For some reason I was thinking you had the older Gemini model (you are entitled to slap me silly
)....since you have the round DIN motor connectors...you are sweet..all is ok.
You can still use the opto isolator in line with your auto-guider. It doesn't need it but some people like the extra "isolation" protection it gives.
Sorry I have never used CCDOPS so I cannot comment on it's effectiveness. I still suspect a good drift align is the best method.
Ok your image is not that bad. It's a bit hard to tell but it looks like some field curvature and maybe the CCD is not quite square on. Is your spacing to you CCD correct from the .63 flattener? This is the biggest influence of field curvature. I have made this mistake myself , recently
.
As I said It doesn't look like miss-tracking as I would have suspected that the self guiding head would be the bee's knees' in tracking. I suspect the guide head also does it's own calibration routine but you can go into the Gemini menu and set the Gemini guiding rate between 0.2x sidereal and 0.8x sidereal. This may help your guide head corrections (although I don't think this is the issue).
Something that may also help is the
Gemini ASCOM driver. You can use this like the Gemini GCC software..but this one is free (you need Microsoft net 3.5 and ASCOM 5.0a (or more) installed for this to work, just read the requirements).
With this you can check all your Gemini settings with this ASCOM driver and save them to and from the Gemini/Computer.
Well that's about all I can add apart from saying it looks like field curvature and I'm probably completely wrong!
but don't rule out polar alignment.