Wimshurst Machines: High Voltage From The Gods

Wimshurst machine demo
Wimshurst machine demo

The Wimshurst machine is one of the oldest and best known electrostatic machines, consisting of its iconic two counter rotating disks and two Leyden jars. Most often you see someone hand cranking it, producing sparks, though we’ve seen it used for much more, including for powering a smoke precipitator for cleaning up smoke and even for powering a laser.

It works through an interesting sequence of events. Most explanations attempt to cram it all into one picture, requiring some major mental gymnastics to visualize. This often means people give up, resigned to assume these work through some mythical mechanics that defy a mortal’s ability to understand.

So instead, let’s do a step-by-step explanation.

The Beginning: Charging The Sectors

Overview of sectors
Overview of sectors

Each disk is covered in metal sectors on their outward facing sides. The sequence of events begins at any sector that has an unequal amount of positive or negative charge. As long as the sectors are clean and dry then there’s usually at least one that’s charged. Let’s say for example that one has a net negative charge and is on the front disk.

That net negative charge influences the nearest sector on the rear disk, repelling negative charge to the far side of it leaving the near side with a positive charge. That’s called electrostatic induction, and it’s for that reason that the Wimshurst machine is called an influence machine since the charge on one sector influences the charge distribution on another sector.

Next, let’s switch to the rear disk and look at what happens to that sector that’s been influenced.

The neutralizer bars neutralizing
The neutralizer bars neutralizing

The next thing that happens is the real genius. Each disk has a neutralizer bar facing it. Each end of a neutralizer bar has a brush that touches the sectors as they pass. And there are an even number of sectors. That all means that when a brush is touching a sector, the neutralizer bar is now electrically connecting that sector with another sector at the other end of the neutralizer bar. It shorts them out.

Let’s say a neutralizer brush is touching the sector that’s been influenced, the one shown above that has had its charge redistributed such that it’s positive on the side facing inward and negative on the side touching the brush. Even though the sector is neutral overall, the neutralizer bar sees only the side that’s negatively charged. It now sees an imbalance between the two sectors that its two brushes are touching. That causes a current to flow in order to restore that balance. Some of the negative charge will flow from our influenced sector to the other sector. From the neutralizer bar’s persepective, it has now neutralized the charge on the two sectors.

Influencing other sectors - rear and front views
Influencing other sectors – rear and front views

When the disk rotates the sectors away from the neutralizer bar, the first sector is left positively charged having just had some negative charge taken from it. And having received that negative charge, the other sector is left negatively charged.

These charged sectors are rotated more to where they face sectors on the front side of the disk just when those sectors are touched by the brushes of the neutralizer bar on that side. And so the newly charged sectors influence charge in more sectors, and so on.

A helpful realization is that this influencing and neutralizing event causes one sector to make the sector facing it on the other disk become charged with an opposite charge. Our negatively charged sector created a positively charged sector. That positively charged sector, once the disk was rotated, went on to create a negatively charged sector.

Charges Whirling To Collectors

Whirling charges and collectors
Whirling charges and collectors

The front and rear disks (which are rotating in opposite directions) result in the charges as shown above.

It may take a moment to convince yourself (since you’re seeing the front and rear views side-by-side), but all negatively charged sectors are headed to the left collector and all positively charged sectors are headed to the right collector. You’ll also notice that the sectors that have just passed the collectors have had their charges, well, collected. They’re neutral overall again until they get to the neutralizer brushes, where the influencing and neutralizing we covered above recharges them.

The collectors don’t touch the sectors. Instead they have sharp points that face the sectors and have an air gap between them. This is a familiar technique which we’ve seen before in the functioning of Van de Graaff generators. Each collector has sharp points facing sectors on both disks which facilitate the transfer.

Using the left collector as an example, the negative charge on the sectors repels electrons from the points, leaving behind a positive charge. Since they are sharp points, that positive charge is crammed together resulting in a strong electric field in the gap near the points. That strong electric field tears air molecules apart and begins the process of making the air conductive, forming a bluish corona near the points. It’s that conductive air that causes the negative charge of the sectors to cross the gap to the collector. That leaves the sectors neutral again.

A similar thing happens at the right collector, just with opposite charges. Since those sectors are positive, the sectors will receive electrons from the collector, making those sectors neutral again.

But where does all that charge used for neutralizing the sectors go to and come from? That’s where the rest of the circuit plays a part.

The Leyden Jars And Spark Gap

The Wimshurst machine circuit
The Wimshurst machine circuit

The rest of the circuit consists of a spark gap and two Leyden jars. The two Leyden jars are just two cylindrical capacitors connected in series. The spark gap can also be thought of as a capacitor, albeit one with a dielectric that breaks down easily and that has a low capacitance compared to the Leyden jars. The spark gap is in parallel with the Leyden jars, and both are in parallel with the collectors.

That means the collectors are connected to each other through the disks but also through the Leyden jar/spark gap capacitors.

Charge that’s collected from the sectors charges up the Leyden jars and the spark gap. The Leyden jars are designed to withstand a higher voltage than the spark gap so it’s the spark gap that breaks down first. When it does it produces a short circuit. All the accumulated charge in the Leyden jars quickly dumps through the spark gap as a spark, neutralizing the Leyden jars until the charging process starts again.


