Site home page
(news and notices)

Get alerts when Linktionary is updated

Book updates and addendums

Get info about the Encyclopedia of Networking and Telecommunicatons, 3rd edition (2001)

Download the electronic version of the Encyclopedia of Networking, 2nd edition (1996). It's free!

Contribute to this site

Electronic licensing info

 

 

Cable and Wiring

Related Entries    Web Links    New/Updated Information

  
Search Linktionary (powered by FreeFind)

Note: Many topics at this site are reduced versions of the text in "The Encyclopedia of Networking and Telecommunications." Search results will not be as extensive as a search of the book's CD-ROM.

This topic is presented in its entirety with illustatations as a sample topic.

Current networking trends favor an integrated network that can support data, as well as multimedia, voice, and video. Fortunately, new structured wiring and networking standards have been defined to help network designers to plan, install, and test cable systems that support gigabit- and multigigabit-per-second data rates.

In the 1970s and 1980s, coaxial cable was the preferred LAN medium. But by the late 1980s, data-capable twisted-pair wiring emerged as the predominant network cabling scheme. While twisted-pair wire has cable distance limitations, a hierarchical wiring scheme, initially built around hubs (and more recently switches), overcomes those limitations. Workstations are attached to workgroup hubs/switches in nearby wiring closets and those hubs/switches are attached to wiring hubs/switches at centralized data centers via twisted-pair cable, or via fiber-optic cable over long distances.

Today, there are a variety of standards that define cable and component specifications, including the configuration, implementation, performance, conformance, and verification of cabling systems. The most prominent standards are listed here:

  • United States    TIA/EIA-568-A (Telecommunications Industry Association/Electronic Industries Association-568-A), defines how to design, build, and manage a structured wiring system. Note that the specification is also called the EIA/TIA-568 in some references. Refer to "TIA/EIA Structured Cabling Standards" for more information.

  • International    ISO/IEC IS 11801 (International Organization for Standardization/International Engineering Consortium) defines generic cabling for customer premises. It is being used in Europe, Asia, and Africa. See ISO/IEC-11801 Cabling Standards.

  • Europe    Cenelec EN 50173 was derived from ISO 11801 and defines generic cabling and open-market cabling components.

  • Canada    CSA T529-Canadian Standards for Telecommunications Wiring Systems that closely follows the TIA/EIA-568 specifications.

  • Australia and New Zealand    SAA/SNZ HB27:1996. This standard is based on the TIA TSB67 standard. It specifies field testing of balanced copper cabling and the methodology of specifying field tester accuracy

You can learn more about international cabling standards by visiting the Agilent's wirescope.com Web page at http://www.wirescope.com/.

Transmission Media

This section outlines a variety of "guided" media for network communications. "Unguided" communication techniques are related to wireless networking. Figure 1 illustrates the primary types of cable used for data transmissions. These cable types are described here:

Figure 1

  • Straight cable    This is the simplest type of cable. It consists of copper wires surrounded by an insulator. The wire comes in bundles or as flat "ribbon" cables and is used to connect various peripheral devices over short distances. Cables for internal disk drives are typically flat cables with multiple transmission wires running in parallel.

  • Twisted-pair cable    This cable consists of copper-core wires surrounded by an insulator. Two wires are twisted together to form a pair, and the pair forms a balanced circuit (voltages in each pair have the same amplitude but are opposite in phase). The twisting protects against EMI (electromagnetic interference) and RFI (radio frequency interference). A typical cable has multiple twisted pairs, each color-coded to differentiate it from other pairs. UTP (unshielded twisted-pair) has been used in the telephone network and is commonly used for data networking in the United States. STP (shielded twisted-pair) cable has a foil shield around the wire pairs in a cable to provide superior immunity to RFI. Traditional twisted-pair LANs use two pairs, one for transmit and one for receive, but newer Gigabit Ethernet networks use four pairs to transmit and receive simultaneously. UTP and STP are constructed of 100-ohm, 24-AWG solid conductors.

  • Coaxial cable    This cable consists of a solid copper core surrounded by an insulator, a combination shield and ground wire, and an outer protective jacket. In the early days of LANs, coaxial cable was used for its high bit rates. An Ethernet Thinnet (10Base-2) network has a data rate of 10Mbits/sec and implements a bus topology in which each station is attached to a single strand of cable. Today, hierarchical wiring schemes are considered more practical, and even though more twisted pair wire is required to cable such a network, cost has dropped, making such networks very practical.

