Testing, Troubleshooting and Network Design
Domain 4 tests whether you can prove a link performs, diagnose a failure, and design to a budget. It links Tier 1 loss testing, Tier 2 OTDR testing, and network design margins.
Test methods and interpretation
- Tier 1 uses an optical loss test set (OLTS) to measure insertion loss and length at the operating wavelength. Set the reference according to TIA-526-14 (multimode) or TIA-526-7 (single-mode); the one-cord and three-cord reference methods give different included losses. Multimode testing should use an encircled-flux-compliant launch.
- Tier 2 uses an OTDR to characterize events: splice loss, connector loss, reflectance, and optical return loss. Use launch and receive cables longer than the OTDR dead zones; test bidirectional and average for splice loss. Ghosts are echoes from high-reflectance events, not real fiber events.
- Clean and inspect every connector end-face before testing; a single contaminated connector can dominate loss.
Design and troubleshooting
- Link loss = fiber attenuation × length + connector loss + splice loss + passive component loss. Compare to power budget = minimum transmitter output − maximum receiver sensitivity, then subtract a design margin (often 3 dB).
- Troubleshoot from the physical layer outward: power, polarity, connector damage, bend loss, then OTDR events. A link that passes OLTS but fails traffic may still have high reflectance or a marginal budget.
Worked example
A 2 km single-mode link at 1310 nm uses fiber with 0.35 dB/km, two connector pairs at 0.5 dB each, and one splice at 0.1 dB. Link loss = (0.35 × 2) + (2 × 0.5) + 0.1 = 0.7 + 1.0 + 0.1 = 1.8 dB. Add a 3 dB design margin: required budget = 4.8 dB. If the transmitter/receiver budget is 8 dB, the design passes with 3.2 dB spare. Reasoning: the spare is not "extra loss" to spend blindly; it covers aging, repairs, and temperature.
Traps that cost marks
- Testing only one direction on OTDR and reporting that splice loss.
- Using a launch cable shorter than the dead zone, hiding the first connector.
- Forgetting that OLTS reference method changes the reported loss.
- Designing to the exact power budget with no margin.
- Confusing event loss with cumulative loss on an OTDR trace.