Selected Results from the GPC1 Bad Pixel Mask creation

DARKMASK 19 was created from a set of six 100-second darks taken on 20070913. The ccd temperature for these darks was around -85. The image below shows chip 67 cell 43, which contains a particularly prominent feature.

DARKMASK19.XY67.xy43.png

DARKMASK 22 was created using a different set of darks from the same night, when the ccd temperature was around -95. The image below shows the same cell.

DARKMASK22.XY67.xy43.png

Taking the difference of these two images (diff = DM22 - DM19) we see a ring of negative pixels, indicating that the bad pixel area is larger in the warmer DARKMASK 19:

DIFF.DM22-DM19.png

One might imagine that the temperature change over the sequence of six images could be more dramatic for the DM19 set at -85 than the DM22 set at -95. A larger dispersion in the input image set would cause more pixels to be flagged as having an "uncorrectable" dark current. However, the temperature change in each set of six images is the same: about 2 degrees.

This case suggests the need for a bad pixel mask that is built from darks that span the range of expected GPC1 operating temperatures. We supposedly eliminated the temperature dependency by selecting a limited set of input darks, but we still see temperature-dependent differences in the bad pixel area.