A Big Wehrsen Machine (or a sectored Holtz machine)

By August 2003, I could test a large Wehrsen electrostatic machine, that I started building more than 2 years before. The construction was interrupted for long periods, while I built a smaller version to test the construction methods, and directed my research about high-voltage generators to other directions, as variations of the Tesla coil. A good introduction about the Werhsen/Wommelsdorf machine is in the page about the prototype machine. The Wehrsen machine is equivalent to a Holtz machine of the first kind, using a rotating disk with internal sectors and inductor plates encased in solid insulation. The machine, as initially planned, didn't work adequately, and is now provisorily working in a simplified form.

Original design for a Wehrsen machine

The machine was constructed essentially as a scaled-up version of the prototype, but with some improvements in the construction methods and variations due to the larger size. It had a fixed disk with 60 cm of diameter, made with 3 mm black acrylic plate. Mounted on it were two inductor plates, behind spark shields made with white 2.5 mm acrylic plates, fixed with screws to the back plate. The inductors had the same construction used in Voss machines, made of varnished paper, with strips of aluminum foil ended in disks at their centers to distribute the charge. The paper and metal assemblies were then covered with two layers of adhesive plastic foil, to ensure good insulation. The complete encasing of the inductor plates is essential for this construction, or the machine doesn't produce long sparks due to leakage by corona around the spark shields. The rotating disk had 54 cm of diameter. It was built as a stack of three 2.5 mm acrylic disks, with the two posterior disks holding 16 sectors of adhesive aluminum tape each, interleaved. The orientation of the three disks was selected for as uniform weight distribution as possible, so the disk didn't vibrate significantly even when rotated very fast. The 32 sectors were accessed through buttons in the frontal disk. The buttons were made of chrome-plated steel rivet heads, glued in place and contacting the sectors through small springs. Due to the large size of the disk, I could not glue the disks together with hot glue, as I did in the prototype machine. I tried to just join the disks with 8 small screws close to the edge of the disk, crossing all three disks, leaving the insulation between the sectors made by their separation alone. The effect of screws crossing the three disks was difficult to predict, but I made the sectors close to the screws a bit smaller, to increase the distance between the sectors and the screws. The construction of the disk in this way didn't work well. See below.

The rotating disk was supported from behind only by a strong turned wood support, having at the top a wood cylinder holding a steel axle through two ball bearings. The disk was mounted on a cylindric wood boss with three screws that pressed it against a washer of flexible material, allowing precise adjustment of the disk position for true rotation. I decided for manual operation for the machine, as it allows better control during experiments. In this machine I used commercial aluminum pulleys instead of wooden pulleys, because they don't deform with time. The driving cord was a sewing machine round leather cord.

The machine had two Leyden jars made with tall polyethylene cups, with 200 pF each (maybe too much). Their outer plates were interconnected with wires below the machine's base, with a small spark gap at the center that can be used as pulsed output. The jars could be switched on or off through switches connecting them with the charge collectors. Other switches could connect the charge collectors with the output terminals and with the inductor plates, for startup. The switches were made all in the same way, pivoting on brass cylinders that could rotate around the charge collector tubes. They remained fixed at any position by friction, created by rubber pads in the holes that hold the switch arms, pressed against the tube by springs and metal pins mounted in central holes in the arms. The charge collector brushes could be moved in and out through handles at the front side of the machine.

The frontal structure that holds the terminals was supported by two acrylic round bars with turned wood bases, holding a horizontal rectangular acrylic bar at the center, and continuing as tubes. At the center of these tubes, aluminum tubes made the connections with the terminals. The terminals were made by metal spinning, in brass. The horizontal bars could slide on aluminum tubes inside the wood bals that made the connection with the upright supports. The horizontal acrylic bar holds the charge collectors and the neutralizer bar.

