I've read multiple posts to try to understand what should be the proper settings, and at this stage my conclusions are:
1- All exposures should be longer than 0.2 seconds, as the sensor is not consistent under that.
2- Take light frames as usual, at lowest temperature reasonable (-15C for me these days), with proper gain and offset (gain 200 and offset 50 for me, as I do narrowband), and for me exposures are determined using help from the tables in this post.
3- Take matching dark frames: same length, same gain, same offset, same everything as the lights.
4- Take flat frames: adjust gain as needed so that exposures of over 0.2s are achieved, giving a SGP ADU readout of around 12,000-16,000
5- Take dark flat frames: same gain and offset as the flat frames, and same length. For me, this means one set of dark flats per filter.
6- No bias frames
7- In BPP, put Dark Frames in Darks, Dark Flat frames in Darks, nothing in Bias, Lights in Lights, Flats in Flats. Dark Optimization set to OFF. What I understand this does is:
a. Create a master dark of same length as light frames
b. Create a master dark flat of same length as flat frames, for each filter
c. Flat frames for each filter are calibrated with the master dark flat that corresponds to the length of each filters' flat exposure
d. Flat frames for each filter are calibrated into a master Flat
e. Light frames are calibrated with Master Dark (from step a.) and Master Flat (for each filter)
f. Light frames are star aligned/registered
g. Light frames are integrated into a Master Light
8- If needed, manually perform a drizzle or Local Normalization integration
For flats I am using a technique using the daylight instead of a light panel, tests have proven that the quality is much better then a light panel, and its easy to do.
Cover the telescope, Filter Wheel and camera to avoid light penetration to the sensor.