I did manage to clean up most of the problem. It seems I had moved some bones in Pose Mode (I’m assuming this) before attaching the meshes, so detaching them, snapping things into their proper places, and then reattaching them went a long way toward getting things to move and track correctly.
I also made changes to anchor points for so-called target bones. Part of the problem was that the (track-to) target bone for the piston wasn’t snapped to the rotation point of the base. That helped a lot. I was trying to keep the target below the base to avoid crazy behaviour in the rig’s rest position. Turned out it wasn’t necessary for some reason although it was in the test rig I built originally to prove the concept.
But even with everything snapped into place, rotating around the correct point, etc. etc. it’s still off by one or two degrees. It’s close enough for what I’m doing, so no worries there, but if I were showing a close-up of that part of the piston for any amount of time I’d be more concerned, but since it will only be seen for a very short time in a shot where the audience’s focus will be on a character peeking under the bed, it’s no big deal.
It does have me wondering, though, how this type of thing could be made as accurate as possible.
I tried to upload the file so you could see what I’m doing, but this site doesn’t let me upload any more. Don’t know why. I doubt it’s anything to do with file size or type because some of the files I’ve tried to upload recently were ~270k and I’ve tried ZIP, Z7 and BLEND files.
Anyway, for now I guess I can say the problem is solved; it’s close enough for rock-n-roll, as they say.
Thanks for the help (and it’s not the first time, I might add) mathiasA.