 |
 |
 |
 |
|
|
|
|
The first variation on the original variety of Ethernet was simply to use a thinner coaxial cable and relax the constraints on how and where transceivers can connect. 10BASE-2 does this with coaxial cable that looks just like the cable used for receiving cable television or hooking up a television set to an antenna. The only difference in the cable itself is the impedance rating. A television cable is rated at 75 ohms and a 10BASE-2 cable is rated at 50 ohms. In a pinch, a small length of one can be substituted for the other.
The connectors used in 10BASE-2 are called BNC connectors for Berkeley Nucleonics Co.: they were originally used in nuclear physics electronics (at least that is what I was doing when I first used them). A transceiver is typically located on the same circuit board as the network interface and a male BNC connector protrudes from the device containing the network interface.
 |  |
Just as with thicknet, the ends of the coaxial cable are capped with terminator that connects the core to the shield through a resistor. If either terminator is removed, the entire network ceases to function because of the break in the circuit. Similarly, if any cable is removed from the tee that connects it to a transceiver, the circuit is broken unless terminators are installed.
As a practical matter, since the actual coaxial cable is typically installed where folks can see and touch it, rather than buried in the floor, thinnet tends to fail rather often. Often the failure is a break hidden inside the connector where it joints to the end of the cable. While thinnet is probably the cheapest form of Ethernet to install from scratch, poorly trained technicians are likely to make errors in the making of cables. Purchasing machine-made cables with secure end connectors is worth the extra expense.
A short story from the Ball State Computer Science Department will illustrate some of the flavor of working with 10BASE-2 Ethernet: When I first came to the department, the entire University's Ethernet consisted of a 10BASE-5 cable run by University Computing Services (UCS) which came to the Computer Science Department (CSD) via the orange cable illustrated in the previous section. A tap and transceiver was connected to a bridge with some transceiver cable so that UCS traffic stayed on one side of the bridge and CSD traffic stayed on the other. Only traffic that ran between the two groups went over the bridge.
The CS department originally ran a very long transceiver cable from the tap to its machine room where it had its very own VAX, but by this time a 10BASE-2 cable had been run from the lab area on third floor west, through the closet holding the tap and bridge, up through conduit to the closet on the fourth floor above, and through the ceiling above the hallway up to fourth floor west to a couple of offices on the south side. My office is on the north side of fourth floor west, so an addition piece of cable was run to it from the south side where the 10BASE-2 transceiver pictured above was located.
As soon as the new cable and transceiver were hooked up to my new computer, the entire network failed. Since, my computer was new, it got the blame. The terminator for the network was moved back to the south end of the fourth floor east and I was left without network connectivity. Several weeks later, technicians from UCS arrived to help us out and they discovered that the total cable length was about 300 meters and when my office was added it became 320 meters. Then they left, I was still without network connectivity, and we were all amazed that a 10BASE-2 network functioned with 300 meters of cable when it should only have 200 meters of cable.
The solution to the problem was fairly simple, we needed to put a repeater near the middle of the cable. However, we had no money to buy a repeater. We did, however, have a couple of transceivers and a bit of transceiver cable hanging about. So, we did the very non-obvious thing: we hooked the transceivers to each other with the transceiver cable and hooked the contraption next to the bridge as pictured below. The transceiver pair acted just like a repeater and enabled the cable run to my office and not cause the network to fail.
A few months later however, the network failed again. However, this time the failure was limited to the fourth floor and disconnecting my computer did not help. Eventually it was discovered that one of the computers on fourth floor east was had an unsoldered joint between the coaxial cable's core and the connector stuck on the end. This joint broke, which was not obvious by looking at the cable, and caused the failure the entire fourth floor. It took several days to track down the problem. When this was fixed, the problem reappeared in another joint several weeks later (causing another several days of loss of network connectivity), and then it happened a third time. Eventually, UCS came in and installed 10BASE-T Ethernet to eliminate the problem.
|
|
|
|