Page 1
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
LHCb Vertex Detector System:Status Report
J.F.J. van den BrandSubatomic Physics Group, VUA - NIKHEF
• Milan design
• Optimized design • mechanics
• vacuum system
• cooling system
• Summary
Page 2
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Mechanics: “TP” design
Side flange
Bending hinges
Detector support and cooling
Bellows (22000signal wires)
Support frame
Si detector
moves by 30 mmonly two positions:open or closed !!
See LHCb 99-042/VELO
top half = bottom half
Page 3
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Milan design
VELO Design:
• Single flange • XY table• CO2 cooling• WF suppressors• Second. vacuum• Studied
• assembly• alignment
• To do• further design• FEA
Page 4
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Detector and support frame
• both halves on same side• VD easier to mount and position in the tank• install complete VD at once• the two halves are no longer interchangeable
Page 5
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Vacuum vessel
• Employ top flange
• Easier installation• Shorter cables• Length 2000 mm• Width 1200 mm
Page 6
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Top flange
770 19
00
Lif
t 60
0
• Length 1500 mm• Distance from
ceiling 1900 mm• Install using wires• Baking to 60o C?• Regenerate NEGs
after every access to Si detectors
Page 7
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Optimized System
1250
1820
2150
1450
Total length: 1750
• Two detector boxes
• Baking up to 150o C
• Decouple access to Si detectors
Page 8
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Support system
bellows
chain/beltcooling/bake out
gearbox 1:40
ball spindle 16x25 mm
linearbearing 2x
30 +5
motor
• Microswitches at out position
• LVDTs• Steel frame• Alignment:
– 2 planes
– 3 points each
– define IP
Page 9
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Support system
• Alignment pins for reproducible coupling
• reproducible positioning
Page 10
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Vessel Installation• Move bellows to
in-position• Install vessel
from top• Align vessel• Mount vessel to
frame• Mount bellows• Pump-outs visible
Page 11
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Install 2nd-Vacuum Vessel• Remove upstream
flange• Need 2 m access• Rectangular bellows
– 60 mm stroke– normal 35 mm– lateral 6 mm
• Fabrication– Palatine, Bird– Calorstat, MB– cost
Page 12
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Vacuum vessel / Positioning system
• Moving parts not in vacuum
• Thin vacuum container
• Special bellows construction
Secundaryvacuum
Primaryvacuum
Page 13
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
After installation• Detector system
separated from vacuum system functionality
• Mount positioning system to detector housing
• Install– pump-out, valves
– turbos, damping
Page 14
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Connect inner system to motion drives
• Mount M8 through side flanges
Page 15
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Detector Installation
• Install detector halfs from sides
• Decouple detectors from box
• Tooling needed
Page 16
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
VELO Assembly
• Detectors mounted
Page 17
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Wakefield suppressors• Mount screens
after mounting 2nd vacuum container
• Mount through top flanges– seal with view
ports?
• Upstream: mount with large flange off WF screens
420
910
IP
Page 18
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Wakefield suppressor: downstream• Up/downstream
suppressors are identical
• Material: CuBe• Length: 179 mm• Thickness: 100
m• 16 segments• Mounting to box
non-trivial
Page 19
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Wakefield suppressors• Segments deform
differently during movement
• Coating needed on suppressors
• Press-fit to beam pipe structure
• Anneal CuBe, deform, harden at 400o C
Page 20
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Detector Mounting
Install Modules
3D alignment
Mount References
Page 21
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Thin Vacuum foil
• Beryllium expensive: k$ 500 per container
• Aluminum– welding 250 m Al
is possible– press-shaping
being developed
• FEA ongoing
Page 22
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Thin Vacuum foil
• Labour intensive: press, anneal, etc.• welding 250 m Al is possible• Extensive prototyping program
Chiel Bron
CP?!
Page 23
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Thin Vacuum foil• Increase radius: 10 20 mm to avoid
folding• Crystal structure is affected• Employ Al with magnesium alloy• Deform at higher temperature: 150 - 200o
C
Page 24
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Foil design ongoing (continued)
Page 25
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Foil design ongoing
Page 26
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Control of Vacuum System
• Group active with experience at former NIKHEF accelerator
• Propose meeting in Q1 2001
Page 27
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Vacuum Tests
• Self-regulating valve behaves as advertized
• Various gas flows have been characterized
Page 28
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Vacuum constraints
LHC:• beam life time: static density of 10-7 mbar 2 m (H2 300K) 0.01 % of LHC limit for integrated density ( 2.7 106 cm 1.6 109 molecules/cm3 )
• beam stability: dynamic effects must be taken into account
LHCb:• 10-7 mbar 1.2 m (H2 300K) 1.5 % of LHCb nominal luminosity
Difficult to achieve with silicon detectors, electronics and signal wires directly in LHC vacuum ! differential pumping.
