We’ve decided that the filter wheel has failed. This results in several things:
- bad flats. If you look at
- Index of /oc61
- and in particular, the files labeled oc61_yymmdd_report.html, you will see where the bad flats started.
- extra light in each image. Note that for many long-exposure images, the background looks mottled; in fact, there appears to be a definite pattern. This is probably caused by the LED that is used to identify proper positioning. If the filter wheel is not positioned correctly, that LED may stay lit.
- poor darks. As mentioned earlier, the July darks with this system were compromised by extra light. CMOS cameras do not normally have mechanical shutters, so to take darks, I typically wait until nighttime, move the spectrograph pick-off mirror into the optical path, and move the filter wheel to the dark or SU positions. One or more of those actions was not taken in July. The August darks are much better.
So it is possible that all July and August data from OC61 is compromised. Check your images carefully. Is the sky background a few 10’s to a hundred counts or so? Does the difference between comp1 and comp2 match the catalog difference for the particular filter? This latter test is the most important, as it tells you whether the correct filter was chosen.
We are sending a new filter wheel to NZ, but it may be a couple of weeks before the system is operational. The observatory staff will look at the filter wheel and see if it can be repaired there.