As noted in Section 2.1
, the post-SM4 WFC electronics have the property that differences in CCD readout timing can result in a significant difference in bias structure. This is observed between the WFC full-frame readout and the various WFC subarray readouts. It is also the case that the profile of CTE trailing is markedly different with differences in readout timing, particularly in the dwell time between parallel shifts. Lastly, the pre-Cycle 24 subarray modes were reading out all 4096 columns of the CCD, though retaining only a small portion of them. Because of this, some readout overheads were actually larger than the full-frame readout overhead.
The full 2048 columns are retained for all subarrays, so that the scene-dependent bias shift (ACS ISR 2012-02
) can also be corrected exactly as with full-frame readout. Because the bias striping effect (ACS ISR 2011-05
) requires more than one quadrant's prescan for precise measurement, WFC subarrays must still be de-striped outside of CALACS (see Section 5.2.6
), using the science region rather than the prescan region. As the subarray readouts all involve a single amplifier, there is no need for a crosstalk correction as done with the full-frame (four-amplifier) readout.