Bidirectional OTDR Splice Loss and IOR Fault Location
An OTDR trace at 1310 nm on a 12.4 km single-mode OSP link records a splice at 6.180 km as 0.42 dB from the A end and 0.05 dB from the B end. The fiber IOR is 1.468, but the OTDR was set to 1.500. A second event at 9.840 km shows a 0.08 dB non-reflective loss, yet field records show no splice there. After a reported outage, a technician tests from A and sees a reflective end at 4.250 km with about 14 dB return loss; the far end at 12.4 km is absent. The link has splice closures at 6.180 km and 9.840 km and a mid-span aerial section. Acceptance for splices is 0.30 dB maximum bidirectional average.
1. Compute the true splice loss at 6.180 km and state whether it passes. Explain why the two directions differ.
Read the worked answer
Bidirectional splice loss is the average of the two directions: (0.42 + 0.05) / 2 = 0.235 dB. The acceptance limit is 0.30 dB, so the splice passes. The one-way values differ because OTDR splice loss includes mode-field diameter mismatch and backscatter-direction effects; the bidirectional average cancels most of that bias. Accepting the 0.42 dB reading alone would overstate the splice loss and could lead to unnecessary rework.
2. What is the true physical distance of the fault, and how does the IOR error affect the reading?
Read the worked answer
OTDR distance is computed from the entered IOR. True distance = displayed distance × (IOR set / IOR actual) = 4.250 km × (1.500 / 1.468) ≈ 4.343 km. Because the OTDR was set to a higher IOR than the fiber, it under-reports distance; the fault is about 93 m farther than displayed. The reflective 14 dB event is a break or connector, not a splice. Excavation or fault locating should use about 4.34 km, then confirm with a fiber finder or toned cable before digging.
3. Is the 9.840 km event likely to be real? What could cause it, and how would you confirm it?
Read the worked answer
The 9.840 km event is probably not a splice because records show none. A 0.08 dB non-reflective event can be a macrobend, a microbend, or an OTDR ghost. Ghosts usually appear beyond a strong reflective event at a multiple of its distance; 9.840 km is not a simple multiple of 6.180 km, so a ghost is less likely but not impossible. Test at 1550 nm: bend loss will increase markedly at the longer wavelength, while a ghost will not behave as a real loss. Also test from B and compare. If the event persists and grows at 1550 nm, locate and relax the bend; if it changes with pulse width or disappears, treat it as an artifact.