  • Fiber-optic cable    This cable consists of a center glass core through which light waves propagate. This core is surrounded by a glass cladding that basically reflects the inner light of the core back into the core. A thick plastic outer jacket surrounds this assembly, along with special fibers to add strength. Fiber-optic cable is available with a metal core for strength if the cable will be hung over distances.

Copper cable is a relatively inexpensive, well-understood technology. However, it has various electrical characteristics that impose restrictions on its use. For example, copper resists the flow of electrons, which limits the length of cables. It also radiates energy in the form of signals that can be monitored, and it is susceptible to external radiation that can distort transmissions. In contrast, fiber-optic cable transmits light (photons) through a core of pure silicon dioxide that is so clear, a three-mile-thick window of it would not distort the view. Thus, fiber cable has high transmission rates and is used for long distances. Photonic transmissions produce no emissions outside the cable and are not affected by external radiation. Thus, fiber cable is preferred where security is an issue.

A characteristic of cable that must not be overlooked is its fire rating. Cable installed in the plenum space, which is the airspace between the ceiling and the next floor or roof, must be installed in metal conduit, or must meet local fire codes. In the event of a fire, the cable must not produce noxious or hazardous gases that would be pumped to other parts of a structure through the plenum. Nonplenum cables have PVC (polyvinyl chloride) jackets while plenum-rated cables have jackets made with fluoropolymers such as Du Pont's Teflon.

The remainder of this section covers so-called guided media copper cabling, as opposed to unguided media. For a discussion of unguided media, see "Wireless Communications." Optical cabling is covered under "Fiber-Optic Cable." Also see "Optical Networks" and "WDM (Wavelength Division Multiplexing)."

Copper Cable Characteristics

Information is transmitted over copper cable by applying variable or discrete voltages at one end of the cable and reading those voltages at the other end. Data signals are discrete pulses of electricity (or light in the case of fiber cable). As mentioned, this discussion is oriented toward twisted-pair cable. Some of the characteristics discussed here only apply to wires that are twisted.

The following relationship exists between the frequency of the electrical signal and the rate at which data is transmitted:

  • Bandwidth    The bandwidth of a communication system is the highest frequency range that it uses. This is defined by the engineering specification of the particular network. Some examples are listed in Table 1.

  • Data rate    The actual data throughput of a cable, after applying encoding and compression schemes to more efficiently use the bandwidth of the cable.

Data rate is a more accurate measure of a transmission system's capabilities, but the term "bandwidth" is often used in a general way.

Network Type

Maximum Frequency
Allowed

Actual Data
Rate

10Base-T (Traditional Ethernet over twisted pair)

10 MHz

10 Mbits/sec

100Base-TX (Fast Ethernet)

80 MHz

100 Mbits/sec

ATM-155

100 MHz

155 Mbits/sec

1000Base-T (Gigabit Ethernet, four pairs)

100 MHz

1,000 Mbits/sec

Table 1: Frequency range and data rate of communication systems

The relationship between bandwidth and data rate is illustrated in Figure 2. You can think of the cable as a pipe, where bandwidth is the size of the pipe and the data rate is the amount of information you can push through the pipe. An encoding and compression scheme lets you use the pipe more efficiently. See "Signals."

Figure 2

For example, 100Base-TX uses an encoding scheme called 4B-5B, which first maps data to a table of efficient codes and then transmits the codes over the cable using the NRZ signal encoding scheme. Thus, an 80-MHz signal supports a data rate of 100 Mbits/sec.

In contrast, Gigabit Ethernet (IEEE 802.3ab or 1000Base-T) transmits on four wire pairs in full-duplex mode, meaning that signals are transmitted in both directions on each wire pair. With encoding applied, each wire pair supports a data rate of 250 Mbits/sec in each direction. See "Gigabit Ethernet" for additional information.

Data transmissions over copper cable are subject to attenuation, delay distortion, noise, and environmental problems. Qualified cable installers should test cable both before the cable is installed (to test for cable quality) and after it is installed (to test for proper installation). A table of test parameters is provided later. Also see "Testing, Diagnostics, and Troubleshooting."

