LDR with three (3) terminals

[candidate]

Some optocouplers have three leads, as seen in the mxr phase 100, and the ehx polyphase/chorus.

Can these optocouplers be constructed at home?  Or is it necessary to acquire some vactrols?

Thanks

[petemoore]

  Opto-coupler pretty much means light and LDR.
  If the two components can each have one of their leads share a connection [such as Gnd. or whatever] the fourth lead wouldn't be needed and = saving copper if mass produced with 3 instead of 4 leads.
  Not sure tho...I'd test for resistance between 2 of the leads, then do diode tests --l<--  -->l-- across the 'other' 3 leads.

[candidate]

My first post was not clear enough.  The optocouplers in question have five (5) terminals all together - two of them belonging to the LED and the other three belonging to a "center tapped" LDR.  Similar in fashion to a potentiometer which typically has three terminals.

[R.G.]

You can just use two LDRs with one light source. The two "halves" won't match, but maybe that won't matter.

[Taylor]

I originally assumed that the 2 LDRs in one of these VTL5c3/2 or similar optos would be setup so that one rises in resistance as the other falls. This would replace a voltage divider easily. After reading Perkin-Elmer's long document on their vactrols, I know otherwise, but I do wish that someone would make an optocoupler that worked that way - it would make replacing voltage dividers so much easier and without the "will this work right?" I'm sure they wouldn't track well enough that the overall resistance would stay the same, but it would at least be a step in the right direction.

The Mouser catalog has a baffling variety of exotic optoelectronics. I bet somewhere in there, there are things that could be extremely useful in building effects, but it will take somebody who gets it better than I do to come up with applications.

[R.G.]

I originally assumed that the 2 LDRs in one of these VTL5c3/2 or similar optos would be setup so that one rises in resistance as the other falls. This would replace a voltage divider easily. After reading Perkin-Elmer's long document on their vactrols, I know otherwise, but I do wish that someone would make an optocoupler that worked that way - it would make replacing voltage dividers so much easier and without the "will this work right?" I'm sure they wouldn't track well enough that the overall resistance would stay the same, but it would at least be a step in the right direction.

Two LDRs, two LEDs, one differential amplifier, LEDs in the collectors.

[Taylor]

Well, ok, but wouldn't a single part be nice? 

Is there a chemical reason that an LDR can't be made to have higher resistance with more light?

[PRR]

> Is there a chemical reason that an LDR can't be made to have higher resistance with more light?

There's no definitive Theory of photoresistors.

They are made by secret recipes discovered over the years.

The general theory is that the electrons are mellow and sluggish, but light stirs them up and they move better. Lower Resistance. (It is not so easily explained why some complex compounds give a large and useful change of resistance; the keepers of the secret go on hunches and lots of experiments.)

If there were another process, it would be quite different, and unlikely to be "complementary" enough for a useful voltage-divider.

Note that in Silicon photo-diodes, which do have a comprehensive Theory, more light is more current.

If you want the other way also, R.G.'s diff-amp is quite simple and direct.

It won't "track", maintain constant overall impedance, but you say that is not vital. (and yet sometimes it is.)

You adjust the worst-case overall resistance level with the tail current, the control sensitivity with emitter resistors.

It can also be done with a chip, but hardly any simpler.

If you want "good" tracking, find semi-matched LDRs, use one for effect, the other as feedback to correct the drive signal. That's where the "3-leg" LDRs are handy: both halves came from the same pot cooked the same amount, and they try for the same thickness both sides. However the common leg may complicate things.

> I bet somewhere in there, there are things that could be extremely

I been looking for 30+ years.

There is an opto-FET. Its "linear resistor" zone is very small, even for an FET, hardly 0.1V.

The various opto-diodes, opto-transistors, opto-SCRs, opto-Triacs are as useful (useless) as their non-opto kin. The only special feature is that you can control across a ground-break: another rack or at a deadly voltage.

[spectraljulian]

My trick is to have two LEDs and two photocells.  You invert the response of the 2nd LED, and then you wire the two photocells like they're a potentiometer. 

I made an AVR expression pedal controller (still in development/on hiatus while I work on less overly ambitious projects) and use that wiring on the expression out with two 10k LDRs.  Works like a charm on the Development Board prototype I made.   

[R.G.]

My trick is to have two LEDs and two photocells.  You invert the response of the 2nd LED, and then you wire the two photocells like they're a potentiometer.    

That will work!

I jumped directly to the implementation when I said put the LEDs in the collectors of a diffamp. This does the current/inverting directly for you with two transistors. Same thing.

[candidate]

You can just use two LDRs with one light source. The two "halves" won't match, but maybe that won't matter.

The appeal of the prepackaged designs is that each half is "identically matched"?

[R.G.]

Yep. As matched as you're likely to get.

The prepackaged varieties come in two flavors: centertapped, where the two LDRs are actually one LDR with a literal center tap, not separable, and the venerable Hamamatsu used in the Mutron, where there are two separate LDRs.