LAN testers look beyond Category 5

Patrick McLaughlin

When telecommunications systems bulletin TSB-67 was issued by the Telecommunications Industry Association (tia--Arlington, VA) in September 1995, it set performance criteria for installed links of Category 5 cabling. Specifically, it named four tests--attenuation, cable length, near-end crosstalk, and wire map--that installed cable would have to pass in order to be certified.

The document also marked a turning point in the way in which installers chose the products they would use to conduct those tests. Before September 1995, manufacturers of local area network (LAN) testers claimed that a tester`s most important functions were whichever ones that particular manufacturer excelled in. TSB-67 silenced those claims by specifying tests and their pass/fail requirements. But the document also set the stage for the next generation of superior-performance proclamations and market positioning by tester manufacturers.

The five manufacturers of Category 5 LAN testers--Datacom Technologies Inc., Fluke Corp., Microtest Inc., Scope Communications, and Wavetek Corp.--are now jostling for market position by increasing the efficiency with which they conduct tests required by TSB-67. They are also providing functions that this telecommunications systems bulletin does not require, but that they claim the market demands. So the users of these testers are again sifting through performance claims on tests not mandated by a standard.

The four tests required by TSB-67 are often grouped together and referred to as an autotest. Tester manufacturers compete with each other in terms of the speed with which their devices can conduct an autotest. Scope Communications product manager Eric Olson says the autotest speed is "a very important criterion for people. They look at the autotest as a big differentiator between testers. That`s the time that the instrument is actually doing all the measurements. And the quicker you can do that, the quicker you can get on to your next test."

Sue Gallego, product marketing manager with Microtest, says there is more to testing than simply autotest times, however. "There`s more to it than how fast the test takes to run. Consumers should delve into how fast it takes to set up, run, and store an autotest." She says automatic incrementing is a function that can help speed the overall testing process. "What we`ve seen from job sites is that these installers have a lot of cable identifications or increments. One office will be labeled AB1, the one next to it will be AB2, and so on. With auto-incrementing, if I run a test in one office, the tester will save it as AB1, then the next test will be saved as AB2."

Olson agrees that automatic incrementing saves time. "If you have the auto- increment feature, it makes saving tests a lot easier," he says. "That can cut your overall testing time."

Level II accuracy, and beyond

TSB-67 defines two cabling configurations: the basic link and the channel. The basic link is commonly called horizontal cabling and ends at the wallplate termination. The channel includes the cabling in the basic link as well as patch cords.

TSB-67 also establishes two levels of testing accuracy: Level I and Level II. According to Paul Kish, manager of ibdn systems and standards for cable and component manufacturer nordx/ cdt (St. Laurent, QC, Canada), Level I accuracy--which is less stringent than Level II--is indicative of the level of performance of testers in existence at the time TSB-67 was being drafted. He states that the Level I specifications essentially allowed for "grandfathering" of existing testers and adds that Level II accuracy was established to provide a reasonable indication of the performance of the channel. He states, however, that achieving Level II accuracy at the channel is difficult because of the 8-pin modular connectors included in the test equipment. These connectors are known sources of near-end crosstalk and attenuation.

As Olson explains, "The Level I limit for near-end crosstalk is plus or minus3.4 decibels. To achieve Level II on crosstalk, it`s plus or minus1.6 dB. To meet Level I accuracy for attenuation, a tester must achieve plus or minus1.3 dB. To meet Level II, the tester will need plus or minus1 dB." All five tester manufacturers achieve Level II accuracy for the basic link.

Little support exists for the idea that a user benefits from purchasing a tester that measures the basic link with accuracy more precise than that needed to meet Level II requirements. Chuck Citron, senior product manager for LAN products at Wavetek, says, "There is some value in going beyond Level II accuracy, but exactly how much is arguable. The real question is: Could a particular cable pass with a less accurate instrument, and fail with a more accurate instrument? Theoretically, yes, it`s possible. But it`s very rare and very unlikely. The real world doesn`t run into much of that. And if you`re coming that close to failing, you`ll want to take a close look at the cabling anyway."

