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.
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.
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.
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. ”