I was wondering if anybody had any suggestions for my application. What I am trying to do is with using an I/O link, indicate that my horse fence/ hotwire is on or off, the fence puts out around 9kv, it pulses the fence for a second and then is off for a
second. I was hoping to find some type of relay that would close but wait a few seconds before opening back up. Not sure if anything like this exists.
Any suggestions would be greatly appreciated.
Brian

I have investigated this before although I never finished wiring mine up. It's not a simple problem because you need to detect a high voltage while protecting your monitoring electronics (which ultimately provide a path to ground). Also you need to be careful not to put too much of a load on the fence - even though it is a high voltage there is very little current behind it, so if your tester loads it down too much, the fence will not be effective.

Hand-held electric fence testers that I have seen use one of three designs: 1) a high voltage step-down transformer driving an analog meter, 2) LEDs with a large resistance in series, or 3) neon bulbs.

I figured the LED technique would probably work for this, so what I did is use a pair of optical s/pdif (TOSLINK) transmitter and receiver modules. The transmitter is wired in series with some large-value resistors across the fence. The whole transmitter and resistor assembly needs to be potted to prevent arcing - I used hot glue for that, although a commercial potting compound would be better.

Then a plastic toslink cable connects it to the receiver. This inherently provides the opto-isolation needed to ensure your electronics are safe.

The receiver needs 5VDC power and will output short pulses (maybe 50ms). I got as far as that, which seems to confirm that this design will work. These pulses probably are too short for an IOLINC to detect so some additional circuitry (eg a 555 or 74HC123) is needed to turn the pulses into a longer signal for the IOLINC to detect.

Barnman wrote:I was wondering if anybody had any suggestions for my application. What I am trying to do is with using an I/O link, indicate that my horse fence/ hotwire is on or off, the fence puts out around 9kv, it pulses the fence for a second and then is off for a
second. I was hoping to find some type of relay that would close but wait a few seconds before opening back up. Not sure if anything like this exists.
Any suggestions would be greatly appreciated.
Brian

There have been some discussions elsewhere on the forum of LED detection mechanisms. I beleive smart home actually sells a kit that will detect whether an LED is on. Most electric fence chargers have a light that flashes when it pulses the fence. Maybe you could take advantage of that?

I like the led idea at the box, the problem I see is that both my energizers have 5 LEDs a piece to indicate if there is a short to ground in the fence, so it may not be as easy that way. I have a little battery operated led light that hangs right on the wire it doesn't go to ground at all. When there is no power on the fence the
Led flashes to warn you the fence is
not on. I wonder if somehow I could rig that to work.
Thanks for the input guys
Brian

Barnman wrote:I like the led idea at the box, the problem I see is that both my energizers have 5 LEDs a piece to indicate if there is a short to ground in the fence, so it may not be as easy that way. I have a little battery operated led light that hangs right on the wire it doesn't go to ground at all. When there is no power on the fence the
Led flashes to warn you the fence is
not on. I wonder if somehow I could rig that to work.
Thanks for the input guys
Brian

That would be ideal, as you would be getting feedback as to the actual fence condition as opposed to what the controller "thinks" is going on.

here is the product I was thinking of.

http://www.smarthome.com/24950A2/I-O-Linc-INSTEON-LED-and-Light-Detector-Probe-Kit/p.aspx

Once you get a signal to the I/O Linc, latching an indicator light on should be easy in Indigo. Simply have an Indigo trigger that looks for the I/O Linc input to turn, then you would need 2 actions: 1) to turn on some indicator light; and 2) to turn the light off after some delay, say 4 seconds, and check the Override previous delay box. As long as your I/O Linc keeps on sending ONs every second, the lamp will stay on.

berkinet wrote:Once you get a signal to the I/O Linc, latching an indicator light on should be easy in Indigo. Simply have an Indigo trigger that looks for the I/O Linc input to turn, then you would need 2 actions: 1) to turn on some indicator light; and 2) to turn the light off after some delay, say 4 seconds, and check the Override previous delay box. As long as your I/O Linc keeps on sending ONs every second, the lamp will stay on.

