Upgrade tracker of the HL-LHC Sergio Dez Cornell, Berkeley Lab - - PowerPoint PPT Presentation

Silicon strip staves and petals for the ATLAS Upgrade tracker of the HL-LHC

Sergio Díez Cornell, Berkeley Lab (USA), On behalf of the ATLAS Upgrade strip tracker Collaboration

HSTD-8, Taipei, Taiwan, Dec 5th-8th, 2011

Motivation: ATLAS Phase II Upgrade (HL-LHC)

 Numerous challenges for silicon sensors on ATLAS Phase-II Upgrade

• Higher granularity to keep same low occupancy
• Higher radiation tolerance to deal with increased radiation environment
• Novel powering solutions to power efficiently x7.5 more channels
• Maintain low cable count to keep detector performance
• Reduce cost per sensor to cover larger area (~ 200 m2)

 Replacement of ATLAS Inner detector by an all-silicon tracker:

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

2

Strips tracker:

3 layers of short strips (2.5 cm) staves 2 layers of long strips (9.6 cm) staves 10 disks of endcap petals

Si tracker (Utopia Layout)

300 cm 75 cm

Stave concept layout and current prototypes

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

3

1.2 m 12 cm Ti coolant tube Carbon honeycomb Carbon fiber facing Readout ICs Si Strip sensor Kapton flex hybrid Cu bus tape

Barrel strip stave (short strip version):

• Designed to minimize material
• Shortened cooling paths
• Module glued to stave core with embedded pipes
• No substrate or connectors, hybrids glued to sensors
• Designed for large scale assembly
• Simplified build procedure
• All components testable independently
• Aimed to be low-cost
• Minimize specialist components

Short strip module:

• 1 n-in-p strip sensor with

4 x2.5cm strips

• 2 hybrids, each with 10

ABCN130 (256 ch) + 1 HCC/hybrid

• Binary readout
• Current prototypes:

ABCN250 (128 ch/chip) + BCCs

“Stavelets”:

Stave cross-section:

• Stave prototype with 4 modules per side
• Single-sided stavelets (serial and DC-DC powered)

already built and under test at RAL[1]

High T conductivity foam

Stave/petal powering

 LV: Two powering distributions under study for n hybrids, each with current I  HV: Parallel power limited by cable reuse and/or material limitations

• HV rad-hard switching for multiplexing under study recently (early stage)[2]

 Current module and stave prototypes have proven to be a powerful test bench for the different powering options considered

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

4

……

Constant current source

1 2 3 4 5 6 n-1 n

Constant voltage source

1 2 3 4 5 6 n-1 n

……

+

• Serial powering
• Total current = I
• Different GND levels per hybrid
• AC coupling of data lines
• Bypass protection required
• DC-DC powering
• Total current = n·(I/r*)
• Switching system
• Can be noisy
• High mass

*r = voltage conversion ratio

Other components of the stavelet prototypes

 Basic Control Chip (BCC) boards for data I/O (1 per hybrid)

• AC coupled multi-drop system LVDS reception
• Generates 80 MHz DCLK and handles 160Mb/s

multiplexed data from each hybrid

 Serial powering: Power Protection Board (PPB)[3]

• Fast response and slow-control bypass of modules

within an SP chain

• Allows alternate SP shunt circuits
• Excellent performances demonstrated on SP stavelet
• SPP ASIC submitted Aug 2011

 DC-DC powering: buck DC-DC converter

• Custom low-mass inductor and shield[4]
• AMIS 4 ASIC:
• Over current, over temperature, input under-

voltage, and soft start state machine for reliable start-up procedure[5]

• New prototype circuits underway

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

5

All hybrids on V = 22.7 V, I = 5.09 A Slow control disables odd hybrids V = 12.7 V, I = 5.09 A 39x6 mm2 13x28 mm2 AMIS4

Stave modules production and tools

 Scalability for large scale production even at prototyping stage

• Panelization of laminated hybrids
• Designed for machine placement of passives

and solder reflow

• Tools developed for controlled gluing and wire

bonding of ABCNs

• Conservative design rules for high yield and

volume, and low cost

• Final hybrids testable on panels, ready for

module assembly

• Diverse tools developed for uniform gluing of

hybrids to sensors

• Numerous options investigated: glue spread on

sensor or hybrid backplane, different glue stencils,…

• Optimized glue thickness for best module

performances: ~ 120 μm

• Automated wire bonding of ASICs to sensor and

hybrids to test frames

• Fully testable modules, ready for stave assembly

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

6

Stave module testing

 PCB test frames: cheap and flexible test benches for testing

• Different power configurations, G&S, added circuitry …

 DAQ system for stave modules and stavelets: HSIO

• Generic DAQ board (ATCA form factor) with single (large)

Virtex-4 FPGA for data processing & connection to controller PC

• Interface board: connectors & buffers for connectivity to FEE
• Currently supports up to 64 streams (>64 streams with larger

FX100 FPGA in future)

 Upgraded sctdaq software

• Allows standard 3ptGain, Response Curve, Noise Occupancy, DT

Noise,… on ABCN-250 modules

 Expected noise performances for parallel, serial, and DC-DC powered modules

• Similar ENC noise performances obtained at the different sites

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

7

Berkeley, serial Freiburg, serial Liverpool

Stave module construction and test

 Numerous institutes involved in the construction and test of stave modules and stavelets[6]

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

8

Up to 31 modules built so far (Nov 2011)

