Mechanical Music Registry Project

National Automatic Music Company Pianos

Manufactured by the National Piano Manufacturing Company

Music Changed Weekly...

A rare National Piano Mfg. Co, token that could be dropped into a nickel coin slot and was "Good For 1 Piece of Music."

The National Automatic Music Company coin pianos are well known for their characteristic boxy appearance, the distinctive “Music Changed Weekly” sign centered on the inside of the upper clear glass window, for the eight-slot coin mechanism, and for the intricate looking mechanisms scattered over the “keybed” area that would normally be home to a piano keyboard. Intriguing as these captivating upright looking pianos are, it is often the case that an owner of a National automatic piano knows little, if anything, about the company behind these unconventional but musically appealing instruments.

The National Automatic Music Company was founded in Chicago, Illinois, in May of 1909. The founding officers were:

• Carroll F. Sweet, Grand Rapids, Michigan, president;
• Clarence U. Clark, Grand Rapids, Michigan, vice president and treasurer;
• Walter Ioor, South Haven, Michigan, secretary;
• I.D. Taylor, Chicago, Illinois, assistant secretary and assistant treasurer.

 One month later, in June of 1909, the National Piano Manufacturing Company was founded. Its founding officers were:

• Clarence U. Clark, of Grand Rapids, Michigan, president and treasurer;
• Walter Ioor, of South Haven, Michigan, vice president;
• I.D. Taylor, of Chicago, Illinois, secretary;
• Carroll F. Sweet and Joseph Renihan, of Grand Rapids, Michigan;
• Robert Mills, of Springfield, Ohio;
• T.W. Sheriffs and F.B. Geisler, of Milwaukee, Wisconsin.

The two companies were to be intimately allied, with the National Automatic Music Company contracting for all output and being the sole distributor for the National Piano Manufacturing Company, the latter company being the manufacturer of what is commonly known today as simply the National Automatic Piano. In other words, the National Piano Manufacturing Company did the physical work of making the National coin pianos, along with any ancillary equipment, music rolls, and spare parts, while the auxiliary National Automatic Music Company owned and serviced all the National pianos, kept the music repertoire up to date on a weekly basis, and collected the nickels earned.

Here, in a few brief paragraphs, is the beginning of the National story, but in the actual words of the company itself, and as published in 1930:

"In the year 1909, with a few hundred dollars, the organization that later was to grow into the Automatic Musical Instrument Company was formed. The assets of the company consisted of one automatic electric piano and two applications for U.S. patents (then not yet allowed) for a selecting device by which any desired music roll contained in a magazine could be played.

These applications constituted a wide departure from the automatic pianos then in existence and installed largely in saloons and lower class dance halls. The advent of the company immediately gave to the automatic piano field a different quality, and the persons who started the original company were somewhat surprised by the progress that was made in securing better and higher types of location.

An organization was perfected early in 1910 to take care of the future growth of the company, whereby its functions were divided into two parts, namely the manufacturing and the operating phases of the business."

For reference, copies (in PDF format) of the principle patents assigned to the National Piano Manufacturing Company, and that relate to coin pianos, are available for viewing by clicking on the detail link below.

In 1912 the National companies were on the move, both literally and figuratively. On May 16, 1912, there appeared the following article in The Presto:

And so, the National companies moved to Grand Rapids, Michigan. According to the Labor Statistics from the 1912 Michigan Department of labor, the National Piano Manufacturing Company had a total of 30 employees, there being 27 male employees and 3 female employees of record.

Because the National companies did not sell the coin pianos to the public, but kept ownership to themselves, they never promoted the National automatic coin pianos through myriad colorful catalogues or advertising campaigns. Consequently, publicly disseminated printed material relating to their coin pianos is rare, to the point that the existence of the National companies, except for the surviving pianos, might seem doubtful. But one notable exception to this dearth of company produced literature is The Nimble Nickel, a simple foldout pamphlet extolling the value of the nickel, and, of course, the National Automatic Piano, owned and controlled by the National Automatic Music Company of Grand Rapids. The pamphlet goes on to explain how orchestrions and their ilk soon fall out of use once the novelty wears off, and only the piano, kept properly tuned and regulated, can capture the of soul of the music and have lasting value. There may be some merit to this idea, because National only had one model coin piano from its very beginning up to the end of the coin piano era, and the company was both profitable and very successful with some 4,000 coin pianos out on location by 1925. It takes a well-oiled organization to keep this many coin pianos serviced and fresh with new music once a week, so as to keep a flood of nickels constantly flowing into the company coffers.

The National Automatic Music Company/AMI’s profit sharing policy, as explained by Nick Fink, the proprietor of the Riversite Hotel, might be unique amongst coin piano operators, and perhaps helps to explain National/AMI’s phenomenal success. Nick Fink was the proprietor of the Riversite Hotel, which over the years operated as a bar, later an an ice cream and soda bar, a hotel, barber shop, and a post office. The hotel was located near Grand Rapids, Michigan, across the river in the Comstock Park area (previously known as Mill Creek). At some point during the 1920s Nick Fink’s tavern hosted a National coin piano, which many years later was to catch the attention of John Perschbacher, a Grand Rapids resident and an avid mechanical music collector, and who also provided the majority of the National/AMI material presented on this page. Perschbacher met with and talked extensively with Nick Fink about National and other locally oriented topics. But regarding the profit sharing policy, as the story goes, when a National field agent located a potentially lucrative site the agent would contact whomever was in charge and offer a proposition for the placement of a National Automatic Piano. Whether, or not, the same proposition was offered across the board, or just to certain highly successfully and busy locations is unknown, but for the Riversite Hotel the proposal set a base rent for the piano, with a 50/50 split of nickel profits over the base rent. This meant that National/AMI always received a base amount no matter the patronage given the piano. Moreover, the proprietor had no piano or music rolls to buy and nothing to maintain, but had plenty of incentive to induce patrons to enjoy automatic piano music—a kind of marriage that seems to have brought forth a flood of profitable nickels.

The coin piano being manufactured had the overall shape of a standard upright piano, but was deeper and with no keyboard. Where a keyboard would have normally been located was instead a sprawling and intricate system of interrelated shafts, chain sprockets, gears, cams, and other complicated looking contrivances, all of which served to control an elaborate 8-roll revolver magazine (roll changer), which was unique because each music roll was fully independent of all others with its own spools and rewind system. The magazine held eight single-tune length music rolls and played them as selected by a patron who could drop a nickel into one of eight different coin slots, the vertical slots arranged in a horizontal row located immediately above a numbered music program.

