Maker Pro
Maker Pro

Electric Fence Ground Fault Detect Circuit

Grnadpa

Mar 13, 2010
5
Joined
Mar 13, 2010
Messages
5
Hello Folks,

I'd like your thoughts on developing an indicator (e.g. light or some such) that triggers when a section of electric fence grounds. Ideally, it would also disable that section allowing the other sections to function.

I've wired the paddocks (corrals) independently, but as it stands, the whole fence is grounded when any part of any section grounds. (The commercial unit that generates the current intends that the fence ground only when an animal touches the attached wire).

Flunked an electronics course 40 years ago. And that was my last electronics encounter. So if you are kind enough to respond, keep in mind that I wouldn't know a PNP from an NPN transistor.

We live on a farm with 5 paddocks ("corrals" for your Western Pleasure folks) which I've wired with an electric fence (in parallel I think). The wire is not grounded so when the horse touches it, it grounds through the horse -- at a respectable voltage (though the amps are probably very low as I survived <grin>).

The problem is that anything can ground the wire, such as a tree branch or a weed or something laying on it. When that happens, the fence is useless until I find the grounding point. And I'm not always aware that the fence has grounded, nor able to locate the point without following the entire fence line.

Again, unencumbered by any electronics recall, I'd think this would be a fairly simple circuit.

Any thoughts?

Regards,

Grnadpa
 

Mitchekj

Jan 24, 2010
288
Joined
Jan 24, 2010
Messages
288
I have no experience with electric fences, but it seems to me that you'd be able to monitor the current coming out of the supply. If there is any current flow, then you know that something is providing that path to ground.

If that's the case, I'm sure it'd be pretty easy to come up with something simple to get a light to light. :) What kind of voltage / current is the supply rated at? Should be able to put a current sense resistor in series with the output, and use an op-amp or similar across it as a comparator to switch a transistor/FET which would drive your light.
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
It appears that you have the fences connected in parallel from a single controller. In this configuration, if any fence grounds, the whole system is going to go down. There is no way to get around that with a single controller.

Think of it another way. Assume you had one guard dog that could sense when there was a problem. When he senses a problem, he would go to where the problem was and bark. While the guard dog is busy barking in one padlock, no one is monitoring the remaining padlocks. They only way you can have those padlocks protected is if each padlock had its own guard dog. That is basically what you will have to do.

I know this is not the answer you wanted to hear, but the only option is to get a separate controller for each padlock.

Your second issue is finding out when there is a short. Do they make controllers that have an alarm? It has been a while since I looked at one up close, but I recall them having a fence OK light that flashes typically once a second to indicate that the fence has been energized, typically to a few thousand volts. If the light stops blinking, it means the fence is shorted. The bigger problem is if the light is not getting as bright as usual, which indicates a partial short.

It can be possible to make a photo-detector circuit that can be attached to the LED. If there is no flash in 5 seconds, the thing can alarm you. The problem is that the circuit is going to require a bit of futzing around. I have no idea how bright this light is and how to attach something to it. Once you figure out how to attach to it in such a way that sunlight does not trigger the photo-detector circuit, you then have to set the threshold so that the photo-detector will get triggered when a flash of sufficent brightness happened. You then have to build the circuit to detect the flashes and raise an alarm when they stop happening. Once you have the alarm set, you need to figure a way to get the alarm to you so you can act on it. All of it is doable, though not particularly simple. Are you up for it?

And you get to do it over for each padlock. And the photo-detector has to be attached in a way so you can remove it to see the light and then put it back on it without a major problem.

Another remote possibility is if there are fence controllers that have a built in alarm. You may want to look into that.

Keep us posted.

---55p
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
If there is any current flow, then you know that something is providing that path to ground.
The current flow is so small that it will be virtually impossible to detect. The system basically charges up a cap at 1Hz to a few thousand volts and connects it to the fence. The cap discharges quite a bit through self discharge. At least that is what it was doing back when I was up close and personal to one.

---55p
 

Mitchekj

Jan 24, 2010
288
Joined
Jan 24, 2010
Messages
288
Aaaaah, I see. That makes more sense than the current coming directly from the supply, with the distances involved and all.

So maybe something that will monitor the voltage across the caps? If there is, say, a tree branch keeping the system down then there wouldn't be the normal voltage across the caps, right? Being in the thousands of volts, may need a voltage divider, etc.?
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
At 5,000V, even a 10Meg resistor will drain 0.5mA which is likely to be a huge percentage of the current the device can put out. You will also have a hard time finding a resistor that is rated to 5KV. Sure, they exist, but it is not something you will find at Digikey or similar places.

But the bottom line is that you or I can not design anything till we get our hands on a device and see exactly what it puts out. I certainly do not have tools to work on the hot side. My 100:1 scope probe is only rated to 2.5KV.

---55p
 

Grnadpa

Mar 13, 2010
5
Joined
Mar 13, 2010
Messages
5
Thank you for your thoughtful responses.

What kind of voltage / current is the supply rated at?

It's a Dura Shock Model 525 Input 110-120 VAC, .027A, 60HZ, 10 Watts. Assume the latter three arguments are the output.

How much juice does a LED require?

It appears that you have the fences connected in parallel from a single controller. In this configuration, if any fence grounds, the whole system is going to go down. There is no way to get around that with a single controller.
...
I enjoyed your watch dog metaphor. And I am aware that -- as it stands -- once any sector is grounded no sector is "live".

What I was thinking is that I know bathroom and other circuits have a "ground fault detector" that will close down the faulty outlet, but leave the remaining elements, such as lights, operational. As far as I know, these ground fault outlets do not require a separate circuit breaker at the fuse box.

