Homemade Capacitors Of A Mad Scientist

Once upon a time I was a real mad scientist. I was into non-conventional propulsion with the idea of somehow interacting with the quantum vacuum fluctuations, the zero point energy field. I was into it despite having only a vague understanding of what that was and without regard for how unlikely or impossible anyone said it was to interact with on a macro scale. But we all had to come from somewhere, and that was my introduction to the world of high voltages and homemade capacitors.

And along the way I made some pretty interesting, or different, capacitors which I’ll talk about here.

Large Wax Cylindrical Capacitor

As the photos show, this capacitor is fairly large, appearing like a thick chunk of paraffin wax sandwiched between two wood disks. Inside, the lead wires go to two aluminum flashing disks that are the capacitor plates spaced 2.5cm (1 inch) apart. But in between them the dielectric consists of seven more aluminum flashing disks separated by plain cotton sheets immersed in more paraffin wax. See, I told you these capacitors were different.

I won’t go into the reasoning behind the construction — it was all shot-in-the-dark ideas, backed by hope, unicorn hairs, and practically no theory. The interesting thing here was the experiment itself. It worked!

I sat the capacitor on top of a tall 4″ diameter ABS pipe which in turn sat on a digital scale on the floor. High voltage in the tens of kilovolts was put across the capacitor through thickly insulated wires. The power supply contained a flyback transformer and Cockcroft-Walton voltage multiplier at the HV side. As I dialed up the voltage, the scale showed a reducing weight. I had weight-loss!

But after a few hours of reversing polarities and flipping the capacitor the other way around and taking plenty of notes, I found the cause. The weight-loss happened only when the feed wires were oriented with the top one feeding downward as shown in the diagram, but there was no weight change when the top wire was oriented horizontally. I’d seen high voltage wires moving before and here it was again, producing what looked like weight-loss on the scale.

But that’s only one of the interesting capacitors I’ve made. After the break we get into gravitators, polysulfide and even barium titanate.


The gravitator capacitor was created by T.Townsend Brown to control gravity and is described in UK patent GB300,311. My implementation was a 30cm (12 inch) long chunk of Bondo resin with two aluminum electrode plates and 29 more isolated plates evenly spaced in between. In one of the photos you can see it under construction. It was made in two pieces, each with a white plastic mold into which a plate and resin was added. The resin then hardened, the mold was raised, then more plates and resin was added, and so on until each piece was half the length of the final capacitor. They were then glued together using more resin to produce the one long piece you see in the photo of the test setup.

The test was a horizontal one this time with the gravitator suspended as a pendulum. No movement was ever detected. However, usually when this test is done, one or both of the feed wires is a small diameter wire with a thin enamel coating i.e. magnet wire. At these voltages, that enamel breaks down easily and ionization to the air results, acting as a jet and creating some form of propulsion. We’ve seen this type of ion propulsion before when we talked about the homemade flying machines called lifters.

The movement is usually small but the experimenter typically cycles the power supply on and off in time to the movement, creating a resonance just as does a person on a swing when they pull on the ropes and swing out their legs at just the right time. The result is a big movement, but not one that has anything to do with gravity control. In my case you can see I use feed wires with insulation thick enough to avoid breakdown, and so I got no movement.


One feature that was supposed to be beneficial in these non-conventional propulsion experiments was to have a high K dielectric, one with a high relative dielectric constant. Someone back then had found that polysulfide had a K of 2260, which is very high. For most materials the K is below 10. I managed to find a polysulfide sealant called Deck-O-Seal, a liquid plastic filler for cement joints around swimming pools. The diagram and photos show what I came up with.

Initially the brass wire was immersed in the polysulfide and the whole thing was suspended on the end of a rotor arm. But with high voltage applied there was no movement. From further research I found that the polysulfide product may contain electrically conductive material and so I moved the brass wire out of the polysulfide in hopes that the air would act as an insulator. This time I got ionization at the ends of the wire in the form of a bluish corona and a hissing sound. And as with the gravitator’s feed wires above, that produced a jet and resulted in a little movement. But again, I was after movement due to interacting with quantum vacuum fluctuations and so I abandoned that one.

Barium Titanate

However, I pressed on with my quest for a high K dielectric and managed to find a source of 99.9% pure barium titanate powder from Atlantic Equipment Engineers (product no. BA-901 in case you want some). Barium titanate can have a K in the thousands if it is at the right temperature, with the right electric field strength and with the electric field in the right orientation.

But the problem is turning that white powder into a solid dielectric with no air in it. One way to do that is to compress it under heat, or to sinter it, but I didn’t have the means to do that. Instead I experimented with mixing in paraffin wax as a binder, knowing that the resulting dielectric constant would be lower than with pure barium titanate. The best I got with this was a relative dielectric constant of 12.5 to 18.6.

