Forward Pixel Survey: Status and Plans

1
Forward Pixel Survey Status and Plans

• Andrei Gritsan, Mikhail Kubantsev, Chung Khim
  Lae, Nhan Tran
• CMS Alignment Survey and First Geometry Workshop
• June 27, 2007

2
CMS Detector Forward Pixels

• 4 Disks with 24 Blades on each
• Each Blade consists of 2 Panels back-to-back
• 4-Sensor Panel and 3-Sensor Panel

3
Forward Pixel Survey Survey Hierarchy
Panel Survey
Half-Disk Survey

• Survey Hierarchy
• Sensor
• Panel
• One side of Half-Disk
• Half-Disk
• Half-Cylinder
• Pixel Endcap

Half-Cylinder Survey
4
Survey Analysis and Software Hierarchy

• Since we expect that the blades will be the more
  stable object over time, we use this hierarchy in
  software.
• Slightly complicates survey analysis and survey
  error analysis
• Survey Analysis package in CMSSW using rigid body
  formalism to match fiducial points to ideal

5
Panel Survey

• Over 100 Physics Run panels have been surveyed
• RMS Values for typical displacements during
  assembly relative to nominal positions (in local
  frame)
• RX 68 µm
• RY 49 µm
• RZ 17 µm
• OX 0.71 mrad
• OY 1.51 mrad
• OZ 2.16 mrad
• Pixel Sensors are flat to within 5 µm

6
Panel Survey Error Analysis

• Two types of error
• Measurement 0.5 µm in XY, 5 µm in Z
• Time/Installation 5 µm
• Use Toy MC to simulate errors and effect on
  sensor displacements
• Error few µm
• Little correlation

7
Half-Disk Survey
-Z1 PR Half-Disk

• Each side of a Half-Disk is surveyed
• Complications - Not all panels can be seen, not
  all fiducials can be seen for every panel
• Sides are joined using reference targets (next
  slide)

8
Half-Disk Survey Reference Targets

• Three types of reference targets
• Stainless steel spheres (SS) on HD support frame
• Ruby Balls (RB) on HD frame
• PHG targets on the HD frame
• Current errors on reference targets from survey
  analysis (SS balls) 10 µm
• Connect SS and PHG methods by taking CMM and PHG
  measurements of 1mm targets
• Agreement of 10-15 µm on the half-disk (300 µm)
  between two techniques

9
Half-Disk Survey Software Hierarchy

• Using reference targets, we join the two sides
  of a half-disk (SS balls).
• With full half-disk created, we can define
  blades.
• -Z1 and Z2 half-disks anlayzed, data available
  for Z3 and Z4 half-disks
• http//www.pha.jhu.edu/ntran/cms/physics_run.pdf

• -Z1 Half Disk
• Panel displacements within a blade
• Blade displacements within a half-disk

10
Half-Disk Survey Error Analysis

• Consider panel error within a blade and blade
  error within a half-disk
• More complicated error analysis
• 2 hierarchies studied at the same time convoluted
  by reference target error (10 µm) in addition to
  measurement and time error
• Hard to pin down reference target error due to
  low statistics
• Different panels have different amounts of points
  surveyed more points surveyed ? less error
• Consider different typical scenarios
• The position of the panel or blade along the
  half-disk gives different correlations
• Each panel/blade position has a different
  correlation matrix associated with it
• Result is different covariance matrix for each
  panel/blade position
• Possible deformations of the disk during
  installation into half cylinder are bringing some
  additional systematics and are under study.
  Effects of temperature distortions are also under
  study.
• With more half-disks surveyed and analyzed, error
  knowledge will improve

11
Half-Disk Survey Panel Error

• Study errors for different fiducial patterns
  using Toy MC
• Found correlations
• For panels belonging to the same blade, there are
  some opposing correlation matrix elements

12
Half-Disk Survey Blade Errors

• Studied errors using Toy MC (same as panels)
• Error on the order of few µm, will increase with
  time/temperature dependence
• Found correlations, dependent on position of
  blade on the half-disk
• Example of correlations

13
Half-Cylinder Survey

• Physics Run Half-Cylinder Survey to take place at
  CERN
• Photogrammetry survey at TIF will define
  positions of the disks relative to each other in
  the service cylinder and in the supporting
  structure (feet, rails).
• Final survey of the service cylinder relative to
  the Tracker will be done after installation in
  the CMS at P5.
• Accuracy under study with preliminary estimates
  of 50 - 100 um
• Software framework ready

14
Summary/Status

• Hierarchy sensor, panel, blade, half-disk,
  half-cylinder
• Survey is ongoing, procedure is set for sensor,
  panel, blade, and half-disk
• Error Analysis complete for these structures as
  well
• Half-Cylinder survey is now planned and will
  start at CERN next month

15
backup
16
Survey in HIP Algorithm

• In Alignment/SurveyAnalysis, SurveyDBUploader
  uploads survey information into two POOL-ORA
  objects
• Survey Values - TrackerSurveyRcd
• Survey Errors - TrackerSurveyErrorRcd
• SurveyInputTrackerFromDB uploads results from
  survey analysis in a text file
• SurveyMisalignmentInput uploads results from
  misalignment scenarios
• Implementation to HIP Algorithm by C.K. Lae
• MORE HERE?