GLAST LAT Project Tower A software alignment 28

GLAST LAT Project Tower A software alignment 28 Jan 2005 Tower A software alignment 28 january 2005 M. Kuss, A. Brez 1

GLAST LAT Project Tower A software alignment 28 Jan 2005 MCM side Ladder alignment on a tray The ladders alignment is measured with a CMM with optical probe (3 points per ladder) in G&A. The small dispersion of the data (s=18 mm, planarity ~ 0. 1 mm) allows to consider the 4 ladders of a tray plane as a single detector. We do not need to introduce ladder per ladder alignment corrections M. Kuss, A. Brez 2

GLAST LAT Project Tower A software alignment 28 Jan 2005 Tower B ladders alignment M. Kuss, A. Brez 3

GLAST LAT Project Tower A software alignment 28 Jan 2005 Tray thickness SSD plane Pins plane The tray thickness is also measured. The results are well within the design specs * Distance from the SSD to the alignment pins on thermal boss M. Kuss, A. Brez 4

GLAST LAT Project Tower A software alignment 28 Jan 2005 Software tray alignment: rotation residual/mm x • y orth. coordinate/mm profile histogram residual/mm A parallelism error generates an error which is proportional to the orthogonal coordinate (v) of the event. v has to be extrapolated from the fitted 3 dimensional track. The fit of the profile histogram gives immediately the rotation tan( ) of the plane (parameter p 1). p 0 again is a measure horizontal shift, but here at v=0 (the preceeding methods give the shift at the center of the wafer). orth. coordinate/mm M. Kuss, A. Brez 5

GLAST LAT Project Tower A software alignment 28 Jan 2005 Software tray alignment: vertical residual/mm In similar way to the horizontal case, it is possible to determine the relative height in the tower of a SSD plane studying the residual distribution as a function of the incident angle of the tracks. x=tg( ) • z Theoretical SSD height cot( ) True SSD height The fit of the profile histogram gives immediately the vertical shift (parameter p 1), and once again the horizontal (p 0). residual/mm z profile histogram cot( ) M. Kuss, A. Brez 6

GLAST LAT Project Tower A software alignment 28 Jan 2005 Software tray alignment: horizontal before shift after A shift of a plane with respect to the other ones generates a residual distribution not centered around 0 (measured points - reconstructed trajectory). An iterative algorithm allows to align the detectors so that all the residual distributions are 0 centered. M. Kuss, A. Brez 7

GLAST LAT Project Tower A software alignment 28 Jan 2005 Coordinate shifts X Sidewalls assy Vibe test TV test Shipment to SLAC Sidewall removal M. Kuss, A. Brez The tower jigs aligns the pins that have been used to align the ladders (the best reference) When the pins on the alignment wall of the tower jig are removed, the sidewalls partially realign the trays to their external dimensions with a max alignment loss ~0. 1 mm 8

GLAST LAT Project Tower A software alignment 28 Jan 2005 Coordiante shifts Y These trays are aligned on the tower jig on a structural closeout. The relative alignment of the ladders is poorer, but the max loss of alignment after sidewalls assembly is only 0. 05 mm Reference wall of the tower jig M. Kuss, A. Brez 9

GLAST LAT Project Tower A software alignment 28 Jan 2005 Ladder parallelism Parallelism for X trays ± 0. 1 mm Parallelism for Y trays ± 0. 2 mm except for layer Y 2 (light tray 016, top side) with an anomalous parallelism of 0. 45 mm The general difference in parallelism between X and Y trays can be related with the tower jig. No explanation for the Y 2 anomalous parallelism. M. Kuss, A. Brez 10

GLAST LAT Project Tower A software alignment 28 Jan 2005 Coordiante shifts Z The software determination of the tray position in Z is poor as shown by the differential plot Z error ~± 0. 2 mm M. Kuss, A. Brez 11

GLAST LAT Project Tower A software alignment 28 Jan 2005 Tower B results M. Kuss, A. Brez 12

GLAST LAT Project Tower A software alignment 28 Jan 2005 Conclusions Good data! Good alignment and parallelism of the trays, typical values in the ± 0. 2 mm range over the whole tower Very good mechanical stability: • limited alignment loss after sidewalls assy • no alignment loss after vibe, thermo-vacuum, a single sidewall dismounting and shipment to SLAC • Tower B data are slightly better than tower A data M. Kuss, A. Brez 13