Recommendations and simplified testing methods challenge decisions to eliminate AXT field testing for Cat 6A UTP.
by Betsy Ziobron
Category 6A testing has come under intense scrutiny over the past few years, mainly due to the need to test for alien-crosstalk (AXT) performance. AXT testing can be a costly, time-consuming process, which has led to the adoption of sample testing methods and has compelled some users to consider shielded solutions.
Now, several cabling and connectivity manufacturers have decided not to require AXT testing as part of their warranty programs, despite recommendations from test-equipment manufacturers and the development of simplified testing methods.
More points and parameters
The internal test parameters for a Category 6A permanent link are identical to those required for Category 6 up to 250 MHz, but are tested to a higher frequency of 500 MHz to support a data rate of 10 Gbits/sec. Electrical characteristicstested include attenuation, return loss, pair-to-pair near-end crosstalk (NEXT), power sum NEXT, attenuation to crosstalk ratio far-end (ACRF), and power-sum ACRF.
"As frequencies approach 500 MHz, there's more crosstalk and attenuation, and the limits are tighter," says Peter Schweiger, marketing development manager for Agilent Technologies'instruments division (www.agilent.com). "As a result, cable and connectivity manufacturershad to improve the quality of their products, and testing manufacturers had to improve their accuracy."
Testing to 500 MHz necessitates more data pointsbeing collected and processed at several more frequencies, but today's testers have evolved to avoid increases in test time. "Testing internal parameters for an individual Category 6A link brought our 9-second test for Category 6 up to about 20 seconds," says Hugo Draye, product manager for certification products, Fluke Networks (www.flukenetworks.com). "But if you look at testers from 3 or 4 years ago, 30 or 40 seconds was more common, so testers are much better today." According to Agilent Technologies, their WireScope Pro has the capability to test a Category 6A link in just 10 seconds.
While improvements in accuracy and test time forinternal parameters is evident, testing for the new AXT parameters is an entirely different story.
"Alien crosstalk is nothing new; we've known about it for years," says Draye. "We just didn't need to be concerned with it because it didn't show up until about350 MHz, and the previous standards weren't concerned with what happened at those higher frequencies."
AXT testing parameters include power-sum alien near-end crosstalk (PSANEXT) and power-sum alien attenuation to crosstalk ratio far-end (PSAACRF), which was once referred to as power sum alien far-end crosstalk (PSAFEXT). PSANEXT is the sum of the crosstalk experienced by the victim cable from signals on pairs in adjacent cables at the same end. PSAACRF is the sum of the crosstalk experienced by the victim cable from signals on pairs in adjacent cables at the opposite end. In laboratory testing, AXT testing is done using a 6-around-1 configuration, in which one victim link is surrounded by six disturber cables and bundled together in a straight run. This method simulates a "worst-case" AXT scenario and is used by cabling vendors to test and certify their solutions.
A whole new setup
Testing an individual link involves placing a main tester at one end of the victim link and a remote device at the victim link's other end. With AXT testing, each adjacent disturber cable within a bundle now needs to be considered.
To measure ANEXT using Fluke Networks' DTX 10-Gig-over-copper test solution, the main unit is connected to the victim link at the near end, the remote unit is connected to the disturber link at the near end, and the two cables are termi-nated with special Link Terminators to avoid reflections.
"We still need to synchronize the main and remote units. Because they can no longer communicate over the cable under test, we now connect the main and remote units with a patch cord," explains Draye. The main DTX unit is also connected via USB to a Windows-based PC or laptop running the DTX AxTalk Analyzer software, which controls the tests and uploads the measurements. Additional Link Terminators can be purchased to save time and avoid having to move them from disturber to disturber on the remote end.
"The remote unit is moved from cable to cable within thebundle in order to gather the data needed to calculate PSANEXT at one end of the victim link," says Draye. "We can't predict that any onecable is going to have a quiet environment, so we need to calculate that combined effect." To measure AACRF, the DTX remote unit is then plugged into adjacent cabling links at the opposite ends of the bundle under test. In this scenario, a spare channel can be used to connect and synchronize the main and remote units.
To test for alien crosstalk using Agilent Technologies' WireScope Pro, the main unit is connected to the victim link at one end, the remote unit is attached to the other end, and both ends of the adjacent disturber cables are connected to AXTalk Stimulators that terminate the disturber link and produce test signals at each frequency. In this procedure, both ANEXT and AACRF measurements are gathered from both ends of the link without moving the main unit or remote unit.
