Welcome to Fluke Fiber 08: OTDR Trace Analysis & Troubleshooting
Course Duration: 35–45 minutes
Running an OTDR is easy. Reading the trace is the skill. This course covers the anatomy of a trace, event shapes, loss and reflectance values, the traps (gainers, ghosts, dead zones), turning optical distance into a location on site, and a repeatable troubleshooting workflow using the OptiFiber Pro.
What You'll Learn:
- Read a trace from front end to tail end
- Identify connectors, splices, bends, breaks, and APC events
- Judge event loss and reflectance against the limits
- Recognize gainers, ghosts, and dead-zone masking
- Turn OTDR distance into a location and troubleshoot to a documented repair
Course Structure
- Anatomy of a Trace
- Reading Events
- Reflectance & Loss Values
- Gainers, Ghosts & Other Traps
- From Trace to Location
- Troubleshooting Workflow & Case Studies
- Final Assessment — 10 questions, 80% (8 of 10) to pass
Each section ends with a knowledge check. You get two attempts, and a first miss gives you a hint, not the answer. Finish the check to unlock the next section. You'll enter your name and email just before you submit the final assessment. Your results are emailed automatically to the training administrator, with a copy to you, whether you pass or not.
Section 1: Anatomy of a Trace
A trace plots returned power (dB, vertical) against distance (horizontal). Read it from left to right:
| Section | What you see |
|---|---|
| Front end | A large initial reflection from the OTDR port, followed by its dead zone |
| Launch cord | Clean backscatter sloping gently down |
| Link start | The first link connector: a reflection and a loss step. This is why the launch cord exists. |
| Link | Backscatter slope, with events along it: splices, connectors, bends |
| Link end | The last link connector, measured against the tail cord's backscatter |
| Tail end | The fiber end reflection, then the noise floor |
The slope of the backscatter is the fiber's attenuation. For example, a 0.5 dB drop over 1.5 km is about 0.33 dB/km, which is normal for singlemode at 1310 nm. A slope that's much steeper than the fiber's specification points to a bad section, stress, or the wrong fiber type.
The OptiFiber Pro's EventMap turns all of this into icons with pass/fail colors. Learn to read the raw trace anyway, so you can confirm what EventMap says.
Section 2: Reading Events
| Event shape | Typical meaning | Check |
|---|---|---|
| Spike, then a small step down | A mated connector (UPC) | Loss and reflectance against the limits |
| Step down, no spike | A fusion splice, or a bend | Compare 1310 vs. 1550 nm: a bend is much worse at 1550 |
| Spike with a large step | A dirty, damaged, or poorly mated connector, or a mechanical splice | Inspect and clean, then re-test |
| Large spike, then noise | End of fiber, or a clean break | Compare the distance with the expected length |
| Gradual drop with no reflection, then noise | A crushed or shattered break, or a bad bend that kills the signal | Locate it with the distance; use the VFL if it's close |
| Small or no spike at an APC connector | Normal: APC angles the reflection away | Don't mistake it for a splice; the OFP lets you edit the event type |
Section 3: Reflectance & Loss Values
Event loss
The step in backscatter across an event. Compare it with the limit for that component. TIA-568.3 allows up to 0.75 dB for a mated connection and 0.3 dB for a splice, and many project specifications are tighter.
Reflectance
How much light an event reflects, as a negative dB number. More negative is better (−55 dB reflects far less than −35 dB).
| Connector | Typical project-spec maximum reflectance |
|---|---|
| Multimode | −35 dB |
| Singlemode UPC | −40 dB |
| Singlemode APC | −55 dB |
These are typical specification values. Use the limits in your project specification. High reflectance (a number closer to zero) usually means a dirty, cracked, or air-gapped connector, even when the loss looks acceptable. It matters most on singlemode, video, and high-speed links.
Section 4: Gainers, Ghosts & Other Traps
| Trap | What it looks like | How to handle it |
|---|---|---|
| Gainer | An apparent gain at a splice | Fibers with different backscatter are joined. Average both directions (SmartLoop) to get the true loss. |
| Exaggerated loss | The same splice tested from the other end shows too much loss | Same cause as a gainer. The bidirectional average is the true value. |
| Ghost | A reflection with no loss, at a multiple of a real reflection's distance, sometimes past the fiber end | Light bouncing between two strong reflectors. Confirm by testing from the other end or changing the settings. Clean the strong reflectors that cause it. |
| Dead-zone masking | Two events close together look like one | Use a narrower pulse (Expert mode) or DataCenter mode on short, dense links |
| Noise | A ragged trace far down the link | Use longer averaging, a wider pulse, or test from the other end |
Section 5: From Trace to Location
The OTDR gives optical distance. Turning that into a spot on site takes care:
- Launch cord: if launch compensation is off, subtract the launch cord's length. Example: an event at {EVT} m with a {LAUNCH} m launch cord is about {INTO} m into the link.
- IOR: set it from the cable data sheet. A wrong IOR scales every distance.
- Fiber vs. sheath length: stranded and loose-tube fiber is slightly longer than the cable sheath (the helix factor), so the fault may be a little closer than the optical distance suggests. Use the sheath's length markings to close in.
- Slack and routing: slack loops, splice-case storage, and vertical runs all add fiber that isn't obvious on the drawing.
- Close in: test from both ends, compare with known landmarks (splice cases, panels), then use the VFL for the last few meters.
Section 6: Troubleshooting Workflow & Case Studies
Workflow
- Tier 1 FAIL or a service complaint? Inspect and clean the end connectors first. That's the most common cause.
- OTDR both wavelengths (SmartLoop if possible) to find the event and classify it.
- Locate it using distance, landmarks, and the VFL.
- Repair: clean, re-terminate, re-splice, or relieve the bend.
- Re-test with Tier 1 and Tier 2, and save both the before and after results.
Case studies
| Symptom | Trace clue | Cause and fix |
|---|---|---|
| Backbone fails at 1550 nm only | An event with much more loss at 1550 than at 1310 nm, at 88 m | A tight bend at a tray entry; re-route it within the bend radius |
| High loss on every fiber at one panel | A large spike and step at the same distance on every fiber | Contaminated patch-panel adapters; clean, inspect, and re-test |
| One splice reads 0.6 dB from one end, a gainer from the other | The bidirectional average is about 0.15 dB | Different backscatter between fiber types; the splice is fine |
| Link shorter than drawn | A large reflection then noise, well before the expected end | A break; locate it from the distance and repair it |
Final Assessment
Answer all 10 questions. You need 80% (8 of 10) to pass. These questions are different from the section knowledge checks. Enter your name and email below the questions, then submit.
Your Information
When you submit, your results are emailed automatically to the training administrator, and a copy is sent to the email address you enter here — whether or not you pass.
Assessment Complete
Score: 0 / 0
Percentage: 0%
Status: —
Time in course: —