TSB-155 on track for imminent publication

Document provides guidelines for 10GBase-T operation on installed cabling systems.

In March 2003, the 802.3 working group of the IEEE LAN/MAN Standards Committee requested the assistance of TIA TR-42 to help investigate the feasibility of 10-Gigabit Ethernet operation on 4-pair 100-Ω horizontal copper cabling, and to assist the IEEE group in the study of the impact of alien crosstalk impairments. In response, TIA initiated two projects.

Project PN-3-0134 was initiated to investigate the installed balanced cabling performance for 10GBase-T applications, including the mitigation of cabling impairments and field testing. This effort resulted in the development of ANSI/TIA/EIA-TSB-155, Guidelines for the Assessment and Mitigation of Installed Category 6 Cabling To Support 10GBase-T; expected publication of this document is this month.

For harsh alien-crosstalk environments, such as those using Kellums grips that secure cable bundles in the riser space between floors, TSB-155 calls for cabling lengths of up to 37 meters over Category 6. Click here to enlarge image

Project SP-3-4426-AD10 was initiated to develop cabling and component specifications and test procedures to support the operation of 10GBase-T over 100 meters of structured balanced twisted-pair copper cabling. This effort resulted in the development of ANSI/TIA/EIA-568-B.2-10, Transmission Performance Specifications for 4-Pair 100-Ω Augmented Category 6 Cabling; expected publication for this document is sometime this year. This specification includes accuracy requirements for Level IIIe field testers.

The 802.3 working group has completed the development of 10GBase-T. On June 8, 2006, the IEEE-SA standards board approved the standard for 10GBase-T to be published as IEEE Std. 802.3an-2006.

What’s different?

From a historical perspective, TSB-155 is the fourth in a series of TIA/EIA documents addressing field testing of twisted-pair cabling. Although the purpose and scope of each of these documents are different, collectively they represent the incremental advances in field testing of twisted-pair cabling over the last 10 years.

Given that the basis for many aspects of TSB-155 is founded in the prior TIA/EIA field-test documents, here is a brief summary to aid in the understanding of TSB-155 and to assist in pointing out the similarities and differences among this series of specifications:

TSB-67 and TSB-95 have been incorporated into TIA/EIA-568-B.1 Commercial Building Telecommunications Cabling Standard Part 1: General Requirements and TIA/EIA-568-B.2 Commercial Building Telecommunications Cabling Standard Part 2: 100-Ω Balanced Twisted Pair Cabling Components.

• TSB-67 Transmission Performance Specifications for Field Testing of Unshielded Twisted-Pair Cabling Systems, published in 1995, provided the first detailed specifications for the post-installation performance of installed cabling and field-tester requirements, including detailed procedures for verifying field-tester accuracy levels against levels of performance (e.g., Level I and Level II). Additionally, TSB-67 introduced the basic link and channel test configurations and established the test parameters of wiremap, length, attenuation, and near-end crosstalk.

• TSB-95 Additional Transmission Performance Guidelines for 4-Pair 100-Ω Category 5 Cabling, published in 1999, provided guidelines to further characterize Category 5 cabling systems to ensure the operation of 1000Base-T Gigabit Ethernet over the installed base of Category 5 cabling systems. TSB-95 defined the additional cabling parameters of return loss, equal-level far-end crosstalk (ELFEXT), propagation delay, and delay skew, as well as field test procedures and corrective actions. In addition, TSB-95 included accuracy recommendations for enhanced Level II (Level II-E) field testers.

• TIA/EIA-568-B.2-1 Transmission Performance Specifications for 4-Pair 100-Ω Category 6 Cabling, published in 2002, provided the accuracy requirements for Level III field testers.

