[sdiy] Quick query about 2164 VCAs

Mattias Rickardsson mr at analogue.org
Mon Mar 18 23:07:13 CET 2019

On Sat, 16 Mar 2019 at 22:56, Mattias Rickardsson <mr at analogue.org> wrote:

> On Sat, 16 Mar 2019 at 22:07, Tom Wiltshire <tom at electricdruid.net> wrote:
>> Thanks David. I couldn’t see why not either. I think I’ll go ahead and
>> try it.
> Another quick answer: I don't see why.
> How could the current know that it should be divided into two equal parts?

OK, turning on the dusty old brain for a little while:

The virtual ground input of op-amp applications (and perhaps the 2164 VCA)
is a phenomenon created thanks to the very high open-loop gain and the
feedback, which regulates the voltage of the input terminal to very close
to ground. But not exactly ground - I guess it's still a voltage changing
linearly with the signal voltage or current used in the circuit.

So if we'd have two identical op-amps connected to identical circuits, we'd
have two identical near-zero voltages at the virtual ground input terminals
- and we could connect them without any current flowing between them. We
could also put them in the same circuit in stead of two identical circuits
- they would still act the same, doing the same job as before.

In the case of the 2164, two identical VCAs would behave identically,
having identical tiny voltages at their virtual ground inputs, and they
would do half the job each in the parallelled VCA circuits we're talking
about. And no, the current doesn't need to know that it should be split in
two identical halves - it just happens. It's almost even trivial.

BUT it assumes identical units! Offset voltages, offset currents and other
imperfections of the real world would break the spell, possibly forcing one
VCA to react much much more and gulp up much more of the input current than
the other VCA, leaving us with a circuit more similar to a single VCA
without higher current capability or lower noise.

So, in the real world, even with VCAs on the same chip: How close and
well-matched can they possibly be, in the worst case? Is it reasonable to
believe that all the error differences are orders of magnitude smaller than
the already tiny (but crucial for the function) non-zero voltages at the
virtual ground inputs? What do you op-amp experts say, can you connect
op-amps fully in parallel without worries? :-)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20190318/cedc4ac7/attachment.html>

More information about the Synth-diy mailing list