Since then, continuous "zero count" intensities signalled a new problem, which was finally found to be located in the high voltage power supply. Both problems were repaired during a maintenance campaign in late August 2019 which re-initiated trouble-free operation.
Previously (in March 2018), the instrument had stopped to produce airglow data due to a shutter problem.
With a brief exception near the end of the night of 14 March, the shutter stayed closed, so that the only information obtained was on nightly variations of dark count, a slowly increasing number of events with missing motor steps, and the (very good) performance of the housekeeping data communication between the control PC and the internet server.
The only evidence available to assess the source of the shutter problem was the report of an anomalously low DC input voltage to the power supply serving the stepping motor system and the electromechanical shutter.
On arrival at the site on May 4, this low voltage condition was confirmed, but found not attributable to line transformer or rectifier bridge faults. Finding the "smoking gun" took another day, until, at the evening of May 5, replacing the 4700 μF, 50V electrolytic filter capacitor directly after the rectifier diodes by a new one brought the solution (and the first new complete airglow data night).
Unfortunately, maybe because of a faulty line power connector, the serial communication between control and internet service computer failed and it took two more days to diagnose and repair this fault. We had to switch to the alternative control computer COM port (with DB25 connector), the DB9 on the server machine, and make the corresponding software adjustments in both computers.
The instrument presently looks like this (status: 10 May 2018). See this link for earlier technical problems.
