Electric Fence Zapper

[rossw]

Conclusion - am putting 600 volts into a 450 v AC rated capacitor. There may be some internal damage to the capacitors, it is after all these things that hold the energy.

sqrt(2) * 450(Vrms) = 636V(peak)
You are still comfortably within their specification (and that's neglecting ANY safety margin)


To what has been mentioned previously - quantifying the output. Do you have a CRO?
Preferably a DSO or one with long-persistance phosphor?

If it were me (it's not me!), I would set up a simple test, and it would go like this.
1. a CAPACITIVE divider to measure volts. Some HIGH VOLTAGE caps, something like 10pF for the "high side"
and perhaps 10nF for the low side (low side won't need to be high voltage), will give you a 1000:1 divider.
So 20KV to the "hot" side will give you 20V to feed to your CRO. You can measure the rise time, pulse width and peak voltage this way.

2. an inductive loop CURRENT sensor. Basically, make a CT. If you wound say 1000 turns of fine wire around a core, passed your high voltage output (via suitable high voltage cable (sparkplug lead?)), and put a burden resistor across its output, you should see 1mA for every 1A in the "primary". Choosing a suitable resistor (10R lets say) would give you 10mV out per amp in the primary.

Between them, you should be able to see the volts and amps at the same time, to quantify the output.

[David HK]

Am catching up on what I proposed to do on Saturday last.

The following details set out the technical parameters of both machines:-

Dave

[David HK]

Just tested a brand new 450v AC 50uf capacitor.

The result is exactly what happened with the very first test of machine B.

All connected up, new Capacitor in place, output lead connected to a necklace containing 3 x 47 Ohm 2 Watt resistors in series. Power on and instantly for the first 4 pulses sparks fly out of the resistors.

Thereafter no more spark emission from the 3 resistors. Disconnect test necklace and insert one used 47 Ohm resistor between the output wires and there is much cracking and sizeable blue sparks from the resistor. In other words the same results seen during other tests on both machines.

Then to satisfy curiosity I replaced the 3 resistors in the test necklace with another new set and the results are the same as written above.

I then cut out one resistor and left two in the test necklace and heavy sparks flew out of both for a couple of pulses.

The conclusion is that both machines seem to be working fine. The way the resistors react might be due to different batch production, damage done internally that affects the way in which the current flows or does not flow, humidity absorbed due to outside storage and so on.

As you said Oztules it may be a non problem.

I note Ross' comments about an oscilloscope and I do have one - given to me many years ago. I will dig it out and see what I can do to set up the circuits mentioned by Ross. I may need some help and guidance as I go along.

Dave

[David HK]

I mentioned in a previous note that I had made a new bobbin for Machine B.

I have since made a better second one which is almost a clone of the first except that it is built with intelligence. The second bobbin shown below is a jigsaw construction in that the centre core pieces are fashioned with what can be described a mortice and tenon joints. The material is 1.5 mm thick fibreglass sheet which is easy to saw and file. The ‘long’ pieces are wider than required and each end contains two ‘prongs’ that prevent the main side walls from slipping off.

This second bobbin will be fitted with a 1.6 mm diameter 9 turns of two-in-hand primary, followed by 200 turns of 1 mm secondary and will eventually be retro fitted to Machine A.

Good fortune seems to be on my side. The transformer I purchased from a local scrapyard a few weeks ago has provided the laminates for the first bobbin and as luck would have it countless cuts with hand shears has provided the laminates for the second. Photographs below

Dave

[David HK]

Here is a photograph of Machine B.

The capacitor is mounted on a mini gantry over the printed circuit board which significantly reduces the size of the mounting board. The PCB is identical to that on Machine A. Only the main transformer coils differ slightly between the machines - details in an above post.

