Source for good quality photodiods?

olleorama · May 13, 2009

Hi, I'm planning to build a shutterspeed tester to test some of my older cameras. But to save me the trip to the electronic parts store I thought I could salvage some photodiods or phototransistors. Anybody know any electronic stuff that contain hi-quality photodiods? I read about cordless computer mice, but that those weren't so good. Anybody got more suggestions? ANy household appliances? I've got a full container full of electronic junk (computers, household stuff, some tools) in my apartment complex so I think there could very well be some lying around there.

John Koehrer · May 13, 2009

Probably more effective & less stressful to just buy from Digikey or +. Certainly take a lot less time than searching through stuff.

bdial · May 13, 2009

John makes an excellent point.
But there are bulb socket adaptors and similar gadgets with light sensors to turn on outdoor lights, and some of the little solar powered LED garden lights have photo cells to turn them off and on. Many of these are 10 $'s or so at home centers.
However a photo cell from Digikey is probably less than the shipping cost with no experimentation needed to re-purpose it.

glbeas · May 13, 2009

I've noticed that Frys has discrete components like that in one aisle. Have seen such since Radio Shack cut back on thier stock.

olleorama · May 14, 2009

Yeah, I know that it would be much cheaper. But still, I have some kind of idea not to buy new stuff all the time. I will probably do this time anyway, since I want it to be a high quality item.

sv@diycamerakit.com · May 14, 2009

Because I was planning to offer shutter tester DIY kits, I built a few prototypes myself, so I can speak from experience. One, you should use an infrared sensor on one side and an infrared diode on the other. Using random light sources for the light to illuminate the sensor yields inconsistent results. The good news is that many devices (like scanners, copiers) employ optical moment limit sensors, which are basically a photo diode and a LED, in most cases, infrared. Many of them can be separated in two by just cutting it, and many have mounting holes which then can be used to easily mount them on a stand. The more difficult part is what you use to measure the time, you really should be using a Schmitt trigger circuit, otherwise the measurement will be imprecise, especially for the high speeds.

The other part to consider is that for lens shutters, the real exposure time is not easy to measure, as during the time the shutter takes to open and then to close the amount of light passing through varies and it adds to the light passing through when the shutter is open fully. The mechanical shutters were adjusted to take that into account, so it may seem like then run slightly slow if you just measure the time the center of the shutter stays open.

Fotoguy20d · May 14, 2009

sv@diycamerakit.com said:
The other part to consider is that for lens shutters, the real exposure time is not easy to measure, as during the time the shutter takes to open and then to close the amount of light passing through varies and it adds to the light passing through when the shutter is open fully. The mechanical shutters were adjusted to take that into account, so it may seem like then run slightly slow if you just measure the time the center of the shutter stays open.

If you use a computer with software that captures and displays the change in voltage against a time scale (audio software for instance), it's a relatively easy matter to determine the area under the curve (integrate, if you remember Calc 101) and compare it to the expected result. Although doing this reasonably quickly will yield a somewhat imprecise result, it'll certainly be within a 1/2 stop or less without too much work.

Dan

BetterSense · May 14, 2009

But there are bulb socket adaptors and similar gadgets with light sensors to turn on outdoor lights, and some of the little solar powered LED garden lights have photo cells to turn them off and on. Many of these are 10 $'s or so at home centers.

Those aren't the same thing, and they won't work. Most daylight-detecting circuits use CdS LDRs, which are far to slow.

You can buy phototransistors at Frys and Radioshack. They will work almost anywhere a photodiode will.

I have several hundred photodiodes in a big bag in my mad scientists' lab. If you PM me I'll send you some through the mail.

I have 2 homemade shutter testers, an analog one that uses a phototransistor and a digital one that uses a photodiode and microcontroller. I use external sources of light for both. Any concerns about the unevenness of the shutter tripping can be taken care of by putting the measuring photodiode at approximately the film plane, which you should be doing anyway. Since the photodiode is approximately a point source, any concerns about the shape of the shadow cast by the shutter are taken care of. You are welcome to worry, I suppose, about the light pulse being not-exactly-digital due to lens effects from leaf shutters, but I'm sure not.

Steve Smith · May 14, 2009

BetterSense said:
I have 2 homemade shutter testers, an analog one that uses a phototransistor and a digital one that uses a photodiode and microcontroller. I use external sources of light for both. Any concerns about the unevenness of the shutter tripping can be taken care of by putting the measuring photodiode at approximately the film plane, which you should be doing anyway. Since the photodiode is approximately a point source, any concerns about the shape of the shadow cast by the shutter are taken care of. You are welcome to worry, I suppose, about the light pulse being not-exactly-digital due to lens effects from leaf shutters, but I'm sure not.

