ITS Algorithm Review Introduction

1
ITS Algorithm ReviewIntroduction

• A. Badala, R. Barbera, B. Batyunya, Y. Belikov,
  M. Bondila,
• N. Bustreo, R. Caliandro, P. Cerello, E. Crescio,
  R. Fini,
• E. Fragiacomo, J. Hubbel, E. Lopez, G. Lo Re, M.
  Masera, B. S. Nilsen,
• G. S. Pappalardo, A. Pulvirenti, W. Peryt, S.
  Radomski, F. Riggi,
• T. C. Randles, P. Skowronski, R. Turrisi, L.
  Vannucci, T. Virgili.

2
Agenda

• Introduction B. S. Nilsen
• Infrastructure-Organization. B. S. Nilsen
• Geometry R. Barbera (afternoon)
• SPD T. Virgili (B. S. Nilsen)
• SDD P. Cerello
• SSD E. Fragiacomo
• Tracking A. Badalá/Y. Belikov
• PID B. Batyunya

3
ITS Algorithm ReviewInfrastructure - Organization

• by
• Bjørn S. Nilsen

4
Philosophy

• Keep it as general as possible
• ITS geometry has gone through many changes
• Simulations?Reconstruction has maximum
  flexibility
• Keep things modular
• Easier to make changes
• Minimize code duplication

5
Which are theNon-general Routines

• AliITSv
• Defines Geometry/Material of the ITS
• AliITSsegmentation
• Defines Geometry of detector layout
• AliITSresponse
• Defines electrical and related properties
• AliITSClusterFinder
• First level Reconstruction

6
One Last Assumption
Z axis same direction as ALICE Z axis.
Violated by AliITSv3.
7
Order of Operations

• AliRunInitMC
• Executes Config.C
• Defines gMC AliGeant3()
• Defines all detectors new AliITSvPPRasymm()
• gMC-gtInit
• For each Detector
• Create Materials AliITSCreateMaterials
• Create Geometry AliITSCreateGeometry
• FinishGeometry
• For each Detector
• Detector-gtInit AliITSvPPRasymmInit
• Sets up and fills AliITSgeom class (GEANT3.21
  only), Optionally may read or write AliITSgeom
  to/from file.
• Sets default segmentation, response, but not
  simulation or reconstruction (at present). See
  later.
• Init for Step Manager, sensitive volume
  (AliITSInit) and ITS mother volume
  (AliITSvPPRasymmInit) ID numbers.
• Detector-gtBuildGeometry AliITSBuildGeometry

8
Comments about Materials
galice.cuts ? Config.C only ?
30KeV e- in Si has a range of 14µm
? In Config.C
9
Still some moreabout materials
10
Detector Simulations

• AliITS(TObjArray) fDetType one for each
  detector type.
• AliITSDetType classes.
• AliITSresponce fResponce
• AliITSsegmentation fSegmentation
• AliITSsimulation fSimulation
• AliITSClusterFinder fReconst
• And the Digit and Cluster branch names
• Default values set in AliITSvPPRassymInit().

11
Hits to SDigits/Digits

• Set default or other simulation, get default or
  change response and segmentation classes for each
  detector.
• Set up SDigits or Digits Tree.
• All hits copied into memory sorted by module.
  AliITS(TObjArray) fITSmodules one for each
  module 2198.
• AliITSmodule
• Module index number ? Layer, Ladder, Detector ?
  Detector type
• TObjArray of hits
• Arrays of track index and hit index numbers
• The location in the TreeH of a specific hit.
• Merging at hit level easy. Just add new hits to
  their module.
• Size of Hitnumber of hits lt Size of
  SDigitNumber of SDigits.
• Loop over these modules
• Determine detector type (SPD,SDD,SSD,)
• Get copy of simulation, response, and
  segmentation
• Execute simulation to make SDigits or Digits.
• SDigits and/or Digits automatically sorted by
  module.

12
SDigits to DigitsUses SDigit Merger

• Set default or other simulation, get default or
  change response and segmentation classes for each
  detector.
• Must be the fully compatible with that used in
  Hits To SDigits!
• Set up Digits Tree.
• Read in SDigits (already module ordered).
• Loop over these modules/SDigits
• Determine detector type (SPD,SDD,SSD,)
• Get copy of simulation, response, and
  segmentation
• Execute simulation to make Digits.
• Digits automatically sorted by module.
• Merger loops over SDigit files. The first file
  processed can be used to set up a region of
  interest cut.

