Cat6A Permanent Link Certification and TIA-568

Copper cabling is the layer everyone assumes is fine and nobody re-tests, which is why intermittent 10GBASE-T problems so often trace back to a link that was never certified. ANSI/TIA-568 defines two link models — permanent link and channel — and field certification to the permanent link is what proves the installed cable, jacks and terminations meet Cat6A performance up to 500 MHz. This guide walks through what each test parameter actually measures, the field-tester settings that produce defensible results, and the alien crosstalk requirement that separates a genuine Cat6A installation from a hopeful one.

Permanent link versus channel

A permanent link is the installed cabling between the patch panel in the telecommunications room and the work area outlet — no patch cords. A channel adds the patch cords at both ends. Certifying the permanent link is the correct target because the installed cabling is the part that cannot be changed later, while patch cords are consumables you can swap.

Permanent link test adapters must be Cat6A rated.
Category 5e/6/6A patch cords must NOT be used as permanent link adapters:
their performance degrades with use and produces false return loss failures.

Build a reference link and run a daily self-test on the tester before testing the customer's cabling. A tester whose own adapters have drifted will fail good links and send crews chasing ghosts.

The parameters that fail installations

Field tester configuration for Cat6A

Required tester capability (per ANSI/TIA-1152-A):
- accuracy level IIIe (or 2G)
- permanent link adapters with a centred Cat6A reference plug
- measurement over 1-500 MHz
- frequency step: 150 kHz (1-31.25 MHz), 250 kHz (31.25-100 MHz),
  500 kHz (100-250 MHz), 1000 kHz (250-500 MHz)
- result storage for >10,000 records
- stores time-domain crosstalk data for marginal NEXT results
- stores TDR data for marginal return loss results

Both worst-case value and worst-case margin must be reported, in both directions for NEXT and return loss. Any margin inside the tester's accuracy window is flagged with an asterisk — treat those links as marginal and fix them while the crew is still on site.

Alien crosstalk: the Cat6A-specific test

Because 10GBASE-T uses all four pairs at high frequency, energy leaking between adjacent cables becomes significant. Cat6A therefore adds Power Sum Alien NEXT and Power Sum Alien ACR-F. Alien crosstalk is tested on a sample of links rather than every link, but the sample must include the worst case: the densest bundle, the longest parallel run, and the links in the middle of the bundle. Cable construction — separate foils, increased pair spacing — is what makes a cable compliant, which is why a "Cat6A" cable without the construction carries a limit that the installation cannot meet.

Field documentation to hand over for acceptance:
- one test report per link (not per cable), including the limit selected
- tester model, serial number, calibration date, firmware and limit database version
- adapters used, operator name, date/time stamp
- TDR/TDX records for marginal or failing results

Diagnosing failures in the field

1. Single link fails NEXT only: re-terminate the jack with the pair twists maintained to within 13 mm of the termination point.
2. Return loss fails after a successful test yesterday: suspect the tester's patch cord or adapter, not the link — swap adapters and re-test before touching terminations.
3. Insertion loss consistently marginal on 90 m runs: measure the actual route. Bends around cable trays add length quickly, and the limit is applied to the measured length, not the drawing.
4. Alien crosstalk failures across a whole bundle: the cabling is not compliant. Separation or a compliant cable type is the only fix; re-termination will not change ANEXT.

Best practices for installation and handover

• Label every outlet and patch panel port before testing, and make the reported circuit ID match the label identity — a test report you cannot correlate to a port is useless during an incident.
• Test every link. Sampling in-channel tests is not acceptable for certification.
• Keep the tester database and upload results daily; a project database is the difference between a warranty claim and an argument.
• Match the standard to the application: Cat6A for 10GBASE-T to 100 m, Cat6 for 1/2.5/5GBASE-T, and fibre for anything above 10G. Where the run crosses between buildings, the fibre link budget rules apply instead.
• Record patch cord make and model. Patch cords are the weakest component in a channel and the first thing to replace when a link misbehaves — the mated NEXT and return loss performance between plug and jack dominates channel margin.

原文链接：https://assets.tequipment.net/assets/3/7/WhitePapers-HowToTestToTIAEIA.PDF