The Invention Of Premises Cabling

The runaway success of PC sales in the 1980s led to the development of today’s computer networks, what we now call local area networks (LANs) or enterprise networks. The need for allowing PCs to communicate quickly followed their sales success. Sharing information with others involved saving a file to a disk, first 5¼-inch then 3½-inch floppies, then carrying it to another PC where it could be read. Sharing files by “Sneakernet,” as it was called, was so frustrating that networking development quickly followed PC acceptance.

Several options were first considered. Ethernet, developed by Xerox Palo Alto Research Center, and Token Ring, developed by IBM, were the early contenders. However, both used expensive, specialized cables that were hard to install. Nobody thought wireless was possible in those days.

Ethernet used a large stiff coaxial cable adopted from cable TV that required a special adapter for connecting devices. The adapter clamped onto the cable and connected to the center conductor and the shield. Electronics in the adapter converted the FM signal in the coaxial cable to voltage pulses compatible with the PC’s data requirements.

When the large, unwieldy, expensive coaxial cable proved unpopular, a new version—using smaller RG-59 coaxial cable directly connected to network interface cards (NICs) in personal computers—gained some acceptance.

Token Ring never caught on in the PC market except among big IBM customers. It was pricey, inflexible and expensive to license, unlike Ethernet, which was already in widespread use.

Here is where crosspollination between the computer world and the telecom world took over. The telephone system was transitioning from analog to digital, and companies were using computers to control the digital signal switching necessary for operating the traditional phone system with digital lines. They were also learning how to send digital signals over twisted-pair cables.

The problem with twisted-pair cables is noise. Coaxial cable has an outer conductor that acts as a shield and prevents the emission of electrical noise from the cable and the interference of external sources of noise with signals in the cable itself. To deal with noise problems, twisted-pair was sometimes enclosed in an outer conductive shield, but that made the cable bigger, more expensive and harder to install.

A better method of using twisted-pair cable is called “balanced transmission.” Used for long-distance copper phone lines, the wires in balanced transmission are twisted more tightly than conventional phone wires, and equal but opposite signals are sent on each wire of the pair. Instead of using one wire as a ground and one as a signal—as is used in the shield and center conductor of a coaxial cable—each wire essentially carries half the signal. At the receiver end, the input is also ungrounded and sees the total signal as the sum of the two pairs.

This method has equal but opposite electrical signals on the two twisted wires, so the electrical emission of each wire cancels out the other. External noise affects each wire equally, so at the receiver, the noise cancels out. That’s a win-win situation.

AT&T used what it had learned from digitizing the phone system to create StarLAN, a 1 megabit-per-second (Mbps) LAN on twisted-pair. At about the same time, several employees of Xerox who had worked on the original Ethernet left to form SynOptics, initially intending to create a LAN based on optical fiber. But SynOptics decided instead to pursue a cheaper 10 Mbps LAN over unshielded twisted-pair cable, which they called LattisNet.

Both StarLAN and LattisNet used a star network configuration based on a central electronic “hub” that provided the connection for all devices on the LAN. The IEEE 802.3 committee that created Ethernet standards adopted both standards, but StarLAN was too slow compared to LattisNet, which evolved into the first Ethernet on unshielded twisted-pair (UTP) cable.

Of course, it was not long before problems began surfacing with 10 Mbps over UTP cable. UTP cable needed to be consistently and properly made, as its performance characteristics were critical for high-speed transmission. Signals degraded quickly over the length of the cable, too. Even termination was tricky. The connector being used was an eight-pin modular connector similar to the RJ-45 one used for phone systems but with the arrangement of the pairs in the connector modified to reduce crosstalk. It was discovered that maintaining the twists in the cable all the way down to the connector was very important.

An even bigger problem was with the varying characteristics of the cables manufactured by different companies. Acceptance of this new networking technology depended on getting reliable cable specifications, which is a topic for next month’s column.

About the Author

Jim Hayes

Fiber Optics Columnist and Contributing Editor
Jim Hayes is a VDV writer and trainer and the president of The Fiber Optic Association. Find him at .

Stay Informed Join our Newsletter

Having trouble finding time to sit down with the latest issue of
ELECTRICAL CONTRACTOR? Don't worry, we'll come to you.