1 James N. Bellinger University of Wisconsin-Madison 26-March-2010 Transfer Line Fits Transfer Line Fits James N. Bellinger 26-March-
Jan 03, 2016
1
James N. Bellinger
University of Wisconsin-Madison
26-March-2010
Transfer Line FitsTransfer Line FitsTransfer Line FitsTransfer Line Fits
James N. Bellinger 26-March-2010
2
Contents
1. Determine that my transfer plate centers are better than Dave’s
2. Hand fits give me something to validate Cocoa with
James N. Bellinger 26-March-2010
3
Whose TP centers are better?
• Fit a set of SLMs with Dave’s and with my transfer plate centers at 0T
• Find alignment pins positions wrt the disk
• Compare with PG of pins wrt disk
• Select ME-2 and ME-3 since I did not test them before
James N. Bellinger 26-March-2010
4
Alignment pin PG looks OK
James N. Bellinger 26-March-2010
Alignment pins outline the SLMs neatly.One not measured
5
CMS X comparison
James N. Bellinger 26-March-2010
X_mine-X_PG mean 0.1mm RMS 1.1mm
X_dave-X_PG mean 0.6mm RMS 1.9mm
ME-2/SLM2 fits poorly for Dave’s #’s. Excluded here.
6
CMS Y and Z comparison
James N. Bellinger 26-March-2010
Y_Dave-Y_PG Mean 1.2mm RMS 2.6mm
Y_Mine-Y-PG Mean 0.3mm RMS 1.2mm
dZ is different: clumpy; but note that my clumps are tighter
7
Conclusion
• My calculated TP centers clearly result in better fits, even if the Cocoa model is not perfect
• Note that the Cocoa model does not account for the different chamber distances from the SLM: these need to be introduced by hand
James N. Bellinger 26-March-2010
8
Reason for Z “clumps”
James N. Bellinger 26-March-2010
Chambers are at different distances from the SLM laser
248mm
SLM line
Cross section of SLM
Cocoa model has all chambers in the same plane: Wrong but easily fixed up.
9
How Did I Generate Positions?
James N. Bellinger 26-March-2010
V direction
H direction
Three PG targets on top of DCOPS
Center of TP is hereSimple and short translation from target to center
Transfer DCOPSSLM DCOPS
Zsensor
10
Hand Fit Comparison
• I create 12 independent fits for the transfer lines
• 2 for each transfer line, in the H and V directions (illustrated in next slide) native to the transfer line coordinate system
• 0T data from June 2009: Oleg pointed me to a range and a BFI search found some good data: no profiles, unfortunately
James N. Bellinger 26-March-2010
11
Transfer Line Coordinate system
James N. Bellinger 26-March-2010
V direction
H direction
CMS X
CMS Y
Transfer plate “X” direction varies with position
12
Hand Fits Illustration
James N. Bellinger 26-March-2010
Transfer Line 1Horizontal (Rphi) coordinateFit for two laser beams and offsets for each DCOPSPosition is wrt center of Transfer Line
Blue points are raw dataBlack are fit positions
Distance between points is invariant with fit
CMS Z (mm)
13
Changes in Cocoa Model
• Single Transfer Line
• Fix DCOPS internals
• Fix DCOPS mounts
• Fix disks
• Widen TP error
James N. Bellinger 26-March-2010
14
Offsets for Line 1
James N. Bellinger 26-March-2010
StationdX (mm) Cocoa
dX (mm)Hand fit
dY (mm) Cocoa
dY (mm)Hand fit
ME+3 2.07378 2.520 -5.58365 -6.072
ME+2 2.06415 2.474 -4.66432 -5.098
ME+1 1.00499 1.220 -5.41692 -5.732
MAB+3 1.32987 3.855 -3.47646 21.982*
MAB+1 -2.30884 3.528 0.357306 25.469*
MAB-1 -1.29895 4.907 -2.51285 -0.214
MAB-3 1.74611 4.546 -0.113382 1.540
ME-1 1.02517 0.893 -5.25257 -5.361
ME-2 1.16804 1.462 4.50519 4.383
ME-3 0.11244 -0.040 3.85222 4.062
MA
B calcs n
ot readily com
parab
le yet
15
First pass at residuals
James N. Bellinger 26-March-2010
No cuts on quality yet 570 microns for this, 440 if exclude point at -3
Residuals still far too large, but agreement with hand fit says we’re on the right track.
16
Conclusions
• Cocoa reproduces the Endcap part of the hand fit successfully– Possibly better, since it includes beam fan tilt
instead of using averages
• More work required on MAB: different sizes
• If continue to get agreement, this will validate the Cocoa calculation
James N. Bellinger 26-March-2010
17
Plans
• Re-expand scope of fit
• Find MABs I can use to constrain the lines
• Check MABs for consistency
• If agree, calculate Transfer Plate positions in X and Y
• Review SLM models with Himali, and generate chamber positions
• Converge on a plan for Z
• Generate chamber Z and angles
James N. Bellinger 26-March-2010
18James N. Bellinger 26-March-2010
BACKUP
19
Predicting MAB positions
James N. Bellinger 26-March-2010
34.78 34.78 34.78 34.78 34.78 34.78 MAB center to MAB sensor center (fixed)
2736.35 2736.35 2736.35 2736.35 2736.35 2736.35Radial lever arm: center to DCOPS (fixed)
-33.14 -33.14 -33.14 -33.14 -33.14 -33.14 MAB sensor center to target (fixed)-6698.93 -6694.2 -6694.74 -6694.49 -6691.12 -6693.94 MAB center in Z
-0.003264 -0.003333 0 0 0 -0.003403 MAB y-angle rotation-6706.2214 -6701.6803 -6693.1 -6692.85 -6689.48 -6701.6118 Predicted target position-6671.4414 -6666.9003 -6658.32 -6658.07 -6654.7 -6666.8318 My predicted MAB center
-23.480 16.328 29.950 23.078 36.226 20.019 How far from ideal is rear?-24.743 13.167 27.616 26.346 47.551 24.589 How far from PG is rear estimate?
Try working backward-10487.737 -10485.839 -10486.666 -10492.268 -10500.325 -10493.57 Rear TP from photogrammetry-10519.287 -10517.389 -10518.216 -10523.818 -10531.875 -10525.12 Distance to IP side of TP4-9629.771 -9628.595 -9629.196 -9631.867 -9644.918 -9635.249 To IP side of TP3-8018.672 -8017.435 -8019.927 -8023.371 -8035.466 -8024.924 To IP side of TP2-6887.060 -6885.315 -6888.109 -6891.053 -6903.913 -6893.197 To IP side of TP1-6681.478 -6678.050 -6683.246 -6683.125 -6698.899 -6690.420 Estimating MAB target gives
24.743 23.630 9.854 9.725 -9.419 11.192 Estimated MAB-found MAB