When it comes to transmitting signals over twisted-pair cable, one of the most important cable characteristics is signal crosstalk. Signal crosstalk occurs when the signals in one wire are electromagnetically coupled, or crossed over, into another wire. This happens because wires in close proximity to one another can pick up each other’s signal. In a twisted-pair Ethernet segment, excessive crosstalk can result in the signals from the transmit wires being coupled into the receive wires. This increases electrical noise levels and signal error rates, and can also cause problems with collision detection on segments operating in half-duplex mode.

The way to avoid excessive crosstalk is to use the correct type of twisted-pair cable, and to ensure that each pair of wires in a twisted-pair segment is twisted together for the entire length of the segment. Twisting the two wires of a wire pair together minimizes the effect of electromagnetic signal coupling between pairs of wire in the cable. This helps make sure that any interference between the wire pairs is below the level of crosstalk that the twisted-pair transceivers are designed to ignore.

One major difference in construction between the various categories of twisted-pair cable has to do with how many twists per foot the wire pairs have been given. The wire pairs in a voice-grade Category 3 cable typically have two twists per foot. This is lightly twisted wire, and you may have to strip back a good bit of outer insulation on a Category 3 cable to reveal the twists.

The wire pairs in higher-grade cables have progressively more twists per foot. Category 5e and 6A wire pairs are tightly twisted, which results in improved crosstalk performance at higher frequencies. For instance, one vendor specifies the wire pairs in a Category 5e cable as having from 19 to 25 twists per foot. Individual pairs within the cable have a different number of twists, which helps reduce the amount of signal that transfers between pairs.

Another characteristic of twisted-pair cables is the type of insulation used on the wires and the cable jacket. Plenum-rated insulation is more stable at high temperatures and provides superior electrical characteristics. Standard PVC insulation will perform as rated for normal room temperatures, but at temperatures above 40°C (104°F) the signal attenuation of PVC-insulated cable increases markedly. Therefore, plenum-rated cables provide better temperature stability and help ensure that the signal quality of your cabling system will remain high. A form of Teflon® called fluorinated ethylene propylene (FEP) is often used for the outer jacket of plenum cables. FEP, the most common form of Teflon, is also used as insulation on the individual wires inside cables to improve signal quality and stability.

Plenum-rated cables are typically required for installation in building air-handling spaces (also called plenums) to meet fire regulations. The reason for this is that different kinds of plastic cable insulation behave differently in a fire. PVC insulation is “fire retardant” in comparison to plain polyethylene plastic, but PVC will still burn and produce smoke and heat. Teflon FEP insulation produces much less smoke and heat when burning, and does not support the spread of flames.

Most structured cabling systems are installed by professional cabling contractors. These contractors have the expertise and equipment required to correctly and safely install the hundreds of twisted-pair cables that a typical office building can require. Cable contractors are also familiar with the structured cable standards and will ensure that the cabling systems that they install meet the specifications.

Should you decide to install a small twisted-pair cable system or a few horizontal segments, the ANSI/TIA/EIA standards provide cable installation guidelines. These guidelines are intended to minimize any effect on the wire twists inside the cable. To support high-speed signals, the wire twists in the cable must remain tightly twisted and not be disturbed anywhere along the length of the cable. Cable ties and fasteners that are too tight, or outer cable jackets that are excessively twisted, can affect the wire twists inside the cable. The installation guidelines include the following:

The words tip and ring are used to identify wires in a wire pair. Most single analog telephone circuits require just two wires to deliver what is known in the telephone industry as plain old telephone service (POTS). These two wires are identified as “tip” and “ring” by the industry. These names date from the earliest days of manual telephone switchboards, when operators made connections between telephone lines using patch cables with plugs on the end.

The plugs had a tip and a ring conductor on them; hence the names for the two wires still used to make a basic analog telephone connection. Each pair of wires in a modern communications cable is still considered to have a designated tip conductor and ring conductor, labeled T1 and R1 for the first pair, T2 and R2 for the second pair, and so on.

To help identify all the wires found in a multipair communications cable, the telephone industry developed a widely used system of color coding. This system uses a pair of colors to identify the individual wires in each wire pair. The primary color group consists of white, red, black, yellow, and violet. The secondary group uses the colors blue, orange, green, brown, and slate. These colors are used to identify the wires in the majority of twisted-pair communications cables, from two-pair cables on up to larger cables.

A primary color is paired with one of the secondary colors for each wire in the cable. For large cables, the primary color is used until it has been combined with each of the five secondary colors. Then the next primary color is paired with each of the five secondary colors, and so on. In a typical four-pair cable, the primary color is white, and no other primary color is needed because there are only four pairs.

