PHOS calibration objects (see DataDormats/Detectors/PHOS/)

The following objects are used for PHOS calibration

Pedestals

Calculated during dedicate runs with and used for FEE configuration in later physics runs

CalibParams

Include channel-by channel gain parameters, High Gain/Low Gain ratio and time offsets for High Gain and Low Gain channels separately.

TriggerMap

Object contains trigger bad map and 14 parameterizations of L0 turn-on curves (one for each DDL)

Runbyrun

Time-dependent energy scale corrections, one number per module (plus uncertainty of the fitting procedure)

BadChannelsMap

Bad channels map used in reconstruction and simulation. Caclulated by combining inputs from pedestal, noisy, LED and physics runs and DCS configuration

Calibration procedures

Pedestal calculation

Dedicated pedestal runs (trigger rate ~10 Hz) and length ~100 s to collect ~1000 events. Analyzed with class PHOSPedestalCalibDevice which uses as input vector of o2::phos::Cell, produced with raw to cell converted with option –pedestal. In this conficuration Cell::Energy contains mean and Cell::Time contains RMS of pedestal.

 cpp
o2-raw-file-reader-workflow --input-conf PHSraw.cfg |
o2-phos-reco-workflow --input-type raw --output-type cells --disable-root-output --pedestal on |
o2-phos-calib-workflow --pedestals

High Gain/Low Gain ratio calculation

Dedicated LED runs (trigger rate ~1000 Hz) and length ~100 s to collect 10^5 events. Analyzed with class PHOSHGLGRatioCalibDevice which uses output of standard reconstruction but wich switched off merging of HG and LG channels:

 cpp
o2-raw-file-reader-workflow --input-conf PHSraw.cfg |
o2-phos-reco-workflow --input-type raw --output-type cells --disable-root-output --keepHGLG on |
o2-phos-calib-workflow --hglgratio

Turn-on curve and trigger bad map calculation

Run over physics run, selecting events not marked as PHOS L0 trigger. Performed with classes PHOSTurnonCalibDevice and PHOSTurnonCalibrator. Stores transient object to CCDB if sufficient statistics not collected yet or calculates TriggerMap objects and stores it at CCDB. Uses list of Cells and FullClusters as input. Running on MC digits

 cpp
o2-phos-reco-workflow --input-type digits --output-type cells --disable-root-output |
o2-phos-reco-workflow --input-type cells --output-type clusters --fullclu-output --disable-root-output --disable-root-input |
o2-phos-calib-workflow  --turnon

Run-by-run correction

Calculated pi0 peak position in each module in current run (or equivavlt period). Use list of FullClusters as input and produce 8 numbers: peak position and fit uncertainties for each module. Run as

 cpp
o2-phos-reco-workflow --input-type digits --output-type cells --disable-root-output |
o2-phos-reco-workflow --input-type cells --output-type clusters --fullclu-output --disable-root-output --disable-root-input |
o2-phos-calib-workflow  --runbyrun

Energy and time calibration

Relative energy scale per channel and time offsets for High Gain and Low Gain channels is calculated with iterative procedure. At the first step physics (better PHOS triggered) data are scanned and list of digits contributed to PHOS clusters are stored in local root files on EPN. digits are packed as std::vector<uint32_t> and simutaneously a list of inv. mass and time histograms are filled. Then new calibration coefficients are calculated from collected histograms and list of clusters re-scanned with new calibration. Few (~5) iterations is necessary to reach final accuracy. First scan is ran as

 cpp
o2-phos-reco-workflow --input-type digits --output-type cells --disable-root-output |
o2-phos-reco-workflow --input-type cells --output-type clusters --fullclu-output --disable-root-output --disable-root-input |
o2-phos-calib-workflow --not-use-ccdb --energy

Bad map calculation

Several methods are used to produce bad map. Method-based maps are stored in PHS/BadMap/Occ for cell-occupancy, PHS/BadMap/Ped for pedestal RMS based bad map and PHS/BadMap/Chi for chi2 of sample fit. Then bad map is calculated by combining different approaches and stored in PHS/Calib/BadMap. The latter is used in clusterization, e.g. in AOD analyses.