> > Next question, how much more accurately (if at all) is the diaphram > controlled on an A lens rather than a K lens.
Simple question, tricky answer. First of all: there's no requirement for the aperture response of any K lens to be the same as a different lens; all that is required is that the lens stop down to the aperture selected on the aperture ring when the mechanism is actuated (within a given maximum angle of travel). In practice most of the K (and M) lenses seem to share the same basic mechanism, so the response is somewhat uniform across the range. But there could well be a small number of lenses that behave differently. There was also no requirement that the lens actuation mechanism had to be at exactly the same place on the lens mount, as long as it was capable of stopping the lens all the way down within the permitted angle of operation. Again most K and M lenses do seem to be the same. Only with the introduction of the "A" lenses was there any requirement for a systematic uniform response across the range. "A" bodies do not use bang-bang push-it-to-the-limit diaphragm controls; in program or shutter-priority modes they position the diaphragm control lever to intermediate positions, and expect any "A" lens to stop down to the same f stop (if possible) for a given position of the control lever. The response curve that the "A" mount specified was not the same as the empirically-observed behaviour of the "K" lenses; the "A" mount required the same angle of travel to change the aperture by one stop from any aperture. This is not how the K lenses behaved, but it did greatly simplify the actuating mechanism in the "A"-enabled bodies. It also allowed more precise control of smaller apertures; half the travel of the actuating lever on the K lenses was spent on the first two clickstops down from full aperture, and so on. If the "A" lenses had followed this pattern it would have been very hard to accurately control the smaller apertures. With the design chosen more of the range of actuation is available at the higher-numbered f-stops, and any error (due to misalignment, etc.) will be uniform across the whole range of apertures. Now I've actually done the calculations, it appears to me that the actuating lever of an "A" body in automatic mode does *not* move far enough to stop down a typical "K" lens to the right position at any aperture except wide open and stopped all the way down, so trying to fool the *ist-D into thinking that an old "K" lens is, in fact, an "A" lens on the "A" setting will result in overexposure.
