1. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 1 Incorporating FSI with the Offline Alignment Overview ATLAS Group, University of Oxford Stephen Gibson Brief overview of FSI Demonstration System Length Measurements to Module Co-ordinates Grid simulations Future Work

2. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 2 Frequency Scanning Interferometry Alignment grid of length measurements help constrain the SCT shape to ~ 10 mm. Each line of the grid must be measured to ~ 1 mm. ~800 simultaneous length measurements. Components: rad-hard, low mass, operate >10 years.

3. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 3 FSI Length Measurement TUNABLE LASER sweep  To interferometer with OPD to be measured DETECTOR M1 M2 Reference Interferometer with fixed OPD I MEASURED   I REF   Ratio of phase change = Ratio of OPDs  /c]D   /c]L 

4. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 4 Interferometers inside ATLAS Each line of the alignment grid inside ATLAS will consist of a quill (two optical fibres & beam splitter) and a retro-reflector. quill jewels beam splitter variable path fixed path delivery fibre return fibre support structure

5. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 5 Demonstration system: Square Grid 6 simultaneous length measurements made between four corners of the square. +7th interferometer to measure stage position. Displacements of one corner of the square can then be reconstructed.

6. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 6 Square Grid Jewel Reconstruction Results Std Dev = 400 nm

7. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 7 Current Work: Tetrahedral Grid Stage raised up by 100mm to form tetrahedral grid. Currently investigating the reconstruction of three dimensional jewel co- ordinates.

8. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 8 From FSI Measurements to Module Co-ordinates Problem – how to incorporate the FSI information into the offline alignment? First an overview of how FSI can be used. Then some current work on the reconstruction process. Essential steps to reach module co-ordinates:  FSI scan  Grid lengths  Reconstruct node coordinates  Interpolate nodes to give module co-ordinates  Use tracks to refine the module co-ordinates The reality is more complicated…

9. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 9 Phase unwrapping Fringe fitting Dual-laser drift correction Subscan linking Refractive index correction FSI Grid Lengths Reconstruction Software FSI Scan Module Co-ordinates Node Co-ordinates Opto-geometrical model of system Degrees of freedom definition Database of nominal jewel coordinates Node topology Jewel internal dimensions Reconstructed Jewel Co-ordinates Interpolation Software Shape parameterisation Data from initial X-ray survey FEA: barrel and disc eigenmodes (Check with ESPI studies) Calibration Offline Alignment Level-3 trigger? Quasi real time

10. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 10 FSI & Offline Alignment Integration FSI will produce quasi real time module co- ordinates with associated errors. Ultimate precision on module co-ordinates will come from tracks. How can FSI help? Offline alignment could use the FSI measurements of module co-ordinates:  as the starting co-ordinates for iterative analysis.  to correct for short-timescale motions in the analysis of long- timescale track data.  as a cross-check of the final calculated co-ordinates.  to help with those distortions that tracks are less sensitive to: Sagitta Z motion (affects rapidity) Multipole radial distortions (elliptical/pear shaped) Relative rotations of distant sections of InDet (invariant masses) ref1

11. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 11 Barrel Grid Simulations FEA model of carbon fibre support structure 70 35  m0m0m

12. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 12 Simulgeo ref2 model of Alignment Grid nodes (jewels) Z X Y ASSUME: end flanges are rigid rings & central jewels constrained in rotation

13. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 13 Lines of sight for one quadrant of Alignment Grid

14. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 14 Single Barrel Grid Simulation Results NB: rigid end flanges assumed – currently repeating with increased number of degrees of freedom. 1 micron precision assumed throughout. Fixed inner barrel. Central jewels constrained in rotation Result without radial lines to modules

15. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 15 Cross-check of Grid Simulations Full barrel grid simulations should predict errors on all nodes of grid, for given measurement precisions. Idea:  Take FEA model of perfect barrel Extract grid line lengths (add random errors to lengths) Pass to reconstruction software for calibration of model  Distort FEA model eg, twist and/or multipole distortions Extract new lengths (add random errors to lengths) Pass to reconstruction software Calculate reconstructed node co-ordinates and compare with those in FEA model Repeat later including interpolation software.

16. Stephen Gibson, ATLAS Offline Alignment, 2 nd July 2002 16 Future Work Continuing 2D and 3D grid testing & modelling. More detailed, full barrel simulations in progress. Check reconstruction software model with known distortions. Tony Weidberg + new post doc:  Interpolation software  Continuing FEA/ ESPI studies References D.F. Howell et al.,’ATLAS-SCT-Alignment Overview’, University of Oxford, ATL-IS-ES-0026. P. Coe Doctoral Thesis, University of Oxford 2001. S.M. Gibson, ATLAS-SCT-Alignment Forward Grid Simulations, ATL-IS-AP-0054. ref1 S. Haywood, ‘Alignment, Stability and FSI’, RAL, SCT End-cap Engineering, 6 December 2001. ref2 used with kind permission of the author: L. Brunel, ‘SIMULGEO: Simulation and reconstruction software for opto-geometrical systems’, CERN CMS Note 1998/079.