Sloping bias frames

[roger smith x294]

Paddy,

I have seen exactly these symptoms, with one important difference:  the
slope on the bias only occured in those frames taken immediately after
a flat field (and was seen even more strongly in the overscan of the
flats themselves).  I'm not too optimistic that the situation we had
would be reproduced in your spectrograph, but perhaps this account will
suggest a line of investigation to you.

My bias decay occured in an old back-illuminated Tek 1K.  By projecting
a thin slotted mask onto the CCD and rotating the mask, I showed that
the effect was produced by illumination of areas outside the imaging
area of the CCD.  The serial register in general was susceptible and
particularly the output amplifier.  If only the image area or the sides
without serial registers were illuminated the effect was not seen.

My conclusion was that charge detected in the relatively field free
areas surrounding these structures was diffusing into the serial
register and output amplifier over periods of 10-20 seconds.  This
accounts for why the effect extends into the overscan, and why you
wouldn't find anything wrong with your electronics to explain it.  The
amplitude of the effect was very similar to what you report, as
was the shape of the decay and the extent to which it entered the
image area.  Regions of interest showed similar symptoms since the
vertical transfers did nothing to sweep away the stray charge.

I peered down the microscope into the blackness of various
back-illuminated CCDs, and compared notes via email with John Barton
who looked at a more recent version of the same chip for me.  We
concluded that there appeared to be a mask missing from the periphery
of the older CCD.  I'd be grateful if there was someone reading this
from SITe who could confirm my interpretation.

Your situation is a bit different since the bias decay effect is not
present in the overscan of your 1 second exposure.  Does it occur in
the lab with the window securely covered?  Do you leave the serial
register static during exposure or do you flush it.  If the latter,
maybe your erase waveforms (between exposures) are messed up.  How does
a 1 sec dark compare to a zero second dark?   I presume the readout
amplifiers are on the lower edge of the image: can you read from a
different amplifier to see if the effect changes?  


Good luck. 

	Roger Smith,  CTIO