The following article appeared in Electrical Business Magazine

How to Wire Ethernet

By David Herres

An important focus in upscale residential and most commercial building today is networking of computers and related equipment, notably an Internet connection and one or more printers.

This type of setup is known as a LAN (Local Area Network) and it is generally confined to one building, although a campus matrix is also common.

The vast majority of this type of connectivity is accomplished through an ingenious protocol known as Ethernet. The name (reminiscent of the nineteenth century Ether that was thought to permeate all of space and carry vibrations central to the wave theory of light) evokes the idea of universality. That is very appropriate considering that this important tool has been used worldwide for many years and shows no sign of being replaced by any competing technology.

It is not at all necessary for electricians or cabling technicians who want to wire Ethernet to understand the inner workings of this protocol but a little background is interesting and provides perspective on the undertaking.

At the heart of Ethernet is, of course, an acronym, CSMA/CD, which means Carrier Sense Multiple Access Collision Detect. This is a description of the Ethernet process. Each connected entity is always listening to determine if there is data activity. If there is no traffic, it is okay to transmit. In the event that two or more members of the network begin to transmit at about the same time, data streams will collide and neither transmission will be successful. Both halt transmission, wait a random amount of time, then retransmit. It is this protocol that makes Ethernet communication possible. There are other specified elements as well. One of these is the notion that transmissions take the form of frames, which consist of Preamble, Start-of-frame delimiter, Destination address, Source address and Frame check sequence.

There are two types of Ethernet elements: Data terminal equipment (DTE) and Data Communication Equipment (DCE). You can think of the former as providers and consumers such as computers and printers and the latter as facilitators, such as hubs, switches and routers. Which of these are being connected determines the wiring pin-out, as we shall see.

The other part of the equation is the Ethernet medium, most commonly Unshielded Twisted Pair (UTP) or fiber Optic (FO). The original Ethernet ran on a serial coaxial bus line, but this setup is considered obsolete. Today's Ethernet is implemented on a star topology.

Traditional electrical work is serial in nature, colloquially termed "daisy chain". Branch circuit power comes from an overcurrent device in an entrance or sub panel, goes to the first device such as a receptacle, then to the next, and so on, typically six or seven in number. Spiderwebbing, the practice of bringing power to a single junction box and then splicing out lines to the outlets, while Code compliant, is considered bad form.

In contrast, the preferred configuration for telephone and Ethernet is star topology. Even though it takes more cable, each branch is wired back to the source. That way there are not any extra splices. One of the reasons splicing is undesirable is that it disrupts the twisting of paired conductors. The twisting is crucial since it counteracts inductive loss. As a matter of fact, when terminating UTP it is essential that no more twists be eliminated than necessary. You can successfully wire Ethernet without knowing IEEE 802.3. It is sufficient to understand that this protocol is contained within and implemented by Ethernet cards (NIC) which are required at both ends of any Ethernet link and which interface to the media (cabling). Ethernet cards are backward compatible, but there is a specialized version for fiber optic, which terminates differently.

The precise mechanics of optical fiber is beyond the scope of this article. For now we shall stick to category cable, which, in Cat 5e and later versions, is capable of speeds and reliability appropriate to most networks currently being installed. .

The first thing to consider is what type of cable to use and how to install it. In regard to installation method, raceway is not required for most environments. You can secure UTP cable to any wood finish surface but for drywall it is necessary to use screw clips fastened through the plaster layer into studs. Do not pinch, kink or in any way deform the cable. Where a finish appearance is important, Wiremold works well. If possible, it is best to conceal the cable using fishtapes, chain, magnets and other electriciansÕ tools. In new construction drill through studs and other framing prior to insulation and wallboard. Do not place Ethernet wiring in the same hole or close to power wiring.

In commercial work, though not required, premium results can be obtained by using Electrical Metallic Tubing (ETM). This versatile metal raceway is coupled and terminated with easy setscrew fittings (or for wet areas, compression fittings). A conduit bender creates gentle sweeps and EMT connectors go into box knockouts. Grounded metal raceway protects against RF radiation including harmonics both incoming and outgoing. EMT provides fire protection as well and another advantage is that if the cabling becomes obsolete in the future, it can be used to pull optical fiber into place.

However the installation is accomplished, it is essential not to damage the media. UTP cable is fairly rugged, but any kinking, pinching or abrading can at the very least compromise connectivity where the same fault would not seem to degrade a lower frequency telephone hookup. Also, donÕt let anyone step on the cable as it is being run out of the carton. Do not use previously installed cable Ð these pieces can be used for telephone circuits. Similarly, small wirenuts or crimpable bugs are suitable for telephone work but Ethernet circuits require splice-free runs.

