Csc Cluster Reconstruction in new simulation (13.0.2) Woochun Park - - PowerPoint PPT Presentation

Csc Cluster Reconstruction in new simulation (13.0.2)

Woochun Park University of South Carolina @Muon Reconstruction Meeting September 20, 2007

Introduction

• Gas gain in simulation is reduced from 1.0x105 to 0.58x105.
• Strip charges are reduced accordingly.
• In cluster reconstruction, threshold algorithm is used. We

need to make sure that all the parameters are reasonable for the new simulation.

Cluster Charge Distribution

QL+QP+QR 12.0.6 simulation

13.0.20 simulation

#Strips of Cluster Depending on Cluster Charge

12.0.6 thr=20k max=5

13.0.20 thr=20k max=5

Cluster Charge Distribution on #strips

• The more cluster charge distribution spreads, the more strips in cluster.

Black Line : 3 strips Blue Line : 4 strips Red Line : 5 strips QL+QP+QR 13.0.20 12.0.60

ClusterStatus Depending on Cluster Charge

12.0.6 thr=20k max=5

13.0.20 thr=20k max=5 1.2/0.1 1.0/0.1 6 4.1/0.1 5.9/0.2 4.7/0.1 0.4/0.0 1.5/0.1 82.2/0.6 13.0.20 2.6/0.1 7.9/0.2 1.0/0.1 0.1/0.0 1.8/0.1 85.6/0.7 12.0.6 11 5 4 3 2 %

Unspoiled Edge Multipeak Narrow Wide Skewed Inconst.

Pull Width Dependence

eta Measured Strip Position Unpoiled + Spoiled Clusters Unpoiled Clusters 13.0.20

Residual on Cluster Charge

Cluster Charge (ADC counts) 500 1000 1500 2000 2500 3000 3500 Clusters/(40 ADC counts) 50 100 150 200 250 300

|<0.15 φ 0.05<|tan

Resolution (mm) 0.05 0.10 0.15 0.20 0.25 0.30

12.0.6 thr=20k max=5

13.0.20 thr=20k max=5

max : max # strip in cluster Cosmic result QL+QP+QR

In cosmic data, spoiled clusters are included which make resolution worse at the high tail.

Calculated Error vs. Cluster Charge

• Error is smaller than the residual.
• Regardless of release, error estimates depending on cluster charges are similar.
• This can be shown in pull width dependence on cluster charge in next page.
• We may want to introduce additional error which is proportional to cluster charge.

12.0.6 thr=20k max=5

13.0.20 thr=20k max=5

Pull Width Dependence on Cluster Charge

max : max # strip in cluster QL+QP+QR 12.0.6 thr=20k max=5

13.0.20 thr=20k max=5

Good efficiency on Cluster Charge

12.0.6 thr=20k max=5 0.995±0.001

13.0.20 thr=20k max=5 0.986±0.001

Good effi 12.0.6 thr=20k max=5

13.0.20 thr=20k max=5

max : max # strip in cluster QL+QP+QR Efficiency 1.000 0.999

Good efficiency dependence on #strip and cluster charge

• If cluster has 4 or 5 strips, more chance to be bad

measurement.

13.0.20 thr=20k max=5

Black Line : 3 strips Blue Line : 4 strips Red Line : 5 strips QL+QP+QR

Good Efficiency and #Strips

• The cluster with more

charges may suffer from delta electron nearby.

Bad Efficiency on Cluster Charge

13.0.20 thr=20k max=5 0.014±0.001

12.0.6 thr=20k max=5 0.005±0.001 12.0.6 thr=20k max=5

13.0.20 thr=20k max=5

QL+QP+QR Mostly from 4 strips and 5 strips (previous page) Bad effi

Spoiled Fraction on Cluster Charge

13.0.20 thr=20k max=5 0.184±0.003

12.0.6 thr=20k max=5 0.144±0.003 12.0.6 thr=20k max=5

13.0.20 thr=20k max=5

QL+QP+QR

This behavior is understood in page 19.

Spoiled Fraction

Eta Resultion

• Position residual is correlated to strip charge amount. The more

charges are deposited, the better resolution.

• Threshold is 20k e-.
• max # strips per cluster = 5
• Compared to 41microns in 12.0.6 simulation, the resolution becomes

worse.

61.4µm

Eta Pull Distribution

• Threshold is 20k e-.
• max # strips per cluster = 5.
• Compared to 1.004 in 12.0.6 simulation, pull width is comparable as 1.035.

• ClusterStatus

– 0: Unspoiled – 2: Edge – 3: Multipeak – 4: Narrow – 5: Wide – 6: Skewed – 11: Inconsist.

ClusterStatus Depending on Cluster Charge

13.0.20 thr=20k max=5

The more cluster charge, the more failed

• wing to too wide.

# Strips Depending on Cluster Charge

13.0.20 thr=20k max=5

The more cluster charge, the more number strips cluster has.

Number of strips vs cluster charge.

Orange line indicates the current cuts [3-5] for # strips of cluster.

# Strips Depending on Cluster Charge

12.0.6 thr=20k max=5

The more cluster charge, the more number strips cluster has.

Number of strips vs cluster charge.

Orange line indicates the current cuts [3-5] for # strips of cluster.

Conclusion & To do

• Gas gain is reduced to 0.58x105 from 1.0x105 in Csc simulation (13.0.20).

– Cluster charges are decreased accordingly. – Cluster position resolution is 61.4 microns (compared to 41 microns before). – Pull distribution width is not changed. – Temporarily, we suggest to keep using the previous parameters until we decide them.

• No calibration is done yet.
• Strip charge error (4300e) and angular error term (0.57mm) are unchanged.
• Less Threshold is suggested but the previous one (20k) is okay.
• Bad cluster fraction is increased (0.5% 1.4%) and spoiled fraction is increased

(14.4% 18.4%)

– Method to discriminate bad clusters (residual > 5 sigma) needs to be developed using extra information such as total charge, 4th strip charge, skewness, number of strips, left-right inconsistency. – Cut needs to be decided on the plot of bad fraction vs. spoiled fraction.

• Maximum # of strips.
• Left-right significance.
• Threshold
• Cosmic data should be compared to the MC simulation

– MuonCalibration is ready to subtract pedestal (Ketevi) and it’s available 13.0.30.