2

Background: the Uverse guy gave me a big spool of cat5e "in case I needed it later" which was great. Here is the setup:

My uverse router is in my house. One cable goes from it down into the crawlspace, into a 10/100 switch which feeds various ethernet ports in the rooms in my house.

One cable however, runs out to the shed behind my house (underground cat5e) and for the last few years I've been happy to plug into the switch in the shed (office) when I've been working.

I also, occasionally would bring the 2nd Uverse box out there to the shed, and plug it into the switch in the shed. Worked fine; I could watch live TV, etc. all over the ethernet. (I wish I could find out how much bandwidth that device actually needs).

So to summarize:

Uverse in house House---------> Shed # works fine. switch in shed

BUT my brother lives next door and he wondered if every now and then, on a big game day, he could borrow the 2nd uverse box to watch football in his house. He lives right next door, our houses almost touch. The most unobtrusive path to his house was from the shed, so I used the cat5e cable left by the uverse guy to make a run from the switch in my shed to his living room.

using the primitive cable tester I have (that simply lights up to indicate each pair is functioning) it seems fine. But I can't acquire an ip when I plug the run into my macbook. Even when I hard assign an ip that should work within the uverse subnet, it doesn't work.

If I walk back to the shed and plug into the switch with a short station cable right at the desk, it works fine. acquires an IP from the uverse router, works great.

I'm mystified that the cable tester was green, but nothing is working. What am I doing wrong?

What can I do to narrow down the problem? I haven't gotten into the nitty gritty of this stuff for a long time.

Phil
  • 21
  • Can you check if it works if you use an extra cable and a switch (or a router)? If it does then you might need to check if it is a straight though cable or a cross cable and which of the NICs support MDI/MDX and which do not. – Hennes Sep 11 '13 at 01:37
  • I have to agree with @Hennes it appears your trying to use a T568A to connect same layer devices. By the sounds of your cable tester it can't tell the difference between T568A and T568B so it isn't identifying the fault. – 50-3 Sep 11 '13 at 01:43
  • well it is a straight thru cable, just like all the ones I usually make. My only way to test whether it is "good" is with the cable tester that tests all four pairs. That cable tester shows green on all four pairs. – Phil Sep 11 '13 at 01:43
  • Ok I googled the T568 standard and now I understand a little. I wired both ends according to T568B I'm pretty sure. I think what I should do is strip the ends off and do it again, very carefully. Is there any other way to test the cable besides "seeing if it works" it is driving me nuts. – Phil Sep 11 '13 at 01:46
  • If you have a multiple meter you can just test if current can pass from one end to the other - T568A is easy pin 1 to pin 1 through pin8 to pin 8 for T568A to T568B (Crossover) it's a bit harder: 1-3 2-6 3-1 4-4 5-5 6-2 7-7 8-8 – 50-3 Sep 11 '13 at 01:57
  • Have you tried the cable inside your house, just connecting between the house switch and a computer? – Daniel R Hicks Sep 11 '13 at 02:36
  • 1
    It doesn't make any difference whether the ends are wired T568A or B, so long as both ends are wired the same. It does make a difference whether the ends are wired T568A or B vs some scheme that does not assign pairs of wires to the appropriate pairs of pins as shown in the A or B diagrams. Ie, pairs of wires must be assigned to 1-2, 3-6, 4-5, 7-8, with colors matching (not mirror image) on each end. – Daniel R Hicks Sep 11 '13 at 02:39
  • (It should be noted that the max total distance between active repeaters in a Cat5 line is 100 meters.) – Daniel R Hicks Sep 11 '13 at 02:45
  • 1
    "Is there any other way to test..." -- In-circuit is always the final, authoritative test. How good is your "cable tester"? Maybe it only does DC continuity like a multi-meter would and also checks for shorts? Can it do full spectrum tests and check for cross-talk? IIRC a sophisticated tool for checking cables is some kind of reflectometer (stuff that telco guys would have for buried or strung lines). – sawdust Sep 11 '13 at 09:21

1 Answers1

0

Many factors can produce this problem.

  1. 150 foot from the switch to your brother's home?

  2. 100 Mbit Ethernet supports only 300 foot. Apart from that, if the cable has mechanical issues, it might not work any more. Strong mechanical load on the cable during installation can influence the electrical data of the cable. Simple testers might not see this because they only test with DC voltage, but Ethernet is a complex modulation.

  3. If you connect both ends of the cable to a Ethernet device, do you get a link (green LED)? If yes, then the cable is OK. Then the issue is the port of the switch.

  4. Sometimes cheap switches do not support such long cables, or this port is damaged.

Daniel Andersson
  • 24,335
user266169
  • 1