Starting with the first wire in the first wire pair of a cable (T1), the insulation is given a base coat of the first primary color, white, with a stripe or dash of the secondary color blue. This is written as “white/blue” and is abbreviated as W-BL. The second wire in the first wire pair (R1) is given a base coat of the secondary color, blue, with a stripe or dash of the primary color white, written as “blue/white” and abbreviated as BL-W (or sometimes just BL). In the first wire pair, then, the T1 wire is white with a blue stripe, and the R1 wire is blue with a white stripe. In the second wire pair, wire T2 is white with an orange stripe, and R2 is orange with a white stripe, and so on.

The term wiring sequence refers to the order in which the wires are terminated on a connector. There are two wiring sequence options provided in the ANSI/TIA-568 standards. The preferred wiring sequence according to the standards is called T568A, and the optional wiring sequence is called T568B.

Figure 16-2 shows the preferred and optional wiring sequences for an eight-position jack connector. Which sequence you use is a local decision. Note that the words “preferred” and “optional” may not reflect reality at your site.

Figure 16-2. The TIA T568A and T568B wiring sequences

The optional wiring sequence is widely used, and many cable installers use it as their default cabling standard. That’s because the optional T568B sequence is also known as the AT&T 258A wiring sequence, and has been widely used for years in AT&T cabling systems. It’s up to you to find out what wiring sequence is widely used at your site, and to make sure that your cabling system adheres to the local standard to avoid confusion.

The center two positions in both wiring sequences, pins 4 and 5, are always used for pair 1—this is where telephone voice circuits are wired if the link is used for analog voice service. That’s why the 10BASE-T standard originally specified the use of positions 1, 2, 3, and 6, avoiding the use of pins 4 and 5: that way you could run a 10BASE-T service and analog voice service over the same four-pair cable if you wished. Although most installations preferred to keep the analog voice and data services on separate cables to avoid the problem of noise from telephone ringing circuits affecting the data service, subsequent Ethernet twisted-pair media standards based on two pairs followed the 10BASE-T wiring scheme for cabling compatibility.

Keeping the wires correctly paired together for the entire length of the horizontal channel is critically important to maintaining signal quality for Ethernet signals. As it happens, there is an older wiring sequence that you may encounter in existing cabling systems that does not provide the correct wire pairing and that can lead to problems for Ethernet signals. The older wiring sequence that results in incorrect wire pairing is called the Universal Service Order Code system (USOC). Despite the name, this is not a universally adopted system, but it was used in older telephone systems. The USOC system deals with the pairs differently, and the wire identification used in the old USOC system is often based on an older color scheme as well.

Because of the way the pairs are wired in the USOC scheme, you will end up with a split pair if you try to install a twisted-pair Ethernet segment on a cable using the USOC wiring sequence.

Figure 16-3 shows an eight-position connector with USOC wiring based on the older color-coding scheme that is frequently used with USOC systems. For comparison, the T568A wiring sequences is also shown. Notice that the wires connected to pins 1 and 2, which are paired together in both the T568A and T568B wiring sequence, are not paired together in the USOC scheme. If you plug a twisted-pair Ethernet station into a cabling system that is wired using the USOC scheme, the Ethernet segment can end up with excessive signal noise and crosstalk because of the split-pair wiring.

Figure 16-3. Split pairs in USOC wiring

It might not be immediately obvious that there’s a problem with the segment, because a simple wiring test of the connection would show that basic wire connectivity from end to end is OK. In other words, the USOC wiring sequence provides wires between every pin of the eight-pin connectors located at each end of the link, so that checking for connectivity between the pins at each end of the link will not detect the problem. What USOC doesn’t provide is the correct pairing of the wires on the wire pairs used to carry Ethernet signals.

It may seem odd that just twisting the wires together in a pair would make this much difference, but it does. Ethernet signals operate at high frequencies, where the lack of twists on a pair of wires makes a big difference in the electrical characteristics of those wires. If the correct wires are not twisted together for the full length of the segment, the segment will experience excessive signal noise and crosstalk, and may fail to operate properly.

You can buy patch cables from cable suppliers at reasonable cost. Because it’s easy to buy ready-made patch cables, many sites choose to avoid the problems with building their own patch cables by purchasing them from suppliers. This also takes advantage of the fact that a good-quality manufactured patch cable will be built using the correct connectors and stranded cable and according to standardized manufacturing and test procedures.

While it’s not impossible to build a good homemade Category 5e cable, it takes a lot of attention to detail, and it may cost more than you might expect to do the job correctly. For example, while all RJ45 connectors may look alike at first glance, there are small differences in the way they are built and the way they fit into a crimping tool. Finding the exact match between the crimping tool and the connector you use can be difficult.

There are a lot of RJ45 crimping tools on the market, too, and a number of them are of low quality, with flimsy plastic or lightweight metal frames that may not provide enough force to produce a really solid crimp. High-quality crimping tools are expensive, and often require special crimping dies designed for vendor-specific versions of RJ45 connectors.

