OTDR Trace Reading: Three Mistakes Even Senior Field Techs Make

I've watched a fifteen-year fiber technician misread a ghost reflection as a real splice loss. He took the OTDR screen at face value, called the trace a bad splice, ordered the splicing rig back out for re-work, and rebuilt a perfectly good splice that wasn't the problem. The instrument didn't lie. His pattern-matching did.

Three mistakes show up over and over in field trace reviews. They're not beginner mistakes. They're the kind of mistakes you make after enough experience that you stop reading the trace carefully.

Mistake one: ghost reflections read as real events

A ghost reflection happens when light bounces back from a connector, returns to a previous connector, bounces forward again, and shows up at the OTDR as a second "event" at exactly the distance the round trip would suggest. It looks like a real reflection. The location is wrong but plausible. If you have two strong connectors in the link, you'll get a ghost at the sum of their distances.

How to catch it: compare bidirectional traces. A ghost only appears in one direction, because the geometry is asymmetric. A real event appears in both. The trace from end A might show a "splice loss" at 12.4 km that the trace from end B doesn't reproduce. That's the test.

Most field techs run a single-direction trace because it's faster. They save the time and pay it back five-fold when they're chasing a phantom.

Mistake two: dead-zone clipping on short patch leads

The OTDR has two dead zones: the event dead zone (where two close events merge into one) and the attenuation dead zone (where the receiver hasn't recovered from a reflection enough to measure loss). For a typical OTDR, event dead zone is 1 to 3 meters, attenuation dead zone is 5 to 15 meters.

Run an OTDR through a patch panel with 2-meter patch leads on each side, and the events at the panel are inside the dead zone. The trace looks clean. It is not clean. The losses are there, the instrument just can't see them.

How to catch it: use a launch lead. A 100-meter launch fiber spool in front of the link pushes the panel events out of the dead zone. The trace now resolves the patch panel correctly. Senior techs sometimes skip the launch lead because the link "looks fine without it". The link looked fine because the OTDR was blind for the first 15 meters.

A second symptom: traces that show suspiciously low loss across known long fibers. If you're reading 0.15 dB/km on a link that should be 0.21 dB/km, the math is off. Check for dead zones eating the first or last event.

Mistake three: misreading gainers

A "gainer" is an OTDR event that shows a negative loss — the apparent signal level after the event is higher than before. New field techs see a gainer and assume the trace is broken. It is not broken. Gainers are real.

They appear when two fiber sections with different backscatter coefficients are spliced together, and the OTDR is reading the splice from the side with lower backscatter. The light power going through hasn't increased. The reflected power (which is what the OTDR measures) just gets stronger on the other side because that fiber backscatters more.

The fix is the same as for ghosts: bidirectional averaging. A real splice has an actual loss that's the average of the two single-direction reads. If the forward direction shows -0.05 dB (gainer) and the reverse shows +0.25 dB, the real loss is 0.10 dB.

Senior techs who haven't run bidirectional averaging in a while will see a gainer, distrust it, and re-splice the connection. They've solved nothing — the gainer comes back, because it isn't a fault. It's a backscatter mismatch.

What changes when you trust the instrument and verify the read

Bidirectional traces take roughly twice as long as single-direction. The launch lead adds a setup step. Bidirectional averaging takes a calculator pass after the test. All three add maybe twenty minutes to a commissioning test.

The cost of skipping them is a re-work crew. One re-work crew costs more than the entire test setup, in time and in confidence. I've never met a fiber team that ran bidirectional traces by default and complained about the time. I've met several teams that ran single-direction by default and explained, at length, why their last fault investigation took three days instead of three hours.

The wider point about senior expertise

The mistakes here aren't about reading the OTDR wrong. They're about reading the OTDR fast. Experience makes you faster. Faster makes you skip steps. Skipping steps makes you wrong in ways you can't see until somebody less experienced runs the bidirectional check and tells you what the trace actually said.

Senior techs aren't immune to this. We're more vulnerable to it, because we've earned the right to trust our pattern-matching, and the pattern-matching is exactly the part that fails on ghosts, dead zones, and gainers.

Run the bidirectional. Use the launch lead. Average the splices. Even when you don't think you need to.

Especially when you don't think you need to.