• CalorimeterCalibration

ILD Calorimeter Calibration

There is a Marlin calibration processor from V. Hangartner, who used to be summer student at DESY in 2007. The basic idea is valid energy conservation. If you have, e.g. e+e- -> ttbar -> 6 jets at sqrt(s)=500GeV, you can expect a maximum of 500GeV energy deposited in both, the ECAL and the HCAL. Therefore, a simple calibration can be performed by fitting

• E_CM = c₁ * E_vis1(ECAL) + c₂ * E_vis2(ECAL) + c₃ * E_vis3(HCAL)

The tarball includes the Marlin code, as well as a documentation and some root and gnuplot scripts.

The CalibProcessor.cc will plot E_vis1(ECAL) + E_vis2(ECAL) + E_vis3(HCAL) versus E_vis3(HCAL) - [ E_vis1(ECAL) + E_vis2(ECAL) ] in 8 bins. The bin with the maximum number of entries is taken as total energy sum. In principle this is not the best, but the easiest solution. A more appropriate way is implemeted in the root script, which is also included in the tarball. Here the total energy sum is fitted several times with a gaus-function to extract the mean value. The root script can be run with the root tree that is written out from the marlin processor.

$ root output.root
$ .x executeScriptSlice.C

In addition you can have a look at the result of your calibration with a gnuplot script:

$ . gnuplot_CalibProcessor_output.sh CalibProcessor_data_histo

Here is one example of the reulting plots from LDC01_06Sc_test with the calibration constants displayed below:

CalibProcessor_data_histo_contour.eps

In a modified version of the CalibProcessor.cc a gaussian fit is performed in the Marlin processor to extract the mean energy.

Be aware that the input calibration constants (your first guess) might influence the output. So make sure you chose at least some reasonable ratio between the three sampling coefficients.

For the moment I have obtained following calibration constants with this processor: