I do AF runs with a red filter, and pause during autoguiding. I also use backlash comp in SGP. What I’m seeing is that after the AF run completes, PHD2 seems to restart before the imaging filter is put back into place, or at least before the backlash comp move for that filter is completed. Because I’m using a standard SCT - meaning mirror flop/image shift - sometimes PHD2 fails to recover the guide star. That’s because when guiding restarts before the backlash comp move is done, the guide star may be outside the search box, and PHD2 might grab a hot pixel instead.
Can SGP be set to not restart autoguiding until the new filter is in place and the backlash comp move for it is completed? As long as it does that, the guide star should always be pretty much centered in the search box.