[sdiy] optocouplers/Vactrols

[Paul Higgins]

Well, probably the solution to #1 is to always use the same LED.  
Although the LED is pretty unimportant if you're just using the DIY 
"Vactrol" as a switch.  (At least in high-Z tube circuits anyway, which 
is what I tend to use them for, the variation in "on" resistance is 
usually negligible).  But for more critical (variable resistor) 
circuits, always use the same LED.

#2 is similar: always use the same cell.  There is of course a pretty 
large tolerance spread in something like an LDR, so you'll still have 
to do some matching.  But you'll probably have to do some matching even 
with commercial optocouplers.  Silonex claims that the best you can 
hope for is about 5% tolerance, and I'm sure that it's worse than that 
over a wide range.

#3 is fairly well taken care of by my little super-glue/catalyst 
technique.  The real bonus is that the super-glue dries water clear 
rather than cloudy when you use the accelerator.  (This same product is 
used for finish repairs all the time; it can even be wet-sanded and 
rubbed out like a clearcoat).  I just use a cotton swab to put a little 
accelerator on one side, put a drop of super-glue on the other part, 
and slap 'em together.  BTW, the super-glue system is made by the 
Satellite City company.  It is a little bit expensive, though 
($12/bottle or so).

The good thing about "Vactrols" is people tend to want to exploit their 
less-than-ideal control law and tolerances.  I think that trimming the 
circuit would go a long way toward fixing the really bad problems.  
(One thing might be to have a biasing trim for the LED, maybe 
stabilized with a zener or something).

On Tuesday, October 19, 2004, at 01:20 PM, Tim Parkhurst wrote:

> So, what are the big variables in DIY optocouplers? As far as I can 
> tell,
> there are three main areas:
>
> 1) LED current vs. output. Manufacturers variation from one part to the
> next.
>
> 2) Photocell illumination vs. resistance. Again, variation from one 
> part to
> the next.
>
> 3) "Manufacturing tolerances." I.e., variations in position, distance,
> opacity, when the whole thing is being put together.
>
> Okay, so #3 is at least under our control. Getting things positioned 
> and put
> together consistently just takes some care and practice. But how bad 
> are the
> variations in items 1 and 2? Could a simple trimming circuit (say a 
> pot in
> series with the LED) be used to get a fairly close min/max match 
> between
> several DIYOCs? (Note the new acronym. Remember where you saw it first 
> ;) )