Build an Efficient Electrostatic Generator: Tips, Tricks, and Resources

[Niki]

Hi all

Interested in making an electrostatic generator, anyone have experience of building one and also what type would be the most efficient build induction or friction based? Lots of info on the net regarding the principles but and good books for this or websites?

Also how best to accurately record the voltage produced.

Cheers...Niki

 

[sophiecentaur]

Look up Wimshurst Machine. That’s a truly impressive machine and people have made some very successful versions. Unlike a Van Der Graaff generator, it can blow your socks off. (Treat with great care!) they don’t allow them in UK schools.

 

[rwooduk]

Look up Wimshurst Machine. That’s a truly impressive machine and people have made some very successful versions. Unlike a Van Der Graaff generator, it can blow your socks off. (Treat with great care!) they don’t allow them in UK schools.

Hah, I remember my physics in teacher in school, really old school teacher, used to get us to stand in a circle while one boy was touching th Van Der Graaf generator then when there was a static discharge when he moved his hand away the shock would go around the circle, it was like a Mexican wave but combined with little boys screaming! Then he said you shouldn't participate if you have a weak heart. He had a very 'natural selection' outlook on life.

 

[tech99]

Hi all

Interested in making an electrostatic generator, anyone have experience of building one and also what type would be the most efficient build induction or friction based? Lots of info on the net regarding the principles but and good books for this or websites?

Also how best to accurately record the voltage produced.

Cheers...Niki

I think it is very difficult to make one. The commercial V de G machines sold for schools are quite simple but not easy to copy. Another approach is to make a safe 5kV power supply using a Cockcroft Walton voltage multiplier. The voltage can be measured by a microammeter with a large series resistor. For a V de G machine I suppose we would measure the spark length between spheres, maybe 2 inches, 20kV/cm = 100kV.

 

[spareine]

Instead of building a generator yourself you might like the 'fun fly stick', a convenient toy version of the Van de Graaff generator (~20 kV, battery powered). See internet (or amazon.com) for details.

 

[Rohan]

Another electro-static machine, which is just a fluid version of the Van der Graff generator;
is obtained by replacing the rubber belt and brushes with a pump circulating a non conducting liquid;
in non conducting tube ( eg polyethylene) but with metal sections replacing the brushes.
This also nicely explains why hoses on petrol pumps need to be earthed !
...strongly advise against using any combustible liquid ! ;-)

 

[Julian Phillips]

This channel (Plasma Channel) has some DIY electrostatic generators, including a cool desktop Wimshurst machine.
https://www.youtube.com/channel/UCg45A-ph7Eu8jQgfrwDkHLg

 

[ADesilets]

There are a lot of different types, the triboelectric ones like the van der graaffs can be quite powerful, but they are higher maintenance and tend to be somewhat current limmited at hobby scales.

Wimshursts and sectorless wimshursts are good options, but the counter-rotation can be a design challenge. One of the later stage developements was to use a sectored disk with fully encapsulated sectors and brushes making edge contact with stationary static inductors on both sides of the disk.

Then to keep the induction level stable the output brushes or combs are separated from the brushes/combs that charge the static inductors.

The output brushes are 45 degrees from the neutraliser brushes and the brushes that charge the static inductors are 90 degrees from the neutralisers.

The disk rotates from static inductor brush to neutraliser brush with the outputs in the middle.

The output voltage is significantly more consistent and is equal to the breakdown voltage between the inductor brushes and whatever is the shortest path to the other inductor brush on the other side of the machine and the current produced is the capacitance of a sector in relation to the inductors multiplied by the output voltage since the output is sector to sector and the number of times a sector passes underneath an output brush in one second.

 

[sophiecentaur]

but the counter-rotation can be a design challenge.

It's a very long time since I used the one at my school. As I remember one rotor was driven straight by a belt and the other rotor used a figure of eight belt. But you may be referring to something a bit more powerful than that.
That one at my school was a lot more scary than the safety conscious VDG that I used to teach with.

 

[ADesilets]

Yeah, the design I'm talking about is quite a lot more powerful than a wimshurst for the number of disks. Because disks are sandwiched between two inductors, you get ~4x the amount of charge per sector compared to a wimshurst machine. Womelsdorf machines are set up with a sandwiched disks, but they aren't the only one that use that technique, but if you refer to https://www.coe.ufrj.br/~acmq/ptable.html you can see that it produces a significantly higher current for the number of rotors at the same rotor size compared to the wehrsen machine which uses a inductors on only one side of the disk rotor.

The womelsdorf doesn't necessary separated the output and inductor plate combs/brushes though. That was done to toepler machines later on as a means to stabilise the voltage and increase opperational efficiency.

You have double the inductance and, because the sectors on a wimshurst are the inductors for the sectors on the other rotor, you don't actually benefit from a current production perspective from the second disk unless you offset the output combs so that they aren't puling charge off of directly opposite sectors.

So the wimshurst looks cool, but if you're after practicality and high voltage power production, stationary inductors are the way to go.

Simplifies the drive train as well allowing you to mount the disks on the input shaft directly. The disks can be stacked like pancakes with inductors in between as well and each disk and inductor pair added to the stack proportionally increases current production.

But wimshurst machines are beautiful, mesmerizing and fairly useful machines too.

 

[ADesilets]

Just wanted to add in a real life example for you, I just finished building a modular and exapandable, 3D-printed electrostatic machine that uses #10-32 and foil tape for all of the electrical interals. It uses 120 mm disks since they are the disks that are used in the "sparkit" DIY wimshurst machine kit.

Sparkit claims a power production of 3 uA @ 30 KV and a maximum spark distance of 20 mm, which actually corresponds to a voltage of 30 KV - 60 KV depending on humidity, in a very humid room it will make 30 KV, in a super dry room it will make 60 KV and, because the machines function by capacitance, 6 uA, . Because of the arrangment of brushes in machines as I described, you can use a visual inspection of the sparking and corona discharge jumping off the rotor to the brush to determine at what electrode distance ~100% of your current flow is going through the output brushes.

Using my designe with a single disk and 4 inductors plates, In my environment at 40% humidity and STP I make approximately 30 - 60 KV with the disks, getting visual continuous sparking at without capacitors 6-7 mm between electrodes and visually apparent streamer flow at up to 20-30 mm, after which the current flow prefers the neutralisers and stops majority flowing through the outputs. Using an analog ameter, at a rotational speed of 2,700 RPM I measured a peak current flow 40 uA when the needle electrodes where arranged in at 90 degree angles and separated by the maximum distance that visiual streamer flow was detectable.

In my experience with my other machine. If you can see a visual streamer flow without capacitors attached at a whatever distance, it will become sparking if capacitors are added to the system.

Using a 9 nF capacitor bank, and a spark gap set to just bellow the threshold where I burnt out the other 11 1 nF 20 KV that used to be in the bank. counted a spark frequency of 1 every 6.8 seconds, which works out to a current flow of 26.47 uA. At no point during the charging cycle did I see the current flow switching to the neutralisers or even shifting towards the neutralisers (god it's cool having voltage so high that you can visually inspect the opperation of the generator) so I would say that between the ameter and capacitor bank timing, the capacitor timing is more accurate.

That with a single disk taken from a known wimshurst design. The known design uses two disks to produce a current flow of 3-6 uA @ 30-60 KV and the design I made uses a single one of the exact same disk and produces 26.47 uA @ ~45 KV @ 2700 RPM.