1. LHCb UT UPSTREAM TRACKER – Mechanics and Cooling –
Ray Mountain Syracuse University
Burkhard Schmidt CERN
for the LHCb UT Group

2. Outline Introduction Stave design and construction Modules
Simulation & prototypes
Stave components
Stave construction
Modules
Superstructure
Frame and Box
Summary and plans
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

3. Upstream tracker: introduction

4. LHCb Detector Upstream Tracker (UT)
Four planes of silicon strip sensors
To be located upstream of magnet, between VELO and Tracking System
Replaces current TT
Installation scheduled 2018–2019
Future location of
Upstream Tracker
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

5. UT Schematic Representation (1)
RAIL
SERVICES
ELECTRONICS
VOLUME
SERVICES
READOUT ELECTRONICS
COOLING MANIFOLD PS
LV/HV PS
LV/HV
END-MOUNTS
FRAME
FRAME
BOX A
BOX C
SENSORS
STAVES mounted
on FRAME
@ service boxes
ELECTRONICS
VOLUME
SERVICES
READOUT ELECTRONICS
SERVICES
Four planes of sensors – Box/Frame retract – beam pipe is captured by UT
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

6. UT Schematic Representation (2)
4 planes constructed using “staves” with silicon on both sides, partially overlapping in the x-direction to ensure 100% coverage
higher segmentation in the region surrounding the beam pipe
Inner detectors with circular cut to maximize acceptance
Electronics located near sensors to allow segmentation & improved signal/noise
Basic mechanical unit “stave”
UTaX
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

7. STAVE DESIGN
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

8. UT Stave Design (1) Stave Components of fully-loaded stave
Main mechanical element of the UT
Provides for the mounting and precise positioning of the silicon sensors
Comprised of competing mechanical, thermal and electronic elements
Vertical support
Stiff sandwich structure
Integrated with the cooling system
Components of fully-loaded stave
“Bare” stave: basic innermost structural support / cooling tube
Data-flex: signal readout / power distribution / control lines
Module: Sensor / hybrid / stiffener, mounted on both sides of stave
Adapting ATLAS-type Integrated Stave concept
“FULL-LOADED” STAVE
DATA-FLEX
HYBRID-FLEX
ASICs
SENSOR
MODULE
99.5 mm wide
~1640 mm long
COOLING TUBE
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

9. UT Stave Design (2)
SENSOR
~3.5 mm
thick
ASICs
STIFFENER
WITH
ANCHOR TABS
EXTERIOR
(FULL-LOADED STAVE)
INTERIOR
(BARE STAVE)
COOLING TUBE
“SNAKE”-SHAPED
(displaced for clarity)
FOAM (STRUCTURAL)
FOAM (THERMAL)
99.5 mm wide
~1640 mm long
~8 mm thick
BEAMPIPE
CUTOUT
CARBON FIBER SHEETS
(BOTH FRONT AND BACK)
HYBRID-FLEX
DATA-FLEX
Bare Stave: CFRP (Carbon Fiber Reinforced Polymer) face sheets epoxied to foam core in sandwich structure with embedded cooling tube, all-epoxy construction Foams: thermal foam for heat transfer, lightweight structural foam for rigidity of sandwich structure Cooling Tube: Ti mm OD, 135 um wall, “snake” shape, runs under all ASICs and edge of each sensor Goals: Keep sensors at –5°C or below, uniform ΔT=5°C across sensor, ASICs < 40°C (6 mW/ch, W/asic)
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

10. SIMULATIONS & PROTOTYPES

11. Thermal Simulations Snake Tube Design Straight Tube Design
S. Coelli
M. Monti
Snake Tube Design
Straight Tube Design
FEA analysis to aid in stave design work
Comparison of two major designs: Snake vs Straight CO2 Cooling Tube
Thermal Loads:
6 mW/ch est. yields 68 W/stave max
Self-heating in silicon after 50 fb-1 is small, ~260 mW/sensor max
Data-flex heat load ~10% power dissipation in stave
Analysis of full stave model (highest power case, including thermal barriers) indicates both designs would meet our goals, with snake design slightly better
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