Then I tried with an epoxy resin as the binder. With a lot of experimenting I got best results by mixing the resin and barium titanate in such a proportion that I got balls mostly 1mm or smaller in diameter, as shown in the photo. The capacitor I was after at the time was a cylindrical one. I used a 1/4″ diameter copper rod for the center electrode and aluminum mesh for the outer one. I fashioned a mold from two pieces of a plastic tube slit lengthwise and with the copper rod running through the center. I poured a little of the barium titanate and epoxy mix at a time into the mold and tapped it well in place, while it was still soft. With 86% barium titanate by weight I got a K of 27. That was the best I could do with this method, but it wasn’t in the hundreds or thousands as I would have liked. However, it was still impressive when compared to plain resin or wax capacitors whose K is usually around 2 or 3.

Two-Dielectric Capacitor

But the barium titanate one wasn’t my most ambitious one. That honor goes to a cylindrical capacitor whose dielectric was actually two separate pieces running down the center. One piece was made of epoxy resin and the other was made of paraffin wax. Full calculations were done for the dimensions and materials to match a hypothesis produced by a theory and of course that meant I couldn’t just use whatever I had on hand. As you can see I not only filled half the capacitor’s interior with wax but encased the whole outside of it too.

With the capacitor oriented with the wax piece on top, there was supposed to be a net upward thrust. Tests were done on a digital scale and also a triple beam balance, but there was no change in weight, and yet gently placing a playing card on top produced a weight change. As you can see, the scale was completely covered in grounded aluminum foil for shielding purposes. The voltage was only 8kV before sparking happened inside the capacitor but it was enough to test the hypothesis. The theory was found to be wrong.


So while I didn’t get the type of propulsion I was after, I did have a great introduction to working with high voltage, capacitors, new construction techniques and had a lot of fun along the way. Have you made any funky capacitors or done any non-conventional propulsion experiments of your own? Let us know about them in the comments below. If you tend to stick to the more conventional, there’s a lot to learn from our article on everything about commercially made capacitors.

26 thoughts on “Homemade Capacitors Of A Mad Scientist

  1. How is K measured? I always thought that dielectric strength was measured in volts/mil (thousandth of an inch), but I suppose there’s an SI unit for that (K?).
    When I was 11 or 12 I built my own Tesla coil using a capacitor made from aluminum foil sandwiched between panes of window glass. It worked well at 7500VAC.

    1. The K is the relative dielectric constant, which has no units. It’s the ratio of the capacitance of the capacitor’s capacitance to the capacitance of the same capacitor with vacuum as the dielectric instead. K is the symbol that’s often used in formulas, so sometimes we just refer to the K, instead of saying relative dielectric constant.
      Dielectric strength, or breakdown voltage, on the other hand is the voltage at which the dielectric breaks down and is often given in volts/mil, as you point out.

    2. “K” is often also called “Epsilon_r”, with “Epsilon” meaning the greek letter and “_r” meaning a subscript “r”. But of course “K” is more easy in ASCII.
      Dielectric strength or breakdown-field strength is normally given in V/m. Of course the Meter can be converted to other, stranger units of length. :-)

  2. Really good experimental work. The capacitor with the isolated intermediate disks can be approximated by a set of series connected capacitors. Shame you couldn’t get the barium titanate into large crystal form – the resulting capacitors would have been fun to experiment with because of the ferro electric and piezo electric properties.

  3. The Casimir Effect has to do with plates attracting. Is this force affected if there is a charge on the plates? What about magnetism? Since the Casimir Effect is thought to be caused by virtual particles that we are only beginning to understand, I believe these are questions that must be thoroughly explored. We really don’t know that much about the details of the Casimir Effect beyond the most basic.

    Are there any PhD’s in physics here that would like to chip in, or share my curiosity about the properties of how the Casimir Effect works?

    1. Problem is, the plate separation required to observe it in measured in nanometers – in a vacuum…. not a bath of paraffin wax sitting on a bathroom scale where atmospheric Brownian effects will skew everything.

    2. I had always thought the Casimir effect was quite well understood, as it is a hypothesis that was proven by experiment, rather than a physical phenomenon that required crafting a theory to explain it.

      1. That’s my understanding too. The hypothesis came in 1948 and was first accurately tested in 1997. That’s from https://en.wikipedia.org/wiki/Casimir_effect. However, also from that page, it looks like alternate theories have been proposed to explain it.
        Funny, that. Usually new theories are developed if there’s a discrepancy between experimental result and original hypothesis. In this case the result supported the original hypothesis. I guess the effect exists at a level of reality that’s still fruitful for theoretical development. For example, some of quantum field theory’s development happened after 1948.
        PS. I’m not a PhD in physics.

