Outer Stave Prototype Update

E. Anderssen, M. Cepeda, M. Garcia-Sciveres, M. Gilchriese, N. Hartman, J. SilberLBNL

W. Miller, W. ShihAllcomp, Inc

ATLAS Upgrade WeekApril 2010

2

Basic Concept

• Foam 4.8mm thick• Cable 0.5mm• In slot in foam• Cable width restricted

Note there are4 separate cablesper stave

0.2, K 40 W/m-K

K13D2U 0-90-0 0.25mm thick

SS or Ti 2-3mm OD

Must have foam here

4cm

3

• Use old 1m stave(K7 foam) made by Allcomp, Inc• Cables designed for full length stave(1.3m) ie they are about ½ the 1.3m

length.• Cut slot for cable (would not do this for real stave). • Glue in cable and cable pieces. Bend tabs and glue. • No connectors yet.• Mount dummy silicon heaters one side.• Water at 20C. IR image of heaters

Tabs

End of cable. To use full length in central part of this stave

4

Thermal Results

• Discovered delamination on part of stave• Lack of glue• Fabrication error(missed)• T better than cable on top• As expected(for good region)

0 2 4 6 8 10 12 140

2

4

6

8

10

12

14

EmbeddedCable

Cableon top Embedded cable

0.43 W/cm2

5

Short Prototypes• All about 12cm long. All of same basic construction.• Hysol 9396/BN(30% by weight) to bond tube to foam• And same for foam-to-foam, facings to foam• K7 foam – heaters both sides

– One SS tube(2.8mm OD)– One Ti tube(2.2 mm OD)– Thermal cycle and irradiation(time sequence)

• 900 cycles (20C<->-35C) then• 1 cycle to -70C then• 1 cycle to about -175C with LN2 then• Irradiation to 50 MRad, then to 150 Mrad total

• K7 foam – heaters on one side– One SS tube(2.8mm OD)– One Ti tube(2.2 mm OD)– Irradiation and thermal cycle(time sequence)

• Removed one heater(and glue) after 1st thermal measurement• Irradiation to 1 GRad then • Put back removed heater• 250 cycles (20<->-35)

• K9 foam – heaters on one side– One SS tube(2.8 mm OD)– Thermal cycle only: 350 cycles(20<->-35)

• IR measurements, water cooling at 1 liter/min at 20C• Power is 0.5 W/cm2 on a heater• SE4445 used to mount dummy silicon heaters• Average temperature of dummy heaters will be shown

6

Summary of Results

• No difference in SS and Ti tubes (with given ID/OD). Not a surprise (from FEA).

• No significant change in thermal performance for any sample after any thermal cycle sequence, including LN2

• Effect of irradiation up to 1 GRad is <10% increase in T.

• Thermal performance with K9 foam is significantly better than with K7 foam, by about 25%

• Some details follow in next slides

7

• Average temperatures• Error (over time) at least ±0.5C• No difference between Ti and SS• No change from thermal cycle• After 150 MRad, < 1C degradation(for 0.5 W/cm2)• Visual inspection of ends(tube exits foam) shows no damage• Good agreement(pre-rad) with simple FEA(shown many times before)

Cycles K7TiAT K7TiAB K7TiBT K7TiBB K7SSAT K7SSAB K7SSBT K7SSBB Average0 8.62 9.03 9.07 9.42 8.88 9.14 9.0

100 9.35 8.87 8.74 9.05 8.86 9.16 8.92 9.19 9.0101200 9.27 9.11 8.89 9.22 9.07 9.36 8.76 8.71 9.0300 9.24 9.05 8.74 9.14 9.23 9.59 8.96 9.24 9.1350450 9.38 9.17 8.8 9.12 9.37 9.42 9.02 9.05 9.2600 9.18 9.02 8.88 9.23 9.44 9.59 8.64 8.95 9.1750 9.03 8.87 8.52 8.88 8.9 9.11 8.7 9.09 8.9900 8.71 8.85 8.39 9.22 8.74 9.09 8.84 8.84 8.8

