This gets discussed in roughly every thread about shutter testing. If one puts the phototransistor directly behind the lens on-axis, or if one puts the phototransistor on axis behind a naked shutter w/o lens elements, it "sees" light from the first instant the shutter opens a little bit at the center, to the last instant it closes. So it doesn't compensate for the "shutter efficiency" where the shutter is open across the outer parts of the pupil for a much shorter time.
Putting the sensor well behind the lens at the focal plane and illuminating the lens uniformly from the front may show a slower ramp up, but you do still have to worry about whether the phototransistor is linear or a switch as someone else mentioned.
I wonder what the result would be if you compare a normal Compur with an electronically controlled one like in a Fuji GX680 or Bronica.
Expert in non-working solutions.
You might also test at various apertures. I recently read a paper that claimed that leaf shutters were more 'efficient' at smaller apertures. The reasoning being (I think) that the shutter clears/occludes the smaller aperture faster. There is also a ramp up/dwell/ramp down cycle in the shutter action that might make faster speeds somewhat more challenging to measure- the dwell portion may be quite short compared to the open/close times. I think this is one reason a high speed video would be interesting...
In the shutter tester I built, I use a phototransistor (a Vishay BPW85B) and it detects pulsed light through the lens at 10K Hz. There is a dramatic difference in amplitude in the signal representing this pulsing during the time the shutter is open. Here is an example of the signal for 1/100 sec. The opening and closing parts of the shutter action are visible. The entire opening time is 1/111 s. Without the ramping up and down, it is 1/114 s. The stop was f8.
The capacitance for the phototransistor at the voltage I run it at (9 VDC, C to E)) is 2.5 pF -- very small. The bandwdith is about 180 KHz
I read the voltage with a sound card, which only goes to 20K Hz. I set the 555 timer to 10K Hz so that I would have double the bandwidth to better capture the signal.
The signal is more than precise enough to capture the shutter action.
Mike, I could be mistaken but I don't believe that a Supermatic opens and closes as fast -- in 0.000237079 seconds combined -- as your measurements show.
Thanks for this. I've read quite a few threads but don't recall ever seeing this bit of information.
I diffused the light going into the lens and placed the sensor near the focal distance of the lens and got this:
Not perfect, but much closer to expected. Close enough for me to chalk up the difference to experimental error (a lot of eyeballing going on) and satisfy my curiosity.
I agree, the time difference is very small may probably wrong, but I am not sure why. It could easily be 2X the 0.0002 measure, as 10 KHz cycle is 0.0001 s.
My measurements could be biased to the len's central axis, as reddesert mentions in post #12. The lens elements remain on the shutter, but the pulsing LED is about an inch from the front of the lens, and the phototransistor is within a 1/4 inch of the rear element. I do it this way to make sure I get a high signal-to-noise ratio. The light is so out of focus though, that the on-axis effect is probably negligible. The light circle on a surface very near the rear lens element (as close as I can get it and still see it) is about the same size as the rear element, so we are not dealing with a narrow light source at the sensor location.
What I was mainly trying to show is that the phototransistor response is fast enough to capture a 10 KHz rise and fall, and using pulsed light gives a more intuitive sense of the shutter opening and closing, and finally, amplitude linearity is not that important when interpreting a pulsed-light signal.
Measuring the phototransistor's voltage amplitude linearity to light amplitude would be worthwhile (the device's datasheet does not show this), and may not be that hard to do, so I may try it. The other aspect is the angle of sensitivity -- it's +- 25 degrees for 50% sensitivity for the device I am using, so off-center light is detected (but the angle is reduced by the entrance hole in the transistor mount). I could get a sense of the on-axis effect by moving the sensor off-axis and see what happens.
I get good exposures using my corrected shutter speeds, and I mostly use slow speeds and f-stops of 16 to 22, so never was that concerned with shutter efficiency.
Bookmarks