Finding break in wire

[phoenox]

I have an old network cable with a broken wire somewhere along its length(50 feet).
I was thinking it just might be possible to find where the break is by moving a small device that will generate noise (transmitter) along the length of the cable, then moniter the end of the cable with something that will see the noise generated(reciever). When the transmitter moves past the break in the wire, the noise will no longer be seen by the reciever.

For a transmitter I am thinking of sending a high frequency signal through a spool of wire and passing the network cable through the centre of the spool.

The reciever might be more difficult. The first thing that came to mind was to connect an oscilloscope probe to the wire and refence to something stable. Or maybe I need to build something like a radio antena reciever circuit. But I have no idea what is actually involved in that.

What do you think?

 

[Mike_2545]

It may be cheaper to buy a new cable.

 

[arhi]

Most new ethernet cards offer tool to find where the ethernet card is cut. Donno how it works but the app will tell you how far along the cable you have a break with few cm precision.

I was looking for the circuit to do that with coax cable (I have some coax cable going trough walls, not easy to replace) but was unsuccessful.

 

[user_88]

Take a look at this .....
Time-domain reflectometer - Wikipedia, the free encyclopedia

.... not sure if you need a commercial device, or maybe can put together something that will just be adequate ...... basically just sending a pulse signal down the wire and measuring the time for the reflected signal to return.

 

[arhi]

interesting link, thanks. I'm sure that I do not want to give 200+ usd for the commercial device and I'm not sure I can whip up something with enough accuracy to be of any use so I gave up long time ago (I have few coax 75R cables in the walls that are broken somewhere, the imbecile who put them into wall did not use canal so they are unmovable... I'd need very precise device to tell me "where to break the wall in order to find a problem and fix it") ... it's cable TV antenna cable and I do not watch TV (did not have one until few weeks ago, but my father in law died and wife did not wanted to throw his tv to garbage so now I do have one - not connected thoe)...

All in all, I was just interested in how the TDR generally works as afaik signal travel at speed of light, measuring "break point" with few cm precision (that's what my eth card can do for e.g.) seams just impossible with hardware I can see on the eth card .... c = 299792458 m/s so 10cm is 1/2997924580 sec 333.6 picoseconds, you need few GHz to measure travel time with 10cm resolution ?!?!?

Now, I'll read the wiki article there might be something interesting there to explain it

 

[arhi]

hm .. not getting the idea after reading the article .. ok signal is not propagating at speed of light but **broken link removed** that's not a problem but then again, coax have VF between .60 and .80 and for VF .70 (middle) the propagation speed is 210,000,000 m/s or 4.76 ns to travel one meter. ( Velocity of propagation - Wikipedia, the free encyclopedia ) ... so, how the hack can a 5$ eth card measure this time with few cm precision (0.05ns ?!?!?)

 

[arhi]

I'm whipping up this later today (6am now)
TDR Circuit

if someone have "simpler" or "better" schematic - please post a link (I have some usb osc, 200 MS/s real-time sample rate, 25 MHz bandwidth - hope it will do)... will post results when I'm done

 

[mneary]

The more conventional way is to put the noise into the cable from one end and listen with a receiver as you walk along the cable. The cable will probably emit less noise where it is balanced (before the break) and more noise after the break where only one conductor has survived.

A TDR can be fun to use, but electricity doesn't travel in cables at the speed of light. It can be as slow as 75%, depending on the dielectric.

Another way to find a break in cable is to use a variable frequency oscillator and find the resonant frequency of the stub. You still have to deal with the velocity factor. Yet another way is to build a simple one-transistor oscillator, using the stub as the principal resonator. Then measure its oscillating frequency.

Edit: Simple TDR shown above is the best idea. I'm going to build one for myself in case I ever need it!

 

[user_88]

I have not tried this...So not sure what pulse width and frequency would actually be of practical use.
If you could in fact observe a reflected pulse of any sort, you could compare this time delay measurement with that of a short length of cut to length cable ... maybe 1 Meter long.

Guess that a 1 meter test length would be too short for TDR circuit shown above....
Specification says 5 meters is minimum cable length .....

 

[arhi]

donno what to say ... I'm still fairly confused but ordered few 74HC14 and 74HCT14 (no AC family in local store ) and will give it a go later today when the order comes. I just hope this osc will be fast enough to show the signal

 

[Leftyretro]

At our refinery we use to trace broken connections in instrumentation wiring that would run for hundreds and thousands of feet through intermediate junction boxes, etc. We would use ohm meters of course but where spare cable pairs were wrapped and difficult to attach to we would use a 'fox and hound' device. The sender device would be clipped to a lead somewhere and to ground. The inductive pick-up would make a tone when the signal was close to the wire, but no electrical connection was needed. It did take a little practice when working with very densely bundled cable but was a real time saver. Didn't help if the break was inside a conduit but such cases were rare.

Tone Generator, Triplett 3399 FOX 2 and HOUND 3 KIT

 

[crutschow]

A Google search for "Fox and Hound circuit" gave this do-it-yourself verison **broken link removed** which uses an AM radio as the hound.

 

[phoenox]

Quick and dirty fox and hound:

I attached the end of wire to a microcontroller pin outputting 300 ns pulses at 2500 Hz. Probably any square wave generator with fast transitions will work.

Use an osciloscope to measure the voltage across a .1 Ohm sense resistor. Put the wire beside the sense resistor. Set the osciloscope trigger level so it triggers when the wire is near the resistor but not when it is past. about 7 mV in my case. If you run the wire through a spool and conect the end of the wires through the sense resistor you get a stronger signal.

Then run the wire past the resistor untill the osciloscope stops triggering. This is where the break is.

I managed to find the problem in my network cable using this method. Two wires were was broken at the end in the plug. Now I just have to crimp a new end on and it will be good to go.

 

[Hero999]

I've found that connecting the wire to a live mains wire and holding a neon lamp right next to the wire works. Providing the insulation isn't too thick there's normally enough capacitive coupling to make the neon light dimly. Move the down the wire until it goes out or suddenly gets even dimmer.

Obviously be careful and only do this when the wire is not connected to anything. I don't know if it will work with the wimpy US voltage of 110V. Connecting a large metal plate to one lead of the neon and using it as an antenna and holding the other lead might help.

If you cant get it to work with the mains voltage then try a CFL inverter. I wouldn't use a TV flyback because the voltage is much too high and most modern flybacks produce DC as they have a built-in diode and multiplier.

Another option is to build a simple electric field detector with a JUGFET. I'll post a schematic if you like.