9.3 Fibre Optic Maintenance, Termination, Splicing and Inspection
Key Takeaways
- A field termination is strip, clean, cleave square, seat in the ferrule, polish to the specified PC/UPC/APC geometry, then strain-relieve; a crushed or angled cleave dumps light out of the acceptance cone.
- A fusion splice welds the cores with an electric arc and typically adds about 0.02–0.1 dB; a mechanical splice uses a sleeve and index-matching gel at about 0.1–0.5 dB and is only used if the AMM permits a splice rather than jumper replacement.
- Inspect every end face with a fibre microscope before mating and never look into a live fibre; a particle on the core is a fail, and unfiltered cabin air or circular wiping scratches the ferrule.
- Minimum bend radius is a limit: macrobends at a clamp or lightening hole leak modes into the cladding and can look like a connector fault as the loom moves.
- Typical aircraft uses include IFE backbones, optical physical layers for AFDX-style Ethernet or ARINC 818 video on some types, and fly-by-light as a design concept — none of these is the exclusive architecture on every aeroplane.
9.3 Fibre Optic Maintenance, Termination, Splicing and Inspection
The pre-12 June 2024 detailed Appendix I description for 5.10 named terminations and aircraft applications alongside the physics; current Appendix I retains only the Fibre optics heading and levels. On the line, most “fibre faults” are end-face, bend-radius or coupler-budget problems, not a mysterious change in glass index. This section is the B2 workshop sequence: terminate, splice, inspect, then recognise where fibre actually sits on the aeroplane — and where copper still owns the system.
Termination: strip, cleave, polish
A connector termination is a mechanical and optical process, not a crimp copied from a copper pin:
- Strip the jacket, buffer and primary coating to the drawing length without nicking the cladding. A nick becomes a crack under vibration and then a delayed open.
- Clean the bare glass with the approved wipe and solvent. Skin oil is a failure mode, not a lubricant.
- Cleave with a purpose-made tool so that the end is flat and perpendicular. A crushed, lipped or angled cleave scatters light out of the acceptance cone the moment the fibre is mated.
- Seat the fibre in the ferrule with epoxy, anaerobic adhesive or a crimp-only mechanical terminus, as the component maintenance manual specifies.
- Polish to the specified PC, UPC or APC geometry on the correct films. Over-polish dishes the core; under-polish leaves a protruding shard that shatters on first mate.
- Strain-relieve the jacket in the connector body so that the glass does not carry loom tension.
Field-fit avionics termini (MIL-PRF-29504 style pins in a circular connector, ARINC 801, expanded-beam inserts) follow the same physics even when the polish step is replaced by a factory-cleaved contact or a lens. Expanded-beam connectors tolerate more contamination because the beam is wide at the interface, at the cost of higher insertion loss — a design trade, not a licence to skip caps.
Fusion splice versus mechanical splice
When a harness is cut, two fibres must be rejoined with cores coaxial and an index match at the joint.
A fusion splice uses an electric arc to melt the two cleaved ends into a continuous glass path. A fusion splicer that positions the two cores on a common axis typically yields about 0.02–0.1 dB insertion loss and a mechanically strong joint after a heat-shrink protection sleeve is applied. It needs electrical power, the splicer, cleaved ends, and a clean environment. The joint is permanent: you cannot demate it the way you demate a connector.
A mechanical splice butts two cleaved ends in a precision sleeve, usually with index-matching gel to fill the residual air gap. Loss is typically 0.1–0.5 dB, gel can dry or pick up dirt, and vibration performance is poorer. It is the usual field joint when a splicer is not available — if the AMM permits a splice at all. Many airframe procedures prefer replacing the jumper to splicing flight-critical fibre, because a workshop splice in a damp bay is a quality bet the type design may not accept.
