Detector controller

On a few occasions the actual exposure times delivered by the array controller were found to be only a fraction of the requested exposure times, with the array controller deciding to read out before it should, while the header information contains only the ``commanded´´ exposure times. When taking short exposure this problem can actually be easily missed.

A program was made that for every exposure checks that the time it takes to complete the command, i.e. exposure time plus the expected overheads, is above a minimum number. If not, an error message is issued with the content ``You should power-cycle the array controller´´. The program has been now running for several weeks with no problems detected so far. The program actually records the overheads for each exposure, and can be used to systematically monitor if, and how the overheads vary with time.

Since the introduction of the Sequencer we have had spurious occasions where the detector control software apparently ignores exposure commands from the Sequencer. This has been seen to happen only for NOTCam. After waiting for a certain amount of time it gives up without any warning or error, effectively causing the skipping of exposures (e.g., in a dither pattern). Extensive tests show that, although the problem is intermittent, it has become worse over time, and there is a strong direct correlation with the exposure time. The cause is not understood, but a modification inside the detector controller software was made to detect the problem and retry taking an exposure. This is done up to a 10 times, and an error is given if it still does not work. Though this causes some overheads, it works for most exposures.