The neutralizer and charge collector brushes were make with a new technique, by wrapping a thin nickel-chrome wire around a silicone monofilament line. The idea was to obtain a flexible brush that doesn't break easily but stays in place. The result was not very good, as the brushes were still breaking after some time and were heavy enough to mark the disk. The buttons in the frontal disk would then be better if almost at the level of the disk, for minimal irregular pressing on the brushes. For the brushes taking charge from the back of the rotating disk to charge the inductors, as similar technique was used, but the nickel-chrome wire was wrapped around silk threads, and fixed with some glue. This resulted in lighter brushes and no problems. Brushes like these at the front side didn't work well, because they were easily moved away by the buttons in the disk

The first tests showed poor performance. The machine was self-starting easily and producing the expected short-circuit current (100 µA), but was not producing long sparks. The reason for this I could trace to intense corona and sparking between sectors in the spaces between the stacked disks, probably helped by accumulation of ozone in the narrow spaces between the disks. Testing the machine with the back disk of the three only, it showed much better performance. It worked even better with the central disk alone, possibly because it was made with blue acrylic instead of black. The holes in this disk didn't appear to be a problem. A stack of the two sectored disks, without the frontal disk with the buttons, worked a bit less well. With the three disks, the performance was poor. Taking into account the good performance of the prototype machine, it's evident then that a machine with embedded sectors must have them really completely embedded in solid insulation. The insulation of the inductor plates as I made it, with adhesive plastic foil, is also not very satisfactory. Significant leakage can be seen at the internal edges close to the neutralizers, by looking at the machine in the dark. It's really necessary to avoid bubbles in this construction. I can see the bubbles under the foils lighting up in the dark..

Sectored Holtz machine

I decided then to abandon the complicated rotating disk until I find an effective method to join and seal the disks together without deforming them (also because I deformed the back disk when trying to join the disks with wax and heat...), and made a 55 cm Wimshurst-type disk with transparent 3 mm acrylic, with 32 sectors. I would then have also something to compare to the future disk with embedded sectors. This disk worked as well as the sectored original disks. The machine was then transformed in to a kind of sectored Holtz machine, with a startup system as in the Wehrsen machine. The plate for the disk was not very uniform, and I had to balance it by gluing a long string of lead blocks to the lighter edge of the disk, a solution that I have used in other machines, that although not very elegant is simple and doesn't appear to introduce significant losses. The alternative would be to sand one side of the disk until its thickness is uniform and then polish the sanded area. A lot of work. I mounted brushes made with silk threads wrapped with nickel-chrome wire in the frontal brushes too, and they worked well with the flat disk.

The startup system, by closing the switches that connect the charge collectors with the inductor plates, can easily start the machine even in very high humidity conditions (tested at 90% of relative humidity in the air), but it's important to guarantee good contact in the switches and in the connections to the inductor plates. The machine works better when the inductor plates are at a considerable distance, 1.5 cm, from the rotating disk. The machine produced easily sparks with 18 cm of length, eventually going to 20 cm. It's expected that in drier air it will exceed this. The shape of the sectors leave a minimum distance of 24 cm between the terminals, that is the expected maximum spark length for the machine.

Created: 19/08/2003
Last update: 23/12/2003
Developed and maintained by Antonio Carlos M. de Queiroz

Return to Electrostatic Machines

Lamento informar que o Prof. Antonio Carlos Moreirão de Queiroz faleceu há algum tempo.
Sei que esta página é visitada constantemente. Assim, gostaria de saber se temos algum visitante (interessado) que seja da UFRJ. Se for, por favor, envie um e-mail para watanabe@coe.ufrj.br.
Comento que é impressionante ver o que Moreirão foi capaz de fazer. Ele não só projetou os circuitos, mas também fez todo o trabalho de marceneiro (melhor que muitos que já vi e eram profissionais).
Segundo Moreirão contou em uma palestra, ele só levou choque uma vez. Sem querer encostou o dedo médio em um capacitor com alta tensão que se descarregou através do dedo. A corrente ao passar por uma das articulações a danificou e doía sempre que dobrava esse dedo. Mas, segundo ele, já tinha acostumado.

E. Watanabe (ELEPOT)