The inaccuracy of fast speeds on mechanical cameras.

Paul Howell said:

Given the years, state of lubricants, tension of springs it is not all that uncommon. My mechanical shutter cameras are all over the map, as much as 3 stops off. When I had a Nikon F and F2 I had them serviced once a year by Nikon and my Pentax Spotmatic maybe every other year to keep within tolerance of 1/2 stop.

Click to expand...

No wonder I get better consistent results with electronic shutters

loccdor said:

Do the leaf shutters keep their high speeds better than focal plane?

Click to expand...

No. Leaf shutters need CLA. What sucks about leaf shutters is the difference is inconsistent and can vary seconds

Alan Edward Klein said:

My Nikon N6006 electronic shutter seems to be right on as well. Amazing for a 35 year old camera.

Click to expand...

Same experience with my electronic shutters. especially the copals

Les Sarile said:

Your CL has a horizontal shutter doesn't it?

A new Nikon FM2 reviewed by Pop Photo 12-1982 shows the Nikon FM2 they tested looks better above 1/125 all the way to 1/4000 then in the low speeds. I wonder how consistent these performance tests are across more than 1 sample?

PP1982-12 by Les DMess, on Flickr

Click to expand...

My tester has 3 diagonal sensors covering the 35 frame. So they should work regardless

SalveSlog said:

I believe that leaf shutters at high speeds are slower for small apertures than fully open.

Because if you need the lens to be fully open the film will only get enough light after the leafs has traveled fully out of the way. And as soon as they start traveling back, the light will be reduced.

With the smallest aperture set the film will get the needed light as soon as the leafs start to open and receive light all the time until they're almost closed again. That takes longer time.

This difference will be irrelevant at slow speeds, but at 1/500 or 1/1000 probably significant (?)

Click to expand...

This is called "shutter efficiency" and you have it roughly correct although perhaps could be stated the other way: testing a leaf shutter with a single-detector shutter tester, or unfocused light onto the tester, will yield the small-aperture speed, and the effective shutter speed is faster (less exposure) for wide-open apertures than for the small aperture.

Frequently when people complain about the highest speed of a leaf shutter being slow, it is clear that they don't understand this effect.

When testing fast speeds for leaf shutters with the lens wide open, I hold a small piece of cardboard in front of the lens.

SalveSlog said:

When testing fast speeds for leaf shutters with the lens wide open, I hold a small piece of cardboard in front of the lens.

Click to expand...

Why? Please explain.

Dan Fromm,
see post #24 and #30.
(That's what my last post was an answer to.)

Dan Fromm said:

Why? Please explain.

Click to expand...

If one does what a lot of people do:
shine a light into the leaf shutter lens,
hold a shutter tester with a small photosensitive detector some random distance behind the lens,
read speed/time off shutter tester.

then one is testing the speed with roughly a pencil beam of light from the source to the detector, and this beam is often smaller than even the smallest taking aperture in common use, so it doesn't model the shutter efficiency at typical taking apertures correctly.

There are ways to get around this, like having the light source far away and putting the tester at roughly the focal plane of the lens, so that light is focused from the entire taking aperture onto the tester's detector. SalveSlog's approach is another way of trying to enforce use of more of the taking aperture.

Radost said:

I finally purchased a 3 sensor shutter meter. My leica/Minolta CL were pretty much the same at fast shutters.
This is what i found on all 3:
1000 is 600
500 is 400
250 is 250
125 is 110

+- 5%
I have maybe 25 more mechanical cameras to check

Are mechanical cameras all slow over 250?

My 3 Minolta CLE were +-5%

Click to expand...

It's more useful to view the error is terms of f/stops using the formula:

log(actual/expected)/log 2

So, for example the 1/500 figure is off by -.32 f/stops - which isn't a big deal

After having measured dozens upon dozens of mechanical shutters, albeit it with a single sensor tester, I have yet to see one hit the top speeds with any real accuracy. HOWEVER, if the other speeds are OK (within 1/3 of a stop or so), I am reasonably sure this is mostly a measurement artifact.

But I think it's overkill to worry about it. Using a single sensor timer, a brand new Hasselblad lens will still show 1/3 to 1/2 stop slow at 1/500th. I just got a Hasselblad lens CLAed and everything measured within 0.2 stop or better but 1/500 was "off" 1/2 stop.

With large format leaf shutters, the effect is even more pronounced because you're trying to move a lot more mass. The older Ilex shutters had enough adjustments and mechanism in them that you could get both high and lower speeds reasonably close. But my newer Copals cannot hit max speed reliably at the fast settings. At least that's been my experience when comparing nearly new Copals against just-CLAed Ilex shutters.

For large format, this tends to be a non issue. It's worth noting that mechancial shutters read somewhat differently if you test them after having not been used for a while, vs. firing them a few times and then testing them.

