Analysis on TB0904 Data

Analysis on TB0904 Data

Jianchun Wang

Syracuse University

09/08/09 Jianchun Wang 2

The Telescope

PixelVELO

Pixel

YX YX

120 GeV proton beam

Pixel

Y

Scint

RR(

X

Z

Y

Pixel & VELO use independent DAQ systems, sharing trigger signals from the same source.

VELO events and pixel events are matched offline. Pixel stations / modules are aligned within its own system and generate tracks. Velo sensors are aligned with respect to the pixel tracks. Pixel track event, corresponding Velo event ID and Velo alignment parameters are

saved in tracking data files. Pixel tracking data are fed to Vetra. Real VELO studies of R/ and irradiated RR sensors.

Just

a re

mind

er


Introduction

Track is reconstructed from pixel hits and fit to straight line. Pixel hit resolutions. Track fit probability issue. Track projection error and multiple scattering effect. Alignment precision (not estimated yet).

Analysis of R/data. Cluster size vs pitch and track angles. Eta curve of charge sharing. Velo R resolution.

Analysis of irradiated RR sensors ( ongoing, not included).


Residual On the 5th Station

Measurement – Track Projection (mm)

Num

ber

of E

ntrie

s (A

rb.

Uni

t)

Ncol > 1

Nrow > 1

Ncol = 1

Nrow = 1

DifferentScale

Resolution (m)

ResidualRemove

track

Ncol > 1 7.6 5.8

Ncol = 1 120.0 119.8

Nrow > 1 8.3 6.6

Nrow = 1 12.7 11.7

Binary readout5 pixel stations

Simulated through iterations track proj. error ~ 4.9 m


Track Probability Issue

Simulation

parameters

N Row N Col

1 >1 1 >1

Probability 0.759 0.241 0.978 0.022

Resol (m) 11.7 6.6 119.8 5.8

Type Z (mm) X (10-3 X0)

X-pixel -450 9.5

Y-pixel -444 9.5

VELO 0 6.4

Y-pixel 317 12.5

X-pixel 514 9.5

Y-pixel 520 9.5

Non-gaussian

Prob (2, ndof)

Tra

cks

(arb

. U

nit)

Exclude Ncol = 1

With multiple scattering

Prob (2, ndof)

Tra

cks

(arb

. U

nit)

Expect

Seen Uniform dist for Ncol=1Gaussian for the rest


Tracking Error

Multiple scatt.

Include Ncol=1

Residual (m)

x y

5 station

s

No No 7.47 6.32

No Yes 7.42 6.30

Yes No 7.73 6.53

Yes Yes 7.69 6.51

Multi-ccatt. only 2.07 1.74

4 station

s

Yes No 7.71 7.72

Yes Yes 7.68 7.68

Multi-scatt. only 1.85 1.85

5 pixel stations

Tracking Error from Pixel (m)

Y

X

Log

( nu

mbe

r of

tra

cks

)

Calculated without multiple scattering

Multiple scattering contributes 1.7-2.1m to track projection error.

One can select events of better tracking error.

Measurements of Ncol=1 improve track projection precision, although distort the track probability distribution.


Look at R/Data

X ( mm)Y

( m

m)

Effective Track Angle (degree)

Signal (ADC)

Matched Hits

We took data at nominal 0, 4, 8, 12 degrees rotated around horizontal axis.

The effective angle is smaller due to concentric strips.

Pixel coverage


Effective Track Angle (Degree)

Per

cent

age

of H

itsCharge Sharing (I)

Cluster Size

All pitches & track angle

Seed threshold = 6 ADC ~ 9.6 Ke

Side threshold = 3 ADC ~ 4.8 Ke

Strip pitch (40, 50) mNstrip = 1 Nstrip = 2

Nstrip = 3

R sensor of R/pair

Range: angle0.5


Charge Sharing (II)

Pitch ( m)

40 – 5050 – 6060 – 7070 – 8080 – 90

90 – 100


(Nst

rip >

1)

/ N

tota

l (%

)

R/ data is split into 1 of angle & 10 m of pitch sub-samples.

Sub-samples of 0, 3, 7 and 11 are with reasonable large statistics.

