The Wehrsen Machine


Wehrsen machineIn 1907 [p77], Heinrich Wommelsdorf described an electrostatic machine that was a variation of the Holtz machine of the first kind, but with a rotating disk made of ebonite having embedded sectors accessed through buttons at the disk surface, and inductor plates embedded in celluloid attached to a fixed ebonite disk. Machines with this structure were popularized by the instrument builder Alfred Wehrsen, in Berlin, that also patented this form of disk in 1903 [DE154175], and so it is sometimes mentioned as the "Wehrsen machine" [1] (see note in [p80]), I find this name convenient to separate this machine from the "Condenser machine", that is more associated with Wommelsdorf's work and is quite different, and so will call the machine here by this name.

The machine described in 1907 had a rotating disk made of three or four layers of ebonite vulcanized together, with sets of intercalated sectors distributed through them, in two or three planes separated by thin ebonite disks, for high insulation. The idea was described in a patent by Wommelsdorf issued in 1908. The fixed disk was a bit larger, and had inductor plates made of paper with a central metal strip, embedded in celluloid. As celluloid is a flexible material, the inductor assemblies were mounted with screws in a rigid ebonite disk. The inductors had exposed brushes or points at one side, used to take charge from the bare back side of the rotating disk at a small angle before the charge collectors, as in a Holtz machine of the first kind. At the front side of the rotating disk, there were charge collectors with adjustable brushes connected to the output terminals and Leyden jars through switches, an adjustable neutralizing bar with brushes, and switches that could connect the inductor plates directly to the charge collectors for easier startup as a Belli machine.

The machine could be motorized, and could run at high speeds, producing a relatively high current due to the efficient use of the rotating disk area by the embedded sectors, and high voltage due to the good insulation.

Versions of the machine were built at least until the 1920's, in several models and sizes, with one or two rotating disks, but almost always with the characteristic appearance given by the white celluloid inductor plates and the buttons in the rotating disks.

Some examples:
An early machine [p77], with switchable sets of Leyden jars, direct motor drive, and no apparent celluloid plates.
Another machine with direct motor drive [26], but with celluloid plates and startup switches connecting the charge collectors to the inductors or to the output.
A small machine [22], without visible startup switches.
A machine with segmented Leyden jars [22], switches only for the Leyden jars, and the neutralizer in a strange position.
A machine with vertical output switches [22] and motor drive.
A machine with two rotating disks [1], motor drive, and a complete set of switches, mentioned as the "Wehrsen machine".
A nice big machine [22] with segmented Leyden jars and no startup switches.
Wehrsen's "Mercedes" machine [34], with many switches. A version with 2 rotating disks [34], also with a complete set of switches and Leyden jars insulated from the machine's base.
 A machine with two rotating disks, dated from 1911 that exists at the Cavendish Institute, in England. Detail of the switches (possibly incorrectly assembled, as there is no way to connect the charge collectors to the output terminals in this way). Back view, showing the back neutralizer and the startup switches. The fixed disk, showing the celluloid inductor plates. Another view showing the switches, and another, from the other side.
A similar machine, but with an electric motor, exists at the Technical University of Clausthal, Germany. Left view. Right view. Pictures sent by Prof. Friedrich Balck.
A still functional machine is used in demonstrations at the Bonn University, Germany. It's the machine on the picture above. Side view. Back view. Photos sent by Michael Kortmann. He sent also this catalog from the Alfred Wehrsen company.

A model of the Wehrsen machine

By August 2001, I started to build a large Wehrsen machine, with a 60 cm fixed disk and 55 cm rotating disk with two layers of intercalated internal sectors. However, I become worried about the methods to use to construct the disks and several other details of the machine, and decided to make first a smaller machine to experiment with adequate techniques and see how this kind of machine works.

I started by the disk, that is composed by 3 blue acrylic plates with 2.5 mm of thickness and 30 cm of diameter, with 32 sectors in two layers of 16 separated by the central disk. The sectors are as the sectors of a Wimshurst machine, dimensioned to look exactly side by side when seen through the disk, with 7.5 cm of length, at 5 mm of the border of the disk The frontal disk has 32 holes for access to the sectors, aligned with 16 holes in the central disk for access to the sectors in the second layer. After the holes were made and the sectors, made of adhesive aluminum foil, were applied, the three disks were glued together with hot glue. The glue was spread with the regular application pistol in continuous beads around the sectors, connection holes, outer edge and inner edge of the sectored area, and the disks were then heated in an oven between glass plates, pressed by weights. I inserted some metal blocks for aligning the holes, at the center and three of the holes crossing two disks (the blocks had threads tapped on them, so they could be easily pulled out with a screw if becoming glued). I heated the disk to 200 degrees for about 20 minutes, until I noticed that the glue had melted completely, and all the moisture of the assembly had disappeared. I then let the assembly cool in the oven, and removed the excess of glue from the holes and edges. The result could be better if I had inserted aligning blocks in all the holes crossing two disks, and had used less heating. It would be probably enough to let the oven on until all the moisture disappears and the acrylic starts to soften and adhere to the glass (easily visible through the glass) and then turn it off and let the heat spread through the assembly while it cools. The disks deformed a bit, and some of the holes got somewhat misaligned. The glass plates kept the surfaces flat, although some irregularities appeared, specially in the upper disk. I could, however, trim the edge of the assembled disk and polish it, with a good result. The 32 buttons were made of rivet heads glued with cyanoacrylate glue to the holes, with small springs inside to ensure good contact with the sectors. I tested all the sector pairs for insulation, trying to force a spark between the sectors. I had to repair two points with more glue through the holes, and one defect in the back sector group could not be repaired. These defects could be prevented with more attention while spreading the glue. In other points, the insulation was perfect.

