Cat6 vs Cat6A: how to choose for your building

Cat6 vs Cat6A is one of the few cabling decisions in a building that is genuinely hard to reverse. The cable itself is a minor line on the estimate; the labour to pull it, and the finished wall or hard-lid ceiling in front of it, are not. So the useful question is not which category performs better — Cat6A does — but which pathways in your building justify paying for it once rather than paying for the whole pull twice.

Cat6 vs Cat6A: the differences that change a decision

Four differences actually affect a specification: how far 10GBASE-T will run, how the cable behaves in a dense bundle, how much room it takes in the pathway, and how long it takes to terminate. Everything else is detail.

Diameters and bend radii vary by manufacturer and by whether the cable is shielded. Confirm both against the datasheet for the product actually being installed, because pathway sizing depends on them.

The line in that table that does the real work is alien crosstalk. It is interference between adjacent cables in a bundle rather than between the pairs inside one cable, and it is the reason a Cat6 channel cannot be relied on for 10 Gb/s at full length. Cat6A controls it in the cable itself — a thicker jacket, greater separation between adjacent cores, sometimes a foil screen. That is where most of the extra diameter comes from, and it is not something a better patch panel or a careful terminator can add later.

The cable is the cheap part: what a re-pull actually costs

On a horizontal drop, the cable is a small fraction of the installed cost. The rest is the pull, the terminations at both ends, the testing, the labelling, and the trades that close the wall behind it. Upgrading a run from Cat6 to Cat6A raises the material cost and adds some termination time, but it does not double the drop.

Pulling that drop a second time is a different exercise. It means a lift or scaffold, ceiling tiles or drywall out, a fire-stop reinstated, patching and painting, and — in an occupied building — coordination with tenants or a strata for access. That work is often a multiple of the original drop cost, and it is disruptive in a way the first install was not.

So the test for any given pathway is simple. Ask what it would cost to get back into it in year seven, and how likely you are to need to. Where the answer is "expensive and likely", specify Cat6A. Where it is "cheap or unlikely", Cat6 is a decision you can put your name to.

Where Cat6 is genuinely sufficient

Cat6 remains the right call for short, accessible drops serving equipment that will not exceed 1 Gb/s within the life of the cabling. In residential and multi-family work that covers most in-suite outlets: a TV, a desk, a printer, a games console. The run is usually well under 40 m, there is one cable per box rather than a dense bundle, and the endpoint has a 1 Gb/s port on it.

It also covers most single-drop office and retail outlets where the desk equipment is a laptop dock or a phone. If a given position later needs more, one drop is a far smaller re-pull than a whole floor.

Cat5e's remaining niche

Cat5e still has two honest uses. The first is short auxiliary runs to devices that will never need more than 1 Gb/s and draw modest power — card readers, request-to-exit devices, door controllers, intercom stations, some sensors. A certified Cat5e link carries 1 Gb/s and standard PoE without complaint, and a reader does not care about headroom. This is common on access control rough-ins, where the device count is high and the data requirement is trivial.

The second is existing plant. In a retrofit, certified Cat5e that tests clean is usually worth keeping for 1 Gb/s endpoints rather than replacing on principle. Test it, document it, and spend the budget on the runs that genuinely need upgrading.

What has largely disappeared is the case for new Cat5e horizontal drops. The material saving over Cat6 is small next to the labour, and fewer cable types on site means less chance of the wrong one landing in the wrong pathway.

Where Cat6A earns its cost

Wireless access point feeds. This is the clearest case. Modern access points have multi-gigabit uplinks and draw from the higher PoE tiers, they live in ceilings that are awkward to reopen, and they are replaced two or three times over the life of a building. Many designs also feed each position with two cables so a future dual-uplink AP does not require a return visit.

Dense commercial ceilings. Where dozens of cables share a tray or J-hooks above an open office, alien crosstalk and bundle heating are both real, and both worsen as the plant fills up. Cat6A is specified for exactly that condition.

Long runs and risers. Any run pushing past roughly 40 m loses the option of 10GBASE-T on Cat6, and runs from a floor closet to the far corner of a large plate are precisely the ones you cannot shorten later. Vertical and inter-closet links are the same argument with worse access.

