Electrostatic Machines

Electrostatic machines are electromechanical devices that produce "static electricity", or electricity at continuous (DC) high voltage. They were fundamental in the early studies about electricity, started in the XVII century, in the form of "friction machines", and their development culminated at the end of the XIX century with the development of powerful "influence machines". Today, some specialized uses for them continue to exist, but they are mostly seen as demonstration devices in physics laboratories, with much of their history forgotten.

I started experimenting with these machines by 1973, building a first series of machines. With this I learned a lot about electricity, and I still think that all people interested in electricity or electronics shall try these machines to get a real feel of the subject. At least, high voltage static electricity is something that you can see and feel. Eventually I abandoned the subject for several years, but in 1996 I renewed my interest in this subject, started to study and build new machines, and set up these pages.

Below are pictures and descriptions of my old machines, of machines that I built more recently, of machines built by others, pictures from old books and papers related to electrostatic machines and other high-voltage devices, and also some pictures from museums. There are also extensive references, covering classical and new materials. This site is always in construction. I plan to add more details about the machines depicted and historical material, as soon as I find or receive more data from interested people, build and experiment with new machines, and have time.

Está também disponível uma seção em português.

To navigate through this site efficiently, use the "open link in a new window" function of your browser to see the pictures in the links, and use its search function for searching. Recent changes.

"Ignis ubique latet, naturam amplectitur omnem"

Machines of Toepler, Bonetti, Voss, Bohnenberger, and Nicholson

My machines

A Wimshurst machine [1][2] that I built in 1974. Front view, back view, and with two Leyden jars. A schematic diagram, with the disks represented as cylinders, and a description of how the Wimshurst machine works.
A Ramsden friction machine [2], built in 1975. Small (18 cm acrylic disk), but useful to test the insulation of materials and for for starting the influence machines in humid days.
A Lebiez machine, or simplified Voss machine [p31], in front view and in back view. Built in 1975 as a kind of Voss machine, and rebuilt in 1996 in this form. Schematic diagram, with cylinders instead of disks for my machine. This machine is equivalent to Lord Kelvin's "replenisher" (see below), with better insulation. The classic Voss machine, also known as Toepler-Holtz machine, is better and is built in this way, with charge collectors and inductor plates separated. A possible similar true Voss machine is shown here and here.
A small cylindric simplified Voss machine built in 1997, with the same structure of the previous machine. Side view, and another view. It is similar to a Dirod machine [10].
A symmetrical 2 disks Toepler machine [4], with some modifications, built in 1997. Side view, another view. A drawing. A schematic diagram, with the disks shown as cylinders. This machine has excellent performance, and can generate higher voltage and even more current than a conventional Wimshurst machine with the same disk size. In 1999, I built a larger machine. Look at the bottom of the page here for a description.
The first classic Toepler machine (1865) was built in this way [4][9], with a different interconnection and disks with two sectors only. Toepler described also a symmetrical machine (1866) that is very similar to my machine (the picture shows a sectorless machine and a similar device used as voltage multiplier). schematic diagram [p39].
My first sectorless Wimshurst machine, or Bonetti machine [4][5][8] (ray-tracing drawing), built in 1997, with 31 cm disks. A drawing. Another view. Pictures of the actual machine, in front view, and back view. A detail of the charge collectors and neutralizers. Pictures (video frames) of sparks from this machine: A short spark, a long spark with a loop, and a longer one. The original Bonetti machine (1894) [31] used series of brushes as neutralizers instead of combs.
A Holtz machine [2][4]-[7] of the first kind, that I built in 1997. A drawing. Another view, and a schematic diagram. Pictures of the actual machine, in front view, and back view. This machine was the first really powerful influence machine, invented in 1865, and was very popular, even requiring external excitation to start. An apparently complete actual machine is here. Some additional pictures, from books by H. Pellat: A Holtz machine [6] (the fixed disk is in a wrong position), a better picture [7], a double Holtz machine [6][7], and a machine with neutralizer and friction starter [6][7]. And another good picture [14]. This picture shows a machine with a more modern structure [22]. A multiple machine [24]. Here is a picture of the Holtz machine of the second kind, that uses two counter-rotating disks, as the Wimshurst and Bonetti machines. A drawing of a possible machine.
The Leyser machine (1873) [4][19], is variation of the Holtz machine with the output taken at positions that would be under the inductor plates in the regular machine, and the neutralizer bar where the original output circuit would be. This is the schematic diagram of the machine, with a cylinder instead of the disk. This is a different design due to Weinhold (1887), with wood inductors and no insulating plate [19]. This diagram [19] shows how it operates. Initial plans for a machine that I have built are here, in front view and back view. The version that actually worked was somewhat different.
A double Voss machine, or double Toepler-Holtz machine, with classical structure, built in 1998. A drawing, and a photo of the machine. This is a good self-exciting machine, invented in 1880. With 27 cm rotating disks, it produces sparks with up to 10 cm and more than 50 uA of short-circuit current.
A Bohnenberger machine (1798)[4], that I built in 1998. A drawing, and a picture of the machine. An ancient machine of the "doubler" type, it is not a powerful machine, but is very interesting. See more about "doublers" in the section about influence machines below.
In June 1999 I made experiments with a bipolar Van de Graaff generator, (drawing) similar to the original machine, but smaller.
This is a large Bonetti machine, that I started to build in December 1999. The disks are old ebonite disks that come with the Radiguet & Massiot Bonetti machine that I recently restored. A drawing of it. Front view, and back view. Another view of it, and another. Some sparks, that may reach 20 cm..
In April 2000 I finished a Wimshurst triplex machine. (drawing). It's a double Wimshurst machine, using the close proximity between the central disks to increase the output current, through greater induction and mutual shielding. Pictures of the machine, in front view, back view, and side view. The machine produces a high current (100 uA with the 36.5 cm disks rotating at 16 turns per second, 4 times more than a single Wimshurst machine with the same disks). With the original design, it reached only 8-12 cm sparks, eventually reaching 14 or 15 cm on dry days, because with the rather small separation of the sectors it easily sparked through them and the neutralizer bars. With half of the sectors removed, it produces consistent 15 cm sparks.
By the same time, I made also a working version of Bennet's doubler, a curious simple influence machine.
In August 2000 I made a version of Nicholson's doubler, the first automatic influence machine (1788).
In January 2001 I completed a double Wommelsdorf machine, following closely the original design of [p84] (1920), but with modern materials. Front view, and back view. A collection of parts. Partial assembly. Assembling the neutralizer. Neutralizer and other details assembled. Back view. The machine, almost complete. Back. The disks, and the switches. The complete machine. back view, side view, other side, and another view, with only painting missing. The machine works quite well (13.5 cm sparks, 100 uA current) for the two 28 cm disks.
In March 2001 I made a curious AC electrostatic machine, apparently new, that I named as "half Wimshurst machine".
And by April 2001 I extended the same idea to a three-disks machine, that I named as "unfolded Wimshurst machine".
A Wehrsen machine, completed in April 2002. It is a prototype for a large Wehrsen machine (see below), that I started to build in August 2001. Some parts for it at the start of the construction. Almost complete, by March 2002. Back view. Working, it performs quite well, with 11 cm sparks and 70 µA of short-circuit current.
An electrostatic linear motor, completed in January 2002.
A large Wehrsen machine, first tested by August 2003. Almost complete by July 2003. Another view, back view. Ray-tracing picture. Just before the first test. First test. The machine is still without its definitive rotating disk due to construction and insulation difficulties.
This is a Toepler-Dirod machine that I was building by March 2004, still without spark terminals, and with terminals. It's connected as the symmetrical Toepler machine, but uses Dirod-type disks. A drawing of the final machine. The machine works, but is weak.
Bohnenberger's Bennet's doubler. A curious little machine that I built in April 2004.
Bohnenberger's Nicholson's doubler. A version of Nicholson's doubler with back-and-forth movement, built in May 2004. So far not so good as the other doublers.
A rotating Bennet's doubler, with a new construction using sectored disks, built in 2006. Photo of the machine. Another view.
Wilson's machine, the oldest influence machine with symmetrical output, in a reproduction made in 2007. A good and interesting machine.
A Van de Graaff generator with external belt. A big machine for "hair rising" demonstrations, completed in 2007.
Bennet's doubler with reciprocating levers. A new structure for Bennet's doubler based on Wilson's machine, built in 2007.
An "electrostatic orrery". A classical demonstration device, built in January 2009. It is quite heavy, but works. This site shows some similar old devices. Video.
Henley electrometers built in 2009. Classical instruments first described in 1772. A video of them operating with an electrophorus. Videos of experiments with an electrophorus: 1, 2.
A Wimshurst machine with insulated sectors, built in 2008-2009. This machine is highly insensitive to humidity.
See also the comments about machines that I have restored, in the section about influence machines below. Of special interest, the Ducretet and Roger and Radiguet and Massiot machines.

