How can I check the actual shutter speed of a Nikon FM2, and can I calibrate it myself?

Calibrating your Shutter Speed Unless you are performing photographic telemetry or a variety of other niche scientific measurements, the only reason to know the exact length of your exposure is to achieve accurate exposure. Although it is rare for a modern camera's shutter to drift by a stop or more and still continue to function, old war horses with spring based shutters can vary dramatically in accuracy while still functioning like new.

A qualified technician can maintain a spring based shutter and a qualified repair facility can replace an electronic shutter. If you are not sufficiently qualified or funded to fix the shutter you can instead characterize it. Once you know how inaccurate the shutter is, the inaccuracy can be compensated by dialing the appropriate exposure compensation into the camera or meter. If your camera lacks exposure compensation it is probably old enough to have a manual ISO setting so just adjust the ISO the appropriate number of stops instead.

Long shutter speeds (1/30th or slower)

Find an object that blinks. I like to use my flash on strobe mode since it is highly adjustable. Any other blinking device which blinks sufficiently fast will work. One option is a fluorescent light which flickers at 120hz (120 times per second) in areas that have 60hz AC.

1. Set up your camera on a tripod or rest so that it can image you or an assistant standing several feet away. Focus the lens reasonably well.
2. Set your camera to time delay mode if conducting the test by yourself and set the desired shutter speed.
3. Start the countdown and stand in front of your camera
4. Trigger the blinking before the exposure and then swing the object back and forth. Make sure that the object stays in the frame, moves continuously, and does not perfectly retrace its steps.
5. Continue swinging the object until after the exposure ends.
6. Count the "blinks" of the object to determine shutter speed.

For example, if you set your shutter speed to 1/4th and wave a fluorescent light in front of it (assuming you are brave enough) then count 28 flashes of the light. You know that the exposure is actuall 28/120 = 7/30th or about .93/4 of a second. Your shutter would therefore by 6.7% fast.

Application notes:

• This technique is only accurate to 1/n where n is the number of blinks. so if there are only 4 blinks then your measurement is only accurate to 25%. In the example above this would mean that you would actually know that the shutter is between and 5.0 and 8.4% fast. You can make this a bit more accurate by making 10-20 exposures and averaging the results.
• If your object is large enough that the "blinks" overlap each other you may want to mask off most of the object and leave a small hole to create a small point of light
• The image will be most readable if you use a small aperture to darken the background so you just see the bright flashes.

Short shutter speeds (1/30th or faster

If you own an old school CRT television this is quite easy. If your TV was made for the USA it draws 525 lines 29.97 times per second. Therefore 15734.25 lines are drawn every second or if you like integers, 62937 lines are drawn every 4 seconds. What does this matter? Each line is 4/62937th of a second. This means that that if you made an image of the tv screen at shutter speeds greater than 1/60th of a second, you could simply count the lines present to know how long the exposure was. Let's see an example:

You set your shutter to 1/125th (.008s) and make an image of the tv. You count the lines in the lines in the image and get 130 lines. 62937/(120x4)=121. In decimal values .000063556 is one line so 130*.00006356=.00826s In either case you can see that your shutter is miscalibrated by 3.2% or 1/10th of a stop. As you can see in the example, the math can be done by multiplying and dividing integers or multiplying decimals. I like to store the decimal value for the time of one line in my calculator and then call it up to multiple by the number of lines.

Application notes:

• To measure between 1/30th and 1/60th you will need to count the "interlines." Even though a frame consists of 525 lines, the CRT only draws half of them on the first pass, leaving a dark area inbetween lines. It will then return to "scan" the interlines. If you shutter speed is between 1/30th and 1/60th (inclusive) you may see the interlines. At 1/60th you'll only see them if the exposure if longer than intended. At 1/30th things could get complicated for a long exposure since you'll see the start of a second frame. If you see the interlines it is easier to count the number of "missing lines" drawn from the last line drawn to the first.
• It's easier to count the lines if they are located in the center of the screen. If you allow the lines to go off the bottom and down from the top you will need to count the blanking lines, which you can't see. This is a particular concern when measuring 1/30th to 1/60th second.
• A much easier way to count the lines is to calibrate photoshop or imageJ. Either software allows you to calibrate units of measure per pixel. Follow that link for details.
• This technique also has accuracy equal to 1/number of lines. It is often less of an issue since you can count "partial lines" At 1/8000 second you will only get 2 lines but you could count, for example, 2.25 lines.

Other ideas:

• A Record is 33.5, 45. or 78 RPM (draw a line on it and see how far the line blurs)
• Kitchen aid mixers are generally quite accurate
• Stroboscopes are intended for this task
• A PWM LED can be setup and frequency controlled with an arduino/Rpi

Credit for some above techniques for measuring shutter time goes to the Photographic Instrumentation course at RIT's College of Imaging Arts and Sciences as taught by Andrew Davidhazy. Since this was a hard-copy course I cannot provide web citations.

~~Disclaimer: I am not responsible for injuries due to waving light bulbs around nor seeing spots for days from staring into a blinking flash. Make sure your assistant is not epileptic.