Cat 6 Bid on a 10GBASE-T Backbone Near VFD Busway
A 60-cabinet data hall is planning a 10GBASE-T migration. The winning bid specifies Cat 6 UTP with 95 m permanent links from the telecom room to each cabinet, using 6 m of patch cord at the telecom end and 5 m at the cabinet end. For most of its length the main horizontal tray runs parallel to a 480 V busway that feeds several VFD-driven air handlers, at roughly 12 in of separation. The customer's IT manager asks you to confirm the design is ready for 10GBASE-T and to sign off on the media selection before the purchase order is issued.
1. Evaluate the proposed link against 10GBASE-T channel requirements. What has to change?
Read the worked answer
Length fails even before media is considered.
- Permanent link: 95 m against a 90 m limit - over.
- Channel: 95 + 6 + 5 = 106 m against a 100 m limit - over, and the cords alone use 11 m of the 10 m allowance.
Media is also wrong for the application. Cat 6 is not a 100 m 10GBASE-T medium in a worst-case alien crosstalk environment; it is generally limited to around 55 m and requires alien crosstalk mitigation. Cat 6A is the correct choice for 10GBASE-T at the full 100 m, and a shielded Cat 6A variant buys margin where the environment is noisy.
What to specify:
- Cat 6A (or shielded Cat 6A) UTP/F-UTP throughout, including patch panels and cords.
- Re-survey and re-route so the permanent link is 90 m or less, with cords and equipment cables inside the 10 m channel allowance.
- Cords and patch panels rated to match the cable; a Cat 6A cable terminated on a Cat 6 panel caps the run at Cat 6.
The answer to the IT manager is that the bid is non-compliant on two independent grounds, and that swapping only the cable or only the length fixes half the problem.
2. The 12 in parallel run past the VFD-fed busway is questioned. How do you assess and mitigate it?
Read the worked answer
Twelve inches of parallel separation from a 480 V busway is not automatically a code violation, but VFDs are broadband noise sources and the parallel geometry is the worst case for coupling. Treat it as a real alien crosstalk and EMI risk and design it out.
Mitigation, in order of preference:
- Increase separation and shorten the parallel run. Distance is the cheapest and most reliable fix.
- Change the geometry. Cross the busway at 90 degrees rather than running alongside it.
- Contain it. Metallic conduit or a closed metal tray with a cover attenuates coupling substantially, provided the enclosure is bonded.
- Shield the cable. Shielded Cat 6A/F-UTP with proper bonding and grounding to the telecommunications bonding infrastructure at both ends; a shield terminated at one end only is a antenna, not a shield.
- Verify, don't assume. Require field testing that includes alien crosstalk on the worst-case runs, plus a sample of runs at the closest approach to the busway.
Document the final separation, the routing, and the test results in the as-builts, so that future work in the tray does not quietly undo the mitigation.
3. The customer asks whether they can just accept the bid and let field testing catch any problems. How do you respond?
Read the worked answer
Explain the difference between testing and design. Field testing verifies that an installed link meets a performance limit; it cannot make a 106 m channel shorter, and a failing link discovered after the ceiling and cabinets are closed means re-pulling at the customer's cost and schedule risk.
Also point out that alien crosstalk problems are not reliably caught by routine per-link autotests unless the test configuration includes them; the bid could produce a folder of passing reports and still deliver a link that will not run 10GBASE-T at the required error rate.
The correct path is to reject the bid on the record now, with the two specific non-compliances cited, and require a revised design before the purchase order. That protects the customer and protects you from being the person who signed the media selection.