Mineral oil breakdown voltage

[MRMILSTAR]

I plan on using food-grade mineral oil for my CW multiplier project as insulator. For purposes of parts separation I am assuming that the breakdown voltage of mineral oil is 10 KV/mm. Does this seem about right?

[klugesmith]

First caveat: in oil, just like air, sharp edges and points concentrate the E-field. 
Breakdown voltages given in kV/mm are for slightly convex electrodes much wider than the gap distance.
They need to be derated for less favorable geometries.

10 kV/mm sounds like a good ballpark value for dry, gas-free mineral oil, in my unpracticed opinion.
In the trade, insulating oils are specified and tested according to IEC and ASTM standards, which all use AC voltage.
It seems reasonable to expect 40% higher breakdown voltage in DC tests. 
Anybody here know better?  How about AC at tens of kHz?

Here's a brochure about an oil tester,
https://www.instrumart.com/assets/Megger-Oil-Testing-Datasheet.pdf

and a snip from an oil datasheet, in which the density values look unrealistically high (by factor of 10^9).



Personally I have plan to test any HV insulating oil for hobby use.  Can a 1 mm thickness hold off NST voltage?  Commercial BDV testers immediately turn off the voltage when breakdown occurs.  It wasn't hard to find cabinet door knobs with curvature matching the test standards.  Here shown with a caliper that is resistant to oil and vinegar.

[MRMILSTAR]

Thanks for the info Rich. There's definitely something wrong with that density number. Maybe they were thinking of neutron star material! 

I need to pack the following into a 4" x 36" tube to achieve a reasonable size and weight for my 14-stage CW multiplier.

* 28 doorknob capacitors (1.88" x 0.88" each)
* 28 HV diode strings (3 diodes per string)
* 28 HV resistors
* About 2 gallons of oil

I have been doing some preliminary layout and it looks like everything will fit. I estimate that the weight of the tube, components, and oil will be about 30 pounds. Its going to be a tight fit so I have to carefully consider any pointy things or sharp curves. At least the tube is made of clear PVC, actually tinted light blue, so I will be able to see if there is any arcing.

I really hate the idea of using oil because its so messy but I don't see any alternative if I want something that is reasonably compact, portable, and not terribly lossy. I thought about using paraffin wax instead of oil but then it would be practically unservicable and I wouldn't be able to see anything.

[johnf]

I would use Shell Diala oil
It is very cheap new about US$1.00 per litre
or US$0.60 for rerefined

Shell_Diala_OilHandbook_e.pdf

[MRMILSTAR]

I would use Shell Diala oil
It is very cheap new about US$1.00 per litre
or US$0.60 for rerefined

Shell_Diala_OilHandbook_e.pdf

Do you know if it is compatible with PVC? Where do you get it? I found a 55 gallon drum on E-Bay for about $1000. That's a little more than I need and little more than I want to pay. 

[johnf]

That price is outrageous
you could probably get it for free from a scrap yard that deal with electricity supply equipment transformers circuit breakers etc
But here in NZ I have bought it recently hence I know the price from our local re-refiner who also tests new oil drums for the local electrical supply aurthority.
My need was to repair the HV tasnk on a SEM where the diodes had gone short so I decided to replace the oil while I had the thing apart

[klugesmith]

Let us know what you find, Steve.

I have 1 gallon of Diala Ax that was given to me years ago.
By a generous online forum member named Carl Pugh, in the next county, who had procured a whole 5 gallon pail for his HV transformer project.

Lots of people sell hard-to-get materials online at ridiculously high prices.
I just encountered that while looking for some sodium.  More than a dollar per gram in small quantities. $90 for a pound shipped from India. 
Huh?  On a volume basis, in industrial quantities, sodium is the cheapest metal on the planet.  Barely half the price of new steel or aluminum, for example.

[davekni]

Steve,

Have you considered air insulation?  With careful layout, I think that could be possible.  The diode sets would need some local insulation, such as the adhesive-lined heat-shrink tubing over each string of three diodes.  Their cases tend to be too short to handle full voltage without additional insulation.  The door-knob capacitors should be able to handle the voltage as long as their surfaces are clean and dry.  At least they have in the bit of playing I've done.

If normal 100kPa air isn't quite sufficient, how about making your container air-tight, and pressurizing it to say 300kPa absolute (200kPa gauge if you are close to sea-level)?  Or, if you can find some, fill it with SF6.  Either option would provide about 3x breakdown voltage compared to atmospheric pressure (100kPa) air.

If you do go with mineral oil, it's going to be much higher dielectric breakdown strength than air.  I doubt you need to be concerned about the exact voltage for any specific oil.