[sdiy] FET VCA in Music Easel timbre generator

[Aaron Lanterman]

Another student, Luke, is banging on the Music Easel timbre generator:

http://rubidium.dyndns.org/~magnus/synths/companies/buchla/Buchla_2080_9_200.jpg

He's got the main nonlinearity working fine, using 4558s instead of  
the weird 4136. We put various waveforms into it (where "In" is  
defined as the big vertical line in the middle of the page, connected  
to resistors connecting to the negative inputs of the 5 op amps, aka  
the output of the single op amp in the middle of the page, which is  
part of the VCA. We change the amplitude, watch it on the scope,  
listen to it, bask in its glory.

We can't seem to figure out the VCA. Well, we have it partially  
figured out.

I see that that R17 and R19 form a divider that chops the 4 v input pk- 
pk signal down to a 4*(3.3/(68+3.3)) = 0.2 V pk-pk input signal, to  
avoid distortions in the FET-based VCA.

If we disconnect the 680K resistor (which we disconnected because we  
can't figure out why it's there R18 - more on that later), and apply a  
voltage we pick to the base of the JFET Q6 directly, we get a nice  
VCA. As we change the input voltage from 0 volts to -2 volts, the gain  
changes as the effective resistance of the JFET changes. Beyond -2  
volts we don't get any change - I'm guessing that is its "pinch off"  
voltage.

At -2 volts or thereabouts, the FET has essentially infinite  
resistance, and the
op amp has a gain of (1 + 33/6.8) = 6ish, which would give a 1.2V pk- 
pk signal. As you move the voltage up from -2 towards zero, the  
effective resistance of Q6 decreases to whatever it's "on" resistance  
is, and the gain goes up.

Here's the things we don't get:

1) What in the world is R18 for? It seems to simultaneously couple  
some of the signal into the control voltage on the FET (which it seems  
you wouldn't want) and couple some of the CV on the FET into the  
signal (which it also seems you wouldn't want).

2) We did the experiments of directly applying a voltage to the gate  
of the JFET since when Luke built the rest of the CV circuitry,  
consising of the 100K pot, R24, Q5, and R23, he couldn't get any  
change as he changed the CV or the pot. There was probably a wiring  
problem somewhere, or an understanding problem... in any case we  
couldn't really debug it effective since we weren't sure how it worked  
and what it was supposed to be doing. So... uhm... what ia all that  
"Timber In" circuitry supposed to be doing? Why the BJT?

3) We assume that the base of Q5 is supposed to be grounded, but it's  
drawn
kind of oddly. Is our assumption right?

Sorry I've been asking a lot more questions lately than answering  
them. At least it sounds like my questions are generating some  
interesting discussions.

The SDIY list is the reason my class was able to exist. :)

- Aaron