Strip Pitch (m)(N

strip

> 1

) /

N to

tal (

%)

Angle ( )-0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5


The Eta Curve

Track Hit Fraction

Center of Strip N Center of Strip N+1

Only Strip N has Charge

Clu

ster

Fra

ctio

n

Only Strip N+1 has Charge

( )

i

Cluster Fraction

ADC i N

ADC

One strip shift due to tracking precision

All pitches & angles

Nstrip = 1 removed


The Eta Curves Of Small Pitches

Clu

ste

r F

ract

ion

Track Hit Fraction

Angle=0 Angle=3

Angle=7 Angle=11

Pitch = (40-50) m

Nstrip = 1 removed


The Eta Curves Of Small Pitches

Clu

ste

r F

ract

ion

Track Hit Fraction

Angle=0 Angle=3

Angle=7 Angle=11

Pitch = (40-50) m

Cluster fraction=0 or1 correspond to nstrip=1, indicating how charge sharing varies with hit position.


Velo Resolution Measurement

<Resid = 19.2 m

trk> = 8.0 m

Nevent = 175K

Rvelo – Rtrack (m)

Resid = 18.0 m

trk> = 5.1 m

Nevent = 12.5K

Rvelo – Rtrack (m)

Trk error = (pixel)1.85m (multi-scatt.)

<trk> = quadratic average over all trks

Tracking Error from Pixel (m)

Error < 6 m

To improve tracking precision one has to sacrifice statistics.


Resolution vs Pitch

R sensor of R/pairV

elo

Hit

Res

olut

ion

(m

) Preliminary !.Angle ( )- 0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5

Strip Pitch (m)

Seed threshold = 6 ADC ~ 9.6 Ke

Side threshold = 3 ADC ~ 4.8 Ke

Tracking projection uncertainty removed from resolution.

Tracking precision is determined for each point ( ~ 4.7–5.4 m).

Error bar represents only statistic error.

Linear charge weighting, eta-correction not applied yet.


Tracking Precision

For each track the projections on Velo and projected errors in both X and Y directions are calculated using the corresponding pixel resolutions. R and error in R is calculated from X/Y.

For each sample (point), the projection error is quadratically averaged over all tracks used.

Projection error due to multiple scattering is ~1.85 m obtained from simulation.

The alignment error is to be determined.

Angle ( )- 0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5

R E

rror

Fro

m T

rack

Pro

ject

ion

(m

)

Strip Pitch (m)


Resolution vs Track Angle

Pitch ( m)

40 – 5050 – 6060 – 7070 – 8080 – 90

90 – 100


Vel

o H

it R

esol

utio

n (

m)

Effective track angle is determined in plane perpendicular to the strip.

Sub-samples of 0, 3, 7 and 11 are with reasonable large statistics.

Other angles are due to concentric strip, thus with small amount of hits.


Summary

Pixel charge sharing is consistent with expectation.

Velo hit resolution is measured. The trend is as expected. But the value is not as good as we hoped.

RR studies are on going.

Explanation on Fanl tb09 VELO data, pixel tracks and supporting software can be found in 08/21/09 presentation. http://indico.cern.ch/getFile.py/access?contribId=5&resId=1&materialId=slides&confId=66615

Velo data, pixel data and pixel tracks (pixel track files will be updated with better alignment) are in /castor/cern.ch/lhcb/testbeam/lhcbvelo/fnaltb09

Track tool and a small example code ( Example.cpp, PixelTracks.cpp, PixelTracks.h, IPixelTracks.h ) can be found at http://phy.syr.edu/~jwang/projects/tb09/


Data Sets

Detector Angle HV Setting Stations ADC

90 Lars90 Kazu

90, 20, 40 Chris P2V1 20.15 - 20.18 9ChrisLarsChrisLarsChrisLars

500 Lars500 + scan Chris500 + scan Kazu

8 500 + scan Kazu 5 P3V1 24.07 - 24.18 514 500 + scan Kazu 4.5+thin P4V1 26.06 - 26.15 630 500 + scan Kazu P5V1 26.19 - 27.09 77

500 + scan Kazu500 Chris

4 500 Kazu P5V3 27.20 - 27.22 99KazuChris

23

26

88

106

First

1

16

20

21.16 - 21.18

22.07 - 23.14

27.11 - 27.18

28.09 - 28.22

Time Range (date.hour )

21.08 - 21.10

21.12 - 21.14

19.13 - 20.11

VELO Pixel

0

8Rbottom

4

0RR

bottom

4

12

500

5

2-bit

binary

90

90

90

8RR top

RR middle

0

Geom Config

P2V2

P2V3

P1V1

P2V4

P2V5

P5V2

P5V4

Analysis on TB0904 Data

Documents

pixel track event

pixel tracking data

pixel events

pixel tracks

track projection error

track projection precision

pixel hits

track probability distribution