Basics: Testing Cables

A network consists of end-user workstations connected to servers by what might appear at first to be a tangled web of wires and cables. If the building or campus is wired correctly, however, this is not a jumble of cables joined together in a spaghetti fashion, but is an orderly collection of components much like a spider's web, fanning out to connect everyone in a hierarchical manner. In addition, wireless networking components have added an entire new territory, and tools are currently being developed to address troubleshooting this new area of network technology.

For the most part, when you begin to build a network, the first thing you have to do is install the cables that will connect the servers and workstations. This can be done when a building is being constructed, as is the case in most office buildings today. Or it can involve placing cable ducts in ceilings and knocking out areas of the walls to install faceplates where the cables terminate. Either way, before you begin to connect end users to the network, you first have to test the installed cables to be sure they are performing as expected.

Devices that can be used to test cables (both copper wire and fiber-optic cables) range from very inexpensive handheld devices that a cable installer can use to check his work, to very expensive devices that require a skilled technician to perform the tests and understand the results. Things that are usually tested include the following:

• Cable length The physical network topology restricts the length of certain segments in the network. If you make your own cables, a common error may result from trying to stretch the limits of the topology and create a cable that's just a few meters too long. If a desktop is just a few meters farther from a switch than the standard allows, you may get complaints from that user!

• Resistance Electricity encounters resistance as it travels along a copper wire.

• Noise Interference can come from other cables that are bundled together or from outside sources, such as fluorescent lighting, nearby welding, strong sources of electromagnetic frequencies, and other high-voltage electrical sources located near the network cabling.

• Attenuation As the cable encounters resistance traveling down the wire, and as part of the signal radiates out of the wire, the signal weakens. This is a normal side effect of using copper wiring instead of fiber optics. You can expect copper wiring to work best at the standardized lengths, and take your chances at extending that length.

• Near-end cross-talk (NEXT) From the transmission end of a cable, it is necessary to remove the surrounding material that encloses the copper wires and attach each wire to a pin in the cable connector. Because the strength of signal is strongest at the end of the cable where the electrical signal is generated, there is a greater potential for interference between the wires at this end of the cable.

To get a better understanding of the kinds of problems you might experience, see Chapter 6, "Wiring the NetworkCables, Connectors, Concentrators, and Other Network Components."

Two basic instruments are used for testing cables. The first is the simple cable checker, which is used to determine that the cable actually provides an electrical path from here to there. The second is the cable tester, which determines whether the cable has been installed correctly to support the topology of your network, taking into consideration things such as cable length and cross-talk.

Handheld Cable Checkers

A cable-checker device is usually a small battery-operated unit that is used to check STP or UTP cables. This simple test is usually done when cables are first installed as a quick check to be sure that the process of pulling the cables through the ceiling or walls has not damaged them.

If the cable is already attached to a network device, you have to disconnect it and attach it to the unit. A cable checker operates by placing a voltage on a wire and determining whether it can be detected at the opposite end. This can be used to determine whether the cable has a break anywhere along its path and whether you are looking at the same cable on both ends when several cables are traversing a single path. Most cable checkers consist of two components, which you attach to opposite ends of the cable.

Cable Testers

A cable tester is a small step up from the basic checker. This device can be used to measure NEXT, attenuation, impedance, and noise on a line. Some cable testers even perform length measurements, of both the total cable and the distance to a fault on the cable, such as a kink in the wire that is causing reflections of the signal to radiate back to the transmitting side of the cable. Another function you might see is wire-mapping, which checks to be sure that the correct wire-pairs in a cable have been mapped to the correct pins on the connector attached to the end of the cable. In cables used for 10BASE-T networks, for example, the standard specifies specific pairs of wires in the cable that must be used for transmitting and receiving data. The actual decisions about which pins are chosen for a particular connector are not made arbitrarily. If the wires are not correctly mapped to the pin-out on the connector specified by the standard, the cable might generate errors due to noise or cross-talk.

Small handheld instruments like these usually have LED lights that indicate a pass or fail condition for the test you are performing. They do not require a keyboard or monitor to display data. Some have a small screen that displays limited text, sometimes showing the suspected type of error that has caused a fail condition. Most are battery powered and can use an AC adapter, which makes them useful portable instruments for installing or troubleshooting cabling.

When you begin to go up the price ladder for these types of instruments, you will find some that can perform more advanced monitoring functions, such as showing network use and Ethernet collisions. Another useful feature to look for if you can afford it is the capability to log data to a memory buffer for later review. Some cable testers are even capable of connecting to a PC or printer to produce a written report. This allows you to leave the device connected for a while to monitor a line.

Depending on the capabilities of the particular device, you can expect to pay from several hundred dollars up to a thousand or more for a good cable tester. When evaluating products, be sure to compare features. Price doesn't always reflect the quality of a device. And you should carefully check the literature and documentation that is available for each device when making a purchasing choice. Although some features, such as the capability to produce a written report, might sound great, do you really need that capability? In a large network, probably so; in a small one, probably not.

Bit Error Rate Testers (BERT)

Data travels through the wire (or the fiber) as a series of signals that indicate a single bit, representing either zero or one. The statistic called bit error rate (BER) is calculated as a percentage of bits that have errors when compared to the total number of bits sampled:

        number of bit errors during sampling interval BER =  ---------------------------------------------         total number of bits transmitted 

Whereas LAN analyzers operate on data captured from the wire in units of frames (depending on the LAN protocol, such as Ethernet or Token-Ring), a bit error rate tester (BERT) performs a more basic function to determine whether the line is capable of carrying the network signaling at the bit level with a minimum of errors.