Gigabit Ethernet can use existing high-performance Category 5 cable (described later) if the cable passes appropriate testing. It requires much better cable performance than was defined in the original Category 5 specification, so existing cable must be retested to ensure it supports higher frequencies. The minimum recommendations for Gigabit Ethernet cabling are outlined in TIA TSB-95 as described under "TIA/EIA Structured Cabling Standards." Test equipment from Fluke (http:/www.fluke.com), Agilent (http://www.wirescope.com), and other vendors can help you determine the performance of a cable installation.

High-performance cable requires special handling procedures. The physical shape of the cable cannot be drastically altered, meaning that it should not be stretched, twisted, or bent beyond a radius that is 10 times the outside diameter of the cable. Figure 3 illustrates what can happen to wires that are excessively bent. The twisted pairs are pushed closer together, which causes signal interference between wire pairs and signal distortion.

Figure 3

Cable quality may be poor if a cable manufacturer has substituted some material because another material is in short supply, as happened several years ago during the worldwide shortage of FEP (Fluorinated Ethylene-Propylene Teflon). According to Anixter, there now exist more than 45 plenum and nonplenum cable designs that exhibit varying electrical performance characteristics but are still labeled Category 5 compliant. Moral: Have all existing cable installations tested and certified for use with the networking technology you plan to use. You may find that some existing cable runs will support high-speed networks, while others may need connector replacements, and still other require complete replacement.

The following sections discuss various cable characteristics and environmental conditions that affect performance and how those characteristics affect cable and network design.

Attenuation

Signal transmissions over long distances are subject to attenuation, a loss of signal strength or amplitude. Attenuation is also caused by broken or damaged cables. Attenuation is the main reason why networks have various cable-length restrictions. If a signal becomes too weak, the receiving equipment will interpret it incorrectly or not at all. This causes errors, which require retransmission, and loss of performance.

The following shows the weakening of signal due to attenuation:

Attenuation is measured in dB (decibels) of signal loss. For every 3dB of signal loss, a signal loses 50 percent of its remaining strength. Attenuation can be measured by cable testers that inject signals with a known power level at one end of the line and measure the power level at the other end of the line. A typical readout from a Fluke tester is shown in the following illustration. Note that the cable is tested at increasing frequencies and the margin between the tested attenuation and the maximum allowable attenuation (based on Category 5 specifications) is listed.

Attenuation is fairly easy to understand. The illustration is primarily meant to show how cable testers graphically display the test values of cable over the entire frequency range against the TIA-rated values.

Attenuation increases with frequency, so 100Base-TX at 80 MHz has higher attenuation than 10Base-T at 10 MHz. Attenuation also increases with temperature, so cable installers may need to plan shorter cable runs in hot environments. Metal conduit also increases attenuation and should be considered when planning cable length. Cable vendors should provide you with technical specifications for their cables.

Capacitance

Capacitance is the ability of a material to store a charge. Copper cables have capacitance that can distort signals by storing some of the energy of a previous signal bit. Capacitance is a measure of the energy that a cable and its insulator can store. Adjoining wires in wire bundles also contribute to the capacitance of a wire. Cable testers can check capacitance values to determine if a cable has kinks or has been stretched. All cable has known capacitance values that are measured in pF (picofarads). Twisted-pair wire used for network cabling is rated at 17-20 pF.

Test equipment from Fluke, Agilent, and other vendors can test for capacitance in order to identify cable link faults or installation problems. However, a TDR (time-domain reflectometer) is usually a better tool for finding such problems. Fluke testers use capacitance to determine if wiring problems like split pairs exist by testing at only one end of a cable (an adapter is not required at the other end of the cable).

Impedance and Delay Distortion (Jitter)

A signal is prone to delay distortion caused by impedance, which is resistance that changes at different frequencies. It can cause the different-frequency components within a signal to arrive out of step at the receiver. The effect is more problematic on high-data-rate networks that use high frequencies. Impedance may change abruptly due to kinks and excessive bends in the cable, which cause signal reflections that distort data. That leads to retransmissions and a loss in network performance. In the worst case, the network may not operate.

