| 3 | | All done on chip XY21 and cell xy53 (chosen somewhat arbitrarily). |
| | 3 | The edge rolloff was checked for a variety of science and flat field input exposures selected with observation dates after 2009-12-09. The rolloff was checked on all sides of cell xy53 on chip XY21. The main points are: |
| | 4 | |
| | 5 | * Although all sides appear to have some degree of rolloff in the raw exposures, the flat fielding corrects the effect for the right and left sides. This is consistent with the magic results, in which the new streaks were oriented horizontally. |
| | 6 | * The level of the rolloff scales with the background level, as can be seen in the clear ordering in the raw image profiles. This is true for both the science and flat input images. |
| | 7 | * The rolloff on the top is generally well corrected. Two science exposures have upturns in the flat corrected profile. One has a star in the profile box (exp_name = o5180g0252o), and the other has a substantially higher background level than the others (exp_name = o5192g0138o). PNG images of these cells are at the bottom. The flat input exposure with an upturn (exp_name = o5195g0002f) does not have a significantly different background that the other flat exposures, but does show an odd defect towards the top of the cell. |
| | 8 | * The rolloff at the bottom is the likely cause of the majority of the magic detections. The science exposures have an upturn in the bottom few rows that does not appear in the flat field exposures. This upturn carries through to the flat corrected profiles, causing a 5-10% spike for these rows. |