I did the same with mine. I have an IR photodiode mounted in a plastic plate and I place the camera with the back open onto this which places the diode at the film plane (or close enough). I use it on my kitchen worktop directly under a light fitting and I have found it works just as well with or without a lens fitted (35mm SLRs). It works fine with in shutter lenses too.

Mine has a 7 digit (7 segment LED) counter and a 100kHz oscilator which is gated by the photodiode. It gives the shutter time in milliseconds to the nearest one hundreth of a millisecond.

Steve.

olleorama · May 14, 2009

Steve Smith said:
I did the same with mine. I have an IR photodiode mounted in a plastic plate and I place the camera with the back open onto this which places the diode at the film plane (or close enough). I use it on my kitchen worktop directly under a light fitting and I have found it works just as well with or without a lens fitted (35mm SLRs). It works fine with in shutter lenses too.

Mine has a 7 digit (7 segment LED) counter and a 100kHz oscilator which is gated by the photodiode. It gives the shutter time in milliseconds to the nearest one hundreth of a millisecond.

Steve.

Okay, can you explain the last part? Or do you have a wiring diagram? I'm not a electronic engineer, I'm more into phusics, so I thought I'd use a simple circuit which I found on the net and just plug it in to the soundboard and use MATLAB to analyze (or audicity, but I feel matlab could integrate better..), but your versions sound much more accurate. I have had some doubt which gives the better readings, photodiods or phototransistors. But now it seems as IR is the way to go? Why would that be more accurate than for instance an ordinary laser pointer?

How would a circuit using a schmitt trigger look like? And how should I ground it? My original design was just a few diods and resistors in a battery box... As I said, I have only studied electronics on a high school level.

And btw, since I live in europe I don't have a radio shack close to me, the nearest electronic parts shop is an hour away. But that wasn't the issue anyway, it just would seem like a good idea to recycle, even if it's just a tiny tiny component, I feel bad everytime I pass the electronic junk pallet, there is so much stuff that could be used, instead it's gonna sit on some deposit, poisoning the earth. Well, when the metal prices are high enough, those circuit boards are gonna be hard currency, at least that's a comfort. Maybe I should start saving old 486 motherboards in my shed?

domaz · May 14, 2009

See this blog post about making a light meter from a cheap light meter cell you can get it at Sparkfun. It looks like if you follow the articles in that blog you will be able to make a light meter fairly easily.

Steve Smith · May 14, 2009

olleorama said:
Okay, can you explain the last part? Or do you have a wiring diagram?

I will try to explain it. I don't have a diagram as I tend to just build things... but I could do one for you if you want one.

Imagine a counter with a display consisting of seven digits with a decimal point after the fifth digit. e.g. 00000.00 (a bit like a sports stopwatch display).

Every time the input to this counter gets a pulse, it increments by 0.01 e.g. increment it 137 times and the display will read 0001.37

There is also an oscilator which gives 100,000 pulses per second. If it is connected to the counter directly, the count will count up in milliseconds. i.e. after 1 second it will read 1000.00, after 2 seconds it will read 2000.00, etc.

Now if I ad a circuit which only allows the counter to run when there is light on the photodiode then it will only count when the shutter is open and light is falling on the diode.

We then have a simple conversion to do to work out shutter speed:

1 second = 1000.00mS
1/2 second = 0500.00ms
1/4 second = 0250.00ms
1/8 second = 0125.00ms

and so on.

You compare your actual shutter speed with the theoretical speed in milliseconds to see how close you are.

The final part is a reset switch which sets the counter back to 000.00 just before you fire the shutter being tested.

Steve.

olleorama · May 14, 2009

Thanks Steve. I got the oscillator and the counter, but I don't get how you trim it down so the counter shows exactly the mS. How do you get the voltage... Wait, I get it! Sort of I think! Electronics is a bit like programming, but IRL, voltage and cycles being important variables

Most of the oscillators I see when scimming through my kind-of-local electronic parts shop has a lot of legs, most of them have four, some are like ICs. Some have variable frequencies... Which one to choose? Is there some kind of standardisation, like there is with resistors? There is some different kinds of counters available here, mostly 'impulse counters', that would be the correct type, right?

But again, this would in theory have the same accuracy as the version using the soundcard? Since I can sample in 99.6kHz? Or do I loose something on the way? Which is the most accurate? The sensors are essentially the same, right?

On the microcontroller thing in the blog link, i think that is above me.

Steve Smith · May 14, 2009

olleorama said:
Okay, can you explain the last part? Or do you have a wiring diagram?

olleorama said:
Huh? How is that gonna help me?

