well, i suppose, if you can, you can finish the job of calibrating and integrating the lights with CCDStack. past that point PI should be compatible with the stacks; might require tweaking the fits reader again, or if you can convince CCDStack to output in the range [0,1] then you are golden.
if you don't own CCDStack... well... hmm. i don't know offhand about any full data repositories that are out there with i16 subs of bias, dark, flat, etc... that does not mean they are not out there, though. i am not sure anymore what's available in the hubble legacy archive but perhaps there is some useful data there.
is your goal to learn calibration and integration in PI or to do post processing?
how was your own data acquired? if you have light pollution gradients and vignetting, that may account for the wild colors even after BN. in that case you should probably do DBE or ABE before attempting to neutralize the background.
or, it could be that an STF which was computed before BN was performed was still applied to the view, and you see crazy colors; in that case doing another AutoSTF after BN should get you back to something that looks better.
rob
