Here is one word of advice I offer to anyone asking about maintenance of fiber optic networks: DON’T! Some people have suggested fiber optic networks need periodic inspection of connectors, mating adapters and even testing or taking optical time domain reflectometer (OTDR) traces. That advice is misguided. It could hurt the network or cause you to be sued by an irate owner whose network you bring down or cable plant you damage.
Do you think the telcos have crews out checking fiber networks to see if the connectors and splices are OK? How about the military on tactical systems in Iraq? Is there a Captain Nemo aboard the Nautilus checking submarine cables? Of course not.
Fiber systems are designed to be installed and never touched unless something damages them—e.g., the infamous “backhoe fade” of buried outside plant cables.
In the early days of fiber optics, some network owners tried building automatic monitoring systems to keep tabs on the loss of the cable plant. That idea faded when fiber proved to be more reliable than copper cabling and the network communications manufacturers built into their equipment monitors for data transmission, a better indicator of problems.
Let me give you some reasons why you do not want to try to perform maintenance on any fiber optic network.
Most inspection procedures require bringing the network down, unacceptable in almost every instance. Telcos have backup links running alongside operational links, and the equipment will switch over to the backup if it senses high errors on the main link. Do you know any premises networks set up like that? Want to bring down a gigabit LAN backbone fast? Unplug a fiber optic connector to inspect it with a microscope. See how long it takes the network manager to find you.
Most harm to installed fiber optic systems (and copper also) is done by unskilled or clumsy personnel during handling. I heard of one network that crashed when a company executive disconnected a fiber connector to show it to a visitor being escorted around the facility. I know of workers accidentally backing into patch panels and breaking cables at the junction to the connector. I have seen connectors dropped on the floor, breaking the ceramic ferrule. I have helped troubleshoot broken fibers in splice closures caused during repairs of other fibers.
It is easy to get dirt into mating adapters or on connectors whenever they are exposed to the air. Fiber technicians are taught to keep connections clean after termination, cover connector ferrules and mating adapters with dust caps and clean the ferrule end whenever it is opened to the air. If dirt is such a big problem (and airborne dirt is the size of the core of single-mode fiber), why risk contaminating operating connectors by exposing them to the air to see if they are dirty?
Mating and unmating may wear the connector interfaces, affecting optical performance. Ferrule end faces rubbing against the mating connector and the outside of the ferrule scraping materials off the alignment sleeve in the mating adapter—especially with adapters using cheap plastic alignment bushings, which are good for only a few mating cycles—can cause higher loss.
Links operating at gigabit and higher speeds generally use 850 nm VCSELs, which are relatively high-power lasers at a wavelength near the high end of human eye sensitivity, still visible to some people. Using a high-power microscope, such as a 400×, concentrates the light into the eye, increasing the risk of eye damage, especially if you are not able to see this wavelength. If a link being inspected is “hot,” the consequences could be bad. Anyway, a 400× microscope is overkill—it’s the maximum magnification you would use to inspect single-mode connectors during termination; 100–200× generally is considered the maximum for connector inspection.
The fiber link loss may be different when a link is reassembled after inspection, especially with connectors that have spring-loaded ferrules like STs. Inspecting a connection could lead to higher loss than initially measured and potentially affect data transfer on systems such as gigabit Ethernet and 10G Ethernet where loss margins are very low.
As for testing with an OTDR for maintenance inspection, well, some telcos do that automatically on spare fibers in outside plant cables that run tens or hundreds of kilometers through desolate regions. An OTDR is inappropriate for most premises systems under any circumstances (as I have discussed in several columns this year) and often causes more problems than it solves.
Finally, if you have a problem with dust in a telecom closet, room or data center, you have a poorly designed facility that should be fixed with proper sealing, filtration and air conditioning. You should not try to fix it with a feather duster.
One more time: What periodic maintenance should be done on fiber optic networks? All together now: NONE! EC
HAYES is a VDV writer and trainer and the president of The Fiber Optic Association. Find him at www.jimhayes.com.