The CRIRES detector arrays become non-linear for exposure levels above about 4000 ADUs. This behaviour is assessed by a series of dark and flat exposures with increasing DIT which maps in the flats the gain behaviour from the linear to the non-linear regime. After dark subtraction, the function

ADU = A + B * DIT + C * DIT^2

is fitted for each pixel of the flat sequence. The coefficients A, B, and C are saved into output images. The deviations from the fit are saved in the Q image. Above about 10000 ADUs, a quadratic fit is not sufficient any more. Such exposure levels should, therefore, be avoided.

The application of the coefficient images to correct science frames is described on the science recipe pages.

Example fit. A fit of the increasing flux for a single pixel (x=511, y=256). The measured values are plotted as blue circles. The red line is a linear fit to the first two exposures; the green line a quadratic function with A = 29, B = 787, and C = -2.5. Both fits begin to differ above about 4000 ADU.

Example frame for A coefficients. Enlargement of the centre region of an image of the A coefficients.

Example frame for B coefficients. Enlargement of the centre region of an image of the B coefficients. These coefficients show the pixel-to-pixel gain variation in the linear regime which can be corrected by flat-fielding.

Example frame for C coefficients. Enlargement of the centre region of an image of the C coefficients. These are the remaining pixel-to-pixel gain variations in the non-linear regime.

QC1 parameters

QC1 parameters for CRIRES flat-field exposures are measured for each detector array separately. Since pipeline version 1.5, the calculation of the QC1 parameters is restricted to the illuminated area.

Trending

Dedicated trending plots for DETLIN parameters do currently not exist.