The fixed disk is made in black acrylic, 2.5 mm thick. It has a hole at the center for the boss that holds the rotating disk, and two white acrylic spark shields fixed by small brass screws, with the inductor plates mounted on their under sides. The inductor plates are made in paper, with a strip of aluminum foil at the center, with lateral connections to the switches that make contact with the charge collectors, and brushes of thin metal foil strips (silver) to collect charges from the back side of the rotating disk. The inductors are insulated by two layers of adhesive plastic foil, that cover the entire back surfaces of the spark shields. In this way, the inductors are completely encased in solid insulators. The layers of adhesive plastic foil are essential. Without them charges leak abundantly through the internal edges of the plates, opposite to where the brushes are mounted, and the machine doesn't produce high voltages.

I made a light base board in plywood, leveled, polished, and painted in black. The rotating disk is supported by a single support turned in hard wood, that has a conical upright part and a horizontal cylinder at the top. The cylinder is crossed by a 6 mm steel axle, running on teflon bearings, that has at the front side a nylon boss that supports the rotating disk, fixed by three screws through a compressible plastic washer, and at the back side a pulley. Above the same support, there is an adjustable brass bar crossing a wood cylinder, that supports the fixed disk, holding it through a screw mounted in a nylon block glued to the disk. The bar can be removed by unscrewing it for disassembly of the machine. A wood ball with a threaded rod fixes the bar to the support structure. The fixed disk is supported below by two sections of PVC tube, with slots for the disk and internal wood cilynders at the lower ends for fixation, that can be mounted at adjustable depth (they are mounted over slots in the base), completing the system that allows adjustment of the distance between the fixed and rotating disks. The two terminal supports at the front of the machine are PVC tubes inserted in wood bases. Also mounted in the base are the Leyden jars and the motor. All the parts are fixed to the base through 3/16" threaded rods, with washers and nuts below.

The charge collector and terminal assemblies are quite complicated. A transparent acrylic bar fixed to the two vertical supports holds at its extremities, aligned with the buttons in the rotating disk, two 1/4" brass tubes (drilled rods). These tubes are crossed lengthwise by 1/8" brass rods, held by light friction, that hold the charge collecting brushes at the back end and insulating handles at the front end. This allows the brushes to be pulled out for distance adjustment and maintenance. Over the tubes are mounted three switches, that allow contacts between the inductor plates and the charge collectors, the charge collectors and the spark terminals, and the charge collectors and the Leyden jars. The switches turn around brass rings with cores of nylon. Inside the nylon cores there are holes with small springs to keep the switches in place and ensure good contact between the rings and the tubes. The switches have insulating handles made in nylon, and contacts with aluminum balls. At the center of the horizontal bar, there is a neutralizer bar, with brushes at the ends touching the buttons on the rotating disk, and a button with a dial that allows precise positioning, rotation, and distance adjustment of the neutralizer. A thumb screw allows the fixation of the neutralizer. The horizontal bar is fixed to the upright supports by short 3/16" brass bars crossing the assembly, terminated in aluminum balls. Inside the PVC tubes, they support vertical bars that make contact with the terminals above, centered in the tubes by small wood cylinders that they cross, at the two extremities. The terminals slide on wood balls mounted over these rods, pressed inside the wood balls by steel balls mounted over springs that insert in holes in the bars inside the tubes. The terminals are of the double ball-plane type, made by metal spinning, and have handles turned in nylon. All the parts are fixed together by screws or glue.

The charge collector and neutralizer brushes were made from brushes taken from the output slot of a discarded laser printer. The brushes that charge the inductor plates were made with thin silver strips. These materials proved to be the most adequate so far, not breaking easily and causing reliable startup of the machine. The Leyden jars were made from tall acrylic drinking glasses, mounted on cups made of soldered brass plates, interconnected by a wire under the base. Their capacitance is of 57 pF.