Dave

[David HK]

I mentioned in a previous note that I had spent some idle time playing around with the layout to achieve 9 (not eight) and 12 Triac PCB layouts. At the same time I included everything else I could think of to produce many options. In the main these are:-

Spade connectors or conventional wire connections to and from the PCB.
A clip on heat sink for the TIP 31 - if its there it can be used.
Proper sized holes and space for the large diodes.
Numerous options on the number of triacs needed.
Jumper holes to connect 'live' triac tabs (A2) centre leg to the PCB next to the toroid. X marks the triac   for the multiple jumpers.
The planes have been chamfered. 
Heavy duty triacs with spade connectors could be connected if necessary.

What have I missed?

Dave

[David HK]

Hello Ross and Oztules,

Thanks for the note on test circuits.

I plan to follow this up.

I do have an oscilloscope model Hitachi V-212/211 CRT – a gift of old and I have never used it. Bandwidth DC to 20 MHz. Sensitivity  1 m/V/div

I plan to obtain the various parts in your note and then draw up the circuits, perhaps you could shroff them over when they are ready.

With regard to the 1,000 fine wire turns on a core, can you confirm a ferrite rod 100 mm x 10mm would be okay, and will each layer require a paper separator?

What would be the Watt value for the 10R resistor?

Dave in HK

[rossw]

I do have an oscilloscope model Hitachi V-212/211 CRT – a gift of old and I have never used it. Bandwidth DC to 20 MHz. Sensitivity  1 m/V/div

The pulses *MIGHT* be hard to catch on a CRO of that type - purely because of the low repetition and short duration.
That's why I indicated long persistance phosphor or DSO.

I plan to obtain the various parts in your note and then draw up the circuits, perhaps you could shroff them over when they are ready.

Sure.

With regard to the 1,000 fine wire turns on a core, can you confirm a ferrite rod 100 mm x 10mm would be okay, and will each layer require a paper separator?

A core is probably irrelevant in this context. I'd wind an air-cored coil you can pass a decent insulated conductor through.

What would be the Watt value for the 10R resistor?

Well, lets assume it makes 400 amps peak.
At 1000:1 you can expect 400mA in the secondary.
400mA through 10R = 4V and 1.6W. But for a poofteenth of a second. A 1/2W resistor should easily manage that.
Allow some safety margin if you like, use a 1W   I doubt it'll even get detectably warm.

[David HK]

Thanks Ross,

Very helpful guidance. The CRO is the best I can do, its worth a flash in the pan for a try. I don't know anyone with a DSO.

Dave

[rossw]

Very helpful guidance. The CRO is the best I can do, its worth a flash in the pan for a try. I don't know anyone with a DSO.

If you get the triggering right (should be fairly easy, it's a nice big step-change in input), and turn the brightness right up and do it somewhere in subdued light, you might be able to see it well enough...

[madlabs]

It never ceases to amaze me how many nice folks you meet on the internet, if you hang out in the right places like this forum. As y'all may remember, I started this thread because I have a bear I need to keep out of my garden.

David in HK has sent me one of the machines he built! And refused to take anything for it! Dave, folks like you make me think there is hope for humanity, you built it for the puzzle of it and then gave it away.  I'm a guy with more projects than can be done in a lifetime, and having one knocked off the list is huge. And this was one I had to get done before fall, when there is a lot of honey and fruit in the garden.

Not only that, but he sent amazing documentation, parts, etc! As I told him, my docs usually consist of a couple of coffee stained pages of scribbles. His work is truly outstanding and the machine itself is beautiful!

Thanks again Dave! I'll post up some pictures of it installed and ready to teach a bear a lesson!

Oh, I do have a digital scope. Just a cheap USB one, but it works. To be honest I don't have time to catch up on this thread this morning, but I'll of course be happy to test whatever whenever the machine gets here.

Jonathan

[madlabs]

OK, I took a sec to scroll up...

I have the high voltage caps needed to make the divider. I have some 50kV wire as well, better than spark plug wire, it's copper. The coil might be a problem, my winder is still misplaced after a move. Might be able to salvage the coil from a solenoid or relay though.

Jonathan

[David HK]

Not forgetting Oztules who is the brains behind it all.

Dave

[David HK]

I am still in the process of searching for and collecting components for measuring voltage, wave form and amperage. While that is going on I have also been engaged in searching for lighting equipment that will indicate when the electric fence wire is ‘live’.

