Re: Last Night's Run

[<aah@nofs.navy.mil>]

I will say, just one more time, that the bias drift can
be removed easily if you implement overscan.
 Tom's scheme on image collection (unless he has changed
it recently) is to read out the previous frame, then
expose, then read out the new frame.  The two reads
make it so that each pixel has the same amount of dark
current + a shuttered exposure.  However, what is variable
is the time from the previous read until the shutter opens
for the current read, since you might move the telescope,
store the last frame, etc.  So the amount of dark current
is not always obvious nor constant, even if the CCD
temperature is accurately controlled.
  A single bias frame is not the way to go, since the
noise in that single frame will propogate into all data
frames.  You should create a 'master' bias by averaging
many such frames.  Doing this at twilight is not the
right way to go either, since there are light leaks
in the system.
Arne