12. Demonstrator Comparison
W. Lentz
Snake Tube Design
Straight Tube Design
Positions of RTDs on Demonstrators
Full-Scale Prototypes (with CO2 blow-through system) were constructed to compare the two designs
Both demonstrators had the TDR module design, both used same construction techniques (though Straight Tube demonstrator used thermal vias and additional carbon foam, compared to FEA model)
Flow-rate and pressure were adjusted for each run, stability reached, then repeated for different settings (equilibrium temperature is sensitive to these settings)
Results from the comparison of the cooling behavior of the two designs:
Both designs successfully reached target temperature (–5°C on sensors) and below
Both designs achieved approximately ΔT=5°C across sensor
Snake-Tube demonstrator reaches target temperature with lower flow and higher pressure (i.e. more “easily”) than two straight tube demonstrator (compare red and blue curves for similar settings)
RTD = resistance temperature detector
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

13. STAVE COMPONENTS

14. CFRP Face Sheets (C) George (D) Ringo abraded surfaces 1550 mm
K13C2U high-modulus carbon fibers in EX1515 epoxy matrix, 45gsm
Layup is a three-layer stack of prepreg in 0/90/0 orientation
This emphasizes stiffness of the sheet while allowing for thermal conductivity in both in-plane dimensions
The facings are easily cut and formed to the sizes we need
We have obtained several batches of facings and have found them to be of good quality and good uniformity
Fabrication by Composite Workshop at University of Liverpool (Tim Jones)
abraded surfaces
1550 mm
CFRP = Carbon Fiber Reinforced Polymer
(C) George
(D) Ringo
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

15. Thermal Foam carbon foam carbon foam mass (long)
Thermal foam core component
Allcomp K-9 carbon foam
High thermal conductivity (~35 W/m.K), low mass density (0.2 g/cm3)
Open cell foam
Machinability is good
Ideal for spreading the heat transfer from a small tube to the large area required to cool the ASICs and sensors.
The machining of the foam is easy and the resulting surface can be made clean for epoxy.
Epoxy uptake measurements
Takes epoxy up into carbon foam to a depth about equal to the size of voids, typ um
Consider it a layer of ~0.5 mm which is a mix of epoxy and carbon foam, 70:30 roughly
carbon
foam
carbon foam
mass (long)
5% spread
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

16. Structural Foam Rohacell rohacell mass Structural foam core component
Evonik Rohacell 51 IG, a commercially-available polymethacrylimide (PMI) polymer foam
Solid, not thermally conducting, very low mass density (0.051 g/cm3)
Closed-cell foam
Machinability is good
Typically used in aerospace and industry as core material for a sandwich structure.
It is used as core in this design, and not as a structural element by itself.
Machined surface is left with opened foam bubbles so epoxy contact area increased, increasing the adherence.
Rohacell
rohacell
mass
2% spread
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

17. Cooling Tube Cooling Tube
Titanium CP2 alloy
OD mm, 135 um wall thickness
Manufactured by High Tech Tubes Ltd. UK
Development by ATLAS (Richard French, University of Sheffield) and we have benefitted greatly from their expertise and help
Snake shape has been designed to run under all ASICs and one edge of all sensors in a stave.
Tube Issues:
Bending
Fittings
Deformation
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

18. Tube Bending Tool based on design by Ian Mercer (Lancaster)
spools & handle
3 m tube
adjustable stops
Prototype tool: Bend one “wave” (four bends), then index and repeat Production tool: Bend entire tube with single tool (no indexing), will ensure better repeatability, compensate for neutral axis shift and springback
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

19. Bending Deformation: Longitudinal
Measure wall thickness
Bend sections of tube, encase samples in epoxy
Grind radial and longitudinal cross sections, polish surface
Image under digital microscope
Measure points on radii, bin
Wall thickness as a function of bend angle
Fish-shaped (expected)
Inner bend thickening, wrinkling
Outer bend thinning
Bending bunches material forward into the bend
Consistent with studies carried out by H. Yang¹, N.C. Tang²
J. Servatius
photomicrograph
¹ Yang et al. / Chinese Journal of Aeronautics 25(2012)1.
² Tang / International Journal of Pressure Vessels and Piping, 77(2000)751.
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

20. Bending Deformation: Transverse
Technique is the same
Measure points on inner/outer surfaces at even intervals on circumference
Wall thickness as a function of azimuthal angle
Slight egg-like shape (expected), follows from longitudinal results
J. Servatius
photomicrograph
(at 45° into bend)
cross section
at 45° into bend
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

