by Martin Norberg
(original article at http://www.geocities.com/martinnorberg/stester.htm)

I am always talking about how much I like using simple old cameras. I am sort of technically inclined and I find it wonderful to be able to fix most camera trouble with just a small selection of tools. Mechanical cameras seldom really break, they just get out of sync, they jam, slow down or exhibit some other problem, but they almost never become hopeless cases if you take reasonable care of them. Internal cleaning and new lubricants every 10th year is the norm for keeping them alive and well. Can you learn to do it yourself, so much better for your wallet. One very useful tool for checking out the general health of your camera bodies is a shutter tester. If you have some electronic junk around, you'll make one in half an hour, for zero money, like I did.

First, you need some basic things, like a computer with a soundcard and line level input. You'll also need a sound editing program, like Wavelab or Soundforge. And, like I said, some electronic garbage. The most important of which is a photoresistor - found in old diskdrives, VCR's, CD players and, not surprisingly, in old cameras. You need one with a fast response time, the faster the better. These are also available in electronics shops for very little money. You also need a battery of around one volt, an 1,5v AA is quite OK (as line level inputs are OK up to about 2 volts) but I used a small watch button cell of the same voltage (and a bit of the watch too, to hold the battery). The last thing is the connecting cord taken from an old pair of headphones.

The theory behind this device is this: Sound is recorded as alternating current, AC. The battery is DC, but as light hits the photocell it opens the 'tap' and lets current through. When the light is removed, the current gets cut off. Both these events are changes in the DC current, which is basically what AC is - so whenever light starts to shine on the cell, or is removed from it, your soundcard records this as a 'click' sound. When the camera is placed between the light and the photocell, guess what happens? CLICK - the shutter opens. CLICK - it closes. The time between the two clicks is the shutter speed.

So, here are the parts again:
- 1 photoresistor (smaller is better, as it takes less time for the shutter blinds to travel across)
- 1 battery (1-2v)
- 1 connecting cord (stereo or mono, but you only use one channel)
Various other bits and pieces: something to mount everything on, like a plate of plastic or thick cardboard with a hole for the light to shine on the photocell... Epoxy glue. A soldering iron. Black tape. There is no need for a battery switch, as long as you unplug the device when it is not in use (the circuit isn't completed until it is plugged in, so no battery drain). This is the (very ugly) wiring diagram:



And this is how it turned out:



So, how to use it? Easy. Just plug it in, start the recording program, uncheck the box saying 'adjust for DC' and start recording a file at the highest resolution your card can handle (usually 44kHz, 16bits - but more is better as it gives more precise results). Position the camera body, lens removed, under a strong light. Hold the tester in front of the film gate (with the back open, of course). Fire at all speeds. Stop the recording and look at the series of clicks you just recorded - you might have to do some changes in magnification and/or volume before you can see them properly. Set the display to 'seconds' and start marking up the time between the peaks (use highest magnification here). 1/1000 is supposed to be 0,001 second, 1/500 is 0,002 and so on (just solve the calculations to get what the shutter speeds should be in seconds). See?

In this next image you see how different the response characteristics are at the lowest and the highest speed. The jagged line is just noise (50Hz hum) because of a less than perfect cable. Still, it seems my photoresistor is able to keep up - the sloping curves at the 1/1000s could very likely be the curtains traveling over the resistor cell (the travel time is supposed to be 14ms across the whole frame on a Nikon F). I put the markers halfway down and up the slopes, another way would be to measure from where the cell starts to react, to where it just begins to cut off (but I can't tell which way is more accurate).

It turned out the camera I first tested - a beat up Nikon F - was more accurate than I could imagine but the two highest shutter speeds were a bit slow, so 1/1000 was exactly 1/500 and 1/500 was more of 1/300 instead. The long speeds were spot on. Still in the ballpark for a 40 year old camera, I'd say.

Now, to REALLY test the shutter, including separate curtain travel time, I would have to make a speed tester 'in stereo', with two photocells positioned at each side of the film gate. But that will be another day.