# Please Critique my HV Measuring Setup

At the moment, I am working on a device that will be able to measure the out put of a high voltage DC power source with an adjustable output of 0 - -60 kVDC and 0-15 mA (the typical power being 600 Watts at maximum output). However, I will need to be able to monitor the voltage being drawn by the load that it shall be connected to during operation, which will require me to build a voltage measuring probe to get an accurate readout.

Below, I have included a schematic of the apparatus that I am considering to build and will use a list of the parts that I am buying:

simulate this circuit – Schematic created using CircuitLab

The 390 M Ohm resistor shown in the image will be made of 100 3.9 M Ohm resistors connected in series. Each resistor - comprised of a carbon film - has a power rating of 1/4 watt and has a tolerance of 5% (which shall be measured upon arrival). The item may be seen here: https://www.ebay.com/itm/3-9-meg-1-4-watt-5-Carbon-Film-Resistors-100pcs/400005279803?hash=item5d222c303b:g:pbMAAOxygo9Q8MMZ.

The 200 K ohm resistor comprised of 2 resistors in series, the two being "4Pcs 5% 100W 100K Ohm Shell Power Aluminum Housed Case Wirewound Resistors." These resistors can be seen here: https://www.ebay.com/itm/4Pcs-5-100W-100K-Ohm-Shell-Power-Aluminum-Housed-Case-Wirewound-Resistor/183271207134?epid=15020349719&hash=item2aabd0b0de:g:NLsAAOSw3uFbIG-d.

The volt meter used here is a 2.0 G Ohm meter that can handle up to 500 VAC/VDC. The meter may be seen here: https://www.acehardware.com/departments/lighting-and-electrical/electrical-tools/meters-and-testers/3251113.

The entire apparatus will be submersed in mineral oil as a means of electrical insulation (to the exception of the meter-portion of the voltmeter).

With all factors considered, would the apparatus function as intended and how could I further improve it?

• what is your target measurement accuracy? – JonRB Sep 16 '18 at 18:54
• Preferably, it should be within 2 kV DC of the actual voltage. – Super Nerds Team Sep 16 '18 at 19:02
• Please stop giving us links to ebay and telling us that the seller won't provide a voltage rating. This is dangerous stuff. Find the manufacturer's data sheet and link to it. If you can't find the manufacturer's data sheet then don't use the component! – Elliot Alderson Sep 16 '18 at 20:25
• @SuperNerdsTeam You're probably getting this, but part of the issue here is that 60000V is potentially very dangerous, as you're aware, and a tremendous amount of engineering goes into designing a high voltage project that is even theoretically safe. Your third alternative is to purchase resistors that are specifically meant for high voltage, or to make your own. They aren't cheap, but neither are decent multimeters or human lives. Putting a resistor in mineral oil will only help if the resistor is using air as an insulator in the first place. – K H Sep 16 '18 at 22:36
• You're probably aware that most resistors are either a coil or a zigzag, and because it produces a voltage difference across it's entire length, if the voltage applied to the whole resistor is high enough, the voltage difference between a zig and it's neighboring zag, or between loops of a wound resistor will become great enough to jump the tiny distance between them. – K H Sep 16 '18 at 22:38

Couple of considerations:

1) Most 1/4 Watt resistors have maximum voltage rating of 300V. You should choose either specifically-designed high-voltage resistors or greatly increase the number of resistors in series. Be sure to have sufficient headroom to handle any transients.

2) There is absolutely NO need to use a pair of 100W resistors for the bottom portion of the voltage divider. A pair of 1/4 Watt resistors is just fine.

From your comments you are ideally looking for 60kV +-2kV from the true DC measurement

Consideration #1.

With 100off 3.9M$\Omega$ at 5% and 2off 100k$\Omega$ at 5% The worst-worse is outside of this tolerance.

At 60,000V worst-case maximum measure voltage is 33.988V (66,612V) and the worst-case minimum voltage being 27.825V (54,288V)

It isn't until the voltage being measured drops to just under 20kV do you meet this limit.

HOWEVER, this is worst-worst tolerance and the probability of all the resistors being biased the same and wrong way is slim. Batch characteristics would actually increase this probability as you cannot gain credit for the randomness from mixed batching (which would pull the overall value closer to the mean).

With a bit more work a distribution curve could be produced to provide some statistical information with regards to precision

• Thank you for the response. Would all of the components mentioned in my post function when used in the discussed apparatus? Would any types of resistors need to be changed out for new ones? – Super Nerds Team Sep 16 '18 at 20:01
• do you really want 2kV of the measured? – JonRB Sep 16 '18 at 20:08
• If it is possible, however, I am willing to go to a variance of upto 8 kV – Super Nerds Team Sep 16 '18 at 20:09