Above-and-beyond recommendations

Recognizing that 10GBase-T is dependent on cabling performance characterized to higher frequencies than specified for Category 6 and alien crosstalk performance of the cabling, TSB-155 provides the recommendations to deploy 10GBase-T over installed 100-Ω, 4-pair Category 6 cabling, including the following:

Further characterization of existing Category 6 cabling plants to support 10GBase-T applications;

Performance recommendations for alien crosstalk of channels and permanent links;

Field-test configurations for alien crosstalk measurements of channels and permanent links;

Field measurement procedures for alien crosstalk;

Alien crosstalk marginal computation;

Internal parameter mitigation;

Alien crosstalk mitigation procedures.

Based on the initial discussions in the 10GBase-T task group on the tradeoffs between the signaling, modulation rate, and Category 6 channel bandwidth, the 10GBase-T task group determined that, to support 10GBase-T, the Category 6 channels needed to be characterized up to 500 MHz for both the internal cabling parameters (e.g., pair-to-pair crosstalk) and alien crosstalk.

Transmission impairments that are internal to a twisted-pair cable include crosstalk (pair-to-pair near-end crosstalk, power-sum near-end crosstalk, equal-level far-end crosstalk, and power-sum equal-level far-end crosstalk) as well as echo and insertion loss. Click here to enlarge imageTSB-155 includes recommendations for channels and permanent links for the Category 6 internal cabling parameters extended to 500 MHz. Parameters are: insertion loss (IL), pair-to-pair near-end crosstalk (NEXT), power-sum NEXT (PSNEXT), pair-to-pair equal-level far-end crosstalk (ELFEXT), power-sum ELFEXT (PSELFEXT), and return loss.

A single cable in a bundle is subject to all internal transmission impairments as well as alien near-end and far-end crosstalk. Click here to enlarge imageA 10GBase-T topology consists of four pairs simultaneously transmitting bi-directionally-full-duplex transmission. The transmit signal is attenuated by the cabling insertion loss. The noise from three NEXT disturbers, three FEXT disturbers, and the echo combine with the receive signal. The NEXT and FEXT combine at the receiver on a power basis, referred to as power summation (e.g., PSNEXT and PSFEXT). FEXT is highly dependent on length. For cabling specifications, the insertion loss is subtracted from the FEXT to eliminate the length dependency; the resultant parameter is defined as ELFEXT, where ELFEXT (dB) = FEXT (dB) - IL (dB).Alien crosstalk performanceBecause 10GBase-T cancels a significant portion of the noise due to the internal cabling impairments, the usable bandwidth of twisted-pair cabling is, therefore, primarily limited by the cabling insertion loss and the alien crosstalk coupled from cabling in close proximity. TSB-155 includes alien crosstalk recommendations for Category 6 channel and permanent links. The use of “alien” does not imply extra-terrestrial crosstalk, but distinguishes the crosstalk between pairs within a cable from the crosstalk between cables. Alien crosstalk is the unwanted signal coupling between cables.

Click here to enlarge imageAlien crosstalk is specified for both near-end and far-end, as alien near-end crosstalk (ANEXT) and alien far-end crosstalk (AFEXT), respectively. It is also specified on a power-summation basis as PSANEXT and PSAFEXT. In addition, due to the statistical nature of the 10GBase-T receiver, the alien crosstalk is averaged and specified as average PSANEXT and average PSAFEXT.For harmonization with related national and international cabling standards, TR-42 has recently redefined PSAELFEXT as “power-sum attenuation to alien crosstalk ratio at the far end” (PSAACRF). Please note the definition of PSAACRF used in TSB-155 corresponds to PSAELFEXT in IEEE 802.3an 10GBase-T.

