9.5.1 The Extra Correction

When all the science images (from night 1, 2, 3, 4, 6, & 9) were reduced using the fat correction mentioned above, it turned out, that it did not remove the fat zero stripes for the images from night 1-4. The fat zero effect had changed with the CCD temperature change between night 5 and 6, and since the images used to determine the fat zero effect were from night 8 and 13 (i.e. both after the temperature change), the established correction worked only for night 6-14.

Since we did not have the same type of calibration images taken before the temperature change, the extra correction needed to remove the fat zero effect from the night 1-5 images needed to be modeled from the science images.

The approach was the following: For all the science images from night 1-4: for all the 20 fat zero columns, and for the 6 background areas, the level (one number) representing the sky background (not the objects, such as galaxies and stars, which happens to be in that column somewhere) was calculated. This was done by a 3 step process: First guess, N₁, was midpt from imstat using no cuts. Second guess, N₂, was midpt using cuts $\pm 5\sigma$, with $\sigma$ being the theoretical standard deviation corresponding to that level, $\sigma = (N/\mbox{CF} + \mbox{RON}^2_{\mbox{\scriptsize ADU}})^{1/2}$, with N = N₁ in this step. Third and final guess, N₃, was midpt using cuts $\pm 3\sigma$, with $\sigma$ being calculated using N = N₂. When having the level in the background and in the fat zero column, the difference can be calculated. This process was implemented in the cl-script extra_fat.

An example of the extra fat effect can be seen in Figure .

 

The data were fit by eye with a 3 piece linear function (example of fit also shown in Figure ). 13 out of the 20 fat column had a significant (> 2 ADU) extra effect. b_fatzero.cl was modified to be able to make the extra correction.