In summary, through induction, the neutralizer bars are tricked into charging the sectors. The collectors collect that charge and store it in Leyden jars and the spark gap. When the charge has resulted in a sufficient potential across the spark gap a spark occurs, shorting out the Leyden jars until enough charge can be collected again for another spark.

But as we said, they can be used for more than just producing sparks. Two examples we’ve seen here on Hackaday are for powering a smoke precipitator and for powering a TEA laser.

38 thoughts on “Wimshurst Machines: High Voltage From The Gods

  1. Years ago I saw a more modern design of this type of generator that used metal rods sticking radially from a PVC hub. It produced some rather impressive discharges.

  2. I have never seen anything like this in my life. It’s said that a person should learn something new every day. Well, this sure filled today’s requirement. Thank you!

    1. Sounds like the machine on display in one of the halls of the science building at nearby small iniverity.. In the past is was a puce of medical equipment use in a small town doctors office. Some of the the “accessories” looked a bit freaky to me. While the weather has to be just right it will produce a fat spark, or it did a few years ago.

  3. In the ’30s Germany developed the first electron microscope which was powered by a lot of Wimshurst machines. That was during Nazi time so there’s little information about this available on the internet.

  4. I don’t need to build this, I already have plenty of projects.
    I don’t need to build this, I already have plenty of projects.
    I don’t need to build this, I already have plenty of projects…

    1. A quick search doesn’t show that HD has ever considered a Kelvin water-dropper generator. Google shows a lot of instances of this neat little gizmo. I made one back when (c. 1962) orange juice came frozen in handy little cans. It worked fine and once baffled three PhDs (physics) simultaneously.

  5. This is one of the BEST HaD postings I’ve seen in several years. Furthermore it is the first time I see such a clear and well presented explanation of how the Wimshurst machine works generating high voltage starting from a small initial spontaneous imbalance. EXCELLENT!!!

    1. Thanks! I’d started doing a video about this a few years ago, that’s where the 3D model came from for the diagrams, but it got bogged down trying to show what was going on on both sides of the disks at the same time. So I dusted the model off and tried it in text form this time. I’m glad to hear it worked out!

        1. Hmmm… My first thought was it’d probably be confusing but seeing the same animation of the charges moving on both sides of the split screen would make the viewer instantly understand that they were seeing two views of the same thing. Definitely worth trying. Thanks.

    1. So did i, saw one on Hackaday years ago, Inspired me to build one. Used brass parts, stained all the wood to make it look old. After I was finished the wife had me put it on display. It has been used to teach a lot of kids (my wife works with children ) about static electricity.
      so build one,lots of plans on the net.

    1. I just measured the capacitance of the whole Leyden jars and spark gap system as around 14 picofarads. With the spark gap set to 2.5cm or 1 inch it took around 2 seconds of moderate cranking to produce a spark. Using the rough formula voltage (kV) = 76.2 x spark_length_in_inches, that’s 76.2kV (we talked about that formula here http://hackaday.com/2016/12/08/measuring-high-voltage-in-millimeters-and-other-hv-probe-tricks/).
      The energy stored in the Wimshurst’s capacitor network just before the spark was 1/2CV^2 = 1/2 x 0.000000000014F x 76,200V^2 = 0.04 joules.
      Brian’s recent article about powering a laptop with supercapacitors (http://hackaday.com/2017/03/03/powering-a-laptop-with-supercapacitors/) mentioned 2.7V, 500F supercapacitors. Using 1/2CV^2 again, charged up to 2.7V, those can store 1/2 x 500F x 2.7V^2 = 1822.5 joules.
      So comparing the two, 1822.5 joules / 0.04 joules that’s 45,562 times as much energy.
      I don’t have resistances, otherwise I’d try doing the calculation with time constants, but very roughly that means it would take 45,562 x 2 seconds = 91,124 seconds = 25 hours to charge up the supercapacitor using the Wimshurst machine.
      That’s actually much faster than I was expecting.
      In any case, Wimshurst machines are very inefficient due to all the corona losses around the sectors and when transferring the charge from the sectors to the collectors so it’s not a good solution to start with. But it’s interesting doing the calculations!

      1. Steven’s comments are most interesting as numerical data on performance are rare. I was interested in the current generated by his generator. To charge a capacitor to a given voltage use the formula V=IxT/C. Using your figures of 14pF, 70KV and 2 S indicates that the charging current is about 1/2uA. I have experimented with Van de Graff generators where direct measurement of the charging current is possible and typically 10uA. With 300mm spheres, I measure 100mm (300KV) sparks every 1 second. Using the same formula the same formula results indicates a sphere capacity of 30 pF. So far so good. If I bring a needle point towards the charged sphere, the sparking stops. The reason being that the high field strength at the needle point initiates a corona discharge which bleeds current from the sphere, reducing the voltage. Even a blunt conductor, such as a finger will do this- quite handy as you can tough the sphere without a sever shock. How is it then that a Wimshurst can support such long sparks between needle points with such a small current.

