how to slow a fast photocell?

[JisforJustin]

Can you simply connect a capacitor in parallel with a photocell to slow it's rise and fall times?

I am just using a photocell to replace a knob so you can waive your hand or foot over it, but it changes to quickly to be really useful. The photocell I am using, in case anyone is looking for a fast responding photocell with a range anywhere from about 5k or 10k to about 200k, is the first listing on this page:

http://www.elexp.com/opt_5516.htm

[JKowalski]

No. The photocell changes resistance, and a capacitor would not affect the resistance. The only way that would work would be if the photocell was part of a DC voltage divider, then the speed that the voltage divider changes could be delayed by a capacitor in parallel. If you are putting a AC signal through the resistance, a capacitor would not slow the effects of the resistance change.

Maybe you could put the photocell in a voltage divider with the capacitor to slow down the response, and use the CV from that voltage divider and a LED to control a second photocell.

[R.G.]

There's not much you can do to slow down the light response of a photocell. The response to a change in light is pretty much fixed when they're manufactured, a property of the mix of materials and processing.  About all you can do is slow down the change in the light source, which you can't do, unfortunately.

A capacitor across the cell itself only passes treble through, it doesn't slow down the light response.

[JisforJustin]

Well, I just did a quick and dirty test on a breadboard. I hooked up the photocell in series with an LED. Waiving my hand over it slightly affected the brightness of the LED. It was a very fast change, like flipping a switch. Then I put a 100uf cap in parallel with the photocell and did the same thing. This time the change WAS delayed. Now, I still haven't tested it in the actual circuit, but that makes it seem to me like it could work. Are there any major flaws in my thinking? EDIT: Yes. 

If you are putting a AC signal through the resistance, a capacitor would not slow the effects of the resistance change.

Hmm, I am not 100% sure if there is an AC signal running throught it. The knob I am replacing is sort of a feedback loop within a distortion circuit. I guess you might be right. I'll have to try it and see what happens.

[Mark Hammer]

There's not much you can do to slow down the light response of a photocell. The response to a change in light is pretty much fixed when they're manufactured, a property of the mix of materials and processing.  About all you can do is slow down the change in the light source, which you can't do, unfortunately.

A capacitor across the cell itself only passes treble through, it doesn't slow down the light response.

WHoa camel.  Ah sed whoooooaaaaaa camel.  WHOAA CAMEL!! Now just a cotton-pickin' minute there.

Of course you can slow down the light source.  Or rather, you can slow down the voltage source that is driving the light source.  You may not be able to change the speed of light, or the speed of the photoresistive reaction in the LDR, but you can sure as shootin turn the light source on more slowly.  Otherwise how could we have autowahs with slower and faster attack times?

[JisforJustin]

Of course you can slow down the light source.

Unfortunately in this case the light source is ambient light and the change is coming from my foot. 

But what about a transistor, could it me rigged to replace the resistance the photocell provides and then use the photocell to control the base of the transistor?

DISCLAIMER: Sorry for my ignorance, I haven't played that much with transistors as voltage controlled resistors.

[R.G.]

WHoa camel.  Ah sed whoooooaaaaaa camel.  WHOAA CAMEL!! ... Now just a cotton-pickin' minute there.

Of course you can slow down the light source.  Or rather, you can slow down the voltage source that is driving the light source.  You may not be able to change the speed of light, or the speed of the photoresistive reaction in the LDR, but you can sure as shootin turn the light source on more slowly.  Otherwise how could we have autowahs with slower and faster attack times?

I'm chuckling because I had my "you can only slow down the driver to the light source" reply typed when I did a ... hey, didn't he say he was using his hand or foot to change the light?  I barely but barely danced to the side of that tank trap. 

But what about a transistor, could it me rigged to replace the resistance the photocell provides and then use the photocell to control the base of the transistor?
DISCLAIMER: Sorry for my ignorance, I haven't played that much with transistors as voltage controlled resistors.

You could use the photocell as a sensor, then do any number of things to change gain, driven by the electrical signal derived from the photocell, and easily enough slow down the electrical signal, in concept anyway.

However, it is going to be a tricky balancing act to use just a photocell and a transistor and nothing else. There's a ton of special cases involved there. A big one is that bipolar transistors are only "voltage controlled resistors" for tiny signals. They can be used that way (EH Pulsar tremolo does), but it's tricky. JFETs act like voltage controlled resistors for bigger signals (up to about 1/2 to 1V) but you also have to use the photocell to make their gates happy. It's easier than a bipolar, but still tricky.

There's a ton of stuff that could be done to correct the range of signal, make an appropriate voltage, match it to the variable element, and so on, but it rapidly loses the neato ability of the photocell to just do what you want with one part. It's probably simpler to track down a slower photocell. Getting a slow photocell isn't nearly as hard as getting a fast one, so the tide is with you on that one.

[Mark Hammer]

Looks like MY photocell went off too fast.

