A "Triplex" Wimshurst Machine

TriplexThe "triplex" Wimshurst electrostatic machine [5] is basically a Wimshurst machine with three disks. The two outer disks turn in one direction, and the central disk turn in the other direction. Neutralizing bars cross the two outer disks in parallel and another neutralizer, with opposite inclination, touches the sectors of the central disk. The output is taken from the three disks at the horizontal diameter. The central disk has its sectors mounted between two insulating disks.For simplicity of construction the central disk can be made as two conventional Wimshurst disks placed at small distance. The machine built in this way is then just a pair of Wimshurst machines operating side-by-side, at small distance, and interconnected in parallel.

The machine can be built with a central disk with embedded sectors accessed through the edge, as in this highly efficient machine [46], designed by H. Wommelsdorf in the early 1900's, but the construction with two central disks allows the construction of a sectorless machine, as this, built by S. Klein by 2000.

The interaction between the electric fields around the two disk pairs allows a greater charge density in the surface of the central disks than would be possible in a single machine. The result is that the triplex machine produces three or four times more current than a single Wimshurst machine with the same disks and speed, while two separate machines at most double the current.


In the first months of 2000 I made one of these machines, reusing some of the materials from a dismantled symmetrical Toepler machine. The machine has four identical white acrylic disks with 2.5 mm of thickness and 36.5 cm of diameter. The two outer disks were mounted with screws on nylon bosses with pulleys, that turn on pairs of ball bearings over a central steel axle. The two central disks were mounted on a short brass cylinder, that was fixed to the axle by a set screw. The axle turns on ball bearings mounted in the upright wood supports. Three pulleys were mounted in another axle below the disks, supported by nylon bushings in the upright supports, and turn the two outer disks in one direction, and the upper axle with the two central disks in the other direction, through a crossed cord in the pulley that also serves as crank, at the outer side of the upright supports. The disks turn all at the same speed. The cords were made of leather. The lower pulleys were made in wood, with central brass cylinders, and fixed to the lower axle by set screws. The speed multiplication in the pulley system is of 4:1.

The disks were sectored with 32 sectors in each disk, each sector 7.5 cm long, with maximum width of 2 cm, keeping constant distance along the radius of the disks. One pair of disks was sectored with adhesive aluminum foil, and other with foil taken from discardable pizza pans glued with contact glue. I wanted to see if the two materials show some difference in performance (no difference was observed).

The neutralizers were mounted on brass rings that can be rotated over brass pieces screwed to the upright supports, that also fix in place the ball bearings that support the upper axle. Set screws at low pressure allow easy positioning of the neutralizers at any angle. The neutralizers were made of brass bars screwed in the rings, that support other bars fixed to them by aluminum cylinders with screws. The neutralizer brushes for the outer disks were made of thin nickel-chrome wire fixed in place by plastic beads. The brushes for the central neutralizers are double, and mounted on additional thin brass bars fixed to the horizontal sections of the bars by aluminum blocks with screws. The central brushes were made of brushes of fibrous material taken from the output slot of a discarded laser printer. Wire brushes there would be too short and break easily. Small plastic beads were placed at the ends of the central neutralizer bars. The central collectors can be rotated to experiment the idea that more output current can be obtained with the charge collectors displaced in the direction of the adjacent neutralizers [5] (works) or even removed, what should keep practically the same output current (it drops a bit).

The charge collectors were made of brass bars interconnected by wood balls and a short brass rod, and terminated in aluminum balls at the outer sides and by small plastic beads at the inner sides. The central charge collectors were mounted in aluminum balls that slide over the rods that interconnect the outer collectors, fixed in place by manual screws with aluminum ball heads. The collectors have brushes of the same fibrous material used in the central neutralizers, ending close to the disk surfaces, instead of points. They appear to work as well as metal points, and there is no risk of scratching the disks in accidental touchings. I used four brushes in each bar.

The charge collector assemblies are fixed by aluminum blocks with manual screws to aluminum tube conductors. The conductors cross and are supported by wood balls mounted over long insulators made of PVC tubes closed at both ends with nylon blocks. The lower blocks serve as base for the columns, and the upper blocks insert in holes below the wood balls. Set screws in the balls fix the conductors in place.

The spark terminals were made with brass balls made by metal spinning, with 3.2 cm of diameter. The balls are screwed into short brass plugs that slide into long aluminum tubes. This allows for easy replacement of the balls for experiments (note the different aluminum balls in some of the spark images). The tubes cross wood balls and are connected to slotted brass plugs that insert into the horizontal conductors, and can be rotated to any position. The insulating handles are sections of thin PVC tubes, adapted by nylon blocks to the tubes.

The support structure was made of four wood bars fixed with screws, and two upright supports fixed to the base by pairs of screws from below. Front view, back view, and. side view. of the complete machine.


The machine self-started easily, and produced a lot of current. Its output current reached 50 ľA at two turns per second at the crank. This is four times the expected current for a Wimshurst machine with the same disks and speed. Really, with one of the lateral disks stopped, the output current reached about 12 ľA only. At the maximum possible crank speed of 4 turns/second, the machine reached 100 ľA. This current was measured connecting a microamperimeter across the terminals, through two long wood bars, serving as resistors. The same current could also be measured between one terminal and the neutralizers (ground). This last form of measurement keeps one of the sides of the machine at high voltage.

It was soon evident, however, that the sectors were too close together, as the machine was easily sparking through the disks, from one charge collector through a series of sectors to a neutralizer, and from one of the other neutralizers through another series of sectors to the other charge collector. Long sparks at the output could only be obtained with the neutralizers at an almost vertical position. In a dry day, the machine could produce eventual 15 cm sparks, but normally not more than 8-12 cm sparks before sparking through the disks.

In October 2000, I removed half of the sectors of the disks, leaving only 16 widely spaced sectors in each disk. This eliminated the problem of sparking through the disks completely. The machine then produced consistently sparks with up to 15 cm with its normal terminal balls and neutralizers at little more than 60 degrees with the horizontal. Attempts to produce longer sparks result in sparking through the center of the machine, directly between the charge collectors. The output short-circuit current dropped to one half of what it was (25 ľA at 2 turns/s at the crank), but the current measured in high-voltage conditions, from one terminal to the neutralizers, reached 35 uA at the same conditions. This indicates that at low output voltage the charges are transported in the sectors, but at higher voltages some of the area around the sectors is also effectively used. A dislocation of the central charge collectors in the direction of the adjacent neutralizers, of just 2-3 cm, increases the output current back to the original level, at the expense of a small decrease in the maximum spark length.

In May 2002, I replaced the leather cords (shoe strings) by 3 mm polyurethane cords, to solve a constant problem with the cords becoming loose or breaking. The crossed cord driving the central disks, in particular, was problematic because the tension on it is at least twice higher than in the other two, and this machine is hard to crank when charged.

The replacement solved the problem, but the machine still has some mechanical problems, that will be solved in a future rebuilding. The pulleys should be larger, and the support structure should be stronger for a machine of this size. Also, the distance between the outer brushes and the disks should be adjustable, and the charge collectors points could look better.

The machine, in front view, back view, and side view.

Multiple sparks

Created: 22/10/2000
Last update: 16/2/2003
Developed and maintained by Antonio Carlos M. de Queiroz

Return to Electrostatic Machines