Therefore, it’s a good idea to buy high-quality patch cables ready-made from a reputable manufacturer. This is especially true for any system that will be supporting higher-speed Ethernets, such as Gigabit and 10 Gigabit Ethernet, which send signals over all four wire pairs simultaneously. Also, 10 Gigabit Ethernet requires Category 6A cables, which are more difficult to build correctly. Maintaining the best possible signal quality for 10 Gigabit Ethernet is critically important to the operation of the link, which argues against using anything but the best available manufactured cables.

There are a lot of patch cables on the market, so you need to make sure that the patch cables you buy are rated to meet Category 5e or 6A specifications, to match your cabling system. Very low-cost or generic patch cables may not be carefully built with quality components and may not meet the specifications or maintain their rating over time.

Beware of using standard telephone-grade patch cables for twisted-pair segments. One common patch cable used in the telephone industry goes by the name of silver satin, which describes the outside color of the cable. This is the flat patch cable that is often used to connect an analog telephone to a wall jack, and this cable is widely stocked in ordinary hardware or office supply stores.

The biggest problem with this type of patch cable is that the conductors in silver satin cords are not twisted together, leading to excessive levels of crosstalk on the wires in this cable. This can potentially cause spurious frame errors on your segment. Another problem with silver satin is that the conductors are quite small, which causes higher signal attenuation. Therefore, using silver satin significantly reduces the distance that a signal may travel.

One of the worst problems with silver satin cable is that despite all the signal errors, it may work OK at the lowest supported speed (10BASE-T) when it is used in an Ethernet segment. However, the silver satin patch cable may still be causing data errors and lost frames. These problems can be masked because the Ethernet system will keep trying to function despite the errors, and the problems on a single segment may not cause the rest of the network to fail.

That, coupled with the fact that each station’s high-level protocol software will keep retransmitting frames until something gets through, tends to hide the effects of a poorly functioning media system. However, the higher the traffic rate gets, the more these errors will occur, often leading to complaints of a slow network.

As things progressively get worse, you will be forced to find all of the silver satin patch cables and replace them with the right kind of twisted-pair patch cable. A better approach is to simply forbid the use of any wire or other component in a horizontal cabling system that does not meet the category specifications for your cabling system (e.g., Category 5e or 6A). Also, make sure everyone understands that silver satin patch cables are something that must be avoided in any structured cabling system designed to carry data signals.

A twisted-pair Ethernet transceiver is often attached to a twisted-pair segment with a patch cord connected to an RJ45-style modular jack in a wall outlet. One RJ45 modular jack looks a lot like another, and you can mistakenly connect a transceiver to a telephone outlet instead of the correct data outlet.

The center two pins of the RJ45 jack (pins 4 and 5) may be used by analog telephone services. Therefore, to avoid a conflict with telephone services, the 10BASE-T and 100BASE-T systems do not use pins 4 and 5. These days, however, all Ethernet segments are wired with all four pairs, as required to support 1000BASE-T and faster Ethernet. This makes it much likelier that these Ethernet cables, when mistakenly installed in a telephone jack, could receive analog telephone signals.

The telephone battery voltage is generally 56 VDC, and telephone ringing voltages include an AC signal of up to 175 V peak with large transient voltages at the start and end of each ring interval. Thus, there is a possibility that an Ethernet transceiver in a device could be damaged by these voltages. The standard notes that while Ethernet equipment is not required to survive such wiring hazards without damage, the equipment manufacturers must ensure that there will be no safety hazard to the user from the telephone voltages.

According to the standard, an Ethernet transceiver typically appears as an off-hook telephone to the analog telephone system, meaning that the telephone is in use. Because the telephone system will not send ringing voltages to an off-hook telephone, this should help prevent any damage to an incorrectly connected Ethernet device.

You may encounter older 10BASE-T Ethernet switches that are equipped with 50-pin connectors instead of RJ45-style jacks. This approach was used in older Ethernet switches when a manufacturer wanted to accommodate a large number of connections on a switch panel or modular card for a chassis switch.

In that case, a single 50-pin connector was used to provide 12 four-wire connections, allowing a vendor to support 24 connections on a single interface board with just two 50-pin connectors. The 50-pin connectors, and the 25-pair cables they connect to, were traditionally used in voice-grade cabling systems and were typically rated for Category 3 performance. Therefore, this approach was more popular back when 10BASE-T Ethernet was new. Note, however, that newer versions of these cables and connectors were developed that were rated for Category 5 use.

