Prevent animal poaching by sound: Please help!

Thread Starter


Joined May 17, 2010
Hi Guys!

We've come up with a concept of detecting the illegal cutting of game fencing (a savere problem in terms of animal poaching and theft in South-Africa), and we could really use some help..

Background: Game fencing are 20 -30 strands of steel wire aproximately 2.5mm thick, under severe tension, stranded parralelle. Some game fences are up to 200km in distance.

Several "alarm" options have been tried with little sucess. Not going into the details of how we plan to convey the signal... we'll cover that, but we need a pulse trigered by sound.

See, when any of these steel wires are cut, they make a very destinktive "TWANG". It MUST be possible to have this "TWANG" recognised by a simple inexpensive device, filtering unimportant frequencies, and trigering a pulse. It should be possible to detect this "TWANG" over vast distances. (Think of the two cans conected by a tensioned string "telefone").

Please guys, any assistance is much appreciated.

John P

Joined Oct 14, 2008
You say the fence can be 200km long but I bet the individual wires are nowhere near that. They must have numerous intermediate endings between the ends of the fence, though I suppose they could be spliced or clamped together to make up much longer lengths. I think you need to explain more about the mechanical construction of the fence. Any chance of pictures?

And speaking of shortened individual wires--there must be occasional gates in the fence. How are those built?


Joined Jul 7, 2009
An interesting engineering problem! I assume you're talking about fencing like what's shown here, right? Have you contacted the various vendors and asked them about solutions? Clearly, it's in their best interest to know what works and what doesn't.

When the wire is cut, it will no doubt have a compression wave travel down it (i.e., the "sound" is being conducted in the wire's metal at around 4 km/second). I would worry about this signal being attenuated by the wire's supports over a long distance (not to mention impedance problems at splices). Clearly, unless you already have this data, it requires an experiment to see how far these waves can be reliably detected. Then you'd put signal detector stations at smaller distances apart that would somehow signal when a wire has been cut. No doubt this work has been done numerous times before, so your first place to visit should be the library.

Besides the attenuation, you'd also want to make sure you ignore false signals, such as a limb falling on the wire.

But there are other techniques. A force monitor could detect loss of tension in the wire and send a notification. In other words, if the wires are cut on one side of a post, then that post is going to lean. Maybe a simple level detector could be put on the post and it squawks when its level changes sufficiently. Of course, if the bad guys recognize a sensor, then just cut the wire on both sides. That means you'd have to put a thing that looks like a sensor on every post.

A problem, of course, is that these sensors have to be put out in the wild and the poachers will no doubt learn of them quickly and figure out ways to quickly disable them (such as smashing them with a club, shooting them, or burning them). So that's also a problem, perhaps an even bigger one than how to detect the wire breaking. Besides, I'd imagine a bunch of bad guys could cut the wires in numerous places and it would take the good guys quite a while to figure out where the "active" break is.

Another approach is to use TDR, which basically sends an electrical signal down the cable and looks for reflected energy -- a break causes a discontinuity in the cable, causing energy to be reflected. The characteristics of the reflected energy can tell you where the break is located (i.e., the distance down the wire). Unfortunately, the cables are probably connected to metals stakes that are driven into the ground; being uninsulated from the posts makes the problem a rather intractable electrical problem I would think. If you were to use this technique, it would require a special cable that will be easy to recognize. And a knowledgeable poacher could probably quickly cut and terminate the cable, so the TDR signals would have to have short periods. And, anyway, a number of poachers could coordinate a bunch of cuts on the wire over many km of fence and, again, you'd have no easy way to find where the active break is.

Another technique would be to have alternate "pluckers" and "sensors" spread out on the fence. The pluckers would "pluck" the cable at a known time and the sensors would expect to see a pluck signal at such-and-such a time. If not, then they'd squawk that there's a problem.

There exists technology to locate a gun from the sound of a shot being fired (again, impulse sound). I believe it works basically by triangulation, but the problem is made more complex by the acoustics, especially in urban environments. Thus, I think they first go in and measure the acoustical characteristics and can then allow for them. Perhaps something like this could be used along the fence, but it isn't going to be cheap and the bad guys could probably disable it pretty quickly.

I hate to sound negative, but I doubt there's a cheap, simple, and reliable solution. And even if you do come up with something technically viable, it's only a matter of time before the poachers learn about how it works and a way around it.


Joined Jul 17, 2007
I was thinking along the lines of TDR as well, but using the acoustics of the wire instead of electrical signals.

If a cable under tension were cut, a very strong acoustic wave would travel down the wire in both directions until it was attenuated by passing through mounts, etc. If the speed of the wave is known, by receiving indications from two or more sensors indicating the event placed a known distance apart on the cable could give a close estimation of the distance and direction to the break from the location of the sensors.

Rich (BB code):
                    break           sensor1   sensor2
more fence <----------x----------------o---------o---------> more fence