Fibre Splicing Explained: Fusion vs Mechanical Splicing

A connector introduces an air gap, or at best a polished physical contact, between two fibre ends, and this junction always adds some optical loss and a potential point of contamination. A splice, by contrast, joins the glass fibres directly, which is why a good splice typically loses far less light than a connector pair.

Splicing is used wherever a continuous run of fibre needs to be longer than what comes on a single reel, wherever a damaged section of cable needs to be cut out and repaired, and wherever a pre-terminated pigtail needs to be joined to field-installed cable rather than terminating a connector directly onto the fibre by hand.

Fusion splicing uses an electric arc to melt the ends of two prepared fibre strands together, fusing the glass into a single continuous strand. A fusion splicer aligns the fibre cores automatically or manually under magnification before firing the arc, and the result is typically a very low loss joint, often well under 0.1 dB when done correctly.

This method requires a dedicated fusion splicing machine, which is a significant piece of equipment, along with a controlled, relatively clean working environment since dust or moisture on the fibre ends during preparation will show up as increased loss or a failed splice. Fusion splicing is the preferred method for permanent, high-performance joints, including backbone and external cable work.

Mechanical splicing aligns two fibre ends inside a small precision-made housing, using an index-matching gel to reduce the light loss at the junction, and holds them in place with clamps rather than fusing the glass. No heat or specialised splicing machine is required, which makes mechanical splices faster to complete and useful for emergency repairs or low-fibre-count jobs.

The trade-off is generally higher loss than a fusion splice, and mechanical splices can be more susceptible to loss increases over time from vibration, temperature cycling or gel degradation. They remain a legitimate choice for temporary repairs, testing, or smaller jobs where the cost of a fusion splicer is not justified, but most permanent commercial installations favour fusion splicing where volume allows.

Once a splice is made, the joint itself is fragile and needs mechanical protection. Fusion splices are typically covered with a heat-shrink splice protector, which is then laid into a splice tray that holds the fibre in a gentle loop, avoiding sharp bends while keeping every joint organised and individually accessible.

Splice trays sit inside a splice enclosure, which can be a wall-mounted box, a rack-mounted unit in a communications room, or a sealed external closure for underground or aerial joints. External enclosures need to be properly sealed against moisture ingress, since water reaching a splice tray is a common cause of long-term fibre failures in buried or exposed cable runs.

A pigtail is a short length of fibre with a factory-terminated connector already fitted on one end and a bare fibre end on the other. Rather than trying to polish and terminate a connector by hand in the field, which is slow and prone to inconsistent results, installers commonly splice a pigtail onto the field cable and rely on the factory-made connector for the termination.

This approach, sometimes called splice-on connectivity, has become the standard method for most commercial fibre terminations because it produces more consistent, lower-loss results than hand-polished field connectors, while still allowing a splicer to complete terminations on site without needing a fully equipped polishing workshop.

The quality of a splice depends heavily on how the fibre is prepared beforehand. This includes stripping the outer coating cleanly, cleaving the fibre end to produce a flat, perpendicular face, and thoroughly cleaning both the fibre and any tools that touch it. Even a small amount of dust or a slightly angled cleave can noticeably increase loss or cause the splice to fail outright.

Experienced technicians treat fibre ends as effectively contaminated the moment they are exposed, cleaning immediately before splicing rather than in advance, and working in conditions where dust and moisture are minimised as far as practically possible, particularly important on outdoor or construction sites.

A well-executed fusion splice typically shows loss figures well under a tenth of a decibel, while a mechanical splice more commonly sits somewhere between a few tenths of a decibel, though this varies by product and installer skill. These figures matter because every splice consumes part of the link's overall loss budget, and a project with many splices needs each one kept as low as possible to leave headroom for the equipment at either end.

When a splice comes back with unexpectedly high loss on testing, the usual response is to redo it rather than accept a marginal result, since a borderline splice today can become a fault later as the cable ages or experiences temperature cycling.

Every splice on a commercial job should be tested, typically with an OTDR for longer external runs or a power meter and light source for shorter in-building sections. The OTDR trace shows the exact loss at each splice location along the cable, which is invaluable both for quality control at handover and for troubleshooting years later if a fault develops.

Good practice includes keeping a splice record showing tray and fibre position, loss result, and date, alongside the OTDR trace itself. This documentation becomes part of the project's as-built records and saves considerable time if a fault needs tracing in future, since a technician can compare a new trace against the original to see exactly what has changed.

Not every fibre problem needs a splice, and not every fibre join should be a connector. Splicing makes sense when a permanent, low-loss, low-reflection joint is required and the connection will not need to be moved: joining reels to make up a long external run, repairing a damaged cable, terminating a lead-in inside a building, or breaking a multi-core backbone out into individual fibres at a distribution point.

Patch leads and adaptors are the better choice wherever a link genuinely needs to be re-patched, such as between a patch panel and a switch. Connectors are designed to be mated and unmated repeatedly, whereas a splice is intended to be made once and left alone inside a protective enclosure for the life of the cable.

For Australian businesses running fibre between buildings on a campus, into a warehouse mezzanine, or out to a detached workshop, the practical pattern is usually the same: external-grade cable spliced onto pigtails at each end, terminated into a rack-mounted fibre enclosure, then patched into equipment with standard leads that can be changed without touching the permanent infrastructure.

Splicing is a specialist task with specific site requirements, so it pays to plan for it rather than treat it as an afterthought. Splicers need a stable, reasonably clean working area near the enclosure location, access to power or charged batteries, and enough slack cable to work with, which means leaving a service loop at each end and at any intermediate joint location during the pulling stage.

The most common cause of avoidable delay on fibre jobs is insufficient slack. If the cable has been cut tight to the enclosure, there may not be enough length to bring the fibre into a splicing rig comfortably, and the fix can involve pulling in additional cable. A metre or two of coiled slack at each end costs almost nothing at installation and saves significant time during both the initial splice and any future repair.