Curious as to whether whatever electronics is involved in these circuits could be adapted to each of the parallel branches thereby breaking the circuit to the grounded sector without shutting down the whole system.
Your second issue is finding out when there is a short.
The Dura Shock controller has a separate unit with five lights that connects to a ground. Determining whether the system is live or dead is strictly a manual process. There is no automatic indicator. To determine if the system is operational, I touch the unit to the terminal on the controller. If all five light, the system is fully functional. If no lights, the system is grounded somewhere. The only time I've seen anything but all five or no lights is during and immediately after a rain or snow storm.

My thought was to install a LED or some very low amp light to each parallel circuit so when that circuit grounds, the LED would light. I understand the other circuits would not be energized, but I was thinking that at least the light would go on for the circuit that grounded.

Again, folks, I greatly appreciate your thoughts. Please continue to point out the error of my ways.

Regards,

Grnadpa
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
What I was thinking is that I know bathroom and other circuits have a "ground fault detector" that will close down the faulty outlet, but leave the remaining elements, such as lights, operational. As far as I know, these ground fault outlets do not require a separate circuit breaker at the fuse box.
The bathroom GFI (ground fault isolator) is located in the outlet. When it detects a fault, it disconnects that outlet form the power.

An analogy would be a circuit that sits between the controller and each section of the fence and when it sees a section activate, it will shut the section down. Are you willing to go and reset a section every time an animal touches it?

The reason this approach is not going to work is because in the GFI in your bathroom, the system grounding out is a fault condition and always a fault condition. In this system, a section is grounded when it works and it is grounded when it has a fault. How do you tell them apart? You can do that by injecting a shock on the line, but at that point the isolator becomes more complicated than the controller and you might as well add a controller.

---55p
 
Last edited:

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
It's a Dura Shock Model 525 Input 110-120 VAC, .027A, 60HZ, 10 Watts. Assume the latter three arguments are the output.
That is the input specification. You have not provided any output specifications and I suspect that the manufacturer will not provide that specification because they consider it a trade secret.

---55p
 
Last edited:

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
My thought was to install a LED or some very low amp light to each parallel circuit so when that circuit grounds, the LED would light. I understand the other circuits would not be energized, but I was thinking that at least the light would go on for the circuit that grounded.
The problem with that approach is that without isolation, all circuits are grounded when one is grounded. You can not diode isolate them because you do not know if the pulse is a positive pulse or a negative pulse. Without much more detailed information about the inner workings of the circuit, it will be hard to design something that I know will work. Others may be more adventurous.

---55p
 

Grnadpa

Mar 13, 2010
5
Joined
Mar 13, 2010
Messages
5
The problem with that approach is that without isolation, all circuits are grounded when one is grounded. ...
---55p
This is not a pulse unit. I believe the theory is that the circuit remains broken until something grounds the circuit -- such as a horse or, far too often in my case, the horse-keeper.

Are you willing to go and reset a section every time an animal touches it ---55p

I didn't think about that, but yes I think I would be willing to reset it every time as opposed to the present all-or-nothing.

Thanks again.
 

Grnadpa

Mar 13, 2010
5
Joined
Mar 13, 2010
Messages
5
You can not diode isolate them because you do not know if the pulse is a positive pulse or a negative pulse.
---55p
This is not a pulse unit. I believe the theory is that the circuit remains broken until something grounds the circuit -- such as a horse or, far too often in my case, the horse-keeper.

Are you willing to go and reset a section every time an animal touches it ---55p

I didn't think about that, but yes I think I would be willing to reset it every time as opposed to the present all-or-nothing.

Thanks for your analysis.

Grnadpa
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
This is not a pulse unit. I believe the theory is that the circuit remains broken until something grounds the circuit -- such as a horse or, far too often in my case, the horse-keeper.
It is my understanding that all cattle fences are pulsed. The controller delivers a pulse of energy to a capacitor connected to the fence at a fixed time interval, maybe once a second. If they were a continuous system, the animal that touches it would likely die. This way energy is delivered to the output, it starts to decay on its own, maybe half of it is left when the next pulse is delivered to replenish it. If an animal touches the fence, it gets hit with whatever energy is left. It then gets hit with a fresh pulse some time later, then another pulse a second later and another one... till it is no longer grounding the fence.
I didn't think about that, but yes I think I would be willing to reset it every time as opposed to the present all-or-nothing.
I am missing something. Are you saying that you are OK with the system shutting down every time it is grounded (a fault) or every time an animal touches it? That is compared to the current situation of the system becoming inoperative when there is a fault and continuing to work when an animal touched it and backs off?

The more I think about it, the more I am convinced that you will be better off just buying 4 more controllers and isolating the padlocks. There is just no practical way to have them connected and isolate a fault.

The alarm features you are looking for will be easy for a manufacturer to add to their product. It may be possible for an aftermarket manufacturer with the right equipment to design an aftermarket fence monitor. One may already exist. Unless your fence is really badly installed, others should be experiencing the same problem and there should be a market. You should search for it online, if you have not already done so. It will be very hard for a hobbyist to design one, which is basically where you stand.

I believe your best bet may be to look for an existing alarm product (if you have not already done so) or take your idea to the fence manufacturer and sell them on the idea of having an alarm. You should also have an expert look at your installation in case this is an installation problem like having the bottom of the fence too close to the ground. Others here may have other ideas.

Good luck.

---55p
 

Grnadpa

Mar 13, 2010
5
Joined
Mar 13, 2010
Messages
5
Thank you for your patience. You saved me a great deal of frustration.
 

neon

Oct 21, 2006
1,325
Joined
Oct 21, 2006
Messages
1,325
consider putting LED at specific points if the 10ma draqin is allowed to light them up then when one circuit dies the LED will be not lighted so you know where the fault is. consider also a solar panel to supply the current that becomes expensive kind of but it will work.
 
Top