Pileup%20study

1
Pileup study

• R.Sawada
• 10/May/2007
• MEG-J Xenon analysis meeting

2
Simulation condition

• Mixture
• RD (gt 52.8/2 MeV)
• RD (gt 528 keV)
• Michel (gt 30 MeV)
• Mu stop 1e8
• Trigger simulation
• Pre-selection with energy deposit gt 30 MeV
• Xe sum waveform threshold 3 V

3
Problems of this study - 1
Kinematics distribution of RD event generation

• There was a bug in event generator in gem. There
  was a cut for positron emission angle.
• This problem was found by Ootani-san and it is
  fixed.
• But I didnt have time to recreate simulation.
• Study written in this slide is done by biased
  kinematics.

4
Problems of this study - 1I

• Time width for charge integration of PMT may be
  too short. It was 150 nsec from the beginning of
  a pulse.

5
Problems of this study - III
Qsum2 (150 nsec)

• Low energy gamma rays are triggered because
  voltage threshold was rather low.

Energy deposit in DRS time window (0-500 nsec)
6
Modification from the previous study

• It was found that Sub-events before DRS start
  were not mixed at the previous study. Due to
  that, number of pileups are underestimated.
• This problem is fixed in this study.

7
Number of gammas
72
0.04
24
4
0.3
All triggered events
8
Energy deposit
7055 events without cut 2688 single events (MC
truth)
No cut
Single gamma event
Pileup event
Pileup rejection with time and space distribution
9
Energy deposit by each gamma
Pileup events Energy sum of all gammas is in
signal region (48-54 MeV)
Probability of pile up with very high and very
low is large. But (0.5 0.5) is not negligible.
10
Similar plots by using qsum2
No cut
Single gamma event
Pileup event
Pileup rejection with time and space distribution
11
Pileup ID efficiency
N number of gamma (MC truth) ID Pileup ID
ID failed
ID failed
OK
OK
OK
OK
Miss ID
Miss ID
Time distribution
Time space distribution
12
CPU time / Disk
13

• Similar study will be done by modified softwares
• RD kinematics
• Xe trigger threshold
• Time width for charge integration
• Investigation of multiple gamma reconstruction
• Waveform analysis for each gamma
• Full waveform can not be contained in DRS window
  when timely well separated pileups.
• For doing multi gamma analysis, re-binning
  waveform tail part in online is not good. (Peak
  search for small pileup in online will not be
  easy)
• Rate of high low pileup is bigger than mid
  mid.
• Strategy to evaluate pileup probability for each
  events.