Thoughts about the inherent inacuracy or Kiev shutter

[Spyderman]

This post has been inspired by the results of measuring the accuracy of the shutter of my 2 Kiev 4 bodies. (I built myself a shutter tester out of 2 phototransistors and a PC sound recording software to record the output of the transistors.)

I always thought (or maybe I read it somewhere long ago) that in Kiev shutter there can be no "capping" or iregularity of exposure between bottom and top part of frame.

The measurements show that (at 1/1250 setting) the top part of the frame is exposed for 1/334 and the bottom part is exposed for only 1/621 of a second. That is a difference of almost 1EV . Both my bodies behave similar at higher shutter speeds (only slightly different numbers). Increasing the shutter curtain spring tension seems to pronounce the difference between the speed of the 1st and the 2nd shutter curtain.

But why is it so ?
Kiev shutter uses 2 ribbons to pull the closing (2nd) curtain down. The opening (1st) curtain is pulled down by the drum, but also rides on the ribbons of the closing curtains.

In the fast speed range 1/125-1/1250 the action of the shutter can be described in 3 phases:
1) the opening curtain starts moving. it is pulled down by the spring inside its drum, the slit between the opening and the closing curtain is set before the exposure, and the opening curtain is held on the ribbons (pulling the closing curtain) by friction
2) both curtains travel from top to bottom of film frame exposing the frame through a slit between the opening and closing curtains. As the curtains travel they gain speed being accelerated by the shutter spring preload.
3) after the opening curtain reaches its stop, the movement of the ribbons pulling the closing curtain is being slowed down again by the friction between the ribbons and opening curtain.

EDIT: I changed the above paragraph so as not to confuse readers with my wrong assumptions - I rather corrected it.

I think this explains also result of my measurement with the shutter tester. The top part of film gets more exposure because the curtains are just accelerating in that region. The exposure is shortest at the bottom where both curtains reach maximum speed.

In the slow speeds range 1/2-1/60 the differences between exposure of top and bottom part of film frame are up to 1/3EV. The action is slightly different here:
1) the opening curtain travels from top to bottom of frame, constantly accelerated by the spring and slowed down by the friction of the ribbons.
2) the clockwork measures the time until the closing curtain starts moving
3) the closing curtain travels from top to bottom of frame, constantly accelerated by the its spring and slowed down by the friction of the ribbons. Thus the exposure should be much more consistent and regular within top and bottom of the frame.

PS: funny thing is that I haven't noticed this exposure irregularity in real world photos (though I must confess that havent shot many films with Kievs). My thoughts were inspired only by the desire to explain the measured numbers. Probably because most photos with sky in them are horizontal, and it's difficult to see the exposure difference of less than 1EV in anything else than equaly lighted sky or large water surface. On photos from Zorki or other horizontal curtain shutter any such irregularity is easier to spot.

No cameras were harmed during this shutter testing.:angel:

 

[Grytpype]

As far as I understand it, at the faster speeds (1/125 to 1/1250) both curtains should run simultaneously, held together by the friction of the ribbons, the exposure being controlled by the width of the gap. That's why the pre-war Contax and the Kiev shutter shouldn't cap at high speeds.

Since you are seeing significantly shorter exposure at the bottom of frame (the end of the exposure), it suggests your curtains are accelerating too much through the frame. Perhaps the tension is too high.

 

[Spyderman]

Maybe you are right, but at the 1/1250 setting the shorter exposure is only 1/600 which would indicate too low spring tension... or too wide slit between the curtains. But because the slow speeds are approximately OK, I assume the timing of the shutter clockwork is also OK.

Anyway thank you for patience to read it all and reply.

 

[Grytpype]

I'm assuming that by "top of frame" you mean the top of the film gate: not top of the image. Either way, the speeds should be much faster. Have you got the full area of the phototransistors of your tester exposed? At fast speeds you really need a very narrow slit in front of them to get a reasonably accurate result. I found this out the hard way when I built a 3 x phototransistor tester (to give a centre-frame reading as well)!

 

[Spyderman]

Good point. Thanks. The transistors are in a plastic box behind 0.8mm diameter holes. I might make a mask for the holes out of aluminum can (by a needle - similar way as pinhole camera holes are made).

 

[Grytpype]

You may find that a pin-hole doesn't let enough light through to give a readable spike in your audio program. I used a mask of thin black plastic (about 1/2 mm), cut slots with a craft-knife and used a .010 inch feeler gauge (.25mm) to open them out so they were the same width as each other.

 

[Spyderman]

Thank you for the useful tips. I will probably try the coke can material, sewing needle and a microscope to check the hole size.

BTW: how much error could the size of the hole cause ?
for example Leica shutter travels 32mm in 11ms which is speed of about 2.9 mm/ms. It takes the shutter cutain edge 0.25ms to travel 0.8mm across the hole in front of the phototransistor. Thus if the exposure time at 1/1000 should measure 1ms, the shutter tester with the error will measure 1.25ms or about 1/800.

 

[Grytpype]

Your calculations look right to me, though the Leica shutter is faster than I'm used to.

Most of the horizontal cloth focal-plane shutters I have met have run at around 1mm/mS, if I am remembering rightly, with a 1mm gap (approximately) at 1/1000 sec, so a 1/4mm hole/slot would show an exposure time 25% higher than the true figure, assuming that the hole is smaller than the sensitive area of the phototransistor, and correctly positioned above it. I think the curtain gap on the Kiev at 1/1250 is also in the region of 1mm.

The error would obviously decrease rapidly as the shutter speed is decreased and the curtain gap increased. The Leica shutter you mention, having a curtain speed about 3 times faster, would have a wider gap for the same shutter speed, so the error would be less. Is that a bayonet-mount Leica?

Another point is that the light source should appear as much like a point-source as possible, as far as the shutter tester is concerned, so a small powerful light at the other side of the room is best. I think you are better with a slit in the mask than a hole. You will get a higher output and you only have to bother about the positioning in one direction.

 

[Spyderman]

The measured speed is from a M6. I measured the time between the start of rise of signal of the 2 phototransistors (11ms) which are positioned 32mm apart.

I don't remember the actual speed of the Kiev shutter.