Thanks for the data, Erik. It all looks OK. You've done one thing differently to the way I used to build my calibration masters - you haven't calibrated the master dark to remove the bias. I can tell because your master bias has a mean value of 2803 and the master dark has a mean value of 2873. If you'd calibrated the darks to remove bias your master dark would have a much smaller mean value (around 2873-2803 = 70 ADU).
Both your master bias and your master dark include bias. That's why you need to tick the "Calibrate" box for the master dark when you calibrate the lights. That tells PI to calibrate the master dark with the master bias before applying it. Otherwise, it will subtract the bias twice and you get clipped data.
It's perfectly valid to do things the way you have, so long as you tick the right box
FYI, your read noise is 14.71 ADU. I calculated this by subtracting the master bias from the single bias frame and calculating the standard deviation of the result (which is just a frame full of read noise!)
These days I don't bother with manual calibration and integration to create master frames and calibrate the lights. The BatchProcessing script does this just as well with less effort. I always do my final integration of the lights manually though. That's the part of the process where careful tweaking of the rejection algorithm and parameters can make a big difference to the quality of the result.
Cheers,
Rick.