The most common causes of fiber optic malfunctions
Anyone who’s ever done network troubleshooting knows it’s a complicated process, so it’s helpful to know where to start looking for a problem. To help you make an educated guess about the cause of your network’s troubles, here are some of the most common fiber optic cable problems with their possible causes:
- Broken fibers because of physical stress or excessive bending
- Insufficient transmitting power
- Excessive signal loss due to a cable span that’s too long
- Excessive signal loss due to a contaminated connector
- Excessive signal loss due to faulty splices or connectors
- Excessive signal loss due to having too many splices or connectors
- Faulty connection of fiber to the patch panel or in the splice tray
Typically, if a connection is completely dead, it’s because of a break in the cable. However, if the connection is intermittent, there are several possible causes:
- The cable’s attenuation may be too high because of poor quality splices or too many splices.
- Things like dust, fingerprints, scratches, and humidity can contaminate connectors.
- There is low transmitter strength.
- There are bad connections in the wiring closet.
When I’m called in to fix a problem in an environment that I’m not thoroughly familiar with, the first thing I do is gather some basic information about the problem’s symptoms and possible causes. As with any other medium, the key to getting good information is knowing the right questions to ask. Here are a few questions that can get you started.
Has anyone disconnected, reconnected, or moved the PC recently?
It’s important to find out if the PC has been disconnected or has moved recently. If the fiber optic cable has been disconnected from the PC, it could be that the cable was never reconnected, was reconnected incorrectly, or was contaminated during the time that it was disconnected.
Have there been any changes to the PC’s hardware?
Upgrading a PC’s hardware can also cause problems. The cable could have been disconnected and either not reconnected, reconnected improperly, or contaminated prior to being reconnected. It’s also entirely possible that the cable was never disconnected during the hardware upgrade. If that’s the case, there’s a chance that the cable could have been overextended when the PC was moved, or the PC could have been accidentally smashed against the back of the desk or the wall, damaging the connector.
It’s also possible that the cable was never damaged or hooked up incorrectly, but rather that the new hardware is preventing the NIC from working correctly. The exact method by which you’d determine whether or not the NIC is having problems varies with your operating system. However, if you’re using Windows 9x, Me, 2000, or XP, you can use the Device Manager to examine each of the system’s hardware devices to see what is and isn’t working.
I also recommend temporarily removing the new hardware and returning the system to its previous state, if possible. This could show if the system’s new hardware is causing the problem, or if the problem is related to something else.
Has any furniture been moved recently?
I can’t count the number of times that I’ve seen copper network cables damaged because someone scooted a desk out from the wall or because the cleaning crew moved a desk to vacuum. When you move a desk without disconnecting the network cable, the cable may become overextended, or the desk could smash or crimp the cable. If something like that can destroy a copper cable, imagine what could happen to a fiber cable, which is made out of glass.
Has anyone, such as the telephone company, been performing any work in the building?
This may seem like a strange question to ask, but it’s my experience that network problems often coincide with visits from the phone company. I’ve had phone company technicians cut cables, disconnect cables, and do just about anything else that you can imagine. I won’t pretend to understand the logic behind this behavior, but if you’re having network problems and someone tells you that someone was in the building doing phone or electrical work yesterday, then you’ve got a good place to start your troubleshooting.
Has the cable been stepped on, had a chair rolled over it, or suffered any other physical stress?
It never fails to amaze me what end users will admit to when you ask the right questions. If you ask a user in a nonthreatening manner about any physical stress that the cable might have endured, they’ll usually tell you. Remember that most users have no idea that a fiber cable can be destroyed if stepped on or bent too sharply, or if they roll a chair over it.
A quick-and-dirty test
Time is usually a critical factor when I’m troubleshooting an on-site problem. Users need to be back online quickly, and I usually have plenty of other jobs waiting. Because of this, I try to diagnose the problem as quickly as humanly possible.
One way is to begin the process with a highly unscientific but effective test. I disconnect both ends of the fiber cable and hold a laser pointer up to the cable to see if the light comes out the other end, as shown in Figure A. If you don’t have a laser pointer, a bright flashlight works just as well, as shown in Figure B. Fiber optic cable is designed to conduct light, so you don’t have to worry too much about exactly lining up the cable with the light source.
Figure A: Use a laser pointer to test to see whether or not a cable has suffered a complete failure.
Figure B: If you don’t have a laser pointer, a bright flashlight will do.
