Category 6A vs category 6_A: what’s the difference and why should I care?

New standards for copper cabling smooth the way for high-performance data transmission in buildings, industrial plants and data centres. However, in order to have a long-term guarantee for the desired performance, the pitfalls in nomenclature need to be considered during project design and component selection.

With the advent of 10 Gigabit Ethernet over copper twisted pair cabling, new classes of cabling standards were introduced. EIA/TIA released the category 6A standard in February 2008 and ISO/IEC the Class E_A channel standard in the same time frame. Unfortunately, these two standards do not define the same performance, leading to confusion in the market. This confusion is compounded when looking at the components, especially the connecting hardware. EIA/TIA and ISO/IEC specify different performance for the modules, but the component naming is very similar. Therefore, special care must be taken or users will not obtain the performance they expect.

The IEEE protocol for 10 Gigabit Ethernet over copper twisted pair cabling (IEEE 802.3an) was released in July 2006. Because it increased the used frequency range to 500 MHz and category 6 cabling was only defined to 250 MHz, it was quite clear that new cabling standards would be required to support this protocol. Of course, category 7 cabling, which is specified to 600 MHz, was an option right from the beginning, but with a worldwide market share of only 4%, it would not ensure the success of this new ultra-high-speed ethernet protocol.

In the 802.3an standard, IEEE specified the minimum channel requirements that the cabling would need to achieve in order for the protocol to work. In fact, a good category 6 shielded cabling system, which operates stably at higher frequencies, could meet these minimum requirements.

However, alien crosstalk posed a significant problem for unshielded cabling systems. Due to the higher transmission frequencies used, as well as the complex coding methods, the low signal strength associated with 10 Gigabit Ethernet is much more susceptible to outside disturbances than was the case with previous, lower speed ethernet protocols. This translated into a length limitation with standard category 6 unshielded systems.

The various cabling standards bodies thus began work on specifying new cabling classes rated up to 500 MHz that are based on conventional RJ45 technology. The EIA/TIA category 6A standard and the ISO/IEC channel requirements for Class E_A were published around the same time frame in 2008. Unfortunately, however, these standards do not specify the same performance characteristics. Figure 1 shows the differences with the Channel NEXT parameters outlined in these IEEE, ISO/IEC and EIA/TIA standards.

Figure 1: The differences with the Channel NEXT parameter.

The EIA/TIA category 6A channel requirements show a relaxed slope of 27 dB starting at 330 MHz, whereas the ISO/IEC Class E_A channel specifies a straight line, thus offering the highest performance available using RJ45 technology. At 500 MHz, this means that Class E_A requires 1.8 dB better NEXT performance than a category 6A channel. In practice, this higher performance translates to higher operational reliability, which, as a consequence, minimises errors. In addition to this, the life of the cabling infrastructure is also maximised.

Component standards to follow

With the channel standards now clear and well defined, the next step is to define the standards for the components that make up the channel. EIA/TIA defined channel, permanent link and component specifications together and all are included in the ratified category 6A standard (568B.2-10). ISO/IEC defined the channel specifications in Amendment 1 and is working on the permanent link and component definitions which will be released in Amendment 2.

The need for standards to define component characteristics first arose with the end user’s requirement for interoperability - in other words, the ability to mix and match components from various vendors and still be assured of achieving the corresponding channel performance of that category. For example, a category 6 module from vendor X, a category 6 installation cable from vendor Y and a category 6 patch cord from vendor Z should all be able to be combined to achieve a Class E channel performance.

In order to ensure this end result, the concept of ‘de-embedded testing’ was introduced in 2003. In this testing procedure, a defined ‘known’ reference jack is used to test plugs in a mated plug and jack connection. The effects of the reference jack are subtracted or ‘de-embedded’ from the mated connection values, giving the NEXT characterisation of the plug.

This method is used to qualify 12 reference plugs in the low, mid and high ranges, which are then used to test the connecting hardware.

In the case of 10 Gigabit Ethernet, initially systems were offered which would meet the channel requirements of the protocol. The new component requirements will enable system interoperability, that is to mix and match systems, as in the past. For category 6A (EIA/TIA) and category 6_A (ISO) components, ‘re-embedded testing’ has been introduced.

The overall idea of ‘re-embedded testing’ is similar to ‘de-embedded testing’, but in this case a reference plug is first qualified by a new and more accurate measurement set-up called ‘direct probing’. The difference between this reference plug and the 12 de-embedded reference plugs is then calculated. At this point, the connecting hardware is tested with the one reference plug. Afterwards, the results that would have been obtained with the 12 de-embedded reference plugs are calculated, rather than individually tested. In essence, the process of finding and testing the 12 de-embedded reference plugs is replaced by one accurate measurement and the subsequent difference calculations, thus ensuring faster, but also more consistent, testing results.

