Category-5 cable is what you use to wire an Ethernet network. Also known as 10BaseT/100BaseTX cable. Think of it as telephone wire on steroids. Sure, it's got the same twisted-pair concept behind it, but the material is of far greater quality, and more pairs are used for a connection (4 pairs for Ethernet, as opposed to one pair for one telephone line.)

AFAIK, there's two different methodologies here-

568A wiring layout

      |white-green  |-
      |green        |-
cable |white-orange |-  
------|blue         |- contacts  
------|white-blue   |-  
      |orange       |-
      |white-brown  |-  
      |brown        |-

568B wiring layout

      |white-orange |-
      |orange       |-
cable |white-green  |-  
------|blue         |- contacts  
------|white-blue   |-  
      |green        |-
      |white-brown  |-  
      |brown        |-

Both 568A and 568B are wired the same way on both ends. And of course, because we'll all need it someday:

Crossover (for connecting two machines without using an ethernet hub)

wiring layout- one end

      |white-orange |-
      |orange       |-
cable |white-green  |-  
------|blue         |- contacts  
------|white-blue   |-  
      |green        |-
      |white-brown  |-  
      |brown        |-

other end

      |white-green  |-
      |green        |-
cable |white-orange |-  
------|blue         |- contacts  
------|white-blue   |-  
      |orange       |-
      |white-brown  |-  
      |brown        |-

How to Crimp Cat-5

Tools and Materials:

Step 1.

Measure your Cat-5 Cable carefully. Give yourself a couple of extra feet. It would suck if you went through all the work and made a perfect Cat-5 cable and then found it wasn't long enough. Too long is better than too short, however, remember that the maximum operable length of Cat-5 cable is 300 feet.

Step 2.

Strip about 3/4 inch of outer plastic shielding from the end of the cable-to-be. This extra amount is for adjustments in length later on. Be careful that you do not cut or nick any of the smaller wires inside. Even if you only nick the shielding but do not cut the wire itself, you have already degraded your cable to the point where it may be unusable. Do it over as many times as necessary. That's what your couple of extra feet are for. I recommend using single sharp razor, or one blade of a pair of scissors to score the outside of the shielding and then pull it off with your hands. This minimizes the danger of damaging the wires inside.

Step 3.

Once you have the outer shielding off, decide on your wiring scheme and line the wires up appropriately, making sure to smooth out the twists in the wires so they lay flat. This may take some practice.

There are various different wiring diagrams for non-crossover ethernet cabling. They really don't matter at all in the least, as long as the heads are absolutely identical on both ends. The electrons don't care what color the wire they are traveling over is.

Due to convention however, there are some different methodologies. These are viewed from left to right when you are looking at the bottom of the RJ-45 head. (The metal pins are on the bottom and the plastic clip thing is on the top)

568A goes W-Green, Green, W-Orange, Blue, W-Blue, Orange, W-Brown, Brown.

568B goes W-Orange, Orange, W-Green, Blue, W-Blue, Green, W-Brown, Brown

I usually arrange it as Orange, W-Orange, Blue, W-Blue, Green, W-Green, Brown, W-Brown. But as I said earlier, it's just a matter of convention and there is no technical reason to do it in these prescribed orders.

Step 4.

Once you have the wires laying flat, it's time to cut them all to the same length. This is alot harder than it seems. It may take 3 or 4 cuts to get them to a satisfactory sameness because the twists in the wire shift as you move and as you try to fit the wires in the head, and if your wires are not all pushed up to the end of the head, and snagged firmly in the pin-prongs when you crimp, your cable will not work. Take your time. This isn't a race.

Step 5.

Final step. Time to crimp. Before you crimp, make sure the wires are all in the correct order (this is more important for the second head than the first, as you can simply emulate your mistake in the second head if you make one in the first) and pushed all the way up to the end of the head. If you crimp before you are ready, you will need to cut the ruined head off and start all over again. Once an RJ-45 Head has been crimped it cannot be reused.

Once that's all settled, shove the head in the crimper, being careful not to dislodge anything, and squeeze hard. If you don't squeeze hard, the prongs won't penetrate the wire and your cable won't work. Whale on that thing.

Repeat these steps for the other end of the cable.

Don't forget to test your cable with the cable tester before you run out to do anything important.

It's not quite true that the color scheme doesn't matter. Each of those twisted pairs can be thought of as carrying a signal in one direction and its return (you can think of it as the ground wire) in the other direction. For the ethernet to work as designed (and especially if you're going to be using it as 100BaseT) it's vital that each signal and its return be part of the same twisted pair.

