[sdiy] SSM2164 VCAs

[George Hearn]

If it is of any use I've had a good play with the SSM2164 and can share my
experiences! 

The transfer function of an SSM2164 deduced by experiment is approximately:

Iout=Iin*exp(-3.489*VC).  

This is approximately 33mV/db so limiting the control range to 0-3.3V would
yield 0 to -100db attenuation.

By using the linearized VCA combination circuit (two Lin VCAs from one
SSM2164) you get some very nice side effects.

1.	Temperature independence.  The VCA gain has a tiny (in theory zero)
temperature co-efficient.  This gives it very good DC performance making it
good for modulation amount VCAs etc.
2.	The linearized circuit has a non inverting summing control port.
You have a VCA with the transfer function.  Iout=Iin*(Ictrl/Iref).  Where
Iref can be fixed to give Iout=k*Iin*Ictrl.  The current inputs and outputs
make circuit integration nice and easy.  Both in/out and control port are
all ground referenced nodes with the inputs allowing summing.
3. 'Off' really is 'off'.  The VCA has a very high maximum attenuation.

There is also a not so nice side effect that the gain response becomes
non-linear below a certain level.  In most applications this is not
significant as it is at very low gain.

One other thing.  In the EDN schematic there is no resistor from IC1 opamp
output to the control ports of the ssm2164 (pin 3,6).  This means that when
the VCA is 'off' 15V is produced across the two 5k input resistances in
parallel ( eq. 2.5k) dissipating almost 100mW of power in the device.  This
produces noticeable heating and can be fixed by simply adding a 9k1 resistor
in series with the output of IC1.  For 15V supplies a larger resistor here
would reduce the maximum attenuation.  George

http://www.edn.com/contents/images/233728f1.pdf

-----Original Message-----
From: synth-diy-bounces at dropmix.xs4all.nl
[mailto:synth-diy-bounces at dropmix.xs4all.nl] On Behalf Of David G. Dixon
Sent: 09 June 2009 07:05
To: 'Scott Juskiw'; 'sdiy DIY'
Subject: RE: [sdiy] SSM2164 VCAs

Scott,

That's awesome!  I will have to study your circuit more closely.  I'm pretty
much sold at this point on the simple switched log/linear Bareille version
of the Gallo/Irwin circuit, and I'm thinking of just a dual VCA on a 1U
panel, and then just cranking a few of these out as needed.  Alternatively,
I might do a 2U panel and put some mixing capability in there (again, a la
Bareille, albeit a bit less complicated).  This would achieve what your
TLN-132 can do; namely, the blending of both log and linear VCA responses
for the same input signal.  I've got lots to think about, and it's making me
very sleepy...

Cheers,
Dave


> -----Original Message-----
> From: synth-diy-bounces at dropmix.xs4all.nl [mailto:synth-diy-
> bounces at dropmix.xs4all.nl] On Behalf Of Scott Juskiw
> Sent: Monday, June 08, 2009 10:07 PM
> To: sdiy DIY
> Subject: Re: [sdiy] SSM2164 VCAs
> 
> With regards to the lin/log argument for VCAs, I don't think it's a
> good idea to assume that straight line envelopes should always feed
> log VCAs. I believe you should do what "sounds" best. Particularly
> with log VCAs, there are instances where 10db/volt works well, and
> other instances where 20 db/volt works well, or something in between.
> I built some VCAs that support both linear and log inputs at the same
> time which I find quite handy when I want to get really picky about
> envelope shape. Scroll down to the bottom of this page and listen to
> the two panning examples. One is 100% linear input, the other is 100%
> log input (with 16db/volt response IIRC).
> 
> http://www.tellun.com/motm/diy/tln132/TLN-132.html
> 
> In both cases a triangle wave (very linear 0-5 volts) is used to
> modulate the VCA amplitude. The results are quite different for the
> linear and log cases. If I was using this in a recording I'd probably
> use a mixture of 80% linear with 20% log.
> 
> On 8-Jun-09, at 10:25 PM, David G. Dixon wrote:
> 
> > Thanks Dave, Aaron and Tim for their excellent advice and links.  It
> > would
> > appear that the Irwin-Gallo-Lamm-Bareille axis is all I need to go
> > forth and
> > amplify!  It's interesting that the Irwin method of linearizing the
> > 2164
> > appears to be very similar to the Sims method for linearizing the
> > 13700....
> >
> > To answer Tim's concern, I was going to put in switches to have the
> > option
> > of linear or log response, for the very reasons you alluded to (analog
> > envelopes should feed linear VCAs, and triangles (such as my lovely
> > triple
> > LFO) should feed log VCAs).
> >
> > Now I've got to go and study (and possibly simulate) both Sims' and
> > Irwin's
> > schemes to convince myself that I really understand what the heck is
> > going
> > on there, and to determine whether it's strictly necessary for VCA
> > applications and therefore justifies doubling the IC requirements.
> >
> > Cheers,
> > Dave
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy

_______________________________________________
Synth-diy mailing list
Synth-diy at dropmix.xs4all.nl
http://dropmix.xs4all.nl/mailman/listinfo/synth-diy