9.08 8.89 8.61 9.28 9.47 9.64 8.75 8.92 9.1 After dry ice to -70C

9 8.95 8.49 9.34 9.16 9.13 8.92 9.07 9.0 After LN2 to -175C

9.26 9.1 9.53 10.01 10.18 10.04 9.71 9.75 9.7 After 50 Mrad

9.45 9.1 9.13 9.51 9.9 9.68 9.72 9.8 9.5 After 150 Mrad

Ti tube 4 heaters(two sides) SS tube 4 heaters(two sides)

8

• Average temperatures• Error (over time) at least ±0.5C• No difference between Ti and SS• Top heater and glue removed before 1 GRad• Replaced after 1 GRad. Can’t see T vs B difference.• Change after 1 GRad small, again <1C(including error)• No change from thermal cycle• FEA (pre-rad) same as other K7, good agreement

K7100SST K7100SSB K7100TiT K7100TiB Average8.84 9.41 9.31 9.7 9.32 Before 1 GRad9.15 9.51 9.79 9.44 9.47 After 1 GRad

9.23 9.56 9.8 9.48 9.52 1 GRad+250 cycles

9

• Average temperatures• Error (over time) at least ±0.5C• No change from thermal cycle• Results better than FEA prediction

• FEA comparison of K=22(K7 foam) and K=40(K9 foam) predicts improvement of 1.3C in average temperature.

• Measured is about 9C-7C=2C or more, ie. better. • Measurement errors? We think not • Why? No quantitative explanation.

– 130 ppi has smaller pore size => better contact with tube?– Simple bond line model between tube and foam is not correct– Glue penetration involves more high K ligaments?– Complicated FEA to get agreement

• More prototypes needed– Will include also high K facings (heat treated)

Cycles 130#1T 130#1B Average0 6.31 6.73 6.52

100 6.43 6.77 6.60100 7.08 7.17 7.13 Repainted heaters350 6.84 7.13 6.99

10

Summary of Other Irradiation(1 GRad) Results• Hysol 9396 – shear strength measurements done improperly, no

result. Thermal already described.• K7 foam

– Limited sample size, remember spread in results(10%)– Tensile strength: before(195 psi); after(200 psi)– Compressive strength: before(195 psi); after (179 psi)

• K9 foam samples not irradiated yet but don’t expect difference from K7 (but have to check)

• K13D2U laminates– Few facing samples, thin, density varies by ± 8%– Modulus: before (44 Msi); after (44 Msi)– Strength: before (91 ksi); after (98 ksi)

• Bottom line: OK to use unirradiated properties. Changes less than about 10%.

11

Support Studies• Presentation given to engineering group about 2 months ago• Have initial concept for support of complete SLHC pixel tracker: outer

staves, disks and inner replaceable system• Shell support of outer staves with minimal rings• Inner pixels on rail system (not shown below)• Weights and load points estimated. • Preliminary FEA encouraging for relatively simple design.• Work on hold pending understanding of upgrade schedule and constraints

Interior Composite rings

Shell end supports

Ring supports

12

Next Steps• Other side of 1m stave looks OK, move heaters to this side…then• Electrical tests of cable inside 1m stave…then• Some thermal cycling to investigate reliability of cable gluing. Could use

CO2 at SLAC or other methods.• Decision to be taken about making full-length prototype with K9 foam.• Make ½ length active stave(s) (see session on “Pixel Stave 2010”)

– Will use K9 foam. Tube type (Ti or SS) to be chosen.– Good cables– Mount chip modules and eventually real modules

• Plan for mechanical R&D depends some on outcome from this meeting– Will do under any circumstances

• Continue prototype studies related to detailed understanding of foam, foam/glue and other stave interfaces both thermally and mechanically

• Small prototypes with K9, including irradiation to 1 GRad (can take other’s samples again)

– Construction of next long staves depends on overall design goal• “Early” replacement of pixel system: current design may not be optimal• Would like to prototype “bent” staves (to eliminate or reduce size of disks) – see talk in other

session