Home-made Vactrols - trials and tribulations

[Mark Hammer]

So today I was busy tidying up some perfed projects that had been populated and wired but but never REALLY finished.  One was an Anderton Bi-Filter Follower (great little pedal that one).  I had a batch of LEDs and had picked up some LDRs, and made some of my own dual optoisolators/vactrols.  One of the nifty things about either square-shaped or flat-type LEDs is that LDRs can snuggle up right against them, making it easy to produce solidly packaged dual Vactrols for cheap.

Maybe....

Earlier last year, I had whipped up a bunch of these babies, attempting to roughly match the LDRs for light/dark resistance when making dual units.  I figured that a dual unit would be great for reducing current consumption since you'd get the light of one LED to do the work of 2 optos.

I installed one of these dual homebrews into the pedal (it normally calls for a pair of CLM6000s), and no go.  Checked everything every single way I could think of and it all seemed to check out fine but had very little sweep.  Perplexed, I unsoldered it the homemade optos from the board and decided to break down and sacrifice a pair of my precious CLM6000's (only a half dozen left in the bin :cry: ).  Fired it up and it worked flawlessly.  Clearly, whatever the envelope voltage was doing to the original LED, it  clearly wasn't doing enough to the LDR.  I got only the barest of barely audible sweeps with the homemade one.  It wasn't because of light leakage because light would have made the filter frequency higher up than it was.

This is not necessarily a reason to give up on homebrew vactrols since they CAN be produced cheaply and if you use one of them square LEDs, you could even make a quad unit with a small footprint for phasers, Univibes, etc.  (Use a superbright and current requirements can be reduced even more,  Something like an 8-stage optical phaser running off a single 9v battery becomes a possibility.)  The issue would seem to be knowing what the parameters of each unit are in advance, though.  For envelope-control purposes, one can always increase the gain of the envelope follower, but how much do you NEED?  That's the million dollar question.

I suppose what one needs to have, maybe is some standardized testing module on a perfboard or PCB, with a socket for the vactrol, a momentary switch to feed a trigger impulse to an envelope follower, a pot in the feedback loop of the gain stage to adjust detector sensitivity, and some pins or terminals to connect your meter up to.  The idea is that you feed a pulse/transient of known amplitude into the LED, and measure the resistive consequences, or else measure that gain needed to produce a given resistive consequence.

I wish I could post such a schematic right now, but I'll need to set this one on the back burner for the time being.

I think it is worth considering though, that simply fitting some heat shrink over an LED and LDR does not make an optoisolator that will do exactly what you want or what the circuit calls for.  It WILL produce light-dependent resistance changes, but the bigger question is whether those changes meet your needs and whether there will be anything you need to alter to the circuit to GET what you need.

[Dan N]

oops

[Dan N]

Korg must have been testing batches of optos. Note the cap choices for color coded optos in the lower right of this schematic:

http://users.rio.com/senorris/junk/pe1000_11.gif

A way to get uniform results from non-uniform batches of optos?

[Rob Strand]

attempting to roughly match the LDRs for light/dark resistance when making dual units.

The key thing you are missing is you want to match the behaviour of your DIY unit a CLM6000.  At minimum that entails making the resistance of the DIY unit close to the CLM6000 at a *given LED current* which is in the normal operating region of the circuit.  You actually want to get a good match  over a range of currents but that is not possible once you fix the LDR.  In fact once you choose the LDR you can only play with the LED.  Light and dark current aren't that useful although you will have no hope all if you try to use a 10k light resistance LDR to replace a commecial unit with 500 ohms light resistance.

The variables you have control of are:

- The LED itself:  brighness at a given current, size, shape, colour.  Size can effect the resistance because the beam spot is often smaller than the LDR area.  Colour has an effect because the LDR has different sensitivitites at different wavelengths.

Square LED tend not to be that bright. So while the physical contact looks good the actual light getting on the LDR might be less.  Using one LED for two LDR's provides even less light for each LDR.

- The physical arrangement: how close, the angle.


Even for a given model of LDR there are tolerances.  There are also tolerances in CLM units.  You want the average DIY to be in the ball park of the CLM.