Machines built by others

A big Wimshurst machine built by Jim Banas.
A sectorless Wimshurst machine with 60 cm disks. This machine was built by Ed Wingate. A spark from this machine. Another sectorless Wimshurst machine, with 30 cm disks. Similar to the one described by R. A. Ford in [8]. A more recent picture. Side and base view. Neutralizers, Charge collectors. Another view. This machine was built by J. Hardesty and Ed Wingate. Photos sent by Steve Cole.
An old Wimshurst machine repaired by Johannes Zolk in 1996, with the original broken disks replaced by LP records, with good results. Front view, and back view. Photos sent by J. Zolk.
A "shake-sphere" machine [10], built by Joachim Bolz and his students in 1997. It is an influence machine using two balls in a tube, moved by shaking the tube, instead of disks. It works as my Toepler machine above. An schematic diagram of it. Photo and drawing by J. Bolz. Operation of the machine.
Complete plans for a beautiful Wimshurst machine, built by J. M. S. van Gelderen in 1997. Plans for the disks, a top view, a back view, a side view, and details of the terminals and Leyden jars. Pictures from the machine, seen from the front and back sides, and from above here and here.
Ricardo "Rike" built this Wimshurst machine in 1997, using LP records for the disks. It produces 7 cm sparks. Another view.
A beautiful large Wimshurst machine (40 cm disks), built by James T. Garavuso in 1998. A frontal view, another, a side view, a back view with the terminals in storage position, another, and a view from above. Details if the charge collectors, neutralizers, and secondary spark gap. This machine produces 12 cm sparks.
A Toepler machine, built by Maximiliano Guzman, from Spain, in 1998. The disks have 27 cm of diameter. A later version used larger shields and a speed multiplier in the crank.
A Wimshurst machine, built by Raymond Zaborski, from the USA, in 1999. The small intersector distance and the neutralizers at low angle result in intense current and easy self-excitation, but relatively small spark length.
A big motorized Bonetti machine, built by Emery Wayman, from the USA, in 1999. The machines has disks with 61 cm of diameter, and produces sparks with up to 28 cm of length. Some sparks from the machine: 1, 2, 3. The terminals balls have 7.5 cm of diameter. Mr. Wayman has built also a similar, smaller, machine with motors driving directly the disks.
A motorized 2 disks Toepler machine, built by Roger Magnuson in 1999. The disks have 20 cm of diameter. Another view. Note the small Leyden jars built in the terminal supports.
A classic Wimshurst machine, built by Ronald Coleman in 1999. Detail of the crank, and of the charge collectors. The machine is prepared for an upgrade with larger disks.
A Wommelsdorf condenser machine with double rotation, built by Bert Pool, following plans in a thesis written by Wommelsdorf in 1904. Another view. This machine is a compressed version of a multiple Wimshurst machine, with sectors mounted between pairs of insulating rings, interconnected through the external or internal edges of the rings. One set of rings/sectors is held by the inner side and the other by the outer side, and both turn in opposite directions.
A well built small Wimshurst machine, made by Harry Boneham, from Canada. The support structure was machined from aluminum, with the disks having 18.5 cm of diameter. Another view.
A Wimshurst machine, built by Terry Baines, from England, in 1999. With 30 cm disks, it produces sparks with 3 cm.
A Wimshurst machine, built by Alex Rice, from England, in 1999. The machine has 32 cm disks, and produces sparks with 10-11 cm of length. The spark is a double exposure. A spark from the machine. In 2000, he built an improved machine.
A Wimshurst machine, with 18" acrylic disks, built by John Clark, from England, in 2000. It produces 3" sparks.
Dan Bowlds, from Kentucky, USA, designed this original machine. A bare disk rotates behind an insulating plate, that holds four wood blocks painted with conductive ink. The lateral blocks are connected to blades collecting charges from the back surface of the rotating disk, and to Leyden jars made in the supports. The upper and lower blocks are inductors, and are charged from the terminal blocks through single corona points, also made of wood. Opposite to the inductor blocks there are interconnected neutralizer blades. The terminals are directly connected to the Leyden jars in the supports. An elegant structure for a small motorized machine (the disk has 6" of diameter) that works essentially as the Voss machine. The machine requires an initial charge to start, and produces sparks with 1" of length. Back view, Lateral view.
A Wimshurst machine with acrylic structure made by Scott Nagel in 2000. With disks with 14.5" of diameter, it produces sparks with up to 6". Another spark. Note the separated small balls in the positive terminal, and the good dimensions for the sectors in the disks. The charge collectors, with some sharp corners, were later modified.
What is probably the largest working Wimshurst machine was built by Paul Hendriksen in 2000, for a technical show by ROVC , in the Netherlands. The machine uses two glass disks with 2.15 meters of diameter (2 cm more than the large machine built by Wimshurst in 1884), 12 mm thick. The output voltage reaches 1 MV, producing sparks of up to 1 meter. It turns at up to 100 rpm, producing 10 uA of current. The output voltage is too high for Leyden jars, and so two copper globes are used as distributed capacitors. Details of the driving pulleys and a curious discharging mechanism. A long spark, another, and more sparks. Eventually (by 2008), the machine was installed in the Technorama museum, in Winterthur, Switzerland. Mr. Hendriksen also build several other machines: A large Van de Graaff generator, 2.5 meters tall, that produces 80 cm sparks. A sectorless Wimshurst machine with 50 cm disks, that produces 28 cm sparks. It uses an adjustable capacitor, seen here with its minimum, medium, and maximum capacitance, that allows control of the intensity of the sparks, between a minimum and a maximum, continuously. The machine uses charge collectors at just one side, and to start it a simple friction rod is used. A hand-cranked Van de Graaff generator, with an unusual toroidal terminal. A small friction machine, with a 25 cm disk, that produces 2.5 cm sparks. A curious electric clock, powered by high-voltage between the two balls below. Detail. Several high-voltage demonstration devices.
