OTDR Trace Interpretation: Events, Loss, Reflectance

An OTDR trace is the only record that tells you where a fiber span actually lost its light, and it is routinely misread: splices get blamed for connector loss, ghost reflections send technicians to the wrong kilometre, and a macrobend looks like a bad splice until you test a second wavelength. This guide covers how to read a trace event by event, the loss and reflectance values that mark a pass or a failure, and the settings that decide whether the trace is even meaningful.

How to read the trace in 30 seconds

• Spike up, then step down – a reflective event with loss: a mated connector pair, a mechanical splice, or a break.
• Step down with no spike – a non-reflective event: a fusion splice or a macrobend.
• The slope is the fiber itself – roughly 0.35 dB/km at 1310 nm and 0.22 dB/km at 1550 nm. A steeper slope means stress, damage, a tight bend radius or the wrong fiber type.
• End of trace – either a spike (connector or clean break at the far end) or a collapse to the noise floor (cleaved or crushed fiber). The distance is the span length.

Single-mode fiber shows a very small reflective spike at every event; multimode traces have far larger reflections, which is why event detection thresholds have to be set per fiber type.

Expected values per event type

Practical thresholds: a connector pair above about 0.75 dB is dirty, damaged or misaligned, and a fusion splice above 0.1 dB is a candidate for re-splicing. Measurements below about 0.5 dB between endpoints are rarely certified by single-ended traces – bidirectional testing and averaging is the accepted method for splice-loss acceptance.

Bend or bad splice? Use two wavelengths

A macrobend gets dramatically worse at longer wavelengths, so a span measured at 1310 nm and 1550 nm tells the two apart: a bend shows a much larger loss step at 1550 nm, while a splice loses roughly the same amount at both. This single test resolves most "mystery" loss in a new build.

Settings that change the answer

• Pulse width – short pulses give fine resolution but a shorter dynamic range; long pulses reach far but blur closely spaced events into one.
• Index of refraction – a wrong IOR scales every distance measurement, so verify it against the fiber datasheet.
• Launch and receive cables – always use them; the first and last connector of the span are unmeasurable without.
• Reflectance thresholds – too aggressive and you invent events, too lax and you miss a dirty connector.

Ghosts and other artifacts

A strong reflection (a dirty connector or an open end) can bounce back, reflect again at the launch end, and appear as a ghost event further down the trace. Ghosts have two signatures: they appear at multiples of a real distance, and there is no change in the backscatter level around them. Do not dispatch a crew to a ghost.

Field checklist

• Clean every connector before testing; contaminated end faces change the trace and the measurement.
• Record both directions and store traces with the correct IOR and wavelength metadata.
• Compare against the as-built loss budget rather than a single-pass number.
• Flag any step over the splice standard, and any reflectance above −35 dB for UPC anywhere on the span.

Related: Optical power budget calculation, reading transceiver Rx/Tx dBm, and DWDM system components.

原文链接：https://shopfiberoptic.com/blog/otdr-trace-interpretation-guide