The machine is powered by a sewing machine motor, with a pedal for speed control. The motor is fixed to a wood block at a corner of the base by its normal fixation structure, and powers the machine through a polyurethane cord. The maximum listed maximum speed of the motor is 7000 rpm, what is enough to turn the disk at 21 turns per second through the 1:5.5 diameter ratio of the pulleys.


The machine self-starts easily. In conditions of high humidity, or after a long time without operation, it is necessary to close the switches that connect the inductors to the charge collectors. After the startup it works better with the switches open, or each spark discharges the inductors. The ball-plane gap only produces long sparks with the balls being at the positive side. If the polarity is inverted (a characteristic hissing noise can be heard in this case), it can be reversed by moving the neutralizer beyond the spark shields for a moment and then returning it. Actually, any sudden movement of the neutralizer causes a reversal when the gap is widely open. The machine produces more current when the neutralizer is at low angle, but longer sparks without polarity reversals only when the neutralizer is at high angle. For reliable operation, the distance between the disks must be quite high, 9 mm. With shorter distances the spark length is smaller and frequent polarity reversals are a problem. The output current is not significantly affected by the distance. It reaches 70 µA with the motor at full speed (measured from one terminal to the neutralizer bar). The maximum spark length observed was 12.5 cm, that is the maximum distance allowed by the terminal assembly. At this distance, however, there is significant leakage between the terminals and the neutralizer and disks, and sparks occur only in dry air. Consistent output is obtained up to 11 cm sparks. With the neutralizer at high angle, it keeps the same polarity for indefinite time if the neutralizer is not moved and the speed is not reduced. Humidity appears to have little effect on the machine, at least much less than in an open machine. It also produces little ozone, as there is almost no sparking on its structure.

The switches that connect the charge collectors to the terminals have little utility. They increase the spark length when the machine is with reverse polarity (negative at the ball terminal) when open and reduce a bit the occurrence of "failed sparks" when the terminals are at large distance if slightly open, but their main utility appears to be to allow touching of the terminals while the machine is still running without shocks, but the neutralizer can be used for this, short-circuiting the charge collectors when almost horizontal. The switches for the Leyden jars serve to disconnect them. The terminals produce a nice display in the dark when the jars are disconnected, with a thick plume of corona flowing from the positive ball terminal to the plane terminal.

The spark length and current observed are consistent with what can be expected for this machine. The current is quite high, but I don't have yet a good measurement of the relation between current and disk speed. The ball-plane gap doesn't require much voltage to generate long sparks, and so the actual voltage generated by the machine may be not so high. More measurements will eventually be added here.


The machine works well and reliably, with impressive output. The motorized operation and the ball-plane terminals, however, turns difficult a clear comparison with my other machines of similar size. It doesn't look better than a double Voss machine, that is self-starting without switching, and reaches almost the same output with manual operation. The machine would work well with manual cranking too, as it's very easy to turn. Its output current is theoretically similar to what can be obtained with a sectorless machine (as a Bonetti machine) with one or two rotating disks, turning at the same speed, but the output voltage of a Bonetti machine is higher. It is certainly better than a Holtz machine, that has a similar structure but is quite unstable, and is also better than a Wimshurst machine, that produces less current and less voltage. The high insulation of the machine turns it suitable for demonstrations where the humidity level is high, as it is practically insensitive to it.

The structure has some problems, particularly of keeping the brushes touching the buttons. The disks and brush supports are mounted on separate structures, and natural dilation and shrinking of the wood base causes the distance to change a bit as the machine is moved from one ambient to another. The fixation of the back disk could be more solid. It vibrates significantly when the machine turns at high speed, even after the rotating disk was balanced by gluing lead blocks to its edge. The disk has also a tendency to turn a few degrees out of position, because its back support is close to its center. Some wood parts had to be readjusted some time after construction, also due to shrinkage of the wood. In particular, the balls that hold the terminals, that deformed visibly and locked the terminal rods. I had to experiment with several materials for the belt that drives the disk. Rubber cords broke easily, and sewing machine leather cords were too thick. A leather shoe string joined with cyanoacrylate glue gave good results, but got loose after some time. I ended using a green polyurethane cord, that can be joined by melting the ends. Adequate material for the brushes required also some experimentation. I started with thin ni-cr wires, but they quickly broke. I tried then silver foil strips, but they were soon cut too short by the buttons. With carbon fiber or conductive rubber brushes, the machine refused to start. In the pictures (2002) I was using using brushes of a springy grey unknown material, found in a discarded laser printer, that worked well but didn't last much. I am now (April 2006) using ni-cr wire wrapped over embroidering line, as in most of my other machines.

More photos: 12 cm sparks, front view, side view, back view.
Drawings: Front view, back view, top view.

12.5 cm sparks

Created: 23/04/2002
Last update: 29/11/2010
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
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)