A big Wommelsdorf condenser machine with 10 55 cm disks was built by Serge Klein, in France, in 2000. It can produce 25 cm sparks and up to 0.7 mA of current. Frontal views, from the left and right sides, a view of the motor that turns it, detail of the disks and inductors, and another view. The disks are composed of three disks glued with epoxy glue, with the central disks separating two sets of intercalated sectors. The inductor plates are also enclosed between plastic sheets glued with epoxy glue. It works better with the neutralizer brushes removed, with the gap between the disks and the neutralizer bars making the role of the gap in the neutralizer circuit of the classic machines. The machine was later upgraded to 12 disks, with better brush supports, in an attempt to increase the output current. A spark from the machine. Mr. Klein has also built other machines, as a Dirod, a Wimshurst machine, a big Bonetti machine, that produces 30 cm sparks, a machine similar to a Felici machine with disks and operating in open air, and a triplex sectorless Wimshurst machine. Another view.
A nice Wimshurst machine, built by Julian Phillips, in New Zealand, in 2000. With 30 cm disks, it can produce 7 cm sparks. Another spark, and a description of it.
A very simple setup was developed by Michael Foster, in Los Angeles, USA, in 2001, to produce long sparks by frictional electricity. He used nothing more than a long PVC tube, a paper towel, a very simple Leyden jar capacitor, and a special positive terminal to excite long sparks. A description of his procedure.
a Wimshurst machine, built by Luca La Valle, in Rome, Italy. He built also other high-voltage devices, as a Van de Graaff generator and a Tesla coil.
A curious small Wimshurst machine, designed by Fausto Gazzi, in Bologna, Italy. Mr. Gazzi deals with ancient instruments, and frequently makes restorations, as of this 4 disks Wimshurst machine.
A nice Wimshurst machine, built by Chris Kitching, from England, in 2001. Top view, detail from the charge collectors, and a spark with 14.5 cm produced by it. The acrylic disks have 36 cm of diameter and 4 mm of thickness, and are mounted on nylon bosses. The balls at the spark gaps and joints are softened steel balls.
This and this Bonetti machines I found at eBay. They are similar to the machine described by R. A. Ford [8]. Builders unknown.
Tony J. Meijers, in the Netherlands, built this nice Wimshurst machine. With 37 cm disks, it produces 14 cm sparks. Note the driving system, without crossed cords. Front view. Back view. He built also this Triplex Wimshurst machine, in 2000, that with 41 cm disks produces 24 cm sparks. It also has a curious implementation of the driving system, with the driving axle making an angle of 10 degrees with the upper axle, so the crossed cord that drives the central disks don't touch itself at the crossing. Front view. Back view. Side view. Other view. Assembly of the disks. A thick 24 cm disk at the center and disks at the outer sides impede sparking to the center of the machine. The Leyden jars also have increased insulation.
Georges Hublart, from France, built this Wimshurst machine, motorized and with a curious construction. Side view. With 33 cm disks, it produces 16 cm sparks. Note the chains driving the disks. He has also other high-voltage devices, as a Van de Graaff machine.
A Wimshurst machine, with conductors insulated within PVC tubes and LP record disks covered by adhesive plastic foil, built by Ben Noviello, USA, in 2002. It produces 10 cm sparks.
A Wimshurst machine, built by Rod Heidel, from the USA, in 2002. With 20 cm disks, it produces 5 cm sparks. The frontal structure is a capacitor.
A beautiful Wimshurst machine, built in cherry wood and brass by Gerald J. Schaefer, from the USA, in 2002. The disks have 18" of diameter. Side view, Frontal view, With two demonstration devices. An intense spark from it.
A symmetrical Toepler machine, built by J. Keverline, from the USA, in 2002. With 30.5 cm disks, it produces sparks with up to 16 cm. The disks have increased insulation with a material used to insulate tool handles. This resulted in voltage high enough to pierce the spark shields, that had to have their thickness increased to 4 mm.
A Wimshurst machine that was once used for demonstrations at the Science Museum, in London, England, restored in 2002 by Rob Skitmore.
A large Bonetti machine, built by Karl Kehrle, in Germany, in 2003. With 80 cm polystyrene disks, it produces 63 cm sparks, between a pair of aluminim balls (8, 12 cm) at the positive terminal and a 30 cm styrofoam ball covered with aluminum foil at the negative terminal. The glass Leyden jars have 720 pF each. Mr. Kehrle wrote a book [49] showing experiments with a similar sectored machine, that with 90 cm disks produces 47 cm sparks.
A Toepler machine with 48 cm disks, built by Alain Tramasaygues, from France, in 2003. This improved version, with the inductor plates mounted inside a box, worked better. This is a curious Van de Graaff generator also built by him, that can produce 30 cm sparks. This is his Van de Graaff with external belt. He also built a Dirod machine. Another view.
