Module Bonding Working Group

1
Module BondingWorking Group

• Salvatore Costa
• Università di Catania and INFN Sezione di
  Catania

2
Bonding Facilities
3
Bonding Machines
Center Bonding Machines available to CMS
Bari Delvotec 6400
Catania Hughes 2470-II (11)
Firenze Delvotec 6400
Padova KulickeSoffa 8090, KulickeSoffa 4123, Hughes 2470-V
Pisa KulickeSoffa 8090
Torino Delvotec 6400
Fermilab KulickeSoffa 8090 (3)
Santa-Barbara KulickeSoffa 8090, KulickeSoffa 8060
Aachen HesseKnipps 710M
Hamburg Delvotec 6400
Karlsruhe HesseKnipps 710M
Strasbourg Delvotec 6400
Vienna KulickeSoffa 4123
Zurich Delvotec 6319, Delvotec 6320
4
Bonding Times per Module

• Pre-bonding visual inspection 10min
• ARC test 15-30min
• Bond 5-35min
• 5min for KS
• 10min for Delvotec and HesseKnipps
• ? Double for 2-Sen Modules
• 35min for Hughes.
• Post-bonding inspection 10min
• Pull test if required (pull, cleanup, rebond)
  20min
• DB entry 5min
• TOTAL 45-90min excluding pull test
• NOTES
• ARC test times vary with Vpsp adjustment
• Some op may be performed in parallel by different
  people

5
DB Interface
6
Bonding WG Web Site
7
Production Rate
8
Production Rate
9
Bonding Failures Repair Actions
10
Maintenance, Failures, Downtime

• Center Maintenance schedule Failures Downti
  me
• TIB
• Bari             yearly, no
  contract      No                      
  N/ACatania       on demand,no
  contract     Once                   1 month
  ()Firenze           yearly by contract       
  2 times             2 months ()
  Padova            regularly by bond op.   
  No                      N/APisa             by
  bond op. as needed    No                     N/A
  Torino            yearly, no contract     
  No                       N/A
• TECAachen          regularly by
  bond op.   No                    N/AHamburg      
    on demand,no contract    Once     3 days
  Karslruhe        on demand,no contract   
  Once     1.5 daysStrasbourg   yearly by
  contract       No                     N/AVienn
  a            yearly by contract      
  No                     N/AZurich          
  on demand,no contract    Once                 
    2 weeks
• TOBFNAL             4/year by contract      
  Damage by None power
  failures
• UCSB  4/year by contract 1/machine 1-2
  week/machine
• () Attempts (failed) to fix in house, to save
  money, made because prod was stopped            

11
Bonding Failures/Rep Actions

• Bonding failure detected during bonding or at
  post-bonding inspection
• Wire wont stick
• Bond pops up (lift off)
• Foot is misplaced and causes a short somewhere
• We repair these right away if we can and/or list
  as unbonded the strips that are such as a result
  of the failure or of the repair action.
• (Bonding) failure detected during post-bonding
  ARCS or LT tests
• Pinhole (caused by bonding/handling accident or
  pre-existing and either undetected in Sensor test
  or on untested Sensor)
• Short (caused by bonding/handling accident or by
  misplaced foot)
• Open (bond popped up)
• Same action.
• Electric failure detected with ARCS at any stage
  (pre-bonding, post-bias-pre-readout-bonding,
  post-bonding or LT tests which dos not exhibit a
  clear link to bonding
• Increased LV current
• Anomalous IV
• We post the problem for the Debug/Repair Center
• Electric failure detected during post-bonding
  ARCS or LT tests, which might be traced back to a
  bonding failure
• Example pseudo-pinhole At post-bonding ARCS
  test, a PHL- is found. We pull the readout bond
  as prescribed for pinholes but the test keeps
  saying PHL-, so the Hyb ch is damaged. Since it
  was OK at ARCS test on arrival perhaps we killed
  it with bonding.

12
Bonding Failure Rates

• Global percentage of damaged channels (Hyb ch
  and/or Sen strips) because of bonding failures
  lt0.01
• Based on a compilation of unrepairable bond
  failures reported by individual centers on
  20Apr2004. At that time 1180 Modules had been
  bonded (800,000 bonds).
• Since then reports at Meetings have indicated
  improvements, therefore although a more recent
  quantitative estimate is not available, it can
  only be less.
• Global percentage is the average between 0 from a
  few centers (such as Pisa) were not a single bond
  failure leading to loss of channels has been
  observed and 0.08 in Catania see next slide
  for analysis of failure types where the oldest
  bonding machines, also lacking modern amenities
  such as automatic correction of bond parameters
  with varying substrate quality, are.
• Not a single module lost yet (i.,e. no module
  with gt 2 affected channels) because of bonding
  failures out of 2,700 Modules bonded so far.
• Only 1 module unbondable for dirty Pitch Adapter.

