I would just re-emphasize that the author of BASS is re-writing the calibration part of the software. So expect changes to BASS calibration, and I’m not sure where the deficiencies currently lie in the BASS calibration routines. That is why I was suggesting MaxImDL. Once John has finished with the BASS enhancements I am hopeful that will contain useful features.
I usually take a minimum of 20 calibration frames. This reduces the noise when they are combined, which is an unavoidable element of calibration frames.
-5C is OK but if you can go colder then that will reduce the noise in your spectra. Just emphasizing again that the temperature of your CCD when taking the calibration frames must match the temperature when acquiring spectra. So you might need one set of calibration frames at -5C for your older spectra. Then another set at a cooler temperature if your CCD can go colder and hold a stable temperature to within at most a few tenths of a C.