I learned out how to do this, when I was fast reflector fan, but I don't remember the exact details of the procedure. The trick has to do with the reflection of the primary in the diagonal when sighted through the focuser. It should NOT be symmetrical and concentric, but should be shifted back (towards the primary) ever so slightly.

The reason for this is because the right angle reflection at the secondary makes the "aft" part of the rays have to travel further to the eyepiece than the forward ones, and the secondary is big enough to subtend a bigger angle at the focus point. Consequently, the "fore" and "aft" rays are not perpendicular to the optical axis of the telescope. You could solve this by tilting the focusser forward (with shims), or more practically, by collimating it so it looks like it is out of alignment. The root cause of the problem is that in a fast scope the big diagonal is very large compared to its distance from the focusser. It's a parallax issue,

Follow that? I didn't think so. I can't really follow it myself. But I remember where I learned the trick, it's in the data/tips section of the old Norton's Star Atlas (where all the fine print is at the back).

Good luck.