EEVblog #1277 - Electric Fence Controller Teardown

[EEVblog]

Random Bunker item teardown. An Australian designed and made electric fence controller.
Repair guides are here: https://pakton.com.au/support.php

[tautech]

Battery fence units typically are permitted to produce up to 10 kV and mains units with considerably more drive and therefore more Joules output only 5-7 kV.
This unit rated at 12 Joules is quite high output for a battery powered unit and you can see from its 12V supply current draw rating even big LA batteries won't last terribly long without solar support.

I've worked on many fence units and 30-40 uF 900 V Plessey pulse grade caps paralleled are the norm in these. Six are the most I've seen paralleled up. Over decades of use they degrade and these pulse grade caps need replacing.
Mains units use rectified AC, downunder all 230VAC so their working voltage is 325 VDC.
Circuitry is normally constructed in a way there is always another pulse after power is removed to discharge the caps.

MOV's on the output are most likely for lighting protection which fries many units.
3 terminals on the outputs are common, ground, 1/2 power and full power and in the case of this unit capacitive coupling that we don't see in NZ.....the home of electric fence units !
Here they use a resistor ladder to limit current to the 1/2 power output.

When ON and working properly there is a substantial clunk each time the caps are discharged into the primary, normally ~1 pulse/s.
Oh, and another little gem I found when trying to source pulse caps was they were being made in Rotorua NZ....not sure if they still are.

[johnlsenchak]

Differently    want to see a reverse  engineering  of that  shock  unit

Those capacitors are   seriously  "Sex on a  Stick"

Enjoy  your vacation in  Tasmania  !

[Razor512]

If trying to keep a place secure, can't they make upgraded models that would work on keeping criminals out of private property?

12 joules may not do much, other than give a slight tingle, but 2000-3000 joules is enough to temporary disrupt muscle use enough to make it so that someone is unable to climb a fence effectively while being shocked, or at least slow someone down long enough for the police to arrive.

[tautech]

If trying to keep a place secure, can't they make upgraded models that would work on keeping criminals out of private property?

12 joules may not do much, other than give a slight tingle, but 2000-3000 joules is enough to temporary disrupt muscle use enough to make it so that someone is unable to climb a fence effectively while being shocked, or at least slow someone down long enough for the police to arrive.

Nope, voltage is everything when attempting to modify behaviour.

Commercial agricultural units max out ~60 Joules but use smart technology to restrict the single pulse energy supplied. My unit here on a farm in NZ is some 40 Joules and you won't come back willingly for a second zap !
No freaking way !:scared:

But hey, it only lasts for 3ms so come be my guest and come try it !
Great for keeping kids in....or out ! 

[NiHaoMike]

If trying to keep a place secure, can't they make upgraded models that would work on keeping criminals out of private property?

12 joules may not do much, other than give a slight tingle, but 2000-3000 joules is enough to temporary disrupt muscle use enough to make it so that someone is unable to climb a fence effectively while being shocked, or at least slow someone down long enough for the police to arrive.

2kJ would be a bit overkill given a defibrillator is on the order of 300J.

[JustMeHere]

How do these things survive a thunderstorm?

[tautech]

How do these things survive a thunderstorm?

Not well if one of the fence wires takes a direct hit. You can install lighting arrestors with spark gaps and such but they're no guarantee the unit will come away unscathed.

Decades ago we had an overland 11 kV line drop a phase onto our fence in a storm and it fizzed most of the traces on the output PCB yet with a cleanup and bodge wires I got it going again. 
I doubt you would be so lucky with MV's ! 

[SparkyFX]

When i was a kid and wanted to find out if the fence was hot, taking a green blade of grass and barely touch the fence with it at some distance gave you an idea if it was safe to touch.

