Still partly shrouded in mystery, one point is certain: It is up to the end user to specify the amount of testing that may be required.
September 1, 2008
When you walk into cabling research laboratories in our industry you cannot help but notice the rows of cable trays filled near to capacity with bundles of cables connected to racks of patch panels with a variety of different connecting hardware types and configurations.
It is representative of many hours of testing and hundreds of gigabytes of data. What is it that we are looking for, what is it that were trying to understand, what is it that is still partly shrouded in mystery?
Would you believe that we are looking for aliens, well let’s say more precisely that we are studying the alien crosstalk behavior of Category 6A cabling components in different cabling configurations. Alien crosstalk is a new parameter that is specified in the Category 6A standard (TIA 568 B. 2-10) that was published in April 2008.
It is a critical parameter to support 10 Gigabit Ethernet over copper for distances up to 100 meters. There is a lot that we have learned about alien crosstalk behavior and in this article I give some insight into what to look for and whether it is something to be concerned about.
First, what is alien crosstalk? The TIA standard defines alien crosstalk as an unwanted noise coupling from a “disturbing” 4-pair cabling circuit into another “disturbed” 4-pair cabling circuit.
A cable pathway contains many cables that are run close together and a patch panel has many connectors that are side by side, so there are many disturbing circuits that can have an effect on any disturbed circuit.
In a real life installation it becomes impractical to test all these disturbing noise sources. That is why the onus is on the cabling vendor to design their products and to do sufficient testing to ensure that alien crosstalk requirements are met for all the worst case cabling configurations that are defined in Annex D and Annex J of the TIA 568 B. 2-10 standard.
One of the most significant sources of alien crosstalk noise is between connector modules in a patch panel. In analyzing a lot of measurement data for different vendors’ Category 6A products, we found that there is considerable variability in alien crosstalk performance at high frequencies due to the method of compensation/isolation that is used for the connector design.
Here are some things that we have noticed:
• The PSANEXT for a channel in a cross-connect configuration (2-connectors in close proximity) can be considerably worse than for an interconnect configuration (only one connector at near end) or for a permanent link.
• Short channel configurations with multiple connectors in close proximity can exhibit a sharp drop off in PSANEXT performance at high frequencies.
• The stability of transmission performance for Category 6A patch cords due to bending and mechanical stress is very important.
Is alien crosstalk something that an end user should be concerned about? Given the fact that there is so much difference in performance depending on the design of the cabling components, as an end user making an investment in a Category 6A infrastructure, I would want to do some pre-qualification testing as well as some sample testing in the field after installation.
The pre-qualification testing would be on both short and long channel configurations, including a cross-connect if applicable. The TIA Category 6A standard does not specify the extent of testing for internal or external (alien crosstalk) parameters in the field.
It only provides general test strategies for selecting the disturbed and disturbing cabling. It is up to the end user to specify the amount of testing that is required.
For alien crosstalk, it is recommended to test 1% of permanent links up to a maximum of 30 links after installation.
A sample size of 30 would give a very good indication of statistical variance and worst case performance (CPK = 1).
Finally, it should be noted that the use of shielded Category 6 / Category 6A cabling does not preclude the need for alien crosstalk testing in the field.
I have heard on several occasions from a neutral party that the alien crosstalk performance of shielded cabling can be inferior to UTP cabling, due to the effects of the shield terminations or due to the connector design.
It would be prudent to do a sufficient amount of alien crosstalk testing to validate the design and to verify the installation.