home-made vca

[Martin Czech]

I guess, this mail got lost, if not excuse me for sending it twice :

>      Also - I want to find out if anybody has a high-performance discreet 
>      VCA design. I'd like to know if there is a design that uses 
>      easy-to-get parts, and would be something that a modular synthesizer 
>      power-user would welcome into his system. You know, wide dynamic 
>      range, ultra-low noise specs, things like that. Any help there?

I tried the old tripple current mirror design (two mirrors stacked on
another mirror, today they call it gain core cell) with BC549Cs.
But this was for an analog multiplier, so I had a cheap differential
log circuit to compensate for the exponential control.
Without that stage, it would be a exponential controlled
two quadrant multiplier which is what you wanted.
This multiplier works good enough as a "ring modulator"  with
extremely low noise. One can taylor all the bias currents for exactly what
the need is. But
since it is really discrete one has to match all the differential pairs
which is a bad job.

(For this purpose I took a transistor which was in the mean value of the
beta distribution and the DUT together in a diff amplifier. The DUTs with
the smallest diff voltage "match". I also used a little cooling fan
to speed up the thermal settling process during the measurement and a pair of
tweezers to prevent heating up the DUT with my fingers.
And still I had to wait a minute for every DUT ...
Otherwise the measurement would have been totally currupted.
Funny thing : they do millions and millions of those BC549 here at Intermetall,
but it was really difficult to get a box of 1000, since they don't sell
small quantities :( ).

So , you see, the design has of course a thermal drift problem,
so I had to put all the BCs into an isolated housing, to keep them at
about the same temperature in order to stop that damned drifting.
After 15 minutes warm up or so the design works good enough for a ring modulator,
extremely low noise, with good
carrier suppression if properly trimmed (the trim pots should be normal
frontplate pots, since one has to once per session).

But I wouldn't recommend this stuff for an precision vca.

I would rather recommend the SSM2018T (ok it is not discrete)
or the SSM2118T. Very clean, noiseless, not expensive, low feedthrough, 
exceptional wide gain range,  very low dc offset,
and the "T" version comes already laser
trimmed, so one might get rid of those trim pots as well.
I use them in my vca circuit and will also use them in my biquad vcf.

I combined this with a home made log circuit,
so I have both linear and exp control.

If I was able to get them from the shelf here in Germany, it should be no problem in the rest 
of the world.

See :
"http://www.analog.com/adibin/locate?SSM2018T"
In the datasheet I think, you'll find also the above described gain core cell.

If a could get that damned Valid Ged to spit out postscript files, I could
post my schematics, if somebody wants it.


m.c.


----- End Included Message -----