Foward Ecal Reco
Updated fEcal in DD4Hep
Code is at
EICrecon GitHub FEMCupdate20250501 branch
Updated "FEMC" in EICRecon
Added switch between Homogeneous and ScFi geometry
Homogeneous : 11.333/sqrt(E) + 3% energy smearing
Homogeneous : Sampling Fraction = 100%
ScFi : Summing all fiber hits to a tower
ScFi : No energy smearing
ScFi : Sampling Fraction = 3%
Added SiPM Saturation in calorimeter DIGI code
src/algorithms/calorimetry/CalorimeterHitDigi.cc
2 new parameters totalPixel(default=0) and nPhotonPerGeV
By default (totalPixel=0) no attenuation will be applied
src/detectors/FEMC/FEMC.cc
nPhotonPerGeV = 1500
totalPixel = 4*159565
default for FEMC is saturation ON for both Homogeneous and ScFi
Use "-PFEMC:SiPMSaturation=OFF" option when running eicrecon to turn this off
There are 9 kinds of clusters exist in EICrecon/FEMC
IslandProto, CoG, SplitAndMerge
Truth (using MC info?) and not-Truth for each algo
Shape analysis added
It seems like CoG(Center of Gravity) is only one working as is among not-Truth one
Throwing 1 to 80GeV Gamma
Hit and cluster distribution for 2GeV
CoG cluster energy for single 1~80GeV gamma events
Homogeneous
Homogeneous
ScFi SF=3%
ScFi SF=2.9%
Smearing On
Smeasring On
Smeasring Off
2.2% Smeasring
SiPM Saturation OFF
SiPM Saturation ON
SiPM Saturation ON
SiPM Saturation ON
ADC Threshold=2
ADC Threshold=2
ADC Threshold=2
ADC Threshold=3
Energy linearity and resolution
Event Display with CoG cluster
Filled color represents energy of the tower.
Colored boundary lines shows the tower belongs to a cluster with the same boundary colored towers.
particle
20GeV
2GeV
2GeV with
EMaxCluser/Einput < 0.6
gamma
Homogeneous
,
Scfi
Homogeneous
,
Scfi
Homogeneous
,
Scfi
mu-
Homogeneous
,
Scfi
pi-
Homogeneous
,
Scfi
pi0
Homogeneous
,
Scfi
Escfi/Ehomo (both with SiPM saturation) gives 0.9681 at high energy end with SF=3.00% for ScFi.
SF is chanegd to 3% * 0.9681 = 2.9043%
Scfi gives ~2% constant term in energy resolution
Adding sqrt(3%^2 - 2%^2)=2.2% smearing (constant term only)
ADC threshold was 2
100[GeV] / 16384 [count in 14bits] = 6.1 [count/MeV]
This was based on 15[MeV] noise, or 15[MeV] / 6.1 [MeV/count] = 2.46 [counts]
ADC = llround(dE [MeV] / 6.1 [MeV/count]) where llround is nearest integer (and not floor).
ADC < threshold is zero suppressed (not <=)
Threshold = 2 is dE < 1.5 [count] * 6.1 [MeV/count] = 9.15 [MeV]
Threshold = 3 is dE < 2.5 [count] * 6.1 [MeV/count] = 15.25 [MeV]
Threshold is changed to 3
OGAWA, Akio
akio@bnl.gov