The tracker bar, in all but possibly the earliest models, was mounted on a movable framework suspended above the revolver magazine, and is lowered onto the roll to be played once the magazine had been rotated into the correct playing position. Each of the eight independent roll stations had its own feed and take-up spool, and rewind system. As a roll was played, a feed spool follower worked its way along the length of a threaded shaft, while simultaneously either a spring barrel or coiled torsion type spring was gradually wound ever tighter. Then when the music roll was finished playing the tracker bar and the take-up spool drive pinion were quickly lifted up and away from the revolver magazine. Now the revolver magazine was free to rotate in preparation for playing whatever the next music selection might be. And once the tracker bar and the take-up spool drive pinion were raised the recently played music roll was free to rewind, and do so independently of all other mechanisms, it being powered by the spring motor rewind system that had been wound up while the music roll was being played.

National pianos were very widely distributed, with the West Coast being a particularly lucrative region. Some of the piano routes were large while others were relatively small, but nevertheless all were owned and operated by the National Automatic Music Company. To keep track of the profitability of any particular piano route the company maintained local bank accounts in various cities, so that each route manager could make a direct deposit of the money collected. If there were expenses incurred by the route operator they were paid out of the same local account. A certain percentage of the receipts were used to keep the route pianos in good working order. This methodology allowed the head office in Grand Rapids, Michigan, to easily keep track of the viability of each route, and where best to focus their marketing expertise.

The business prospered and by the summer of 1918 the company announced that it would be moving to a new factory, one of sufficient size to accommodate its expanded coin piano manufacturing and distribution business. Then in 1922 there was another move, also in Grand Rapids, Michigan, when both of the associated National companies moved to 1500 Union Avenue, a factory building earlier occupied by businesses owned by Walter Ioor.

In 1924, the Multi-Control, a derivative revolver magazine product for home use was marketed. It consisted of a compact mahogany furniture cabinet, or mahogany console, that contained a roll changing mechanism holding eight Welte (Licensee) reproducing piano rolls and that could be located at any distance from a piano, it being electrically connected to the piano. Because of the compact nature of the revolver mechanism, it left the major portion of the console for the storage of surplus music rolls. The desired music roll could be selected at will by means of eight buttons in the console, or by a remote push-button control box, which had a series of eight labeled buttons. The Multi-Control was marketed by Bush & Lane. Another much larger and more elaborate version of the Multi-Control system was made specifically for pipe organs, and held ten music rolls. (A version was reported in original press releases that held twelve rolls, but whether, or not, such a version was actually marketed has yet to be confirmed.) The merits of the Multi-Control, it was written, allowed any pipe organ to be transformed into reproducing instruments, operated from a push button located in any or several rooms, thereby providing not the mediocre in music, but the perfect reproductions of the exact expressions of the world’s greatest pipe-organists. Currently unknown are the specifications for the music rolls used with the pipe organ version of the Multi-Control.

By 1925 the company reportedly had some 4,000 pianos on location, spread over 32 different states and the District of Columbia. In 1930 the number of pianos on location had risen to 4,200 pianos. This geographically widespread operation involved about 300 people in the field as route operators and supervisors. To keep the popularity of the pianos alive, the company paid close attention to producing exceedingly well-arranged music rolls featuring the latest popular tunes in addition to old standards and favorites. From roughly 1922 to 1930 it is calculated that about 3,000 new roll titles were issued, which provided a superb library of music.

The Automatic Musical Company (AMI)

Prior to the reorganization in late 1925, the National operation had been divided into two separate corporations, so as to facilitate administration, service and repairs through an auxiliary company, while the principle company did the manufacturing. On November 9, 1925, Walter Ioor, then in charge of the whole operation, merged both National companies to form the Automatic Musical Instrument Company (AMI), also based in Grand Rapids, and using the already existing factory. Sometime after the close of 1925, the company issued a 16” x 20” broadsheet titled: Nickels Piled Up into Millions of Dollars Under the Management of this Organization. This revealing document was probably directed toward stockholders and various people involved in company affairs. Its purpose seems to be the selling of the Automatic Musical Company to investors, with little or nothing to do with selling coin operated automatic pianos. It is a remarkable presentation, and it gives some interesting details about how much money the company was taking in from nickels alone, as well as reveal some working details about how the various parts of the organization functioned. The document begins with a chart showing the nickel derived income per year, starting with the year 1909 and going up through 1925.

Nickels Piled Up into Millions of Dollars Under the Management of this Organization
There are two essentials to making profit from coin-operated musical instruments. The first is a manufacturing problem. The second is a problem of management and service for the operating instruments. By the mastery of both these essentials on an extensive scale, the Automatic Musical Instrument Company has gained a distinctive place for itself in American Industry, as a brief reference to the history of the Company will show. The business of Automatic Musical Instrument Company (formerly National Piano Mfg. Co.) was founded in 1909. The present management of the Company took charge of the business in 1911, when the real growth of the business began. From the outset the Company operated in conjunction with an auxiliary company. This auxiliary company purchased instruments from the fundamental company and contracted with the fundamental company to handle all financing, installation, servicing, collections and business management in consideration of a percentage of the receipts. Thus it has been the responsibility of Automatic Musical Instrument Company (National Piano Mfg. Co.) to produce a coin-operated musical instrument that would stand up and meet all of the exacting requirements of its field; to handle all financing for the auxiliary company, that it could pay for the instruments as produced; to find locations for the musical instruments; to service these instruments and keep them in tune and repair; to collect all coins from the instruments; and after deducting their percentage for this service to pay over the balance to the auxiliary company as surplus profits, free and clear of all expense. It will be noted that all problems relative to efficiency of management, manufacturing and service in connection with the two-company arrangement have rested entirely on Automatic Musical Instrument Company (National Piano Mfg. Co.) and that all credit for knowledge, skill and organizational ability are therefore due to the organization of Automatic Musical Instrument Company. The chart adjoining tells the story of these operations, showing the growth of the business under this method of operation. The figures show the receipts in nickels from coin-operated musical instruments manufactured, managed and serviced by this Company for each year over a period of seventeen years. Beginning with a few instruments, there are now over 4,000 of these instruments being managed and serviced by Automatic Musical Instrument Company. So Successfully has this business been handled in every particular by the organization of Automatic Musical Instrument Company that (in addition to its own profits from the sale and servicing of the instruments) it has turned over sufficient surplus profits to the auxiliary company to buy ample new instruments to liberally cover depreciation and to pay regular dividends to all stockholders at the rate of 12% per annum, with numerous extra dividends, aggregating a grand total of $2,376,809.25 in dividends. Attention is particularly called to the notable consistency and evenness of the growth, receipts and profits of the business. As administered by Automatic Musical Instrument Company, this business runs with a regularity that is astonishing—rarely found in any business—and is entitled to the distinction of being one of the most seasoned and conservative businesses in the United States. The results shown in the accompanying chart were all obtained from a single type of musical instrument, a 5-cent coin-operated Automatic Piano, the designs and patterns on which are owned by Automatic Musical Instrument Company. The facts are here presented, because of their bearing upon the merit and future of the new instruments announced on the following pages. “By their fruits ye shall know them.” From these facts the new instruments may be judged with the assumption that practical knowledge of the business is built into them. NOTE: For the management and servicing of the Automatic Pianos, Automatic Musical Instrument Company received 50% of the gross receipts from the Pianos. Provision has been made whereby the stockholders of the auxiliary company may exchange their stock for stock of Automatic Musical Instrument Company. (Source material courtesy of John Perschbacher.)

