Measuring pulse from electric fence energizer with oscilloscope

Thread Starter

icecats

Joined Feb 20, 2023
27
Hi,
I am trying to measure the pulse output by an electric fence energizer commonly used to control/contain livestock. Based on this article, I expect to find "very brief (less than 3/10,000 of a second in duration), high-voltage pulses (usually 2,000–6,000 volts) of electrons down the conductor every 1–2 seconds."
Since an oscilloscope obviously can't handle 5 kV, I am using a voltage divider to get the voltage to a manageable level. I am using a high voltage 10Mohm resistor with a 2Kohm resistor for a division ratio of ~0.0002. I have the oscilloscope probe hooked up as shown in the diagram.
The oscilloscope reading surprises me and makes me question if my measurement setup is correct. It is a very brief pulse of ~22 volts that seems to resonate afterwards. 22 volts is unbelievable to me since this would imply that the energizer is outputting 110 kV. I also don't understand the resonant pattern which seems to be on too brief of a time scale to be the actual pulse. I have also tried hooking up an LED in place of the multimeter and it flashes briefly with each pulse; this implies to me that the voltage is not 22v since the LED doesn't burn out.
I would expect to find scope output that looks like the diagram on page 29 of this paper or the output shown in this video.
Any thoughts?
Thanks!
 

Attachments

BobTPH

Joined Jun 5, 2013
8,813
Did you make sure the scope and probe multipliers are set the same? It could be that the scope is set for 10X and the probe at 1X, so the voltage is actually 2.2V.
 

BobTPH

Joined Jun 5, 2013
8,813
Is the charger connected to a fence when you do this? I would not be surprised if an unconnected device produced a much higher voltage.
 

DickCappels

Joined Aug 21, 2008
10,152
Your physical layout is not conducive to getting a clean waveform. Something is ringing, but that might be because the fence looks like a transmission line (when combined with the effects of the earth.

Did you try this probe on a pulse generator to make sure the divider is properly frequency calibrated? With resistances like that compensation capacitors will be necessary to assure accurate pulse amplitude readings.

file:///Users/bob/Downloads/electronics-11-03446.pdf <== This article gets pretty deep. You can usually achieve decent compensation by adding ore or two capacitors.
 

ronsimpson

Joined Oct 7, 2019
2,988
In the resistor divider, are you certain the top resistor is rated for that much voltage. I use 5 resistors or more.
What resistor values are you using.
 

tautech

Joined Oct 8, 2019
383
Member icecats may have already damaged his scope input and for this should be using a 1000x probe.
Without a suitably rated probe please keep away from the output side of electric fence energisers.

One needs ask the question, what is the objective of this measurement ?
 

Thread Starter

icecats

Joined Feb 20, 2023
27
Is the charger connected to a fence when you do this? I would not be surprised if an unconnected device produced a much higher voltage.
No, my test setup was not connected to a fence.

Connect capacitor 1...10 nF, 10kV parallel to energizer output for imitation of fence capacitance.
Then you will see curve with right voltage and shape, like on page 29.
I will give this a try. I was thinking that a high voltage resistor would be a good proxy for a physical fence, but I didn't consider fence capacitance. Time to order some high voltage caps.

Your physical layout is not conducive to getting a clean waveform. Something is ringing, but that might be because the fence looks like a transmission line (when combined with the effects of the earth.

Did you try this probe on a pulse generator to make sure the divider is properly frequency calibrated? With resistances like that compensation capacitors will be necessary to assure accurate pulse amplitude readings.

file:///Users/bob/Downloads/electronics-11-03446.pdf <== This article gets pretty deep. You can usually achieve decent compensation by adding ore or two capacitors.
I'll give this a try.

In the resistor divider, are you certain the top resistor is rated for that much voltage. I use 5 resistors or more.
What resistor values are you using.
I am using a high voltage 10Mohm resistor rated at 10 k with a 2Kohm resistor for a division ratio of ~0.0002.

Member icecats may have already damaged his scope input and for this should be using a 1000x probe.
Without a suitably rated probe please keep away from the output side of electric fence energisers.

One needs ask the question, what is the objective of this measurement ?
What makes you say I have damaged the scope? If the voltage divider is functioning properly, I wouldn't expect to see voltages that are too high for the scope input. Additionally, I have placed in LED in place of the scope probe and it pulsed without burning out. My objective of this measurement is to understand the shape of the fence pulse as I am building a fence monitoring system. I want to characterize the fence pulse on an oscilloscope before moving to something like the STM32 for project implementation. I'm looking at doing something similar to this project.

Thanks all for the input!!!
 

BobTPH

Joined Jun 5, 2013
8,813
I will give this a try. I was thinking that a high voltage resistor would be a good proxy for a physical fence, but I didn't consider fence capacitance. Time to order some high voltage caps.
The resistance of the fence to ground should be very high, until some creature touches it.
 

Thread Starter

icecats

Joined Feb 20, 2023
27
Connect capacitor 1...10 nF, 10kV parallel to energizer output for imitation of fence capacitance.
Then you will see curve with right voltage and shape, like on page 29.
I got a hold of some 1 nF caps and put 10 in parallel to the energizer output to get 10 nF equivalent. I also adjusted my resistor divider to use 20 Mohm resistor and a 4.7 kohm resistor to get a division ratio of ~0.000235.
I have a few different views of the waveform. It certainly looks more like the curve on page 29, although more oscillatory than I was expecting. I have a couple questions.
1) I observe two distinct types of peaks, which I have labeled as primary and secondary peaks. Based on the division ratio, the primary peaks would indicate a fence voltage of ~425 v and the secondary peaks indicate ~725 v. The secondary peaks are very short duration. What do you think these are caused by?
2) Why does the voltage dip below zero after the initial peak.
 

