Further investigation in to the reported zero point variation has indicated that this might be related to the rotation of the filter in the slot, however this is not well understood. The twilight flats from 2007-2009 indicate that the light distribution changed in B and V about the same time in mid September 2007. The V-band came back to normal at the end of January 2008 and the B-band at end of February 2009. The current flat fields are very similar to the ones before September 2007 and provide a good correction of the raw data.
The tests have shown that rotating the instrument between flat field exposures does improve the resulting flat field by up to a few percent (the report zero-point problems were in the order of several percents).
To make standard star observations easier and more reliable we are planning to make a peak count calculator for the standard stars. The idea is to have a web-page similar to the Exposure Time Calculator where for a given `ALFOSC' standard star field, filter, seeing and integration time, the peak counts are estimated.
One of the problems with obtaining flat fields during twilight is that is it not always clear what is the most optimal way to obtain flat fields in many filters. Currently we have a script that takes an optimal set of flat fields for a given filter, and the idea is to provide a wrapper script around this script that accepts a series of filters as arguments, and figures out the optimal order of the filters, and then takes sky-flats for all filters in the list.
Both these facilities should make things easier for none ALFOSC users, but also allow for a better definition and planning of observations (e.g., service or ToO observations which require observations of well exposed, but not saturated standard stars).
Thomas Augusteijn 2010-02-09