"I`m not sure users get caught up in manufacturers` claims of being a fraction of a decibel better than the rest," says Jim Bordyn, Datacom Technologies` director of marketing. "The standards body didn`t want to make Level II accuracy so tough that only one or two companies could meet it. But they wanted to make it tight enough so that installers had reasonable assurance that what they measured was accurate enough so they weren`t unnecessarily reinstalling cabling."

Fluke and Scope claim to meet Level II accuracy for the channel as well as the basic link. Scope`s Olson explains, "There is some controversy around this issue. Theoretically, no tester can meet Level II accuracy for the channel because of the crosstalk associated with 8-pin modular connectors. But overall, it is possible for a tester to achieve Level II for the channel."

Carolyn Carter, senior product specialist with Fluke, maintains that "the channel is really what`s important for network installers and end-users because if the channel has a problem, the network`s not going to run. In order to substantiate our accuracy claims, we contacted UL [Underwriters Laboratories--Melville, NY], the leading independent-testing lab. UL randomly selected products and classified them as having Level II performance on both the basic link and the channel."

According to Gary Goetzinger, operations manager with River City Communications Corp. (Sacramento, CA), there is a need to test the link or channel with Level II accuracy only if the customer demands it. Goetzinger, who has purchased Category 5 LAN testers from three different manufacturers, says, "It`s all customer-driven," when asked if the capability to measure with Level II accuracy is a necessity. "The customer`s specifications will indicate what the test parameters will include," he adds. He says that if end-users are convinced that the link or the channel should be tested with Level II accuracy, they will demand that installers use products capable of doing that, which will create demand for it from manufacturers.

Extended-performance cable

As high-speed networks make use of all four twisted pairs in Category 5 cable, the resultant electrical characteristic of power-sum near-end crosstalk has gotten increasing attention. Although TSB-67 does not address testing power-sum near-end crosstalk, tester manufacturers are building this capability into their units.

"Determining power-sum near-end crosstalk in the field would give the user more information than they`d get by running just the tests called out in TSB-67," says Datacom Technologies` Bordyn. "It`s possible that a link could pass the tests in TSB-67, but not have the performance to run 10Base-T4, or 100VG-AnyLAN, or the anticipated Gigabit Ethernet, all of which are expected to run on Category 5 cable with signals on multiple pairs."

Scope`s Olson agrees that this feature, and others that measure the performance of enhanced Category 5 cable, are valuable. "For a lot of people who make purchasing decisions, the question in their minds is, `Is this tester going to support me for some of the newer test functions that will come out as enhanced-performance cables are approved? Am I going to have to buy another tester three or six months from now?` So some of the newer testing features, like power-sum near-end crosstalk, are good differentiators between testers."

Fluke`s Carter adds that attenuation-to-crosstalk ratio (acr) is another measurement that should be tested if extended-performance cabling is installed. "It`s a very good measurement for extended-performance Category 5 cables because there are no standards," she says. "If you look at the frequency range and acr, you`re getting a way to compare the extended-performance Category 5 cables against one another."

The Institute of Electrical and Electronics Engineers (ieee--Piscataway, NJ) group that is currently developing a Gigabit Ethernet standard is looking at power-sum near-end crosstalk and acr as possible parameters to be tested within gigabit-speed networks. In addition to those two characteristics, the ieee group is also considering far-end crosstalk and return loss as mandatory tests for these networks.

Testing beyond 100 MHz

Carter touches upon a controversial subject that has made the marketing of Category 5 LAN testers look something like rival political campaigns, with conflicting claims backed up by seemingly justifiable facts.