A sectorless Wimshurst machine, with 60 cm disks, built by Grant Vincent Wells, in New Zealand. It can be operated by hand or by a motor, has an electronic startup system, and produces sparks with up to 24 cm.
These two machines were built by Alan Kerley. The larger machine is a Voss machine with a 21" and 17.5" disks, and the other is a small Wimshurst machine made from CD disks.
This is a Wimshurst machine made by Keith Stuart, by 2000, in New Zealand. It produces 10 cm sparks. He also restored an old machine (probably German, from around 1900) for the Auckland Museum of Transport and Technology. Front view, back view. By the end of 2003, he made a curious combination between a symmetrical Toepler machine and a Dirod. Side view, other side, top view, end view. With 12 cm disks, it produces 4 cm sparks.
A motorized Wimshurst machine, made by Thomas Rapp, in Munich, Germany, in 2004. Another view. The disks have 30 cm of diameter. More informations and other projects can be found at the author's site.
A Van de Graaff generator, made by Richard Linder, in Burlington, USA. The terminal is a stainless steel sphere with 45 cm of diameter. The bottom roller is made of Nylon, and the top roller of Teflon. The belt is made of 0.4 mm Mylar foil. Mr. Linder makes demonstrations using it at the Burlington Science Center. For the 2004-2005 school season, he built a larger machine, with a 36" terminal. The 6" belt is made of vinyl impregnated nylon. It produces arcs with 18" to 24" to a 1.5" grounded sphere.
A Wimshurst machine, built by Ricardo Triches, in Brazil, in 2004. Another view.
A big Van de Graaff machine, built by Harold Pollner, in California, USA, in 2004. The terminal has 30" of diameter, the comumn is 9" PVC, the belt is made with 4" Neoprene, and the machine is powered by a 1/4 hp 1725 rpm ac motor. Excitation is by rolling friction between the belt and the lower roller, that is a 4" PVC coupling mounted over a wooden core. It produces 22" to 27" sparks, but from the rim of the sphere opening to a grounded target electrode positioned below the sphere, (as in the picture). Sparks from other points of the sphere reach only 6" to 7".
A small Wimshurst machine, with 20 cm disks, built by Hannu Eloranta, from Espoo, Finland, in 2005.
A nice Wimshurst machine, belonging to Dr. Alistair Miller, England. The machine has 19" disks and produces 6.5" sparks. It was built by Anthony Swift, that runs a museum dedicated to Victorian science in North Yorkshire, England.
A motorized Wimshurst machine, built by Peter Bradley, in England. Spark picture. Another spark.
A curious friction machine built as a Gramophone, by Kaj V.M. Heiden, in the Netherlands. A spark.
Several Wimshurst and Bonetti machines built by Jarrod Kinsey. Another view. He experimented also with Wimshurst machines with sectors made with conductive ink with excellent results. Some laser experiments powered by electrostatic machines. An interesting comparison of sparks. This is a Holtz machine completed in January 2009, with simple construction but excellent performance, as seen on this video. Another view.
A Wimshurst machine, built by Christophe Branger, in France, in 2006. Another view, and another. Spark, another spark.
A Wimshurst machine, built by Emiliano Salinas Covarrubias, from the Universidad de Sonora, Mexico. The acrylic disks have 40 cm of diameter, and the structure is made of polystyrene. It procuces 6 cm sparks.
A big Bonetti machine, made by Hal Pollner, in the USA, in 2006. With 25" disks, it produces 11" sparks. A Van de Graaff generator is used to excite the machine. Another view.
A well built Wimshurst machine, built by Leonardo Giacomelli, in Italy, in 2006. All parts are made in machined metals and acrylic. Front view, charge collectors, lower pulleys, upper pulleys and insulated neutralizers, and top view. The disks have 40 cm of diameter, and it produces 16 cm sparks.
Two Lebiez machines, one hand-cranked and other motorized, made by Milan Manchich in 2007.
Two Wimshurst machines made by Brian Philips in 2007. First machine, another view. It uses a flat capacitor instead of Leyden jars. Second machine, another view.
Several machines and other devices, built by Luiz Alberto Feijó Junior, in Brazil.
A Wimshurst machine built by Vaughn P. McDowell by 1986. Back view, side view.
A set of machines, Wimshurst, Voss and Ramsden, built by Leonardo Cannone, from Italy.
A big Wimshurst machine, with 61 cm disks, built by Haywood Turner, from the USA.
Wimshurst machines and Van de Graaff generators, built by Harry McCarty, from the UK.
An electrostatic motor, built by Dan Bowlds, from the USA.
A Wimshurst machine, rebuilt by Rod Heidel in 2008. The original is described above. One of the Leyden jars exploded in a test due to glue vapors inside the PVC tube. So, take care with ventilation on these constructions...
A Wimshurst machine, built by Carlos Alberto Vargas Alfaro, from Peru, in 2008. There are some videos here.
A Wimshurst machine built by Kevin Acres in 2008. Side view. The machine, compared to another machine built by his grandfather 70 years ago. Back view. The machine has 27 cm disks, and produces 8.5 cm sparks.
A sectorless Voss machine, built by David Hodges, in 2008. It uses combs in the charge collectors, neutralizers, and inductor chargers.
A Wimshurst machine, built by Rosalino Trobbiani, from Italy, in 2008.
A beautiful Wimshurst machine, built by Jon Bodsworth, in England. The machine was originally built 25 years ago, and recently (2008) reformed. Front view, side view. The glass disks, coated with shellac, have 22 cm of diameter. The machine produces 6 cm sparks. The structure was made with mahogany and brass, the insulators were made with polyester resin, and the sectors were cut from aluminum litho plates.
A Voss machine, restored by Alan Recce in 2009.

