[sdiy] Linearized 2164 VCA Warning

[Neil Johnson]

Hi Paul,

> A picture being worth a thousand words, has anyone got drawings of
> what's going on? I don't quite follow the words.

This is the original design:

http://www.sdiy.org/philgallo/EDN-072402.pdf

Mike Irwin's linearisation method for the SSM2164.  The issue is to  
do with the input offset voltage of the linearising op-amp (IC1A in  
the schematic pointed to above).  A little note of caution: while  
Mike mentions the LF323 and TL072, the prototype was built with an  
OP290, which has a room-temp typical input offset of 125uV,  
significantly better than you'd get from a TL072.

> I'm using the SSM2164 in something I'm making and I've followed the
> datasheet, so if there's additional "wisdom" to be gathered from
> previous experience, I'm keen to understand it.

One useful nugget from George is to take care with driving the  
control inputs.  They have a 5k internal resistance, so at 15V it  
will cause an internal dissipation of 90mW or so (2 x 45mW).  That is  
enough to heat up the silicon and cause the gain to drift off.

George suggests putting a series resistor to limit the current flow.   
I would be tempted to add a 5V1 zener diode as well as there really  
is no need to go beyond a CV of 5V -- at which point the attenuation  
is approximately 151dB, well beyond what any synth audio path could  
achieve.

I'd consider the OP177 (available from Rapid, nice) which is about  
1/4 the price of the OP290 and has a pretty good input offset voltage  
(25uV max).  Its bipolar, not FET, so you need to be a little more  
rigorous in circuit design, but you'll get far better DC performance.

Cheers,
Neil
--
http://www.njohnson.co.uk