Joining two fiber strands into a single continuous optical path can be done two ways: fusion splicing, which permanently welds the glass fibers together, or mechanical splicing, which aligns and clamps them without melting anything. Both get light from one side to the other. They are not interchangeable for most real-world work.
Fusion splicing
A fusion splicer strips, cleaves, and aligns the two fiber ends under magnification, then uses an electric arc to melt and fuse the glass into a single, continuous strand. Done correctly, the splice point has almost no measurable difference from the surrounding fiber — typical loss runs well under 0.1 dB per splice.
Where it's the right call:
- Permanent installations — outside plant, in-building backbone, anything expected to stay in place for years
- Data centers, campus networks, and any application with a tight loss budget
- Long-haul runs where losses from multiple splices would otherwise add up
Trade-off: it requires a fusion splicer (a real piece of calibrated equipment, not a low-cost tool), a stable, relatively clean work environment, and a trained technician. It's not something to attempt as a five-minute field fix.
Mechanical splicing
A mechanical splice holds two cleaved fiber ends in precise alignment inside a small mechanical component — often using an index-matching gel to minimize the air gap — without fusing anything. No electric arc, no bulky equipment.
Where it's the right call:
- Temporary or emergency repairs, especially in the field with no power source for a fusion splicer
- Very quick turnaround situations where a permanent fix will be scheduled later
- Low-strand-count, non-critical runs where a slightly higher loss budget is acceptable
Trade-off: typical loss runs higher (often 0.1–0.3 dB or more per splice, sometimes higher depending on component quality), and the splice is more sensitive to temperature changes, vibration, and time — it's a connection, not a weld. For anything permanent, it's a stopgap, not a solution.
The practical answer
If a contractor proposes mechanical splicing for a permanent backbone run to save time or cost, that's worth a direct question: is this the final splice, or a placeholder? For anything that's supposed to last the life of the building, fusion splicing — fully OTDR tested afterward — is the standard we use, and it's the standard worth insisting on from anyone doing the work.
Need a run spliced, tested, and certified the right way the first time? Reach out for a quote.