 |
Global Engineering Documents published the set of new UTP cabling standards for commercial buildings, TIA/EIA-568-B.1 and B.2, in mid-May. These new UTP standards only extend to Category 5e. At the present time, the documents have not recieved wide circulation. Cat 5e was published a year and a half ago, and this market is also unlikely to be saturated yet. The next obvious question, however, is "When will the Cat 6 standard be published?" At the moment, there are a few obstacles. Although the drafts of Cat 6 electrical characteristics (currently Draft 8) have remained substantially unchanged since 2000, precise details of test standards to which manufacturer's products must adhere need to be clearly defined and repeatable. There remains some variation in test methodology among manufacturers. Once the standard is published, any manufacturer anywhere can compete in the same market. That's a good reason the test requirements need to be rigorous, well-defined and repeatable. There will be two more meetings of the TR-42 stadards committee in 2001. It is possible that Cat 6 may be ratified by the end of the year, after which you may expect a publication delay of 2-3 months. So the estimate is early 2002.
In the meantime, be aware that two methods are being used in the marketplace to exaggerate the capability of products being sold.
The first questionable method involves meeting specification for Channel requirements only, while ignoring the requirements for individual component testing. The Anixter Levels program requires that all components (cable, connectors, patch cord) must individually pass the rigorous requirements of Cat 6 or better, as required by the standards, before being tested again in a Channel configuration. By ignoring the component testing, it is possible to pass aggregate Channel tests even though individual components would fail.
The second questionable method involves averaging and/or ex post factocurve fitting. In this method, the limit curve is not drawn until after the test data is collected. The standards define a limit curve for various electrical characteristics. Products must test without exceeding the limits, so there normally exists a margin of headroom between the specified curve and the test results. Consider this a safety zone that allows for some variation and manufacturing tolerances. With good product this safety zone can be considerable, because the product passes with room to spare. In the questionable curve fitting method, the test is performed, the data is graphed and the limit curve is then drawn as close as possible to the data. Sometimes, as in the case of Return Loss, the data graphs were actually once averaged to form smooth curves. Thats how some products exaggerate their claims to exceed the highest standard by outrageous margins. They haven't found a new secret in cable design, they merely fudge the test limits.
The Anixter Levels Lab doesn't average the test results, or use curve-fitting techniques. If a manufacturer's product tests with one blip over the curve, it fails. As a UL Certified testing lab, Anixter must maintain rigorous scientific lab standards. Anixter Levels curve limits are higher than TIA and ISO standards, and were established years before ratification of the equivalent standards.
-Frank Dickman, May 2001
| |