How to read an OTDR trace — events, artifacts, and what insurance pays for.

An OTDR (Optical Time Domain Reflectometer) is a fiber technician’s most important certification tool and one of the least-understood artifacts it produces. The trace — a plot of optical power return versus distance along the fiber — contains a complete record of every event on the span: splices, connectors, bends, breaks, and the artifacts that don’t represent real fiber events but look like they do. Reading it correctly is a skill that takes time to develop, but the principles are not complicated once you understand what the instrument is actually measuring.

This article covers the trace anatomy, the event types you will see on a real fiber span, the artifacts that confuse new readers, and the acceptance criteria that matter for new construction certifications and insurance claims.

What an OTDR actually measures

An OTDR injects a short pulse of laser light into the fiber and measures the light that returns to the launch end over time. Two physical phenomena cause light to return:

• Rayleigh backscatter: As the pulse travels through the fiber, a tiny fraction of the light scatters backward continuously from microscopic density variations in the glass. This produces the sloping baseline on the trace — the straight-line decay in optical power as you move farther from the instrument. The slope of that line is the fiber’s attenuation coefficient (dB/km).
• Fresnel reflection: At discontinuities in the glass — an air gap, a connector end-face, a mechanical splice, a break — a fraction of the light reflects sharply back. These appear as vertical spikes on the trace.

The OTDR converts the round-trip travel time of the returned light to a distance (using the fiber’s known index of refraction) and plots the return power in dB on the Y-axis against distance on the X-axis. Everything you need to know about the span is in that plot.

Trace anatomy — the three zones

Every well-acquired OTDR trace has three zones:

The launch box: The beginning of the trace is dominated by the reflection at the OTDR’s output connector and by the instrument’s recovery time after the pulse injection. This zone is blind — real fiber events within the launch box distance cannot be distinguished from the instrument’s own noise. The solution is a launch cable (also called a lead-in or pulse-suppressor cable): a spool of fiber, typically 100–500 m, inserted between the OTDR and the span under test. The launch cable pushes the launch-zone blind area out beyond the first real connector on the span, making the near-end connector visible and measurable.

The measurement zone: The main span. This is where real events appear against the backscatter baseline. The baseline should be a straight, uniform slope. Any deviation from that slope is either a real event or an artifact. Real events are characterized by a loss (a step down in the baseline), a reflection spike, or both.

The tail box: The far-end connector creates a reflection spike, and after that spike the trace continues for a short distance before dropping to the noise floor. This tail region, like the launch box, can obscure events near the far end. A tail cable (a second spool at the far end) extends the measurement zone past the far-end connector, making events near the far end measurable.

Event types — what you will see on a real span

Events fall into two categories: reflective and non-reflective.

Reflective events produce a spike on the trace followed by a step loss. They appear at locations where there is a physical discontinuity in the fiber geometry — an air gap, a mechanical interface. Typical reflective events:

• Connectors: Every mated connector pair produces a reflection and a loss. A clean, well-polished, properly mated PC or APC connector should show a reflection peak followed by a loss step. The loss should be 0.3 dB or less for a single connector pair under TIA-568 acceptance criteria. APC (angled physical contact) connectors produce much less reflection than PC connectors — the angled end-face deflects the reflected light away from the fiber core.
• Mechanical splices: Mechanical splices produce a small reflection (from the index-matching gel gap between fiber ends) and a splice loss. Acceptable mechanical splice loss is typically 0.5 dB or less, though modern index-matching gels can achieve lower values.
• Fiber breaks with gaps: If the fiber is broken cleanly (a cleave that separated), there will be a large reflection at the break point followed by no further trace. If the fiber is broken with a crush or bend, you may see only a loss step with no reflection.

Non-reflective events appear only as a step loss in the baseline, with no associated spike. Typical non-reflective events:

• Fusion splices: A well-made fusion splice is nearly invisible on the trace — 0.02–0.05 dB loss for a good splice, sometimes with no visible step at all. TIA-568 specifies 0.3 dB maximum per splice for acceptance. A fusion splice that shows a pronounced step (0.1 dB+) warrants re-inspection; values above 0.3 dB should be re-spliced.
• Bends and micro-bends: Exceeding the fiber’s minimum bend radius produces loss that appears as an elevated attenuation over a distance rather than a sharp step. A bend-induced loss region on the trace looks like the baseline slope increases over a short distance, then returns to normal. Look for this at conduit bends, at cable management clips, and at any location where the cable changed direction.
• Splitter events (PON architectures): Optical splitters appear as large non-reflective losses. A 1:8 splitter produces approximately 10.5 dB of loss. Not an error condition; just the expected loss of the splitter.

Ghost reflections — the artifact that catches new readers

Ghost reflections are one of the most common sources of confusion in OTDR trace reading. They appear as reflection spikes at locations where no physical event exists. The cause: a large reflection earlier in the span (typically a connector end-face with poor contact, a broken fiber face, or a bare fiber end) reflects light back toward the OTDR. Some of that reflected light travels back through the launch cable, reflects again at the OTDR output connector, travels back out to the far end, and returns to the OTDR a second time. The OTDR measures the round-trip time of this double-reflected light and plots a “ghost” event at twice the distance of the real reflection.

Diagnosing a ghost:

• The ghost event is always at exactly twice the distance of the real event that caused it (measured from the OTDR, or from the near end of the launch cable if one is in use).
• The ghost has no associated step loss — only a reflection spike. Real connector events have both a spike and a loss step.
• The ghost disappears when the source event is corrected (cleaning the connector, re-polishing the end-face, or capping the fiber end).

Acceptance criteria and what documentation must include

For new construction fiber acceptance testing, TIA-568.3-D (the fiber standard) specifies maximum attenuation per link, maximum connector loss, and maximum splice loss. The OTDR test documentation that satisfies a GC or owner acceptance requirement and supports an insurance claim must include:

What insurance carriers want when a fiber span is damaged and a claim is filed: evidence of what the span looked like at acceptance, so the repair scope is bounded by the pre-damage condition rather than by what the carrier can reconstruct from an undocumented installation. Baseline OTDR documentation is the pre-damage record. Without it, the carrier has no way to verify that the span was correctly installed before the damage event — and that ambiguity is resolved in the carrier’s favor.

Bottom line

An OTDR trace is a complete record of the fiber span at the moment of measurement. Reading it correctly requires understanding the difference between real events and artifacts, knowing what the acceptance thresholds are, and knowing how to produce documentation that is usable for both acceptance and future troubleshooting. The technical skill is learnable — the principles in this article are enough to read most commercial building fiber traces correctly. The discipline of documenting fiber spans at acceptance, with launch and tail cables and bidirectional measurements, is what separates installations that can be certified, insured, and troubleshot from ones that can only be replaced.