A Modified Van de Graaff Generator Part I

Fig 1. The world's largest air-insulated Van de Graaff generator, built by Robert van de Graaff in 1931, now housed at the Boston Museum of Science. Image from http://en.wikipedia.org/wiki/Van_de_Graaff_generator.

Original Van de Graaff generator

The original Van de Graaff generator design is a good example of how electric charge continually brought onto the inside surface of an already charged spherical-shell conductor can be added to the shell for no additional energy penalty, hence increasing its voltage, no matter what voltage the shell has already reached. The charge is initially placed onto a motor-driven belt made of insulating material, at a location outside the shell. The belt transports the charge upwards into the shell, or "dome", where it is collected.

In a small to medium-sized laboratory Van de Graaff generator being charged by external excitation, charge would typically be sprayed onto the low-voltage end of the belt at say 10kV and 20μA. Ignoring losses, it would also be collected at the high-voltage end at 20μA, where the dome voltage can reach say 350kV. So if all energy losses could be eliminated, we would expend 0.00002 A × 10000 V = 0.2 W to get 0.00002 A × 350,000 V = 7 W, i.e. a 35 : 1 power gain.

A 7 watt output is of little practical use, and it's understandable in such a low-power device that not much attention would be paid to the possibility of generating excess energy. But if the current could be increased to say 200mA, still a very modest value in electromagnetic terms, the power output would be 70 kilowatts. That would certainly be a useful result, if it could be obtained without paying a full energy penalty to get it.

Even higher power gains would be possible for larger generators, able to operate with higher dome voltages.

Energy loss

For a Van de Graaff generator, the major energy loss associated with bringing charge onto the dome is obvious. Charge carriers on the belt have the same polarity as the charged dome, and so there is repulsion opposing the belt's movement towards the dome. The charge carriers have to be transported "exposed" against the entire voltage gradient between the dome and Earth.

Pelletron

Fig 2. Basic principle of the Pelletron. See http://www.pelletron.com/charging.htm for this image as an animated GIF.

This major energy loss also occurs in the "Pelletron" modification of the original Van de Graaff generator, which uses a chain of alternating insulating and conducting segments, instead of an insulated belt. (The conducting "pellets" can be conveniently charged/discharged by induction instead of by charge spraying or triboelectrification). In a Pelletron, not only are like charges repelling as the charged chain moves towards the dome, but unlike charges attract as it moves away, with a major energy loss in both cases.

Generation of charge within the dome

At first sight it would seem very easy to generate charge within the dome, rather than to generate it externally and then bring it in against the full voltage gradient. For example, simple metallic "comb" corona discharge generators could be connected directly to the inside of the charged dome (and would thus operate at its already high voltage). Or, a heated filament could be placed within the dome to generate electrons by thermionic emission, a method which can easily produce a current of 200mA or more.

Charge separation

However, we cannot produce any net charge by such methods, either within the dome, or elsewhere. Charge can only be separated, and the problem then arises of how to deal with the charges of unwanted polarity. A corona discharge will act on air molecules to create charges of the desired polarity; but also an equal quantity of charges of opposite polarity. These cannot be allowed to drift internally to the dome, where they would tend to neutralise the desired charge on it. Worse, if they are expelled outside of the dome, they will be attracted back to it, again tending to neutralise it. To get rid of them by neutralisation against Earth, they would have to be moved against the entire voltage gradient between the dome and Earth, thus destroying any possible net energy gain.

If a (negative) charge of electrons is expelled from a filament by thermionic emission, then the filament and whatever is energising it will develop an equal and opposite positive charge. This would soon cause a voltage breakdown (flashover) if the positive charge was not neutralised somehow, e.g. by connection to Earth. But electrons emitted from an earthed filament, no matter where it was positioned, would once again have to act against the entire voltage gradient between the dome and Earth, to reach the inside surface of the dome. This would again destroy any possible net energy gain.

So, charge separation within the dome certainly presents difficulties. Next time I'll look at the more promising method of bringing externally-generated shielded charge into the dome.