LHCb Silicon Tracker hybrids - PowerPoint PPT Presentation

About This Presentation
Title:

LHCb Silicon Tracker hybrids

Description:

LHCb Silicon Tracker hybrids – PowerPoint PPT presentation

Number of Views:32
Avg rating:3.0/5.0
Slides: 22
Provided by: lhcbPhy
Category:

less

Transcript and Presenter's Notes

Title: LHCb Silicon Tracker hybrids


1
LHCb Silicon Tracker hybrids
  • overview LHCb
  • ST Detectors IT TT
  • hybrid design
  • procurement testing
  • words on vendor issues
  • conclusions

Frank Lehner U Zurich LECC Heidelberg Sep 14,
2005
2
(No Transcript)
3
(No Transcript)
4
Detector overview
IT ladder
  • IT-detector module
  • one (10 cm) or two (20 cm) sensor ladders, 320 or
    410 ?m thick silicon with pitch 200 ?m
  • hybrid at ladder end, supported on aluminum
    balcony
  • TT-detector module
  • full length 160 cm
  • 500 ?m thick silicon with pitch 180 ?m
  • two half-modules w/ 7 sensors each are mated to
    full length
  • full module has at both ends hybrid stacks (out
    of detector acceptance)
  • hybrids supported on carrier pieces
  • further readout via 5 m long cables to service
    boxes (digitization, see talk by A. Vollhardt)
  • Total electronic channels 280k
  • IT 336 readout units w/ 3 chips each
  • TT 280 readout units w/ 4 chips each

TT half module
5
Beetle chip
size 5.5 x 6.1 mm2
  • common development for LHCb silicon detectors
  • designed by ASIC lab Heidelberg NIKHEF to meet
    LHC requirements
  • fabricated in 0.25 ?m CMOS radiation tested up
    to 40 MRad
  • 128 channel low noise preamplifier w/ CR-RC
    shaper
  • analogue pipeline, programmable latency up to 4
    ?s
  • L0 trigger rate 1.1 MHz
  • multiplexed deadtimeless R/O within 900 ns
  • input pads four row staggering, effective pitch
    40 ?m

6
IT hybrid
  • three Beetle chips on flex (polyimide) PCB 130 x
    77 mm2
  • opening in bottom solder resist _at_ beetle
    positions for thermal and GND contact
  • ceramic pitch adapter
  • flexible pigtail
  • connector single row 60-pin, pitch 0.5 mm
  • PCB mounted on aluminum heat spreader
  • 386 hybrids needed (incl. spares)
  • 15 as pre-series delivered

7
TT hybrid
  • four Beetle chips on flex PCB 67 x 70 mm2
  • stack of up to three hybrids at both ends of a TT
    full module
  • three variants/flavor with different size
    necessary long/medium/short
  • lowest level aluminum nitride carrier
    pitchadapter to silicon
  • upper two levels copper plate pitch adapter to
    flexible interconnect
  • robust dual row 80 pin connector w/ 1 mm pitch
  • 322 hybrids needed (incl. spare)
  • 15 as pre-series delivered

8
Pitch adapters
  • pitch adapters (PA) as fan-ins from 200 ?m
    silicon strip pitch to 40 ?m?narrow Beetle input
    pitch
  • all three types of PAs for IT TT
  • sizes from 77 x 11 mm2 to 94x24 mm2
  • 0.25 mm Alumina in thin film
  • metal TiW (50 nm)/Au (2 ?m)
  • Al wire bonding very satisfactory
  • no severe degradation after accelerated aging
  • fan-in structure is mirror image of fourfold
    staggering of Beetle
  • Trace width/separation
  • less than 10 ?m in limited areas
  • 30-40 ?m trace width elsewhere

Vendor RHe Microsystems GmbH
9
Flexible PCBs
  • four layers (from top routing/routing/GND/VDD)
    flex PCB polyimide 50 ?m, Cu 18 ?m
  • sizes 67 x 60 mm2 (TT) up to 130 x 77 mm2 (IT)
  • Au/Ni surface for Al bonding
  • min. trace width/separation 75 ?m/75 ?m
  • solder resist on top (laquer), polyimide bottom
    layer
  • only one active component (Beetle)
  • passive components (all SMD) Rs, ceramic Cs,
    PT-1000, B2B connector
  • design core successfully used since 1st
    testbeam 2001
  • prototypes, pre-series and series PCBs
    successfully produced at Optiprint AG

10
Hybrid assembly
  • all assembly steps done at single company (RHe
    Microsystems)
  • procedures, tolerances, acceptance criteria
    defined in specs document
  • manual chip placement, accurate 20 ?m
  • aluminum wire bonding
  • 25 ?m wire for chip PA bonds
  • 30 ?m wire for flex PCB - chip
  • regular pull tests performed
  • thermal cycling 10x from -20ºC to 60ºC, 5
    minutes at each endpoint
  • test stand set up at vendor
  • test of basic digital and analog functionality of
    all beetles on hybrid
  • repair (chip replacement) still possible

11
Hybrid assembly
  • full documentation of which chip goes to what
    position on hybrid from wafer maps (-gt hybrid
    database)
  • excellent documentation of pull tests, electrical
    tests repair work

