Electric Fence Zapper

[oztules]

That looks to tick all the boxes.
I just used a top3 outline for the triacs, and that doubled the realestate, plus making the track very thick in the Z plane (third dimension).

I'm pretty sure I ended up buying the tabs from HongKong from RS components ... 40 dollars /1000.


.................oztules

[madlabs]

Oz, I usually work like you and build things out of what I have. My schematic generation and product development are like yours. But David's board looks so sexy! David, any chance of a to-scale jpeg, or I'll see if your CAD is compatible with mine. Those boards will be a snap for me to make here at home.

Jonathan

[David HK]

Oztules and Madlabs,

If you can down load free software from Express PCB ---- this URL   http://www.expresspcb.com/expresspcbhtm/download.htm     I can send you my files for the Schematic and PCB directly. From these you will be able to print out actual size and enlarged sizes up to A3 if you prefer.

Direct from the horses mouth will avoid losses in clarity through different software and scanner mediums. Please PM me with your E-mail address and I will send when ready.

You may be surprised how nice and clear the circuits appear on-screen. I have checked the PCB layout against the Schematic and everything links exactly.

Please note that my AutoCad file of the circuit layout is a bit rough and ready and was done for the purposes of circuit tracing only.

Okay, so far so good. Next step, assembly of parts and coils. Thanks for the tip on the PCB spade connectors, I also have a fair bit of gear on order from RS components and it should arrive this week. I will try another visit to local scrap yards and see what I can find.

Dave

[David HK]

Here is my end of this week roll up.

In the above note I mentioned that I would devote time to collecting parts and other related work.

Most of the parts have been ordered through RS Components Hong Kong and will arrive in the middle of the week ending 27th April 2013.

I did discover one mistake with the triacs I ordered. I selected a type that had insulated tabs whereas in this design I need non insulated tabs. You will see later how I link the tabs with a bolt and copper spacers to a thick copper winding wire that lands on the board. In helps make up for insufficient contact area on the middle leg of the triacs.

Documentation has been useful and I now have an Excel file with multiple sheets showing different things. In one sheet I have drawn up a table showing wire diameters and the theoretical number of turns that can be done on the width of the ETD 49 bobbin. I have subsequently wound a bobbin with 0.8 mm wire and discovered that instead of the theoretical 20 turns of 2-in-hand, I can only achieve 18 turns in a single pass. The difference is of course due to enamel thickness on each wire plus a minor amount of slippage space due to the single pass and non over-lapping requirement.

Other documentation covers capacitor energy values in Coulombs and Joules. Another deals with the inescapable fact of dollar cost which does not bother me too much since this is an advance birthday present.

The coil winder metalwork alterations are more-or-less completed except for an arm to hit the counter each revolution. This should be in place by the weekend.

The PCB layout has been refined with minor cosmetic improvements but without change to the overall layout. Warning text has been added for obvious reasons.

A good size mounting board has been obtained and is yet to be equipped with stand off rubber feet.

The elusive 50uf 600 volt capacitor continues to be a problem and in the whole of Hong Kong I have found one shop that sells high voltage capacitors which are mostly for very expensive sound systems. The cost is in the Cleopatra’s dowry range. As an interim measure I still have a 50uf 450 volt capacitor which is easily purchased locally. 

On the subject of AC working voltages and DC working voltages for capacitors, this URL touches on the subject http://armyintelligence.tpub.com/IT0351A/IT0351A0015.htm

One other thing on large AC capacitors - the one I have is filled with vegetable oil. Does it make any difference if this is laid horizontally or vertical for operation?

On the subject of testing I have some 1 Ohm wire wound Powerohm resistors which I found in a local scrap yard. One is used as a dump load for Ghurd’s famous contribution to electronic science, and the others are waiting for rainy day use. The relevant website is here:-   http://www.powerohm.com/resistorLanding.aspx?partNumber=WR100

Are three of these any use for testing the main coil output? If not, the local pigeons that roost in my mango tree and crap all over the backyard floor are in for a shock.

Back to documentation again. I have cut out most relevant thread entries and pasted them into an MS Word file. Thereafter I have spent a couple of hours amalgamating everything into a logical story based on a typical 12 VDC system. It covers quite a lot of what has been written above in note form.

I have also taken the time to study other commercial electric fence systems sold as well known brands in the US of A and Australia. I have created an Excel table for Model, Distance and Cost for Gallagher (USA) electric fence machines. The longer the distance the more expensive they become.

The PCB design now has six mounting holes – one in each corner and two roughly middle board.

The wiring links to the PCB have been “planned” to cater for two types of connection – solid and reliable soldered wire connections, and, push-on and pull-off spade type. I am wary of the push and pull spade type because enormous pressures can be applied to the circuit board if they are stubborn in each direction. The presence of two centre board mounting and locking down  holes will go some way to alleviating this problem.

