Also, just another observation: even with the global delay, the slewing problem is definitely better but not completely gone. I experimented with the mount quite extensively today. Turns out, even with the global delay, I'm seeing the mount go out of control in about one in 4 to 5 manual slews.

Well, removing all my changes, and just adding a small delay after the StartMotor command, in the SlewDE and SlewRA methods, improves stability without impacting guiding.

However, there's something wrong with the mount control dialog in Ekos I think. It seems to be missing some key presses, resulting in occasional runaway behavior. The logs show the driver is receiving a start command from ekos, but no corresponding stop command (sometimes). I also see sometimes multiple stop commands in a row, with no corresponding start command. So something is definitely wrong there.

I think this might run deeper than the amount of time I have right now to dedicate to tracking this down.

May I suggest not using manual slewing for now?

Of course!! Thank you so much for your help so far. This is really helpful. I'm also more familiar with the indi-eqmod code base now that I can keep tweaking it if I have time. I have mostly got rid of my manual slew needs. About the only time it's necessary is occasionally the mount would wrongfully believe it is pointed below the horizon, therefore refusing to turn on tracking, which in turn prevent the alignment module from syncing to plate solving solutions. But I can slew from my desktop. It looks like it's my iPad setup that is causing the spurious second slew which is really the cause of the problem.

Thank you again! I wish there is a way I can send you a beer