Bonding WG

1
(No Transcript)
2
Bonding web site, renovated

• Finished renovating Bonding WG web site, a
  tasks handed down to me by Alan last July 2003.
• Both site infrastructure and contents renovated
• Uniform graphic style
• Automated handling of bookkeeping info (such as
  Meetings archive, mailing lists and now also
  Center facilities).
• Revised and updated documents. Organized into
  Procedures and Tech Info sections.
• Some docs were in need of major updates,
• for which Group consensus is in order
• (discussed later during this Meeting)

3
Bonding web site, renovated
4
Center facilities archive

• Uses info gathered by Alan with a questionnaire
  sent to Institutes on 26 Apr 2000
• I will ask all Level-3 Responsibles to update the
  info
• Mechanism similar to Bonding I/F
• Update older info about existing machines
• Enter new info for a new machine
• Access controlled
• username and pw will be distributed to L3s
• I hope working together with the L3s we can keep
  this info up-to-date at all times from now on

5
Center facilities archive
Machines are identified by the DB tool_id
6
(No Transcript)
7
Official Documents
8
Procedure Document

• Source
• Based on Alans original 24Sep2002 doc
• Includes
• additions specific for TIB/TID developed in the
  INFN community
• items from the Oct 23, 2003 official doc of the
  Module Test WG concerning all tests correlated
  with Bonding
• my own additions mainly concerning interactions
  with the DB.
• Technical structure
• Global and comprehensive (the Reference Bible
  of CMS Module Bonding), i.e. has all procedures
  for all Module types
• Paragraphs referring only to specific Tracker
  components (TEC, TOB, TIB/TID) are coded for an
  automated program which can extract and present
  to the user a subdocument with only the relevant
  info
• Same technique can be used to extract info
  tailored to specific operations or audiences
  (Klaus comment)

9
Document approval path

• Advertised 2 weeks ago
• Gathered comments until yesterday (but only
  Susanne Kyre reacted)
• Still has missing/not yet available info
• Will go thru it now and get live comments, if any
• Formal approval.
• With approval it becomes a living document, open
  to changes as need arises with growing experience
  and to the addition of info missing or not yet
  available today
• (subject each time to the group approval, by
  e-mail unless revolutionary changes).

To doc ?
10
Carrier/Transport/TO plate
11
50th wire pull test

• Small addition (by S.Kyre)

12
Specs Document

• Source
• Based on Alans original 29 Oct 2000 doc
• Includes
• additions specific for TIB/TID developed in the
  INFN community
• Recent additions and changes by Alan
• my own additions mainly concerning specs
  implemented within DB I/F.

13
Document approval path

• (same as for Procedure doc)
• Advertised 2 weeks ago
• Waited for comments until yesterday (none)
• Still has missing/not yet available info
• Will go thru now and get live comments, if any
• Formal approval.
• With approval it becomes a living document, open
  to changes as need arises with growing experience
  and to the addition of info missing or not yet
  available today
• (subject each time to the group approval, by
  e-mail unless revolutionary changes).

To doc ?
14
(No Transcript)
15
Bad strips study conclusions (from last Meeting,
21 Oct 2003)

• 1) Practically all good-IDIEL strips have
  0ltIDIELlt0.1 nA
• 2) Most bad-IDIEL strips have higher IDIEL, but
  16 have the good
• 0ltIDIELlt0.1 nA
• 3) Having constructed the relative CAC
• rCCACOfBAdCACStrip/AvgCACOfAllGoodCACStripsIn
  ThatSensor,
• all isolated bad-CAC strips have rClt1
• 4) the consecutive bad-CAC have a complex distrib
  with 2 parts,
• one for rClt1 which is very similar to the
  isolated ones, thus
• indicating these are not shorts but (mostly)
  pairs of strips
• affected by the same defect which causes a
  low capacitance (a
• scratch?) and one with rCgt1, with peaks at 2,
  3, 4 and a marked peak
• at 2.6. These are probably shorts although I
  can't find an obvious
• explanation for the peak at rC2.6

16
Bad Sensor strips

• Presented study at ModProd Meet 23 Oct 2003
• Got response from Manfred, Sensor Group
  Coordinator
• The technical reasons why 16 of bad-IDIEL strips
  have just good IDIEL and rC has a peak at 2.5
  remains unexplained, but Manfred said they will
  investigate. As a general remark, they
  occasionally have uncertain results because of
  flaky probe contacts.
• However, we cannot put Bonding on hold until we
  get an explanation.
• Ambiguous cases account for only 0.038 of all
  tested strips
• Finally, an agreement was reached on strips to
  leave unbonded within a de-facto commettee made
  of Manfred, Alan, Tony and myself ?

17
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
18
New release of DB I/F (v. 3.2)

• Unfortunately, we will need a new release
• To implement new skip rule to produce new lists
  of strips to skip and corresponding red boxes in
  checkerboards
• In v. 3.1, if no test data by QTC Centers (e.g.
  if no rows with status reference) are found in
  DB for a Sensor, we suggest to bond all strips.
  However, we are moving to a sampling test mode of
  sensors ? most sensors will never have rows with
  status 'reference but just 'valid (data
  inserted by Manufacturer, which we will simply
  have to trust), so I/F will look first for
  'reference' rows, then for 'valid' rows to
  implement skip suggestions, and finally suggest
  to bond all strips only if no rows whatsoever are
  found for a given Sensor.
• Release will also include 1 additional variable
  about Hybrid Quality Control (requested by
  Strasbourg).
• Center names will have to change (requested by
  TrackerDB Administrators).