Attachments

Danko

Joined Nov 22, 2017
1,829
1) I observe two distinct types of peaks, which I have labeled as primary and secondary peaks. Based on the division ratio, the primary peaks would indicate a fence voltage of ~425 v and the secondary peaks indicate ~725 v. The secondary peaks are very short duration. What do you think these are caused by?
2) Why does the voltage dip below zero after the initial peak.
Long wires of your setup is transmitting line with its own inductance
and capacitance, so you have decaying 2.3 MHz oscillations.
"Secondary peak" is single short train of decaying oscillations with
very high frequency. This train is generated on start of HV pulse
at short wires inside energizer and then is repeatedly mirrored
in transmitting line.

Try setup shown below:

1678458969798.png
 
Last edited:

Thread Starter

icecats

Joined Feb 20, 2023
27
Long wires of your setup is transmitting line with its own inductance
and capacitance, so you have decaying 2.3 MHz oscillations.
"Secondary peak" is single short train of decaying oscillations with
very high frequency. This train is generated on start of HV pulse
at short wires inside energizer and then is repeatedly mirrored
in transmitting line.

Try setup shown below:

View attachment 289413
Ok I will give that a try. However, given that an electric fence is basically a really long transmitting line and I am trying to simulate the condition that the energizer is connected to a fence, perhaps this is actually a good representative signal... Or maybe I'm thinking about it wrong.
This makes me question how accurately off the shelf electric fence voltmeters are able to measure.
(also I would totally test this on an actual fence except that I am currently located in the middle of the city and it is simply not practical)
 

Thread Starter

icecats

Joined Feb 20, 2023
27
Long wires of your setup is transmitting line with its own inductance
and capacitance, so you have decaying 2.3 MHz oscillations.
"Secondary peak" is single short train of decaying oscillations with
very high frequency. This train is generated on start of HV pulse
at short wires inside energizer and then is repeatedly mirrored
in transmitting line.

Try setup shown below:

View attachment 289413
I shortened my wires as much as possible and am getting similar results. I've attached two captures at various time scale zooms. Since the fence is inherently a long wire, I guess I would expect the waveform of an actual fence to be similar.
Let's say that an animal were to touch the fence. Which voltage would the animal "feel"... the peak voltage of the "primary" oscillations or the peak voltage of the high frequency secondary oscillations. I wonder which part of the waveform would be the most representative voltage for measuring and characterizing fence strength (like this commercial device does).
 

Attachments

Danko

Joined Nov 22, 2017
1,829
@icecats,
Worst case may be, probe divider or even oscilloscope input,
as said @tautech in post #8 are damaged.
Then observable decayed sine is ringing of coaxial cable after every HV pulse.
Сheck how oscilloscope reacts on its calibrating voltage.
 

Janis59

Joined Aug 21, 2017
1,834
First: oscilloscope must be handheld battery powered. For example Hantek or even smaller - Owon, Fluke, Fnirsi, or other DSO. Main task is stay as far as possible from the grounds. From oldy goodie analogue may use the C1-94, C1-95 as they have a regime be powered from +12V.
Second: never use the factory made dividers. When they touch anything faster like DC, their voltage capability is at least 10x less than noted in manual. I have the large box full of 10 kV and 50 kV probes what was killed just with the 3 kV at 100 MHz containing the China wonderous superproducts what 3 kV probe exploaded just by 300 V..... Seems, producers (and even experienced and not cheap producers) ignores the reactive current thermal effects in the divider input. Capacitors are exploding. May form the resistive divider DIY, if bit read about how to compensate the input capacitance.
 

tautech

Joined Oct 8, 2019
383
First: oscilloscope must be handheld battery powered. ............
Main task is stay as far as possible from the grounds.
Double why ? Really why ?

FYI most electric fence installations worldwide use a ground/earth return......the exact same as mains ground !
Sure you can work directly onto the energiser output terminals however you then omit an important part of an electric fence system, the return path.
In real world installations there are many possible loadings on the energiser other than discharging into the critters you're trying to contain/keep out.
They can range from grass to twigs to UV decayed insulators or even one where a nice moist bird dropping provides a discharge path over an insulator......all these add up to substantial loading in large installations where to have the energiser remain effective it needs be upsized to effectively deal with these loads and not fall on its face.
Of course all these unknowns will change the energisers output waveform and the design and reserve energy will determine by how much.

Some years back at NZ Ag shows energiser manufacturers enjoyed placing 500 Ohm resistors across output terminals and everyone knew where they were from the healthy CLACK that could be heard for 100's of meters. It was a representative load of a NZ installation which often approach 50km of electrified wire.
 

Janis59

Joined Aug 21, 2017
1,834
RE:""Double why ? Really why ?""

Never didnt explored what is the reason in-deep, but I have killed at least dozen of osciloscopes having mains zero on the input plug briding (I every day am working with 3-30 kV). The same way many was killed taking feeding via the isoleeringtransformer when input braiding had few kilovolts on it. Probably moisture, probably bad capacitors, probably some designer failure. In second case (of transformer) sure the problem was caused by parasythic capacitance between trafo bobbins. So, my receipt for absolute safety is - or use very cheap oscillo, or take battery powered one. And, avoid damn much the rapid voltage spike on the mass (case) of oscillo, that is killing them just instantly. Can only guess why. And never never never trust the marked max voltage on the probe. It lies in 120% of cases. I met NONE what was been true.

And, hench signal is very killous, having very sharp fronts and short length. Thus the minor parasythic capacitance makes very big stress.
 
Top