The fence line testing instrument sold in Australia seems like a nice bit of kit and someone has obviously put a lot of thought into its design.

I note in one comment on this topic Oztules mentioned the use of lower value capacitors for shorter distances. Is it possible to make up any form of table relating one to the other? Also, apart from different capacitor values, what would be the knock on effect in changes to the circuit generally?

So it has not been an unproductive week and the next one should see a significant move ahead.

David in HK

[oztules]

Yikes... I just build the thing and turn it on.... wait for the smoke to clear, then rebuild until it works.
450vac cap is perfect for the job, and all I use. The link you pointed to at least shows that is correct.

You could use lower volt caps in series config for the same effect. The DC in this case is behaving like low freq and high freq ac, but the electric field is continuously changing.

Smaller caps only change the rate of charge the oscillator runs at. Small caps, you will have to change the divider network on the transistor as mentioned previously in the text.

MUST keep rate of charge to >= 1 second for health and safety reasons. Dont let it pulse fast, you need time to get away if you tangle in it.


The resistance to mimic the transformer is difficult to determine, as it is the impedance to the pulse that stops the triacs from over current collapse. Their surge is very high, but a dead/near short rather than an impedance will behave differently. The impedance sees it as half a pulse of MHZ frequency voltage, so the impedance will look very different to the resistance in this case. Even winding 20 turns around some heavy steel will help I suspect.

Having said that I=E/R so 600/3=200A, the traics should handle that fine as a short surge.

"The longer the distance the more expensive they become."... I have seen one of these over power a high end unit ($3000).... it's not rocket science....otherwise I would not be able to do it I guess....


.........oztules

[David HK]

Saturday the 20th April 2013.

Most of the hardware assembled. In the photograph are:-

the main mounting board with power supply connectors at one end and the mains transformer at the other, a full size print of the PCB with the ETD 49 coil containing the two primaries, the capacitor, symbolic form of the mains transformer, the coil winder just in view to the right, and the copper pipe is for spacers between the tabs on the triacs.



More next week.

Dave in Hong Kong

[David HK]

For those that are interested there is an article on electric fences at this URL


www1.foragebeef.ca/$foragebeef/frgebeef.nsf/all/frg36/$FILE/fenceenergizer.pdf
cut and paste this into the menu bar and it should work

What caught my eye was the very rough rule that one Joule  is equivalent to about 6 miles distance. I have already cut and pasted a lot of data for Thunderbird, Gallagher and Patriot electric fence manufacturers and the above rule seems to be ~ about ~ the norm.

Dave in HK

EDIT: I made the link clickable. Tom

[David HK]

A bit more from my studies this Sunday morning.

Joule range: This charger puts out a 0.19 joule pulse on a 500-ohm circuit and a 0.16 joule pulse on a 100-ohm circuit

The above statement can be found in some website advertisements for electric fences.

Oztules, do you have any idea how an individual machine could be selected for one or the other? Or, is it a play on words meaning two separate machines?

Dave

[oztules]

"Oztules, do you have any idea how an individual machine could be selected for one or the other? Or, is it a play on words meaning two separate machines?"

Not sure what your saying/getting at here.... so I'll try this scattergun approach.
1 Stored joules are just that.... stored joules thats the 1/2e^2x Farads

2. Energy is energy, the rate it is expended is power.

3. If you have different impedance fences ( not resistance) you will have different characteristics to contend with power wise, but the energy delivered into the load will be dependent on the device matching the load... heard that before haven't we.

4. A scabby transformer with too little iron, will suffer disproportionately from loaded to unloaded fences than a decent core transformer.

5. Once the energy stored is known, and is a constant, and the switching losses are known, and called constant.. then the load matching is all down to the transformer.
 
6. If you look at the midrange units that claim long fence lines, they all seem to have under cooked transformers. They get around this expensive mismatch by using loads of farads, and just assuming/allowing the losses in the transformer matching, using lots of stored joules, and then losing lots of the joules in the transformer.
I have seen some units of modest power, using 300 uf or more, just to match my little one with decent microwave transformers.

7. For my money, there is an awful lot of rubbish written about electric fence units. They are fundamentally simple, and there is nothing mysterious about their workings, or energy / power outputs.... it is simple physics.



not sure if any of that covers your question/s, if not ask it a little differently, and i will try again.


Edit: I said impedance not resistance. Although this is essentially a DC device, it behaves just like ac because of it's rising and falling wave front... so it is frequency dependent on it's transmission in a lot of ways ( other wires in the fence not connected to the unit, the ground etc, capacitances and inductances all over the place... hence impedance not resistance.

That said, resistance is important too, as the grass growth loads the fence up, the  resistance to ground decreases, and this also changes the inductance properties too.... it is a suck it and see situation, and you can calculate roughly with resistance, but you won't then be able to explain the readings when you go out into the real world....... but make no mistake, a low resistance across any part of the fence makes a huge difference.... but not as much as I expected either.

