Measuring LDR Ranges

[Timko]

I have been working on a few LDR projects while at home.  One of them is a double Meatball style envelope filter, and I've been having issues getting the envelope sounds to actually trigger.  I began to think that the issue may be due to the LDRs in my build so I tried testing the light and dark resistance of the extra ones I had.  The light resistance measured up, but the dark resistance was massive - it started around 10M and kept increasing until my multimeter was out of range to read it.

I constructed a measurement jig on my breadboard with an LED and a CLR connected to a switch.  I then put the LDR I Want to test next to the light, and cover it with a toilet paper cardboard roll with tape over the top.  I used a multimeter to measure the resistance of the LDR when the LED was on, and when it was off.   I ran my tests at night as my office has windows on 3 sides. 

Since an LDR is in a box, the lux value will be really close to 0 if the LED that's shining on it is off.  I'd like to match 4 of these for this build (and I think another 12 or 16 for the Mutron BiPhase that's almost done), but want to know if I should only concern myself with the resistance under light conditions or if I should also concern myself with the dark.  And if I should concern myself with the dark resistance, how should I go about measuring that?

For reference, I've run these tests on some GL5516s (5-10k /.5M), GL5528s (10k-20k/1M), and Tayda KE-10720s (5-10k /.5M).  All had light resistance in the general ballpark of what I expected but dark resistance way past the expected value.

If anyone could provide some insight I'd appreciate it!

[PRR]

The dark resistance always goes to infinity, in shunt with package leakage. It can take hours or days or forever.

It should never matter to the device.

The Spec-Sheet number is to reject leaky devices. It is usually time-limited: they don't want to hold the production line all day waiting for the meter to max-out.

[idy]

Important to notice in the meatball and many/most other uses the ldr is in parallel with a resistor. When it is dark, in the megaohms range, the resistor is more or less what the circuit "sees", in the meatball 220k or 120k.
When the LDR goes down, it predominates. The curve will combine the curve of the ldr/led combo with the resistor, which is kind of like that "secret life of pots" trick where you turn a linear pot into a *sort kinda* reverse log.

so the circuit works a funny taper dual shaft 120k/220k pot....

[Mark Hammer]

What idy said.

[Timko]

Thanks!  I'm building Diablochris' Duelin Herbies, which is 2 envelopes going in opposite directions (one of the LEDs is off, one of the LEDs is on upon startup).  It appears for each audio path section, there's an omp amp, then an LDR in parallel with a 220k resistor, then another op amp, then a 2nd LDR in parallel with a 120k resistor, then a final op amp.  Each LDR in the section is acted upon by the same LED.  From some quick math, I see that if the LDR is dark (I used 10M), than the resistor in parallel takes over.  If the LDR is dark (I used 1k), then the LDR takes over.

I'll do some more prodding around, but it sounds like the the dark resistance is fine.  I'll try to test the other LDRs I have and ensure I'm finding a match in the 1k range as specified by the build doc.

BTW, this is the schematic I've been working with.
https://drive.google.com/open?id=0B7QeV_7r70dTdGFzbHk2alRfWDg

[idy]

Maybe look into that speed issue, return to dark resistance time. Just to be sure one or more isn't crazy slow. But maybe it's not return to near infinity, maybe past like 1M there is no effect on the filter...
Does this have a manual setting? Seems a shame to build such a fancy filter without a way to just turn a knob and sweep manually

[ElectricDruid]

+1 totally what IDY said.

The maximum dark resistance doesn't matter, *or* how long it takes to get there. In any practical circuit, you *need* a resistor in parallel with the LDR that sets the maximum resistance, and the only question is how long it takes the LDR to get so large that the resistor becomes the effective resistance. As long as you keep that resistor in the "few hundred kilohm" range, that's usually a pretty reasonable and reasonably repeatable figure. If you wait for the LDR to hit 50Mohm...well, you'll get ten different answers from only two or three different LDRs....

Edit - As an example, the FilterFX pedal I did uses 120K in parallel with the LDRs for exactly this reason:

https://electricdruid.net/filterfx-lp-bp-hp-lfo-filter/

[Timko]

That's some fantastic information!  Thanks all for helping me understand the practical applications of LDRs.

IDY, what exactly do you mean by manual setting?  The effect has all sorts of controls for the envelope detector (band for the amount of signal into the detector, as well as attack, decay, and intensity on the LED) as well as some toggle switches and a resonance pot that control the frequency of the signal.

[idy]

Manual settting meaning turn the envelope off and just sweep the filter with a pot. Fixed filter mode. %^&*ed wah sounds, phasers nice that way on distorted guitars...

[idy]

Also kind of useful for adjusting the sweep of the ldr: if you know what the filter is capable of you have some goals for how you want to envelope/ldr to drive it. Plenty of knobs means it may already do anything you can imagine if its working right though...

[Timko]

It has controls that adjust the strength/attack/decay of the LED change in the envelope detector range, but no way to turn off the envelope. 

I took out my LDRs from the circuit and replaced them with some sockets.  I should be able to use those to ensure that the resistor values in the two stages of the signal (the 220k and 120k) are correctly connected, then insert the LDRs with light on them to make sure they're performing as expected.  This weekend, when measuring some of those values, I saw some odd things.  Hopefully the sockets will let me do a better job of troubleshooting!