Mills AC/DC Rotary Converter
Invented by H. K. Sandell
Mfd. by Mills Novelty Co., Chicago

us001471067-001b.jpg (25 kb)
110 VAC input ("Line") is applied to the brush holders, terminals 29, nearby the slip rings.
110 VDC output ("Load") is presented at terminals 28 nearby the segmented commutator.

Images of the original patents, obtained from the U.S. Patent and Trademark Office (USPTO) web site,

U.S. Patent 1381505  Patented June 14, 1921.
us001381505-001.tif ( 25 kb)
us001381505-002.tif ( 70 kb)
us001381505-003.tif ( 55 kb)
us001381505-004.tif (233 kb)
us001381505-005.tif (102 kb)

U.S. Patent 1471067  Patented October 16, 1923
us001471067-001.tif ( 40 kb)
us001471067-002.tif ( 65 kb)
us001471067-003.tif (114 kb)
us001471067-004.tif ( 60 kb)

The image files are 7.75  x 11.36 inches  (19.7 x 28.9 cm) at 300 dpi resolution, stored in Tagged Image File Format (TIFF) using compression method CCITT Group 4.  A very nice image viewer plug-in for Internet Explorer and Netscape Navigator is available free from AlternaTIFF  at

Analysis of operation
by Robbie Rhodes

According to the 1923 patent description, the machine is essentially a 4-pole self-excited DC generator, powered by an integrated 4-pole AC motor with an auxiliary split-phase starting winding and centrifugal starting switch, as in an induction motor.  It starts as an induction motor but at normal running speed it operates as a synchronous motor, an innovation claimed by the inventor.

The motor is "inside out" as compared to a typical motor, since the main motor winding (the motor armature) is on the rotor instead of the stator.  The AC starting current also flows through a stator (field) winding, which implements one of Sandell's patent claims.

Assuming the starting winding or switch isn't damaged, the rotor will spin the same direction each time it starts.  If the starting switch is bad or the start winding is open, then I suppose the rotor just sits there and hums and overheats, just like tired motors in orchestrions.

D'ya remember long ago when we bought a rebuilt generator for the old 1939 Ford or Chevy?  The paper tag said to "flash" the field with battery current before use, because most likely the residual magnetism had been degaussed during testing.  The flash of current was to give the steel stator (the field magnet) a bit of residual permanent magnetism of the correct polarity.  Then it would start generating with the correct polarity.

The generator field electromagnet of the Mills converter is the steel stator, which probably retains some magnetism after shutoff, but the winding carrying 60 Hz AC starting current also runs throughout the stator.  Thus the stator is thoroughly degaussed during starting.  When running at normal speed the stator magnetic field is not alternating, so that it can provide the DC magnetic field needed for both the generating and the synchronous motor functions.

An oscilloscope displaying the voltage at the output "Load" terminals would show 60 Hz AC, phase modulated by the rotor position, until suddenly regeneration takes effect at random polarity when the starting switch opens, and the DC voltage builds up quickly.  Sandell doesn't mention this polarity randomness in the patent.

Installation remarks:

For safety and reduced radio frequency radiation, the AC power wiring should have a 3rd "green wire" to connect the frame of the converter to external earth ground via a 3-contact plug.  1 kV .001 ufd ceramic disc capacitors may then be connected from each AC input terminal to the frame to further reduce radio frequency interference developed at the slip rings.  Similar capacitors may be connected from the DC output terminals to the frame.  But add the capacitors only if the frame is grounded via a "green wire".

The AC power line should be protected with a "Slo-Blo" fuse or appropriate circuit breaker.  The converter can be protected from damage due to a shorted load by installing a fuse at the DC output terminals.

Reducing the Converter Noise
by Hal Davis

I've resorted to several different tricks to reduce the noise from the vibration of the converters and finally wound up buying an assortment of  springs from a spring company in southwestern Michigan.  Among these I found some that were excellent when used with the converter sitting on a flat board which in turn sat upon four of these springs, which were about 3/4" diameter and 2-1/2" long uncompressed.  These springs sat in the small wells in the floor of the Violano cabinet.

The fun thing was to watch the converter doing a little dance to the music when it was running, due, of course, to the variations in the load to the output of the converter.  But it did reduce the transmission of the converter noise to the floor of the cabinet to the point where it was almost  inaudible with the cabinet doors shut.

Another scheme I used was making a larger board for the bottom and making hooks mounted at the corners, then hanging the converter from those hooks.  That one would also bounce around to the music but, due to the larger surface of the board holding the hooks, it transmitted more of the vibrations to the floor of the cabinet.

Converter sizes:

The converters for Mills Violano Virtuosos generally came in three different ratings:

1. The Single Mills used a 2.0 ampere converter.
2. The Double Mills used a 2.5 ampere converter.
3. The Double Mills with drum cabinet, which is what I used to have, used a 3.0 ampere converter.

A Single Mills with a drum cabinet would probably have used a 2.5 ampere converter, but as I've never seen one of these models I can't say for certain.  The converter for the Mills Magnetic Expression Piano that I had was, I believe, a 2.0 ampere model.

Of the fifty-plus Violanos that I've worked on over the years, including five Single Violanos and a Double Violano of my own, plus a Mills Magnetic Expression Piano, I don't recall a convertor ever turning at a random direction.  As far as I ever observed they always ran in the same direction every time they started.

rev. 16 January 2003, 22 January 2003