Pre-embedding QA for W analysis II

Motivation:  Verify compatibility of MC signal W events with real background events (zerobias) before embedding begins. 

Things to check for TPC (just check low HV settings for now):

• dE/dx
• chi^2 / dof

Data Samples

• MC : Pythia QCD sample with partonic pT > 20 GeV.  Using new TpcRS simulator in CVS.  Jobs run in SL09e Nov. 19th. bfc.C(100,"MakeEvent,ITTF,NoSsdIt,NoSvtIt,Idst,VFPPVnoCTB,logger,-EventQA,-dstout,tags,Tree,EvOut,analysis,dEdxY2,IdTruth,useInTracker,-hitfilt,tpcDB,TpcHitMover,TpxClu,McAna,fzin,y2009,tpcrs,sdt20090410.060000,geant,geantout,beamLine,eemcDb,McEvOut,bigbig,emcY2,EEfs,bbcSim,ctf,CMuDST","/star/data05/scratch/balewski/TpcRS-rev2/test_QCDprod.fzd") 

          TimeStamp used, sdt20090410.60000, is during run R10100015.  Kumac and Logfile used to generate .fzd file are attached below with ckin3=20.

• Data: SL09b production of L2W stream events from Run 9 pp500 R10100015.  (This includes preliminary TPC calibration)

  bfc.C(200,"DbV20090817 pp2009a ITTF BEmcChkStat QAalltrigs btofDat
  Corr3 OSpaceZ2 OGridLeak3D beamLine -VFMinuit VFPPVnoCTB -
  dstout","st_W_10087028_raw_4180001.daq")

Track Selection:

• Select vertices with rank > 0 and |zVertex| < 100
• Look at global component of primary tracks associated with each vertex that passed the above condition
• Require global track flag == 101, pT > 0.4 GeV and nHitsFit/nHitsPoss >= .51

For reference Jan did a similar study in June with the old TPC simulator.

Note:  For all 2D plots data is on the left and MC is on the right.

Figure 1:  Chi2 / dof distribution for Data and MC

Figure 2:  dE/dx vs p for Data (left) and MC (right)

Figure 3:  dE/dx for multilple pT "slices"  for Data (black) and MC (red)

Figure 4: 1/pT vs chi2/dof for Data (left) and MC (right)

Figure 5:  nHits vs chi2/dof for Data (left) and MC (right)

Figure 6:  Last Hit on track phi vs eta for Data (left) and MC (right)

Conclusions: 

1. chi2/dof is ~40% larger in the data than the MC -> Clusters are better aligned in MC than Data
2. dE/dx is at least factor of 2 smaller in the MC than the data