Similar to the situation with channel performance specifications, a category 6_A connector that is specified by ISO/IEC will be required to achieve a higher level of performance than a category 6A connector that is specified by EIA/TIA. The current draft specifies a 40 dB slope starting at 250 MHz for category 6A and a 30 dB slope for category 6_A. At 500 MHz, this means that a category 6_A module must achieve at least 3 dB better NEXT performance than a category 6A module.

Confusing names

With component standards for connectivity and cabling, it now starts to get confusing. The component specifications needed to achieve a category 6A (EIA/TIA) channel are clearly different and less stringent than those required to achieve a Class E_A (ISO/IEC) channel. Therefore, users who want to ensure Class E_A channel performance must use components that meet category 6_A specifications. A channel composed of components that are compliant to the category 6A (EIA/TIA) specifications won’t be able to guarantee compliance to the Class 6_A (ISO/IEC) performance specification.

Therefore, the difference in the ‘A’ - specifically whether or not it is written as a subscript - is very important:

category 6A ≠ category 6_A

Table 1 shows the two new cabling classes and the associated component naming and terminologies. Further complicating the situation is that the category 6_A specifications have not yet been ratified [as at the time of writing]. Amendment 2 of ISO/IEC 11801 is in progress and the time frame for release is currently unclear.

Table 1: New standards for Channel and Components.

Why does ISO/IEC take longer?

You may be wondering why ISO/IEC is taking so much longer to specify the components criteria than those of the EIA/TIA specification. One reason is because of the different structures in the respective standards organisations. ISO/IEC includes a number of different organisations that are responsible for the cabling, cable and connecting hardware specifications. Coordination between the various groups naturally takes longer than in the case of EIA/TIA, where all of the interested parties are in the same group.

However, another more significant reason is the technical complexity of developing transmission performance characteristics that really amounts to forging a path into new territories. Until now, we have understood the behaviours of the components and how they work together very well up to 250 MHz. We are now effectively doubling the frequency range; and the modelling methodology to be used for those higher frequencies is not sufficiently stable. The modelling must take into account second and third effects such as cross-modal coupling, which significantly increases the complexity. These phenomena are not as prevalent with category 7 connecting hardware due to the contact geometry which separates the cable pairs from each other.

To achieve Class E_A channel performance, a category 6_A module must have 3 dB better NEXT at 500 MHz than a category 6A module. This is a significant issue. To achieve this, it means that new modules need to be developed from the ground up as opposed to modifying existing designs, which is often the case with current category 6A modules on the market. Specifically, more compensation elements are needed to compensate for the additional coupling experienced at such high frequencies. More care is needed to be taken to separate the pairs from each other in the termination. Also, the termination process should be very precise and error-proof to ensure consistent performance. This no so much an issue in the laboratory for product development or testing, but more of a concern in the real world, where cabling contractors must install and terminate cables and connectors in a wide range of conditions and endeavour to maintain the high level of consistently good terminations to ensure compliance to EIA/TIA category 6A or ISO/IEC Class E_A channel performance criteria.

Recommendations

Today, a Class E_A channel is the highest performance channel available, based on the prevalent RJ45 technology. It not only ensures support for the 10 Gigabit Ethernet application, but it also extends the life of the cabling and ensures higher operational reliability. For these reasons, R&M recommends installing a Class E_A channel in new installations.

If interoperability is a requirement, then it is important to choose category 6_A connecting hardware. Category 6A modules simply cannot guarantee the higher Class E_A performance. Although the category 6_A component standards are taking longer to develop and subsequently publish, it will be worth the wait for the added performance that the cabling system will be able to achieve, which will then translate into fewer headaches for the end users who will build their high-speed data networks onto their cabling platforms.

Editor’s note: Under the World Trade Organisation (WTO) agreement, Australian standards in a wide range of fields follow ISO standards, where they are compatible with the Australian market. In the case of structured cabling, this rule has applied for many years, with our primary cabling standard - AS/NZS 3080:2003 ‘Telecommunications installations - Generic cabling for commercial premises’ following the equivalent ISO standard ISO/IEC 11801:2002, an arrangement that is anticipated to continue for future editions of AS/NZS 3080. This, therefore, underscores the significance of this article to our market.