Homemade Van de Graaff generator

As a fan of nuclear physics and accelerator technology, I am always on the lookout for new sources of high voltage. Usually such a device turns out to be something based on a large transformer, voltage multiplier, or a combination of both. But there are many other methods of generating high voltage, which are often overlooked and considered obsolete. One of them is the good old Van de Graaff generator, invented back in 1929. The assembly of such a device on its own will be discussed in this article.

Van de Graaf Generator (VDG, Van de Graaf Generator) is an amazing machine capable of reaching the megavolt range and used in DC accelerators. The most famous example of such a device is Westinghouse Atom Smasher (nuclear Van de Graaff accelerator), which was used to trigger the first recorded photofission process.

In principle, Van de Graaff generators are very simple, but over time they have undergone many changes and improvements, finding their final form in the form peletron. This modification of them is still used in the creation of small high-voltage accelerators, produced mainly by High Voltage Engineering, founded by Mr. Van de Graaff himself.

Wanting to take a little break from the exams of the current semester, I decided to create a self-excited version of such a generator in its simplest form. The design will consist primarily of 3D printed, machined and laser cut parts, as well as some components from a hardware store and gardening tools.

At the heart of my generator is an acrylic plate, on which is mounted an engine from an old garden tool, bearings for the lower roller and a holder for a tube through which the charge-transmitting belt passes. Here, suspended above the belt and the lower roller, there is a lower brush for charge transfer.

Additional photos

The upper part of the generator is even simpler in design. It consists of a 3D printed part mounted on a pipe, into which threaded inserts are soldered to fix the bearing units under the upper roller, as well as a second brush. The bearing assemblies are mounted on threaded studs, on which is mounted an iron platform holding the upper sphere. The brush is directly connected to the bearing assemblies through a pin. During the experiments, I did not find a difference between whether the bearings are connected to the brush or not.

The charge transfer strap is made from Thera-Band Gold fitness rubber. This elastic is 120mm wide, but my design was 80mm wide and had a closed loop rather than just a straight strap. A circular cutter worked well to cut the tape to the right size (don’t try to cut it with a scalpel or the like), after which I glued the ends in a V shape to more effectively distribute the resulting overlap along the seam. The rollers themselves are made of PTFE and Nylon 6-6, both 80mm long and 25mm in diameter. At the same time, their edges are tapered with a slope of 5 ° to ensure the correct running of the belt. The rollers are fixed on their axes with M4 locking screws.

Additional photos

The last element is the upper sphere, which performs several functions. It stores the transferred charge by accumulating voltage, provides a smooth surface to maintain a high breakdown voltage, and also acts as a Faraday cage to keep the internal electric field free. Its last task is the most important, since it provides the conversion of VDG into a real current source, regardless of the voltage at its outputs.

The main part is a steel hollow sphere with a diameter of 30 cm, bought in the garden decor department. I made a hole in this sphere, and on the resulting edge I glued a smooth edge, which I cut off from a dog food bowl.

The sphere is held on the generator by a magnet attached to a steel plate above the top roller. I fixed the magnet itself with hot glue, centering it relative to the hole from below.
By installing a PTFE roller at the bottom and a nylon roller at the top, I get a positive voltage on the electrode at a current strength of 15 to 20 μA. For better visibility of breakdown discharges, I additionally assembled a stand with a grounded electrode, which is a textolite rod fixed on a wooden base, on which a sphere with a diameter of 100 mm is installed. Depending on the humidity of the surrounding air, discharges with a length of 20 to 50 cm are generated. When this range is exceeded, the discharge strikes into the air, and not into the grounded electrode.

If I decide to build a new version of this generator for use as a power source, then I will do something differently. Firstly, I implement it with external excitation, which will allow you to easily control the incoming current and polarity. The main difference between generators with external excitation is that they do not rely on the triboelectric effect, but use metal rollers and a high-voltage bias on the bottom brush to transfer the charge to the tape.

It is also possible to increase the transmitted current by installing several brushes in different positions, which will ensure that the terminal is charged when the tape moves both up and down.

Another significant improvement would be a more efficient formation of field lines in order to reduce the electric field strength on the electrode and reduce the number of air discharges. Since the top electrode is spherical, its electric field is proportional to 1/r², and therefore it has a high density on the surface.

In addition, the installation of grading rings between the electrode and the grounded base, connected through a high-resistance voltage divider, will make the field strength more linear along its entire length.

Graduating rings can be found in commercial Van de Graaff accelerators. They effectively help to get a higher and more stable voltage.

Commercial VDG accelerator with many grading rings. Photo from Science Museum Group

This linear alignment can be easily demonstrated with a simple Python script that plots a spherical and idealized linear field for any given parameters of voltage, sphere size, and ground distance. Naturally, it contains a number of serious assumptions, but at the same time it reflects the main idea very well. The generally accepted value for the formation of a discharge in air is 1 MV / m.