One of the scripts uses the MEDIAN ADJUST command – it adjusts the key word for exosure time so that a better guess is used, based on the previous exposure time and attained flux level. It seems to work in the DOMEFLAT_COMMANDED.csv script, except for B and VE2 filters, where the exposure times suggested is 180 seconds. This is the limit (to avoid wasting too much time) and the corresponding images are underexposed. This is basically because the CCD camera is RED-sensitive, not BLUE-sensitive, and because the paint on the inside of the dome at MLO is not very reflective at red wavelengths. The V, VE1 and IRCUT images are properly exposed to about 51000 counts as they should be.