In my view, this just isn't that important. It's largely measurement artifact at the higher speeds. Moreover, a slow shutter will give you slightly more exposure which is well within the latitude of most films and is generally far more desirable than underexposure. The one exception would be shooting 'chromes in which case I'd just bracket a half stop and call it a day.

At high speeds, the shutter is open longer in the center than at the edges. For slower speeds, this difference is negligible since it is very small compared to the overall time the shutter is open. But at higher speeds the difference between center and edge open time becomes an appreciable percentage of the short time the shutter is open.

There is probably some formal definition for speed that takes into account edge and center open times. It might even be some kind of integral formula that computes the total amount of light the film absorbs. But a sensor measuring open time in the middle of the aperture isn't going to take that into account. In theory, some of the multi-sensor measurement tools do, but I have no idea how they are taking into account center vs. edge speeds.

In any case, I think this is mostly not much of an issue in practice. Other than transparency films, modern emulsions have plenty of latitude and 1/2 stop or even 3/4 stop of actual error isn't really much of an issue.

It's worth noting that, for monochrome, that's why the whole idea of "Personal EI" came into being. You calibrate everything for your shutters, your meter, your exposure placement, your manner of agitation, your developer, and so on to come up with your effective film speed. None of this contemplated 0.1 stop resolution (Beyond The Zone System notwithstanding).

Small Editoral Note: With the advent of digital displays and cheap measurement tools, there is a disturbing trend of folks becoming obsessed with absolute accuracy with everything they use. These old mechanical devices had a fair bit of tolerance when new, and will now have even more with decades of use behind them. Expecting electronic shutter accuracy is unrealistic.

I see this all the time when I use a laser to align my old Omega D-2. I can adjust it to be "right on", move the head to a new position, and now it's off a bit again. This old stuff wasn't designed for absolute perfection (whatever that might mean). Another good example of this is that people tend to trust digital numeric displays as being more correct than a moving needle analog indicator. This trust is sometimes misplaced.

SalveSlog, reddesert, I'm well acquainted with the idea of shutter efficiency, but thanks for reminding me. Reminders never hurt.

reddesert, when I test my shutters' speeds I don't use a light source that projects a "pencil beam of light." In any case, my Calumet shutter speed tester doesn't measure light transmitted while the shutter is open, it measures how long the shutter is open. It starts timing when it sees light coming through the shutter, stops timing when it no longer sees light coming through the shutter.

Dan Fromm said:

SalveSlog, reddesert, I'm well acquainted with the idea of shutter efficiency, but thanks for reminding me. Reminders never hurt.

reddesert, when I test my shutters' speeds I don't use a light source that projects a "pencil beam of light." In any case, my Calumet shutter speed tester doesn't measure light transmitted while the shutter is open, it measures how long the shutter is open. It starts timing when it sees light coming through the shutter, stops timing when it no longer sees light coming through the shutter.

Click to expand...

Dan -

I have that same timer and I agree with you that it works fine. I have found that an open bulb light source very close to the front of the lens or camera, with that timer as close as possible to the film plane and centered over the middle of the frame, gives me good results. I take 3-5 readings and average the results, expressing the variation in f/stops. I have done this for dozens of shutters of all kinds and keep a spreadsheet with a tab for every shutter ever tested.

As noted in my prior post, it doesn't take into account the difference between edge speed and center speed (aka shutter efficiency). But I don't think it's much of an issue even at higher speeds. I've mostly used mine to spot gross errors to determine when a shutter is really out to lunch and needs a CLA. But I've never thought of this as being a high precision activity. No other part of my workflow from my meters to my enlarging timers are high precision

Dan Fromm said:

SalveSlog, reddesert, I'm well acquainted with the idea of shutter efficiency, but thanks for reminding me. Reminders never hurt.

reddesert, when I test my shutters' speeds I don't use a light source that projects a "pencil beam of light." In any case, my Calumet shutter speed tester doesn't measure light transmitted while the shutter is open, it measures how long the shutter is open. It starts timing when it sees light coming through the shutter, stops timing when it no longer sees light coming through the shutter.

Click to expand...

Dan, it's not just the size of the light source, but the limited size of the detector, that can result in the mis-accounting for shutter efficiency. This also depends on where the sensor is located, forward/at/back of the focal plane of the lens. Depending on the source and detector locations, it is possible for light from the source through the outer regions of the lens (wide aperture) to largely miss the detector, so the tester winds up only measuring the beam through the center. As you note, the design of the shutter tester and what threshold it uses to trigger on the rise/decline of the light source also matter.

This is all somewhat head-of-a-pin-dancing, since it only matters near the fastest speeds of the shutter and the fastest speeds the tester can measure. But if one is going to say that all leaf shutters are off at high speeds (which has been expressed by other posters in this thread), then one must control for these effects.