As part of the Multi-Control line of products, first introduced in 1924, the company developed a new coin piano that was essentially a competitor to the regular National 5-cent automatic pianos that had essentially built and sustained the company since its inception. “The Aristocrat,” as it was called, was a coin operated reproducing piano with a 10-cent coin slot, which utilized the Multi-Control system to select a Welte-Mignon (licensee) roll from a vertically oriented chain type magazine that held 10 music rolls. The roll magazine was hidden in a narrow case extension on the right-hand side of the piano case, and was easily accessible through a side access door. According to the company, the Aristocrat was “designed and perfected to meet a demand which the Automatic Piano cannot satisfy.” Because these 5-cent instruments were “restricted to jazz and the so-called popular music,” they were deemed unsuitable for use in higher class places. Other manufacturers, such as Seeburg and Wurlitzer, had already tried their hand at marketing a coin-operated “reproducing” piano, but with limited commercial success. Apparently AMI’s Aristocrat 10-cent reproducing piano met with a similar lack of widespread interest, because, according to John Perschbacher, only about two dozen Aristocrat pianos were made.

By this time it was becoming readily apparent to management that with the advent of the rapidly evolving electronic age the older electric piano technology would probably sooner rather than later be obsolete, it giving way to the new automatic, electrically-amplified disc phonograph. Although with a new corporate structure the transition to the automatic phonograph had officially begun, the National name continued to be used for a brief time in printed notices. That soon changed, and an Automatic Musical Instrument Company decal was prominently displayed on the front of each new or refurbished coin piano. Then the National Automatic Selective Phonograph was introduced in 1927, which was a novel innovation at the time. The new automatic phonograph stored 10 records, and due to the ingenuity of the mechanism either record side could be selected and automatically played. The music was electronically amplified.

The introduction of the automatic phonograph was a fortuitous turning point for the company, and soon thereafter the manufacturing of new National coin pianos ceased. However, many serviceable National pianos were sent back to the factory in Grand Rapids, where the piano actions were refurbished and the mechanisms cleaned and repaired as necessary. Some of these refurbished pianos also had a dog-racing device (officially termed an “Automatic Whippet Race”) retrofitted, or even fewer yet had a “Dancing Figure” (a patented jointed figure that danced to and fro) retrofitted, either of which were visible through the upper clear glass front. But if the stack was beyond reasonable repair, it has been rumored, it was replaced with a new single-tier Pratt-Reed Unitype pneumatic stack (also used in a few 1928 Seeburg Greyhound models), but no Unitype stack has yet to be factually observed in any National coin piano. Many of the refurbished instruments were also given a coat of light green exterior paint that was highlighted with pinstriping, whereupon the renovated pianos were then shipped out for further use on location.

The demand for National automatic pianos, as well as for all other types of coin-operated pianos, diminished in the late 1920s. This is a time when the radio and amplified music on phonographs was in its ascendancy, and due to their popularity were increasingly made available in public venues. By 1930 most of the company's coin-operated pianos had been replaced by automatic phonographs, with about 8,500 then in use. And so, the birth of a new industry, one founded on electronics and that could supply a wide variety of music more readily and with greater orchestral fidelity (and without the technical limitations imposed by the cumbersome logistics of arranging and cutting music rolls) took the public interest by storm. The National pianos in places distant from Grand Rapids were simply abandoned, and sold for storage charges, or junked. Some instruments in and near Michigan were kept in operation into the early 1930s.

In mid to late 1930 the company issued an illustrated and beautifully instructive booklet, which was titled, History, Organization and Personnel of the Automatic Musical Instrument Company. It was addressed to stockholders and interested members of the organization, who may have from time to time had a curiosity as to the business, its origin, operation, personnel, history and prospects. It covers the history and progress of the company from its inception in 1909 up through 1930. It was a forward-looking booklet, with grand prospects for the future; there were established and profitable operations under license in the Dominion of Canada, and the company’s instruments had been licensed for use in several European countries. The booklet concludes with the financials of the company, and a final statement by Walter Ioor, President, expressing confidence in the future, and in “the tremendous parade of progress that is upon us. Manned, by a loyal and efficient staff, which is our proudest boast, and supported by our friendly and helpful body of stock-holders, no power can retard our progress.”

But the future was not to be as glorious as the booklet published in 1930 had happily projected. The Automatic Musical Instrument Company was forced into receivership in 1931 by a Michigan bank. About this time Walter Ioor left the company and moved to Florida. Cornelius H. Knoll, who had been with the National companies since 1920, and was now secretary and treasurer, tended to finances. One of the pressing issues Knoll had to face was the growing cost of storing hundreds of idle coin-pianos that were being warehoused in various cities around the U.S. The answer as to how he dealt with this vexing problem came to light in April of 1980, when the late Willard E. Burkhardt, a mechanical music enthusiast, noticed a telephone listing for a “C. H. Knoll” in the Grand Rapids telephone directory. He called Mr. Knoll and interviewed him by telephone. Here is the rephrased gist of the interview:

In 1931 Cornelius H. Knoll took on the daunting task of disposing of all the useless coin pianos for which there was now no market due to the increasing popularity of the radio, and, perhaps more directly, because of the coin operated phonograph. This was a device that had quickly taken over the coin-operated market and was something that the company now made and placed on location as it had done with coin pianos a few years earlier. Mr. Knoll traveled to about 70 cities in 100 days disposing of pianos in storage or that had otherwise been removed from use, and practically giving away most of them. Many pianos were stored in warehouses and back storage charges were due. Sometimes Knoll was able to sell the pianos to the warehouseman for storage charges and end up with $10 for a piano that originally cost $1,000 a few years earlier. The warehouseman would then salvage the pianos by removing the electric motors and then burning the pianos to salvage the remaining scrap iron. In New York City, 400 pianos in storage were sold, in Philadelphia he sold 100, plus many more in other cities, such as Los Angeles. Some of the pianos were sold to location owners.