Camera and device runs behind finished surfaces. A surveillance drop to a high-resolution camera or a PTZ with a heater is a long run, a PoE-heavy run, and a run that needs a lift to redo. The same reasoning applies to anything behind a hard-lid ceiling, in a poured slab, or crossing a fire-rated assembly where the penetration would have to be reworked.

What the larger cable costs in the pathway

Diameter is the practical penalty, and it compounds because pathway capacity goes with cross-sectional area, not with diameter. A 7.6 mm Cat6A has roughly twice the cross-section of a 5.5 mm Cat6, so a conduit or sleeve that comfortably takes a given count of Cat6 takes about half as many Cat6A. If the switch to Cat6A happens after pathways are sized, the sleeves and stub-ups are what fails, not the cable.

Bend radius scales the same way. The common working rule is four times the outer diameter during and after installation, so Cat6A needs a noticeably wider sweep — which matters most at the back of a shallow wall box, at the top of a rack, and where cable enters a conduit body. Pull tension limits are similar across categories, but the heavier cable reaches them sooner on a long run with several bends.

Termination adds time per drop. The conductors are stiffer, the pair geometry is less forgiving, and shielded Cat6A needs shield continuity through every connector plus a bonding scheme back to the telecom bonding point. Done properly that is a real labour line; done improperly, a screened system performs worse than a good unshielded one.

PoE heat runs the other way, in Cat6A's favour. Current through a bundle raises its temperature, temperature raises insertion loss, and higher insertion loss shortens the usable length of every link in that bundle. Larger conductors, a larger overall diameter and — where present — a foil screen all help a bundle shed heat, so Cat6A sits at the higher PoE tiers with less length derating than Cat6 or Cat5e. Field practice matters as much as category here: keep bundles modest and spaced apart rather than cinched into one mass. Jacket fire ratings and the fill limits applied to pathways are set by the authority having jurisdiction, so confirm both locally before pathways are ordered.

Category alone does not guarantee performance

A channel performs to its weakest component and to the quality of its installation, not to the label on the box. Cat6A cable terminated on Cat6 jacks and patched with Cat6 cords is a Cat6 channel. Excess untwist at the punch-down, an over-tightened tie wrap, a kink from a tight pull, or a run stretched past its tension limit will fail a link of any category, and none of that is visible once the ceiling is closed.

Copper-clad aluminum cable sold as Cat6 is the other common trap. It fails DC resistance requirements, behaves badly under PoE, and is not a compliant product for a permanent installation. Confirm that what arrives on site is solid copper before it goes in the wall.

The only thing that resolves this is field certification. A wire-map tester proves the pairs are in the right order; it proves nothing about performance. A certification tester reports insertion loss, NEXT and PS NEXT, ACR-F, return loss, propagation delay and delay skew, and DC loop resistance and resistance unbalance — the last two being what tell you the link will carry PoE. On links intended to run 10GBASE-T, alien crosstalk is the additional parameter, and because measuring it in the field is slow it is usually handled either by using a qualifying pre-tested system or by testing only the specific links that need it.

Ask for the results as a deliverable: one report per link, with the link ID matching the label on the jack and the patch panel port. That file lets a future contractor troubleshoot a fault instead of re-pulling the run, and it is the record that settles an argument about deficiencies at handover. How we scope and test drops is set out on the structured cabling page.

What to do next

Walk the drawings pathway by pathway rather than setting one category for the whole building. Mark the runs that are hard to reopen — ceiling APs, camera positions, risers, anything over roughly 40 m, anything behind a hard lid or through a rated assembly — and specify Cat6A there. Leave the short, accessible, single-device drops as Cat6, size pathways for the larger cable before sleeves are ordered, and write field certification and labelled test results into the scope rather than leaving them to be negotiated at handover.

If you want that walked on your own drawings, Orbit Automation is a licensed and insured low-voltage integrator in Surrey serving Metro Vancouver and the Fraser Valley, and can review pathways and a cabling spec on site.