I just read through you posts on CloudyNights. It looks like your RMS guiding in terms of pixels were < 0.15 or so which in my mind is excellent. I don't know what the limits of PHD2's centroid algorithm is, but it seems to me that you would need to change your plate scale on your guiding camera to improve it any more. I think your guide scope has a focal length of 210 mm with a plate scale of 3.68 arc-sec/px. If you could get this to around 2.5 arc-sec/px then you may still see some improved guiding, perhaps < 0.5 arc-sec total RMS.
Director Electrical Engineering
Explore Scientific, LLC.