13
Bond Failure Analysis (Catania)
53 Modules 34048 bonds Unrepairable failures Unrepairable failures Unrepairable failures Unrepairable failures
53 Modules 34048 bonds modules affected channels affected
Totals 10 19 29 0.08

Machine failures 7 13.2 14 0.04
Bond foot misplacement 2 3.8 9 0.026
to stick 4 7.6 4 0.012
Pinholes (caused by bonding?) 1 1.9 1 0.003
Operator mistakes 3 5.7 15 0.04
Mishandling 2 3.8 7 0.02
Wrong ref
Wrong par (lifts, pinholes) 1 1.9 8 0.02
14
The chief Bonding quality indicatorthe Pull
Test
15
Module Pull Tests Performed
Modules pull tested after 01Dec2003 (at last
Mtg)
Center PA TA PA-Sen Sen-Sen
Bari 98 (83 ) 101 (86)  n/a
Catania 51 (51) 51 (51) n/a
Firenze 9 (9) 5 (5)  n/a
Padova 18 (18) 99 (99)  n/a
Pisa 18 (10) 46 (37)  n/a
Torino 119 (114) 46 (46)  n/a
Fermilab 121 (119) 35 (33) 35 (34)
Santa-Barbara 51 (51) 58 (58) 76 (76)
Aachen 37 (0) 41 (0) 41 (0)
Hamburg 69 (62) 50 (45) n/a
Karlsruhe 0 (0) 20 (20) n/a
Strasbourg 20 (20) 65 (65) 65 (65)
Vienna 55 (19) 91 (28) n/a
Zurich 0 (0) 42 (38)  n/a
16
Pull Test Results (Nov 04)
Not angle-corrected yet!
Average forces (g)
5g
BA CT FI PD PI TO FL FL SB
SB AA AA HA KA ST ST VI ZH
.
ss ss ss
ss
17
Pull Test Results (Jul 04)
Not angle-corrected yet!
5g
BA CT FI PD PI TO FL
FL SB SB HA KA ST ST VI
ZH .
ss ss
ss
18
Pull Test Results (Jun 04)
5g
BA CT FI PD PI TO FL
FL SB SB HA KA ST ST VI
ZH .
ss ss
ss
19
Loop Documentation

• Procedure Uses a few of the identical
  symmetrical bonds we place on the PA Tests areas.
• Document the original bond
• Straighten a bond to (near) triangular shape
• Measure the angle at the bond foot
• Determine correction factor for pull strengths

20
1. Document the original bond

• Gently push bond over with the pull tester hook
  until it rests flat on the PA surface, but taking
  care that its shape is not significantly altered.
• Take snapshot of bond under measuring machine.

21
2. Make bond triangular

• Gently pull bond up at mid-span with the pull
  tester hook until it takes a (nearly) triangular
  shape (figure below).

• Gently push bond over with the pull tester
  hook until it rests flat on the PA
• surface, but taking care that its now
  triangular shape is not significantly
• altered.

22
3. Measure angle at bond foot

• Take snapshot of triangularized bond.
• Measure angle using measuring machine.

23
4. Determine correction factor (1)
24
4. Determine correction factor (2)
Correction factor a 1 / (2 sin?)
q (º) corr. factor a
10 2.88
20 1.46
30 1
40 0.79
50 0.65
60 0.58
70 0.53
80 0.51
90 0.5
For symmetrical bond
25
Module Gradingfrom Bonding
26
Module Grading from Bonding (1)

• Grading based on unbonded strips
• First flag in BONDSTATUS_VAL vector
• 0 if n_unbonded lt1
• 1 if 1 - 2
• -1 if gt2
• These are consistent to grades A,B,C of ModTest
  and the module is faulty with gt2 unbonded
  strips.

27
Module Grading from Bonding (2)

• Grading based on
• mechanical quality of the bonds
• Criteria to generate this second flag
• Bias Bonds (Spec 5, 15 for TIB back)
• gt2 grade A (flag 0)
• 2 grade B (flag 1)
• lt2 grade C (flag -1)
• critical readout bonds definition list of
  critical by Alan
• 2 grade A (flag 0)
• gt 2 grade B (flag 1)
• gt 20 grade C (flag -1)
• Operator judgment on all other bonds guidelines
  to issue this judgment by Alan
• A, B, C.
• These criteria will be ored together to generate
  flag.

28
Bonding Specifications Procedures
29
Bonding Specs (1)

• Document on Bonding web site
• cms.ct.infn.it/bonding/bondspec.htm
• Contents
• Bonding Specifications ?
• Specifications on Bonding Jigs
• Specifications on Module Carriers
• Specifications on Module Components and Assembly
  Relating to Bonding
• Specifications on Working Environment
• Specifications on Final Module Bonding Quality
  Acceptance

30
Bonding Specs (2)

• wire thickness and type
• number of bias bonds
• loop height and form
• tail length
• accuracy of placement in X and Y
• pull strength
• break on pull failure
• failure rate
• number of bonds with one tool

31
Bonding Procedures

• Bonding Proc Document on Bonding web site
• cms.ct.infn.it/bonding/bondproc.htm
• Reference doc in 43 steps
• Has many pictures (mainly for TIB at the moment)
• Procedure to perform the 50th wire Pull Tests
• cms.ct.infn.it/bonding/pullproc.htm

32
Skip rules

• Leave unbonded
• All strips in bad-IDIEL list (considered
  pinholes)
• All Isolated bad CAC (believed to have high
  chance to develop into pinholes with irradiation)
• All but lowest in a bad CAC chain (believed to
  represent shorts)
• CAC Example 3 34 35 36 37 ? skip
  isolated 3,
• skip all but lowest in 34-37 chain, or bond
  only 34

• Leave unbonded all strips in bad-IDIEL list
  (although 16 are good), i.e. do not bother
  looking at their actual IDIEL value.
• Bond all isolated bad-CAC
• For the consecutive bad-CAC, compute rC and
  consider them members of a short if rCgt1.5. Then
  bond only the first strip in the short.
• If rclt1.5 bond them, as if they were isolated

New rule
Old rule