Friction machines

The first electrostatic machine [15], was built by Otto von Guericke [16] by 1663, using a sulphur globe frictioned by hand. The globe could be removed and used as source for experiments with electricity. A picture of a working replica of the machine, from the University of Oldenburg.
Another important early researcher was Francis Hauksbee, that built several machines using glass globes [50][53] and cylinders by 1705.
The friction machines were gradually improved through the works of many researchers. This is the machine with a glass globe of the abbot Nollet (~1740) [7]. Eventually, the machines took a stable form, with leather friction pads (Winkler, 1744), glass globes (Bose, 1751), and insulated charge collectors. Demonstrations with these machines were common.
Watson's machine [51][52] (1746) had a large wheel turning several glass globes. The prime conductors were a sword and a gun barrel suspended from silk cords. Watson made many experiments with the Leyden jar, then recently invented.
A Ramsden electrostatic friction machine [2]. Another picture [7], another [12], a good drawing [17], and a picture of a large machine [14]. The first popular machine using a disk (1766). Designed by J. Ramsden, an instrumentist that also designed many other good instruments in the 1700's. A beautiful restored Ramsden machine, found at eBay in 1999. Photos by Fausto Gazzi. This large machine I found in a museum in Geneva, Switzerland. A simpler machine built by myself is shown in the first section of this page.
The machine of Le Roy (1772) [50][p26] was suitable for the production of long sparks, due to the high insulation between the friction pads and the charge collectors (see a more modern version as the Winter machine, below).
This large disk machine (1785) with 1.6 m disks can be seen at the Musée du Conservatoire Nacional des Arts et Metiers, In Paris, France. On its base is written the motto at the top of this page. There is a picture of it in [21].
A brass model of a Ramsden machine. A curious decorative object, possibly from the 1930's or before. The disk has 3.5" of diameter. Photos sent by Blake Awbrey.
A Nairne electrostatic friction machine [7], built in 1770, consists in a glass cylinder, a friction pad in one side, and a charge collector in the other, both connected to insulated conductors. Another one. The machine was used for medical purposes.
The van Marum electrostatic friction machine (1784) [9]. By moving the two curved bars with charge collectors, it was possible to collect charge from the disk (bars as shown), or from the friction pads (bars turned 90 degrees), producing voltage with any polarity, as shown here. Van Marum is also known for the big machine [16][21] that he had made in 1784, that is now in the Teylers museum.
A similar machine, now in the Deutsches Museum, Munich, Germany, belonged to Georg Ohm (1830?)[21]. Another view. Photos sent by Hans Bussmann (I could not find it there in september 2008).
A belt machine [50] built by N. Rouland by 1785, had a charge collector with blades that collect charges from a silk belt rubbed by two grounded tubes covered with hare fur [21].
An old friction machine using a glass disk. Another picture of the same machine. Photos sent by Don Day.
A Winter electrostatic friction machine. One of the most efficient friction machines. A picture from an old book [3], and another, from H. Pellat [7]. This was the last popular structure for friction machines, as shown in these catalog pictures from the 1920's: this and this are from [17], and this is from [22]. The characteristics of the machine are the disk frictioned at one side, at both faces, with a pair of charge collectors at the other side, shaped as rings with points turned to the disks. Sometimes a large wood ring (Winter's ring) with a metallic core was attached to the terminal, increasing its capacitance. A double adjustable version can be seen on the first picture.
The Woodward machine (1840) [43][21] was a modified Ramsden machine, with the prime conductor located above the disk, or disks, saving some space. It could also generate negative voltage, by mounting the upper friction pad in place of one of the charge collectors. This double machine is in the University of Porto, Portugal. Partially disassembled. Pictures by Marisa Monteiro.
The Armstrong hydroelectric machine [2], a friction machine using steam as charge carrier (1840). It is just an insulated boiler producing a steam jet mixed with water droplets. A better picture is here [9]. Very powerful machines of this kind were built for research.
The Lorente generator. A triboelectric machine composed of four cylinders that roll together without friction, under a slight pressure. The two outermost cylinders are metallic, and the two central cylinders are of distinct insulating materials (nylon and teflon). Opposite charges are collected in the metallic cylinders. The basic machine produces voltages of some tens of kV, but several modules can be stacked for more voltage. A coaxial version is also possible. Pictures from actual models are here and here. This device was invented and patented by G. Lorente, who sent the pictures.
Rolling friction is also commonly used in models of the Van de Graaff generator, although the principle of the charge generation system on those machines is a mix of friction and influence.

