Installation Pitfalls in Cat6 Cabling

One of the common question that appears on our mailing list is regarding the use of Cat6 cable in home installations rather than Cat5e. Here’s a very interesting piece which gives some of the problems you might not have considered…

“Due to Category 6 cabling, no longer does the installer have sufficient buffer to allow for expedient installation. A more stringent requirement for link and channel testing calls for more stringent installation practices. This, coupled with the need to revisit terminations when failures occur, has made for tedious and frustrating installations, with lost time and money for all involved. The need for more controlled terminations hasn’t existed for Cat 5 or Cat 5e installation standards. Often work practice variation would amount to some degree of variation within the tested result, but sufficient headroom ensured adequate buffer for such anomalies. But now with Category 6 standards, it is paramount that the structured cabling system of any high-speed network must be capable of minimising or eliminating inconsistent installation methods that cause costly performance variations.

It’s also worth pointing out that many of the items explained below won’t necessarily cause a link to fail a test on its own. But each one eats into the limited Cat 6 headroom by a few points of a dB, often more. It only takes an accumulation of these little things and the link fails.

Before getting in to the installation details, let us understand how the cable behaves at higher frequencies and the wave guide effect of wires.

No products found.

At DC and low frequencies, the current in a wire flows through the wire as you would expect. As the frequency gets higher, like the 33 MHz of 100BaseT the wire starts to radiate or act like an aerial, and the current only flows in the skin of the wire. Some of the energy is actually travelling as radio waves along the length of the wire, and we become susceptible to alien crosstalk.

At 75 Mhz, the frequency used for Gigabit Ethernet on UTP cabling, more of the energy is on the outside and the potential for alien crosstalk increases.

But once we get to the 250 Mhz limit for Cat 6 the wire is acting just like a wave-guide. Most of the electrical energy travels from one end to the other outside the conductor as electromagnetic energy. Quite a lot of it travels within the plastic of the insulation, so these are no longer just mechanical devices for protecting and holding the wires together, they are now an integral part of the dielectric of the cable. The sheath becomes a significant factor in the cable’s impedance make-up.

At these frequencies, any kinking, crushing or elongation of the cable has a proportional effect on impedance and a consequent knock-on effect on transmission performance.

Termination problems
Sheathing cut back too far on Cat 5e-who cares?: But at Cat 6 frequencies, the wave energy is actually flowing through the sheath. The sheath is a dielectric and cutting it away changes the impedance of the pairs causing echoes (return loss), and upsets the geometry affecting the crosstalk performance. Each manufacturer has different recommendations, but insist on not more than 3mm from the jack.

Wire routing, not just pair untwist: Tiny details like the routing of each wire at the back of the jack are critical. So, believe it or not, is the order in which the punch-downs are done. A wire can easily be damaged during termination and/or influence another. The best order depends on which end of the cable you’ve got hold of, and the direction the cable is leaving the jack.

Too much untwist: You know this one. But with Cat 6 there’s quite a bit of field experience now, which indicates that up to 2dB of NEXT can be saved by maintaining the twist up to the point of termination. If the lay of the wires doesn’t suit this, an extra half twist should be added. Why not make up and test some channels in the workshop with any new jack/cable combination before you go anywhere near site-it could save you a lot of rework costs.

Horizontal Cabling
There’s loads here that can and does go wrong. The internal geometry of Category 6 cables is so sensitive that you are talking about several dB’s worth of damage-and probably looking at total replacement of any cable that suffers this fate. We’ve seen all these sorts of things that frequently happen on site-like kinked cables (the installer’s straightened it out, but it’s too late, the plastic separator inside has deformed and will never return to the right shape).

And in fact there are many other little things that affect the dBs-like avoiding the prolific use of nylon cable ties. Nylon cable ties should only be used as a cable retainer at the patch panel and outlet jack in accordance with the manufacturer’s instructions. Never use the nylon type on Cat 6 horizontal runs-they pinch and cause local impedance to increase. Only use ‘hook and eye’ (e.g. Velcro) types but use them sparingly. If you are running cables in tray or basket, tie only where absolutely necessary. Don’t run the cables in nice neat bundles-the more the cables are parallel the worse will be the alien crosstalk effects. We’re not talking ‘rat’s nest’, but when it comes to Cat 6, neat is bad! Also, if you need to use cable ties don’t put them at regular spacings-the impedance effects will cause standing waves, which can lead to transmission problems.

