Hi Harry,
Thank you for the calibration files. An additional 'signed-is-physical' hint was missing for the image integration process, which will be added in the next version. However, even with this improvement your calibration files don't work.
The main problem is in the bias frames. Your bias frames have median values of 0.03 on average, while your dark frames have 0.016. So bias frames are about twice stronger than darks in your calibration set. When your master bias is subtracted, it leads to severe clipping that destroys all the calibrated frames. There are no PEDESTAL keywords in your FITS files. A pedestal of about 1400 ADU would be necessary to prevent a massive amount of negative values after bias subtraction, but this is obviously wrong because it would lead to saturation of many bright image structures. Note that normal pedestals are always in the 50 - 200 ADU range, and are only intended to fix relatively small amounts of negative values after bias subtraction. Pedestals are seldom necessary in practice, in fact.
I don't know what happens with your calibration files. If this set is representative of the data you normally acquire, and you get good results consistently with the software that creates them, then there is something it does that we don't know how to do. Our calibration routines are working perfectly with data acquired with the rest of applications, including Maxim DL and Nebulosity. We use our ImageCalibration, ImageIntegration and StarAlignment tools all the time with data acquired with professional telescopes and CCD cameras without any issue. I'll keep trying, but right now I don't have any clue as to what's going on with your data.