13
Comments on Digits

• Each detector type defines its own digit.
• For each detector, there are two type of digits.
• Basic Digit. Same for all detectors
• Int_t fCoord1 SPD, SDD z cell number SSD P/N
  layer.
• Int_t fCoord2 SPD, SDD x cell number SSD strip
  number.
• Int_t fSignal SDD, SSD ADC value, SPD 1
• Simulation Digit. Detector specific
• SPD Array of track index and hit index
  numbers, signal in SPD cell (Int_t electrons).
• SDD Array of track index and hit index
  numbers, signal in SDD cell per track (Float_t
  electrons), and the signal the dominant particle
  contributes to the total signal.
• SSD Array of track index and hit index
  numbers.

14
ReconstructionDigits to RecPoints

• Set default or other ClusterFinder
  (reconstruction), get default or change response
  and segmentation classes for each detector.
• Should be the fully compatible with that used in
  Hits To SDigits or Hits To Digits
• Set up RecPoint Tree.
• For each module.
• Read in this modules digits
• Determine which detector
• Perform the Digits to RecPoints

15
Comments onAliITSgeom class

• Consists of
• TObjArray of AliITSgeomMatrix class members.
• Transformation from detector local coordinates
  (x,y,z) to global coordinates.
• Minimal set of functions exclusively used via
  AliITSgeom class.
• Number of layers, ladders per layer, detectors
  per ladder, and the total number of modules.
• Flag to describe transformation type. Default
  GEANT.
• TObjArray of the types of shapes of each
  detector.
• Large set of functions related to geometry and
  coordinate transformations. Some just repackage
  AliITSgeomMatrix classes.
• Primarily motivated by needs of ITS Alignment
  work.

16
Fast Simulation-Reconstruction

• Takes hits, smears position, generates RecPoints.
• Position smearing s need review.
• Overlapping tracks have perfect separation.
• Signal effects, noise, response, dead-channels,
  Not taken into effect.
• Uses the same RecPoint data structure as the slow
  simulation.
• Used by some groups. They must verify its
  validity, we will not.

17
Macros

• Simulation and Reconstruction Macros
• AliITSHits2SDigits.C
• AliITSMerge.C
• AliITSDigits2RecPoints.C

18
SPD simulation

• Tiziano Virgili
• University of Salerno
• Present by
• Bjørn S. Nilsen

19
Content

• The Model.
• What is new.
• Results from simulation.
• Comparison with Experimental data.

20
The model

• Two models are available.
• Bari/Salerno (default)
• Geometrical charge sharing.
• Noise and threshold fluctuations.
• Coupling effects.
• Dubna
• Geometrical charge sharing.
• Charge diffusion.
• Noise and threshold fluctuations.

21
Geometrical Charge Sharing

• Find points where particle crosses pixel
  boundaries.
• Distribute energy/charge according to fraction of
  that between boundaries.

22
Simulation Parameters

• Threshold 2000 e
• Noise s280 e
• No noise appears above threshold
• No Coupling used.
• 1 dead pixels, random.

23
Test Beam Comparison

• The Bari-Salerno model (FORTRAN version) is used
  by NA57.
• The number of single pixel clusters simulated is
  more than seen in the Test Beam data.
• The number of 2 and more wide pixels are greatly
  under predicted.

Bari-Salerno Dubna Test Beam
24
Test Beam ComparisonsProblems cont.

• Adjusting parameter of the Dubna model can
  compensate a bit, but leads to unphysical
  diffusion parameters.
• Changes to the Bari-Salerno model would require
  similarly unphysical asymmetric coupling
  parameters.

Bari-Salerno Dubna Test Beam
25
Final Comments

• New ALICE1 chip will be used (not used in NA57)
• Lower noise, thresholds,
• Introducing coupling will increase 3 pixel
  clusters and not 2 pixel clusters (unless the
  coupling is non-symmetric).
• Diffusion exists and so the Dubna model needs to
  be reintroduced with some new modifications (work
  under way).