Most delinquent accounts were cleared, and Mr. Knoll was able to successfully guide AMI through the Depression years, as well as manage the expansion of the company’s popular line of Selective Automatic Phonographs introduced in 1927. The happy result of Cornelius Knoll’s steerage propelled AMI on to great success in later years as a maker of excellent coin-operated phonographs and other devices. By 1936 the formal and elegantly crafted wooden cabinets of the early automatic phonographs had given way to a modernistic art-deco design with the introduction of the “Top Flight” model. In 1946 the company name was officially changed to AMI Inc., and during that same year the “Model A” juke-box was introduced, which began the trend toward the more gaudy, futuristic molded plastic designs.

Meanwhile, the Automatic Canteen Company, a nationwide vending operation undergoing rapid expansion, acquired Rowe Manufacturing Company, Inc., in the mid 1950s. Rowe was a Chicago based company founded in 1926 by William Rowe, a man usually credited as the inventor of the first cigarette vending machine. Then, in early 1959, the Automatic Canteen Co. of America bought AMI, Inc., merging it and the Rowe operation into a manufacturing subsidiary named Rowe AC Services. But Automatic Canteen eventually divested the manufacturing division, whereupon Rowe continued to build juke-boxes under the Rowe/AMI name, along with a line of vending machines. Rowe/AMI was well known as a leader in the development of high quality juke-boxes, and, during the late 1980s and 1990s Compact Disc era, the company commanded an estimated 65% to 70% of the domestic juke-box market share, and 55% to 60% worldwide.

In 2003 Rowe sold its vending business and redirected its engineering and marketing resources to pursue opportunities in the emerging digital juke-box market. As part of that initiative, Rowe launched its AMI Entertainment Inc., subsidiary, which provided and managed digital music content, software, and networking technology for Internet-access enabled juke-boxes. And so, the tiny organization that inauspiciously assembled for the purpose of manufacturing a selective coin piano way back in 1909 survives to this day (after various corporate transitions) as AMI Entertainment, Inc.

The End of an Era

On Tuesday, July 11, 2006, Rowe International staged a celebration at the old National/Automatic Musical Instrument Company (AMI) factory complex in honor of the 1 millionth juke-box manufactured by the company. Long time collector, John Perschbacher, had become friends with many AMI employees during his visits to the factory, and was mailed an invitation to the upcoming affair. He was, and remains to this day, an enthusiastic researcher of all things related to National coin pianos, and first visited the factory in 1972, with several other visits in following years. This celebratory event would be, however, his last visit to the old buildings at 1500 Union Avenue S.E., historic structures that had once been home to the National Automatic Music Company and the National Piano Manufacturing Company. The Grand Rapids Press carried the following article on July 11, 2006, about the grand event and the future of Rowe/AMI:

Then, on February 27, 2009, the Grand Rapids Press carried this article:

Then by the end of the year the historic, rambling office and factory building at 1500 Union Avenue, S.E. was razed, and loads of splintered wood and broken bricks were carted off and buried as unwanted rubble. For John Perschbacher there could be no more visits to the historic factory. Now all that remained for him were some wonderful memories, and a lone souvenir brick salvaged from the factory during demolition—a simple reminder of the end of a golden era.

Technical Details

Music Rolls

National Tracker Bar Scale
1. Accent (Sforzando). 2. Hammer rail down. 3. Hammer rail up. 4 to 68. 65-notes in order from A to C#. 69. Sustaining pedal. 70. Low vacuum (soft). 71. Normal vacuum (loud).
Note 1: This is the tracker scale as published by Art Reblitz. Note 2: Accent or Sforzando expression (tracker bar hole #1) is not used when a balance-beam expression control has been installed. Note 3: The highest piano note (tracker bar hole #68; C#) is often left disconnected. This is apparently because this hole in certain late rolls was used to operate the "dancing figure" retrofit, rather than the piano note.

The 65-note (A to C#) music rolls made for the National automatic piano are always. more or less. the length of a single tune music roll. However, medley arrangements incorporating more than one tune have been noted, and rolls with two or three distinctly different selections have been observed. But some other roll characteristics remain consistent throughout, with the rolls being invariably 12" wide, and cut with a 6 per inch hole spacing. For the most part the perforations are round, but an early perforator used by the company reportedly cut square holes. It is estimated that about 3,000 musical arrangements were churned out between 1922 and 1930. This would average 7 new music rolls per week, which would almost certainly allow the pianos to live up to the company's motto boldly lettered across the clear glass front, which proclaimed: “Music Changed Weekly.” The down side of this up-to-date music policy was that the music rolls were systematically destined for elimination once they had been replaced. This was perhaps a good business decision, but a tragedy from the point of view of today’s mechanical music enthusiasts, because relatively few of the thousands of arrangements have survived, and in some cases only a single copy exists, while most are gone forever.

A great variety of popular music was cut, along with ethnic music and a smattering of classical music. Factory records have been lost as to who supplied the blues arrangements, which are in a style not recognized as being among the great jazz artists of the day, such as “Fats” Waller, J.P. Johnson, Eubie Blake, Lem Fowler, or Jelly Roll Morton. The arranging style appears to be edited from line marked recordings. It is a syncopated and “hot” kind of low down South Chicago style of playing, which is consistently well done and equals the best blues arrangements issued by the Clark Orchestra Roll Company or the Capitol Roll & Record Company. The thing that makes National rolls somewhat unique, other than the fine arrangements, is the sophisticated expression coding located along the margins of the paper, although some might argue that the piano’s expressive capabilities were often less than fully exploited. They used the soft and loud pedal, as did other coin pianos, but in addition National pianos employed a wide range of dynamics and were capable of accent effects by going instantly from soft to loud and back again through manipulation of the stack vacuum level.

