WFPC2 WF4 Bias Level Anomaly
10/25/2005
We have identified a serious anomaly in images from the WF4 CCD
in WFPC2. The WF4 CCD bias level appears to have become unstable,
and we are seeing sporadic images with either low or zero bias
level. The severity and frequency of the problem is
rapidly increasing, and it is possible that WF4 will soon become
unusable if no work-around is found. The other three CCDs
(PC1, WF2, and WF3) appear to be unaffected and continue to
operate properly.
The impacts from "low" and "zero" bias are somewhat different,
but in both cases the effects are immediately obvious.
Images with low bias will tend to have horizontal
(x-direction) streaks and stripes with an amplitude
of ± about 0.5 DN in WF4. We believe these data should be mostly
recoverable with some effort, though at a loss in the detectability
of faint targets. "Zero bias" is a much more serious problem and
is evidenced by images which are blank in WF4, except for showing
occasional cosmic rays, bright targets, and negative pixels
from dark subtraction. These images with zero bias are probably
unusable for most purposes. Both the CCD gain settings of 7 and 14
are affected.
The frequency of the anomaly is rapidly increasing. The first
significant instances of low bias appear to have been in late 2004
when a few images were impacted. However, within the last few weeks
over half the images are beginning to show the low bias problem.
The more serious "zero bias" problem appears to have first occurred in
Feb. 2005, but it is also increasing and now impacts 10% to 20%
of WFPC2 images. At present there are still many images which appear
fine and unaffected, but the situation is quickly evolving.
We believe the science impact for most observers will be minimal.
Targets are by default placed on either PC1 or WF3 which
continue to operate properly. However, observers requiring the
full field of view (survey projects, large targets, etc.) will
potentially lose one-third of their imaging area.
Our understanding of this anomaly is still evolving, and most of the
information is tentative. Additional details will be posted on the
WFPC2 website as they become available.
|