Influence machines

The first rotating influence machines were the "doublers". The first was Nicholson's doubler [p14] (1788). It was a rotating implementation of Bennet's doubler (1787), a device based on Volta's "electrophorus" (1775) [p110], that allowed great multiplication of a small initial charge by a series of repeated operations with three insulated plates. The original machine proposed by Nicholson didn't require a connection to ground, but versions with explicit ground connections are also possible, as this [28], and this (built by Wimshurst) [p14]. An actual machine exists the Musee d'histoire des sciences, in Geneva (the site has a movie of the doubler in operation), that looks as this machine (John Read's doubler) [p106]. See my Nicholson's doubler. A similar implementation, where the two plates that are fixed in Nicholson's device rotate, is Bohnenberger's machine (1798) [4]. Bohnenberger designed several other doublers, as this automated version of Bennet's doubler and this variation of Nicholson's doubler (1801) [p107], both operating with back and forth movements. See my Bohnenberger's machine.
Multipliers based on a different system are also possible, as Péclet's condenser (1841) that increases the charges linearly with the number of operations [p87] and a multiplier with 4 plates invented by Pfaff and Svanberg that combines addition and multiplication [54]. Multipliers directly based on the electrophorus were studied, by Volta himself [p110] and Lichtenberg.
A similar adding device was Cavallo's multiplier (1795), where a movable insulated plate was moved back and forth, alternatively being grounded under the influence of a second previously charged plate, and touching a third insulated plate close to a grounded fourth plate. After some cycles, the grounded plate would be removed, causing the accumulated charge at the third plate to rise its potential to about the potential of the second plate times the number of cycles [4].
Two Cavallo multipliers can be combined in Wilson's machine (1804), a curious machine that incorporates essentially the same idea of the symmetrical rotating machines developed much later.
One of the scientists that studied doublers was Erasmus Darwin. His "commonplace book" contains a sketch of a doubler, that appears also in [p14], that is probably the first drawing of these devices. His book "Phytologia" contains a drawing of a "pendulum doubler", attributed to Bennet (see here).
The next development was of symmetrical influence machines, using influence to generate new charges and Faraday's shielding effect to collect them. The first was Belli's machine [4][p14] (1831), the first symmetrical influence machine. A picture of an elaborated actual machine. Belli developed also a different machine using the same principle, shown here. The same basic structure appears in Lord Kelvin's "replenisher" [2][p92] (1867), in schematic representation, and as constructed [12]. A simple machine built with insulated curved metal plates, used as part of measurement instruments. Here is a ray-tracing picture similar to a machine that I built by 1973. The rotation of the central insulating bar with two metallic carriers, touching the four contacts, causes accumulation of opposite charges in the outer plates. Another similar machine was the Varley machine [26] (1860).
A device that can be considered an influence machine is Einstein's "Maschinchen" (little machine). It was a mechanical voltage multiplier resembling a multistage Belli machine without feedback. This paper describes it.
A curious machine [18] that appears to be similar to Belli's machine, if the lateral brushes are connected to the fixed plates. The same machine appears illustrating this advertisement (1962) but with an added set of brushes installed, to separate the output circuit from the inductor plates, as in the Voss machine.
The Piche machine, or Bertsch machine (1866) [7]. One of the simplest influence machines, uses an insulator plate (I), that is separately electrized by friction, and used to generate charges in the rotating disk by induction. See the original letters about this machine in the references. A similar device is the Dubrowski machine [22]. This type of machine was called "continuous electrophorus".
The Carré machine [6] (1868). A friction machine below charges by induction a fast rotating disk, that transfers charge to the upper conductor. It is similar in operation to the Bertsch machine, but regenerates the charge in the inductor. A better picture [7]. A variation [14] with slanted combs. A ray-tracing drawing. A photo from an actual machine, sent by John Newman. A machine with a double terminal. The Van de Graaff generator [p4][8] is an evolution of this machine, with a belt instead of the disk, and a more efficient charge collector at the top. This machine is in a museum in Switzerland. Another machine, in a Museum in Spain.
A double Bonetti machine from the same museum, with curiously shaped neutralizer combs (?).
Another antecessor of the Van de Graaff generator is Righi's electrometer [p55] (1872), that used a rubber string with brass rings for charge transport, and a hollow sphere as charge collector. A picture of this machine [41]. Similar machines are also discussed in [p59] (1875), as a bipolar machine, that must have grounded pulleys, and another, that adds a neutralizer circuit and can use insulated pulleys.
Righi studied also a belt machine [p59], that antecipates a regenerative charging system used in some Van de Graaff machines, and shows a curious polarity reversal phenomenon, where for some time the belt operates with bands of both polarities.
Righi designed this big Holtz machine [41] (~1875), used at the Regio Istituto Tecnico de Bologna for teaching and research. A drawing of a similar machine [42].
A Lord Kelvin's water machine (1867) [1][p91]. It is an influence machine that uses water droplets instead of rotating carriers. It works in the same way of the 2 disks Toepler machine. Another picture [6] of a similar machine. An improvement of this machine using two additional units for output, avoiding the discharge of the inductors, was proposed by Fuller in 1888 [p103] [ p62]. A different version was proposed by Sylvanus Thompson in 1887 [p104].
The Schwedoff influence machine [9][13][29] (1868). A very strange machine. The lower plates form a modified Holtz machine, with inductor plates replaced by combs (dotted lines) charging the lower surface of the lowest fixed disk with charge taken from the charge collectors. This first machine provides bias for the sectors in the lower plate of the upper assembly (the even-numbered ones with one polarity, the others with another), that form a current multiplier. The output is taken between the two insulated sets of combs over the upper, rotating disk. This picture from the original paper [p46] shows more clearly the connections. This is the charge collector [p46] that completes the machine.
A Toepler-Holtz machine, or Voss machine with classical design. From a catalog from the 1920's. Two more pictures here and here. Pictures found in the Gemmary's forum.
Several Toepler-Holtz, or Voss, machines from [17]: A simple machine, another, a double machine, a quadruple machine (see one here), and a multiple machine. Two more simple machines from [18]: this and this. And another one, from [22].
A magnific quadruple Voss machine, at the museum of the University of Pavia, Italy.
A Holtz-Wimshurst machine [4][23], simple and with a frictional starter [4][23]. These were Holtz-type machines with several disks and improved construction, as the inductors fixed in separate square glass plates, developed by Wimshurst by 1878.
The Kundt machine [4] (1868) was a mixed friction-influence machine, similar to a Bertsch machine with the back side of the disk frictioned by a friction pad with a silk flap attached, as in a friction machine. At the front side are positioned two charge collectors, as in the Bertsch machine. A similar machine was the Cantoni machine (1869), that added a third charge collector at the back side of the disk, so the machine can also be used as a friction machine.
A good picture of a classic Wimshurst machine (1883) [1]. A Line drawing of the same picture [2].
A Wimshurst machine [6], similar to one that exists in the Museum of the UFRJ Engineering School, and that I restored. Another [14] similar machine. That machine was built by F. Ducretet and E. Roger, Paris, and originally should look as in this (front) and this (back) ray-tracing drawings. Here are comments about the restoration and more pictures. This is how it is now, compared with my 1974 Wimshurst. A large spark produced by the machine in a demonstration. Another picture, showing two of these machines connected as a generator and motor pair [24]. These pictures [23] show discharge images on photographic plates obtained with one of these machines. A positive discharge, and a negative discharge. A multiple machine from the same instrument builder, at the University of Porto, Portugal.
Here is another Wimshurst machine from the same museum that I restored (ray-tracing), a picture of it, and some comments about the restoration.
