Problem: UTP only works at 10 MHz

Claudio Bogado Pompa wrote in part:

Which termination pattern did you use? T-568-A or T-568-B? A split pair will sometimes work at 10 but almost always fail at 100.

-- Robert

Why would A vs B make any difference? Neither one splits a pair. Standard cross-over cables, for example, are simply wired to the T-568-A spec on one end and the T-568-B on the other.

A more likely explanation is that he is running on old telephone plant quad pair style cable which will work reasonably well for short and medium length runs at 10Mbps but not at all at 100Mbps.

Cat5 cable, which is required for 100Mbps operation, has very specific specs for the twist on each pair. The no. of twists per unit length is different from pair to pair in order to minimize both near end and far end cross-talk. See: pp 278-280 ISO/IEC 8802-3: 1996(E), the ISO std. for 10Base10.

It was probably a subtle suggestion that if the original poster could answer the question, he would have wired it to one standard or the other.

I'd like to believe that UTP means Category 5+ cable. Again, giving the OP the benefit of the doubt.

I'm guessing split pair.

Carl Navarro

Correct. Of course either will work, but if the OP doesn't know about either, it is extremely likely the OP split a pair. The correct patterns are non-intuitive.

I read the same thing.

Likewise.

-- Robert

Improperly made connectors, like not having a good crimp, or even split pairs can cause this problem. Also plant wiring may not be capable of the speed, depending on the age of the plant wiring, like if there is any cat3 being used. Also check for any patch panels or other termination points that may have degraded or become defective.

>

messagenews: snipped-for-privacy@w3g2000hsg.googlegroups.com...

Thanks to all your answers. I bought a very cheap crimper so that may be the problem. Is there a minimum or maximum resistance that I should check between both ends to ensure the cable will work fine?

On testing the cable for networking applications, since more is involved than just DC continuity, using an ohm meter will not provide a very good indication of your cable plant. Use a dedicated CAT 5 cable test set, or a lan network test set. You can get one for not too bad of a price. While using an ohm meter may indicate that there is DC continuity, it does not test far end and near end cross talk and other cable parameters needed to ensure proper operation of the cable

>

Claudio Bogado Pompa wrote in part:

It most likely is. Crimpers are hard to use correctly. How are you arranging the wires?

There are 40,320 ways of wiring a 4pr cable "straight-thru". All but 1,152 resulting in splitting a pair needed by 10baseT and 100baseTX. Electrons may be color-blind, but they know who their dance [twist] partners are.

It depends on cable length. I measure very close to the theoretical/table values.

I used same order of colors in both ends, brown, white and brown, green, white and green, blue, white and blue, orange and white and orange.

Claudio Bogado Pompa wrote in part:

All pairs side-by-side. While intuitive, this is incorrect for 10/100baseTX because it splits a pair. T-568-B is:

(viewed tab down, cable from right)

8 - brown 7 - wh/br 6 - green 5 - wh/blue ! 4 - blue 3 - wh/green ! 2 - orange 1 - wh/or Note the blue pair is perfectly in the middle and the green pair straddles it. You can swap colors if you like and it will still work. But you cannot move pairing positions. Ethernet relies on pins 1&2 to be on a single pair and pins 3&6 to be on another for noise cancellation.

This counter-intuitive standard evolved for phone compatibility.

-- Robert

Check this against the standard wiring, you will see this is not correct. Correct the pin out and you should be ok.

>

Ya did good and we now know why it doesn't work, which was your first answer.

The cable is about 50 ohms per 1000 feet, and all the meters in the world aren't going to help you

Carl