AliPhysics  3aa38c9 (3aa38c9)
The EMCAL trigger framework

Reconstruction of EMCAL trigger patches from FastOR / FEE data

The EMCAL Level1 trigger decision is based on reconstructed trigger patches. Events are selected once a trigger patch of a given type is found above threshold. Trigger patches are reconstructed using a sliding window.

Currently two Level1 triggers are implemented: A gamma trigger and a jet trigger. The setup of the gamma trigger is unique for all data sets while the setup of the jet trigger might differ for various data sets.

Gamma trigger Jet trigger
Subregion size: 1 Subregion size: 4
Number of subregions: 2 Number of subregions: 2 or 4
Total patch size: 2x2 Total patch size: 8x8 or 16x16

Trigger patches are reconstructed by the trigger maker wagon consisting of the task AliEmcalTriggerMakerTask. The trigger maker task configures and runs the trigger maker kernel for every event. The trigger maker kernel contains the trigger algorithms for gamma and jet trigger. Trigger algorithms are run for EMCAL and DCAL separately.

Configuration of the trigger maker

The trigger maker can be added to the train by using its add macro AddEmcalTriggerMakerNew.C

gROOT->LoadMacro(Form("%s/PWG/EMCAL/macros/AddTaskTriggerMakerNew.C", gSystem->ExpandPathName(gSystem->Getenv("ALICE_PHYSICS"))));
AliEmcalTriggerMakerTask *triggermaker = AddTaskTriggerMakerNew();
triggermaker->SelectCollisionCandidates(AliVEvent::kINT7 | AliVEvent::kEMC7 | AliVEvent::kEMCEJE | AliVEvent::kEMCEGA);

For existing datasets, identified by the run numbers in the dataset, the trigger maker can configure the trigger algorithms itself and no interaction has to be done by the user. Thus it is possible to provide a configuration by the user. This is particularly of relevance if a desired configuration for a future dataset needs to be tested before on existing data. The following configurations are implemented so far:

Configuration Data sets: Detectors Remarks
ConfigureForPP2011 pp, 2011 EMCAL Only level0 algorithm
ConfigureForPbPb2011 Pb-Pb, 2011 EMCAL L1 for gamma and jet trigger (16x16), only high-threshold online bits
ConfigureForPP2011 pp, 2012 EMCAL L1 for gamma and jet trigger (16x16), only high-threshold online bits
ConfigureForPPb2013 p-Pb, 2013 EMCAL L1 for gamma and jet trigger (16x16), high- and low-threshold online bits
ConfigureForPP2015 pp, 2015-2016, p-Pb 2016 EMCAL + DCAL L1 for gamma and jet trigger (16x16), high- and low-threshold online bits
ConfigureForPbPb2015 Pb-Pb 2015 EMCAL + DCAL L1 for gamma and jet trigger (8x8), only high-threshold online bits, background patches

More configurations need to implemented before running by the experts.

The following example configures the trigger maker for Pb-Pb data from 2015:

AliEmcalTriggerMakerTask *triggermaker = AddTaskTriggerMakerNew();

Masking of bad cells or fastors

The trigger maker reads in cell and fastor data from the input event. Handling of bad cells is typically done by the EMCAL tender or the EMCAL correction task. Thus the trigger maker provies the possibility to mask additional bad cells:

  • AddOfflineBadChannel (Short_t absId)
  • ReadOfflineBadChannelFromStream (std::istream &stream)
  • ReadOfflineBadChannelFromFile (const char *fname)

These are functions of the trigger maker kernel Cells with the given IDs are ignored when reading the cell data. The following example configures the trigger maker masking cel 12303:

AliEmcalTriggerMakerTask *triggermaker = AddTaskTriggerMakerNew();

FastORs can only be handled by the trigger maker. Also for fastors it is possible to define certain fastors directly

  • AddFastORBadChannel (Short_t absId)
  • ReadFastORBadChannelFromStream (std::istream &stream)
  • ReadFastORBadChannelFromFile (const char *fname)

For masked fastor channels the L1 ADC (L1 time sums) are ignored when reading the FastOR data, trigger bits however are still handled.

Apart from applying the masking manually, it is also possible to read the masking from OCDB or an OADB container. The masking in the OCDB corresponds to the masking which was applied at hardware level. It is of relevance in simulation for fee / offlinepatches in order to have the same acceptance map as it was applied online. The trigger maker will also take care about mapping issues for run1 and run2. In order to apply the masking from the OCDB, users have to do the following:

AliEmcalTriggerMakerTask *triggermaker = AddTaskTriggerMakerNew();

Note: The CDBconnect task is required when applying the FastOR masking from the OCDB.

