Optical Phasers - Matched LDRs?

[somnif]

Simple enough question, I've acquired an optical phaser PCB and am curious if there is a noted benefit to matching the LDRs. Easy enough to match them, but I'm just curious before I go and buy 20 when all I need is 6.

Any advice for a opto-neophyte?

[alanp]

I've never bothered.

[alanp]

Well, apart from making sure that the LDRs are of the correct general range (ie, don't use 500K dark LDRs instead of 2M dark LDRs.)

[midwayfair]

Maybe.

You have to look closely at the stages and determine: "If one of these stages is partially out of phase with another stage, what will happen?"

The reason for matched FETs in FET phasers is because they will have fairly extreme behavior. What happens is that when you apply voltage to one pin of a transistor, it will vary the resistance between the other pins (the "-istor" in transistor is short for resistor, if you're at all curious about such things). When you take two transistors, if you apply the same amount of voltage to the same pins on each at, say, the base, the resistance might be different between their respective outer pins will be different, unless they're matched. So far we're not really any different from LDRs. Apply the same amount of light to two LDRs are you are sure to get different resistance.

But FETs are going to vary from a couple hundred ohms to ... M Ohms ... with only a very small change in voltage. It just so happens that we can match them for a couple characteristics and get them to behave just about the same with the same amount of voltage.

More importantly, FETs react almost instantanously. LDRs not so much.

So now that we're back to LDRs, what's the main difference here? After all, we could match LDRs so that their resistances are the same in pitch black and under full light ...

But WHEN? Because LDRs, you might have noticed, have more characteristics, that are pretty darned important: Rise and fall time. Just like how your eyes take a second to clear when someone flashes a light in your face, LDRs have a brief memory that keeps them at a lower resistance immediately after they're exposed to light. The time it takes for them to get down to that low resistance is the fall time (resistance going down), and the time it takes for them to "forget" how much light they were exposed to is the rise time (resistance going up). These are measured in milliseconds. The VTL5C1, for instance, has a super fast turn-on time of 3mS, while may discrete LDRs might be more in the couple dozen to maybe as high as 100mS range. Still kind of fast, you might think, but if you start flashing a light on it at 10Hz you might think differently (10 cycles per second = 0.1 s = 100mS = hovering at the low end of the resistance range the whole time!).

So now if you're going to match them, you've got a pile that are all sorted into their dark and light characteristics ... do YOU have a way to measure dynamic changes in resistance in the millisecond range? I certainly don't. I don't trust my eyes to read an analog meter (which is better than digital for accurately assessing dynamic changes) accurately at those speeds.

The way I generally handle this? Make a list of the LDRs I can buy, and then figure out whether higher dark resistance or lower light resistance is more important for the particular build, and figure out what the fastest setting for the LFO is, so that I know whether I need to eliminate slower LDRs (and maybe take the waveform into consideration to determine if a SLIGHTLY slower response will sound nicer). After I determine that, I can look at the datasheets for the LDRs and buy the ones with the largest dynamic range out of those that meet the dark or light characteristics I'm looking for. Dynamic range and speed trump most other characteristics.

Hah, I'm just kidding, I don't do all that. I just buy the $1 LDRs from Smallbear, usually the 9203, or the Hi-Dark if I need really high dark resistance, unless the designer specifies a particular vactrol or LDR, because those work in almost everything.

[zombie_rock123]

Best post ever. Read the whole thing thinking "oooh rly! Hmm I need to...  "

[cajone5]

You bastard.

[somnif]

Hah, works for me. The phaser in question is the Mutron II (or rather the fantastically compact Krypton version of it). He mentioned the LDRs from Tayda are what he used, and given their proclivity for sometimes loose tolerances, I was just curious if it was something I really needed to worry about. A cursory google search just confused things further, as some random folks said matching was essential, others mentioned nothing at all.

Bit of a shame really, I'd already started mentally designing a millisecond scale resistance graphing function using a buddies Raspberry Pi and a lightproof enclosure. That's three beers worth of thought I'm never getting back! 

Looking at the schematic for the phaser, the LDRs are in parallel with a 200k resistor, acting as a ground shunt for pin 3/5 of the opamps. I suppose that rather takes resistance range out of the equation a bit, but I'll just hope the sweep time matches the data sheet. (~35ms rise, ~5ms fall, so they claim).

[thesmokingman]

that's the problem, it was only three beers' worth. now if you had between 6 and 12 invested in this, we wouldn't be talking about this and it wouldn't have been wasted.