Different types of cable should not be mixed along the same signal path, since a change of impedance at the junction causes a signal reflection back to the source. On high-speed networks, a connector used to join two cables will almost certainly cause an impedance problem because of the untwisting of the wire pairs at the connector. Such connectors should never be used when high-performance networking is implemented over Category 5 cable.

Decreasing the cable length and/or lowering the transmission frequency may solve these problems. Note that the impedance value of a cable can be measured to detect breaks or faulty connections. Data-grade cable should have an impedance value of 100 ohms at the frequency used to transmit data. Category 5 cable meant to support Gigabit Ethernet must be tested for delay distortion.

Delay skew is a problem in networks that transmit on multiple pairs in the same direction at the same time, such as Gigabit Ethernet. It is caused when signals travel at different speeds in each of the wire pairs of a cable. A Gigabit Ethernet transmitter will send signals in parallel across all pairs, which must be reassembled by the receiver. Delay skew desynchronizes these signals, which may prevent reassembly. Category 5 cable meant to support Gigabit Ethernet must be tested for delay skew.

Noise

Transmission lines are susceptible to background noise generated by external sources. This noise combines with and distorts a transmitted signal. While noise may be minor, attenuation can enhance its effects. As shown here, the signal is higher than the noise level at the transmitter but is equal to the noise level at the receiver due to attenuation:

Ambient noise on digital circuits is caused by florescent lights, motors, microwave ovens, and office equipment such as computers, phones, and copiers. Technicians can certify wire by testing for noise levels. If noise is a persistent problem in some areas, it can be avoided by running wire away from sources of noise, by using shielded cable, or by using fiber-optic cable.

As mentioned, twisted-pair cable is supposed to form a balanced circuit where one wire in a pair is equal in amplitude but opposite in phase to the wire it is twisted with. If this characteristic changes due to cable distortion or other factors, the cable becomes unbalanced and starts acting like an antenna, picking up noise from all over, including machines, fluorescent lights, radio stations, and alien transmissions.

Inductance and NEXT (Near-End Crosstalk)

Inductance occurs when current flows on two adjacent metallic conductors. Electromagnetic fields created by the current flows can create signal distortions in adjoining wires. The biggest problem this creates is near-end cross talk (NEXT), which is basically the crossing over of a signal on one wire pair to another wire pair (electromagnetic disturbance). NEXT occurs near the transmitter and creates distortions that typically affect signals on adjacent receive pairs, as shown here:

Note that strong fields from the transmit line may overwhelm the weak (attenuated) signal arriving on the receive line, which can lead to intermittent problems, lockups, or complete failure of the system. NEXT should be measured at both end of the cable.

A typical NEXT measurement is shown in the following illustration. This illustration is derived from a Fluke tester. NEXT is measured in dB, with higher values being better. Note that NEXT is measured for all frequencies between 0 and 100 MHz and that crosstalk varies across the spectrum. The lower line indicates the TIA minimum allowed value across the spectrum.

Twisting wire pairs is the primary method for reducing the effects of inductance, but the type of conductor and insulation also play a role. Twisting wire pairs cancels the positive and negative energy on the cable. Because of this, twists in cable must be preserved all the way up to the connection, especially on high-performance networks. In addition, wires cannot be untwisted more than one-half inch from their connection points.

NEXT is measured by injecting a signal on a wire pair and measuring its crosstalk on another wire pair. Every pair must be tested in this way. Fortunately, cable testers make this job easy and automatic. Keep in mind that NEXT refers to crosstalk at the near end, as the name implies. Crosstalk lessens down the cable as the signal strength weakens due to attenuation. However, this also implies that NEXT should be measured at both ends of a link.

FEXT (Far-End Crosstalk)

FEXT is a relatively new cable measurement requirement. It is a measure of the crosstalk noise that exists at the opposite end of a cable (at the receiver) and is only relevant on network technologies that transmit on multiple pairs in the same direction at the same time-that is, Gigabit Ethernet (1000Base-T). What happens is that crosstalk occurs between transmitters as signals are transmitted down the line.

FEXT can be tested by putting a test signal on one pair and measuring how much of that signal crosses over to the other pairs at the far end of the cable. You will often see FEXT discussed in terms of ELFEXT (equal-level far-end crosstalk). ELFEXT provides a standard way to measure far-end crosstalk no matter what the cable length, so that all cabling can be tested to the same certification levels. The ELFEXT test is required for Gigabit Ethernet cable certification.