Cable depth on trays or when laid directly onto concrete floors has issues too. The weight of the cable above causes cable ties, grit, and irregularities in the pathway surface to dig into the lower cables. The cross sectional shape and volume effects the rate of temperature build up from the transmission signal energy. This is insignificant for a shallow total depth of cables so it will not be a problem, but deep pile is out of fashion. Both of these issues effect the impedance characteristics of the cabling.

The other things we’ve found to have significant effects on site are nicks, cuts and grazes in the sheath. On one job, several cables in one run all failed. Checking with a TDR (time domain reflectometer) all showed a major reflection at the same distance. The problem was that another trade had dropped a piece of metal studding onto a cable bundle and then someone had trodden on it. The studding was now hidden amongst the nicked. Most of those damaged cables needed replacement.

We strongly recommend that you lay-in Cat 6 cable in preference to pulling it in. If you’re forced to pull-in, be very careful around corners. It’s a pain at the time, but it’s even worse if you have to go back and replace the cables to get the installation through test.

In fact pre-fabrication or partial pre-fabrication comes into its own with Cat 6, since you can draw the cables off the drums straight onto the bench-and you can terminate them in good clean, light and controlled factory-like conditions.

Obviously, all the points we made about termination problems apply here-the other point is how to best dress the cables into the cabinet/frame. Use as many ties (Velcro type) as you need to install vertical runs. Then, when all the cables are in place, remove any that are not absolutely necessary to support the Cat 6 cables. And on the horizontal runs keep them to a bare minimum and don’t tighten them.

Finally, we’ve found quite a few failures on links that worked fine before the cable was shoved back into the wall box or floor-box. Think what’s going to happen to the cables as you push back. If they are going to kink, crush or over-bend you’re going to have problems.The author is Technical & Product Manager – PremisNET, Krone Communications. ”


Want More? – Follow us on Twitter, Like us on Facebook, or subscribe to our RSS feed. You can even get these news stories delivered via email, straight to your inbox every day

Last update on 2024-04-08 / Affiliate links / Images from Amazon Product Advertising API

7 Comments on "Installation Pitfalls in Cat6 Cabling"

  1. is it really that bad?

  2. This is a good article from someone who understands the theory of transmission. Well done for making it available to all us newbies into the Cat6 world.

  3. I am a a graduate in analogue and digital tranmission technology, and I have to congratulate the author of this article. It is well written, very accurate and presented in a simple and straight forward manner. All the comments and advice given are spot on.

  4. I wired my own house with solid core cat 6, modular jack to modular jack, and to be honest it was not the best of installs as some of the cables go over electrical cables and I still get 90 Megabytes transfer from my PC to my NAS.

  5. I wired my house in Cat6, close to 90% of the cable runs failed testing. Mostly this was due to the installer stripping back too much and untwisting as he was used to with Cat5. In all but one case careful re-twisting at the termination was sufficient to pass the test. In the one case, it was probably a crushed or damaged cable, it met the cat5, but not cat6 tests; as the cable was in the plaster by this point it wasn’t worth ripping out. Next time, i’ll invest more time in ensuring the installer really knows the difference between cat5 and cat6.

  6. William Watson | February 3, 2020 at 12:22 pm |

    ” insist on not more than 3mm from the jack.”
    3mm is an awfully short distance – do you mean 3cms?

    I have a building which had (still has!) Systimax UTP Cable in it installed in the 1980’s. By the early 2000’s we had several links up the 4 floor riser running at 1Gbps with no problem. As we let parts of the building out since 2005 tenants insisted on installing their own 5e cabling… to connect to BT services in the basement (100Mbs, tops)! Each time I pointed out that the original (pre standards) Systimax was more than good enough, the tenants’ IT installers insisted on rewiring!

    I also muse on the fact that ou Open Connect FTTC internet service at home is delivering 80+Mbs down old CW1308, with several junctions on the way. The street cabinet is about 600 m away. Remarkable.

  7. William,
    You are correct up to 1Gbps. The change from 100Mpbs to 1Gbps was relatively simple, use 4 pairs instead of 2.

    However to go to 10Gbps, there are no more pairs, so the cable must be able to run at higher frequency. The challenge is when you to run cat6 at full speed of 10Gbps there is almost no scope for imperfect implementation.

    3mm is the limit at connectors. It was installers stripping back an inch and liberally untwisting the pairs that killed the performance at my place. Everything had to be reconnected with the shortest possible distance and the pairs kept twisted right to the end.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.