I also agree that imaging under the pole can be of interest and should be straightforward with a bent knee pier.
One thing I've noticed if you set your limits on top of each other is that the RA axis gently oscillates at the end of a slew instead of tracking. Like a PID controller that's not quite tuned.
If you don't mind being weights up then perhaps you could take it out of GEM mode. That will get rid of the limit settings.
Unfortunately for dimensional and spacial reasons I need to be weights down when begin to image; be that in the East or the west. Which means that ultimately Don's suggested settings won't work for me if I choose to begin imaging looking west. I need the mount to function as a GEM.
But I'm going to try Don's settings to check if they illuminate the unwanted slewing during plate solving in the east. Either way it will be an important test.
I shall also try with APT instead of SGP, a different plate solver and a different camera. That should rule in\out any possible equipment compatibility issues.
I shall also test using the ASCOM camera simulator. You can load an image into it and run a real sequence which means I can test during the day.
Hopefully soon I'll have some test results.