With the WireScope Pro, says Schweiger, "there's no need to hook up a laptop, and time can be saved by using several stimulators attached to all disturber cables in the bundle versus moving them from disturber to disturber. Stimulators canalso be rented to use with an existing WireScope Pro. We're excited about the speed with which AXT field testing can now beaccomplished, and we believe this will help encourage people to do more testing and reduce their uncertainty."
WireScope Pro also includes a spectrum analyzer to enable external noise measurement capability as part of its Alien Crosstalk Toolkit. External noise testing is performed continuously on a selected cable pair or other devices under test. Measurements are compared to the 10GBase-T limits for quick assessment of possible impact on performance.
Tackling test time
Agilent and Fluke agree that workmanship plays a key role in facilitating AXT field testing. "Fortunately, alien crosstalk is only a problem with adjacent cables in a bundle," says Agilent's Schweiger. "If installers can physically make smaller bundles, they can significantly reduce the testing time because one bundle won't interfere with another. If I have a bundle of 24, Ihave to test 23 disturber cables against the victim. But in abundle of 12, I only have to test 11 disturber cables."
Fluke's Draye says, "We've placed two 6-around-1 bundles right next to each other, and there's no measurable interference from one bundle to the other. To make sure you test all disturbers in a bundle, it's important to know which cables are in which bundle. That can be as simple as always bundling the 12 left-side ports together and the 12 right-side ports together. Once the cables are under the raised floor or overhead in cable tray, they should be free flowing and no longer in bundles anyway."
Because AXT field testing can be time consuming, a sampling method is often deployed, in which only a certain percentage of the cables are tested. Sample testing is not addressed in the standards but recommended by most in the industry.
"The longest links have the greatest losses and are more susceptible to being hampered by alien crosstalk," Draye says. "When we cut a 60-meter link down to 50 meters, we saw a 1.5- to 2-dB better margin. With shorter links of 50 meters or less, there is a greater effect from far end connections." Draye adds, "We, therefore, recommend testing the longest links and the shortest links. If the longest links and shortest links pass, and I'm using the same components and same installation methods, there's no need to spend time testing the other links."
Fluke is recommending a 1% sampling method for time savings, but some in the industry are calling for more. "If an end user wants to require 100% AXT field testing, it's in the statement of work, and everyone bidding knows what that entails, then it becomes an even playing field," says Draye. "It really comes down to time and money."
Following an approximately 25,000-link Category 6A greenfield installation that used 2% AXT sample testing with passing results, a Chicago-based engineering design firm considered moving to 1% to save time on future projects. "We wantedto gather more data before moving to the 1% sampling method," recalls James Sauvageau, RCDD, vice president of Environmental Systems Design (www.esdesign.com). "But then we experienced unacceptable alien-crosstalk failures on a few smaller Category 6A projects due to issues with cable and connectivity as well as craftsmanship."
While Sauvageau was pleased that cabling vendors stepped in to remedy the situation, his firm has since decided to stick with requiring a 2% sampling method on Category 6A installations. "We'd like to reduce down to 1% in the future, but we need to see some changes in the market first," he says. "Hopefully, as the market matures, we'll see both manufacturing processes and installation techniques become more refined."
Skipping it all together
While the time involved in AXT sample testing has come down with improved testers and sampling methods, many cabling and connectivity manufacturers have decided to offer system warranties on Category 6A UTP installations without requiring any AXT field testing. Manufacturers offering warranties without AXT testing include ADC, Systimax, Hubbell, Berk-Tek/Ortronics, Siemon, and Panduit.
"We design our solutions with adequate margin to meet all of our performance claims, and we don't mandate field testing of internal electrical performance or alien crosstalk to certify GigaSPEED X10D solution installations," says Simon Cowley, technical director with CommScope Enterprise Solutions (www.commscope.com). "Performing field tests of alien crosstalk is currently cumbersome and time consuming, and we believe it provides minimal incrementalvalue when dealing with high-qualitysolution providers and installers."
John Schmidt, senior product manager for ADC (www.adc.com), adds, "Alien crosstalk testing is not required by the standards. We know our product design exceeds the required performance, and the perfect 6-around-1 scenario we test for in the lab is virtually impossible to achieve in the field. We've also found that installation issues that contribute to alien crosstalk will most likely be exposed and identified by testing for internal performance parameters."