21. Cooling Tube Fittings Option 1 Option 3 Option 2 VCR fittings
Can consider double-sealing with external sleeve
Mate tube and fitting with insertion joint
Avoid thermal stressing tube, increase ease of construction, cost, etc.
Allows for the possibility of in-situ repairs
Option 2
Brazing is the standard option for joining dissimilar metals.
stainless tube
Ti tube
weld
braze
stainless stub
VCR fittings
VCR blank gland
braze
Ti tube
stop
bored gland
Ti tube
epoxy
sleeve
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

22. Cooling Tube Fittings – Samples
M. Wilkinson
Have several samples of brazed tube in hand
GH Induction Atmospheres (Rochester NY)
LM Ag-Cu alloy
Single-joint, double-joint samples
Brazed samples
Ti tube SS stub
Have several samples of epoxied tubes in hand
Araldite 2011
Armstrong A12
Types of samples made
Single epoxy joint: one simple Ti/SS insertion joint with cap epoxied on end
Double epoxy joint: one simple Ti/SS insertion joint, other Ti/SS insertion joint with sleeve
Single braze joint + epoxied plug + Swagelok fitting
Ti SS
Double braze joint + Swagelok fittings
SS Ti SS
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

23. Cooling Tube Fittings – Pressure Tests
Pressure tests (ongoing)
Raise sample to 150 bar
Then let leak out, measure leak rate of sample plus system (in dead end line which is valved off from step-down regulator)
Repeat with irradiated samples – Irradiation to 100 kRad (60Co -rays)
Repeat with thermal cycling (typ. 5 cycles, RT to –15°C)
Repeat with pressure cycling (typ. 5 cycles, 1 bar to 150 bar)
Samples tested
Braze single
Braze double joint: @RT
Armstrong single joint (IRRAD): @RT, TC
Armstrong double joint TC
Araldite double
 All samples tested reached >150 bar without exhibiting any obvious leak, cracks, or other catastrophic failure modes
M. Wilkinson
S. Guerin
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

24. Traci V1 CO2 Cooling Test Setup
S. Coelli
(INFN-Milano)
Cold box
open
Setup will make detailed and systematic measurements of cooling system parameters
Dummy stave
Instrumented with 20 thermocouples
TRACI v1
2 PACL
Cooling pipe
Titanium snake
central stave
ID 2 mm, th 0,125 mm
Routing snake vertical
Tot length 3 m
cooling lines
With fluid measurement T, P CO2
Inlet dP variable CO2
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

25. CONSTRUCTION

26. Stave Construction (1)
Bare stave construction Procedure is an all-epoxy construction in precision fixturing End-of-stave mounts hold datum serving as master alignment references for all stages in construction. Epoxied to end of facing (Side A) Epoxy applied via stencil over large areas, volume control is important Jigsaw hold-downs position foam core pieces after epoxy application
END-OF-STAVE
MOUNT
REF
DATUM
REF EDGE
VACUUM
PULL-DOWN
FIXTURE 1
FACING
FOAM CORE
PIECES
JIGSAW
HOLD-DOWNS
REF EDGE
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

27. Stave Construction (2)
Trough milled in foam core for snake tube (prototype shown, real assembly will remain on vacuum fixture to maintain alignment) Second facing (Side B) mounted to vacuum fixture 2
Epoxy applied to tube via stencil. Tube laid into trough. Second fixture flipped. Fixtures mated ref alignment pins. Stops assure thickness control. Cure. Closing operations, metrology.
VACUUM
FIXTURE 2
(Facing Side B)
VACUUM
FIXTURE 1
Prototype snake tube
Demonstrator #1
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

28. MODULES

29. UT Module (exploded view)
HYSOL+BN
HYBRID-FLEX
STAVE = SENSOR width + 1 mm/side
CERAMIC STIFFENER
PHASE-CHANGE THERMAL INTERFACE MAT’L
ANCHOR TABS
PC TIM
ASICs
THERMAL VIAS
SENSOR
~500 um thick
COND. EPOXY
gap
fiducials
97.5 mm wide
~ 150 mm long
Design options
hybrid flex vs hybrid ceramic
Pyrolytic BN vs AlN
Electrically insulating, thermally conducting
Slits / no slits
CTE = Coefficient of Thermal Expansion
TIM = Thermal Interface Material
Sensor: –5°C, ΔT=5°C Stiffener: CTE match to Si Protect wirebonds, testing and handling Not mech over-constrain sensor, allows for bow Maximize heat transfer from ASICS to stave, minimize from ASICs to sensor Electrically isolate sensor bias from stave facings (ground) Reworkable epoxy (TIM): allows module removal if needed
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

