T568A or T568B

Generally true. [...]

I imagine in some corner cases this is correct.

Yes, but the twists per foot are different for the different colors. If each pair had the same twists per foot, you would still get cross-talk between the pairs, where the differing twists per foot provide a level of immunity.

Like 1000BaseTX, which already does this...

On Sat, 12 May 2007 18:04:09 -0500 Daniel J McDonald wrote: | In article , wrote: |>Some information sources indicate there is no difference between T568A and |>T568B wiring besides what color is used at which pin position. | | Generally true. | [...] |>But some other information sources suggest that using T568A when T568B is |>in standard use, or the reverse, that the electrical quality is slightly |>degraded even though things generally work. | | I imagine in some corner cases this is correct. | |> If this information is true, |>then that would suggest some kind of _electrical_ or _metallic_ difference |>in the wires, and not just the insulation color. As far as I am aware, the |>additives to insulation to create the color has no effect. |>

|>Could the wires in a CAT5/CAT5e/CAT6 cable be _electrically_ different in |>some way, such as a different impedance (I thought it was all supposed to |>be 100 ohms within a twisted pair, or a different cross sectional area 9and |>thus affect the resistive loss)? | | Yes, but the twists per foot are different for the different colors. If | each pair had the same twists per foot, you would still get cross-talk | between the pairs, where the differing twists per foot provide a level of | immunity.

Ah. Is there any short info on exactly how many twists per foot per color, or do I have to go download that hundreds pages document if I can even find it?

|>I still look forward to the day that computer interconnections begin using |>symmetric bidirectional cabling so we can stop having to deal with various |>wiring schemes and get rid of crossover cables and such. | | Like 1000BaseTX, which already does this...

There is a different crossover wiring for this, too. So no, it isn't really that way. Of course some devices have been made to detect which lines are wired which way, so you could communicate fully regardless of which cable type is connected. But this (Cat5/6 with 8P8C on each end) is nowhere near being a symmetric bidirectional cable/connector scheme.

snipped-for-privacy@ipal.net wrote in part:

I have never heard this. The reasoning for -A over -B in govt work and residential (T-570) is that pair 2 (orange/wt) is in the USOC position for POTS (and other phone stds).

-B is an AT&T legacy. Legacy lives!

I very much doubt it could affect -B vs -A. Both use pairs

2 & 3 (orange & green/wt) for 100baseTX, and all four pairs for 1000baseT (GigE).

Very different question. I've heard some early Cat5 was rather slack (untwisted) on pair 4 (brown/wt) since only pairs 2&3 were expected to be heavily used (100baseTX). A cable with one slack pair will meet spec for XT and actually makes the rest of the cable easier (less problems with twist slew). It will have problems with alien XT (from other cables). It will also have problems with skew -- that pair is shorter and signals will arrive sooner. AFAIR, Cat5e doesn't have a slew spec, Cat6 does.

Rejoice, for that day is nigh! GigE does autodetect, and some more recent 100baseTX equipment do MDI/MDI-X autodetect.

-- Robert

What is "electrical quality"?

snipped-for-privacy@ipal.net wrote: (snip)

The different twist pitch slightly changes the impedance, but the difference should be well within the allowed tolerance. Unless you have cable with the two wires of a pair bonded together it is likely less than the variation along the pair.

I believe this is true for gigabit, which uses all four pairs in both directions. You do still have to maintain pairing, which is even more important at gigabit frequencies. As someone else noted, the reason for 568A vs. B is to agree, for two pairs, with the standard used for voice lines. The third and fourth pairs for voice lines are (2,7) and (1,8), and so don't agree in any case. The connector impedance and crosstalk characteristics aren't good enough in that configuration for ethernet. (Well, it would likely work with 10baseT which is somewhat less sensitive to such changes, and probably still with

100baseTX.)

-- glen

snipped-for-privacy@ipal.net wrote: (snip, someone wrote)

I don't know if it is in the standard, or just the allowable crosstalk at different frequencies. Note, though, that as someone else said the difference between 568A and 568B is the exchange of the green and orange pairs, such that both are still being used. If there were any differences, they would cancel out in a round trip configuration.

-- glen