From the single tune master rolls a few arrangements were punched out on the 9 per inch scale, with standard 11¼” wide paper and boxed under the brand name of “8T8 Har-Mo-Nie" for sales directed to the home player piano market. It is likely that the failure of this venture was due to the player piano rolls not being competitively priced compared to the other more established roll manufacturers who already had extensive distribution networks in place. There is no doubt that some Grand Rapids, Michigan, merchants did stock the “8T8 Har-Mo-Nie” brand, but whatever the case, these 88-note rolls are rare today, with what is probably the largest collection belonging to the late Willard E. Burkhardt, a long-time collector in Grand Rapids, Michigan. The “8T8 Har-Mo-Nie” brand piano rolls were only produced for a short period of time, from 1920 to 1922. An employee who operated the roll perforating department, during this same era, recalled making some ten-tune coin piano rolls using the National master rolls. Although he did not recall who the customer was, he did remember sending out an order about once a month. Except for these few exceptions, all roll arrangements were only issued for consumption on the National automatic piano routes, and by 1925 there were reportedly 4,000 pianos out on location, and then by 1930 there were reportedly another 200, for a total of 4,200 pianos pumping out toe-tapping tunes across the U.S.

In 1925 the National Automatic Music Company and the National Piano Manufacturing Company merged and became the Automatic Musical Instrument Company (AMI). In 1931 the Michigan National Bank forced AMI into receivership. Under the bank’s direction they sought to eliminate the accumulating storage costs associated with pianos taken off routes and put in storage. The potential to revitalize these pianos by adding Whippet race dioramas or other amusement lures was no longer viable. The electric coin operated phonograph, or “Juke-box,” had become the rage. The Automatic Musical Instrument Company (AMI) reported approximately 8,500 coin operated Selective Phonographs on location in 1930, with little or no demand remaining for the older National coin pianos.

The music roll production continued for another three years. Some National pianos in storage or still on location were sold to a willing proprietor for a near give-away price, but this move to private ownership of National pianos greatly reduced the once high demand for new music. After 1933, apparently, the Clark Orchestra Roll Company continued to supply new National music up until 1937. These Clark rolls were distinguished by a five-digit roll number, some of which were proceeded by the letter “X,” and the titles were written in pencil instead of stamped with a mimeograph type labeler as National had done. Both the National and Clark origin music roll were made on perforators using a 2-to-one master roll. The Clark perforators, however, had a slightly slower paper advance, which slightly increased the tempo.

Alternative Tracker Scale

This alternative National tubing and tracker scale diagram was created by Dana Johnson, circa 1968 - 1970 (during the time he lived in Denver, Colorado) and is reportedly representative of a National automatic piano for which he personally carried out certain repair work and then regularly serviced. The drawing is made available here because it may help to better illustrate how the various expression control components, working in conjunction with National built stacks, are tubed together and interrelated. And the story behind the drawing is interesting in and of itself and reveals a time when mechanical music collecting was still in its infancy. Please note, however, that certain tracker bar connections for expression controls shown in the diagram differ from the tracker bar scale published by Art Reblitz (and shown above on this page), which have recently been confirmed by a reliable source as accurate.

The Roll Cutting Department

It was on January 15, 1980, when Michael Montgomery (an accomplished ragtime historian and pianist) telephoned James S. “Rags” Allen and asked how music rolls at the National Piano Manufacturing Company (later Automatic Musical Instrument Company) were made. Mr. Allen said that he helped make the rolls and that he penciled them out—or drafted them on special paper. He used two systems:

• The Meahl System used heavy paper that was preprinted with linear lines coded with the keyboard positions, laid out like the keyboard of a piano. The music arranger made their pencil marks on this paper. At each edge were a row of sprocket holes for precise advancement of the paper, which made this method very accurate.
• The Salier [sp?] System used thinner paper and the arrangers had to draw their own guidelines.

The arrangers started with sheet music or manuscript scores. Then they worked out their arrangement on a piano keyboard, and then lastly came the actual drafting work, creating a penciled out master template. When they finished marking up a master, a young woman did all the cutting out of the perforations. She did this by inserting the master into a machine that could make the cutouts by pressing a button. She never used a hand-held knife to do the work.

Paul Estabrook’s arranging approach was to give his music a hand played feel and sound. He first used a line-marker and then laid the music out in steps to agree with the stepping system of the perforator. Paul then moved to a recording device to mark up the master roll. James “Rags” Allen never used the recording machinery because, he said, it was not accurate enough. It apparently took him more time to correct hand-played errors than to simply arrange a tune by drafting it out. No one trained Allen on how to be an arranger—someone simply showed him the setup and he figured it out for himself. He arranged one or two rolls a week. Paul Estabrook made as many or more. A fellow by the name of Walter A. Goble also arranged rolls, and was working as a roll arranger when James Allen came to work at National. Goble was “an old band man” and made march rolls. Another man (name unknown) also made rolls and possibly made blues rolls. Neither Paul Estabrook nor James Allen made blues rolls.

James S. Allen was born in Vancouver, British Columbia, Canada. He became a working pianist and eventually ended up in Ohio playing shows. There was a saloon in Lima, Ohio, that he frequented. In the saloon there was a National Piano. One day an “inspector” for National heard him playing and invited him to move to Grand Rapids, Michigan, to arrange music rolls for the company. When he visited Grand Rapids, he decided against the move, but was later convinced he should take the job, and did so circa 1920-1921. He was probably about 23 years of age at the time. At National, James Allen earned the nickname “2 Finger(ed) Jimmy, but was also known as “Rags Allen,” an endearment arising from a "Rags" tattoo on his arm; Paul Estabrook, the principle arranger, was affectionately known as “16 Finger(ed) Paul” or sometimes “20 Finger(ed) Paul.”

Ralph Durnbaugh was another “inspector” for National, who was from Lima, Ohio, but he was not the inspector who encouraged James Allen to move to Grand Rapids. However, by the time James was hired by National, Ralph Durnbaugh had become the head of the music roll department. The perforators made fifteen copies per cutting. However, an earlier perforator cut rectangular holes instead of the round holes of the newer perforators. As such, any rolls with rectangularly cut holes will be much earlier than rolls with the more traditional circular cut holes.

When Gale E. Snyder (born in 1901) was interviewed later in 1980, by John H. Perschbacher, he remembered working in the roll cutting department, and operating the perforators from 1918 up through 1921. The perforators cut fifteen copies simultaneously. Using an old printing press, he also printed the little tune strips that slip into the tune title strip slots on the National 8-coin slot casting. Gale also recalled putting the words on the “8T8 Har-Mo-Nie” player piano rolls by air brushing through a stencil overlay.

(Please note: the historical information within this Roll Cutting Department section is a partial rephrasing and/or clarification of text derived from telephone interviews and correspondence conducted circa 1980, and made available through the courtesy of John H. Perschbacher,)

Piano Serial Numbers

There is still a lot of mystery surrounding National automatic piano serial numbers. By 1925, according to company printed material, there were some 4,000 pianos being managed by the organization. By 1930 that number had risen to 4,200 pianos—all National 5-Cent Automatic Pianos.

