Checking darks for dark signal

Adam,
I loved your explanation of the flat dark. I am thinking of ways to check dark current accumulation and the need for a flat-dark versus a bias.
Assumptions:
1. If you subtract a bias from a bias I would expect a clean frame with ADU counts at or near 0. Correct?
2. so if you are worried about whether you can get away with only calibrating your flats with a master bias, could you not subtract your 10 second or 30 second flat-dark from a bias and judge the level of the dark current by checking the average ADU could of the resulting frame?

If so, then what you would need is an algorithm to test whether the residual ADU count was significant relative to the quality of your flat.

Comments

  • Hi Ed,

    1. You would expect to see readnoise- whatever the expected level is for the sensor. (some of the values will even be negative..but representated as zeros unless you add a pedestal.)
    2. Yes, if it does not significantly differ from the read noise...but- there is a gotcha. Small differences may be hard to detect and the accumulated error(s) you get from calibrating and stacking many frames might show a difference that is hard to see in looking at the differences of two (or a few) frames. 

    Since the sensor will be consistent whatever it does- you might want to compare the difference between the STACKED calibrated images being the result of 10 second or 60 second darks. Everything else would be held constant. If you cannot see a difference in the results... then you can be pretty confident- since this is a more sensitive result by combining many frames. 

    -theBlockhead
  • Adam,
    Many thanks for those comments. I will begin teaching the CCD1 course for AAVSO next week and will have the students do some comparisons of stacked darks taken at obviously different integration times. We will see what they come up with. I am also recommending they view your #3 WBPP, very nice discussion.
    Ed
Sign In or Register to comment.