12
Hybrid testing pre-assembly tests chips
  • chip testing on uncut beetle wafer at MPI
    Heidelberg
  • setup wafer prober needle card DAQ
  • test of full functionality of analog and digital
    performance
  • power consumption
  • I2C addressing tests
  • control logic
  • sequential pipeline scans
  • pulse shape scans
  • ()
  • tight quality acceptance criteria
  • typical yield 90
  • 3x more good chips than needed

13
Hybrid testing pre-assembly tests PCB PA
  • flex PCB tests at vendor (Optiprint AG)
  • automatic optical inspection
  • electrical test for contact
  • only perfect PCBs accepted, no bad PCB found up
    to now (gt100)
  • specs (bond pad width, thickness etc.) defined in
    specs document
  • bond tests on lot basis (at hybrid assembly
    company)
  • pitch adapters (RHe Microsystems GmbH)
  • bond pull tests
  • electrical test with needle card (on spare pads)
  • complementary optical inspection
  • several shorted traces allowed prior to repair
  • laser repair of shorts final spec ? 1 short or
    interrupt
  • 70 w/o defect at all remaining have up to 3
    repairs

14
Tests during hybrid assembly
  • test stand set up at vendor (RHe Microsystems
    GmbH)
  • short and important test of basic analog and
    digital functionality
  • test performed before and after thermal cycling
  • 10x cycles 20C lt-gt 60C, 5 min. at each endpoint
    standard
  • test is done after assembly but prior to
    chip-pitch adapter bonds
  • easy detection of missing bonds
  • also full chip replacement still
    possible/feasible at this stage
  • so far no sign of infant mortality of 43 (43)
    thermally cycled TT-hybrids

15
Hybrid burn-in test
  • extensive hybrid testing program after delivery
    to MPI Heidelberg
  • characterize, document and grade hybrids
  • start with visual inspection
  • detailed R/O test
  • at start/middle/end of 48h long burn-in run at 1
    MHz rate
  • R/O tests similar to Beetle chip testing wafer
  • tests in box at room temperature, no thermal
    cycles
  • up to 16 hybrids (any combination of IT TT
    hybrids) can run at same time
  • analysis output linked into database

16
Hybrid burn-in test
  • full digital and analog tests (similar to chip
    wafer tests)
  • I2C addressing, power consumption etc.
  • raw noise baseline subtracted noise
  • mapping of pipeline - relative pipeline cell
    gain 5 variation
  • pulse shapes timing characteristics
  • ()
  • check power-on startup
  • run at 1 MHz trigger rate check Beetle
    synchronicity
  • check HV contacts monitor HV at 600V check for
    sparking effects during burn-in
  • check PT-1000, VDD, VDDD sense lines
  • yield 84 out of 92 hybrids passed on first go -gt
    analysis of failed hybrids not yet concluded

17
Problems
  • IT hybrid has openings under Beetle chip
    positions in bottom solder resist
  • filled with silver epoxy
  • connection to common GND
  • thermal conductive path to heat sink
  • problems on some prototype and pre-series hybrid
  • openings too large problems with insulation
  • potential risk of GND/VDD short repair with HV
    insulating cover
  • fixed for series production

18
vendor issues I
  • choice of right vendors is a big issue for
    small-scale projects with design features that
    are at cutting edge
  • we experienced large difficulties in hybrid
    prototyping phase which caused lots of delays
    fortunately, they were no real show stoppers of
    the project
  • pursued initially approach of two different
    companies
  • ceramic pitch adapter (PA) production in thin
    film technology
  • hybrid assembly, SMD loading bonding
  • problems with coordinating quality standards and
    agreeing on common specs among companies like
    pull strength on PA etc.

19
vendor issues II
  • vendor qualification process of first company
    (dedicated PA production) suffered significantly
    from low quality and large delivery delays of
    ceramics
  • mainly due to overestimation of own abilities at
    company
  • partially due to design/layout that could have
    been made simpler
  • second (dedicated SMT) company lost interest in
    continuing after successful prototypes
  • due to low quality ceramics/PA
  • due to technical difficulties (4x staggered
    chip-PA bonding)
  • luckily, we could identify a single company
    capable of PA production, SMT and wire bonding
    in-house (RHe Microsystems GmbH)
  • developed tight relationship customer - company
    from day one
  • worked/cooperated together on several design
    modifications to simplify fabrication process
  • rapid successful prototypes on very short time
    scale
  • proof of principle achieved, smooth transition to
    series production

20
vendor issue III
  • our lessons
  • start with vendor qualification as early as
    possible
  • be prepared for back up solutions
  • work closely together with companies
  • what can they do? what is achievable at
    reasonable yield/cost? what do they recommend?
  • what can be changed/modified in the
    design/process to make it simpler/fail safer?
  • cultivate a tight connection close relationship
  • understand early if the company shows real
    interest in the project. This interest should go
    beyond commercial interests

21
Summary
  • hybrid assembly outsourced to single company
    capable of pitch adapter production, SMT mounting
    and aluminum wire-bonding
  • hybrid pre-series (15) for IT TT detectors has
    been delivered just recently
  • thorough hybrid testing program is installed
  • 84 out of 92 tested hybrids passed our testing on
    first go
  • testing not yet fully concluded - remaining
    hybrids being debugged
  • expect to give go-ahead for series production
    very soon
Write a Comment
User Comments (0)
About PowerShow.com