[sdiy] optocouplers/Vactrols

[Paul Higgins]

You meant LED/FET opto, right?  I think that's what the H11Fx parts are.

The LDR (photoresistor) optocouplers have much better linearity specs 
than the FET ones, but are also much slower.  I know that there are 
some techniques for using the FET optos with better linearity, usually 
involving putting the FET at a summing node or other virtual ground 
point.  But the LDR optos (like the Vactrols, etc.) can handle much 
larger signal swings.  Mesa/Boogie uses them all over the place in 
their tube guitar amps to do very high-level signal switching.  (I 
don't think they're actually intended to do that, but they do work).

On Tuesday, October 19, 2004, at 01:41 PM, Scott Gravenhorst wrote:

> I would add:
>
> #4 linearity
> At least for H11Fx parts, (this is a LDR/FET optocoupler) there is a
> linearity issue when the input causes a larger than 100mv signal 
> across the
> FET terminals.  This linearity problem caused me to abandon a filter 
> design
> using it.  But I may revisit this with the low signal caveat, since 
> that
> seems to cure it in large part.  In fact, the OTA methods have a 
> similar
> issue, I note that an OTA SVF often uses a 100:1 (or similar) voltage 
> divider
> before the signal hits the OTA input terminals.