Known issues with MC2008

1. For reading 2008 MC data together with correct geometry and conditions data base

• Using DaVinci configurable, v21r0 or higher

  •  DaVinci().DataType     = "2008"

  •  DaVinci().Simulation   = True  

• Otherwise, e.g. GaudiPython

  • from Configurables import DDDBConf

  • dddb = DDDBConf()

  • dddb.DataType   = "2008"

  • dddb.Simulation = True

• Packed DST, third round of test production, requires EventPacker v1r2 to read correctly, in DaVinci v22r0p2, Brunel v34r1p1, Panoramix v16r3 ?

2. Problems

• First round of production, Gauss v35r0, Boole v16r3, Brunel v33r3p1
  • Velo module misalignment parameters not used in Gauss, but used in Boole and Brunel leading to kind of velo misalignment effects. FIXED in Gauss v36r0
  •  Too high Ecal noise reflecting 2008 conditions. FIXED in Boole v17r0 to reflect improved 2009 performance
  • Ecal calibration, OPEN, expected to explain
    • lower muon ID using global likelihood
    • electron reconstruction, J/psi -> ee
  • Rich PID, higher kaon misid
    • bug in Rich reco, aerogel varying refractive index not taken into account, understood, FIXED in next Brunel, some more tuning needed, due to more realistic background, performance worse than DC06, maybe due to Velo problem
  • MC association inefficiency (Wenbin): CLOSED
    • problem of combined association ?
      Was a problem of analyzing private root ntuple.
  • Velo digitization: noise similar to DC06, but larger signal, thresholds not changed, therefore large increase in spill over. New thresholds are being prepared. CLOSED ?
    Detailed presentation at T-Rec 15 December . Procedure used:
    Use the best available data on the expected energy deposit in eV from GEANT4, MPVs from the ACDC3 testbeam, the linearity slopes from JC, the noise from the Liverpool construction and the time curves from Kazu and make the simulation as realistic as possible.
  • OT: some issue with position of stereo layers discovered in the alignment procedure with cosmics/Ted data, OPEN
• Private reprocessing of Bs2JpsiPhi with Boole v17r0 and Brunel v34r0
  •  10% loss in reconstruction / acceptance ? OPEN
  • Kaons from Phi inefficient, muons ok. Does not explain fully above 10%

3. MC verification