N00b alert: Reducing High, Pulsing Voltage

[John154]

Hi all,

I am trying to reduce a pulsing voltage (coming from an electric fence charger) of 5kV down to 5V. I already tried using a voltage divider, but the voltage is coming out very inconsistently—anywhere from 7V to 170V. I found the schematic attached (from a thesis paper) but I'm not too sure how to set it up. I only want the voltage detector, I am not interested in the pulse detector. I tried setting up the circuit by placing the 10M, 10M, 1M, and 6.8k resistors in series, but I am still getting the inconsistency issues.

Any help you could give would be much appreciated. Sorry if I'm missing something extremely simple, I'm very inexperienced and just looking to learn.

 

[berkeman]

Hi all,

I am trying to reduce a pulsing voltage (coming from an electric fence charger) of 5kV down to 5V. I already tried using a voltage divider, but the voltage is coming out very inconsistently—anywhere from 7V to 170V. I found the schematic attached (from a thesis paper) but I'm not too sure how to set it up. I only want the voltage detector, I am not interested in the pulse detector. I tried setting up the circuit by placing the 10M, 10M, 1M, and 6.8k resistors in series, but I am still getting the inconsistency issues.

Any help you could give would be much appreciated. Sorry if I'm missing something extremely simple, I'm very inexperienced and just looking to learn.

Welcome to the PF.

You could just use an off-the-shelf HV probe like the one I link to below. It has a 1000:1 divide ratio, and reasonable bandwidth for low-frequency measurements.

Do you know what the nature of the 5kV waveform is? When you make both DC and AC measurements with your meter, what do you read off of your resistor divider? Do you know how often the pulses happen? You might be able to just add a 1uF non-polar capacitor in parallel with your bottom resistor to get a smoother DC reading. You can also put a 1N4148 diode in series between your last 2 divider resistors (the diode cathode connects to the top of the 6.8k Ohm resistor) to rectify the waveform to give you a better DC voltage.

What are you planning on doing with the 5V, BTW?

http://www.mouser.com/images/fluke/lrg/80k6_02h_p.jpg

 

[John154]

Welcome to the PF.

You could just use an off-the-shelf HV probe like the one I link to below. It has a 1000:1 divide ratio, and reasonable bandwidth for low-frequency measurements.

Do you know what the nature of the 5kV waveform is? When you make both DC and AC measurements with your meter, what do you read off of your resistor divider? Do you know how often the pulses happen? You might be able to just add a 1uF non-polar capacitor in parallel with your bottom resistor to get a smoother DC reading. You can also put a 1N4148 diode in series between your last 2 divider resistors (the diode cathode connects to the top of the 6.8k Ohm resistor) to rectify the waveform to give you a better DC voltage.

What are you planning on doing with the 5V, BTW?

http://www.mouser.com/images/fluke/lrg/80k6_02h_p.jpg
View attachment 204355

Thanks for the response. It is 5kV DC and pulses are happening every half second. I have a zener diode, could I place that in series instead of the 1N4148? My plan is to feed the 5V into a microcontroller (Arduino) and then use the microcontroller to sense whenever there is a fault in the electric fence.

 

[berkeman]

Thanks for the response. It is 5kV DC and pulses are happening every half second. I have a zener diode, could I place that in series instead of the 1N4148? My plan is to feed the 5V into a microcontroller (Arduino) and then use the microcontroller to sense whenever there is a fault in the electric fence.

If all you have is the Zener diode, you can use it temporarily until you get a regular diode. What is the Zener's reverse breakdown voltage? Still put its cathode connected to the top of the 6.8k Ohm resistor.

Is your Arduino running off of 5V or 3.3V? In either case, you will want to clamp that voltage sense input between ground and Vcc to the Arduino with protection diodes (the board may have them already on that input connection point).

 

[jim hardy]

Ordinary resistors have a maximum voltage rating , typical is 400 volts. It's not their power dissipation limit but a dielectric breakdown limit.

To make a whole watt in 10 megohms would take 3162 volts. So it seems safe enough from power handling perspective. But at just 16 milliwatts you hit their 400 volt limit.

What kind of resistors did you use?

To withstand 5KV you'd need about twelve 400 volt rated resistors in series.

I suspect your two 10 meg resistors are arcing through. If you had more current available they'd explode, like the picture in this thread

https://www.physicsforums.com/threa...help-me-identify-a-burned-up-resistor.915397/

Look carefully at your resistor datasheet and find their maximum voltage rating.. Did you exceed it ?
If so, go to the manufacturer's catalog and see whether he makes resistors for higher voltage work.
Or series up twenty one-megohm resistors, they're cheap enough and it'll be safer.

If you did use high voltage resistors, i'd suggest run it at night and look for sparks.

 

[jim hardy]

This from Ohmite's resistor selection hints at http://www.ohmite.com/techdata/res_select.pdf

Vishay datasheets show "Maximum Working Voltage" and make clear that you must use the lesser of power dissipation or working voltage limit.See "Critical Resistance Value" and "Maximum Working Voltage" definitions here

http://www.vishay.com/docs/49873/49873_sg2113.pdf

Typical datsheet

That's the kind of notes you'll look for in your resistor datasheets.