Field testing of Category 5, 5E, 6, and 7 installed links reveals dramatic UTP improvements.
Mark Johnston,
Microtest
Many claims and counterclaims have been made about cabling media that are "better" than Category 5. Adding further fuel to the debate is a recent survey indicating that 92% of major cable manufacturers are specifying the performance of their 4-pair copper cabling beyond 100 megahertz. This performance criterion raises a number of important questions. Are all of these performance claims meaningful? What level of performance is appropriate for future applications? What will this additional performance cost? Do we in fact need performance beyond 100 MHz? How do we verify manufacturers` claims?
Today`s Category 5 unshielded twisted-pair (UTP) cabling more or less meets or exceeds Category 5 requirements. But now there is enhanced Category 5, or 5E, as well as draft standards that define Category 6 and Category 7. What do these classifications actually mean, and what is their actual performance in the field?
It is easy to misunderstand cabling performance claims when you are reading advertising or product data sheets. One common specification trap is "17 decibels at 100 MHz." It`s not the performance at any one frequency that matters but rather the performance across a bandwidth (e.g., from 1 to 100 MHz). Another phrase to watch for is "typical acr." The typical link is much shorter than 100 meters, so you`re going to enjoy an attenuation advantage when computing typical attenuation-to-crosstalk ratio (acr). Insist that acr comparisons be based on worst-case examples--100-meter links, for example.
Connecting hardware is often the weakest point in a channel. You may look at cable specifications and forget that they`re not link specifications. Only the latter includes the effects of connecting hardware and installation practices.
Comparing bandwidths
A good way to meaningfully differentiate links is to compare bandwidths. Bandwidth can be defined as the lowest frequency at which the pair-to-pair acr is 3 dB. This frequency is the maximum at which common local area network (lan) applications can be supported. Link return-loss performance is also an important parameter in supporting new high-speed lan applications such as Gigabit Ethernet and should be included in any comparison.
How do premium links perform in the field? Field testing was performed on basic Category 5, Category 5E, Category 6, and Category 7 links from January to May 1998 in both the United States and Europe using Microtest`s recently introduced omniscanner. The device uses a vector digital signal processing frequency domain design, much like a lab network analyzer. It has a noise floor better than 100 dB, which is necessary for making accurate measurements on Category 7 cabling.
Measurements included near-end crosstalk (NEXT), power-sum NEXT (PS-NEXT), return loss, attenuation, acr, and equal-level far-end crosstalk (ELFEXT). Category 5 and 5E links were tested to 100 MHz, Category 6 to 250 MHz, and Category 7 to 300 MHz (the current bandwidth limit of the field tester). Category 6 was tested to 250 MHz rather than to 200 MHz because the Institute of Electrical and Electronics Engineers (New York City) has petitioned the Telecommunications Industry Association (tia--Arlington, VA) and the International Organization for Standardization (iso--Geneva) to specify Category 6 to a frequency 25% higher than the zero acr point.
Performance was compared against current and proposed standards on both legacy and newly installed cabling. Products from 11 suppliers of Category 5, Category 5E, and Category 6 cabling were studied, as were Category 7 products from three suppliers.
Let`s examine field performance by looking at each measurement in turn for all cable types. NEXT, attenuation, bandwidth, and return-loss performance are compared. ELFEXT performance was directly proportional to NEXT performance; good NEXT always meant good ELFEXT.
In all tests, basic Category 5 links fully met Category 5 requirements. There has also been a noticeable performance improvement in these links in the last four years. Category 5E links performed better, with improved acr, return loss, and NEXT margin. Still, Category 6 is simply the best UTP available. When care is taken in its installation and test, Category 6 appears to offer outstanding performance. With competitive pricing, it should become an extremely popular cabling alternative.
Yet, unlike Category 5 products, which are interchangeable between suppliers, Category 6 currently is supplier-specific. For example, supplier A can provide a Category 6 link, and so can supplier B, but if you connect supplier A`s plug into supplier B`s jack, you`ll probably get Category 5 rather than Category 6 performance. Any tests done above 155 MHz require that the test cords used with the field tester properly match the link under test. Check with Category 6 suppliers for up-to-date information. Otherwise, testing a properly constructed Category 6 link with the "wrong" test cords could lead to a false failure. Also, field testing of Category 6 and 7 links is challenging, and not just because of the higher bandwidth. Equally important is the increased dynamic-range requirement of the field tester that must now accurately measure much weaker signals.
