The Hasselblad Shutters

As I mentioned it before, I had tested the shutter lag on the 500 ELM with the mirror locked up and found it to be around 32 to 40 milliseconds. I referenced the instructions on this page for making an electronic release cord, using also the Hasselblad service manuals here, and using a limit switch I triggered the 500 ELM with the lens shutter set to 1/500 of a second. The circuit was such that one of the poles of the switch would go from 0 to 6 volts when the camera was triggered, so I could see it on the oscilloscope. Shining a flashlight into the front of the lens, and putting a photo-detector behind the body, I could compare the time between me triggering the camera and the photo-detector getting a voltage. This did not measure the lag to full open, which is the time at which the flashes would be fired, but that could be done with the high speed camera or precisely timing a flash pulse as Frans had done in his tests. It didn’t matter, because the timings I was getting were pretty disappointing anyway.

Months after the tests with the Ilex shutter, I repeated the Hasselblad tests, then dug into the “broken” 500 ELX which I bought on eBay (and it turned out to be fully functioning) to look carefully at the shutter firing mechanism. The camera connects to the shutter in the lens via a coupling that sort of looks like a slotted screwdriver (on the body side), and this screwdriver in the camera body is tensioned by a torsional spring on a gear that is held at certain positions by three levers. The gear meshes with the winder motor (and the knob in the manual versions of the camera) and the mechanism which controls both the rear auxiliary shutter and the fold (reflex) mirror. One lever lets the torsional spring uncoil a set number of degrees, which opens the auxiliary shutter and brings up the mirror – this is the mirror lock-up position. The next lever releases the spring to uncoil enough to fire the shutter in the lens. The timing mechanism for the shutter speed is in the lens shutter itself; this connection to the camera body simply trips the shutter. The third lever maintains the shutter coupling gear at a certain angle while the shutter button is kept pressed down, which, in B (bulb) shutter mode, will keep the lens shutter open as long as the button is held down.

But the second lever is the one that matters. This is coupled, via two other levers, to the solenoid which is connected to the shutter button in the EL cameras. All three of these levers hinge on the same pivot inside the camera body. From my observations of the camera taken apart with a lens attached, it seemed as though this shutter lever would actually hold the firing mechanism nearly 25 degrees from the point at which it actually fired – so I figured that if I could modify it to hold the mechanism much closer to the firing point, thus reduce lag.

A few days of tinkering followed, and finally I figured the only way to modify the mechanism appropriately would be to replace the rotating lever with a sliding tab, since when the mechanism was at the firing point, the tab on the sprung gear which is held in place by the lever is nearly level with the pivot point so a rotating lever would not work in holding it at that level (without it having to act against the spring itself to fire the camera). (I realize the mechanism is hard to visualize, but perhaps if you take one of these cameras apart yourself you will see.)

I wanted to keep the rest of the functionality of the camera intact. So I needed to find two pegs in the body which I could use to make a sliding mechanism to replace the middle lever without obstructing any of the other mechanisms. I took the whole body apart to blueprint the location of all the pegs I could get access to, and design a part that could be cut with wire EDM on two axes. Wire EDM Masters gave me a good price ($80) to cut it out of steel. As I suspected, to actually fit it in there, I had to modify it a bit, going at it with the dremmel, but unfortunately I also had to modify the third lever (which maintains the shutter open in bulb mode) – I never like to modify stock parts in these situations, but I was really excited at this point and had no choice. I should have found some washers to take its place, but I would have had to grind at those, too, and, well, I was impatient. The figures below show the new “S-arm” as the service manual calls it; on the left it is mounted by itself and on the right after all the modifications it is mounted with the other two levers. On the left picture you can see the gear that is spring loaded on the lower left; as it spins clockwise, first the mirror moves up and the auxiliary shutter opens (and the lens aperture stops down to its set value and the shutter blades close), then the lens shutter fires, then comes the bulb-mode hold, and after that, the lens shutter is closed and one of the tabs on the gear rests against a tab on the camera body to prevent it from unwinding multiple turns.

Sliding version of the S-arm, straight from Wire EDM MastersThe new S-arm, modified to fit with the other two levers in the mechanism.

I took it home that night and fired it a few times (maybe 30) to loosen it up a bit. Unfortunately I noticed that the “practice” lens I was using – my oldest 80mm, which has a nick on the rear element – was firing after locking up the mirror. This meant that my sliding mechanism was past the point of firing. I tried it with a 250mm lens that I have (which is in good shape) and it didn’t fire prematurely, so I took it back to the lab and measured the delay. To my surprise, it was 47 milliseconds at best.

I took the whole thing apart after a few more tries and saw what happened: the impact of the sprung gear on the sliding S-arm was so strong (and the S-arm so thin over the left-most peg) that it got bent completely out of shape. Looking back, I could have thickened it a bit, then left only the very tip short so that the gear would be held at the right angle.

The new S-arm after a few firings; it could not take the impact.

But this brought another point to light – since I had tried it with two lenses and one fired prematurely, it meant that the angular precision of the point at which the shutter is tripped is not very tight, so this may all be a lost cause anyway. Since there was nothing left to do with that, I decided to try to fire the lens by itself using a servo. For my thesis I was using an RC style servo to move a flapper, so I had written firmware to turn an Atmel USBKEY demo board into a precise (in terms of timing) motor controller, and it even had a trigger out for synchronization. Months before I had bought an old-style Hasselblad bellows, which has a separate firing mechanism for the lens which is in essence the same thing the camera bodies have attached to the sprung gear. I removed the spring so it was just a free-turning axle coupled to the lens trigger, so that by turning it one way I would wind the shutter and then I could slowly turn it back until the shutter was tripped. By advancing the servo minute amounts, I could hold the coupling probably less than a degree from the firing position, then make the servo turn as fast as possible to trip the shutter. I don’t think I even wrote down the numbers – that’s how disappointing it was. I believe that the lag (again with the flashlight and photo-detector) turned out to be around 20 milliseconds.

At this point, things seemed pretty meek. The Ilex shutter did not seem so fast, but now seemed to be the most viable alternative. The Hasselblad lens shutters, which are optically at the best place (the aperture plane) were never going to be fast enough, unless I took the lens apart and manipulated them directly (which I wasn’t willing to try). At this point I came to the realization then that there is no reason not to use a large-format lens, thus maybe employing the Ilex shutter concept with a lens format that makes it every easy to put the shutter at the aperture plane. At the time the Ilex was still too slow, but now, after my last post, I figured maybe it could be solved with more current through the solenoids. In this case, both the shutter lag and shutter time share the same problem – and possibly the same solution. Unfortunately this timing-less shutter seems to be more rare than I thought – I’ve not been able to find any others, much less at different (bigger) sizes.

About this entry