Networking: The Physical Connection
A network device is something that connects directly to a network. It's a circuit that is able to manage and process network packets to ensure that they end up at the right place and, just as importantly, that they don't go to the wrong place.
The transmission medium isn't really a network device but it obviously affects how network devices work so the cables, etc. have to be perfect before anything else can be trusted. The first place the network connects to is a Network Interface Card (NIC) and they come with all kinds of features and capabilities. They are mostly interchangeable for the same network type so the brand and feature-set aren't usually an issue. All ethernet NICs, for instance, will do pretty much the same things. The NIC will support at least 10 mega-bits per second but the newer ones can support transmission speeds to a giga-bit per second. Buy a NIC that will do 100 mega-bits per second and it will work in any ethernet network.
Only the most antique Ethernets will use a BNC (round) connector or an AUI (D shaped, 15 pin) connector. Expect to see either a wireless card having an antenna or a cable with a phone-jack type connector (8 wire RJ45) plugging into the NIC. The NIC will probably offer to Auto-sense or Auto-negotiate its speed and duplex settings. You can accept that but if you suspect that the connection is slower than it should be you can force the NIC to a fixed speed and duplex setting. I prefer to force it since it avoids problems in most cases. As long as the NIC speed and duplex setting matches the setting for whatever device the cable is plugged into, there won't be a problem and it should remain stable and work at the best speed.
The cable from the NIC is going to connect to another device somewhere and that device will probably be either a hub or a switch. Hubs are not being installed much anymore since a switch is about the same cost and offers several advantages. In the first place a hub will make every device (NIC most likely) share the speed of the network. If the network is capable of 100Mbs and everyone is connected to it via a hub, that 100Mbs is shared by everyone. If some number of NICs are busy at the same time that will increase the likelihood of packet collisions and collisions slow down the whole network for everyone.
In any recent installation the NICs will connect to a switch. A switch will take the available network speed and allocate each NIC the full speed. It does this by learning which NICs are using which ports (the receptacle that the cable is plugged into) and passing packets destined for that NIC to just that port and no other. A shared hub "broadcasts" all the packets to all its ports and makes the NIC figure it all out. Switches also offer all manner of interesting configuration options.
It's possible (and likely) that the switch or hub is itself connected to another switch or hub before the packets leave the network through a router somewhere. The reason for that arrangement is that each hub or switch extends the length of the network beyond the basic 100 meters so that larger networks over wider areas are possible. There is a limit however and network performance will suffer if the specifications are ignored. These specifications will vary depending on the network speed and the type of switch or hub is used and if they are "stacked".
Wiring Tutorial for 10BaseT Unshielded Twisted Pair (UTP)
Copyright 1996, Robert Kerr NetSpec, Inc. All Rights Reserved. One of the most common and most puzzling problems a network engineer/technician may face is what is the PROPER way to make up a 10BaseT cable. Usually, to confound the learning process, someone introduces the need for a reversed or cross-over cable at the same time. What these are and how to make them is the subject of this on-line tutorial.
Selection of Cabling Category
the overwhelming bulk of network cabling done today uses Unshielded Twisted
Pair (UTP) wiring that is what we will discuss. The process begins with the
selection of the proper wiring level or category. Today it is basically
inexcusable to use or install anything at less than
While technically Category 5 and Level V are not the same, they are identical in practice. Both support up to 100 megabit per second data transmission, and their physical cable assembly requirements are the same. Throughout this tutorial we will refer to them both as CAT5.
When you order CAT5 unshielded twisted pair (UTP) cable you will receive a cable containing 4 twisted pairs of wires, a total of 8 wires. The strands that constitute each wire will either be a single strand or multiple strands, usually referred to as solid or flex . Typically the solid is used to run through walls and ceilings and the flex is used to make drop cables (the cable from the wall plate to the desktop computer) and patch cables (the cable from the patch panel to the hub). Whether the exterior portion of the cable that contains the 4 twisted pairs, the jacket, is Plenum grade or Non-plenum grade is very important, it refers to the Fire Codes, but is outside the scope of this tutorial.
The pairs of wires in UTP cable are colored so that you can identify the same wire at each end. Furthermore, they are usually color coded by pair so that the pairs can also be identified from end to end. Typical CAT5 UTP cables contain 4 pairs made up of a solid color and the same solid color striped onto a white background. The most common color scheme is the one that corresponds to the Electronic Industry Association/Telecommunications Industry Association's Standard 568B.
The following table demonstrates the proper color scheme.
The cable connectors and jacks that are most commonly used with CAT5 UTP cables are RJ45. The RJ simply means Registered Jack and the 45 designation specifies the pin numbering scheme. The connector is attached to the cable and the jack is the device that the connector plugs into, whether it is in the wall, the network interface card in the computer, or the hub.
Now that we are ready to insert the cable into the RJ45 plug the wire number and color sequence becomes more complicated.
The IEEE Specification for Ethernet 10BaseT requires that two twisted pairs be used and that one pair is connected to pins 1 and 2, and that the second pair is connected to pins 3 and 6. Yes that is right - pins 4 and 5 are skipped and are connected to one of the remaining twisted pairs.
According to the EIA/TIA-568B RJ-45 Wiring Scheme:
It gets even more odd because wire Pair#2 (white/orange, orange) and Pair#3 (white/green, green) are the only two pairs used for 10BaseT data.
The remaining two twisted pairs are connected as such:
This is illustrated in the following diagram:
Now the wires forming the pairs must be gathered together and trimmed so that they can be inserted into the RJ45 plug. The pairs gathered and trimmed is illustrated in the following diagram:
Then when the pairs are inserted into the RJ45 plug they should look like this:
In order to make what is commonly referred to as a "Crossover" cable one must change the pin-out connections on ONE end of the cable. If you do it on both ends of the cable you have crossed-over the crossover and now have a straight-through cable, albeit a very non-standard one. In this case two negatives do make a positive.
You need to make a cable where pins 1 & 2 from one end are connected to pins 3 & 6 on the other end, and pins 3 & 6 from the first end are connected to pins 1 & 2 on the other end. Pins 4 & 5 and 7 & 8 are unchanged.
The following is the proper pin out and
cable pair/color order for the
The crossover pairs are illustrated in the following diagram:
Then when the pairs are inserted into the RJ45 plug they should look like this:
Note: Even though we are only interested in attaching the connectors to the cable in this tutorial, we must take into account the wiring of the jacks as well so that we connect the proper wires from the cable to the proper pins in the connectors. And that is determined by the wiring in the jack the connectors will be plugged into.
Go to the Contents Page