Once upon a time, ethernet used fancy coaxial cable, since coaxial cable has much better propagation properties for electric fields than plain wires do. It turns out that twisted pairs are better at propagating fields than plain wires, too, and eventually someone figured out how to run ethernet over them instead, which was a boon, because twisted pair wiring is much cheaper and easier to install than coax.

So, no, it doesn't matter where the orange, blue, green, and brown go per se, but pins 1 and 2 must be a twisted pair, and pins 3 and 6 must be a twisted pair, and pins 4 and 5 must be a twisted pair, and finally pins 7 and 8 must be a twisted pair.

It's kind of a strange pattern, but there's a bit of a method to it. For one thing, the middle 4 pins (3, 4, 5, and 6) end up being wired just like a standard RJ-14 two-line phone jack, so you can easily run phone signals through your network wiring if you have to.

If you screw the wiring up, for example if you make pins 3 and 4 be the green pair, and 5 and 6 be the blue pair (which is an easy mistake to make), it might work for 10BaseT, but you're skating on vanishingly thin ice / pushing outside the envelope / etc.

...and yes though it is true that a trade magazine was able to run Ethernet over barbed wire at a major trade exposition, it is still not a good idea.

When crimping RJ-45 plugs on cat-5 cabling, it is best to follow a predetermined wiring code, such as the EIA/TIA 568A and 568B cabling standards. One reason for these predetermined wiring patterns is that the lengths and quality of the copper wires in a CAT 5 cable vary by pair. For instance, generally speaking, the brown pair in most brands of CAT 5 has worse transmission characteristics than the other pairs. This is a combination of manufacturing practices and a longer twist. This, however, may not hold true as often on CAT 5e.

Most cabling standards also include plug transmission specifications, minimum cable specifications, and termination best practices (such as minimizing untwisted termination length).

One common problem people have the first time they crimp an RJ-45 on CAT 5 is a weak jacket crimp. The outer jacket of the cable should be in the plug, well past the jacket crimp point (usually a small wedge of plastic in the belly of the connector that is pushed up by the crimp tool). This way, the jacket crimp point is pressed onto the jacket far enough down to have a good grip.

"And remember, 568A is not just a standard, it's a way of life..." -My Cabling Instructor, 1994

...without a crimper.

A true story: About a month ago I needed to network 2 PCs, but all I had was a patch cable.

Well, I had two PCs too. But the point I'm trying to get across is that I didn't have a crossover, which is what I needed.

Now, I'm a bit of a newbie at all this, but I knew I needed at least one RJ-45 terminator. The plan: cut off one end of my patch cable and cross it over.

So I headed down my local corner techie-shop. Yes, there's one about ninety seconds from where I live. I asked for an RJ-45.

"Certainly", said the man in the shop with a big smile, "that'll be 50 pence".

His smile unnerved me. I was about to pay when he asked "Do you have a crimper?"
"What's a crimper?"
"You need it to crimp the wires. You NEED it."
"How much?"
"Sixty-five pounds plus VAT." he said with a look of triumph. That's over 100USD1.
"In that case I think I'll try without the crimper", I said.
"In that case", he laughed, "you can have the RJ-45 for free".

</truestory>

If you want to crimp a Cat-5 and you don't have a crimper, don't despair. You can do it, whatever evil shopkeepers may tell you.

You will need:

  • A pair of pliers (optional) - the small ones on a pen knife are ideal.
  • A thin straight screwdriver, or a blunt, thick bladed knife; thin enough to fit in the gap of one pin of an RJ-45, thick enough that it won't easily bend or break.

Do everything as you normally would or as a guide might tell you, right up to the point where you're about to crimp. Now put the RJ-45 in your pliars and gently push the pins down. Push hard enough so that the pins to go down into the plastic casing, but not so hard that you deform the plastic. The pins are also likely to move forward (towards the end of the RJ-45). We'll fix that in a bit.

Then take your screwdriver or knife and place the end -- carefully --- on one of the metal pins of the RJ-45

Now push

Harder

A bit more

Nice. You should be able to see that the pin has gone down more, and is now kind of sunk in. Ideally, you should be able to see the prongs at the bottom of the pin have made holes on the other side of the wire they are piercing. Try to push the pin down and back (towards the cable) at the same time.

lather, rinse, repeat.

Finally, use your tool to flip the chord grip (the little plastic flap thing) on the RJ-45. A thick screwdriver can be good for this. It should hold the cat-5 tightly in place to stop stress on the cable pulling on the actual wires. And there you have it, a crimped Cat-5.

Now, I wouldn't recommend this to people who make cables daily. It can really hurt your hand and arm. But for the occasional home networker, it's probably the most cost effective solution.

