[sdiy] Simple VCA?

[Tim Stinchcombe]

With the circuit as drawn by Scott (below) my (nearly) considered opinion is
that it is just an inverting op amp set-up, the JFET might well as not be
there.

I got an analysis worked out (not double-checked though), and when I made
all the '1/gain' terms go to zero for an op amp gain of 100,000 (say) it is
just an inverting set-up. I then ran a quick simulation with 1V input, a
2N3819 and a 741, with the feedback resistor at 2k4, and I got -5x gain, as
I swept the gate voltage from -3V up to 0.

There is a little 'kink' at -2V gate, presumably the threshold voltage where
the JFET starts to conduct, and the gain does ramp, but the change in the
o/p voltage (at 0V gate) is around 400uV (i.e. peanuts). Then when I read
Scott's comments about small input voltages, I stepped the input from 1uV up
to 1V, and with 100uV in, it _almost_ looks like it is doing something
interesting, but when I realised I was dealing with input currents in the
region of 200nA, and that changing the sign of the input voltage just
*shifted* the whole output voltage curve up (i.e. rather than flipping it!),
I decided I was on a hiding to nothing (the input bias currents of the OA
probably having some effect...?), and wasn't going to spend any more time on
it!

I predict if anyone actually breadboards it, you'll get pretty much zippo
'VCA' effect out of it (but I'll keep my eye on the thread if anyone can
show me that I'm wrong!). I suspect the reason it isn't to be found anywhere
else is that other people soon realised this, and didn't copy it. Another
similar version I *did* see was to 'split' the input resistor in two, and
have the JFET coming off the confluence of these two. Then it should really
act more like a variable subtractor (and I'll add 'I think'...as I don't
know that I'm that bothered to check!).

Tim
__________________________________________________________
Tim Stinchcombe 

Cheltenham, Glos, UK
email: tim102 at tstinchcombe.freeserve.co.uk


> So it's like this then:
>                                  FB ?
>  
>                         +------\/\/\/\---+
>              510R       |                |
>                         |  |\            |
>      IN *--/\/\/\/---+--+--|-\           |
>                      |     |  \          |
>                      |     |   \         |   
>                    |-+     |    \        |    OUT
>                    |       |     >-------+----->
>  - CV *----------->|-+     |    /
>                      |     |   /
>                      |     |  /
>                      +-----|+/
>                      |     |/ 
>                      |
>                      /
>                      \ 510R
>                      /
>                      \
>                      /
>                      |
>                      |
>                     GND
> 
> When I imagine this without the FET, a 500K FB resistor would 
> give a gain of 1000. 
> Input impedance seems very low (510R).  My sense is that this 
> VCA probably works better with smaller signal voltages.  I'm 
> wondering if JFET input opamps would work, don't see why not. 
>  A dual could provide a high impedance buffer/inverter for 
> it.  Intriguing simplicity though.