(rough!)
See LHCb 99-045/VELO
Page 29
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Static pressure in VDConsider outgassing by: assuming outgassing rates of:
(mbar • l • s-1 • cm-2)
11 m2 Kapton (signal wires, pumped 40 hours) 10-7 H2O 2.3 m2 Al housing (per half) 10-10 H2
1.5 m2 bellows (per half) 10-9 H2
8 m2 SS vessel 10-10 H2
Pumps in detector volume: 140 l/s (per half) H2OPumps in tank: 4000 l/s H2
Bypass tube: 200 mm 4 mm pumped in the middle.
Calculate using a static flow model.Result: 1•10-4 mbar in detector volume
1•10-8 mbar in VD tank2•10-8 mbar • l • s-1 from det. vol. to VD tank
Page 30
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Summary table:(Data are approximate. QLHCb_total = estimate for the full vertex detector, i.e. both halves.)
Item Outgassing rate of item QLHCb_total
[mbar l s -1] Kapton foil, after 40 hrs pumping 1 E-7 mbar l s -1 cm-2 n/a sample Kapton flat cable QPI 3 E-5 mbar l s -1 130 E-4 male/female pair of PEEK D-type 25-pin connectors 6 E-6 mbar l s -1 / pair 50 E-4 male/female pair of stand. D-type 25-pin connectors 1 E-5 mbar l s -1 / pair 100 E-4 Liverpool carbon-fiber Si support 1 E-8 mbar l s -1 cm-2 ~ 1 E-4
Outgassing measurements
Continue: measure all unknown outgassing rates of components in a detector station
Page 31
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Dynamic Vacuum
Beam-induced particle bombardment desorption, emission
Ions, photons, electronsenergies up to keV
• Local pressure runaway (ion/electron-induced desorption)• Local static charge increase (electron multipacting)
LHC beam instability
See Adriana Rossi’s presentation
Page 32
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Dynamic Vacuum (continued)
Perhaps a solution:
use coating of surfaces by Tiadvantages: low SEY , low , local pumping
Design issues: • better surfaces ? (NEG ?)• in-situ coating required or not ? • thickness of layer needed ?• what re-coating rate ?• affordable cathode temperature in-situ ? • wake field / RF properties ?• side effects ? (peeling, ...)
We need , for:• different materials • surface conditions (un)baked, saturated, activated, etc. • different impact energy spectra
Data available only in a few months ! (Mahner et al.)
Page 33
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Cooling system with mixed-phase CO2
Phase diagram CO2
1
10
100
-80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50
Temperature [°C]
Pre
ssur
e [b
ar]
vapor
liquidsolidgas
critical point
triple point
*
Page 34
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
CO2 Cooling Tests
Cooling system
-30o C
40 W/module
Page 35
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
CO2 gas-liquid storage tank57.3 bar at 20 C
CO2 supply line
compresssor
P [W]
P [W]
P [W]
flow restrictions
cooling lines
gas only
pressure (temperature)regulating valve
heat to 20 C
Mixed-phase CO2 Cooling systemSee LHCb 99-046/VELO
cool to 20 C
supply lineexpansion valve
Page 36
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
CO2 Cooling Tubes
Cooling tubes
1.1 (0.9) mm S.St.
Welding and brazing
Page 37
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
FEA ongoing
Page 38
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Tests ongoing
Page 39
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
RF tests at NIKHEF
Simulation with MAFIA
First 3 measured eigenmodes:•220 MHz•270 MHz•380 MHz
Outlook:•Eigenmodes•Short range effects; Z/n•Electric field inside secondary vacuum
Picture 1 of tank removed
Picture 2 of tank removed
Page 40
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Wake field Suppressor
Central cooling line
Temperature sensors (2 per station, 4 wires per measurement
Page 41
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Detector Modules
Number of planes: 25
Discuss
Liverpool delivers modules
40 W, 1 m cable, 50 % isolation thickness, 10 - 15 K T, radiative cooling
44 pins,440 / module
Page 42
vrije Universiteit amsterdam CERN, November 27, 2000 VELO System J.F.J. van den Brand
Summary• Design is based on secondary vacuum system
– Beryllium option: costly and uncertain
– needs approval for TDR
• Current design – allows baking up to 150o C
– decouples Si detectors from primary vacuum system
– employs venting with Argon
– cooling based on CO2 in gas-liquid phase
• Self-regulating valves behave as advertised
• Wakefield excitation under study
• Need information on dynamic vacuum effects
• Propose meeting on control issues (e.g. NIKHEF)