Is the issue that the photocell is too fast, or that it is too responsive?  I ask this because slower LDRs have a way of "averaging" lots of fluctuations, while faster ones show everything.  Perhaps the solution is to use multiple parallel photocells (like a small-scale fly eye) such that the average change comes out more predictable and less wonky.

[JKowalski]

I still think my idea of using the LDR to control a voltage divider, slowed down in response by a capacitor, and use that to control an LED/LDR is a good one.

That way you can easily adjust the speed of the control response to whatever you want. It's not that many parts extra at all.


From what I am getting, you want it to be SLOW. Not just a tiny bit slower, but slower then any photocell out there can actually give you. It would help if you could clarify that, for example, how long do you want it to be until the change is fully implemented, in approximate seconds/milliseconds?

I believe the slowest photocells you can get can only give you around 100ms delay maximum. Most are much less. 100ms is 1/10th second, that's pretty fast. A parameter change at <100ms is going to sound pretty much like an almost instantaneous change

[GibsonGM]

How are you going to keep the ambient light source constant, so you get predictable results every time?  Or are you just making something for 'flash', where that wouldn't matter so much?
I like Chris' idea above, using the LDR to modify a voltage divider; low parts count, adjustable under test, passive and simple.  Just my .02

[R.G.]

I still think my idea of using the LDR to control a voltage divider, slowed down in response by a capacitor, and use that to control an LED/LDR is a good one.

That way you can easily adjust the speed of the control response to whatever you want. It's not that many parts extra at all.

It is a good idea. The only thing it leaves out is that it will be very difficult to drive an LED that way. This is the kind of issue that I was alluding to when I said:

There's a ton of stuff that could be done to correct the range of signal, make an appropriate voltage, match it to the variable element, and so on, but it rapidly loses the neato ability of the photocell to just do what you want with one part.

LEDs are almost-fixed-voltage devices. What you need to do to successfully modulate them is to change their current. Sometimes this is done by driving them with a voltage through a resistor. The current is then the driving voltage minus the LED forward voltage divided by the resistance. The LED voltage is an error term preventing the thing from being a nice start-from-zero, so you have to rig ways to have "zero" be just below the LED forward voltage. That depends on the LED, of course, so it has to be matched to the LED you use. And LEDs need relatively a lot of current. LDRs will conduct enough current to drive an LED, but a voltage divider setup needs more current than the LED even. The simple thing to do is just to connect the control LDR in series with the LED to a voltage source, so that light turns it fully on, the LDR varies the current. The problem is that this doesn't suit slowing down the current change much, and a capacitor doesn't help much because the voltage on the LED doesn't change much - maybe 50mV - over the range of currents.

If I had to do this, I might use an LDR in a voltage divider to generate a control voltage, a capacitor on on the control voltage to slow it down, or better yet a pot in series with the voltage divider to a smaller cap to let me adjust the amount of "slowness" of change, then a buffer of some kind, transistor follower or opamp, to provide the current for the LED. The buffer needs biasing and gain setting. You may have to dink with the DC voltages of the buffer to get dark on the control LDR to be almost zero current on the LED...

It's not really that difficult, but handling each little special case as it comes up rapidly makes for a more complicated circuit.

One reason I'm familiar with these issues is that the circuit to do this is very similar to the LERA circuit at GEO. If you subbed in the LDR+voltage divider for the input pot/switch in LERA, you'd have one way to do this application.

I'm not saying it can't be done more simply, just that one runs into a zillion little gotchas in doing it. One device I saw at a guitar show was  simple - an LDR with a window to ambient light in a guitar cord jack. The LDR was in series with the signal, and you simply covered the window partially with a finger to fade your guitar in and out. It had a noticeable delay/smoothness to it, although not as slow as seems to be needed here. It was patented too, although I'm not sure how they got that past the patent reviewers. 

[JKowalski]

Yeah, that's what I meant. I kind of assumed the buffer, that's basically housekeeping around the basic idea I was trying to get across. I suppose it's always best on these kind of forums to provide full explanations so as to not lead anyone to the wrong assumptions.   

I'm not saying it can't be done more simply, just that one runs into a zillion little gotchas in doing it. One device I saw at a guitar show was  simple - an LDR with a window to ambient light in a guitar cord jack. The LDR was in series with the signal, and you simply covered the window partially with a finger to fade your guitar in and out. It had a noticeable delay/smoothness to it, although not as slow as seems to be needed here. It was patented too, although I'm not sure how they got that past the patent reviewers. 

Wow, that's really patented? That's ridiculous.

[PRR]

Dim the lights. (Or put sunglasses over the foto-R.)

Photoresistors are slower in low light.

Of course the resistance will be higher. Probably MUCH higher to get any interesting change of speed.

If you really need slooooow response, you need to buy different cells (maybe no longer made, certainly far less consistent), or do the indirect sense-delay-affect Rube Goldberg process mentioned above.