Electric fence

[tautech]

A strong enough fence becomes self-mowing - blades of grass touching the fence will get killed off. From memory this works for energisers of 18J or more. But you need to start with short grass!

As for parallel vs series: an electric fence is a high-voltage, low(ish) current device so capacitance/leakage matters at least as much as resistance. So total length still matters whether series or parallel - capacitance and leakage are the same in either case, but fence resilience and resistance are better with parallel wires.

eg 1km of 2mm diameter wire, 0.5m above ground, has a capacitance of roughly 9nF.  It takes almost half a joule to charge this to 10kV.

Which is the max for battery powered energisers.
Mains energisers are not permitted to create voltages to 10kV.

[Marco]

Are electric fences all lossy transformer based or do some also charge/discharge them through an inductor with energy recovery?

[tautech]

Are electric fences all lossy transformer based or do some also charge/discharge them through an inductor with energy recovery?

Typically mains energisers are ~20x step up laminated transformers fed rectified mains via 900V rated pulse grade caps with 3ms pulses.
Big energisers draw some 40W but many are less.

[Marco]

Typically mains energisers are ~20x step up laminated transformers fed rectified mains via 900V rated pulse grade caps with 3ms pulses.
Big energisers draw some 40W but many are less.

Absent weeds and animals, most of the energy seems to be heating up the transformer.

[twospoons]

Mains energisers are not permitted to create voltages to 10kV.

Its more of a practical limit than a legal one. Beyond 10kV insulating the fence wires becomes problematic. When I was designing these things all that mattered was adhering to the current vs pulse length limits.  Mains vs battery is not a concern re output voltage.  The only difference was a requirement for isolation, which was 30kV (from memory).  And we did have a mains energiser that put out just shy of 10kV.

Are electric fences all lossy transformer based or do some also charge/discharge them through an inductor with energy recovery?

The transformers are not 'lossy' as such, just that there's nowhere else for the energy to go if there's no load on the fence.  Energy recovery is not allowed as it would put the energiser back into a state where it could deliver another pulse too early  - the upper limit is one pulse per second. In the big energiser I designed the pulse energy wasn't fixed - it was kept at a lower level to reduce heating, and only ramped up when the fence came under load.

[tautech]

Typically mains energisers are ~20x step up laminated transformers fed rectified mains via 900V rated pulse grade caps with 3ms pulses.
Big energisers draw some 40W but many are less.

Absent weeds and animals, most of the energy seems to be heating up the transformer.

With a 3ms pulse ?
Here, almost all electric fences are ground return so a pulse into an open circuit hardly presents a significant load to heat anything up, except anything that touches it !

I'm not aware of any vented energisers which indicates there is little or no heating in energiser circuits.

[tautech]

Mains energisers are not permitted to create voltages to 10kV.

Its more of a practical limit than a legal one. Beyond 10kV insulating the fence wires becomes problematic. When I was designing these things all that mattered was adhering to the current vs pulse length limits.  Mains vs battery is not a concern re output voltage.  The only difference was a requirement for isolation, which was 30kV (from memory).

For whom may I ask ?

It's some 40 years since we installed a now quite small Gallagher energiser and over subsequent years asked far to much from it, as like spider webs, fence systems continue to grow.
I do remember 11kV overland lines breaking and falling onto a feed line some 2km from the energiser which turned all the traces on the output PCB into fuses. 
Some cleanup and bodge wires fitted to replace the vaporized traces and it went on to give some additional years of service until deemed too small to do what we asked of it.

[twospoons]

I was working for Staffix, Gallagher's main competition.  Later they were bought out by Trutest.  Staffix had a rather nice semi-resonant drive circuit that created a very tidy half-sine pulse on the output.  Farmers seemed to like them at Fielddays - usual comment was they felt "hotter". I wasn't brave enough to try, considering even the little 0.1J strip grazers could hit pretty hard.

I do remember 11kV overland lines breaking and falling onto a feed line some 2km from the energiser which turned all the traces on the output PCB into fuses. 

Ha! You should see the result of a lightening strike - a box full of charred lumps and nice copper plating on the inside of the plastic case.

And the transformers do get hot. All the stored energy has to go somewhere on each pulse - and it ends up  in the tfr primary if theres no load.

[Marco]

I'm not aware of any vented energisers which indicates there is little or no heating in energiser circuits.

Where is the power going then? Leakage? EM? The resistance of the wire doesn't matter much AFAICS.

[twospoons]

Venting isn't necessary. The casing is large enough to dump 20 to 30W without the internals getting cooked.  An 18J energiser is typically storing around 25J, so at one pulse per second thats ~25W going into the case when not loaded. Its not a lot in a big box.

[tautech]

I was working for Staffix, Gallagher's main competition.  Later they were bought out by Trutest. 