You can find old wire from a near fence tangled up in the electric fence wire ( hot wire run along the front of a normal fence etc), and although it is a direct short to ground, and you foolishly think it will be a zero fence... your eyes will still light up if you touch it. The short to ground never seems to pull the voltage down to anywhere near zero.... and of course when you untangle it.............. sigh ... never learn

This is also  noticeable in a long fence when you find the voltage miles from the unit is higher than at the unit. I suspect standing waves in the fence and/or nodal points.



.....................oztules

[David HK]

Oztules,

You really are a wizard on this topic so I take my hat off to you.

I only asked the question above because I was curious about why someone would write about a comparable performance difference which most readers would not have a clue about. Your scatter gun answer does indeed illustrate the more deeper background of physics and well known 'laws' which apply to the topic. The answers do indeed make sense.

With regard to your paragraph 7, how right you are. In paragraph 1 you will be pleased to know that I played with that formula early this week and have an Excel table worked out for various Farad and Voltage values. I surprise myself sometimes.

I am learning fast.


Dave

[David HK]

Tom,

I missed your Edit note in the above post, so thank you for making the URL work by "click" on demand.

I assume your following this thread with avid interest. 

In a week or so I should be able to post the finished project in terms of material assembly. Thereafter you may be reading about the great magic smoke adventure, or, a brief encounter with pain. Chinese swear words are far worse than those used in the Western language. By the way and to avoid confusion, I am Anglo Saxon Brit (white still) and not Chinese although my wife is.

I am still mulling over how to test the output and I have pondered the idea of putting an earth stake into the ground near the roots of my mango tree, and then wrapping the other  transformer "contact" end around the tree trunk. I am rather concerned that the tree might jump out of the ground and attack me.

Unfortunately there are no bears, crocodiles, wolves or such like around here otherwise I would be tempted to tie one end to a cattle prod and chance my luck!

Dave

[oztules]

I have a box of 2w 47ohm resistors. I usually put 3 or 4 in series for 200R,... then put them directly on the output of the microwave tranny.
This usually gives you  2 or 3 pulses which yo will hear, and then by about the fourth, the enamel will start to get blasted off the resistor casing.... starting with blue fingers of spark walking across the surface, and quickly getting to loud bangs, big blue sparks, that you can hear 100 feet away as you approach it.

If you have a fence tester with current abilities, you will probably see 6-8000 volts at 40 amps or so, if you use more cap, then you get up in the 10000@40A and beyond.... my tester only goes to 40A and 13000v

After a minute or so, it settles down to big blue sparks across the resistor bodies, and parts of the enamel blown off until the air takes a lot of the energy as plasma dances along the sides of the resistor.

At this point, you can bet it is working well enough to keep any beast at bay.


I think I'm just misguided somehow.




..............oztules

[David HK]

Oztules,

I have been studying the main transformer windings in advance and have stumbled across a question that I cannot find the answer for via Google.

The primary winding in your notes is Bifilar, 9 turns, 2-in-hand 1.8mm diameter. The transformer bobbin I have is 51.5 mm wide so that offers a single pass of wire at 14.3 turns. However, that is bare wire only, so with enamel on the 14.3 turns could easily drop down to ~ 9 turns.

I have checked your hand drawn circuit sketch and cannot see any special reference to a “start” point for either wire, Am I correct to assume its just two wires with each end of the pair being soldered to the relevant tab on the transformer?

Secondly, I have oodles of 1.6mm wire given to me some years ago which is in want of a job. Could I go with 1.6mm or is 1.8mm essential?


Dave

[oztules]

If your not going for broke, and just want to run 30km of fence.... then a single 1.6 would do fine. It does not seem to matter how you hook the wires up on the primary. You can use two in hand, and get lots of power, or you can series them, and drop power, and widen the pulse width.

The pulse width will be a product of the impedance of the primary. The bigger the fencing wire in there, the skinnier the pulse, the higher the current and voltage, as the rate of change .. discharging the cap will go up markedly with lower and lower impedance in the primary.

A very skinny pulse will show up as a very high output from the unit. but will degrade faster with distance from the unit, as the inductance and capacitance of the fence and general impedance of the  surroundings will have a greater impact.

The closer to a sine wave you get, the better the transmission, and lower noise.... so it is all a compromise.... as usual. The toroid and the 2uf caps help shape the wavr wider just for this purpose.

two in hand of 1.6mm will be fine if you can fit it in, but one will be good too.



............oztules

[David HK]

So there are four choices:-

2-in-hand x 1.8mm diameter wire; or, one single run of 1.8mm diameter wire, or

2-in-hand x 1.6mm diameter wire; or, one single run of 1.6mm diameter wire. each of the foregoing to occupy the bobbin width or thereabouts without over lapping.

I have written this out in the long winded way so that I understand it.

You mention "in series".  What exactly does this mean? This is the only part that confuses me.

Dave