| Previous | Table of Contents | Next |
SIMMs and DIMMs
There are two kinds of common memory nowadays:
A DIMM has contacts on both sides and is therefore more dense pin-wise (168 pins) and data-wise. Although a SIMM has contacts on each side of it, as well, each side leads to the same place on the board. This means that there are actually 60 contacts on a 30-pin SIMM, and 144 contacts on a 72-pin SIMM. SIMMs are also typically slower than DIMMs.
To remove a SIMM from a motherboard, pull outward on the metal clips holding it. This will make it pop from its 90-degree angle from the motherboard to a 45-degree angle. You’ll then be able to remove the SIMM. To insert a different SIMM, reverse the process: Put it in at a 45-degree angle, get it nestled in, and push it to a 90-degree angle. You should feel a click when the metal tabs engage the SIMM’s holes. (If it’s not easy to do, turn it around; you might have it backward.)
A DIMM is even easier: Simultaneously push down (toward the motherboard) the plastic levers on each side of the DIMM. The DIMM will pop up, and you can remove it. To insert a DIMM, simply line it up correctly—it has a different number of pins on each side of its bottom notch—and push it in hard. The plastic levers will engage by themselves.
Depending on which type of memory your machine has, you may have to replace it in pairs or fours—whatever your motherboard documentation refers to as a “bank of memory.” The best practice is to swap the entire bank of memory with another known good bank—that way, you don’t have to worry about whether you need to keep pairs of memory together, even though some computers consider a bank to be one SIMM or DIMM.
Also, you’ll frequently find that network application problems are related to component problems; taking half an hour to swap components can save you blank looks from a repair technician. In other words, an intermittent problem on your network might not occur when an outside technician runs diagnostics on what you have determined to be a problem PC. If you can localize the problem, you’ll probably save time and aggravation in the long run.
There are two major types of network cabling widely in use:
Of the two, UTP is the most common. STP, while less sensitive to electromagnetic interference, is also more expensive and harder to work with than UTP. STP is found only in Token-Ring networks. STP is harder to work with because it’s very thick and not very flexible. Also, the process for crimping the cable involves stripping the outside, stripping four inside wires, crimping each wire separately, and then assembling a DB-9 housing (much like the housing on a serial mouse). Contrast this to UTP cabling, where you strip the outside wire, place all the wires into a telephone-like jack (RJ-45), and then do one crimp.
Why is STP only used for Token-Ring? Well, actually, shielded cable was used at one time for Ethernet as well, but it, too, was horrible to work with, and nobody uses it anymore. (See Hour 9, “Ethernet Basics,” for more information.)
UTP, which is categorized as CAT-III and CAT-V, is used for 10Mb (megabit) and 100Mb Ethernet, respectively. The categories of wiring are an industry-standard way of referring to the manufacturing specifications for the cable: the thickness (gauge) of the wire, the number of twists per foot, and the electronic resistance of the wire. CAT-III and CAT-V can also be used for Token-Ring. (See Hour 10, “Token-Ring Basics,” for more about Token-Ring cables.)
Sometimes, you’ll troubleshoot a problem down to a cable. For example, if a user cannot get onto the network, and you run the network card’s vendor-supplied diagnostic program and find that the network card is okay, the next thing to suspect is the cable. It’s pretty easy to swap a cable, regardless of whether you’re working with shielded or unshielded cable. Therefore, if you swap the cable with a known good cable and this brings the workstation back on the network, you have a broken cable on your hands. In this case, you can simply toss the cable in the trash and buy another.
Unless you make really, really good cables, you’re better off buying your cables ready made. Good cables make a good network.
When troubleshooting UTP wire runs, you’ll want to verify that the UTP is nowhere near motors or fluorescent lights, because EMI (electromagnetic interference) can create very odd problems on your network. Because UTP is so easy to work with, lots of people (trying to save a buck) run their own cable. (It’s actually easy to slice the housing off of a cable, insert it, and crimp down on it.) However, many times, folks run cable without realizing that EMI even exists. Save yourself trouble by verifying that your UTP runs are nowhere near it.
| Previous | Table of Contents | Next |