Calibration Access and Data Handbook

Procedure

In order to preserve the calibrational accuracy of the CTI correction, it should be performed before event recombination. This is due to the way the current values of the CTI correction were derived from analysis of ground calibration data.

This model of CTI correction assumes pulse-height proportional charge loss as a function of distance from the readout node.

The CTI correction of the parallel shifts is carried out first, followed by the CTI correction of the serial register.

The CTI correction for the parallel transfers has a dependency on the column number (DETX), due to lower CTE closer to the sides of the CCD's. The CTI is performed according to

$$PI_{\rm corr}' = PI \times [ 1 + y \times YCTI(x) ]$$

where $x$ and $y$ are the distances of the location of the event from the respective readout node in CCD pixels, starting from $(x,y)=(1,1)$.

The CTI correction for the serial register is

$$PI_{\rm corr} = PI_{\rm corr}' \times [ 1 + x \times XCTI ]$$

with the same convention as before.

$XCTI$ is stored per readout node in CCF CTI, and YCTI is stored per column (DETX-coordinate).

No temperature dependence will be included, however consistency will be checked between the current temperature and the temperature for which CCF CTI is valid. If the temperatures are inconsistent, a warning will be issued.

Note, that in due time the history of the events (column wise) may have to be included: e.g. to apply the CTI correction only to the first pixel of an multi-pixel event in a certain column and readout.