Less Pentax 6x7 focal plane shutter shake with the shutter curtain on the tripod axis
The consensus of Pentax 6x7 users seems to be that the focal plane shutter prevents it from being used with long lenses at less than 1/60th of a second or so because the action of the shutter in motion (a torque) causes a reaction of the lens in the other direction during exposure. As a new user and a physicist I wondered if this problem could be minimized by zeroing out the angle of the lever arm (r) of the first action so that the torque (= r x m change in speed per unit time) would be zero (and hence Newton's reactive torque would be zero).
(The conventional response, interestingly enough, has been to have a very high change in shutter curtain speed per unit time (a large torque) with the hope that the reactive torque would not kick in until the shutter was closed. Alternatively, one could use slow shutter speeds and correspondingly smaller torques, or exposures greater than one second allowing time for the shake to go away for most of the exposure time.)
The cross product r x Delta_v equals the magnitude of r (changing in time as the curtain moves) times the magnitude of Delta_v (changing in time as the shutter starts, travels, and stops) times the sine of the angle between them. If one sets the camera up so that the shutter curtain crosses the tripod's rotational axis as it moves, then r is always parallel to Delta_v and the angle between them, and hence its sine, is always zero. This would yield zero torque and thence zero reactive torque.
In this picture, small deviations from the shutter curtain crossing the tripod's rotational axis would give small torques. The worst possible case would be attaching the tripod to the tripod mount that comes on 400 to 800mm lenses instead of the tripod mount on the camera.
I tested this principle with a Gitzo R-3 tripod head that has a 6 inch long platform with a 2.5 inch slot anywhere along which the camera can be mounted, and an additional mount on the other side of the tripod column. I set up an MLU Pentax 6x7 with 300mm lens in three positions on this head, shot an Edmund Scientific no. 83.001 USAF resolving power chart, and viewed the resolution on the film in my enlarger. The placement of the camera in the conventional position with the shutter as far as possible from the tripod axis gave resolution twice as bad as when the camera's shutter curtain was directly over the tripod's extension column. The additional mount on the other side of the tripod column gave an intermediate distance and an intermediate result. With both a T-6 doubler and a Gray doubler behind the 300mm lens to produce the worst possible scenario of a 1200mm lens, the results were 4, 10, and 7 line pairs per millimeter (LPM) resolved, respectively.
I am convinced enough by my in-house tests that I have modified my R-3 head by unscrewing the bolt around which the H to V tilt is done and mounting the platform backwards. This puts the shutter curtain almost perfectly above the tripod's center. (One must also grind a 3/8 inch deep arc in the corner of the platform to reach a full 90 degree orientation with this tilt.)
This test was done at exposures of about a second due to the dim household lighting I have available, but I haven't had weather good enough yet to test this in the field at 1/30 or a second or so on some real-world subjects. I am hoping that some of you who have Arca mounting plates that slide could try to replicate this experiment in either setting.