I am trying to build an APPM model for my f/10 Cassegrain, now using a mono imaging setup with a 0.40 image scale and variations on 1x1 and 2x2 binning, all with limited success. This leads to two questions.
First, could I be getting spurious results? APPM keeps telling me that my telescope on the west side is off in RA by as much as 12 degrees. This seems very strange. Twelve degrees is a lot. I've checked for cone error and it is under 8 arc-minutes. The errors are only in RA, not DEC. The mirrors in the Cassegrain are fixed -- no mirror flop. The model on the East side is far better -- pointing is off by 5-15 arc-minutes. What explanations are there for such high pointing errors on one side (only) of the meridian? Could this be a result of my settings leading to spurious solutions? Only about half of the model points result in a solution anyway. Note that I've done this twice, with two different cameras, and got similar results.
One option to give up on an automated camera solution and build, instead, a T-point model manually with 20-30 points. I'm not worried about tracking accuracy in this setup as much as pointing accuracy. I then wondered if I couldn't build an APPM model "manually" with a series of Platesolves/Recals. This leads to a second set of questions about how APPM models behave...
How does a Platesolve/Recal executed through APPM work in cases where an APPM-built model is in already place and in use? Does it modify the underlying APPM model in any way or just make a one-time adjustment for where the telescope is at that moment pointing? Would that adjustment carry over to other pointing on that same side of the meridian? In other words, can I gradually refine an APPM model by a series of Platesolve/Recals and then save it as a new model?
Similarly, how does a Recal ("Sync") work in TSX relative to an APPM model? I assume it doesn't change the underlying APPM model at all. But will it apply the Sync adjustment to all pointing positions on that same side of the meridian?