TSB-155 includes requirements to extend the testing process of the channel and permanent link to include testing for alien crosstalk. Click here to enlarge imageAlien crosstalk and insertion loss are the primary parameters to consider for 10GBase-T operation. TSB-155 provides equations to determine the PSANEXT loss and the PSAACRF limits for channels or permanent links based on the insertion loss. In addition, TSB-155 provides limits for the average PSANEXT loss and average PSAACRF “across the four pairs” adjusted for the insertion loss.The derived PSANEXT loss and PSAACRF, and the insertion loss, are applied as limits in qualifying installed cabling for 10GBase-T operation. It is important to consider these parameters together because they can have a significant effect on one another. As an example, 55 meters of cabling can tolerate 15 dB more PSANEXT noise than can 100 meters of cabling, because the shorter circuit has approximately 50% less insertion loss.The channel and permanent link test configurations are defined to enable the creation of performance specifications for the cabling, as well as a reference point for attaching field test equipment. Field measurements are compared to worst-case limits derived from the cabling component performance and the cabling topology-i.e., the lengths of cables and cords and the numbers of connectors.TSB-155 extends the test configuration topology definitions of channel and permanent link to include the alien crosstalk test configuration. The figure (page 51, bottom) depicts the ANEXT and AFEXT coupling between a single disturbed channel and a single disturbing channel. The dashed lines illustrate the reference point for the attachment of test equipment and the demarcation points of the channel definition.ANEXT test configurationsTSB-155 includes recommendations for the field measurements of ANEXT loss and AFEXT loss, and the calculations of PSANEXT and PSAFEXT. The field measurements are to be implemented in accordance with Annex 1 of TIA/EIA-568-B.2 using test equipment that meets the field tester accuracy requirements for Level IIIe that are proposed in Annex E of TIA/EIA-568-B.2-10. Should the cabling fail either the internal parameters or the alien crosstalk parameters, TSB-155 provides annexes containing detailed mitigation procedures.TSB-155 provides a specification for the measurement floor of the test instrument as well as a correction to the measurement floor. The measurement correction is applied because, in the power summation, the noise floor is added (as well as the crosstalk) for each measurement. The power summation of the noise floor is removed from the ANEXT and AFEXT power-summation results by the measurement correction. The correction is derived from the number of disturbing pairs measured.An alien crosstalk significance condition is specified for determination of the ANEXT loss or AFEXT loss measurements that are to be included in the power-summation calculation. The significance condition is applied to determine which of the measured channels or permanent links to include in the power-sum calculation (i.e., what channels or permanent links have significant alien crosstalk). The value of the significance condition is 91 dB.TSB-155 provides examples of how to process the measurement data, including the measurement floor correction and adjustments for asymmetrical channels or permanent links. Additionally, general test strategies are provided in TSB-155 as follows:• Only test links that are intended to support 10GBase-T operation;• Initiate ANEXT and AFEXT testing using disturbing channels that are terminated adjacent to the disturbed channels on patch panels or other connecting hardware;• The number of disturbing ports to be included in the power-sum calculation depends on the configuration. For any given configuration, the determination of which ports to include can be made based on the pair-to-pair alien-crosstalk contribution to the disturbed channel. The power-sum alien crosstalk should be calculated and monitored as each pair-to-pair measurement is power summed until all likely disturbing cables (from knowledge of the cabling topology) have been measured.A list of corrective action options are provided in the event that an existing Category 6 installation fails any of the Category 6 internal parameters when measured from 250 to 500 MHz. The corrective actions outlined were derived to accommodate a 4-connector topology. One or more of the following options can be applied, depending on the cabling topology:Option 1: Replace the work area, patch, and/or equipment cords with Category 6A cords;Option 2: Reconfigure the cross-connect as an interconnect;Option 3: Replace the consolidation point connector with a Category 6A consolidation point connector;Option 4: Replace the work area outlet connector with a Category 6A work area outlet connector;Option 5: Replace the interconnect or cross-connect with a Category 6A interconnect or cross-connect.

This test result is an example of a conductor pair in a permanent link failing to meet one requirement (PSAACRF), but more than making up the difference in margin over another requirement (PSANEXT) and, therefore, still supporting 10GBase-T because the combined noise is within the overall limit. Click here to enlarge imageMitigation procedures have been used successfully in the past in support of 1000Base-T operation over Category 5 cabling as detailed in TSB-95.Computation and mitigationChannel or permanent links can still have sufficient signal-to-noise margin to support 10GBase-T operation, even though the individual pair limit lines are not met. TSB-155 includes the alien crosstalk margin computation specified in 10GBase-T to assess whether the cabling can support 10GBase-T operation should the individual pair limits not be met.