  6. We had a Wimshurst Machine at home when I was a kid and it’s in my garage now. I just checked before posting anything descriptive and I think my Dad made it. The discs are black, like pre-vinyl LPs but smooth and unsegmented – how does that affect the description of how it works?

    It has a completely wooden base and vertical supports, Meccano pulleys and used to have leather belt but I can’t see that now. Dad used to have the sealed glass tubes which lit up in pale colours like pink and green and I may have those too.

    1. I’ve heard of Wimshurst machines with sectorless disks and seen photos but never experimented with them myself. This page has a lot about the construction of it and tips for starting it http://www.coe.ufrj.br/~acmq/bonetti.html.
      It looks like it works the same way but the fact that the electrically conductive metal sectors are replaced by non-conductive plastic necessitates the differences in the neutralizer bar, collectors and how it starts.
      Imagine you modified the machine in the article above by removing the sectors.
      The first thing you’d have to address would be the neutralizer brushes. Since charge doesn’t move around on a non-conductive plastic surface, the neutralizer brushes would take charge only from where they made physical contact. That means they wouldn’t get much charge at all. Making the brushes as wide as the sectors would be if you have sectors still wouldn’t make as much surface contact as you’d expect. So instead you replace the brushes with wide strips of metal with sharp points along the strips. We call those combs. And the combs don’t touch the disks, but have a gap between the points and the disk. Now the air between the combs and the disk becomes conductive (like I describe in the article above for the collectors). Charge from anywhere adjacent to the combs can now leave the plastic and go to the sharp points of the combs.
      But the non-conductive plastic also creates an issue for the collectors. The neutralizer combs resulted in a wide swath of charge on the disk. If you have a narrow set of sharp points for the collectors then you’d collect only a narrow width of charge from the disk. With metal sectors, that wouldn’t have been the case since the charge moves around on the metal, but not on the plastic. So the collectors have to be the same width as the neutralizer combs in order to collect all the charge.
      The last issue is getting it started. The link I gave says his at least wasn’t self-starting. You need a charge source to get it started. You’d spray charge to the disk opposite one of the neutralizer combs. I guess that’s needed because even though you may have a random net charge somewhere on the disk, it won’t spread around like it would on a metal sector and so would likely not be in the right place to start it (my guess anyway).
      Do you remember if yours was self-starting? Or did you need to charge it somehow first?

      1. Sorry for tardy reply and thanks for your fast one Steven.

        I’ve just examined my one again. The collection system comprises numerous copper ‘bristles’ (set in wooden brushes) which do make contact with the disk surfaces and probably cover 60 – 70% of the area. Maybe as it’s simpler, it’s also less effective or the metal sector type would not be more usual type.

        There is no visible starter mechanism (or anything apparently missing, apart from belts) and my recollection was that the handle was simply turned and after a short time the voltage would appear. I think I remember Dad and then me offering a knuckle to the disk surface to get a small spark. Have I remembered this wrong?

        1. I don’t know how well they’d self-start. As I said, I’ve only heard of sectorless ones and not experimented with them myself so all I can go by is what I’ve read. And now I’ve read (from you!) that some do self-start. Thanks! Theoretically I don’t see why they wouldn’t, it’s just how well they do in the real world I wasn’t sure of. I can see it being more difficult to self-start than one with conductive sectors.

    1. The answer must be “yes”, but it’ll also depend on the geometry of the discs, their spacing, and maybe even the humidity of the air to some extent. In short, it would be a try-it-out-and-see situation rather than something you’d want to design around.

      But they’re easy enough to make. Try it out and see.

  7. I purchased one of these from Edmund Scientific in 1975. I think it sold for about $25 at the time. The rotating disks were about 10 inches in diameter. It would produce sparks about 1 inch long every few seconds when rotated..

    The first thing I did in experimenting with it was to run it without the mini Leyden jars connected, which apparently turned it into a corona discharge machine. In a small room for any just a few seconds would produce a very strong odor of ozone that permeated the air very quickly.

    I then built a homemade Leyden jar out of a plastic Tupperware juice container. It was a 1 quart model. I lined it with aluminum foil inside and out, affixed with scotch tape. Drilled a hole in the center of the plastic snap-on lid, and stuck a metal rod through it, which I connected to the inner foil with a piece of wire and more tape.

    I connected it to the two collectors of the machine with alligator jumpers – one clipped to the rod, and the other to the outside foil.

    I cranked away for probably a full minute and nothing seemed to be happening at first, but after about 30 seconds I could hear the relatively loose foil wrap start to make a crackling noise, as if it were moving. (In fact it was).

    The machine came with a discharger, which a a semi-circular pice of stuff wire with small metal balls on the end, and a plastic handle. It was intended to be used to remove any lingering charge on the machine’s own small Leyden jar.

    I held the discharger against the outside foil of my homemade Leyden jar, and rotated the other end toward the central rod. I got the surprise of my life… As the wire began to approach the rod, there was a brief hissing noise, the any enormous fat spark, a good 4 inches long, flashed across the gap, with an extremely loud “BANG”. Scared the holy piss out of me! It was much more than I expected, and I’m lucky I didn’t manage to kill myself.

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.