Really simple DIY LAN tester for long distances

[vanarebane]

Hi all!
So I had a problem with a LAN cable today that run through a floor and did not work. It seemed impossible to test, what's wrong so I had a spark idea on how to make my own LAN tester that can be used over long distances and that is really easy to put together.

This project is really too simple to post here among your great minds perhaps, but I thought to show it off anyway and to get some feedback on my tutorial, of what should I do different and what I do good in it.
You can checkit out here: http://groundify.com/diy-lan-tester/

I'm slowly adding projects and ideas to my blog so there is not much there yet. There are ideas waiting to be published although

Thank you for your time!

[Monkeh]

This is basically what all the cheap cable testers do.

Don't know why you're prattling on about crossover cables, though. They are not required unless you're connecting two old NICs or switches directly. Normal cables are A to A or B to B, and modern switches (anything made since, oh, the mid 90s, anything gigabit) and NICs (anything gigabit, and most recent 10/100) do not give a crap if they're straight or crossover, no matter what's on the other end.

[128er]

If you have nothing to test the cable - yes, thats a good McGyver solution . But after that, you should build a more automated tester or at least buy one for a few dollars.

In the pictures below you can see one, that i have build many years ago in a course for my vocational training. Dirt cheap 

HCF4001 as clock source and CD4017 counter to drive every of the eight wires (+ shielding). The other device is the receiver. One LED after the other will light up. If the wires are reversed, the LEDs will light up in the wrong order. So you can see which wires are wrong. The frequency is adjustable with a potentiometer.

[vanarebane]

Don't know why you're prattling on about crossover cables, though. They are not required unless you're connecting two old NICs or switches directly. Normal cables are A to A or B to B, and modern switches (anything made since, oh, the mid 90s, anything gigabit) and NICs (anything gigabit, and most recent 10/100) do not give a crap if they're straight or crossover, no matter what's on the other end.

Well I did not know that. I guess I learned that they cause problems when they did cause problems and did not learn that there is intelligence built to newer devices Learning something every day. Do you have any source on you claim?

If you have nothing to test the cable - yes, thats a good McGyver solution . But after that, you should build a more automated tester or at least buy one for a few dollars.

In the pictures below you can see one, that i have build many years ago in a course for my vocational training. Dirt cheap 

HCF4001 as clock source and CD4017 counter to drive every of the eight wires (+ shielding). The other device is the receiver. One LED after the other will light up. If the wires are reversed, the LEDs will light up in the wrong order. So you can see which wires are wrong. The frequency is adjustable with a potentiometer.

Nice solutions I before though that these testers were wireless or how they test and now it seems that they are really easy. Thanks for compliments, I was happy with the easy result and I guess I won't be buying LAN tester though as it's then again another tool in drawer that I rarely use

I guess that high-end testers can test the speed and noise and errors of the cable? So when you are in need of high-performance cables installed and need them top shape they test if the cables perform well and if there is any problems with it.

[Monkeh]

Don't know why you're prattling on about crossover cables, though. They are not required unless you're connecting two old NICs or switches directly. Normal cables are A to A or B to B, and modern switches (anything made since, oh, the mid 90s, anything gigabit) and NICs (anything gigabit, and most recent 10/100) do not give a crap if they're straight or crossover, no matter what's on the other end.

Well I did not know that. I guess I learned that they cause problems when they did cause problems and did not learn that there is intelligence built to newer devices Learning something every day. Do you have any source on you claim?

http://en.wikipedia.org/wiki/Medium_Dependent_Interface#Auto_MDI-X

[David_AVD]

A good tester will check for correct pinout as well as split pairs (Google it :-) ), the latter being impossible to detect with very simple testers.  Split pairs result in cables that work to a degree but introduce weird issues increasing to complete usability over longer distances.

[jimjam]

A good tester will check for correct pinout as well as split pairs (Google it :-) ), the latter being impossible to detect with very simple testers.  Split pairs result in cables that work to a degree but introduce weird issues increasing to complete usability over longer distances.

know any project / design for a good tester that can detect this? I was wondering about how to achieve this. Also would be handy to detect the terminator "number" so we can have multiple terminators plugged in on the other end e.g. on a patch panel.

[128er]

 I think detecting split pairs is not so easy. Maybe proper networktester do that with intepreting the crosstalk attenuation. If the signal send over one pair interferes massively on another pair its obvious that the wires are messed up. They can also show you on which end the problem appears (NEXT, FEXT).

[David_AVD]

I have a "Test-Um" branded tester that detects split pairs.  I think I mentioned the exact model etc in another thread about a year ago?  It was around $150 from memory.

[jimjam]

I'd also like to know how to detect the distance to the break in the cable.... like what those fancy testers can do.

[AndyC_772]

It's done by time domain reflectometry. Send a high speed edge up the cable, detect the reflected wave that comes back from the impedance discontinuity at the break, and measure the time interval between the two. Divide the time by the speed of wave propagation in the cable and you have a distance.

[cs.dk]

I have had the LanCaster Pro. It'll detect split pairs, etc. - It shows at which end is faulty crimped, and can show how many meters to a broken wire. Also measure cablelenght.

I can recommend it. Note that there is a new version now. http://www.betterbox.co.uk/products/bi-structured-cabling-testers/lancaster.html

[jimjam]

It's done by time domain reflectometry. Send a high speed edge up the cable, detect the reflected wave that comes back from the impedance discontinuity at the break, and measure the time interval between the two. Divide the time by the speed of wave propagation in the cable and you have a distance.

Know where I can learn a bit more about this?

I have had the LanCaster Pro. It'll detect split pairs, etc. - It shows at which end is faulty crimped, and can show how many meters to a broken wire. Also measure cablelenght.

I can recommend it. Note that there is a new version now. http://www.betterbox.co.uk/products/bi-structured-cabling-testers/lancaster.html

I understand that this can be bought - Fluke makes something similar too. I'd like to know how it's done to try and make something that does the same thing - primarily for fun.

[128er]

It's done by time domain reflectometry. Send a high speed edge up the cable, detect the reflected wave that comes back from the impedance discontinuity at the break, and measure the time interval between the two. Divide the time by the speed of wave propagation in the cable and you have a distance.

Know where I can learn a bit more about this?

[AmmoJammo]

$AU2.78
www.ebay.com.au/itm/301221618632

Yeah, I ordered one just for a laugh :p