4.2 OTDR Trace Analysis

What the OTDR Does

An Optical Time Domain Reflectometer (OTDR) launches short pulses of light into a fiber and measures the light scattered back (Rayleigh backscatter) and reflected back from discontinuities. Because light travels at a known speed in glass (about 2 × 10⁸ m/s), the OTDR converts the time delay of returned light into distance. The result is a trace: distance on the x-axis, attenuation on the y-axis in dB.

The OTDR is unique because it characterizes every event along a link from one end only. An OLTS measures end-to-end loss; an OTDR tells you where the loss happens.

Reading the Trace

A clean link has four regions:

1. Launch pulse / dead zone — the front-end pulse and recovery region; events inside the dead zone cannot be resolved.
2. Backscatter slope — the gentle downward line representing the fiber's intrinsic attenuation (e.g., ~0.35 dB/km at 1310 nm single-mode; ~1.0 dB/km at 850 nm multimode).
3. Events — spikes (reflective) or steps (non-reflective) where the backscatter line changes.
4. End of fiber — a final reflective spike (connector) or a clean drop (cleaved end, broken fiber, or far-end non-reflective termination).

Reflective vs. Non-Reflective Events

This is the most important classification on the exam:

A fusion splice is non-reflective because the glass is continuous — there is no index discontinuity to reflect light, only a small loss. A connector is reflective because the glass-to-air-to-glass interface has a refractive index mismatch that reflects a fraction of the pulse back.

Reading Loss at an Event

• Splice loss (non-reflective): measure the vertical drop in dB at the step. A typical fusion splice should be ≤ 0.3 dB; anything above ~0.5 dB on single-mode is suspect.
• Connector loss (reflective): the spike is the connector itself; the loss is the vertical drop from the backscatter line before the spike to the new (lower) backscatter line after. Total connector loss (the pair, both ends combined) typically ≤ 0.75 dB per TIA-568, but the OTDR shows only the loss at that interface, not the mated pair end-to-end.

The OTDR reports both event loss (the loss at the point) and reflectance (the magnitude of the spike in dB, negative). A clean APC connector has reflectance around -60 dB or better; a UPC connector around -50 dB; a dirty connector may show reflectance as high as -30 dB.

Ghosts

Ghosts are false events that appear at integer multiples of a real event's distance from the OTDR. They are caused by multiple reflections bouncing back and forth inside the link: light reflects off a connector, returns to the launch, reflects off the launch connector, and travels down the fiber again, appearing as a second event at 2× the real event distance.

Signs of a ghost:

• No measurable loss at the event (the backscatter line continues unchanged after the spike).
• The distance is an exact multiple of a strong reflective event.
• The ghost disappears when the strong reflector is cleaned or re-mated.

Ghosts are not faults. Do not chase them. Reduce them by cleaning the connectors, lowering the launch pulse power, or using an OTDR with built-in ghost masking.

Macrobends

A macrobend is a fiber bent tighter than its minimum bend radius. Light escapes the core at the bend, producing loss. On the OTDR it appears as a non-reflective drop — just like a fusion splice — but the cause is mechanical. To distinguish: trace the cable route visually; a macrobend shows where the cable makes a sharp turn, is crushed, or is pressed against a sharp edge. Macrobends are sensitive to wavelength: loss is much higher at 1550 nm than at 1310 nm, so a bend can be invisible at 1310 and visible at 1550. This is why OTDR testing of single-mode plant is done at both wavelengths.

Splice Loss vs. Connector Loss

The exam will show you a trace and ask which is which. The key tells:

• Fusion splice: non-reflective step, no spike, typically small (≤ 0.3 dB), often in the middle of a run.
• Connector pair at a patch panel: reflective spike with a measurable loss, located at a known patch panel distance.
• Mechanical splice: reflective, but with smaller reflectance than a connector and usually higher loss than a fusion splice.

Practical Tips

• Set the refractive index in the OTDR to match the fiber (typical 1.46–1.47 for single-mode, 1.48–1.50 for multimode) or distances will be wrong.
• Use the right pulse width: short pulses give better resolution but less range; long pulses reach farther but blur events together.
• Always test from both ends and compare — an event that looks small from one end may look large from the other (gainers), especially on multimode.
• Use a launch box (a 1 km reference fiber) to move the launch pulse and dead zone out of the link under test, so the first connector is visible.