We found that the fault on the video board in the detector controller used by the detector that showed the recurrent problem was indeed the cause of all the problems. The solution was a swap of the faulty video board with another board in the controller that uses the other half (channel) of its board. During the clean-room tests the detectors were operated warm and this gave bright columns from two chips but at the same relative positions. After spending sometime trying to identify the cause of these features the instrument was cooled and the bright columns disappeared, showing them to be `normal' features of the detectors. Since the solution was implemented MOSCA has been cooled down for three runs and it has been used in eleven nights without any problems with the top right CCD.
The MOSCA quality routine has been integrated into the sequencer environment, to the extend possible, see:
http://www.not.iac.es/instruments/qc/mosca_howto.html