| Joint | How the cores meet | Typical extra loss | Workshop implication |
|---|---|---|---|
| Fusion splice | Arc weld, then protect with a sleeve | 0.02–0.1 dB | Lowest loss; needs splicer |
| Mechanical splice | Sleeve + gel | 0.1–0.5 dB | Faster field joint; check AMM |
| Connector pair | Two ferrules in an adapter | about 0.2–0.5 dB | Demountable; contamination magnet |
| Expanded-beam pair | Lenses, no glass-to-glass contact | higher than PC | Better in dirty bays; still cap unused ports |
Inspection, contamination and bend radius
Inspect before every mate. Use a fibre microscope (often 200× or 400×) with the correct adapter. End-face zone rules treat a particle on the core as a fail; the same particle on the outer cladding may be acceptable. Never look into a fibre that may be live: avionics transmitters include LEDs and lasers. Unfiltered visual inspection of an active 1310 nm link can damage an eye because the wavelength is invisible.
Contamination (dust, lint, hydraulic mist, fingerprint) is the leading cause of sudden high insertion loss and of back-reflection that destabilises a laser. The repair is clean with approved tools, re-inspect, then mate. Blowing with unfiltered cabin air, or wiping in a circle that drags grit across the core, scratches the ferrule. Fit dust caps whenever a connector is open. A missing cap on a star-coupler spare port is a future fault, not housekeeping.
Minimum bend radius is a maintenance limit, not a suggestion. Macrobending at a clamp, through a lightening hole, or around a pulley that is too small couples guided modes into the cladding. The symptom is often a temperature- or door-position-dependent loss as the loom moves. Do not dress fibre like copper. Follow the AMM figure; if none is given, the cable marking or manufacturer data (often at least ten times the cable diameter, or a stated 25–38 mm for a patch lead) applies. Microbends from overtight lacing tape produce the same physics at a smaller scale. After any disturbance of a fibre loom, restore clamps and radius formers before returning the aircraft.
An optical power meter and stable source measure end-to-end insertion loss against the loss budget taught in section 9.2. An OTDR locates a break or a reflective connector along a longer run; on very short aircraft jumpers the OTDR dead zone can hide the fault, so a visual fault locator (visible red laser, used with eye-safety rules) is often more useful for a smashed jumper. A copper milliohmmeter will not ring out a dark fibre.
Aircraft applications (typical, not exclusive)
Fibre is chosen where bandwidth, mass or EMI beat copper, not because every digital system is optical.
- In-flight entertainment (IFE) and cabin networks: long, high-rate video and Ethernet backbones are a common civil application. A failed IFE fibre is an inconvenience; treat it with the same contamination discipline because the same technician may next open a flight-critical connector.
- Avionics data networks: some types use optical physical layers for AFDX/ARINC 664-style Ethernet, fibre channel, or ARINC 818 cockpit or cabin video. Copper 100BASE-TX AFDX remains widespread. Optical variants are typical options, not a claim that every AFDX network is fibre.
- Fly-by-light: optical signalling of flight-control commands is a design concept (and appears on some military or research programmes) that exploits EMI immunity and spark-free routing through noisy or fuel-adjacent structure. Do not state that production airliners have replaced fly-by-wire copper or ARINC buses with fly-by-light as a universal standard. If the aircraft uses optical flight-control links, the AMM and type design define them.
- Sensors and data concentration: some fire, landing-gear and structural-monitoring installations use optical sensors or optical runs through electromagnetically noisy bays, still with copper power to the box.
Criticality belongs to the system, not the medium. An IFE star coupler and a flight-control optical link both need clean termini; only one is operationally essential. After work on a fibre loom, verify the system BITE or a measured optical loss. Do not sign a fibre task on the basis of a continuity beep from a copper-only meter.
B1 level 1 is recognition: do not bend, do not contaminate, inspect with a microscope, know that fusion is the low-loss splice. B2 level 2 is the procedure and the application map, including the honesty that fly-by-light and optical AFDX are contexts, not exclusive fleet-wide facts.
A severed avionics fibre must be rejoined. Which statement correctly compares a fusion splice with a mechanical splice?
Before mating an avionics fibre connector, what inspection and contamination practice is required?
Which termination step most directly keeps launched light inside the fibre's acceptance cone?
Which statement about aircraft fibre applications and loom practice is accurate for topic 5.10?