If no light comes through the cable, then the cable is broken and needs to be replaced. If light does come through the cable, it doesn’t necessarily mean that the cable is good; it just means that the fiber inside the cable hasn’t been completely destroyed. However, if the light comes through the cable and the cable run is shorter than a hundred meters, then the cable will often be good enough to use.
Additional diagnostic techniques
If the laser pointer or flashlight can pass light through the fiber optic cable, then it’s time for some further diagnosis. I tend to favor the process of elimination when diagnosing fiber optic problems.
I usually begin by disconnecting the troubled computer from the fiber optic cable and connecting a PC or laptop--that I know to be good--to the cable. If the good PC is able to function, then I know that the cable isn't the problem. It could be due to a damaged NIC, or a configuration problem could be preventing the computer from recognizing the NIC properly. For example, the NIC might be conflicting with another device in the system, or a necessary DLL file might be corrupt.
If the good computer also fails to function when attached to the fiber optic cable, then the problem is either in the cable itself or in the port that the cable is connected to at the other end. Leave the good PC connected to the cable and attach the other end of the cable to a known good port. If the PC still fails to function, then the cable is bad. If, on the other hand, the PC begins to function, then the port is bad.
Sometimes the process of elimination just isn’t an option. For example, you might not have a spare computer or a spare port available, or the fiber optic cable may span too long a distance to make this type of troubleshooting effective. If this is the case, there are other troubleshooting methods and tools that you can use.
You probably noticed in my list of common fiber optic cable problems that signal loss is a common culprit in fiber problems. Fortunately a technique called loss testing allows you to test a fiber optic cable to measure signal loss.
The technique involves connecting a source generator to a power meter. You can set the dBm range and the wavelength that you want to test on the source generator. You then attach the source generator to the power meter, which measures the signal and reports on the signal’s losses. Typically, the process involves using a cable you know to be good as a reference to establish a baseline measurement before you test an unknown cable. If the unknown cable is good, then its losses should be similar to the losses of the reference cable.
Unfortunately, I can’t tell you how much of a loss is too much because it depends heavily on the cable and on the physical environment. However, there are some guidelines that you can use. Keep in mind, though, that these are just guidelines: Depending on your cable type, your numbers could be radically different from mine:
- Each connector gives about a 0.5-db loss up to a maximum of about .75 dB.
- Each splice results in a loss of about 2 dB.
- If you’re using single-mode fiber, you can plan on losing 0.1 db for every 600 feet of cable.
- If you’re using multimode cable, plan on losing about 0.1 dB per 100 feet of cable.
The actual type of test equipment you’ll need depends on what type of cable and connectors you’re using. Obviously, not every connector will attach to every tester. Likewise, you can’t expect a tester that’s designed for single-mode cable to test multimode cable. There are a wide variety of fiber optic testers on the market. Some test only a single type of cable, while others are designed to function as all-in-one testers. In my opinion, Fluke Networks makes the best network testing tools available.
A fiber optic microscope
As I’ve mentioned, fiber optic contamination is also a problem. Fiber optic contamination is caused by things like dust, scratches, or epoxy on the ends of the fiber optic connectors. The best way of testing for such contaminants is by using a fiber optic microscope.
A fiber optic microscope is a special microscope designed specifically for inspecting fiber optics. Traditional fiber optic microscopes range in price from just under US$100 to about US$800.
If you’re considering the purchase of a fiber optic microscope, there are several things to look for besides price. The first thing that you should look for is a connector to hold the cable in place. Many of the cheaper portable microscopes require you to hold the cable in front of the microscope. At high magnification levels, it’s almost impossible to hold the cable and the microscope still enough to perform a thorough inspection.
The next thing that you should look for is the magnification level. Typically, the magnification levels of fiber optic microscopes range from 100X to 400X. The higher the magnification level, the more detail you’ll be able to see when inspecting the cable.
Finally, you should check to see if the microscope is optical or electronic. Some higher-end fiber optic microscopes actually display the image on a video monitor rather than requiring you to look through a microscope barrel. Such systems typically display a 3 mm cable end magnified to the size of a tennis ball.
A company called Prior Scientific has even produced a software-driven fiber optic microscope. This microscope takes all of the subjectivity out of cable inspections and certifications. The software looks for specific things in the image produced by the microscope and makes a decision based on the image.
It’s not your fault
Like the glass upon which it’s based, fiber optic cable is very fragile. You can tie copper cable in knots and still use it, but you have to treat fiber optic cable with kid gloves. When you start using fiber, you’ll eventually run into problems with it. Knowing why fiber fails and what to do about it will help you solve the problem without tearing the fibers out of your head.