As of this writing:

• On the low end of the scale, known pianos are numbered 240, 348, 794, and 1485, each with a 32-note piano mid-section, a Simplex stack, and a revolver magazine with either of two known variations of the spring-barrel type of rewind system. This spread could potentially represent 1,245 pianos.
• On the high end of the scale, there are extant pianos numbered 6260, probably 6995 (first digit unknown), 7013, 7125, 7150, 7173, 7190, 7207, 7244, 7274. 7352, 7504, 7524, 7535, and 7823, each with a 30-note mid-section piano, a National built stack, and a late revolver magazine with the torsion spring type of rewind system. This spread could potentially represent 1,563 pianos.

The above figures alone count for some 1,245 + 1,563 = 2,808 pianos. Now, as an experiment, let’s assume that the first or earlier serial number range began at number 1, and the later range started at 6,000. By doing this we arrive at this calculation: 1,485 + 1,823 = 3,308 pianos. This leaves us some 892 pianos shy of the approximately 4,200 pianos reported in the circa 1930 company booklet. If we add this number of shy pianos to the piano serial number 1485 (892 + 1485) we get piano #2377 as possibly the last piano to be made before a major serial number break occurred. However, the piano quantities reported by the company in promotional material were only approximations, which were probably enhanced to the upside. Thus, some leeway need be given for any hypothetical calculations, but should the math here be somewhat credible, the serial number range in which the break occurred may have been near or in the 2300 range.

The patent for the torsion spring type of rewind system—which all known pianos in the 6,000 and 7,000 number range employ—was filed on October 23, 1923, and was granted on October 25, 1927. The patent was assigned by the inventor, Clifford H. Green (of Grand Rapids, Michigan), to Automatic Musical Instrument Company, also of Grand Rapids, Michigan. This is interesting in that the Automatic Musical Instrument Company supposedly did not officially exist until 1925. However, even more intriguing, is 1923 the year in which the serial number break occurred, and when pianos with a 30-note mid-section, a National built stack, and a revolver magazine utilizing the new torsion spring rewind system came into existence? Is it accurate to imply that any National piano with a 30-note mid-section, or a National built stack was built circa 1923 or later?

Understanding the evolving mechanical improvements and precisely when new designs elements were implemented is one of the goals of further research. But until many more National pianos can be surveyed, covering a much broader range of serial numbers for both the apparent early and late serial number ranges, many questions will continue to linger, and dating a National automatic piano will remain somewhat elusive.

Art Reblitz thinks it is likely that National would have begun a new numbering series at the same time the company introduced a group of major mechanical design changes that required different repair parts—possibly when the company changed from spring barrels and cast iron parts over to torsion springs and die-cast parts, assuming these changes might have happened at the same time. That way, when an operator had to order a replacement part, the serial number series would tell the factory which style of part to send out, without looking up the specific serial number in a log book. Moreover, in light of the large number of Nationals that were destroyed, it seems logical that most survivors would be the newest ones that were still in better working condition, which would have higher serial numbers.

Piano Plates

To date, in National pianos, three distinctive, differently laid out and decorated piano plates have been observed, but although each is unique in some ways, all are, at the same time, standard 88-note upright piano plates. It seems that National did not order special pianos, other than the enclosing cabinetry design was special, and it was shipped without a keyboard. The piano actions were also standard, but abbreviated with fewer than 88 action notes, omitting at least some portion of the unused piano action parts. There may be more plate variations yet unknown, but, for the time being, the three known, and conspicuously different plates, help shed a little more understanding on the evolution of National coin pianos. But what vendor, or vendors, supplied the different piano/plate designs is still a mystery. Nevertheless, each of the three plate variations observed to date have easily recognized unique characteristics, which makes it rather easy to quickly identify one from another, as follows:

Early (Piano #240, #348, and #794):

• Plate color: Greenish gray, with gold pinstriping (similar in color to early Wurlitzer piano plates).
• Plate lettering: Raised lettering, “National Automatic Music Co,” resting atop an elevated decorative cartouche located in the upper right-hand corner of the piano plate.
• Serial Number: Painted digits on plate surface, located near the plate top and to the right of the tuning pins for the bass string section.
• Piano action: 26-note bass; 32-note mid-section; 21-note treble = 79-note piano.
• Stack: 14-note bass; 32-note mid-section; 19-note treble = 65 playing notes for the 3-tier Simplex stack.
• Unused piano action notes: 14 piano action notes not used = 12 lowest bass piano action notes and the 2 highest treble piano action notes.

Intermediate (Piano #1485):

• Plate color: Light gold.
• Plate lettering: Hand painted ornate calligraphy, “Property of, National Automatic Music Co., Grand Rapids, Mich.,” located in upper right-hand corner of the piano plate.
• Serial Number: Painted digits on plate surface, located near the plate top and to the right of the tuning pins for the bass string section.
• Piano action: 21-note bass; 32-note mid-section; 21-note treble = 74 note piano action.
• Stack: 14-note bass; 32-note mid-section; 19-note treble = 65 playing notes for the 4-tier Simplex stack.
• Unused piano action notes: 9 piano action notes not used = 7 lowest bass piano action notes and the two highest treble piano action notes.

Late (Pianos with 6000 and 7000 range serial numbers):

• Plate color: Gold.
• Plate lettering: Raised lettering, “National Automatic Music Co,” is painted black and is located in the upper right-hand corner of the piano plate.
• Serial Number: Painted digits in oval opening in plate, located near the plate top and between the tuning pins for the bass and mid-section piano notes.
• Piano action: 28-note bass; 30-note mid-section; 19-note treble = 77 note piano action.
• Stack: 16-note bass; 30-note mid-section; 19-note treble = 65 playing notes for the National 3-tier built stack.
• Unused piano action notes: 12 lowest bass piano action notes not used.

The early and intermediate category plates are possibly from the same plate manufacturer, because there are some definite similarities. Both have a 32-note med-section, they have an identical tuning pin layout, both have the same protruding bosses around the holes for the bolts that secure the plate to the piano’s wooden framework, and there are a few other shared casting peculiarities. But why the earlier plates (pianos #240 and #794) had raised letters with the name “National Automatic Music Co.” cast into the plate, while the intermediate plate (piano #1485) had no raised lettering, just some hand applied ornate calligraphy on an unbranded plate, remains an unanswered question. Perhaps the plate manufacturer could not immediately fill an order for more National branded plates, and so shipped some unbranded plates until another batch of National branded plates could be manufactured. The late plates in pianos falling within the 6,000 and 7,000 serial number ranges, however, are very different than the plates that preceded them. They have a 30-note mid-section, the tuning pin layout is different, and there are other easily discerned differences. Thus, it is possible that the early and intermediate plates, with a 32-note mid-section, were from the same supplier, while the late plates, with a 30-note mid-section, came from a different supplier.