Which cable to choose
So what can we conclude? The quality of UTP has improved dramatically, with several grades of performance available. Which should you use? That depends upon your budget and application needs. If you expect to stay with 10Base-T or 100Base-T, basic Category 5 is probably fine, though the improved return loss and NEXT margin of Category 5E would be good insurance. If you`re unsure what applications you will be using or which applications will require Gigabit Ethernet or 1000Base-T in the future, then Category 5E or Category 6 is highly recommended. A number of new Category 6 products will be available during the year, which should further improve its price/performance ratio.
Research has shown that the cost of network cabling represents between 1% and 5% of the total cost of the network, which includes the cost of computers, routers, software, and other network equipment. This cost consideration has influenced many network managers to invest in the best UTP they can afford, since the relative cost premium compared to the total network cost is minimal.
Finally, it`s always prudent to ask your installer for proof of specified performance. This proof should be in the form of certification reports tested to the maximum bandwidth of the link under test. For Category 5 and 5E, maximum bandwidth certification is 100 MHz, and for Category 6, it is 250 MHz. A high-performance link requires high-performance cables, high-performance connecting hardware, and high-performance installation practices. Unless you test the final installation, you could pay for a premium cabling system but get just basic Category 5 performance from it.
Category Classifications with Relative Performance and Cost of Each
- Category 5 -- Unshielded twisted-pair (UTP) cable links that offer no claims of performance beyond Category 5. Basic performance only at lowest possible cost; link often constructed from various suppliers` cables and connecting hardware. Relative- cost factor*: 1
- Category 5E -- Mid-priced UTP links that offer claims of superior attenuation-to-crosstalk ratio (acr) compared to Telecommunications Industry Association or International Organization for Standardization specifications. Includes a system warranty of 10 or more years, and has good link balance, with reasonably good impedance matching of system components. Relative-cost factor*: 1.2
- Category 6 -- Premium UTP, with a positive-link acr at 200 megahertz. Designed with extremely well-matched components for outstanding uniform impedance and very low return loss. Relative-cost factor*: 1.5
- Category 7 -- Four individually screened/shielded pairs surrounded by a shielded jacket. Features positive link acr at 600 MHz but requires proprietary connecting hardware (non-modular 8-pin design). Has the best copper-based electromagnetic compatibility protection but is the most expensive implementation. Strong support in central Europe but low market share elsewhere. Relative-cost factor*: 2.4
*Based on the relative installed/tested cost of 100 50-meter basic links
The testing procedure for Category 5, 5E, 6, and 7 cabling links between two Microtest omniscanners.
The attenuation performance of Category 5 and 5E links was identical because attenuation is mostly a function of the thickness of the copper conductors. In graph (a), there are actually three lines, but the Category 5 and Category 5E lines are so close, they appear as one. In contrast, the Category 6 and Category 7 links in graph (b) have improved attenuation, due principally to thicker copper conductors. In general, attenuation is less of a differen-tiator for high-performance cabling. Therefore, when we look at attenuation-to-crosstalk ratio or bandwidth, we are really looking at crosstalk performance.
All types of links met return-loss requirements, but Category 5E links tended to be much better than low-end Category 5. Category 6 links tended to have return loss that stayed well below 20 dB up to 100 MHz and stayed controlled at higher frequencies compared to Category 5E. Note that for return loss, 2 to 3 dB can make a big difference in performance.
Comparing Category 5, 5E, 6, and 7 next performance
Here are typical examples of basic Category 5 near-end crosstalk (NEXT) performance up to 100 MHz (a). Compared to the basic Category 5 link, Category 5E crosstalk (b) is several decibels lower all across the band, with a comfortable margin at 100 MHz. Compared to the Category 5 link limit, Category 6 links tested up to 250 MHz (c), have much better NEXT performance, with increased bandwidth and better dynamic range.
Because of its internal shielding, Category 7 demonstrates remarkable crosstalk performance (d). Shown are two NEXT curves from two different Category 7 link constructions. The red line indicates a link made with a proprietary connector. The shape of the curve shows most of the crosstalk comes from the cable. The blue line is from a link made with dual 8-pin modular connectors with only the outside 12/78 pairs used. While this design does use the common 8-pin modular plug-and-jack connector, there are only two pairs per connector, so it would not meet Telecommunications Industry Association (tia) requirements. The shape of the curve reveals that most of the crosstalk in this link comes from the near-end connector, not the cable. In both cases, the crosstalk is 20 to 25 dB better than Category 6, but it costs more and its design is not tia-compliant.
Mark Johnston is technology development director at Microtest (Phoenix, AZ) and has been developing communications field-test equipment since 1981.