We also commonly see that the trailing is faint and the star points have pixels values many times higher than that of the trails; from this we can deduce that the trailing is happening over a period of time shorter than the normal galaxy camera exposure time.
My current thoughts on a solution to this problem lie in changing the order in which we image jobs. In the current system we image each colour channel first, and then image the dark frame that is applied to each of the colour channels. This design was to help reduce the memory footprint whilst processing a job and also with the idea that a dark frame could be imaged whilst the mount is slewing to the location of the next job.
|Comparison of old and new imaging order|
(left: old, right: new)