ACR (Attenuation to Crosstalk Ratio)

ACR is the ratio at which crosstalk affects an attenuated signal. In other words, how much does the noise on the cable distort the signal I am receiving? If the noise is high, and the signal being received is attenuated, the bit error rate (BER) will be high and retransmissions will be necessary, which leads to a loss in network performance. ACR is important because it provides a useful indication of a cable's performance and is helpful when making purchasing decisions. High ACRs indicate high-capacity cables.

The ratio is calculated by dividing attenuation by NEXT. Most cable testers gather all the information about a cable and perform this calculation automatically. Figure C-4 illustrates the relationship between NEXT and attenuation.

Note that at the point where NEXT and attenuation meet, the crosstalk and data signal are equal, and the crosstalk exceeds the signal strength at higher frequencies. A cable meter like that from Fluke will measure the ACR and compare it against a TIA limit for NEXT. The margins between the worst-case ACR and the TIA limit are outlined in Table 2.

Parameter

Category 5

Proposed
Category 5E

Proposed Category 6, Class E (Performance at 250 MHz Shown in Parentheses)

Proposed Category 7, Class F (Performance at 600 MHz Shown in Parentheses)

Specified frequency range

1-100 MHz

1-100 MHz

1-250 MHz

1-600 MHz

Attenuation

24 dB

24 dB

21.7 dB (36 dB)

20.8 dB (54.1 dB)

NEXT

27.1 dB

30.1 dB

39.9 dB (33.1 dB)

62.1 dB (51 dB)

Power-sum NEXT

Pending

27.1 dB

37.1 dB (30.2 dB)

59.1 dB (48 dB)

ACR

3.1 dB

6.1 dB

18.2 dB (-2.9 dB)

41.3 dB (-3.1 dB)

Power-sum ACR

N/A

3.1 dB

15.4 dB (-5.8 dB)

38.3 dB (-6.1 dB)

ELFEXT

17 dB

17.4 dB

23.2 dB (15.3 dB

Pending

Power-sum ELFEXT

14.4 dB

14.4 dB

3.2 dB (12.3 dB)

Pending

Return loss

8 dB

10 dB

12 dB (8 dB)

14.1 dB (8.7 dB)

Propagation delay

548 ns

548 ns

548 ns (546 ns)

500 for ns (501 ns)

Skew

50 ns

50 ns

50 ns

20 ns

Table 2: Cable parameters

Categories of Twisted-Pair Cable

Twisted-pair cable has been used for decades to transmit both analog and digital information. The existing telephone system is mostly wired with voice-grade twisted-pair wires (the wire is not twisted in some cases). Twisted-pair wire is now the preferred wire for network cabling. The twisting of pairs, the quality of the conductive material, the type of insulator, and the shielding largely determine the rate at which data can be transmitted over twisted-pair cable.

The following categories of cable are recognized throughout the industry, and Category 3, Category 4, and Category 5 cable are specified in the TIA/EIA 568-A specification.

  • Category 1    Traditional unshielded twisted-pair telephone cable that is suited for voice. Most telephone cable installed before 1983 is Category 1 cable. It is not recommended for network use, although modems do a good job of transmitting over it.

  • Category 2    Unshielded twisted-pair cable certified for data transmissions up to 4 Mbits/sec. This cable has four twisted pairs. It was commonly used for IBM mainframe and minicomputer terminal connections and was also recommended for low-speed ARCNET networks. This cable should not be used for high-speed networking.

  • Category 3    This category is rated for signals up to 16 MHz and supports 10-Mbit/sec Ethernet, 4-Mbit/sec token ring, and 100VG-AnyLAN networks. The cable has four pairs and three twists per foot (although the number of twists is not specified). Costs are around 10 cents per foot. Plenum cable costs about 40 cents per foot. This cable is installed at many sites as telephone cabling.

  • Category 4    This category is rated for signals up to 20 MHz and is certified to handle 16-Mbit/sec token ring networks. The cable has four pairs and costs under 20 cents per foot. Plenum cable costs under 50 cents per foot.

