Do Differently Colored LEDs Change Anything About Vactrols?

[Josh?]

Hey everyone,

I've got some heat shrink tubing and generic cheap LDRs (the package they came in says GM5539, but I can't find a brand name, if that helps any), and I want to try making some of my own vactrols. Will using differently colored LEDs change any characteristics of these vactrols?

From my (currently) limited understanding of physics, red light has a lower frequency than blue light, green and yellow are in the middle, and white light includes both higher and lower frequencies. What, for example, will an led of each of these colors do differently to my LDRs, if anything?

[stallik]

As a general rule, I would imagine that the spectral sensitivity of the ldr  would be the important thing. For example, if the ldr was only sensitive to blue and you hit it with red light, you wouldn't get much reaction.

[Rob Strand]

The spectral response of a CDS/LDR is very similar to the eye, peaks at 555nm which is green
https://www.tme.eu/en/Document/01ba1573cea1124ddd9a55cccc53ed63/all.pdf

It's not just color.  You have to consider the relative brightness of the LEDs, how much of the LED beam captured and how much of the LDR is illuminated.

The easiest way is to power up a few LED with the same current and same physical set-up and measure the LDR resistance.    Then next batch of LEDs or LDRs will be different ...

[Josh?]

The spectral response of a CDS/LDR is very similar to the eye, peaks at 555nm which is green
https://www.tme.eu/en/Document/01ba1573cea1124ddd9a55cccc53ed63/all.pdf

Thanks! being able to see that graph of the spectral response in the datasheet helped things click for me.

The easiest way is to power up a few LED with the same current and same physical set-up and measure the LDR resistance.    Then next batch of LEDs or LDRs will be different ...

I figured it might end up being a case-by-case sort of thing like this, but thought I'd ask anyway.

[darron]

different colours and types of LEDs will have different forward voltage drops too. so without adjusting the limiter resistor, that simple fact can change the overall brightness and response - i would say enough to matter. the forward voltage drop makes a difference if you are using it for an oscillator like in a trem or phaser, where the waveform can bottom out for example. i find the best way to balance this is a series resistor for current limiting and also a parallel resistor for sort of voltage control.

somebody recommended to me years ago to use green LEDs, and I have since then. i think as hinted mostly overall brightness from the LED is more important than the colour. but i understand that's not what you asked.

so yeah, i think it matters a lot. and well worth experimenting.

[duck_arse]

seek out a copy of the PerkinElmer "Photoconductive Cells and Analog Optoisolators (Vactrols®)" datasheet, like the one on this link:

http://pdf.datasheetcatalog.com/datasheet/perkinelmer/VT500.pdf

in it, it says orange best matches the spectral response of their parts. one of our members, whose name I won't mention [prrr] says makes no much difference, what with the led so close and blasting all it's brights. or words to that effect.

[PRR]

It's a minor effect. Most of the optoresistors have wide bands. The several colors AND the historical sequence of LEDs have very different brightnesses for the same current. Today I would expect any recent LED to be plenty bright for any historical opto-app; you probably want to increase the LED resistor for true historical tone.

[[WZ]]

I've got some heat shrink tubing and generic cheap LDRs (the package they came in says GM5539, but I can't find a brand name, if that helps any)

This part of the question is also definitely worth noting: the LDR's themselves.
Brand doesn't seem to matter much, they are always(?) named 55## and generally a higher ## means a higher dark (and light) resistance. Also, response time is (a bit) slower on the lightest rated ones. As a general rule of thumb I'd say go for decently heavy rated ones.

So, you have 5539s, those are quite 'heavy' ones, with VERY high dark-resistance (~5M).
That comes "at the cost" of having also a relatively high light-resistance, but where the charts say 50k~100K, this will be be much, much, MUCH less when point-blank in LED brightness (no matter what color LEDs you use..)
If you want full choppy Tremolo, that's probably good.
If you want some mild modulation, maybe not so much..

Annoying thing with LDRs though is that they don't have the number printed anywhere, so of course I bought a bunch of floavours and then breadboarded a stereo tremolo where 1 channel couldn't go fully 'off' at max depth.
So yeah.. The right LED, but the wrong LDR.

The forward voltage drop makes a difference if you are using it for an oscillator like in a trem or phaser, where the waveform can bottom out for example. i find the best way to balance this is a series resistor for current limiting and also a parallel resistor for sort of voltage control.

Was wondering why the Tremulus Lune schematic has it like that; a 10K in series AND a 10K in parallel to the LED.

Maybe worth to line up a bunch of different color LEDs to a 0V to 9V LFO and do a side-by-side comparison, to check out the differences in brightness and 'off'-point (due to voltage drop differences)...

[Mark Hammer]

I like to file down the round ends of my LEDs, and then sand them down with ultra-fine emery paper and diamond-grit sheets so that they are fairly smooth as well as flat at the end.  This allows one to press the LDR flush against the end of the LED.  Hopefully, the quasi-satin finish at the flat end diffuses the light reasonably well.  It also makes for a more stable and compact package.

[duck_arse]

GL55 series, I don't know the manuf name, doesn't matter. see:
https://www.espruino.com/datasheets/GL5537.pdf

and waitrony and hamamatsu both manufacture ldr's, look them up.

if you have some diff part number GL55xx's, and you carefully compare the colour of the epoxy, you might see that the 5539 is more orange than the 5516, which is more brown. for instance, only. and, do the observations whilst the parts are still in the packets w/ the makers part numbers printed on.

my method is to colour code the leads w/ hook-up wire insulation offcuts - you only need a few mm slid up one or both legs, and it stiffens them some, too.

[PRR]

> high light-resistance, but where the charts say 50k~100K, this will be be much less when point-blank in LED brightness

The data linked in #9 is taken at 10 Lux. This is essentially twilight. "Family living room" is 50 Lux, and we know we can be brighter at the face of an LED. Maybe near 500 Lux, "Office lighting", maybe more. Few-K should be in reach even on a "50K" LDR.

They rate near "twilight" because that is the bulk of the market for these things: dusk-dawn light controllers. Who cares what they do in bright light? we aint gonna do anything until it gets dark. But our toys can make more light than that.

[[WZ]]

Few-K should be in reach even on a "50K" LDR.

Okay, fixed it:

> high light-resistance, but where the charts say 50k~100K, this will be be much, much, MUCH less when point-blank in LED brightness