[EEVblog]

UPDATE: Pakton have responded and unfortunately don't want to release the schematic for the MTB200. They are actually a bit embarrassed by this old unit and think "the plastic case was horrible, we did not own the tool, Daken did".
They have offered a more modern JVA brand unit for a loan teardown and are happy to answer questions on operation etc.
The designer (Paul) has chimed in with a few comments:

"That unit was one of the first prototypes of that model, hence the "not for sale" on the back. I designed that unit. The PCB and firmware were all Pakton. The only thing owned by Daken was the case. At that time Daken did not even have a trained technician let alone an engineer.

JVA is our in house brand. Pakton is an OEM designer and manufacturer, we make and badge product for a lot of the other brands in the industry. 90% of what we make is exported, we are one of the larger manufacturers in this niche in the world.
It was a very reliable model. The output transformer is our most reliable by far. It was designed by myself in conjunction with an incredibly talented engineer in NZ called Bob Smith of Marque Magnetics, since deceased.

One of the key new features in this energiser design (when it was released) was the moving of the filter capacitor (usually across the primary) to the output. I had a patent on that. This improved the pulse shape without increasing I^2R losses in the primary, it also allows the leakage inductance of the transformer to work with the filter capacitor (or more accurately it's reflection in the primary). The goal was to produce a soliton pulse shape when coupled to a real fence. We did a lot of testing on long fences in western Queensland.  Real long fences are mostly capacitive in load. I.e. lots of reactive power flows. We have since patented methods of measuring the real versus reactive so we can isolate faults. Some of our monitors can pin point a fence fault to within 1/1000 on the fence. See the ZM20.


The move in electric fencing now is all towards monitoring, farmers do not have time to test their fences, and it's critically important to know if there is a fault on an electric fence if it is being relied on to keep stock in or bad guys out."

[SeanB]

Here the most common electric fence is for property protection, and is regulated legally to no more than 8J per pulse, and a 1Hz pulse repetition rate.

Most common failures I see are the pulse transformer shorting out, almost always the final row of turns shorting, as that is the highest voltage and also the highest dVdt as well. Second is the MOV units going low resistance, and as they are there to define the peak voltage, and also have to absorb all the energy of the pulse, you have to use multiple 800V units to get the energy absorption ability, a single unit would be expensive and unreliable. Lightning does cause problems, but if installed correctly, with the required number of grounding rods at the correct spacing, they will survive local near hits with no damage, though you will often find that the power supply to them has blown a fuse.  you can not say the same for your local electric supply though, or your phone lines if you still have them.

Almost all of the energisers use a battery, as the first couple of cycles on the internal charger is driving a short circuit, so it does draw a heavy current, in the order of 10A for a very short time, to start charging that capacitor. Might be 1A average over the cycle, but those do not run without a battery, even though the battery charger built into them is capable of delivering 3A, the unit will trip out with low battery.

I have taken apart the old analogue only units, and the LED in the optocoupler is used with the phototransistor to charge a small capacitor on the controller, and this voltage is then used to drive the voltage display, generally a set of comparators driving LED's to show low, med and OK, with the capacitor being discharged by a small transistor before the next pulse. Modern units have the optical path as a box in the case, or as shielded plastic partitions between lid and base moulding, so there is a large gap, as that optocoupler is running at the fence voltage, and no optocoupler cheaply available will handle 10kV pulses.

Designed to be serviced, the parts that fail most commonly are either plug in, or a separate board, or are on large pads so you can easily replace them, and almost all faults are on the high voltage side, the low voltage side is quite reliable and rugged.

[tautech]

Good on them for getting back to you Dave. 
Found their website:
https://www.jva-fence.com.au/
I seems they make more niche products than real grunty units however each have got considerably more complex over the years beyond what they once were.

That has added some difficulty with repairs where once they were very basic ....simply, charge caps, dump charge into the primary and pipe the secondary output to the fence.
At one time the mains supply wasn't even interrupted when the pulse was triggered like they are now.

In their simplest form you only need a 1s oscillator driving a well snubbed SCR to dump the charged caps across the transformer primary.
I've seen all manner of 1s oscillators used from UJT's, 4000 series IC's and micros with buffer bipolars switching the SCR.

