A double Voss electrostatic influence machine

The Voss machine, invented in 1880, is a derivation of the Holtz and Toepler machines (1865), and so is also known as the Toepler-Holtz machine (or even Toepler-Voss machine, as Toepler also developed a similar machine by that time). This design was very popular by the end of that century, being extensively used for laboratory and medical applications, including use as power supply for early X-Ray machines. It consists of a fixed disk and a rotating disk, slightly smaller, that rotates in front of it. At the back side of the fixed disk are pasted two paper inductors with the shape of curved strips covering an angle of about 90 degrees. These inductors are connected to "appropriating brushes" that touch metallic buttons in the rotating disk, as they start to cross the region over the corresponding inductors. At the angle corresponding to the other end of the inductors, a neutralizer bar is positioned in front of the rotating disk. The bar has brushes that touch the buttons in the rotating disk, surrounded by combs with points, to act also on the rotating disk surface. The output is taken by combs positioned horizontally in front of the rotating disk, opposite to the inductors in the fixed disk (schematic). The design includes also metallic strips under the paper inductors, terminated in two disks, one in front of the corresponding appropriating brush, connected to it by another strip, and other close to the other end of the inductor (as here [18]). They turn the operation of the machine less sensitive to the conductivity of the paper, that may vary with the humidity level. The machine can be seen as a Holtz machine, with the inductors charged from the buttons in the rotating disk instead of from the back of the rotating disk, that is not used, as in a Toepler machine, and with a set of round metallic sectors that act as a starter for the machine, also as in a Toepler machine.

In June 1998 I completed a double Voss machine. The machine has two of the basic machines connected back-to-back, for greater output current and shielding of the inductors, with shared driving system and terminals. It has a wood support made of four bars with a cylindric upright support. The two fixed acrylic disks have 30 cm of diameter, and are supported by two acrylic pieces below and one above. They can be rotated to any position, and their separation is of 5 mm. Each has a set of two inductors, covering angles of 90 degrees and 5 cm wide, made of drawing paper, covering conventional metallic strips and disks made of aluminum foil. The two rotating acrylic disks have 27 cm of diameter and are mounted in a single wood boss, screwed one at each side of a ring, part of the boss between them, with three screws each. They turn at 2 mm of the corresponding fixed disks. The boss has ball bearings fixed in cavities at each extremity, that turn over a steel axis fixed to the upright support. Each rotating disk has six round sectors glued to it, with a button stamped in the middle, made of 0.4 mm aluminum sheet. The neutralizers are fixed to the central axis, so they can be rotated to any angle, and made with 9 mm aluminum tubes, 1/8" brass bars that interconnect the combs/brushes in the two basic machines, 1/16" brass wires for the comb points, and brass pieces and screws to adapt the tubes to the brass bars. The brushes were initially made of nickel-chrome wire, but soon replaced by conductive rubber, that doesn't brake easily with the repetitive contact with the buttons in the sectors (the only problem is some startup difficulty in humid days, comparing to the instantaneous startup with metal brushes). The brushes are mounted at the end of screws that cross the aluminum tubes, fixed to holes at their tips with wood pins and glue, to allow some distance adjustment. The comb points are simply glued to holes in the aluminum tubes. The charge collectors and appropriating brushes are of similar construction. All the extremities of the assembly are covered with 1 cm plastic beads, or larger wood balls. The terminals are supported by varnished 1.5 cm PVC tubes. Aluminum tubes connect the charge collectors to the terminals, passing through wood balls mounted over the insulating supports and fixed by screws from above. The spark gap assemblies are connected to these tubes by slotted brass plugs fixed to wood balls, so they can rotate with some resistance. The spark gap rods are aluminum tubes that cross that balls, and make contact with the plugs trough small springs inside the wood balls. The terminal handles are also 1.5 cm PVC tubes, adapted to the end of the tubes by nylon pieces, and the terminal balls are in solid aluminum, turned on the lathe. The machine is turned by a single pulley connected to a crank, mounted in a support at the back side. All the wood parts are varnished with several layers of polyurethane varnish.
My machine can be seen in these ray-tracing drawings, in front view, back view, and side view. Pictures from the machine, as initially built are here (front) and here (back). The final machine is seen here, with Leyden jars. A picture of it used to excite my Bonetti machine by placing one of the terminals in front of a neutralizer, is shown here (this single-point excitation works well only in dry air).
A similar, but maybe more beautiful, machine can be seen here [17] (with the back fixed disk at a strange position). The configuration is more clearly seen in this quadruple machine [17].