  • Category 5    This category has four twisted pairs with eight twists per foot and is rated for signals up to 100 MHz at a maximum distance of 100 meters. Ethernet 100Base-TX, FDDI, and ATM at 155 Mbits/sec use this cabling. The cable has low capacitance and exhibits low crosstalk due to the high number of twists per foot. It costs under 30 cents per foot. Plenum cable costs under 60 cents per foot. This is the predominant cable installed in all new buildings since the early 1990s. Specifications for this cable are outlined in Table 2.

Even though Category 5 is widely used, there are many factors that can prevent a cabling system from delivering the intended data rate. Cable runs should not exceed 100 meters (300 feet). The TIA/EIA specification calls for 90-meter maximum runs from the wiring closet to the wall outlet. An extra 10 meters is allowed to connect computers to the wall outlet and to connect the cable runs to patch panels. Category 5 installations must use Category 5 connectors, patch panels, wall plates, and other components. In addition, proper twisting must be maintained all the way up to connectors.

Enhanced Cabling

Even though Category 5 was considered future-proof, new gigabit-per-second networking schemes have emerged that call for a better class of cable. As mentioned, you can have existing Category 5 cable tested to see if it supports Gigabit Ethernet, but if you are installing new cable for Gigabit Ethernet, choose Category 5E cable, or if you really want to future-proof your installation, consider Category 6 and Category 7 cable. The specifications for these cable types are outlined in Table C-2, earlier in this section.

  • Category 5E (Enhanced)    This cable has all the characteristics of Category 5, but is manufactured with higher quality to minimize crosstalk. The cable has more twists than traditional Category 5. It is rated at frequencies up to 200 MHz, which is double the transmission capability of traditional Category 5. However, at these frequencies, crosstalk can be a problem, and the cable does not have shielding to reduce crosstalk. This cable is defined in TIA/EIA-568A-5 (Addendum 5).

  • TIA Category 6 and ISO Class E    These cable types are designed to support frequencies over 200 MHz using specially designed components that reduce delay distortion and other problems. The TIA and ISO are cooperating on this category.

  • TIA Category 7 and ISO Class F    These cable types are designed to support frequencies up to 600 MHz. Each pair is individually shielded and the entire cable is surrounded by a shielded jacket. Connectors are expected to be specially designed proprietary components. The TIA and ISO are cooperating on this category.

ACR (attenuation to crosstalk ratio) values for the above cable types are listed in the following table. Note that these values for Category 6 and Category 7 are still tentative. The table is meant to provide a comparison at this point in time. Always ensure that new cable is certified for the maximum bandwidth required. Category 5 and 5E cable should be tested at 100 MHz over a 100-meter cable. Category 6 should be tested at 250 MHz over a 100-meter cable.

Category 5 cable

6 to 10 dB ACR at 100 MHz

Category 6 cable

6 to 10 dB ACR at 130 MHz

Category 7 cable

6 to 10 dB ACR at up to 200 MHz

Category 6 and Category 7 specifications have been formulated by the ISO/IEC 11801 groups and the TIA TR42.1.2 committee. One of the best places to find information on these specifications is http://cabletesting.com.

Components of a Structured Cabling System

In the 1980s, vendors and standards organizations saw a need to standardize cabling schemes, and they eventually created the TIA/EIA 568 structured cabling standard. The typical components of a structured wiring scheme are illustrated below. The patch panel provides a place to terminate the horizontal wiring that fans out to work areas. The twisted pairs in the cable are directly attached to the back of the patch panel. The front of the patch panel then provides a place to attach patch cables that connect to network hubs and switches. This arrangement makes moves and changes easily. When someone must be moved to another workgroup or subnetwork, the patch cable on the port leading to his or her computer is moved to another port on a network hub or switch.

As network bandwidth has increased, high-quality cable and components are essential, and they must be installed to exact specifications. A Category 5 cabling system must test within the allowed specifications across the entire cable plant, including all connectors, outlets, patch panels, and cross-connects. This is especially important as networks move to gigabit-per-second speeds. As mentioned, it might be possible to use existing Category 5 cable for gigabit networks, but testing may indicate that some components need replacement.

For more information on cable installation, refer to "TIA/EIA Structured Cabling Standards." Also see "Network Design and Construction" and "Testing, Diagnostics, and Troubleshooting."




Copyright (c) 2001 Tom Sheldon and Big Sur Multimedia.
All rights reserved under Pan American and International copyright conventions.