I’ve been meaning to write this post ever since we moved into our new house eighteen months ago. Before we took up residence, we spent six weeks gutting the place. This meant floor-boards up all over the place, holes in floors, walls and ceilings, and the perfect opportunity to lay gigabit ethernet cable wherever I wanted! The overall structure of the network has changed in the twelve months since, but this is the setup we’ve finally ended up with…

Although I’ve titled this post “Structured cabling”, what I’m doing here is obviously a world away from proper, commercial structured cabling. My aim in laying cable throughout the house was to hide away the miles of untidy stuff we had in our old place, but also to lay it to places I’d never ever get away with a visible cable. A prime example is the TV in the bedroom; I had visions of a PS3 or network media streamer there delivering BBC iPlayer and playing music from the media server. I knew if I raised this with my better half she’d never go for it, but I also knew that if I just laid some cables under the floor she wouldn’t have a clue what I was doing! There were a few things to consider with cabling:

1) Cat5e or Cat6?
Both are suitable for use with gigabit ethernet (GbE), but Cat6 does have more stringent regulations and can therefore be used for connections up to 10Gbit/second. A couple of factors influenced my decision here. Firstly, most home networking hardware is only GbE, so this is what limits the connection speed (not the cabling). Secondly, the bend radius of Cat5e can be smaller than that of Cat6 without affecting the signal. As I wasn’t sure quite how tight I was going to have to run each cable, Cat5e gave me more flexibility. A final factor is that Cat5e is cheaper and easier to get hold of than Cat6.

2) Shielded or not shielded?
The use of shielded cable is a controversial topic. Lots of people suggest that it’s not required and is overkill for all but the most extreme circumstances, while others take the opinion that it can’t do any harm. In my case, the tight timescale for refubishment of the whole house meant that network cabling was very much squeezed in around other more important activities, like knocking down walls and re-building floors. I therefore didn’t have time to make nice dedicated holes in the floor joists, and so consequently the network cable was run through the same holes as the existing mains and/or lighting cables. It’s common knowledge that you shouldn’t really do this – and I separated them as much as possible with wooden blocks and the like – but I figured shielded wouldn’t hurt.

There are several different types of shielded cable – Wikipedia has a good breakdown, but essentially you can shield the entire cable, the individual twisted pairs within, or both. I went for what’s commonly known as FTP, which has a single foil screen around all four twisted pairs. There’s also a ‘drain wire’ in the cable, which provides the ground required for any shield to work. This is a somewhat essential step, as is making sure that the whole ‘chain’ of your network is shielded. There’s no point using shielded cables and shielded RJ45 connectors (technically 8P8C, but let’s not go there) if your patch cables are then unshielded. You’ll also need your modules (the sockets that slot into the face plates on the wall) to be shielded too. All of this stuff was easy enough to get hold of – the cable came from Maplin and the modules, face plates and patch cables from a company whose name I’ve completely forgotten. I’ll update this post if I ever remember. As it happens, I wasn’t able to ground the cables at either end, so my shielding is effectively redundant! However, if you do the job properly, unlike me, note that it’s generally recommended to only ground cables at one end. If you ground at both, you can set up a ground loop. While in an audio equipment context this can cause an audible hum, on a network it’ll just cause dropped packets and general interference, so is to be avoided.

3) Solid or stranded?
Solid cable is harder to work with, but transmits data more effectively. Stranded cable isn’t as good over very long distances, but is easier to work with for things like patch cables. I chose solid for all my main cable runs, and stranded for the patch cables.

Once you’ve decided on your cabling, you just need to lay it. We were having the whole house ripped to pieces, so I was able to hide the cable away pretty effectively. The main setup is the pseudo-patch panel in the loft, where the router sits.

Obviously I could have run the cables directly into the router rather than through the patch panel, but the panel keeps the cables nice and tidy, and also means I can coordinate the cable colours with the cables used downstairs. All the sockets are also labelled, but it’s much more intuitive to use colours. From here, the cables run down through the upstairs ceiling and through some trunking to floor level.

At floor level, they go under the floorboards and split off into the relevant rooms. Unfortunately our renovation project on the house was so manic that I didn’t have time to take any photos of that section of the project.

The trunking (and all the pipes) are being boxed in, so won’t be visible soon, and are hidden by a cupboard at the moment anyway. Finally, they pop up again from the floor wherever required. I managed to get the cables in before the skirting board went on, so that’s nice and neat. Alas, the walls had all been plastered by then, so a little bit of trunking was required, but it isn’t too bad. I’ll replace it with some of the low-profile stuff at some point, which should minimise the impact even further.

Obviously for downstairs a different tactic was required. The cable to the kitchen goes between walls to the cupboard under the stairs, where the media server is based. The one for the lounge actually runs out of the gable end of the house in the loft, then goes around the outside and into the lounge in the corner behind the TV.

Once the cable reaches the required point, and obviously at the patch panel end too, you’ll need to wire it in. There are two wiring standards, T568A and T568B (link includes pictures). Some sockets will come pre-marked with the relevant colours, so you can just match the wires to the terminals and everything’s fine. In essence, it doesn’t really matter which wiring scheme you choose, as long as you use the same one at both ends of the cable! A punch down tool is really very useful for this entire process. Remeber you don’t need to strip the insulation from the individual strands first; the contact points will bite through the insulation to make contact with the conductor. Yes, you can – in theory – use a small flat screwdriver instead of a punch tool, but it won’t trim the ends for you automatically, and there’s far more chance of something going wrong in the hundred-or-so connections you’re likely to be wiring.  Bear in mind that these are generally going to be in awkward places to access, so it’s best to just pay a few pounds and get it right first time around.

So there we go: network cables to all rooms, happily running at gigabit speeds. Out of all of them (96 connections by my reckoning) I only wired one connection up wrong. In these situations there isn’t really much you can do other than open up the relevant sockets when you discover a lack of connectivity. Obviously a network tester can help, but I didn’t find it necessary in my case.