TRD tracking

1
TRD tracking tracking in central barrel

• Marian Ivanov

2
Barrel tracking

• Tracking using information from the barrel
  detector
• Currently TPC ITS TRD TOF integrated
• Kalman filter approach chosen
• Algorithm
• TPC seeding - and tracking
• ITS tracking inside
• ITS back propagation (time of flight integration)
• TPC back propagation (time of flight integration)
• TRD back tracking (time of flight integration)
• TOF (PID using TOF)
• TRD refit inward
• TPC - refit inward
• ITS refit inward

3
Integration of TRD

• First approximation
• Pt and angular resolution using space point
  information from TRD worse results as without
• Periferial Hijing event magnetic field 5 T
  -dPt/Pt 0.7 without TRD 1 with TRD
• ? carefull investigation of the TRD tracking

4
Improvements and bug fixes in TRD tracking (1)

• Kalman filter for TRD
• All Kalman numerical functions checked
• Some of the numerical algorithm exactly the same
  as in the TPC
• AliTRDTrackPropagateTo, AliTRDtrackRotate
• Non algorithmic bug found but
• too simplified description of the material
  distribution space frame and frame for TRD
  modules not taken to the account ?
• FIX - First approximation (BUG fix)
• if the track change the sector in f or theta
  direction (it means cross huge amount of the
  material)
• the track is stopped, not propagated to the TOF
• but information is used during refit inward

TRD segment
5
Improvements and bug fixes in TRD tracking (2)
Cluster position (fykdz)

• AliTRDtrackUpdate
• Update track using cluster position information
• Different from TPC pad tilde angle adding
  correlation between y and z coordinate
  information
• Kalman Matrix multiplication Checked - OK
• 1 bug found
• When the track z information and cluster padrow
  information differs (track points to another
  pad-row) - tilted y correction was not correctly
  calculated

Track intersection (y,z)
6
Improvements and bug fixes in TRD tracking (3)

• Another bug fixes
• When the track was stopped the same material is
  used during track propagation
• Before PropagateTo used without material
• Refitting of the track using the clusters found
  in back propagation implemented
• Before, tracks forgot the clusters, looked for
  them again, if not found track stopped

7
Systematic effects (1)

• Tilted pads
• Supposed to improve z resolution
• dY on the pad length -1.5 mm
• Systematic effects -0.5 mm
• Angular and time bin depending residuals
• AliTRDclusterErrors.C macro used to calculate
  corrections

Time bin
Track angle
8
Systematic effects (Example)
Time bin
Time bin

• Y systematic shift as a function of the time bin
  (tan(?) - 0.4-0.2)
• Left side - before correction
• Right side - after correction

9
Barrel tracking performance TPCTRDITS (1)

• After previously described changes barrel
  tracking performance improved
• Hijing peripheral event magnetic field B0.5 T

10
Barrel tracking performance TPCTRDITS (2)

• High pt event magnetic field B0.5 T

11
Tracking performance TPCTRDITS (3)
Without TRD
With TRD

• Relative Pt resolution as a function of pt
• left side new results TRD with tilted pads
• right side old results (2003, TRD without
  tilted pads)

12
Full event

• Still problem combining TPC and TRD tracking
• Worse resolution even for high pt tracks
• Problems with unfolding
• Possible solution
• New unfolding mechanism planned to be implemented
• Cluster position uncertainty as function of
  measurable overlap factor
• Tracking linear tracklets of cluster inside of
  the chamber to remove fake clusters
• Update of the Kalman track using the tracklet
  position and angle information

13
Next improvements in the TRD tracking

• As mentioned before, uncertainties of the cluster
  points positions are correlated
• The track parameter precision doesnt scale with
  1/sqrt(Npoints)
• Correction factors introduced
• as Xu factor for z position measurement in
  update introduced
• Test beam additional correlations
  (unisochronity not yet included in simulations)
• Angular and y position shifts of all the tracklet
  points depends on the relative position of the
  track to the anode wire
• Signal shape also depends on the drift velocity
  of ions

14
Correlations distance to wireTest Beam
Z position
Z position

• Left side space point residuals versus z
  position according silicon detector
• Right side reconstructed angle versus z
  position according silicon detector

15
Correlation space point reconstruction (Test
beam)

• New method introduced to minimize systematic
  distortions in pad directions
• But
• Angular resolution did not improve
• MIP (Buzuloiu) fit optimized to minimize integral
  residuals ?
• Plan to optimize it to improve tracklet angular
  and position resolution
• More complicated
• Require good calibration

COG
residualspad
Gauss fit
MIP fit
phasepad
16
Conclusion

• TRD tracking checked and improved
• Tracking in the high multiplicities environment
  has to be improved (unfolding, new track update
  strategy)
• Correction for the systematic effects has to be
  introduced
• Fine tuning, wait for the updated simulation
  which will include unisochronity ?