The largest Wimshurst machine ever built is presently at the Science and Industry Museum in Chicago, USA. It was built in England in 1885, with 7 foot glass disks 3/8 inch thick, and produced sparks with 22 inches. This picture is from Engineering, Vol. 39, 1885, scanned from [23] (also appears in [4][5][8][26]). More informations and pictures about this machine.
More Wimshurst machines, from [14]: A large simple machine, a double machine, a quadruple machine, and an octuple machine. And also two machines from [15], with ebonite disks and Leyden jars with two sections, and with glass disks. Two more large machines from [22]: A simple machine and a quadruple machine.
Wimshurst machines, from the collection belonging to Louie Scribner: A French machine (Bonetti), and a German machine (Leybold, 1901).
Wimshurst machines that have full construction details in [4][23]: A laboratory machine, a long spark machine, and a twelve plates machine. Designs by J. Wimshurst.
A "Voltana" Wimshurst machine [34](1921), used to run an electrostatic motor, and a bank of Geissler tubes. A larger machine, a smaller machine, another machine, more Geissler tubes being lighted, lighting a spinning Geissler tube, charging a Leyden jar, and charging a spring, that expands when charged [35].
Several machines, from [38], that illustrate the state of the art by 1900: A classic Wimshurst machine made by Bonetti, with Leyden jars that support the terminals, a complex double machine and a large multiple machine made by Ducretet, and a machine with large conductors, that act as capacitors. The classic Bonetti sectorless machine, a multiple Bonetti machine, and a double Bonetti machine.
Some multiple Wimshurst machines: A double machine, built by the instrument builder E. Balzarini, similar to the machines that appear in its 1907 catalog. Note the triple brushes at the neutralizers. A quadruple machine, built by Newton & C^o. Back view. Detail from one of the charge collector assemblies. Pictures from eBay auctions.
Wimshurst machines with an orthogonal drive system were built by companies as Central Scientific Co. A machine with the Leyden jars serving as terminal supports. A small unusual German machine. A larger machine, and a big machine . One of the pulleys below is driven by the crank. The other runs freely ¹.
A Wimshurst machine with cast iron structure, including the neutralizers ¹.
The Wimshurst machine can also be built with cylinders instead of disks [4][23]. A more practical structure was designed by Lemström [5][8] (1899), with the two cylinders turning around a fixed central axle, that also holds the internal charge collectors and neutralizers. The axle was separated in two sections by an insulating block at the center. The machine was kept warm, dry, and ventilated by a heating system and the sides of the cylinders, shaped as fans. See his patents.
The Wimshurst alternating machine (1891), that generates alternating voltage, synchronized with the rotation, with a polarity reversal at each 3/4 of rotation of the disk. The operation of this machine was considered difficult to explain [5]. A single disk with sectors at both sides, alternating, rotates between two pairs of collectors/inductors. Picture from Engineering.
A triple Bonetti machine [11]. This kind of multiple machine was used in early X-ray work.
By June-October 1999 I restored a similar quadruple Bonetti machine, built by Radiguet & Massiot (~1910), for my university's museum. Some pictures of the machine during initial tests and a report about the restoration are available.
Another double Bonetti machine [26], with similar features.
A cylindric Bonetti machine [11]. A compact design with high current output. This appears to be a large machine built by Bonetti by 1894, where the cylinders had 50 cm of diameter and height [p78]. Another cylindric machine [26].
A machine that appears to be a sectored Holtz machine [22], said to be quite powerful. There are fixed inductors on the back plate, that are charged by sectors at the back of the rotating disk through brushes that project through two holes in the fixed disk.
A motorized Voss machine with a fully sectored rotating disk [26].
P. V. Schaffers (1885) [4][5][p29][p32] described a machine that is essentially a Wimshurst machine with the charge collectors at different positions, with brushes at the charge collectors. The Schaffers machine works as a combination of the Wimshurst machine and the Holtz machine of the second kind, producing higher current (schematic). A (bad) picture of a Schaffers machine [26].
A Wehrsen machine [1][26][27][32][34][p77] (1907). Wommelsdorf's idea, it is a highly insulated sectored Holtz-style machine with sectors embedded in the rotating disk, contacted through small buttons, and inductors [34] also totally insulated behind celulloid plates. Some machines had corrugated sectors for greater surface area, what increases the output current [5], or mounted at different planes for higher insulation [26]. They had switches to allow the direct connection between the inductors and the output circuit, for startup as a "replenisher" machine [32]. Two better pictures from [17]: A large machine, and a small machine. A simple machine, and a large machine, from [22]. A machine with direct motor drive [26]. A similar machine [p77]. Wehrsen's "Mercedes" machine [34][26], with one rotating disk, and with two rotating disks, one at each side of the fixed disk. A similar machine, built in 1911, exists in the Cavendish Laboratories, England. left view, right view, back view. Detail from the switches, and from the central inductor plates. The machine is normally under an enclosure, and was disassembled for cleaning and investigation of why it doesn't work anymore in 1999. Photos sent by Tacye Phillipson.
The Wommelsdorf condenser machine [1][17][18][p84] (1902-1920) was the last of the classical disk machines. It was basically a sectored Voss-style machine with double induction plates, one pair at each side of the rotating disk, and with all the sectors and inductor plates enclosed in ebonite plates. Some models had a switch in the middle of the neutralizer bar. A simpler model [17]. In the last versions, the disk had a full set of embedded sectors, monted in alternate groups separated by one or two thin insulating disks. The sectors were touched only at the borders of the disk by brushes running in a V groove there. The ebonite disk was covered by a material (celluloid or bakelite) resistent to deterioration caused by ozone and other gases produced by electrical discharges. [5][8]. Schematic diagram [27]. Versions with multiple sections were also built, with some versions combining different charging systems for the inductors, with alternate sections charging the inductors from brushes at the edge of the disks and others from brushes touching the sides of the disks, as in a Holtz machine. Another version [22] with older design, totally enclosed and with fixed contacts for the neutralizer. Partial schematic diagram. A similar machine [p80], another, an open machine, and a very large machine [p81][46] along with a small machine. A condenser machine powered by a steam engine [46]. Wommelsdorf designed also machines with pairs of disks rotating in opposite directions [45][p83], similar to multiple Wimshurst machines, as this [46], a triplex Wimshurst machine with a central disk accessed through the edge.
Several of the papers and patents by Wommelsdorf are available here. See also my Wommelsdorf machine.
See some tables with the performances of several machines.
The Pidgeon machine [26] was a Wimshurst machine with fixed inductors positioned in a way that increases the induction effect. Fixed inductors with same polarity of the opposite disk were placed surrounding, insulated, each neutralizer brush. The sectors were embedded in the disks [5][26][p53][p54]. Pidgeon studied also machines based on "triplex Wimshurst" sections (double machines with a single central disk), with enclosed sectors, that produce more current.
Piggott made a set of experiments with radiotelegraphy and "antigravity" using a compact double Wimshurst machine enclosed in a pressurized box. Drawings from his patent (1911) showing the machine. Front view, side view, top view.
A "Dirod" generator. A modern electrostatic machine, designed by A. D. Moore [10]. It is a cylindric machine similar to the Belli machine, or Lord Kelvin's replenisher, with metal rods as carriers. The output is taken at the inductors.
Another machine described in A. D. Moore's book [10] is the "shake-sphere" machine. It is electrically equivalent to the symmetrical Toepler machine or the earlier Wilson's machine.
From the 1940's to the 1960's, Nöel J. Felici, in France, developed a series of high-power electrostatic generators [40], initially for applications in research. This site, by Lyonel Baum, contains many informations about his work.
See also the several machines that I have built, in the section "my machines", above. The linked pages have many informations about the machines.

