This page outlines the development of the W-Algorithm for the run9 500 GeV data taking.

Last updated on 12/08/2008

Version - v1

Blue - Completed

Red - Outdated or Replaced

White - To do

• Changes to Simulation Package
  1. Initial Simple Changes for BSMD Studies
     • Change BFC to use sim and Wbose tables for pedestal/gain/calibration tables
     • Change StEmcSimulator to use constant sampling fraction (across eta) for all layers of BEMC
     • Change StEmcSimulator so that gain does not saturate at 1023 for BPRS and BSMD (now 4095 like for BTOW)
  2. Realistic Simulation and MC/Data Matching
     
     • Put back in realistic pedestals/pedestal widths

• Simulations to Run for Algorithm Vetting
  
  1. 1k Level
     • Single Particle Samples
       • ET = {10,20,40}, [-1.2,1.2] in eta, 2pi in phi
       • electrons, positrons, mu+,mu-,pi0
  2. 10k Level
     • Single Particle Samples
     • W +- Event Samples
     • QCD Background Event Samples in pt-hat bins [5-10]
  3. 100k Level
  4. Full Statistics Set (one for each event to be taken x some factor)

• W-Algorithm Development
  1. Information Gathering
     • BTOW Information
       • Find highest ET 2x2 cluster
       • Use simple gain formula to calculate ET
       • Use linear weighted CG for shower position in eta/phi
       • Use logarithmic weighed CG for shower position in eta/phi (article on logarithmic weighting) The formula used for a position in either direction is X = sum_i {x_i * w_i} / sum_i {w_i} where w_i = max(0,w0+ln(E_i/E_T)) and I have (for no really good reason) set w0 = 3.0
     • BPRS Information
       • Find ADC sum for BPRS in same 2x2 tiles
     • BSMD Information
       • Find BSMDE [-3 strips,+3 strips] ADC sum around central strip
       • Find BSMDP [-3 strips,+3 strips] ADC sum around central strip
       • Position in eta/phi found by logarithmic weighting in [-3,+3] sums
     • TPC Information
       
       • Find highest pT track near BTOW cluster
         
     • Vertex
       • New Vertex Algorithm to keep vertices with 1 track pointing to them
  2. Particle Level Cuts
     • Energy Cuts
       • ET of highest ET 2x2 cluster from BTOW > 20 GeV
       • pT of highest pT track from TPC > 15 GeV/c
     • Shower Shape Cuts
       • BPRS Energy in same 2x2 Cluster < (insert number) GeV
       • BSMDE(and P) Energy in [-3,+3] / [-N,+N] < (insert number)
       • BSMDE(and P) Energy Dispersion (sum{i=strips} En_i * |strip # - center strip #|)/(sum{i=strips} En_i) < (insert number) ... maybe second moment too...
     • Matching Cuts
       • Highest PT track and highest ET Cluster position (at BSMD depth in BEMC) must be within +- (insert number) eta and +- (insert number) phi
       • Highest PT track and highest ET Cluster energies must be within +- (insert number) GeV of one another
  3. Event Level Cut
     • Trigger Cut
     • Vertex Cut
     • Topological Cuts
       
       • Missing ET cut
       • No forward jets in EEMC cut (or just an < some energy cut)

• Algorithm Vetting
  1. Energy/Position Reconstruction Comparison
  2. e/h Separation vs. ET
  3. QCD Background Supression