A second option is to provide an OADB container with a masking. This option is focused on maskin fastors which turned out to be bad during offline QA and were not handled online. In this case the triggered sample is cleaned up from noisy triggers. The OADB container has to follow the definitions

  • The name of the OADB container has to be AliEmcalMaskedFastors
  • The content is a TObjArray with FastOR Absolute IDs, stored as TParameter<int>

The following code example masks FastORs 1211, 1560, 2300 for the run range 195000 to 200000:

AliOADBContainer fastorContainer("AliEmcalMaskedFastors");
TObjArray *fastorData = new TObjArray;
std::array<int, 3> masked = {1211, 1560, 2300};
for(auto m : masked) fastorData->Add(new TParameter<int>(Form("%d", m), m));
fastorContainer.AppendObject(fastorData, 195000, 200000);

The new container needs to be specified when adding the trigger maker:

AliEmcalTriggerMakerTask *triggermaker = AddTaskTriggerMakerNew();

In case users want to apply the FastOR masking also to cell data, this can be specified via SetApplyTRUMaskingToFEE();

Important notice for simulations: In order to have a realistic simulation of the acceptance, the only bad channel map on fastor level should be used, not the one on cell level. As tender or correction task will handle bad cells, the offline bad cell map will be applied. Therefor the trigger maker should run before the tender / correction task in case of simulations.

Selection of trigger patches by the user

Trigger patches are attached to the input event as new list object of type TClonesArray. The list object has the name EmcalTriggers. The AliAnalysisTaskEmcal can import the object automatically once the list object is specified in the constructor:


The following example code selects events with EMCAL gamma trigger patches with a minimum FEE energy of 12 GeV:

TClonesArray *triggerpatches = fInputEvent->FindListObject("EmcalTriggers");
bool triggered = false;
for(auto p : *triggerpatches){
AliEMCALTriggerPatchInfo *recpatch = static_cast<AliEMCALTriggerPatchInfo *>(p);
if(!recpatch->IsEMCAL()) continue; // patch on EMCAL site
if(!recpatch->IsGammaLowSimple()) continue; // patch is a gamma (2x2) patch, reconstructed from cells
if(recpatch->GetPatchE() > 12.){
triggered = true;

Trigger bit setup

Trigger bits are used to store online trigger selection by the STU as well as offline trigger selection. Trigger patches can be offline or recalc patches.

Online Offline Simple Recalc
Patches firing the
trigger in the STU
Recalculated patch
based on FEE energy
Recalculated patch
based on FastOR energy

Accoring to the new definition, where there exists only one trigger algorithm per L1 trigger type, trigger patches are always offline simple and recalc patches at the same time and store both online ADC values and offline calibrated FEE energy. Recalc patches with an ADC value above online threshold applied for the dataset must fullfil per definiton the trigger requirement in the STU and should therefor be online patches as well.

A certain bit number is assigned to each patch type in a bitmap in the AliEMCALTriggerPatchInfo object. Ranges are reserved for certain patch types:

  • Bit 0 - Bit 15 for online patches
  • Bit 16 - Bit 23 for recalc patches
  • Bit 24 - Bit 32 for offline patches

Note that for online bits the bits are shifted in case of data by the amount of online trigger types. The order of the bits is defined by the configuration used in the reconstruction.

3-bit configuration 5-bit configuration
0) Level0 0) Level0
1) Gamma 1) Gamma-high
2) Jet 2) Gamma-low
3) Jet-high
4) Jet-low

The access to the information is generalized by the functions IsLevel0(), IsJetHigh(), IsJetLow(), IsGammaHigh(), IsGammaLow(). Users should not access the trigger bits directly but use the helper functions. In case of a reconstruction with the 3-bit (referred to as old) configuration jet patches will be both jet-high and jet-low patches at the same time, similarly for gamma patches. In case of the 5-bit (referred to as new) configuration it is mandatory to distinguish between high and low threshold.

Offline and recalc bits are set by the trigger maker and independent of the software version used in the reconstruction. The order of the bits follows the 5-bit configuration. Also for offline and recalc patches general getter functions are available making the check for a patch type easy for the user (IsJetLowSimple(), IsJetHighSimple(), IsGammaLowSimple(), IsGammaHighSimple() in case of offline simple patches and IsJetLowRecalc(), IsJetHighRecalc(), IsGammaLowRecalc() and IsGammaHighRecalc() in case of recalc patches). For both offline simple and recalc patches the set of high-threshold patches is automatically a subset of the set of low-threshold patches as the trigger decision to a patch is applied in case the patch ADC is above threshold. Thresholds can be set via

AliEmcalTriggerMakerTask *trgtask = new AliEmcalTriggerMakerTask("NewTriggerMaker", kTRUE);
trgtask->SetTriggerThresholdJetLow(0, 0, 120); // V0-centrality independent threshold at 120 ADC counts

In case the thresholds are set to 0, no trigger decision to patches is applied, and the output contains all possible gamma (2x2) and jet (8x8 or 16x16) patches. In this case the user is responsible for the trigger selection.