Schmidt says ADC offers a shielded solution for those set on 100% AXT field testing, and several other manufacturers donot require AXT testing with their shielded systems becausethe shield virtually eliminates alien crosstalk issues.
"10GBase-T is highly sensitive to alien crosstalk caused by poor installation or substandard moves, adds, and changes that can exceed pathway-fill requirements, kink patch cords when rerouting ports, or damage adjacent channels whenremoving abandoned cables," says Valerie Maguire, globalsales engineer with Siemon (www.siemon.com). "The performance headroom provided by shielded solutions providesassurance that normal wear and tear over the lifecycle of the cabling systems will not degrade performance below thelevels required to support a particular application."
While it's obvious why AXT testing wouldn't be required for shielded systems, Fluke Networks' Draye points out that the AXT testing is a very good shield effectiveness test over the full frequency range of 1 through 500 MHz: "Once you have the jacks terminated properly to the shield, shielded systems will perform better; however, we've tested shielded systems that have failed alien crosstalk testing because the cable shield was not properly connected to the shield around the jack. We don't recommend skipping the testing even with shielded solutions because there's never really been a good shield test at high frequencies."
According to Maguire, Siemon puts a strong emphasis on best practices in the company's certified installer program to minimize the risk of alien crosstalk issues caused by poorshield termination.
To test, or not?
Some do not agree with the idea of excluding AXT testing, and Belden is one of the few manufacturers currently requiring alien-crosstalk testing for 1% of installed permanent links up to a certain number of links based on test conditions.
"Customers who make the investment in 10-gigabit UTP cabling should expect their systems to be tested success-fully for alien-noise performance," says Warren Davies, manager of channel marketing for Belden (www.belden.com). "Conditions affecting alien noise performance are highly variable from one site to the next, and while lab simulations mayestablish a system's general performance in controlled circum-stances, field testing is the only way to verify alien noise performance in customers' live networks."
Davies believes that, in the absence of compulsory testing for PSANEXT and PSAACRF, one could conclude and potentially argue that a manufacturer's certification of TIA 568B.2-10 AugmentedCategory 6 performance is incomplete and unproven and, therefore, of questionablevalue. "Manufacturers who don't require alien crosstalk testing as part of their certification program may be avoiding exposure of performance issues or attempting to improve their perceived price point," he says.
Draye believes there's good reason to be cautious and do the AXT testing: "We're talking about huge investments, and mission-critical networks and data centers. There needs to be some check for alien crosstalk before expensive equipment isbrought in and fired up. I believe manufacturers are being short-sighted when it comes to not requiring alien-crosstalk testing, and I've witnessed specific examples where installers were burned. We know that if the internals don't pass, there's going to be problems with alien crosstalk, but we can't guarantee that if the internal tests pass, alien-crosstalk tests will also."
Sauvageau couldn't agree more: "I would recommend to any end user looking to install Category 6A that they require 1 or 2% alien-crosstalk testing. I believe it's better to know there's a problem today rather than five years from now when it's time for a major upgrade. We'll continue to educate our clients about the risks associated with not testing."
Agilent Technologies' Schweiger points out that no one is eliminating the need to test for alien crosstalk on Category 5e or Category 6 for shorter runs of 10-gigabit. "Over 80% of the cabling is Category 5e and 6, and people don't want to throw that away; if someone wants to try running 10-Gigabit over that cabling, they're going to have to test for alien crosstalk," he says. "Instead of alien-crosstalk performance testing, there's also the option of using a spectrum analyzer that analyzes actual noise from operational cables and the environment to verify 10-Gigabit performance capability on a Cat 6A link."
No need to tiptoe
Still, many manufacturers are holding firm to the idea of no alien-crosstalk testing. "When the IEEE figured out how much Shannon Capacity they needed to support 10 Gigabit Ethernet, they added a factor of safety," says ADC's Schmidt. "Then TIA added their own headroom on top of that, followed by manufacturersdesigning their systems for the worst-case scenario. So, we literally have headroom upon headroom upon headroom in the Category 6A world. People are walking on eggshells around the alien crosstalkissue, and they really shouldn't be."
BETSY ZIOBRON is a freelance writer and regular contributor to Cabling Installation & Maintenance. She can be reached at: [email protected]