I suggest a simple test set-up drive say 100uA, 1mA (and maybe 10mA)  into the LED and measure the LDR resistance.  If it's not even close to the CLM you are wasting your time.   If it's too low change the colour that matches your LDR's peak sensitivity, or use a brighter LED.

As far as time constants go that's yet another parameter to match, but if the resistance is way off you have no hope of it working.

[puretube]

I`ve encountered rather large tolerances in commercially available optocouplers of the same type,
and to make things worse, some well-known units show quite some surrounding-light sensitivity, which doesn`t matter too much once a circuit is placed inside a stompbox, but is a hassle when tweaking schematics...

[petemoore]

I don't know if it matters...but I roll my Vactrols with white paper, it's more reflective than Shrink Wrap...unless of course your'e using white.
 I gave up on Rolling Vactrols after trying them in various circuits...and getting the EZ Vibe and Uglyface working.
 DOD440, design my own tremolo attempts, etc....nogo...then I started measureing the light/dark resistance, and the readings didn't look like I'd get much out of them..the circuits I had couldn't use them.

[RDV]

I've had good success with one of those superbright green LEDs in me homemade vactrols. I don't think any other will do.

RDV

[sean k]

And,of course,you can file off the top of the LED to make it flat then polish it back up with wet 600 grit paper the cutting compound.Though this may only be to advantage if the LED has a wide enough beam.

[Johan]

the LED is really important to have a closer look at when making your own optocoplers.
depending on the application you might want different LEDs in different units. for compressors and envelope driven units you want a led that starts to glow with very little voltage/current and gets brighter the more you push it. these are the kind that gets dimmed when your battery is running low. for switching applications you can go with the kind that just turns on and off. those are usualy a little bit cheaper and do the job as a LED perfectly, but are not very good when you want gradual change, like in the easyvibe for exemple.
when I buillt my compressor a few years back, out of curiosity I disassembled a number of different vactrols just to find out why they behaved so differently and ( somewhat )to my suprize the LED-part seemed to affect the performance much more than the LDR part and for low voltage units like those running on 9volt the VTL5c2 seemed the best and vtl5c10 didnt work at all...so if you build the yourself, you can test the LEDs in your parts-bin by connecting them in series with a big value pot ( like 50k-100k) and slowly turn the pot down. if the LED turns on/off within a very narow resistance range, get another one.

Johan

[puretube]

let me add a little mojo but true:
I select the LEDs for the couplers by Vf,
and take very good care of thermal influence while measuring...

[DiyFreaque]

Mark,

Just as important as the resistance vs LED current is the response time of the LDR itself.  

Each LDR is going to have a certain amount of time to go from dark to light resistance and a certain amount of time to go from light to dark resistance.  The light to dark resistance will always be a bit (or quite a lot more than a 'bit') longer.

If your LDR takes too long to get to the target light resistance during the sweep of your device, the sweep will be shallow.  If it gets to the light resistance in enough time, but doesn't get to the target dark resistance in time, the sweep will be also be shallow.

This is why I mentioned the VTL5C3 as being the 'golden bb' in many of Buchla's designs - its light resistance and dark resistances match, either by serendipity or design, the curve of natural acoustic responses such as striking a mallet or plucking a string, depending on how 'hard' the LED current is pulsed.  

So, the trick would be to find an LED/LDR combo that both matches the resistance and quick enough response times - for both light->dark and dark->light - to your Clairex part.

My friend Motohiko Takeda has written a wonderful page detailing how to go about doing such a thing (including his test setup).  It's his famous "LDR's sitting around a campfire" phase shifter page.  The test would quickly determine if your LED/LDR combo will work - just put your Clairex part on first, note the curve, and find a combo that will match that curve.

http://www.aleph.co.jp/~takeda/radio/phaser/indexE.html

Be sure to hit 'next' at the bottom of the page.

Take care,
Scott

[mikeb]

I`ve encountered rather large tolerances in commercially available optocouplers of the same type,
and to make things worse, some well-known units show quite some surrounding-light sensitivity, which doesn`t matter too much once a circuit is placed inside a stompbox, but is a hassle when tweaking schematics...

The best (phaser) work is done in the dark.  

Mike

PS a thick black sock works wonders also....  