Other high-voltage devices, not electrostatic

Changing of subject, an Induction coil [1], or Ruhmkorff coil (1851). A "fly-back" circuit with a mechanical interruptor, that eventually replaced the electrostatic machines as a practical source of high voltage. Its schematic diagram [1], ommiting the secondary of the transformer, that is wound with many turns of thin wire and well insulated. By 1867 Ruhmkorff (instrument maker in France) was making coils that could produce sparks with 40 cm. A spark image (16 cm) obtained with one of these machines, with the terminals applied to a photographic plate [12]. The positive terminal was in the left.
The Planté rheostatic machine. It produces high voltage by charging a bank of capacitors in parallel and discharging them in series. The connections are made by contacts in the rotating cylinder [24][47]. The input is obtained from a battery.
A resonance excitator [1], with adjustable inductances (L1, L2), Leyden jar capacitors (C1, C2), and a spark gap, used in old experiments about resonance.
Self-resonant coils [1] with different lengths, that emit corona discharges when driven at their resonant frequencies by the excitator above.
Connection between an induction coil (J), the excitator, and the resonant coils (Seibt experiment) [1]. By changing the excitator inductance, it is possible to put one or another of the self-resonant coils in resonance. This experiment is a variation of the "Tesla coil" circuit, using a direct connection instead of a transformer. See my implementation of a similar system, and follow the links to see my other experiments with multiple resonance networks.
Complete apparatus for the Seibt experiment [18]. A variation using a large open primary coil [17]. A curious experiment using a Tesla coil and a long resonator coil, where corona to two wires demonstrates standing waves.
A classic Tesla transformer [1]. A primary coil with a few turns of heavy wire (or tubing), and a well-insulated secondary with many turns. It produces a similar result if the primary is inserted in series with the spark gap in the excitator above. Several Tesla coils, from [18].
Apparatus for Tesla experiments, with an usual Tesla transformer, and with an oil-insulated transformer [18]. A large Leyden jar capacitor is connected in series with a spark gap and the primary of the transformer. The capacitor is charged by an induction coil, or a powerful influence machine. The other devices shown are for demonstrations of the effects of high-frequency high voltage. Similar apparatus with the usual air-insulated transformer [17], and an experiment to demonstrate the high impedance of a wire loop, using a Tesla coil with the primary coil inside the secondary [22].
An Oudin coil (1898) [33], a Tesla coil with the primary and secondary windings forming an autotransformer. Another Oudin coil [36]. See also my Tesla coil, that can also be operated as an Oudin coil.

Half Wimshurst, Toepler, Wommelsdorf, Holtz, and Unfolded Wimshurst machines

Some final observations, about pseudo-science, safety measures, and troubleshooting.
Calculating the output current of electrostatic machines.
References - Books, papers and patents about electrostatic machines, cited between "[]" in the text.
See also my links about these subjects.
Community in Orkut.
Links to e-mail addresses or collaborators had the "@" replaced by "|".

¹ Found in the catalogs here.

Last update: 8 September 2009. Recent changes and updates
© 1996-2009 Antonio Carlos M. de Queiroz
The contents of these pages cannot be reproduced without consent of the author.
Send comments and questions to .