Well, you did ask for an explanation.

olleorama said:
For me to use this information I need a specific type of microcontroller, which I don't have.

Mine doesn't use a micrcontroller.

Steve.

spark · May 14, 2009

industrial limit switches commonly use a packaged photodiode or phototransistor mounted in a plastic housing. Break the plastic away and you have a device to use. IR LED's are found in old remote controls. Almost any of these will be fast enough for checking camera shutters.

The counter approach is a good one. Old counters can be found at electronics flea markets for about $20-$30. A 20 MHz one is very slow by today's standards but fine for 1/1000 sec (1/0.001 MHz!) shutter speeds. For an oscillator- get a 10 or 20 MHz one and add a few cheap CMOS chips to divide by 100 or 200 to get 100 kHz.

Steve Smith · May 14, 2009

Think of it like a very fast version of a stopwatch as used for sports timing. When the shutter opens the watch starts timing, when it's closed it stops.

The only difference is most stopwatches work in hundredths of seconds whereas a shutter tester works in hundredths of milliseconds.

Steve.

olleorama · May 14, 2009

Steve Smith said:
Well, you did ask for an explanation.
.

Sorry for that, that was aimed at the post above you, if you follow his link you will see..

Your post just happened to come between, I just didn't quote him. Ok? Sorry again.

Steve Smith · May 14, 2009

I used one of these for the oscilator: http://www.fairchildsemi.com/ds/CD/CD4060BC.pdf (4060)

And six of these for the counter: http://www.edutek.ltd.uk/Binaries/Datasheets/4000/HCF4026.pdf (4033)

I can't count. It's a six digit display, not seven!

Steve.

olleorama · May 14, 2009

Okay, I'm *very* impressed by the skill and knowledge people have here. But knowing is one thing, understanding is a second, and actually building is a huge third thing. I have ver little experience in building these kind of things. So it will take a long time for me to put this together.

Steve Smith said:
I will try to explain it. I don't have a diagram as I tend to just build things... but I could do one for you if you want one.

Yes, that would be incredibly nice. Then I just need to find the parts. Suddenly everything took a big giant leap from scrapheap sensor to hitech counterbased sensor. But as long as I good something pretty accurate within a reasonable budget and complexity I'm happy.

BetterSense · May 14, 2009

People are making this way too complicated. If you just want to use your soundcard or a scope or some other input, then all you need is a photodiode/phototransistor and a resistor. And a AA battery. I made one like that and used it a lot until I made my microcontroller-based one. It works fine. Note that the simplified diagram shown will dump a square wave into your soundcard, but for the short speeds typical of shutters, it shouldn't be a problem. Mine uses a capacitor and a pot, so it actually outputs two spikes, one when the shutter opens and one when it shuts. Using audacity you can just highlight the distance between the shutter opening and shutting, and it reads out the time value of the highlighted selection in seconds at the bottom of the screen.

olleorama · May 14, 2009

Okay, that is like the design I first intended to build. But if anybody would provide a more complex and more accurate tester I would be very happy.

And again, which one is the better diode or transistor? I've read that diodes give a clearer rise and fall in voltage? And also, why is IR better than any visible light?

paul_c5x4 · May 14, 2009

olleorama said:
Okay, that is like the design I first intended to build. But if anybody would provide a more complex and more accurate tester I would be very happy.

And again, which one is the better diode or transistor? I've read that diodes give a clearer rise and fall in voltage? And also, why is IR better than any visible light?

I used a photo-transistor (part no. SFH300-3) for my simple tester - This was chosen as a: cheap & in stock, b: fairly fast rise time.

The vast majority of diodes/transistors are most sensitive to IR, and even the visable light versions tend to have peak response at around 8-900nm.

As BetterSense pointed out, people are making complicated - For my part, I plugged the SFH300 in to the parallel port and wrote a 200 line C++ program to record the on time. The second version may well use a fast DAC card and quite a bit more code.

BetterSense · May 14, 2009

Diode vs. transistor really doesn't matter here. Photodiodes are faster, but phototransistors are plenty fast for this application. Phototransistors seem to be a lot more common; you can get them at radioshack in a pair with an IR emitter, and also at Fry's. When I was looking I couldn't find photodiodes anywhere, until I discovered a bag of 5000 in the electronics lab here at the university.

As paul_c5x4 said, it's not that IR is better...it's just that silicon PN junctions are sensitive to IR, because of their band-gap. They have a peak in close IR, even the ones used for camera meters.

Source for good quality photodiods?

Privacy & Transparency

Privacy & Transparency