[puretube]

I`ve constructed a little desktop-darkroom with quick-snap "door";
in it: a "reference-LED" with associated rotary switch for 12 point matching

[DiyFreaque]

I've used those plastic containers that 35mm film comes in - poke four tiny holes in the lid, work the leads of the LED and LDR through those, position the two components so they're facing each other, and snap the lid onto the cannister.  After that, it's just a matter of plugging the contraption into a breadboard via the four leads that are sticking out.

Cheerios,
Scott

[toneman]

nice link Freakue!
i used the Vactrols(Allied) recommended on the MCMeat sch.
works fine...
got others 2 experiment with including Cds cells.
But, don't need 2.....since it works fine...
BTW, Lovetone uses CDs's and greenLEDs.
staytrolledintheVAC.
tone

[DiyFreaque]

Hey Tone,

Yeah - Motohiko's a pretty groovy guy.  He's got his fingers in a lot of pies (hey, I'm a poet and didn't know it).

Sounds like you've got some nice projects lined up.  

My main roll-yer-own project was figuring out a good modulation configuration for a PT2399 delay.  I wanted the absolute min and max, and the delay time for the PT2399 is entirely dependent upon that resistance value.  I had several cells in parallel with a jumbo LED inside one of those film canisters, but I just wasn't getting where I wanted to go.  That, and the response time of the LDR's bugged me, so I ended up using a FET based design - just an MPF-102 acting as a VCR.  The range worked out great, and the response time was instantaneous.  Down-side was that it was a bit non-linear, but the advantages more than made up for that.  Modulating a PT2399 with a sample and hold is cool, and the slew of the optocoupler never would have allowed the same effect.

Cheers,
Scott

[R.G.]

To do sensitivity:
Pick a standard LED. Green is good; it's close to the peak of LDR sensitivity. Hook up a gated constant current source to the LED.

Make a dark chamber for testing. The one I used was a 35mm film cannister. The LED is mounted centered, bottom, and gooped in to seal light out. Cut a hole in the lid big enough for the LDRs to stick through with their leads sticking out.

You have to get light-tight when testing, so cut some chunks of black electrical tape and stick the LDR leads through them from the sticky side. This way, you can put the LDR into the hole on the film can lid and seal the tape down, sealing out light.

* first order test
turn on LED. Stick in LDRs, noting resistance on ohmmeter and only recording when the ohmmeter stops changing.

* second order test
Arrange your constant current source to stairstep through a number of currents, like perhaps 1 ma, 2ma, 5ma, 10ma, 20ma, letting each settle for a second or two before recording.

To do a  time response test:
Arrange the LDR leads as a voltage divider with a fixed resistor and a9V source. Watch the voltage on the junction of the fixed and LDR with an oscilloscope lead. Leave the LED full off for ten seconds, then turn it on. Watch the voltage response on the scope trace and calculate time response. Reset the scope from single trace and then turn the LED off. Watch the light-dark transition trace and calculate the time constant from it.

[mikeb]

BTW, Lovetone uses CDs's and greenLEDs.

Yup ... here's a peek inside a Dopp ...



Mike

[petemoore]

Homey Tremolo Works pretty Darn Good.
 I have two LDR';s and in light they measure 2k3 or so, in dark over 1m.
 I have them Ina Raco, under a large punchout in the corner, inside a paint cap [inside cup...the outer cap was cut cut away].
 Signal path is tied to one side Gnd. the other side of the parallel LDR's.
 then...I have an "Onkyo' tuner component that has a nice bushing...and 4 screw shaft mount...this made it easy to get a spinner going, for the disc-template that blocks or lets light into the LDR Cup...I just cut some hardcardboard to shape...there's a flywheel bell...what made the tuner knob have a heavy feel, and cutting a hole in the cardboard a touch large, then spanning elec. tape around to make the hole seem smaller...and to grab the bell...
 Anyway I have a motor hooked up...very 'Spock'-MacGyver' looking with all kind of scientific/mechanical looking items connected to/hanging off the outside of the box ... for home use...at least I got a PDTremolo going.

[Arno van der Heijden]

I ended up using a FET based design - just an MPF-102 acting as a VCR. The range worked out great

What range (resistance) did you end up with?