Proton irradiations of stave modules

 Irradiated at CERN-PS

• 24 GeV proton beam scanned over

inclined modules

• Module biased, powered, and clocked

during irradiation

• Up to 2x1015 cm-2 reached
• Sensor and module behave as expected
• Noise increase consistent with shot noise

expectations

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

9

Slide borrowed from T. Affolder, TIPP2011, June 2011

Stavelets

 Stave prototypes with 4 modules per side  Sensors directly glued to bus tape with “soft” glue for easy module replacement

• r removal

 Key test bed for electrical testing

• Powering, protection, G&S, …

 Single-sided serial and DC-DC powered stavelets built and tested so far

• SP stavelet tested with custom constant

current source (0-6A, OVP), excellent performances[7]

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

10

Custom Cu bus tape Power and PPBs EOS board EOS board BCCs

SP stavelet DC-DC stavelet

Power and Buck DC-DC converters Custom Cu bus tape BCCs

Stavelet bus tape layout

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

11

SP Trace Layer (Cu)

LVDS Clock/Command/Data & NTC SP Current Return HV

11

100μm track/gap over 40cm (1.2m)

SP shield Layer (Al) For DC-DC, the power section of the SP tape is cut off and replaced by a custom section

Slide borrowed from P. Phillips, TWEPP2011,Sept2011

Electrical tests on stavelets

 ENC noise close to noise on individual modules for both stavelets

• Approximately ~ 20e higher in both cases
• SP stavelet: PPB and bypassing hybrids does not affect noise performances

 Double Trigger Noise clean at 1 and 0.75fC with appropriate current routing

• Slightly better DT Noise performances at 0.5fC for DC-DC stavelet

 Still work in progress[1]

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

12

H0 H1 H2 H3 H4 H5 H6 H7

Column

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

ENC

661 623 628 675 650 636 697 760 687 646 640 666 680 661 624 656

DTN @1.0fC DTN @0.75fC DTN @0.5fC

130 40 1 58 3 1 255 1181 32 4 56 102 50 26 50 237 H0 H1 H2 H3 H4 H5 H6 H7 Column 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 dENC 8 1 27 26 11 2 17 26

• 10
• 9

28 31

• 26
• 23
• 2
• 2

DTN @1.0fC DTN @0.75fC DTN @0.5fC 1 6 36 18 5 12 38 12 2 4 9 4

Serially powered stavelet DC-DC powered stavelet

Stave material estimates

 Stave material estimates for 130 nm stave[8, 9]:

• Based on as-built stavelets
• Titanium cooling tube: 2.2mm OD x 0.14mm wall
• Tapes contribution could be significantly reduced (~50%) by removing Al

screen + one glue layer: under investigation

• Sensor dominates module material (~ 63%)
• Power components will add 0.03 - 0.15 %X0, depending on power scheme

(first approximation: changes in bus tape not considered)

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

13

%X0 Stave core 0.55% Bus tapes 0.30% Modules 1.07% Module to stave adhesives 0.06% TOTAL 1.98%

Stave core Bus tapes Modules Module to stave adhesives

Modules 54% Stave core 28% Tapes 15% Adhesives 3%

Endcap petals: Petalet program

 The endcap petal follows closely the barrel stave design  First petal cores already been produced  First endcap hybrids (ABCN-250 ASICs) produced and tested  Petalet prototype underway

• Combines innermost radius sensors and

region where petal splits in 2 sensor columns[10]

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

14

“Petalet” Endcap hybrid

Conclusions

 Stave program has shown significant progress  Module prototypes built and shown to work after irradiation at higher fluences than expected on the Si tracker  Both LV powering architectures being studied in detail with stavelet prototypes  Up to 20 groups involved in the module/stave/petal construction and test  Up to 31 modules and 2 single-sided stavelets, with both powering schemes implemented, have been built and tested so far, more underway:

• Double-sided stavelets at RAL
• Stavelets at other construction sites
• Petalets

 Full-size, next generation stave prototypes will be designed and built as soon as ABCN-130 ASIC is ready (6 months from now?)

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

15

Thank you!

 References

[1] P. Phillips, Stavelet status, ATLAS Upgrade week, CERN, Nov 2011 [2] D. Lynn, Possible Approaches to HV Distribution to Atlas Strip Staves, ATLAS Upgrade week, CERN, Nov 2011 [3] D. Lynn et al., Serial power protection for ATLAS silicon strip staves, NIM-A 633, pp. 51-60 (2011) [4] G. Blanchott, DC-DC converters: gained experience, ATLAS Upgrade Week, CERN, Nov 2011 [5] S. Michelis, DC-DC powering ASICs, ATLAS Upgrade week, CERN, Nov 2011 [6] S. Wonsak, Stave module status, ATLAS Upgrade week, CERN, Nov 2011 [7] J. Matheson, Progress and advances in Serial Powering of silicon modules for the ATLAS Tracker Upgrade, JINST 6 C01019, 2010 [8] T. Jones, Strip stave radiation lengths, Local Support Working Group (LSWG) – Mechanics, Berkeley, Sept 2011 [9] A. Affolder, Material study , ATLAS Upgrade Week, Oxford, March 2011 [10] I. Gregor and C. Lacasta, The petalet, ATLAS Upgrade week, CERN, Nov 2011

 Backup slides:

• Radiation hard n-in-p short strip sensors
• Thermo-mechanical stave demonstrator
• Short strip module
• Stavelets

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

16

Radiation-hard short strip sensors

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

17

Slide borrowed from T. Affolder, TIPP2011, June 2011

Thermo-mechanical stave demonstrator

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

18

Slide borrowed from T. Affolder, TIPP2011, June 2011

Short strip module

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

19

Stavelets

06/12/2011

• S. Díez Cornell, HSTD-8, Taipei (Taiwan)

20

Serial power: DC-DC power: