1

I know of the regular crossover/straight-through assembly. But assuming all the cables are rearranged in order that the cables end up at the same pins but are of different color, what happens?

Here's my problem: I have self-assembled Cat5e cables wired in a house and the guy who assembled thought it wouldn't matter as long as the cable have the same coloring on each side of the cable. After thinking for a bit I am still wondering why the cables max out at around 90 Mbit/s.


thanks for the responses! I compared a few of the cables, most of the are arranged as followed: orange striped, orange, blue striped, blue, green striped, green, brown striped, brown so the only difference to the T568B-standard I see are that the solid green and the solid blue are switched.

  • 2
    How are you measuring that 90Mbps, and what speed are you expecting to see? If you only have 100MBps equipment then a sustained 90Mbps is not unreasonable... – brhans Apr 6 '16 at 14:59
  • I have a 250Mbps connection and when connecting with a new cable I get the full speed – keevw Apr 6 '16 at 15:05
  • 1
    Responding to your edt: Yes, that's exactly the sort of thing that "breaks" the pairing inside the cable. You have one differential signal being driven onto the blue-stripe and green wires, and a different signal being driven onto the blue and green-stripe wires. But the blue and blue-stripe wires are twisted together and the green and green-stripe wires are separately twisted together. – Dave Tweed Apr 6 '16 at 15:06
  • 3
    @keevw It's worth pointing out that reterminating Cat5 is pretty darn easy. Get yourself an RJ45 crimper and a cheap punchdown tool, plus some cable ends, and you can reterminate to your hearts desire. Or, you could call the installer and tell him he did a rubbish job and he needs to fix it. – uint128_t Apr 6 '16 at 15:08
  • I used to do cables myself but I always sticked to the standards, My main question is now: if I fix the cables, will i get the performance boost or is it possible that bad cat5e cables were used and I wont get a better performance at all? – keevw Apr 6 '16 at 18:21
3

The problem with arbitrary wiring is that the wires in the cable are organized as twisted pairs. The intent is that one signal is driven differentially onto the two wires of a given pair, and most of the bad electromagnetic effects will cancel out.

Each pair is deisgned to have a specific characteristic impedance, and the twisting of different pairs is deliberately made different in order to minimize crosstalk between pairs.

With arbitrary wiring, this pairing is no longer happening, and this creates a lot of crosstalk among the signals, as well as making each signal more susceptible to outside interference. The uncontrolled impedance also causes each signal to interfere with itself, limiting the bandwidth of the cable.

  • The question noted "in order that the cables end up at the same pins". The wires are twisted all down the line of the cable. One quarter of a twist less will not have any impact; that would be equivalent to if the cable were cut a little bit smaller. Cat5e can support up to gigabit, so a color mismatch is unlikely to cause such a significant degrade. It is far more likely that sloppy quality (a loose connection within the cable end) would cause this, than that the colors would have such an impact. – TOOGAM Apr 6 '16 at 14:47
  • 1
    @TOOGAM But if twisted pairs are split between electrical diff pairs, there will be a significant performance issue. (Consider the difference in impedance between a twisted pair and two halves of two different twisted pairs). That is what Dave is talking about. – uint128_t Apr 6 '16 at 14:49
  • 2
    @TOOGAM: Ethernet cable contains 8 conductors, which are organized as four twisted pairs. The pairs are twisted individually inside the jacket as I described -- you don't just have 8 wires all twisted together. Maintaining the pairing is CRUCIAL. It's true that the color of the insulation has no direct effect on the signal but it DOES allow the pairs to be correctly identified. – Dave Tweed Apr 6 '16 at 14:49
  • 1
    Okay, I confess, maintaining the pairing matters. Mixing up the green-white and the orange-white, while making the green and orange correct, could have the effect you state. Still, if the blue pair went to where the orange pair goes, and the orange pair went to where the blue pair goes, I wouldn't expect such an effect. So I would say some differences can matter, but this doesn't mean that the cable is doomed just because the colors don't completely match the official T568A diagram. Some instances would have the effect you describe, and others would not. A mismatch doesn't guarantee a problem – TOOGAM Apr 6 '16 at 14:56
-4

The colors have no effect on the electrical signals. They are only meant to help human installers. The cable will effectively communicate the required electrical signals. The blind (literally: no optical sight) computers will not notice any difference. The colors are from the plastic insulation over individual wires within the cable, and don't affect how the cables operate.

If any experienced technician notices the cable, they will be prone to consider the cable questionable. They will be more prone to do unnecessary testing of the cable, or (probably much more likely in a well-financed commercial environment) to throw it away and replace it with something that feels more trustworthy.

  • 2
    You missed the fact that the OP is actually measuring a drop in performance of the cable, which I have explained in my answer. – Dave Tweed Apr 6 '16 at 14:43
  • I think you should edit your answer, adding that colors don't matter as long as twisted pairs are matched. – Dmitry Grigoryev Apr 6 '16 at 15:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.