Graphically depicted here is a computation of alien crosstalk margin across the wide 10- to 400-MHz frequency range; this calculation is another example of how and why a link can support 10GBase-T even though an individual limit line was violated. Click here to enlarge imageThe figure PSANEXT, PSAACRF individual pair measurements illustrates PSAACRF and PSANEXT measured on Pair 2 of a permanent link. The PSAACRF fails the individual limit while the PSANEXT exhibits approximately 10 dB of margin from the limit. The overall SNR margin for this link is sufficient to support 10GBase-T operation because the 10GBase-T receiver function does not discriminate between individual PSAFEXT and PSANEXT noise contributions. The receiver detects symbols for the “combined noise” and the transmitted signals on each of the four pairs.The figure PSANEXT and PSAFEXT combined shows the combined PSAFEXT and PSANEXT measured on Pair 2 of the failed link. The combined PSANEXT and PSAFEXT noise is calculated and compared to the alien crosstalk noise limit determined from the disturbed pair’s insertion loss.The alien crosstalk margin is the average “across frequency” from 10 to 400 MHz of the difference between the combined PSANEXT and PSAFEXT and the limit. The margin must be greater than 0 to use the cabling for 10GBase-T operation. In this example, the margin is 3.09 dB, so the cabling would support 10GBase-T even though the PSAACRF had failed the individual limit line. The alien crosstalk margin is specified for each of the four pairs as well as the average across all four pairs.TSB-155 provides guidelines to mitigate the alien crosstalk between the disturbed pair and disturbing pairs should the alien crosstalk margin computation not be met. The choice of mitigation options is based on the particular circumstances, and options include:

Selective deployment of 10GBase-T in non-adjacent patch panel positions;

Recommendations for testing adjacent patch panel positions;

Reducing the alien crosstalk coupling by separating the equipment cords and patch cords, and unbundling the horizontal cable; or, by using equipment cords sufficiently specified to mitigate the alien crosstalk coupling, such as Category 6 ScTP or Category 6A;

Reconfiguring the cross-connect as an interconnect;

Replacing connectors with Category 6A connectors;

Replacing the horizontal cable with Category 6A cable.

TSB-155 provides additional guidelines for the 10GBase-T-supported distances that recognize harsh alien-crosstalk environments, such as the use of Kellems grips in riser applications. The Kellems grip provides a method of securing cable bundles in the riser space between floors. The grip tightly binds the cables, and can increase the alien crosstalk coupling compared to other cable-binding methods, such as ties.For cabling configurations in harsh alien-crosstalk environments, such as those using Kellems grips, TSB-155 states 10GBase-T cabling lengths of up to 37 meters over Category 6 cabling.Radical advancementsTSB-155 provides the additional guidelines to support the IEEE 802.3an 10GBase-T standard for operation over installed 100-Ω 4-pair Category 6 cabling extended to 500 MHz. TSB-155 embodies radical advancements in field-testing balanced twisted-pair cabling. The pass or fail criteria once restricted to evaluating the cabling performance relative to a limit line for each parameter now includes further data processing to compute the alien crosstalk margin, more accurately reflecting the real performance of a receiver.Field test equipment from multiple manufacturers can perform the extended-frequency measurements for the TSB-155 cabling parameters, including alien crosstalk, and the data-processing functionality to compute the alien crosstalk margin. The primary challenges remaining for field test manufacturers and installers are reducing the time it takes to perform the alien crosstalk measurements and simplifying the mechanics of performing those measurements.

(Editor’s note: The author is currently drafting a follow-up article focusing on hands-on techniques that can be used in the field to implement TSB-155. We anticipate publishing that article in March.)CHRIS DiMINICO is a consultant at Solarflare Communications (www.solarflare.com) and president of MC Communications, a telecommunications consulting firm. He is project leader for the TR-42 PN-3-0134 group developing TSB-155.