Ah yes, a trade name now consigned to NZ history. We had a Staffix energiser some years back but would hardly call them main competition to Gallagher which IMO is PEL....at least in the larger mains powered energisers we use.

[twospoons]

PEL got bought up by Trutest at the same time Stafix did. Stafix branded energisers are still being sold by RD1, so the brand is not gone.  Looks like the case design hasn't changed much in 25 years.

[tautech]

PEL got bought up by Trutest at the same time Stafix did. Stafix branded energisers are still being sold by RD1, so the brand is not gone.  Looks like the case design hasn't changed much in 25 years.

Yes, the Trutest group is certainly good competition to Gallagher....sorta forgot who is now owned by who.

Over the years we have had products from all of them and our current energiser was the 2nd largest from PEL when we got it a few years back.
We had the option to get the remote control for it but our layout stems from the days when such features were not available and has many pre-existing branch switches to simply isolate fence sections for fault finding and/or repairs.
For our switches we use old powerco pole fuses, the now rarer 1/4 turn glazed ceramic types. These don't give shocks in bad weather when wet like the flimsy commercial types. 

Some years back I did some research into replacing aged pulse grade caps to give weak energisers a new lease on life and quickly found 900V ~40+uF pulse grade caps are a speciality product and IIRC a crowd in Rotorua was one of the main manufacturers but the cost..... 
Now I have a selection of old energisers primarily for a small stock of their better pulse grade caps which I should really harvest and store and get rid of the old electronics.....then I should also harvest their SCR's but where do I stop !

[tautech]

Quick question if I may twospoons; totally dead energisers, typical failure points ?

[twospoons]

Back when I was doing repairs it was usually the diodes across the storage caps that blew - the ones that prevent reverse charging. There was a bit of a design flaw there, using two diodes in parallel.  They would fail short and stop the caps from charging.  Second most common was the control board getting zapped by a minor lightning flash over on earlier models that didn't have an internal diverter board to send that spark to ground.  This was only the Stafix units. We didn't fix anyone else's product.

I seem to recall we were using 15uF and 20uF 1200V film caps from Plessy. Their pulse rating was ... dubious, as I found out after destroying a bunch after a week running in  a prototype. After talking to them about what we were doing they quoted an I²t rating that was half what was on their datasheet. I guess a 10ms pulse is not in the same league as a 100us pulse with 10x the current.

[tautech]

Thanks, not seen much in the way of diodes across those pulse caps but drags out an old unit that certainly does have the odd diode on a PCB that presses down onto a string of caps with 6mm press terminals and a total of 6 in this one, all 900V 30uF Ducon from OZ.
Plessey caps are certainly the more common, also from Aussie.

[twospoons]

Internally the Stafix energisers looked like this:

C2 is half the size of C1, D1 prevents reverse charging of C1, and D2 was to clamp the primary to prevent ringing and create the half sine pulse.
D1 was a weak point - the diodes were a bit underrated for the job, and would sometimes explode just after being replaced.

[Marco]

Recovery is not allowed as it would put the energiser back into a state where it could deliver another pulse too early

You can feed it back to supply rather than the discharge capacitor.

[PlainName]

Related, I am thinking about making a kind of tester for the fence - a light which indicates correct activity (at the fence) and/or a WiFi broadcast that it's not working now.

I acquired a Hotline tester which I've verified does actually do what it says it does, and that shows the approach I would like to take. But how does it work? I know I could just take it apart but the theory of operation would be more useful than trying to figure out a specific implementation...

[twospoons]

Most likely a capacitive divider, using the users hand as the ground end. Then you have a little MCU to measure the divided down pulse and drive the LEDs.  Its only likely reading to 10% precision (1kv out of 10kV), since thats all you really need for a fence. 

[PlainName]

10% precision would be more than good enough for me.

AFAIK, holding it between two fingers doesn't make any different to grabbing it as shown, which I would have thought would affect the capacitance. Since it's a pulse, could there be a decent capacitor on board, behind a low-pass filter, that acts as 0V (rather than ground)?

Hmmm. I can feel some microcap usage coming on

[Marco]

Looking at a repair video I don't see any grounding pad from the button, so it seems to rely on the capacitance from the PCB ground plane to your hand. With a resistor divider to the ground plane from the look of it.

[PlainName]

Ah! Couldn't find a teardown originally, but since you mention a video I made an extra effort knowing it exists and found it 

[Marco]

It has a micro, it might actually be far smarter than we are assuming. If you know the pulse waveform (which they do if used with their fence energiser) then you can determine the parasitic capacitance to ground with a bit of math from the shape of the measured waveform (the RC network low pass filters it). Once you know the capacitance, you can determine the pulse voltage from the amplitude.

[PlainName]

It's been accurate (within the 1kV steps) for two completely different energisers I've tried, one 8.5kV and t'other 10.5kV, neither from a company related to it. I will have it apart this weekend to take a closer look