The machine self-excites easily, and in dry air produces easily 6 cm sparks between 2.2 cm terminal balls, going to 9-10.5 cm with the addition of a smaller (9 mm) ball to the positive terminal, and a pair of Leyden jars. In very dry air (as conditioned air), even with the normal balls it goes to 10 cm sparks, and produces them easily even with the disks moving slowly. The short-circuit current reaches 60 uA with the disks turning at about 700 rpm. The current is very high for a machine of this size, but the voltage is just regular. In the dark, it is possible to see significant leakage from the paper inductors to the bars that interconnect the two sides of the neutralizer. Due to this, the highest voltage is obtained with the neutralizer bar moved to a high angle, somewhat away from the area covered by the inductors. In this condition the machine takes some time to reach maximum excitation, while charges spread from the inductors to the surface of the fixed disks under the neutralizers. I added plastic tubes enclosing the connections between the two sides, with small improvement. A different design, without that interconnection is something to experiment. Other leakage points are the connections of the appropriating brushes to the inductors, between pairs of inductors, and to the central boss. The maximum possible spark length, given by twice the distance between the active area of the disks and the central boss, would be 11 cm. The machine shows significant resistance to polarity reversal. Sometimes it reverts polarity after some time of operation at high voltage, or if one pair of inductors is grounded for some time while the rotation speed is reduced, or if turned backwards for some time. The separation of inductors and output avoids the polarity reversal at each spark, that happens with my similar machines mounted with output and inductors directly connected. The machine is significantly less prone to reversals than the Holtz or Toepler machines. In dry air it is even rather difficult to obtain a reversal, and the machine performs as well as a Wimshurst machine, that does not revert polarity while in operation.

The construction of the machine could be simplified, as there is no need of two complete machines. One of the sides could have a bare disk, no appropriating brushes, and no brushes at the neutralizers, with the inductors connected to the inductors of the other side. The complete duplication, however, turns easier the self-excitation of the machine.

This picture shows an old double Voss machine that is on display in a museum in Lisbon, Portugal, that has interesting ideas. Note the assembly of the appropriating brushes, mounted in supports at the frontal hub and connected to the inductors by flexible wires. They have combs covering the outer side of the charge transport area. One of the output charge collectors is broken, and the output terminals are missing. The rotating disks were made with a kind of mica composite, and the inductor plates in glass. View of the cabinet. The machine is not quadruple. The other set visible is a reflection in the glass. Note that there are no buttons or brushes in the back disk, and that there are brushes in the frontal charge collectors (why?). An old advertisement of a very similar machine. These large machines were common as power supplies for early X-Ray apparatus.

In July 2002, this apparently original Voss machine, single, was being autioned at eBay. Another view, that shows details of the construction, and part of a booklet about it. The structure is as in this drawing.

P. Atkinson [44] developed two versions of the Voss machine. A restored single machine with his design [US275347] patented in 1883 was auctioned in eBay in 2000. More views of it. The other was a double machine as mine, described in great detail in his 1885 patent [US331754].

Voss machines were commonly used in the early 1900's for demonstrations of lightning rods. They were characteristically mounted in a transportable box, having attachments representing a "house" and a "cloud", sometimes with also a "fence" and an "animal". The machine in the picture appears to be Baysdorfer's machine [US827497] (1906). A single Voss machine is shown in [US700536] (1902). A double Voss machine appears in [US837178] (1907). Some pictures of a surviving machine of this type, slightly damaged, were sent by Pete Windahl: Front view, back view, with the box removed, and label. Another machine is shown in [US1043040] (1912).

By March 2005 a Voss machine was being auctioned in eBay. Front view, back view, and detail. Note that the fixed disk is incorrectly mounted. By the structure, this machine was built in the 1900's.

Created: June 1998.
Last update: 10 March 2005
Created and maintained by Antonio Carlos M. de Queiroz
Return to Electrostatic Machines