After two days trawling the Internet some success has been achieved.

Firstly, electric fence manufacturers such as Gallagher, Patriot, et al do offer warning light accessories that clip onto a fence wire. Some seem to be powered by the pulses and others will only activate when there is no pulse present. Other kit contains battery power. Trying to find recognizable circuit layouts for this kit has been wishful thinking so far.

I then turned my attention to motor vehicles and in particular ignition wires for which there is testing kit. Most of this is associated with engine timing and I eventually found several circuits that looked promising.

The circuit components are fairly easy to acquire but the main difficulty is with the ‘pick-up’ which usually surrounds the ignition wire. Commercial timing guns feature a clamp which contains a U shaped ferrite with x number of turns of magnet wire. The opening aperture for the wire can be covered by a sliding ‘door’ which also contains a ferrite to make the thing box shaped. I can’t find any of these items locally so I decided to make my own.

Fortunately YouTube contains a video from someone who has been down this road already so please view the following URL:-

The chap also has a web site which contains more detail:-

http://www.howtoalmanac.com/kevin/projects/automotive/timinglight.htm

Trying to find a suitable ferrite that can be cut up for the purpose proved impossible so I rummaged through my odds and sods coils box and dug out two pieces that seemed suitable for the job. Another coil was deprived of its fine wire winding and the first U ferrite contains many turns which were not counted at the time. The second ferrite – the E shaped one - contains 150 turns.

Along comes testing time and Machine A is set up as is the test circuit. Most things that I build always fail first time around so I was not imbued with great optimism. The ferrite was placed under the ‘fence wire’ and the machine turned on and lo and behold the test circuit red LED showed a small flash in tandem with the machine pulse. The red flash was difficult to see so I recalled the comment in the video where the speaker recommends a high brightness LED. Later in the day I made the changeover and the difference is spectacular. A nice bright white flash.

Human curiosity being what it is, I then moved the ferrite from under the ‘live wire’ to test the operating distance and found the following:-

9 inches (230mm)         No perceptible signal to trigger LED

8 inches (200mm) to the wire itself                    An ever increasing signal with proportional LED brightness.

I then tested this in circle form using the ‘live wire’ as the centre and found that the test circuit would work anywhere around the ‘centre’. The non-working and working distances are shown for 9 inches and 8 inches above which are the radius of circles.

Car ignition leads   Tested on a Mazda 323 which appears to have HT coils for each plug –             ferrite adjacent to the wires with corresponding bright LED flashes. Maybe detecting adjacent ignition wires but it does work.

This was interesting and as far as I got in the afternoon.  It therefore seems that the output of Machine B at X voltage and Y amps will send out a pulse that has an energy ring of approximately 16 inches (406 mm)  in diameter around the wire itself. Perhaps readers can tell me more about this.

My next test will involve a long run of insulated wire including ‘connectors’ to represent ‘gateways’ that occur in fences to see if the same energy circle runs through the wire at a greater distance.

I mentioned earlier that I found other circuits during the search and here is another that I plan to build – follow the URL:-

http://www.sportdevices.com/rpm_readings/

I have converted some of the circuits into Express PCB Schematic files so they are available for easy and clear printing. If anybody wants them in soft format let me know when all have been tested and reported on.

Comments welcome.

David in HK

[rossw]

David, I hope this is drawn wrong:


As presented, your coil and cap are purely a capacitive pickup, feeding whatever is detected straight to the gate of your FET with just a small diode clamp to save it!

It would also explain why you're picking up signal inches away from the wires. (You shouldn't!)

The pickup coil should be HIGHLY selective and ONLY pick up current flowing through the core. I think if you look at your source material again, the other side of the coil and cap should go to ground.

Since it's acting now as a CT, you probably should also have a burden resistor across the input. The value will depend on the number of turns you have on the coil. This type of sensor is designed to detect *CURRENT* in the line, not the presence of voltage across it. (A fence "alive" detector should be detecting voltage, not current)