Finding a break in an underground wire

[scott1554]

I have an underground electronic dog fence. The line has been cut, but I can't find the cut. From other research I have:
-wrapped the wire leads from the 1,600 feet of wire to a RF choke lead
(Radio Shack)
-plugged the RF choke lead wires into the transmitter
-Tuned an AM radio to various frequencies looking for static (recommended 600 or 1600 mhz
- i was told i should hear a constant or slightly pulsating signal tone from the radio (i assume the signal tone i should be hearing is static?)
- i should walk the line until the tone disappears, i have found my spot (but all i continue to hear is static)

Am I doing something wrong or is there a better way? Some suggest digging up the line half way, but if I could find the line at all, I could solve my problem by tracing it back.

thanks

 

[Leftyretro]

I have an underground electronic dog fence. The line has been cut, but I can't find the cut. From other research I have:
-wrapped the wire leads from the 1,600 feet of wire to a RF choke lead
(Radio Shack)
-plugged the RF choke lead wires into the transmitter
-Tuned an AM radio to various frequencies looking for static (recommended 600 or 1600 mhz
- i was told i should hear a constant or slightly pulsating signal tone from the radio (i assume the signal tone i should be hearing is static?)
- i should walk the line until the tone disappears, i have found my spot (but all i continue to hear is static)

Am I doing something wrong or is there a better way? Some suggest digging up the line half way, but if I could find the line at all, I could solve my problem by tracing it back.

thanks

Well the pros would use a TDR, time domain reflectometer. Sends a short pulse and then measures the reflected wave returned by the break. The principle is simple but not sure I've ever seen a DIY version, but would be a cool project...

 

[Sceadwian]

TDR's aren't practical for the hobbyist. You'd need really high clock rates to get any kind of resolution, and he says he doesn't know where the wire is located exactly which means even a precise distance to the break isn't going to do you much good. Just as an example, in order to get a resolution of 1 foot you'd need a clock rate of 1ghz.

 

[billericayboy]

When the wire for my mum's mowbot broke i kept it simple by locating bits of the wire and then circuit testing either side of where i inserted a pin into the cable. It only took 5 minutes to locate the break and it was so simple. All i needed was long wire, 2 croc clips, a pin and a multimeter.

KISS

 

[mneary]

Any progress?

 

[Pommie]

Can't you just put the dogs collar on your arm and walk along it.

Mike.

 

[ljcox]

EDIT. I've deleted the previous Wheatstone Bridge idea as I think it will be impractical.

Are there 2 wires undergraound?

If so, you could measure the capacitance between them at each end. The ratio of the capacitances would be roughly equal to the ratio of the distances to the break.

If there is only one wire, you could measure the capacitances to earth. But use a well watered earth stakes at each end.

Another trick that was used by someone I knew who found faults in underground telephone cables was as follows.

As an alternative to the above, you could use a DC voltage and a multimeter rather than the earphones.

If the wire is leaky to earth (due to water ingress), he connected an AC voltage to the wire with the other side of the voltage source connected to earth. This creates a current through the wire and back via the earth.

He had two earth stakes with wires attached to earphones. He drove these into the ground about 3 metres apart near the underground cable and used the earphones to hear the tone due to the voltage drop between the stakes.

This is the voltage drop due to the return current.

He started near the voltage source. He then extracted the stakes and walked abour 3 metreto the mid point of the faulty section of cable, drove the stakes into the ground. Then measured listened again. If he could hear the tone, he walked to the 3/4 point and repeated the process. If he could not hear rthe tone (ie. at the half way point) he walked back to the 1/4 point repeated the process.

He kept doing this (ie. dividing the cable span in half each time) until he found the approximate point where the break was.

This needs conductive ground. So if you are in an arid area, it would be best done after rain.

 

[philba]

I have found a break in burried wire before using capacitance. But there is a simpler way.

Take 1 ft of the wire and measure the capacitance. compensate for the capacitance of the probes and leads. that should tell you the capacitance per foot. Then measure the capacitance of the broken wire. do the math and that should be the distance to the break within a few percent. In my case, it was a sprinkler wire and the break was at about 140 ft. iirc, I calculated the break at about 135 ft. It took about 10 mins to find the break. I didn't know about where the wire ran and it wasn't burried very deep.

Of course, if the wire is shorted, capacitance is out but you might be able to measure resistance.

 

[RODALCO]

If this is a single core wire with possible damaged insulation or a break in the ground we use the pool of potential method in the power industry.

Basically you pulse dc at an unequal ratio say (4OFF:1ON) at about 50 mA's, from a short circuit proof HV supply into the faulty core.
The positive lead goes to the faulty core, the negative goes to an independant earth spike opposite from the direction you are searching from the fault.
With a midpoint microammeter with balance you can locate an earth fault within less than 5 cm of the fault by looking at the deflection of the needle of the meter.

You can try to use a dc DMM instead of the µA meter and look for the pulse.

Good luck
Raymond

 

[ljcox]

This sounds good but there are 3 points I don't understand.

1. What is "a midpoint microammeter with balance" ? Do you mean a centre zero meter?

2. To what is it connected?

3. Why does the source have to be pulsed?

You wrote HV. What voltage does it require?

 

[RODALCO]

1. Yep a centre zero meter, I have a look if i can locate the circuit somewhere and will post it when found
So you can cancel out stray currents which are always in the ground.
AC is of no use because of return currents with the M.E.N. system used in New Zealand.

