The optical time domain reflectometer (OTDR) has been an important tool for fiber optic testing and troubleshooting since its invention 25 years ago. We have discussed this several times in this column and in other Electrical Contractor articles, so I hope you are already familiar with them.

You should know by now that OTDRs use backscattered light from the fiber to imply fiber attenuation and splice or connector loss, and they are not an acceptable substitute for insertion loss testing. You should also know that connectors are characterized by a large reflective peak on the OTDR trace caused by the imperfect joint between two fibers.

That intensity of the reflection from a connector joint can be measured to determine the reflectance or return loss of the joint. If the reflection is too high, it can overload the OTDR receiver and cause errors in attenuation measurements or diminish its ability to resolve close events. You can tell when a reflected peak is overloading the OTDR because the peak will be flat on top.

But, a more confusing effect of reflections is what we commonly call “ghosts.” These are confusing events that show up on the OTDR trace that really aren’t there. Ghosts are caused when the OTDR tests a cable with two highly reflective events, one of which is often at the OTDR interface.

Since the fiber optic connector on the OTDR will have cables plugged into it every time the OTDR is used, it generally becomes dirty and scratched, even if it is regularly cleaned—as it should be. Likewise, the launch cable used to reduce the effect of the reflection at the OTDR connector often suffers from too many connections to cables under test, so it becomes reflective too.

When the OTDR sends a test pulse down the cable, the big reflection from the far end comes back to the OTDR where it shows up on the trace as a overloaded reflection, then is reflected from the OTDR interface back down the cable for a second trip, effectively becoming a second “test pulse.”

From there it is reflected back from the far end yet again, going back to the OTDR to be recorded as a second trace. If the reflections are big enough, this process can happen three or four times, each time producing a “ghost” event on the OTDR trace.

A ghost trace looks like the diagram above, which shows the absolute worst case, a cable plugged directly into the connector on the OTDR. You can see the initial pulse at the OTDR is flat-topped, as is the first pulse at the end of the cable, indicating both are beyond the range of the OTDR and are saturated.

The colored arrows allow you to follow the path of the OTDR test pulse. Red shows the initial outgoing pulse and its first reflection; green is the reflection from the OTDR that makes the second trip causing the ghost. The ghost appears at exactly twice the length of the actual cable on the OTDR trace. This location is the way most OTDR ghosts are unmasked.

The uninitiated OTDR user might think this trace shows a fiber with a break in the middle—a common assumption. I’ve known several users who replaced cables under these circumstances, when they mistakenly used an OTDR where a meter and source should have been used for a insertion loss test.

One user asked me to test such a cable, wondering why it showed 20 dB loss in the middle of the cable on the OTDR trace. When an insertion loss test showed normal readings for a terminated cable, he was surprised.

“Did you check the length on the OTDR?” I asked.

He retested the cable with the OTDR and he was astounded to find the trace to the “end” was twice the actual cable length, exactly what you expect with an OTDR ghost. He was tricked into thinking the cable had a break rather than checking the length and comparing it to the known length of the cable.

Of course, this is another good point: Always compare test data to cable plant documentation. If you know the cable is a given length, you will look at that point for the trace and not be fooled by ghosts. Using an OTDR to measure cable length without documentation is asking for trouble.

And always, without exception, never connect a cable under test directly to the OTDR. Using a launch cable of known length and knowing the approximate length of the cable(s) under test will help isolate ghosts and prevent their confusing the actual test data.

For more information on OTDRs, see the online tutorial at www.lennielight     EC

HAYES is a VDV writer and trainer and the president of The Fiber Optic Association. Find him at