ATLAS Pixel Detector Pixel Support Tube Rails Design

ATLAS Pixel Detector Pixel Support Tube Rails: Design and Prototyping September, 17 2002 Review September 2002 A. Das LBNL

Pixel Detector ATLAS Introduction September 2002 I. Two Rail Designs, V and Flat, are used to provide 4 point, Pseudo-Kinematic Support II. Rails are for delivery of the Detector only not for Support III. Longest rails in actual detector are approx. 9 feet long (2. 5 m) IV. Short test rails were made to check shape and surface quality A. Das LBNL

Pixel Detector ATLAS Rail Design Old Design DETAIL 2 DETAIL 1 New Design DETAIL 3 September 2002 DETAIL 4 A. Das LBNL 3

Pixel Detector ATLAS Rail Tools §Aluminum Tools • CTE = 22. 95 ppm/K §Cut By Wire EDM • Length of 12. 25 in. (311 mm) Flat Rail • Wire EDM is cheap, but surface is poor, so Tooltec release film was used V-Rail §Actual Rails 2. 5 m in length Cross Sections 311. 00 September 2002 A. Das LBNL 4

ATLAS Pixel Detector Lay-Up Procedure (Tool + Tooltec) Teflon Tooltec (5 mil) September 2002 A. Das LBNL

ATLAS Pixel Detector Lay-Up Procedure (Tool + Tooltec +Mat) 17 gsm Prepreg Reinforced Mat (10 mil) Teflon Tooltec (5 mil) September 2002 A. Das LBNL

ATLAS Pixel Detector Lay-Up Procedure (All Layers) 8 Plies P-30 Cloth (1 mm total thickness) 17 gsm Prepreg Reinforced Mat (10 mil) Teflon Tooltec (5 mil) September 2002 A. Das LBNL

Pixel Detector ATLAS Prototype Parts • – – – • – – September 2002 High Strength Carbon Fiber Cloth The cloth is 1 K tow square weave, and the resin is EX 1515, 125 gsm, 5 mil thickness Cloth has lower stiffness than the YSH-80 Print-Through Print-through was a problem on the initial parts, the rough fiber structure on the top of lay-up was pushed through to the smooth side by the pressure during the curing cycle. Since no additional resin dams were laid down and perforated A 4000 R was used as the release ply during cure, much resin bled through to the surrounding bag and breather material. It was believed that the print-through was caused by the lack of resin within the part. Resin dams were laid down on second set of prototypes and non-perforated A 4000 R was used with the result that the tool side was much smoother, proving hypothesis correct. Tool Part Interaction Another possible problem is the tool part interaction associated with the difference in CTE’s between the aluminum tool and the part. Tool Part Interaction cause the part to “bow” in the cross-sectional and or longitudinal direction A. Das LBNL

ATLAS Pixel Detector Tool-Part Interaction (Cross-Sectional Bow) • Both V-Rail and Flat Rail Prototype Parts fit the tool very well with no gaps. • Tool Part Interaction doesn’t seem to be a problem in the cross -section • Will CMM cross-section later to find out for sure September 2002 A. Das LBNL

ATLAS Pixel Detector Tool-Part Interaction (Longitudinal Bow) We need an equation relating radius of bow to max deflection Equation for Circle We can find the radius from heightgauge measurements, and use the equation above to calculate max bow. September 2002 A. Das LBNL

ATLAS Pixel Detector Worst-Case Longitudinal Bow Plot of longitudinal bow versus length of part for the second V-Rail Prototype Plot of longitudinal bow versus length of part for the second Flat Rail Prototype • A curve-fit with radius of 3, 451 in. best approximated the data. • Radius of 2, 352 best approximates data • For the final 9 ft. (108 in. ) long part, the estimated max bow is approximately. 422 in. September 2002 • Max bow of approximately. 620 in. for final part. A. Das LBNL

Pixel Detector ATLAS Conclusions September 2002 • Preliminary measurements using a height-gauge on prototype parts show that a worst-case bow of approximately. 620 in. is expected on the final 2. 5 m long rail • There are three ways to compensate for the longitudinal bow caused by Tool Part Interaction in the final part. I. A Reverse-Bow can be machined in the tool itself to compensate II. The “bowed” part can be bonded to the PST so that it lies correctly. III. In the future, steel tools with much lower CTE will be used in place of aluminum to minimize the effects of tool part interaction. A. Das LBNL