  4. It’s all well and good having huge capacitors but consider the theoretical lowest possible capacitance. Now contemplate the highest possible frequency oscillator you could build. I wonder how sheets of nano tesla coils compare to silicon when harvesting light. Maybe have the substrate egg box shaped to drop the frequency by guiding the wave into phase.

  5. This should have been a Fantastic Fail of the Week. You set out to develop a non-Newtonian propulsion system (fail), but in the process did some amazing builds, and established your own personal foundation for further experimentation. Fantastic work.

      1. It pisses off other people trying to learn from your success too… Like your buddy phones, and is all “How the actual frick did you do X to your car the other weekend because I’m trying to do it and it’s taken me all day.” and you go over and say, “Well you just do this and this and it goes in like that.” and he’s all super frustrated now and “WHAT!!! I swear I did that, WTF did you just do.” and you take it back out and go. “This, this, that.” and then his eyes pop and he goes “Fuggit, it’s in there now.”

        I was struggling to name a specific example, drum brakes are probably a good one.

  6. Barium Strontium Titanate and the compounds in that series have some really amazing properties. Some time ago there was talk and research using them as replacements for gallium arsenide or silicon.

    10 years ago I got to play with superconductors and lifters or asymmetric capacitors. We were using 50Kv DC. They are fun toys but are not defying gravity or physics. At the time we couldn’t get BST in any quantity or purity for our experiments. We were going to use it for the dielectric not as a superconductor.

    High voltage, giant discharge arc’s and beautiful plasma coronas. Boy that was fun.

  7. I made a huge cap in my teens from many rolls of aluminium foil, cling film and a few gallons of vegetable oil. Capacitance measurement unknown, but it was fun for making massive sparks fed by a modified disposable camera flash circuit (and a fair wait)

  8. Had some similar experiences playing with EL, the problem here was that I could never quite get the mix right. The phosphor and dielectric paints have a specific range of conditions and if you deviate even slightly the whole stack breaks down. My best and brightest attempt was a single layer with Superglue as a dielectric and a conductive backing made from tyre cement mixed with graphite, used this one on Halloween as a “custom” display.
    Of course at the time I didn’t know about “Sellotape graphene” or anything similar, a good idea that I couldn’t test was to use tyre cement mixed with ATO powder.

    If anyone is interested I am trying again with out of date EL powder rather than paint, this time using UV cured adhesive to form the layers.
    The trick is to thoroughly dry each layer and never *ever* use anything with “ene” in the name or it simply melts through and fails on the first power-up.
    Also the white formula can be mixed with a solvent then centrifuged to separate the red and blue components!
    If its clumped up this can be fixed using an oven to get the moisture out, then simply (gently) sieve to get the lumps out and check under a UV light for any inactive debris.

  9. What you’re trying to do is already been created it’s called the e/m drive. You need to focus on changing the size of the plates from a smaller diameter to a larger-diameter

  10. Interesting results with the barium titanate. I’m curious, what frequencies did you measure your capacitance at for the barium titanate? At high frequencies, it experiences a significant decline in capacitance. For example, I’ve measured a capacitance 5x lower at 10 kHz versus at 100 Hz.

    1. I just used the capacitance setting on my Fluke meter. I just checked the manual and it gives range and accuracy but no mention of frequency, so possibly none?

      1. Ok! Yeah I’m not sure how DMM capacitance measurements work in relation to LCR meters. Also, did you ever test the breakdown voltage of your barium titanate capacitor? If so, what was it?

        1. No, since the dielectric was lower than I would have liked — probably due to the presence of the epoxy — I didn’t test breakdown voltage.
          I just now did a search and found Fluke’s webpage on how they measure capacitance http://en-us.fluke.com/training/training-library/test-tools/digital-multimeters/how-to-measure-capacitance-with-a-digital-multimeter.html. It looks like no frequency is involved. They charge the capacitor with a known current and then measure the voltage. I guess they’re making use of the basic formula C = q/V, capacitance = charge/voltage. Charging with a known current (q/t) over a known period of time (t), if that’s how they do it, would give you q.

          1. I’m still in the process. At first I tried a mixture of BaTiO3 and oi to try to balanc high dielectric constant with strength, but that didn’t give me the values I wanted. So now I’m going to try BaTiO3 in epoxy, and maybe look into adding other powders as well.

  11. It’s funny. I found this article while trying to find a dielectric for the same kind of experiment. But I’m going for a completely different method of obtaining thrust.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.