You need to latch at the sensor. One insteon command per second is a lot of chatter on your powerline - it will cause serious problem with other switches as Insteon does not handle collisions well.

Just a couple of wraps of magnetic wire around the hotwire might work. # of turns will/should dictate the voltage out. (this is if the 9KV is AC)

I'd still buffer it. The neon bulb is a cool trick. Did that with a phone ring circuit I did many moons ago.

One of these (under $15) might do the trick, though you'd need to check with the vendor to see if it could take 9kv and switch fast enough. Just a few wraps of the fence wire through the core and you'd get a closure whenever the fence is hot. (Essentially what anode suggested but in a commercial package).That would still leave you with the latching problem. But, as seanadams points out, a 555 timer would fix that and they are really easy to work with if you have ever done anything remotely electronic.

So, you would have a current sensing switch firing a 555 timer (with say a 5 sec timeout) feeding the input of the I/O Linc. As long as the fence fired once a second, the I/O Linc input would stay in an "ON" state.

If you wrap magnet wire around the hot wire, the 9KV will instantaneously blast right through that enamel coating and find its way to ground through your measuring equipment, and then your power lines. There is a reason the screw terminals on the energizer are a couple inches apart... this is serious voltage and to transmit information across such a potential requires special considerations. If you're going to wrap anything directly around the hot wire I would suggest spark plug wire, although I still think it's a bad idea. Another possibility would be to use an empty spool of fishing line or something to enforce a minimum dielectric around the wire. Even then I don't think it's a good idea.

But would it work? It would be interesting to try, but maybe not for the reason you're thinking. What you have described is properly called a "current transformer", and its application is to detect a current going through a conductor, NOT a voltage on a conductor.

So what do we actually want to measure here? The problem is, during normal operation current does NOT flow on the fence, it is only when there's a load on it - i.e. either a short or an animal touching the hot wire to complete the circuit. So in fact with a CT, were are looking for an _absence_ of a signal to indicate that the fence is OK!

Also the energizers don't actually switch the fence on for a while, then off for a while. Each cycle is a single high voltage spike from an inductor - like the spark coil in an engine for example.

An electric fence can be broken in several ways. Let's address each one:

1) energizer is unplugged. Or maybe it is on a switch and you want to get a warning if somebody leaves the switch off. This is easy enough, just use the "inlinelinc relay" with sense input to detect the power going to the energizer. (I think we talked about that one before)

2) energizer is not functioning. To detect this we need to look for a voltage on the hot wire where it exits the energizer

3) fence wire is open (not connected). Again we need to detect a voltage on the wire, but in this case if we're going to detect a break in the middle of the fence, we need to be measuring the voltage at the far end, away from the energizer.

4) fence is shorted to ground. Here is where the current transformer might work - in fact this might be exactly what is inside the energizer with the "shorted" LED indicators. When the fence is shorted, the current will increase. We could also do this by checking for voltage - the voltage will go down when the fence is shorted.

So in reality if we want to fully characterize the state of an electric fence, we need to take a few different measurements in a few different places. The single most effective measurement to cover most cases would be if we could non-intrusively monitor the voltage on the far end of the fence wire.

seanadams wrote:If you wrap magnet wire around the hot wire, the 9KV will instantaneously blast right through that enamel coating and find its way to ground through your measuring equipment, and then your power lines. T

Oooh good point! except for the straight to power lines. An EZOI is isolated, and without looking at the data sheet, doubt the fence has enough currents to jump that isolation. (the wire wrap I can see though)

Don't forget there main 3's items to be satisfied, voltage current and resistance. (of course dielectric strength factors in with HV, hence the 'quality' if the insulation)

Got a HV probe here. Whats the uA of the fence? (but no Meg'er)

Though me? I'd pop it open and 'watch' the primary of the HV xfomrer.

Here's a pic of my optical transmitter. I made this a couple years ago just to test the concept and it definitely works.

I never finished the receiver end but I tested it and was able to get pulses.

It's just a toslink transmitter (TOTX177L) wired in series with a couple of resistors. Sorry I forget what kind of resistors I used and now they're potted so I can't tell you for sure.