I'm trying to use the Oculars plug-in in Stellarium to determine the best camera orientation to ensure my guide camera will see an adequate number of stars. I've entered all the paramters into the Oculars Settings dialog as shown below:

Note that the "Position angle" is set to 0 degrees. With these settings, Stellarium shows this for the night sky view:

Within the OAG sensor window, you can see 3 relatively bright stars. In PHD2, I confirmed those 3 stars were visible and in the correct orientation. So far, so good.
Then I slewed to a different target to double-check, and nothing matched up. After fiddling for a while and getting nowhere, I tried changing the Position angle to 180 degrees and then the stars in the PHD2 display matched what Stellarium was showing. Hmmm...
I slewed to several other targets and got highly inconsistent results. I then realized that once I set the Position angle so the views matched, all targets in the same hemisphere (east/west) would be OK. But when the mount slewed to the other hemisphere, the Position angle needed to be rotated 180 degrees. This is 100% repeatable.
I suspect this is because I'm using a GEM and the optical train is rotated 180 degrees in a meridian flip (the mount is defined as a GEM in Stellarium). But if you select a target in Stellarium and watch the Oculars outline as you advance time, it does not rotate as it crosses the meridian. Have I overlooked some setting in Stellarium that enables it to model a meridian flip with equatorial mounts?