ajobacho Yes, that is the solution, but it is a waste of time.
When I do my own Meridian flips, I pause imaging and guiding just before an exposure would have overlapped a Meridian transit, and immediately execute a GOTO to a target with an hour angle that is just a little past the Meridian. That will (with any German mount) execute a Meridian flip.
I then use the "dead time" to recenter the OTA on that virtual target, and refocus the OTA. I then wait (sometimes not even need to wait) for the actual target to also transit past the Meridian, and execute a direct GOTO to it.
Very short time to get there with the direct GOTO, and since the Meridian flip has been already been recentered, the new GOTO takes at most one plate solve to refine its position. Basically, the "wasted" time is taken up by the Meridian flip, refocusing, recentering, etc. So, not really wasted.
If the program supports scripts, programs could automatically pre-stage itself this way too.
I think that 15-20 minutes is enough
I agree. If the mount has been star sync'ed (or east-west leveled), you don't need even a minute. All that has to happen is that both the mount and the controlling program agree that the target has transited past Meridian. If everyone agrees to 1 arc second in Right Ascension, the time discrpancy just needs to be 15 seconds long on the wall clock.
(Don't forget the "other Meridian" too for circumpolar targets.)
When I do use auto meridian flips, I set up my mount to stop 5 minutes (time) before Meridian, and then set it to perform the actual GOTO 5 minutes after transit. That is plenty of time since I make sure that my mount is east-west leveled (my mount has a calibratable index mark).