2. We use 2 spikes to probe in the ground, say 2 metres apart and walk the direction of the cable.
Where the pulsing DC leaks out where the fault is, the microammeter will pick that up by means of the deflection of the needle it is pulsing, and so a very accurate locate can be done.

3. Reason as above. If continuous dc you are unable to pick up the readings.

In our POCO we use a 240 to 4200 Volts supply step up transformator which is controlled via a variac, then a voltage doubler with smoothing capacitor which will get voltage up to 10 kV dc for burning in a high resistance fault.
Then flick it over to pulsing DC via a on off timer and you are in business.

You probably get away with lower test voltages but I use this set up to pre locate 11 kV and 33 kV faults as well.

 

[ljcox]

1. Yep a centre zero meter, I have a look if i can locate the circuit somewhere and will post it when found
So you can cancel out stray currents which are always in the ground.
AC is of no use because of return currents with the M.E.N. system used in New Zealand. I thought it may be because of stray currents. But if his cable is remote from power sources, eg. on a farm, would there be any stray currents?

2. We use 2 spikes to probe in the ground, say 2 metres apart and walk the direction of the cable.
Where the pulsing DC leaks out where the fault is, the microammeter will pick that up by means of the deflection of the needle it is pulsing, and so a very accurate locate can be done. Can you only read the current near the fault? Or is it readable all along the cable until you pass the fault?

3. Reason as above. If continuous dc you are unable to pick up the readings.
But if remote as I wrote above, would DC work?
In our POCO we use a 240 to 4200 Volts supply step up transformator which is controlled via a variac, then a voltage doubler with smoothing capacitor which will get voltage up to 10 kV dc for burning in a high resistance fault.
Then flick it over to pulsing DC via a on off timer and you are in business.

240+V sounds dangerous. I hope he does not electrocute himself.

You probably get away with lower test voltages but I use this set up to pre locate 11 kV and 33 kV faults as well.

Scott1554, if you don't have a high voltage supply, you could make one using a transformer from an old TV or radio (these days they use switching mode power supplies)

Use a rectifier bridge to make it DC and (if necessary - depending upon the answers to my questions in red above) a simple pulsing circuit. A series resistor will be necessary to limit the current to about 50 mA.

I estimate that the power dissipated in the resistor will be about 3 Watt. You need about 6k, so four 1 Watt 1k5 resistors in series would suffice.

 

[RODALCO]

As long you know what you are doing and warn other people off, guard equipment during testing it is a safe and accurate way of finding ground faults.

If you are very remote where your fence wire fault is, and well away of ac mains cables, you can try other options and may get away with continuous dc.
I still would prefer to use a pulsing option.

In testing for the POCO i have been on 3 km long cables where you get no readings at all, unless you are within 50 metres or so off the fault area, where the readings on the µA meter increase.

To check if your system works.

1. Be in a lawn or bare ground where you are sure there are no powercables nearby.

2. Set up your test gear.

3. Put your (-) earth spike the opposite way from where you are fault locating.

4. connect your cable under test to your (+) connection of your tester.

This cable under test can be a 10 metres piece of Romex or single core where you stick the end of in the ground to resemble an earth fault.

5. Warn off other bystanders and treat all as alive.

6. Do your fault location and you will pick up good readings near the end of your cable under test.

Good luck and play it safe.

 

[ljcox]

If you can only detect the current in the ground within 50 metres of the fault, he will have to test about every 25 metre to be sure of detecting the fault.

1000 ft is about 300 metre, so he should have to make no more than 12 tests.

Of course, this all assumes that the fault is an earth fault. But if it is not leaky to earth, then the only option is the capacitance measurement - given that pulse echo testing is not feasible in this case.

 

[RODALCO]

Don't try to reinvent the wheel ljcox

This method is one of many methods used to locate cable faults accurately
I do this as part of my job and my hit rate is 99% !
The POPIE method has been around for 50 years or so.

Regardless of all fancy gizzmo's you can clip on a cable and tell you distrances in cm's, quite often they are only a approx guide.

You explain to the digging crews, o sorry, i was wrong, dig here. and again etc.

Raymond

 

[scott1554]

To all, your knowledge is impressive but is wasted on my limited electroinc experience. The wire in question is 18 gauge, and is twisted coming off the device to cancel the signal going to the dogs collar. Once I am at the beginning of my rectangle the the wires split and cover 1 acre of ground. I've dug approximately where I think the break is, but can't visually find the wire. Thus my interest in finding the break... using limited resources. Thanks, Scott.

 

[ljcox]

Do you have a multimeter?

If so, we can guide you through the construction of a simple power supply and all you need is the multimeter and 2 earth stakes to detect the current near the break. Rodalco has the experience in actually doing this so he can guide you with the technique.

Is the ground damp or dry?

 

[ljcox]

Scott,
Do you know what frequency the dog's collar transmits?
You said the radio you used was 600 to 1600 MHZ (I assume you meant kHz since the MW band is between 600 and 1600 kHz). I have a portable radio that has LW, MW & SW bands.

If you knew the frequency, you could tune a radio to that frequency before you start walking.

Do you have a circuit diagram of the collar transmitter?

Do you know anyone that has electronics experience? A friend, radio ham, TV repairman, teacher at a college who teaches electronics, etc.?

You need to be able to measure the frequency emitted by the dog's collar unit so you can tune a radio to it. Then ask him how to connect the collar unit to one end of the wire so you can trace the wire and find the break.

If you can't find anyone, could you take photos of the circuit boards inside the collar unit and the receiver and post them on this forum? If so, we may be able to guess the approx frequency by studying them.