top of page

Fiber Optics - OTDR Test Equipment

Optical time domain reflectometers (OTDRs) are critical instruments for evaluating optical fiber spans. Like any technology, they are not infallible. As the term ‘reflectometer’ implies, OTDRs work by analyzing optical signals which are reflected (and scattered) back to them. Some people call them “cable radar” because they analyze a reflected signal and not the transmitted signal. Due to this, OTDRs will sometimes generate measurement “artifacts”. One such artifact is a secondary reflection, or an “echo” - often referred to as a GHOST! After reading this article, you won’t be afraid of no ghosts!


The most common ghosts occur on short spans with a highly reflective open fiber end. In this case, the end of the fiber at distance L1 will generate a high-powered reflection that will arrive back at the OTDR connector. Much of it is received, but some will reflect back toward the end of the fiber, will reflect off the endpoint again, and travel back to the OTDR having made two full round trips. In this case, it will be plotted at a distance of L2.

Remember that the OTDR plots a trace based on the time of arrival of any scattered or reflected power. It doesn’t know distance; it converts the time to distance. So even though the reflected photons in both the first and second spike came from distance L1, the OTDR had no idea that some of the photons had taken two round trips, so it plotted them at L2. Note that in all these cases, the backscatter plots normally (produce a normal OTDR ‘trace’) and the reflections, both primary and echo are superimposed on the OTDR signature trace.

This one is simple to analyze. Since it plots at a distance that is longer than the fiber span and appears out in the “noise” it can’t possibly be a real event, so we just call it a ghost and forget about it. Also, note that most modern OTDRs have the ability to identify these logically and there is often a “Ghosts on/Ghosts off” setting in the OTDR menu. Setting this to OFF will not remove the spike from the trace, but it will prevent the artifact from being reported in the table as an event.


If there is a connector at distance L, perhaps at midspan, and the connector has enough reflections, it could generate an optical pulse or spike that will make two round trips and plot again at 2L. But in this case, the spike will fall on the fiber span. It is easy to mistake this for another connector but there are two hints that it is not. First, it will be exactly 2 times the distance from the OTDR as prior large reflections from an actual connector. Secondly, it will not show any loss.


Sometimes two big ghosts can get together and generate a little ghost. These can plot in a number of places depending on the relative distances of the mama and papa ghost, but they will stay near their parents. Consider the span below with a reflective connection at distance X and a reflective connection at distance Y. A ghost that makes a double round-trip between the OTDR connector and X, and the OTDR connector and Y, will plot at 2X and 2 Y respectively. But there can sometimes exist a “harmonic” or a “hybrid” ghost that reflects from the OTDR to Y, back to X, then back to Y, then back to the OTDR.


There’s nothing wrong with ghosts if you can recognize them. But wouldn’t it be nice to eliminate them? Some of the best ways to eliminate ghosts are listed below:

  • Reduce the pulse setting. If possible, reduce the pulse setting to the shortest pulse appropriate to your task. Reflections are directly proportional to pulse energy, so if you limit their energy initially, they can’t travel as far before dissipating.

  • Reduce the reflections at the OTDR. OTDRs with APC connectors limit ghosts significantly compared to UPC connectors.

  • Reduce the far-end reflection. Place a “terminator” or “terminating connector” on the far end of the system, or the far end of the receive cable. These look like a regular connector but act like a “light sump” with ratings of -55 dB reflection or better. They reduce the end spike, and ghosts and tend to reduce noise on OTDR traces.

  • As a last resort, make a hasty, “field” terminator. A patch cord with severe bending will attenuate any reflections from its end. Attach a patch cord to the system and wrap it very tightly or tie knots in it (don’t plan to use it for anything else after this!). This is hard to do with bend-insensitive fiber and especially at 1310nm.

  • Finally, when busting ghosts… DON’T CROSS THE ENERGY BEAMS!


bottom of page