While using prewired 25-pair cables for connections to patch panels and switches can minimize the amount of wiring you have to do in a wiring closet, there are serious drawbacks. For one thing, they are limited in signal quality and cannot support higher-speed versions of Ethernet. Also, it can be more difficult to troubleshoot a network problem in this kind of installation, because there is no easy way to move a connection from port to port of the Ethernet switch. Because all connections are wired simultaneously with the 25-pair cable, you can’t pull one connection out and try it on another hub port as a test, making it much more difficult to isolate a problem to a particular horizontal cable.

A cable harmonica is a small plastic housing equipped with a strip of RJ45-style jacks in a row, so named because the row of RJ45 holes on the housing makes it look somewhat like the musical instrument. The harmonica terminates one end of a 25-pair cable whose other end is equipped with a 50-pin connector for connection to an Ethernet switch. This system typically supports up to 12 RJ45-style jacks per harmonica.

Building a patch cable involves installing RJ45 plug connectors on each end of a stranded cable. Here we describe the process of installing the RJ45 connectors.

Here are the steps to follow:

In the fairly unusual case of networking only two devices, the two Ethernet stations can be linked together with a single cable. This eliminates the need for an Ethernet switch, but also eliminates the signal crossover that is done inside the switch ports. However, if the Ethernet interface in either or both of the devices implements Auto-MDIX, then they will automatically create a signal crossover when connected together. If the two devices do not have the Auto-MDIX option, then you will need to build a crossover cable to make the signals work properly.

Another use for a crossover cable arises when you need to link switch ports together between two older switches that have hardwired signal crossover inside their ports, and that do not support Auto-MDIX. In this case, there are one too many signal crossovers being done, and this connection will not work with a straight-through cable. Therefore, you need to use a crossover cable to link the two ports.

Figure 16-11 shows the crossover wiring required for the original 10BASE-T and 100BASE-T systems, prior to the widespread adoption of Auto-MDIX. Because both of these media systems use the same four wires, a crossover patch cable or a switch port with internal crossover wired in this fashion works on both systems. Modern Ethernet interfaces almost all support Auto-MDIX, and there is generally no need to build special crossover cables anymore.

Figure 16-11. 10BASE-T and 100BASE-T crossover cable wiring

The Gigabit and 10 Gigabit Ethernet systems use all four pairs of wires, and require that all four wires be crossed over correctly to operate. To make this easier, the Gigabit Ethernet standard uses the MDI-X automatic crossover function, which is supported in most modern Ethernet transceivers.

In the automatic crossover system, the transceiver automatically moves the link signals to the correct logic gates inside the transceiver chip. Once a transceiver has moved the signals to different gates, it waits for approximately 60 milliseconds while checking the link for link pulses or data. This provides a mechanism for each end of the link to automatically configure the crossover function as needed. A random startup time is used to ensure that the ends of the link will not start moving the signals in synchronization, and thereby never achieve a correct crossover.

If neither or both of the ports you are connecting implement an internal crossover, then you can provide an external crossover to make the link work. You can provide the signal crossover for 1000BASE-T or 10GBASE-T links by building a crossover patch cable as shown in Figure 16-12. This crossover cable is universal, and will work for all other Ethernet twisted-pair media systems as well. Given that modern cabling systems provide all four pairs on each horizontal link, a four-pair crossover cable is the only kind of crossover cable you need to keep on hand for those increasingly rare instances when they are needed.

Figure 16-12. Four-pair crossover cable

You should know that disabling Auto-Negotiation can shut off Auto-MDIX under certain circumstances, which can cause failures on a link. This happens when you set a fixed speed by, for example, configuring a port to provide 10BASE-T operation only. The crossover failure occurs because disabling Auto-Negotiation removes support for 1000BASE-T on the switch port, which can also disable Auto-MDIX. The result is a link failure when the speed is manually configured, because the auto crossover mechanism that was making the link work is no longer available to automatically cross over the signals.

It can sometimes be difficult to tell how the Auto-Negotiation implementation on a given switch works from the documentation. It’s therefore worthwhile to spend the time running some tests to see whether or not Auto-Negotiation and Auto-MDIX are still working after a manual speed or duplex setting is made on one end of the link.

There are several ways to tell the difference between a normal straight-through cable and a crossover cable. Ideally, a crossover cable will be labeled as such at one or both ends of the cable, making identification easy. However, if there are no labels, then there are a couple of approaches you can take.

A handheld cable tester can be used to generate a “wiremap” of the cable, which typically provides a display that shows which wires are connected to which pins.

You can also try looking at the wire colors inside the RJ45-style plugs on each end of the cable, assuming that the plugs are made of transparent plastic. If you hold the two plugs together—side by side—you can see that the wire colors on the pins at each end of the cable are the same for a straight-through cable. On a crossover cable, the wire colors connected to pins 1 and 2 at one end of the cable will be connected to pins 3 and 6 at the other end.