Barrel Harness Status

1
Barrel Harness Status

• Harness Production Status in Taiwan.
• Harness Functionality Tests at RAL.
• Harness Mounting Trials.
• Release of production.
• Schedule.

2
Taiwan Status

• Sub-Assembly Count
• Optopackages 288/352 in units of equivalent
  harnesses passed all QA.
• Assembled doglegs 64/352 (sufficient for B3).
• i/f PCB 116/352
• PPB1 66/352
• Assembled PPB1/LMT/i/f PCB 46/352
• MT terminated ribbons 342/352.
• Two batches of 8 harnesses sent to RAL.

3
Summary RAL Tests

• First batch of 8 harnesses.
• 4 fully functional harnesses.
• 2 with broken PIN fibres.
• Furcation tubing damaged by sharp object. Not
  understood how this happened. Not found on
  subsequent harnesses.
• 2 with problems on redundant clock
• Two DORIC4As gave unwanted jittery redundant
  clock when SELECT off. This is now tested at the
  dogelg level and the complete harness level and
  this defect has not been seen again.

4
Harness Mounting Tests

• With first batch of 8 harnesses.
• Many problems with clash between opto-package or
  fibres and envelope module.
• Problem due to poor quality of dogleg S bend.
• One dogleg with clash was re-bent on an old
  bending jig and afterwards it fitted fine.
• We have made two new verification/bending jigs
  for LH and RH harnesses. Can use these jigs to
  bend the doglegs on the 4 functional harnesses
  and to verify the next batch of harnesses.

5
Good and Bad S bends

• No prizes for guessing which is the good bend.

6
Harness Mounting Tests

• With second batch of 8 harnesses.
• Quality of dogleg S bends much improved (correct
  use of existing jig).
• A LH and RH harness were mounted on B3.
• Clearances from opto-package and dogleg to
  envelope module was checked on all 12 positions
  and was very good.

7
Harness Mounting Tests
3 fibres together

• One problem with fibre loop from dogleg creating
  clash.
• More care required to ensure fibre loops are
  correct.

1 fibre separate
8
2nd Batch of 8 harnesses

• No broken fibres and no unwanted jittery
  redundant clocks.
• Many problems due to glue added to stiffen region
  around redundancy arm connector.
• Glue not fully cured before connector added ?glue
  wicked into connector and prevents a connector
  being added or worse removed.
• This led to many problems with redundancy arms
  for dogleg 6.
• Several breaks on tracks
• Mainly on tracks for redundant signals
• Those breaks we have managed to see optically are
  near pads.
• All pads are Ni/Au plated. This is optimal for
  wire bonding but it is well known that the Ni
  layer is brittle and can crack ? need stiffener
  under connector.
• We have stiffener under module connector but we
  didnt add stiffener under redundancy arm
  connectors because of tight clearances at end of
  barrel.

9
i/f PCB connector
Redundancy connector
Special redundancy connector dogleg 6
Module Connector with FR4 stiffener
10
Broken Tracks

• 4 harnesses had problems with broken tracks on
  redundancy arm connectors, mainly but not all for
  dogleg 6.

crack
11
Summary 2nd Batch(1)

• two are 100 functional.
• One was 100 functional apart from the fact that
  we couldnt insert redundancy connector for dog
  leg 6. Connector removed and new connector added.
  Tests to be repeated but should be good.
• One was 100 functional but dogleg 6 redundant
  tracks were broken when removing the temporary
  interlink cable.
• One had o/c for two lines ILEDA and SELECT
  ?broken tracks. Also had bad redundant command
  data out of dogleg 6 ? another broken track.

12
Summary 2nd Batch (2)

• Bad data for V0 dogleg 6 ? broken track or bond
  wire.
• Bad redundant data dogleg 3. One side of signal
  missing. Works when connector pushed down ?
  hairline crack.
• s/c SELECT and RESET on i/f PCB due to thermode
  soldering. Rework removed the solder making the
  short. Bad redundant command data out of dogleg 3
  after attempt at rework it works but fails if
  connector pressed down ? hairline crack.

13
Improvements

• Fully cure glue used for stiffener before adding
  connector on dogleg 6.
• Epoxy is similar to cover layer so tracks should
  be much better protected.
• Try this on sample of mechanical grade doglegs.
• Make multiple mate and demate cycles.
• Flex circuits deliberately.
• Repeat on control sample with no added glue.
• Test continuity at cold temperature.
• If this doesnt work we will have to use
  stiffener but this will be very difficult.
• Add permanent interlink cable for end doglegs on
  half of the harnesses. Use Minglees special
  interlink cable for testing the other half of the
  harnesses to allow safe de-mating.
• Better protection for shipping.

14
Schedule

• Sub assemblies very well advanced so bottle neck
  is now harness testing
• Next shipment of 20 harnesses on 10th September.
• With more optimisation of test equipment,
  Radiantech should be able to produce up to 100
  harnesses/month ? finish production by end of the
  year.

15
Harness Release

• Recommend production of barrel harnesses to
  start.
• The high rate of track failures in the last batch
  of harnesses needs to be understood and steps
  taken to minimise it in future.
• Care must be taken when using the harness forming
  jig to
• Ensure that sufficient over-bending is used so
  that a nearly vertical part of the dogleg is
  produced The bends should be parallel to the line
  of the low mass tapes.
• The 3 fibres in furcation tubing should stay
  together and the fibres should dip down after
  exiting the opto-package.
• The protection of the harnesses during shipping
  should be significantly improved.