If your statement (1) is correct, the camera indeed recognizes the lens as a PK/A. Now, what is the difference in moving the aperture lever with the lens set to "A" or to an aperture value?
- If a PK/A lens is set to an aperture value at the lens (or the lens is recognized as a PK lens), the camera will try to move the lever all the way through. The lens itself is mechaniclly limiting the closing of the aperture according to the set value.
- If the camera recognizes the lens as a PK/A type (set to "A" at the lens), the camera processor will calculate a limited move of the lever. This calculation is based on a linear mapping of the mechanical limits of the lever to the aperture value limits of the lens. The limits of the lever are fixed values, given by the Pentax design of the PK/A mount. The limits of the aperture can be asked for by the camera - there is a passive matrix inside the lens, connected to the contacts. The bit code of the electrical resistance between the contacts and/or contact to mount tells the aperture limits of the lens.
I think the camera must somehow get values which are not fitting to that lens. Such an error - no matter whether caused by faults inside the lens, at the contacts, or somewhere on the communication path - would have no consequences for shooting in "M" mode (as long as you set aperture
at the lens, not from camera). But it could cause all kinds of weird behaviour in any mode where the camera tries to actively control the aperture.
I think I am now running out of ideas - the very best solution would be a cross check: try the lens at another Pentax body, and try another PK/A lens at your camera body. You should prefer a true "A" lens: F, FA, DA and DFA lenses may provide additional active electronic circuits instead of mearly resistors, and this could make it difficult to draw conclusions.