Scope Communications introduced a tester that it claimed measured cable to 155 megahertz. Other manufacturers said that testing to this frequency was unnecessary because no standard calls for frequencies above 100 MHz. Bordyn says, "There are no current network applications or proposed network applications, or applications of any form, that require performance beyond 100 MHz."

Olson explains the rationale behind Scope`s introduction of a 155-MHz tester by saying his firm conducted experiments in which two cables passed Category 5 testing at 100 MHz. Test personnel then ran 155-megabit-per-second Asynchronous Transfer Mode traffic over the two cables. According to Olson, one cable achieved a bit-error rate of approximately one error every 10 minutes. The other cable, however, had an attenuation flaw at a frequency close to 115 MHz that caused a bit-error rate of close to 15 errors per second.

He concludes, "There are cables out there that have defects that will affect your network traffic but don`t appear below 100 MHz. The more bandwidth we`re pushing out of our cables, the higher the possibility that you`re going to run into those defects in the higher frequency ranges."

Bordyn responds with a hypothetical example. "If you install Category 5 cable that passes TSB-67-defined tests up to 100 MHz but fails at 125 MHz, based upon nonstandard test limits, what are you going to do? Are you going to reinstall the cable? Are you going to ask the manufacturer to replace the cable that did not pass at 125 MHz? It passed up to 100 MHz, which is all that Category 5 requires."

Fluke has defined a middle ground on the issue. Carter explains that Fluke lost sales because its products could not test to 155 MHz, so the company added the feature although it believes it is not necessary. "We added the feature, but we`re explaining to our customers exactly why we added it," she says. "We will show you actual measurement data out to 155 MHz, but we don`t make a pass/fail call beyond the standard, which is set at 100 MHz."

She contends that although measuring to 155 MHz is not necessary, it is useful. "Looking at acr or power-sum near-end crosstalk at 155 MHz allows you to see how the cable is performing at higher frequencies," she says. "Even though there`s no standard today, customers want to ensure that the cable they install will work for the next 10 to 15 years. Testing to 155 MHz gives a lot of customers comfort."

Microtest takes a different approach to testing beyond 100 MHz. Gallego explains that the company`s extended-performance tester includes the same hardware as its 100 MHz tester. The difference, she says, is a software module that allows for grading the headroom performance of premium 350-MHz cable at 100 MHz. She explains, "Leading cable manufacturers are specifying the headroom at 100 MHz for their 350-MHz cable. Our performance grading allows users to establish the relative performance of different Category 5 links. Higher-performance links offer longer life, lower error rates, and better signal-to-noise ratio and will support higher data rates than minimally compliant links." Gallego adds, "There is no way to measure 350 MHz today. We report the performance increase and relative quality of the link to indicate the headroom available for future growth."

What is Goetzinger`s opinion of the worth of these new testing features? "Nobody has a crystal ball," he says. "At this point, you can only use a tester that complies with the standard. Does a cable that`s touted as being a Category 7 cable comply with the standard? It complies with the Category 5 standard, but not with the Category 7 standard because there is none."

He says there is no guarantee about what will be included in future standards, either. "What is the next step?" he asks. "You never really can tell." He points to Fiber Distributed Data Interface (FDDI) over copper as an example of a technology that many believed would flourish but has not lived up to its expectations. The lesson behind fddi over copper is that it is risky to make assumptions about the future functions of copper cabling.

Goetzinger considers five characteristics when purchasing a Category 5 LAN tester: ability to perform functions required by customers; ease of use; ease of configuration; durability; and ease of test-result manipulation. He asks, "Do you have to take a course in the manufacturer`s proprietary software?"

Of the three products he has used, he would give one highest marks in ease of use and ease of configuration. Another is his favorite in terms of durability and test-result manipulation. And he determined the third to be the most accurate in an unscientific experiment in which he tested the same cable 100 times with each tester and then checked for anomalies in the test results.

So, while manufacturers of Category 5 LAN testers compete on a high-technology level, some users of these products are more concerned with issues that are much less technical.