[sdiy] Using SSM2164 in stereo

[David G. Dixon]

Upon reflection, I don't think it is possible to drive a 2164 VCA with a
current source without frying it.  Consider: When you convert a voltage into
current through a resistor, the current is not predetermined, since the
voltage source (in the Thevenin sense) will source or sink as much or as
little current as is required to maintain its output voltage.  This means
that the VCA inputs can sink or source as much or as little current as they
want.

Charge (unlike potential) is conserved.  The current must go somewhere.
Remember: When feeding a fixed current to an opamp (as with a I-V converter
or an integrator) the current flows around the opamp, through the feedback
loop.  There is no feedback loop around the VCA.  The current has to go
through the VCA, but the VCA has its own ideas about how much current to
pass.

Bottom line:  2164 VCAs must be driven with voltage sources which are
subsequently converted into currents with I-V resistors.  Hence, each input
requires its own stabilizing RC network.



> > The VCA will be fed by a current output, so I don't have the 30K
> > input resistor that the datasheet also shows.
> 
> I would be more concerned about this aspect of the design - this is only
> guaranteed to work iff both inputs have *EXACTLY* the same input
> characteristics so that the input current is perfectly split 50:50 between
> the two inputs.  Trouble is, you can't rely on this - it may change with
> temperature, with control voltage, with age, etc.  The result would be
> audio panning all over the place out of your control.
> 
> Adding the V-to-I resistors fixes this problem.  Then you add the little
> RC compensation network for stability.
> 
> Note: SPICE won't show this, unless you have a _very_ good model of the
> input, and run Monte-Carlo over several parameters.