This kind of instrument is normally used when installing a connection to a network service provider, and it might be used to demonstrate the quality of service that the provider establishes for your link.

The instrument used to perform this kind of error detection usually does so by generating a specific bit pattern on the line and then checking it at another location to compare the generated signal with that which is received. A pseudorandom binary sequence (PRBS) of bits is produced by the instrument. It is pseudorandom because it simulates random data. However, because the pattern is also known by the receiving connection so that it can make the comparison, it's not truly random, but instead is a predefined pattern. Other tests include sequences of specific bits, either zeros or ones, for extended periods, or specific user-defined bit patterns.

When you have a line that exhibits a high bit-error rate, using a slower transmission speed usually improves performance. This is because when you lower the number of errors that occur, higher-level protocols do not have to resend packets as often to compensate. Although one bit error in a frame usually is easily recovered by a network protocol using an error correction code (ECC) technique, multiple bit errors might be all that it takes to cause an entire frame of several hundred thousand bits to be re-sent.

Time Domain Reflectometers

A signal usually propagates down a wire at a constant speed, provided that the impedance of the cable is the same throughout its journey. When the signal runs into a fault in the wire (such as a kink or a splice) or reaches the end of the wire, part or all of the signal is reflected back to its origin. Similar to radar, instruments that use time domain reflectometry (TDR) to make cable measurements are based on precisely timing the signal pulse as it travels through the cable and back.

Of all the instruments you can use to test cables, TDR is one of the most accurate and fastest. It can help locate faults due to various causes, such as these:

• Wires that have been spliced together

• Moisture trapped in the cable

• Cables that have been crushed or have kinks in them

• Short circuits

• Problems in the sheath surrounding a cable

• Loose connectors

You also can use TDR to measure the length of a cable that has no faults. This can be useful for inventory functions because you can even use it to measure the length of a cable while it is still on a reel to determine whether you have enough or need to order additional stock before beginning a major wiring project. TDR can be used to take measurements on twisted-pair cables, coaxial cables, and even fiber-optic cables. Because fiber-optic cabling is the most expensive (but largest bandwidth) media today, you should consider investing in a TDR that supports fiber-optic media if you are planning on a wide deployment of fiber-optic cabling. At the very least, you should expect a third-party installer of a large cable plant to provide such an instrument. You should also make it part of your project plan to record the statistics provided by the installer. Thus, if a future need dictates, you might have a particular cable segment re-evaluated. Be sure to stipulate such things as the performance of any cable segments in your agreement with an installer. If you are performing the installation yourself, be prepared to use a TDR to check a cable segment should performance degrade. For example, for electrical cabling, other devices (even cables) might be installed later that interfere with your original installation. In that case, moving cables to new locations should solve the problem.

The more expensive models of this instrument can be equipped with a CRT or LED display that shows the wave form of the signal and any reflected signals. The more common instrument displays the number of feet to the end of the cable or a fault, and might have an indicator that tells you the type of fault. By showing the number of feet to a perceived defect, you can trace your installed cabling (or unroll your on-the-roll cabling) so that you can get to the point where the defect occurs.

Using a TDR in this method can help you with installed network cabling (where a new problem has been introduced) and help you check out cabling spools before you accept them.

Impedance

When conductors made of metal are placed in close proximity to each other, as in a twisted-pair or coaxial cable, the effect they have on each other is known as impedance. When the wires are perfectly separated by a constant distance, the impedance remains the same throughout the cable. When something happens along the way, such as damage caused by a crushed cable, the impedance changes at that point. Changes in impedance cause parts of the signal to be reflected back to where it started.

Cables that are used in local area networks (LANs) need to be manufactured to strict specifications, ensuring that the dielectric material that separates the wires within the cable remains constant. If there are random variations due to poor manufacturing procedures, the cable will suffer from problems caused by signal reflections, which might render it unsuitable for your network. Thus, TDR can be used not only as a fault-finder when troubleshooting a wiring problem, but also to ensure that you've received what you paid for when you upgrade or expand your network.

Setting a Pulse Width

Most of the good TDR instruments allow you to select the pulse width, which is usually specified in nanoseconds. The larger the pulse width, the more energy that is transmitted from the device and thus the farther down the wire the signal will travel.

A good tip for setting this value is to start with the smallest that the instrument allows and make subsequent measurements, gradually increasing the pulse width. If the fault in the cable is only a short distance away from the measuring instrument, a small pulse width will be adequate to locate it. However, if the fault is minor, a small burst of energy might not be enough to travel to the fault and send back a reflection strong enough to be accurately measured. By varying the pulse width and making several measurements, you can more accurately determine the location of a fault in the cable.

Velocity

Light travels at a constant speed of 186,283 miles per second in a vacuum. When measuring the velocity at which an electrical signal travels through a wire, it is expressed as a percentage of the speed of light, which is considered to be slightly less than 100%, or a value of 1. For example, a twisted-pair cable that has a VOP (velocity of propagation) of .65 would conduct an electrical signal at 65% of the speed of light, or about the speed I drive on the interstate (miles per hour, of course).

Manufacturers usually supply this value to customers, and it will most likely be found on the specification sheet for the cable you are purchasing. Because TDR measures the time it takes for a signal to travel down a wire and make the return trip, you have to know the VOP of the cable being tested before you can make accurate measurements.

If you have cables that you are unsure about, you can test them first to determine the VOP. Do this by measuring a specific length of cable to get its length and then using the TDR instrument to test for the length of the cable, varying the VOP until the tester reads the correct length. Of course, this assumes that the segment of cable you use for this test is in good condition!