Melville Clark Air Motor Designs

Public domain US Patent Application.

Melville Clark Air Motor Patent US 1,230,995 filed on December 16, 1916, and Patent granted on June 26, 1917. It is unknown if this unusual air motor design ever made it into regular production. Nevertheless, in the above patent figure 1 (and referring to the air motor crankshaft at top) the ladder chain drive at left side powers the spool-box transmission, while the chain at the right side connects to a camshaft that runs the width of the air motor assembly. Each of the five positions on the camshaft is comprised of two cams, adjusted approximately 180 degrees apart, each operating against a lever-valve with a center fulcrum point. These lever-valves in turn (when a cam pushes down on one end) expose air ducts connected to primary valves situated inside a valve chamber located at the bottom of the air motor. The atmospheric vents for these primary valves are visible on the removable valve chamber cover (see the following image panel for valve chamber details).

Public domain US Patent Application.

Melville Clark air motor Patent US 1,230,995 filed on December 16, 1916, and Patent granted on June 26, 1917. In the figure 2 cross section (at right) the shape and layout of the lever-valve cams is clear, as is the layout of the primary valves within the vacuum chamber. For each motor pneumatic there are two large ducts, the lower one connects to a primary valve that admits a vacuum, while the upper duct connects to a chamber where a primary valve opens an atmospheric vent. The operation of the cam controlled lever-valves is the inverse of what might normally be expected, inasmuch as opening the duct closes the valve instead of opening it. When a cam presses down on one end of a lever-valve the opposite end is raised, thereby exposing and venting to atmosphere a duct connected to a primary valve. When air rushes into a primary valve pneumatic the valve port is forcefully closed. Figure 3 (at left) shows the obvious offset for each pair of primary valve pneumatics, so that they are in line with the air ducts under each pair of cam operated lever-valves. While this design probably worked, it is quite a manufacturing feat as compared to the industry standard and comparatively simple windmotor of the time, with simple external sliding valves.

Public domain US Patent Application.

Melville Clark Air Motor Patent US 1,265,747 filed on December 8, 1916, and Patent granted on May 14, 1918. This patent drawing most closely resembles the description of the air motor for the Airapollo announced in the Music Trade Review in June of 1917. Here are some excerpts from the announcement: "The motor is constructed with adjustable Pitman lever operated plunger, double seated spiral spring actuated, self-acting valves, which only move three-sixteenths of an inch when in operation and cannot be affected by dust, dirt, or climatic changes. A five-inch fly wheel has been attached to the crank shaft, which steadies the motor under all conditions and prevents hitching of the music sheet as it travels over the tracker range, and causes it to run smoothly and evenly."

Figure 1 is a perspective view from the forward side of an air motor, the same shown in connection with adjacent spool-box chain sprocket which it serves. Figure 2 is a fore-and-aft vertical section of the air motor in Figure 1, through one of the motor pneumatics and its valve chamber. Figure 3 is a section similar to Figure 2 of a modified form. Figure 4 is a horizontal section of the tempo valve chamber, taken above the valve therein. Figure 5 is an section at the line, 5—5, on Figure 4. Figure 6 is an axial section of a fly wheel and spring-connected crank arms by which the fly wheel is connected with the shaft on which it is mounted.

The patent dose not detail exactly how the valves are lifted from their corresponding seats, and states only that "there are two valves, 11 and 12, mounted on a common stem, and on that stem thrust apart by the reaction of the spring, 13, forcing each valve to its proper seat, controlling respectively the ports, 9 and 10.

Another construction oddity is the flywheel assembly. It fits loosely on the crankshaft; it is not pinned to it as might be expected. Instead it is straddled by two loosely fitting collars, each with a radial arm attached, the arms which come up against "abutments' in the rim of the flywheel. Each of these loose collars is more or less connected to the crankshaft by means of coil springs, each of which is wound in the opposite direction. The purpose of this complication is explained in the patent application thusly: "It has heretofore been regarded practicable to provide an air motor which is employed for operating an automatic player with a fly wheel upon the crank shaft, because while operating to give desirable steadiness to the action of the motor it tends to prevent quick response in respect to changes of speed to the speed-controlling valve or tempo valve, tending therefore to cause gradual acceleration and diminution of speed instead of the instantaneous or nearly instantaneous change which is sometime desirable. I have found in practicable, however, to employ a fly wheel and the crank shaft of this motor by properly relating the weight of the fly wheel to the power of the motor. And I have found that practically all the disadvantage which might result from the presence of a fly-wheel for the reason above suggested, may be obviated by the construction shown and which will now be described in which the fly wheel is connected to the crank shaft yieldingly as to relative rotation of the two."

Photograph courtesy of Jere DeBacker.

Notwithstanding the above patent drawings of 1916, above is an air motor as installed in an Artapollo player piano of the same year. Note that it uses the simple sliding valve design common throughout the automatic music industry. The above patent drawings illustrated two distinctly different designs for the Melville Clark air motor (commonly termed a windmotor), one with a large flywheel (to smooth out the air motor's operation) and both with a complex system of valves and interconnected linkages. But what was actually installed in the Airapollo and other Melville Clark pianos may have been quite different. For the Artapollo air motor pictured here (and installed in an electric/foot pumped expression piano built in 1916) about the only thing that distinguishes it from the common 3 crank/pneumatic variety of windmotor is that the Melville Clark design used a 5-lobe crankshaft powered by five motor pneumatics. Whether a piano operator could discern any difference in tempo accuracy between the 3 crank/pneumatic design versus the 5 crank/pneumatic design is anyone's guess. Nevertheless, the 3-crank/pneumatic configuration seemed to work satisfactorily for just about every other major manufacturer of automatic pianos, including the finest quality full reproducing piano systems.