A Kiwi devised probably the first electric fence some 90 years ago and his company is still going strong.
https://am.gallagher.com/nz/about-us/gallagher-story

[NivagSwerdna]

Dave attaches the terminals together using a load.... I'm not sure that's right.  When I was a kid we used to put the fence controller (and that was rather overstating the technology) such that one side was grounded and then the other side went to the rope which had metal strands through it.  So normally it would be pretty open circuit.  When an animal touched the fence it would touch the wire and make a path through ground.

Had much fun with electric fences... showing off to the urban kids with a bit of grass... and a happy memory of accidentally sitting down on some excess wire coiled on the ground only to find it was live. 

So... the resistance between terminals should be one cow + the conductivity of soil back to the unit.... big number ohms?

[tautech]

Dave attaches the terminals together using a load.... I'm not sure that's right. 

It's a completely valid test and fence units do have a 'voltage vs load in ohms spec' in their manuals.
Here in the manual for my PEL 36J on P9 are the load specs:
https://pel.co.nz/sites/default/files/20%2C%2036%2C%2063%20J%20Mains%20Energizer%20User%20Manual%20%28EN%20ES%20PT%20FR%20DE%20SV%20NL%20DA%29_0.pdf

[firewalker]

Are those things send high voltage pulses all the time? Is there any monitoring of the fence? How?

Alexander.

[tautech]

Are those things send high voltage pulses all the time?

Yep @ 1Hz repetition however some units have settings for faster and slower rates depending on whether stock are trained for fences or not.

Is there any monitoring of the fence? How?

Traditionally, no, they were just dumb units sending out pulses.
The unit Dave has, has a display that gets its info fed back from the HV side via a custom opto.

Some modern units optimised for security needs have triggered outputs for alarm purposes should the unit sense a single large current draw resulting from someone shorting the fence or getting a good shock. Once you could get these monitoring units to add onto a security installation but now there are many inbuilt variants available with LAN or cell/sim card options.

[NivagSwerdna]

https://pel.co.nz/sites/default/files/20%2C%2036%2C%2063%20J%20Mains%20Energizer%20User%20Manual%20%28EN%20ES%20PT%20FR%20DE%20SV%20NL%20DA%29_0.pdf

That's interesting.  The extra wire to provide a return... I guess our soil is conductive enough for short runs at least.

[tautech]

https://pel.co.nz/sites/default/files/20%2C%2036%2C%2063%20J%20Mains%20Energizer%20User%20Manual%20%28EN%20ES%20PT%20FR%20DE%20SV%20NL%20DA%29_0.pdf

That's interesting.  The extra wire to provide a return... I guess our soil is conductive enough for short runs at least.

Ground wire return systems need only be used in high security installations and in very dry climates.
Here in NZ for animal control ground wire return systems are rarely used instead several ground pegs are installed deep into the water table and soil conductivity does the rest. 
Then there is just one 'hot' wire on a fence and the risk of shorts to the ground return wire over many km's of fencing is much reduced.

We only hope animals don't learn to wear rubber boots ! 

[texaspyro]

Had much fun with electric fences... showing off to the urban kids with a bit of grass... and a happy memory of accidentally sitting down on some excess wire coiled on the ground only to find it was live. 

Two other fun fence things:

1) get a noob to piss on the fence

2) park you truck with the front bumper touching the wire and wait for your passenger to get out of the truck first

 

[rrinker]

 My dog learned the hard way. Were visiting my GF's parents, and her nephews had a cow in the field that's part of their property, fence energized. Big little tough dog goes tearing after the cow (dog is a pug), ran under the fence - not quite short enough to clear the bottom wire. Last time she went in that field. She stays away, even though that was a couple of years ago, they no longer have a cow, and the fence is not only disconnected, it's been removed. She still won't leave the yard and go into that field.

