Gubi,
The documentation has been published, so changes will go live pretty much immediately:
indilib.org/telescopes/skywatcher/skywat...lt-az-dobsonian.html
My understanding of SYNC and SLEW TO TARGET matches yours. My concern comes from a search of the code to find where the "Sync Failed" message is thrown. It comes from align.cpp and not the driver. Further, there is a notation in the alignment code that indicates handling of Sync has previously been an issue.
In the case with which I am concerned, the plate solve succeeds and outputs to the log the correct coordinates. This is followed by the Syncing Failed message and generates a retry. By now, tracking being off (or perhaps on, I am not certain about this), the target has drifted a few more arcseconds. So, the new image is a bit further away from the target. The retry solves, but "Syncing Failed" repeats and the cycle continues until the retry limit is reached.
I have also seen the issue where something does not solve for no good reason. That is a different issue and may be due to photo quality. When that happens to me, I typically bump the bin up a notch and retry. Then it typically solves.
If you have not tried guiding lately, you should retry. The latest changes made to the driver made a great difference. I did a great deal of testing using the Guiding Simulator driver as a CCD to get my head wrapped around the process during daytime. Guiding really does make a difference. I would try the internal guider first, as it calibrates faster, which for me is less frustrating upon failure. I believe the guiding is very close now to being of great value, if one can just find the right spot to tweak. I have taken as many as 500 shots at 10 seconds with the image remaining centered. Granted, there is a high number of throw-away images due to overly-aggressive adjustments.
A question comes to mind: Do you know if the slew rate impacts guiding at all?
There may be another evil at work in the mix with our scopes as well. I am not convinced that some of the "Freedom Find" or Dual Encoder Technology or SoftPEC isn't getting involved at some level. Too many cooks in the soup! I may try turning some of that stuff off.