Moreover, there also seems to be a relationship between the style of piano plate and the player mechanisms. With still relatively few data points to go by, it is nonetheless beginning to look as though each piano’s plate design and scale characteristics tend to correlate with the specific type of pneumatic stack used, the three options, so far, being (1) a Simplex 3-tier stack, (2) a Simplex 4-tier stack, and (3) a 3-tier National built stack. However, any arbitrary categories, and whatever else we might like to think we know about the pianos and/or their cast iron plates, may well change dramatically whenever more details eventually come to light.

Piano Attachments—Muffler Rail vs. Mandolin Attachment

Mechanical music research is commonly limited to hazy memories, inaccurate word of mouth recollections, and/or oftentimes poor-quality photographs, which, when taken all together, can make it difficult to correctly differentiate between factory original and non-factory alterations. With this in mind, a significant number of National pianos have been identified with what looks like either a felt muffler rail, or a typical curtain type mandolin attachment, or simply a bare metal rod that can be manually lowered and raised but that is devoid of any apparent musical use. And so, considering the uncertainties, did some National pianos come with factory installed muffler rails, while others were fitted with mandolin attachments, or was only one of these things ever factory installed? Taking into account these factors, the most logical conclusion seems to be that it was a muffler rail that was originally installed, which makes good sense for the following reasons:

• National pianos with original looking (muffler/mandolin) rails with a mandolin curtain seem to be in pianos that also have, or once had, a manually operated muffler rail control knob.
• National pianos that have a non-original looking (muffler/mandolin) rail seem to have no telltale signs of ever having had a manually operated muffler rail control knob.
• Manual operation would not be a musically sensible way to turn a mandolin effect on or off. There might be some novelty effect at first, but then it would probably never be used by patrons, or, if engaged, it could inadvertently be left in the on position. It is just not a practical way to add a mandolin effect to the music.
• National music rolls do not feature typical mandolin arrangements.
• Automatically controlled mandolin attachments have been reported, but these are suspected of being installed by a route operator or more recently by a mechanical music collector, with the mandolin attachment reportedly controlled by connecting it pneumatically to the vacuum loud-soft controls.
• National music rolls do not have a provision to control the operation of any type of “extra” instrument, making an originally installed mandolin attachment an unlikely option, especially one that might be haphazardly turned on and off by curious and possibly rowdy patrons.
• A manually operated muffler rail is a sensible option for manual control, and it would be very useful in situations where a loud and raucous piano would be annoying, making it an ideal option for strictly manual operation.

To actuate the muffler rail (or mandolin effect) there is a machined metal knob with a knurled rim that is located on the right-hand side of the piano case a few inches down from the top edge. Turning the knob clockwise lowered the rail, while turning it counterclockwise raised it, and if turned far enough, passing the over-center position, it was locked in the upward disengaged position. On the rail itself, over each piano action bracket bolt there is usually some action felt wrapped and secured firmly around the metal rod, so that when it was lowered and if it bumped up against the action bracket bolts there was no metal-to-metal clanking noise, keeping the muffler/mandolin rail operation silent. Of course, some muffler rails could have long ago been modified or converted by a route operator or some more recent rogue mandolin attachment fanatic. But whatever the case, of the relatively small sampling of attachment rails noted to date, and that look professionally crafted and original, rather than being a kludged together adaptation, they have either a felt muffler curtain glued to the rail, or a mandolin curtain glued to the rail, or the rail (a round metal rod) is essentially bare.

Exactly how many National pianos originally had a muffler rail (or mandolin attachment) installed is unknown, and will probably forever remain a mystery. This is because there are no factory records to examine, and there is neither sufficient photographic evidence nor a large enough sampling of surviving National pianos to predict a quantity or percentage with any sense of accuracy. Adding to the mandolin attachment confusion, some of those reported are said to be automatic, in that they have a pneumatic mechanism installed that lowers and raises the mandolin attachment, casting further doubt on the originality of the installation. It would be easy to convert a felt muffler curtain to a mandolin attachment, by simply ripping off the muffler felt and gluing on a typical curtain type mandolin effect. Unfortunately, it is often difficult or next to impossible to ascertain the originality of a muffler rail or mandolin effect installation without personally being able to carefully examine and touch the piano in question.

Anyone able to shed more light in this subject is invited to share their knowledge at:

Pneumatic Stacks

There are at least two types of pneumatic stacks known to have been factory installed in National coin pianos, (1) the commonly observed National built stack and (2) a seldom seen Simplex style stack. Both are visually quite different, making them unmistakably easy to identify. The National built stack is open with all of the unit control valves and/or unit (piano) pneumatics visible and easily removed and serviced from the front, while the Simplex stack is semi-enclosed, with none of the valve units visible from the front, or even removable without first taking off the front cover, which forms the vacuum chamber that feeds the individual unit valves. Even then, the stack needs to be completely removed from the piano to easily service the valve assemblies.

But there is possibly one more stack type that might be encountered. These are late, circa 1928, Unitype stacks, bought from an outside supplier, and while it is reportedly possible that a National piano might have been factory retrofitted with one of these single-tier stacks, no such retrofits have, as of this writing, reportedly been observed in the wild.

National 3-Tier Stack with Front Facing and Accessible Valve Units

The National Piano Manufacturing Company built pneumatic stack is what is commonly found installed in National coin pianos. It consists of a rather standard looking wooden framework that supports three tiers of unit valve and pneumatic assemblies. There are 24 units making up the top row; another 24 on the middle row, and just 23 on the bottom row, for a total of 71 units. There are also 71 holes in the tracker bar, but tracker bar hole #1 for accent (or Sforzando) goes to the spill valve regulating pneumatic, and not to the stack. This leaves one unit valve control assembly unused for some other non-tracker bar related function, for which this “extra” valve pouch is connected to the rewind trip pallet valve, which opens when a lever riding lightly over the music roll paper falls into the rewind cutout at the end of the music roll.

