My problem was that the primer disk was not turning enough to bring the primer to the correct position. It worked with 1000rds/h that the guide rod made the last few degrees. But with 1500 rds/h the guide rod crashed into the primer disk (I already have the updated disk).
Solution: The inner spring (green arrow) on the compliance rod seems to be out of a cheap pen. So I was searching for another one with similar dimensions but a higer compression force. Found this one with approx 26N total compression force or 2N/mm. Left is the new one, on the right the original one. I think 26N is a good value. Enough to drive the disk without loosing the failsafe function....
Runs like a charme now
Sorry for the metric system guys