p0 Calibration at ECAL Test Beam

1
p0??? Calibration at ECAL Test Beam
M. Gataullin, V. Litvin, Y.Ma, Y. Yang Mar 22,
2007

• p0??? Reconstruction
• two production places
• Calibration
• L3 algorithm
• Estimate of p0 calibration precision
• comparison with TBS1 calibration

2
Experimental Set Up
3 different beam energy 20, 9, 50GeV. Pi0s
produced both at Al target and S1 ( next slide).

3
Reconstruction of p0???

• raw energy of each channel ADC x 35MeV/ADC
• Photon candidates 3x3 matrix made from Island
  basic clusters. Zero suppression of raw energy
  50MeV at crystal level.

A factor 1.424 is multiplied to raw energy. The
relative shift between two mass peak 22 tells
the second peak should come from about 1m
upstream of the beam line. This is the where S1
sit. The peak around 20MeV is because of
overlap between two photon candidates
Assuming pi0 produce at Al target ONLY.
4
Select pi0 according to S1,S2 ADC
No overlap between two photons is required, about
8 events ( mostly 50GeV beam data) lost. No
effect on calibration results
5
Final selected events for Calibration
Low Energy pi0 dominates, similar to MC selected
pi0 for calibration
6
Calibration Algorithm L3 method

• Events within M-2sigma, M2sigma is selected.
• Starting with no calibrations (C1).
• Update photon energy and direction after each
  step with new calibration constants
• Well tested in our MC calibration studies
• ideal calibration should lead to precision

7
Convergence of calibration constant
Relative difference ( upper limit of precision)
,compared to TBS1 constants. Blue dot, one
crystal inside the 9x8 matrix Red square, one
outside. We will estimate pi0 calibration
precision inside and outside the matrix.
8
Quick check on Independent samples
Only 75 of total samples are used for
calibration iteration. And the 25 left is used
to quickly check the calibration performance in
term of pi0 mass resolution, as well as comparing
with TBS1 calibration ( next slide)
9
Comparison with TBS1 calibration

• Good agreement with pi0 and TBS1 calibration. But
  it
• is difficult to estimate calibration precision
• Should divide into at lest two crystal regions
• how calibration precision transferred into pi0
  mass resolution is not exactly clear. So we will
  use electron data at H2.

10
Relative difference between C(pi0) and C(TBS1)
Inner 9x8 matrix
Outside this small matrix
We need two more equations to estimate pi0
calibration precision. Two more measurements
two electron peaks.. (next)
11
Electrons at H2 ECAL

• Clean cuts for electron beam (From Daniele and
  Sasha Ledovskoy )
• At least 1 hit in both WireChamber A and B.
  Exactly
• 1 hit in Chamber C.
• Track matched in A,B -3lt wcAx-wcBxlt4.5 mm

wt is the sqrt(Sum of EiEi)/E Measurement of
electron peak resolution using both pi0 and TBS1
calibrations.
12
(No Transcript)
13
Estimate of calibration precision