19
(No Transcript)
20
Module Repair

• Originally, we were named the
• Bonding Repair Group
• It is possible Management assumes we should take
  responsibility for Module repair in a general
  sense, not just bonding-related
• Do we think we should?
• Which repairs besides bonding?
• Ive asked to put this question on the Agenda of
  the TPO Thursday, but I should also voice there
  the thoughts of this Group.

21
(Bond) Repair document

• Alan is working on a first draft
• Request has been sent to all in Mod Test for
  input on problem/cure cycles
• Alan Procedure for shorts that are found to be
  either on the PA or due to the PA-APV bonding
• If there is a PA short fixed, this will give the
  following "fault" in the readout test the
  channel which had the bond pulled will look like
  a standard open. The other channel will have
  about 2x the normal capacitance so will have
  higher noise. It will likely fail the noise cuts.
  This will also be true of a short on the first
  sensor which has been fixed in the "standard" way
  given in the module bonding procedures. Clearly
  the whole interpretation of the data from a
  bonded module becomes much more complex when
  there are these various faults and repairs.
• Tony Affolder has volunteered to contribute the
  experience they have made in Santa Barbara, e.g.
  on CMN problem/cure pairs

22
(No Transcript)
23
Center Compliance

• Results in DB from 7 Centers

Modules pull tested, separately per bond type
24
Module Pull Tests
25
Module Pull Tests
26
Module Pull Tests
27
Module Pull Tests
28
Module Pull Tests
29
Module Pull Tests
30
Module Pull Tests
31
Module Pull Tests
32
Module Pull Tests
33
Module Pull Tests
34
Module Pull Tests
35
(No Transcript)
36
Reports from Centers
Aachen Bari Catania CERN Fermilab Firenze Hamburg
Karlsruhe
Padova Pisa Santa Barbara Strasbourg Torino Vienna
Zurich
37
Catania

• On 5 Dec 2003 the INFN Consortium has agreed to
  re-designate Catania as Production Center.
• This means we are going to become a full-blown
  Bonding Center
• Previously we were Bonding Repair Center and we
  also already had responsibility for LT Module
  Testing.
• Its foreseen we will bond 500 TIB/TID Modules.
• Correspondingly the load of Module to LT-Test
  will reduce from 1000 down to the 500 we will
  bond.
• Modules for Catania would come from the Bari
  Gantry.
• Sebastiano will now give an overview of the
  Catania facilities.

38
Catania-Recent Activity

• In preparation for mass production, looking
  for best parameter sets, bonded ? pull tested
• Test Str (Hamamatsu, ST)
• PAs
• Module 30200020000008 ()
• () This is a Module with unbonded Hybrid,
  assembled in Bari, originally to be bonded
  (both Hyb-PA and PA-Sen) in Padova, but which has
  shown several problems (foreign material stuck to
  APV pads, scratches across the PA pads facing
  APVs, misalignment of PA to Sen by 1 pad) and
  therefore has been granted to us for exercise

39
Module 30200020000008
Material on APV pads
Scratches on PA pads
40
Catania Pull Tests Best-Performance
(All pull strengths in g)
() same parameters () every 25th wire
41
Firenze

• Has bought automatic Pull Tester identical to the
  one of Pisa.
• From now on, will join Pisa in the systematic
  Sensors Test Structure Pull Test job (relevant
  to Sensor qualification).
• Test Structures will still come into Pisa only,
  but subsets of them will then be moved to Firenze
  for test bonding and pull.

42
Hamburg

• Jigs?
• DB
• Center registration into DB? HAMBURG-U ?
• tool_ids

43
(No Transcript)
44
Schedule

• Alan cannot attend
• Info will be given by Gabriella Pasztor/ Bruno
  Wittmer at Mod Prod Mtg

45
(No Transcript)
46
Problems w/ delayed data recording in TrackerDB
This Module seems to have been assembled
(GANTRYVAL) in Aachen
47
Problems w/ delayed data recording in TrackerDB
For this Module MODPULLTEST seems to have
occurred in Aachen, not Zurich
48
Peculiar TEC R1 Sensor Design

• Klaus Kretschmer of Hamburg pointed out a
  peculiar variable pitch observed on TEC R1 Sensor
  bond pad row.
• Klaus Freudenreich investigated at my request and
  it turns out its a feature
• From Anna Macchiolo The wedge sensors W1
  TEC, W1 TID and W2 are the ones where the strip
  pitch is narrower and this has required some
  adjustment on the DC and AC pad position. For
  these three designs the pad step has been kept
  constant except for one over 16 pads. The first
  pad in each group of 16 has been repositioned in
  such a way to center it with respect to the
  corresponding strip.This was discussed in the
  sensor group at the time of the mask approval.
• We are showing it here to make everybody who has
  to bond R12 Modules aware of this.

49
Peculiar TEC R1 Sensor Design