30. Pyrolytic BN – Typical Properties (@R.T.)
Module Materials
Acquiring and Testing component materials for mechanical / adhesive / thermal properties as well as radiation damage testing
Pyrolytic BN – Typical Properties
   Apparent Density (gm/cc)
  1.95–2.22
   Tensile Strength (MPa)
  40
   Flexural Strength (MPa)
  80
   Thermal Conductivity (W/m.K)
  “ab” 60, “c” 2 *
   CTE (10-6/°C)
  “ab” 2
3.0 (-40 to +150°C) **
   Resistivity (ohm-cm)
  1015
   Dielectric Strength (kV/mm)
  200
   Dielectric Constant
  “ab” 5.2, “c” 3.4
   Total Metallic Impurities (ppm)
  <10
   Outgassing
  >Negligible
   Maximum Suggested Use (°C)
  2,500
* might work as is, but would have to grow thicker and cut slices to get max K
** from Morgan. Other values from Momentive. Compare to 2.3 for Si.
, ρ=1014 Ωcm
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

31. Thermflow Epoxy Film: Shear
M. Wilkinson
Measured ultimate shear strength (USS) for Thermflow films, using sandwich pull method Compared USS under irradiation at 30 MRad (60Co -rays) for different curing conditions: no significant effect of irradiation for any sample
Irradiated Thermflow Samples
T725
T777
TPCM 780 P I n
Holder
FORCE
epoxy
Carbon Fiber
(surface)
Test Epoxy
Kapton
no bake
bake
bake+weight
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

32. Sensor Epoxied to PBN Stiffener
Difference in sensor profile after wrt before epoxy operation: effect of epoxying to PBN Epoxied edges: one less, one more bowed Free corner rotated down by ~250 um (approximate size of free-standing sensor bow)
PBN stiffener
vacuum fixture
(weighted)
reference edge
sensor
shim
jig plate
Sensor MSL3092_10
(after–before epoxy, rotated)
epoxied on right edge X=0 and back edge Y=0
First sensor epoxied to ceramic stiffener
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

33. SUPERSTRUCTURE: FRAME & BOX

34. Stave Mounting to Frame
D. Wenman (PSL)
Kinematic mounts at top and bottom of stave
Fixed at top
Slotted at bottom
Standoffs in EOS mounts prevent crushing data-flex cable when mounting (both sides) Standoffs on frame allow stagger in Z, which provides overlap in X
Custom stud
Locates stave in standoffs
Thermal contraction in slots
Kinematic movement via spring-loaded nut and spherical washer set
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

35. Support, Frame and Box Two halves of UT will retract from beam pipe
J. Batista
O. Jamet
B. Schmidt
J. Andrews
ONE QUADRANT SHOWN
I-BEAM
SUPPORT
PERIPHERAL
ELECTRONICS
Two halves of UT will retract from beam pipe
PIGTAIL
CABLES
FRAME
STAVES
BOX
BEAMPIPE
APERTURE
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

36. Summary & plans

37. Summary and Plans
LHCb Upstream Tracker is comprised of four planes of silicon sensors which are supported by integrated stave structures The UT stave has been designed using FEA model techniques, prototypes, component testing, etc. Many components have been tested and qualified, some work still remains Module design has been refined although several issues remain Frame and box design is progressing, although the end of stave region is dense and challenging
Phase 1 construction planned to start end Summer 2015
Bare stave construction
Will soon freeze all mechanical parameters and design choices necessary to begin construction
This will allow us to construct all bare staves, which frees manpower and resources for subsequent phases
Phase 2 construction will start when data-flex cables become available
Phase 3 construction will begin when modules (sensors, hybrids) are available
Assemble at CERN in 2018
Install in IP8 in 2019
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

38. – FIN –
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015

39. Construction Clean Room
HVAC
Sealing
Electrical
Assembly
Lighting
Services
Testing
Lots of work in progress on clean room…
R. Mountain, Syracuse University FORUM on Tracking Detector Mechanics
6/16/2015