Update: We continue to have the problem (e.g. May 31-June 1) - the intermediate solution (until we can take a look at the hardware) will be to re-initialize the filterwheel positioning regularly, which requires a software update (to be made very soon!).  Until then, we will have the situation that the filterwheel should be OK at the beginning of the night (hardware reset by hand!) but may fail and shift the filter sometime thereafter during the night.  (sigh)

 Update: In the following night (May 29-30) we still had the problem, but now it should be solved.

 Last night (May 28-29) the filterwheel glitch appeared again. Technicians at SAAO power cycled the telescopes  this afternoon and it seems that everything is working fine again. However, your images from that night have most likely been observed in the wrong filter.