Bidirectional splice loss calculator

Enter each splice's OTDR reading from both ends. The true loss is the average, and anything over your pass limit is flagged to reshoot.

Pass limit (dB) Wavelength

Splice	A→B (dB)	B→A (dB)	Avg (dB)	Result	Remove

Splices passing

—

Why one-way OTDR splice readings lie

An OTDR does not measure loss directly. It measures the light scattered back toward it and infers loss from the step in that backscatter. When two fibers with different backscatter coefficients are spliced, which happens whenever their mode field diameters differ even slightly, the step is wrong. Shot from one end the splice looks better than it is, and shot from the other it looks worse by the same amount.

The classic sign is a gainer: a negative reading, as if the splice added light. It did not. The true loss is hidden under the backscatter difference, and only the average of both directions cancels it out.

True splice loss = (A→B reading + B→A reading) ÷ 2

Worked example

Splice #3 reads −0.05 dB shot from end A (a gainer) and 0.21 dB shot from end B. Neither number is the real loss. The average, (−0.05 + 0.21) ÷ 2 = 0.08 dB, is, and it passes a 0.10 dB spec. Judged from end B alone, it would have been cut out and redone for nothing.

What pass limit to use

Use the number in your project's spec. Many carrier and contractor specs call for 0.10 to 0.15 dB average per fusion splice, and modern core-alignment splicers routinely do better than that. TIA-568 sets 0.3 dB as the maximum for any splice, fusion or mechanical, which is a ceiling for design, not a target for workmanship.

Tips for clean bidirectional results

Use launch and receive cables long enough to clear the dead zone at both ends, so the first and last connectors can be measured too.
Match the settings in both directions: same pulse width, wavelength and index of refraction. A different IOR shifts distances and makes events harder to pair up.
Number the splices from the same end. Splice #1 from end A is the last event from end B. Most OTDR software flips the trace for you; when entering by hand, check the distances line up.
Shoot 1550 nm as well as 1310 nm. A splice that is fine at 1310 and high at 1550 points to a bend or a tight spot in the tray, not a bad fusion.

Questions techs ask

What is a gainer on an OTDR?

A splice that shows negative loss from one direction. It happens when the fiber after the splice scatters more light back than the fiber before it. It is not a real gain; average it with the reading from the other end.

What is acceptable fusion splice loss?

Most project specs ask for 0.10 to 0.15 dB average from both directions. TIA-568 allows up to 0.3 dB for any splice. A good splicer on matched fiber often averages around 0.02 to 0.05 dB.

Can I trust the splicer's loss estimate?

It is a useful check that the fusion went well, but it is an estimate from the image of the fibers. Acceptance testing is done with a bidirectional OTDR or a power meter.

Why does the same splice read differently at 1310 and 1550 nm?

Some difference is normal. If 1550 nm reads much higher, look for a macrobend in the tray or the closure: longer wavelengths leak out of tight bends more easily.