There are several UTP parameters to keep in mind:

• Solid vs. Stranded: UTP is available in both versions. Solid is used for a permanent installation whereas stranded is better for short patchcords that will be bent repeatedly. Solid has better high-speed performance. It is important to use the correct connector type. Both are of the insulation displacement type. The stranded variety pierces the conductors whereas the solid version slides past and presses against them.
• Plenum, Riser, General Purpose: the current carried in an Ethernet circuit is not capable of igniting combustible material. But insulation and jacket, ignited by fire that originates elsewhere, contribute to flame propagation and also significant amounts of thick toxic smoke can be generated. It is important to use the correct cable for various environments. Plenum is used within any cavity that also serves to transport air from one part of the building to another. Riser is used for vertical runs penetrating more than one floor. General purpose is for other locations and limited use cable is for dwellings only. Plenum, the most costly of the three, can be substituted for riser and general purpose. Riser can substituted for general purpose. Any of these can be substituted for limited use cable. For small jobs, you may find it easier to keep a single carton of plenum cable on hand and use it for everything.
• UTP cable is further divided into indoor and wet location cable, which can be used outdoor or underground, usually in RNC (Rigid Nonmetallic Conduit). Underground installations should come out of the ground at buildings through 90-degree sweeps, expansion joints where subject to frost action and LBÕs stubbed into the building through a hole in the siding and rim joist.

Finally we come to the subject of terminations. The cable should be run first with plenty of extra wire at each end. Trim off any damaged ends.

Ethernet protocol calls for either two or four pairs to be used depending on the version, but in both cases the wiring is done the same, i.e. unused pairs are terminated anyway for mechanical strength.

First slide on a rubber boot. This optional item improves the quality of the job by protecting the cable where it enters the connector and also preventing entry of moisture or dust.

Trim the jacket back about two or three inches. It is best to use a data cable stripper but an electricianÕs wire stripper 10 AWG cutter will work if it is not closed completely. But beware of nicking the insulation of the twisted pair, in which event high speed performance will be compromised. With sharp scissors, cut off the rip cord. Untwist the pairs back to the jacket, fan them out in a flat plane and arrange them in the proper order (see illustration). Straighten them out and make them parallel and at the right intervals to go into the connector. Then, again using sharp scissors or the cutter built into some crimpers, make the final cut across all eight conductors. The untwisted conductors should be no longer than 1/2Ó. Otherwise you will have crosstalk. Again restraighten the conductors, push them firmly into the connector making sure that they all bottom out, none of them buckle or fold over and none of them exchange places. Crimp them with a good squeeze. It is necessary to use the proper crimper. If you try to squeeze them with pliers, it won't work.

TIA/EIA Standard mandates RJ 45 pinouts for Ethernet connections. Looking at a connector with the clip to the back and wire opening down, there are eight terminals numbered 1-8 starting at the left. For straight through wiring, used to connect a hub to a computer, T 568-B is used at both ends. For crossover wiring, used to connect hub to hub or computer to computer, T 568-A is used at one end and T 568-B is used at the other end.

T 568-A is:

Terminal 1 - green/white

Terminal 2 - green

Terminal 3 - orange/white

Terminal 4 - blue

Terminal 5 - blue/white

Terminal 6 - orange

Terminal 7 - brown/white

Terminal 8 - brown

T 568-B is:

Terminal 1 - orange/white

Terminal 2 - orange

Terminal 3 - green/white

Terminal 4 - blue

Terminal 5 - blue/white

Terminal 6 - green

Terminal 7 Ð brown

A homemade jig is very helpful in preparing the end. Place an RJ 45 connector in a vise and trim off the back end as shown in the photo. Insert a cable end into this device before terminating. It serves to straighten and align the conductors so that they slide in place easier without jamming or exchanging places. Notice that the cable jacket is crimped inside the connector -- this is the strain relief. Slide the rubber boot in place and you are done.

As for the order of the conductors, it is necessary to realize that at both ends there are transmit and receive pins. You have to connect the transmit pins at one end to the receive pins at the other end. Sometimes you use straight through wiring, the terminations being the same at both ends. In other situations you employ crossover wiring. It depends what is connected at each end. If you are going from an Ethernet hub or switch to a computer, use straight through wiring. If you are going from a computer to another computer or from a hub to another hub, use the crossover configuration. If you make a crossover cable, mark both terminations with an X so that there will be no confusion in the future.
-- END --
More About Ethernet