Ghosts and Goblins: How OTDRs Work, Part 4

By Jim Hayes | Aug 15, 2019
Ghosts 1
Web Exclusive Content




One of the first customer visits I made after starting a company making fiber optic test equipment was in the basement of a government building in Washington, D.C. I brought along one of our test kits for demonstration, with a fiber optic power meter, test source and the accessories needed to test the loss of fiber optic cables. 

The customer we were visiting was new to fiber optics. However, he had a big budget for equipment and had already bought one of the early OTDRs. When we arrived, he was in the middle of installing a fiber optic cable between two labs. It was the second time they had done that, he said, because the first cable had a break in the middle of it and had to be removed and replaced.

I asked how he knew it had a break in the middle? He replied that they had tested it with their new OTDR and it showed clearly that it was broken. He even showed me a trace from the OTDR that proved that. It looked like this:

“See the end of the fiber where it’s broken, the big loss after that and the end of the fiber reflection after that?” he said.

Now at this point, we were new to fiber optics and were not that familiar with OTDRs, but I did know how to read the display. “How long was the cable you were installing?” I asked. The answer was 200 feet.

I pointed to the OTDR trace distance scale. This trace showed the first peak, where he thought the fiber was broken, was at 200 feet. The second peak was at 400 feet, twice the length of the cable. They were misreading the OTDR trace and the cable was perfectly OK.

We also tested the cable they thought was bad with a meter and source. The fiber was transmitting light perfectly well and the measured loss was quite low. There was nothing wrong with that cable. What was wrong was their interpretation of the OTDR trace. 

What was happening was really obvious. The light reflected from the connector at the end of the cable was very high. It was reflected back toward the OTDR and then reflected back to the far end and then reflected back again to the OTDR where it showed up on the trace as a peak, bouncing back and forth in the fiber. 

The trace did indeed look like there was a break in the cable, but the location of the first peak was a dead giveaway—it was at the actual length of the cable. The second peak was twice that length, much farther than the cable was long. Here is a the OTDR trace shown above with explanations. 

This anomaly in OTDR traces has become known as a “ghost.” Although OTDRs have gotten much more sophisticated, ghosts remain one of the great nuisances of OTDR testing, confusing practically every OTDR operator and anyone looking at OTDR traces. 

Ghosts are much more common in premises cabling when OTDRs are used on multimode cabling because multimode connectors have higher reflectance and the distances are shorter, so reflected signals are not attenuated sufficiently to fade into the background. But we’ve seen ghosts on long single-mode runs when you have high reflectance events too, as shown in the trace below. 

There are several ways to determine if a peak on the trace is a ghost. The peak’s location is usually the simplest way. As you can see in the trace above, the spacing of peaks in the real cable trace is duplicated in the spacing of the ghosts. In addition, if a ghost shows up on top of a real trace, it will look like a connector, but it will have no loss because there is no connector there. Loss readings won’t make sense either.

Ghosts are an excellent example of why cable plant documentation is so important. With documentation, you know what you should be seeing, so anomalies like ghosts are much less confusing. 

But even documentation won’t solve every problem. One of the Fiber Optics Association (FOA) master instructors had a very good story to illustrate that. He had retired from AT&T after many years of service and was helping clients with testing problems and teaching FOA certification courses. At AT&T he had developed their training program for OTDRs and taught most of their techs in the United States, so he became the FOA’s resident expert on OTDRs.

One of his clients, a metro transportation authority with hundreds of miles of fiber along transit lines, had tech with a testing problem they could not solve. They were installing cables along a length of rail line and were testing at one of the intermediate stations. They called our guy because they thought there was a break in a cable at about 4,500 feet along an 8,000-foot length of cable, according to OTDR tests. The OTDR traces showed the end of all fibers in the cable at 4,500 feet, but when they went to where the break was indicated, there was nothing to indicate any problems. So, they called this FOA instructor for help.

When he arrived, he had them explain the problem, looked at the documentation and then started testing. His OTDR traces showed no problems along the 8,000-foot span. Everybody was confused. Then he looked at the documentation and realized that the cable going in the other direction out of the station was 4,500-feet long. He put an OTDR on that span and it looked good too. Then they all realized what they had been doing—looking in the wrong direction—not paying attention to the documentation they had available.

Sometimes the obvious solution isn’t so obvious.

About The Author

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





featured Video


New from Lutron: Lumaris tape light

Want an easier way to do tunable white tape light?


Related Articles