you see at the output is a load MultiSim wanted)

What's very odd is that if you measure point C to ground, and point A to 
ground, and point B to ground, and then measure C to A, the various 
voltages make sense. But... if you measure C to B and then add the voltage 
from B to A... that's a few volts lower, it seems, than measuring straight 
from C to A. I'm at a loss to explain why... I wondered if there was too 
much current and we were losing voltage in the wires, but measuring 
voltage drops at various connection points along the wires didn't yield 
anything interesting. Confusing...

My second (more complex) idea is here:

http://users.ece.gatech.edu:80/~lanterma/Large_Varactor_Circuit.bmp

>From Moog ladder type of things I got the idea to add the CV to the input 
on one "channel" and then subtract the CV from the input on another 
"channel," and then run each of these through an R-C style structure with 
the varacs as the C. Then, adding the two "channels" together should 
cancel the CVs and give you the filtered signal.

Neither filter seems to do anything to the signal that could be considered 
voltage controlled. I'm not sure if there's a small tweak missing, or if 
there's a more fundamental flaw in my thinking. I'm thinking my thinking 
is fundamentally flawed.

Help!

Thoughts on these designs? Am I seriously misinterpreting how varactors 
work?

Any ideas for other varac-based designs? Brian is frustrated and close to 
ditching the varacs for something else together, but since he's already 
put so much time into it, and he has the varacs, I'd like to do SOMETHING 
VCF-ish useful with them.

Thanks as always!

- Aaron

-----------------------------------------------------------------------------

Dr. Aaron Lanterman, Asst. Prof.
and Demetrius T. Paris Junior Prof.    Voice:  404-385-2548
College of Electrical and Comp. Eng.   Fax:    404-894-8363
Georgia Institute of Technology        E-mail: lanterma at ece.gatech.edu
Mail Code 0250                         Web:    users.ece.gatech.edu/~lanterma
Atlanta, GA 30332                      Office: Centergy 5212