Update 12/11/02 I've now used this method to network a whole house, with basically no problems.
For other minimalist network solutions, see my esteemed flatmate's Running Two Connections Down One Piece of Cat-5

1 Crimping tools can be *much* cheaper than this. ymmv.
Thanks to lj] for suggestions.

I have been told by my salesrep at Graybar that the standard nowadays is to use T568A for phone connections with Cat-3. I scoff at that, and use USOC for Cat-3. However, just like scs says above, using 568A or T568B will work with phones, which use either one pair, or in rare cases, two pairs.

However, since 568B will work the same (and the only difference is color), I use 568B on all Cat-5 cables. One really good reason to follow the color standard for pairs is that you can crimp one end and punch down the other end to a modular jack, such as when cabling an office or house. This way, you have crimp ends on all of the cables in the wiring closet, and modular plugs at all of the stations. Another great reason is that you can re-crimp one end without having to examine the other end.

I have found that Allen Tel Cat-5e patch cables do not follow the pair color standard, and use random colors, while still conforming to the pairing rules. This is probably true of other brands.

Another thing to note is that you can purchase crimp ends which are designed to work with stranded conductor cable in addition to solid conductor. These have a smoky appearance, rather than the clear plastic of the ends designed for solid conductor only. The difference is the shape of the 'teeth' that bite into the wire.

Crimping a Cat-5 loopback

A cat-5 loopback is a very simple device which simulates the behaviour of an active network card. Loopbacks are very useful when running wiring over long distances, as they use no power, require no interaction at the other end when testing a cable run, and are cheap enough that it doesn't even matter whether they are collected again or not.1

Loopbacks are created by wiring the upstream twisted pair to the downstream twisted pair. The easiest way to do this is inside an RJ45 plug - no cable required. Cut about an inch of cat-5 off the end of your spool, and untwist the pairs - this is enough for 4 loopbacks. Fold the wires in half, and insert both ends into the plug, connecting pin 1 to pin 3, and pin 2 to pin 62, then crimp. Though it's not necessary, I recommend using the same colour for both wires, to aid in describing individual loopbacks over the 'phone. Most RJ45 plugs are transparent, so it's quite easy see which is (for instance) the brown or green loopback.

 ______________________________
|                       |______| 
8===                    |      |\__________ 
7===                    | C    |  
6===xxxxxxxxx           | o    | 
5===        x           | r  G |  Cable.
4===        x           | d  r | 
3===ooooooooxoooooooo   |    i | __________ 
2===xxxxxxxxx       o   |    p |/ 
1===ooooooooooooooooo   |______|           
|_______________________|______|

As seen from the wiring side of the RJ45; you can't see the 
catch, because it's on the other side of the plug. 

Loopbacks appear to most ethernet devices as a switched-on ethernet device at the other end of the link. When one is plugged into a network card, hub or switch, its link light should switch on. Devices capable of more than one speed (10/100 switches, and network cards, for instance) should come on at the highest rate supported by the device (It's talking to itself, after all). If the device is capable of noticing low-quality cabling, it may negociate a lower speed - this is a clue that there is something wrong with the cable.

Most dedicated cable testers have a loopback mode, where instead of being attached to both ends of the cable, the tester is attached to one end, and a loopback plugged into the other. Many network cards can go into loopback testing mode, where they either test their own hardware (with a loopback plugged straight into the card), or act as a rudimentary cable tester (with a loopback on the other end of the cable). While a network card can't tell you what's wrong with a cable (for instance it may be wired incorrectly, or one of the wires may be damaged), but it can tell you whether the cable is working or not - the results of the loopback test should be near perfect. If more than 1% of the packets are being corrupted, there's something wrong with the cable.

As well as testing single cables, loopbacks are very useful in running several cables at once to a patch panel - wire all the wall sockets, put a colour-coded loopback in each one, and then after all the wires have been run to the patch panel (and tested using a loopback cable tester), connect them to a hub. 'phone your accomplice, and have them remove loopbacks one at a time.

1 - When telephone companies are testing local loops, they use a similar device - it connects a resistor across each pair, so they can tell from the exchange if the pair is working, open, or shorted. They usually instruct customers who have had their loop tested in this manner to throw the loopback in the bin - it costs much less to buy a new one than to send someone to pick it up.

2 - Less common are fully-connected loopbacks. These test all four wire pairs, and can test if the link is wired correctly for ethernet and voice, or gigabit ethernet. In addition to the megabit ethernet pairs, these also connect pin 4 with pin 7, and pin 5 with pin 8.

Log in or register to write something here or to contact authors.