[tautech]

My dog learned the hard way. Were visiting my GF's parents, and her nephews had a cow in the field that's part of their property, fence energized. Big little tough dog goes tearing after the cow (dog is a pug), ran under the fence - not quite short enough to clear the bottom wire. Last time she went in that field. She stays away, even though that was a couple of years ago, they no longer have a cow, and the fence is not only disconnected, it's been removed. She still won't leave the yard and go into that field.

Yep few dogs ever go back for seconds.
It can be a real problem with working dogs too as after a belt or two they get real fence shy especially if electric fences have been introduced during their working life.
Dad had one that would happily slip through conventional wire and batten fences but after getting a few shocks would run several 100 yards to a gateway and crawl under the gates where he knew he wouldn't get zapped.
Now we just don't train them to slip through fences and instead wait to be shown a place where to get through.
Good dogs are smarter than most give them credit for and also have excellent memories as you've discovered. 

[CatalinaWOW]

As a kid visiting the family farm I had a number of run ins with these things.  Because there were long fence lines and getting from one field to another required a long detour, or rolling under or stepping over the electric fence.  Rolling under was the cowards way out, and was dirty.  Stepping over involved obvious risks to tender parts.  Have no idea of pulse power or voltage, but am quite sure that back then they were totally unregulated.  The deterrent factor is very strong.  One time I was walking along a fence line and felt some pain in my arm.  I still remember very clearly thinking that it was OK because I knew that fence wasn't electrified.  A few steps later the pain convince me to look and I realized I was dragging my arm along a barbed wire fence.  But I was so worried about the electric fences that anything less didn't require attention. 

The ones from that era were extremely simple.  Basically a bimetallic strip oscillator similar to old school turn signal flashers interrupting the current in the primary of a step up transformer.   The cheap ones drew current from a battery.  An upgrade would be a very simple DC supply.   

Everywhere I have been in the US they use a single wire for the boundary fence and depend on ground conduction for the return path.  That includes drier areas like west Texas, New Mexico, Arizona and the like.  As I said, I can testify that it works quite well.  Few animals will mess with it.  Occasionally a bullheaded and heavily furred dog with go back for seconds, but even they will put a lot of effort into avoidance.

[NivagSwerdna]

The ones from that era were extremely simple.  Basically a bimetallic strip oscillator similar to old school turn signal flashers interrupting the current in the primary of a step up transformer.   The cheap ones drew current from a battery.  An upgrade would be a very simple DC supply.   

That sounds familiar... Ours was basically a metal box with an integral metal stake and a small lead acid (motorcycle sized) battery and would make a ticking noise.

[Hydron]

Most fence energizers I'm familiar with use a similar system - charge a bunch of caps to a moderately high voltage, then dump into a specially designed pulse transformer via an SCR. The SCR can get quite chunky - to the point you're bolting on your connections with a spanner, none of this TO-247 rubbish

The transformer etc is somewhat lossy - it's not an easy optimisation problem when you need a high power pulse once a second with physical volume and cost limits. As an example, see the following high power units from Tru-test/Gallagher:

https://www.stafix.com/en/product/m63rs-mains-energizer (made in Auckland by Tru-test, 63J output)
https://am.gallagher.com/global/product/6585/fence-energizer-m10-2c000i (also NZ made, 71J output)
These are both ~70% efficient if you look at the stored vs output energy.

Those products are probably running above a sensible limit for a single unit - at some point you're better off splitting fence systems into smaller segments and maintaining (trimming grass touching the wires etc) them better than just throwing raw power at the problem, but there is always a certain segment that it appeals to.

An interesting alternative topology can be seen in this patent:
https://patents.google.com/patent/US7893521B2/da
Here it's charging the pulse caps in parallel and discharging in series using IGBTs, to eliminate the pulse transformer (mains isolation is in the charging circuit) and giving more pulse shape control. Cool idea, but not sure if it went anywhere commercially (possibly because it just perpetuated the high energy insanity beyond where it was still vaguely safe).