Wireless LAN cable tester

[Kamikaze47]

Years ago I made a LAN cable tester using PBP and its served me very well. It displays on a 2x16 LCD if the cable is standard or crossover, and lets you know if there are broken wires, crossed wires or shorts.

But the main drawback is I have to plug in both ends of the cable to the same device meaning its difficult to test long cable runs in buildings and such.

So I've been thinking about making a "wireless" LAN cable tester where you have 2 units - one to plug into each end of the cable. I thought one device could send different frequency square waves down each wire and the device at the other end could pick them up and work out all the required info to make a diagnosis.

The main snag is where to get common ground from? I cant use just one of the wires as ground because that assumes that that wire is always going to be OK and wired correctly.

Does anyone have any ideas? Better yet, does anyone know how the commercial LAN cable testers solve this?

[mark_s]

Maybe you could make the transmiter and receiver plug into the mains power outlet and utilize the neutral or ground to complete the circuit. They could also talk to each other using something like the X10 format over the same line? Interesting problem.

[rmteo]

Each Airbus A380 aircraft has 330 miles (1,742,400 feet) of cabling - the cabin wiring alone has 100,000 wires and 40,300 connectors. Wonder how they test them.

[Kamikaze47]

Thanks for the suggestions. I don't really want to use mains for ground - mainly because it will make the device much more bulky with a power cord and also less convenient because it would need a power outlet near by.

I'm sure there must be some ingenuous technique that the commercial LAN testers use.

[cncmachineguy]

I don't know much about them, but how about making one end a "one wire" device. then the smart end could send a OWOUT on each pin.

Or even easier, how about just a loop back plug for 1 end. then you could send a signal out each of four wires and expect it back on the coorsponding wire.

[mark_s]

Looks like they use RF.

http://www.google.com/patents/about?id=_yCzAAAAEBAJ&dq=patent:7479776&as_drrb_ap=q&as_minm_ap=0&as_miny_ap=&as_maxm_ap=0&as_maxy_ap=&as_drrb_is=q&as_minm_is=0&as_miny_is=&as_maxm_is=0&as_maxy_is=&num=10

[Kamikaze47]

I don't know much about them, but how about making one end a "one wire" device. then the smart end could send a OWOUT on each pin.

Or even easier, how about just a loop back plug for 1 end. then you could send a signal out each of four wires and expect it back on the coorsponding wire.

Unfortunately, 1-Wire still needs a common ground.

The loop back plug is a good idea. The only thing I can think of that it wouldn't be able to test is if there happens to be a short across one of the pairs. I don't that's going to happen very often however. If I cant think of anything better, I think I'll probably use that method.

[cncmachineguy]

Maybe the loopback plug could employ some sort of passive delay, so you could measure the time between sending a signal and getting it back. That way if there is a short you will see it as instant response.

[dhouston]

The loop back plug is a good idea. The only thing I can think of that it wouldn't be able to test is if there happens to be a short across one of the pairs.Why not? Four reed relays and a PIC in the remote device would allow you to either short or open the pairs.

This URL gives you a better method of checking for a shorted pair...http://www.epanorama.net/circuits/tdr.html