I'm really fighting with field rotation.
If your mount is not aligned perfectly to the polar star, even with good guiding, this puts strong constraints on your exposure times. Even off axis guiding does not solve this problem. If you miss aligned just 1', the guiding star is 2 degrees away from your image and you're 60 degrees above the horizon, field rotation will accumulate on a 30 min exposure to about 5 pixel on my Atik 490ex.
As I build and tear down my setup for every session, there is a limit of time I'm willing to spend for alignment. All Star Polar Alignment on my mount gives me about the 1' precision mentioned above. For now I will limit the exposure times, depending how far up I'll look.
For the future I may investigate drift alignment and how good I'm able to streamline this process.
Anyway, here's my latest result based on things learned from all the previous mistakes. I especially looked for tracking errors, exposure times and focus.
3x15 min each in Ha, OIII, SII; 3x2 min each in RGB for the star colors