The 6 stack unit valves devoted solely to control functions are located on the left side of the stack, and populate the two left-side outermost spaces on each of the three tiers, while all other positions are for unit pneumatics that play piano notes. Notice the difference in terminology used: Unit valves and unit pneumatics. The 6 special unit valves used for control functions do not have pneumatic bellows attached to the bottom of the valve assembly, but instead have a single spacer block of wood with a 3/8 inch hole drilled through to the rear facing side. This hole is fitted with a brass elbow tubing connector, which is in turn connected via rubber tubing to whatever control function pneumatic is appropriate. These control valves perform the following functions: Hammer rail up (and latched up), hammer rail down (locking latch is released), soft vacuum level on, soft vacuum level off, sustaining pedal, and rewind.

Each of the 71 valve units are firmly held in place by threaded metal studs inserted into the stack framework and that are positioned between the valve units, each with a metal washer and nut on the outer end of the stud that can be tightened to hold the valve units in place. These stacks are easily identified by the individual rubber tubes from each of the value units converging on the area of the tracker bar, resulting in what might appear to be a hopeless tangle of rubber tubes.

From a structural viewpoint, the vertical wooden supports on either side of the stack are notably different. The left leg is structurally complex and consists of a foot and attached vertical structure, both of which have internal channeling to distribute vacuum to the three horizontal tiers of unit valve assemblies. The other vertically oriented leg, located on the right-hand side, is simple in comparison, is without any hidden channeling, and serves only to physically support one end of each tier.

These stacks, at quick glance, might all look to be the same or, at the very least, quite similar in appearance. Upon closer look, however, there are some obvious but perhaps not immediately eye-catching differences, such as variances in the wooden spacer blocks that take the place of omitted unit pneumatics where divisions in the piano string sections and/or piano action brackets occur, and, perhaps much more significantly, in the design of the stack’s main structural framework. For instance, end legs and their foot design can vary, and with more information these differences might be useful in understanding the evolution of National automatic piano mechanisms.

The Tilting Yoke Atmospheric-Port Bleed Type Control Valve is the vacuum expression methodology used with the commonly observed 3-tier National built pneumatic stacks.

Simplex 3-Tier Semi-Enclosed Stack with Front Access Cover

Simplex style stacks all share at least two traits: (1) the unit valve design (i.e., unit control valves and the unit pneumatics that play piano notes), and (2) the generalized trunk board hole pattern layout, to which the various unit valve assemblies are attached by a pair of wood screws. But there are also necessary differences that must be accommodated between various player systems and the piano plate scales that will inevitably be encountered. For instance, the trunk board width and the precise layout of valve attachment holes will vary with different piano plate designs, which may group and space the piano strings in similar looking patterns but that require spacing adjustments in the trunk board layout. It is well known that Simplex manufactured a wide variety of stacks with special spacing requirements for many competing player piano brands, and to a lesser extent for coin pianos. And so, was National a big enough customer to warrant a special Simplex stack, ready to mount in a National piano? Or did National purchase some basic Simplex components and finish the work of laying out and building the stack themselves? In as much as there is currently too little evidence to definitively answer this kind of question one way or another, let’s bypass it for now and assume that the Simplex style stack, as installed in National automatic pianos, was made in its entirety by Simplex, but installed, more or less as is, by the National Piano Manufacturing Company.

For sure, the 3-tier Simplex style 65-note stack design used by National bears no visual resemblance whatsoever to the more commonly found 3-tier National built 65-note stacks. Although both stacks provide unit pneumatics for 65 piano notes, along with 6 unit valves dedicated to various control functions, that is pretty much where the similarities end. The unit pneumatics are basically divided into three groups, the dividing line coinciding with the string divisions in the piano plate. There are 14 individual unit pneumatics in the bass section, 32 in the middle section, and 19 in the treble section, for a total of 65 playing piano notes. Curiously, the Simplex stacks examined are laid out to accommodate National pianos with 32 notes in the mid-section of the plate, whereas all of the pianos examined with National built stacks have only a 30-note mid-section plate design. This implies that Simplex stacks may have only been installed in a select production run of National pianos fitted with a specific piano plate design, and perhaps with their own serial number range, making these pianos structurally different from the common variety seen today that have National built stacks. This perhaps trivial seeming difference between the two types of piano plate styles might someday turn out to be a solid clue as to what National was building during its earliest years.

The Simplex stack is built around a central trunk board, which is 3-tiers high and with a width that is somewhat more than enough to accommodate the string layout on the piano plate. It is this trunk board to which all of the unit valve assemblies are fastened, whether they be for playing piano notes or carrying out some important control function. As the name implies, the trunk or channel board feeds necessary resources through channels to the attached unit valves, so that they are functionally useful, such as connecting a tracker bar signal to the appropriate unit valve, or providing a vacuum source to each valve. Additionally, the trunk board is accurately drilled to hold each unit valve, using a pair of wood screws, firmly in its correct position. Covering and protecting the trunk board is a removable wood panel that has a framework around its outer edges, so as to form an air-tight cavity or compartment once the cover is affixed over the front side of the trunk board. The cavity formed by the cover is then connected to a source of vacuum, which, in turn, provides a vacuum source to all the attached unit valve assembles. This design is elegantly simple and durable, but has its drawbacks. To remove any unit valve for servicing the front cover must be removed first in order to access and unscrew the pair of wood screws that hold it attached. But then, once the unit valve is loose, that stack must still be taken out of the piano to gain easy access to the unit valve assemblies.

But there is more to the stack than the trunk board and its front cover, because it needs to be held rigidly upright and in the correct position. This is accomplished by hefty and blockish wooden vertically oriented sections at each end of the stack, also adding depth to the stack approximately equal to the length of the individual unit valves sticking outward on the backside of the trunk board. The end pieces also support a top shelf, to which is attached a wooden rail with 65 wood flanges, each flange acting as the pivot point for a movable finger that is raised at its outer end by a piano pneumatic. This finger also has a capstan screw near the outer end used to adjust the contact height with the piano action, so that the stack accurately conforms to the piano action wippens, thereby eliminating any lost motion between the stack components and the piano action. This self-contained arrangement makes it relatively easy to slip the entire stack out of the piano for servicing, and then put it back in without undue effort.

Screwed to the front side of the stack, and to the left of center on the trunk board cover, is a large block of wood bored and channeled to accommodate the needs of a poppet type spill valve, which is controlled by the cam-work on the Rewind/Play control system located to the left of the roll changer and mounted on the flat “keybed” area shelf. This spill valve is either wide open or closed tight, and it possess no vacuum regulation capabilities at all. Near the left side of the front cover there is a medium sized elbow jutting out with a rubber tube that curls upward toward the top of the piano action. The elbow is inserted into the vacuum cavity made by the trunk board cover, while the rubber tubing connects to the tune indicator mechanism to the left of the piano action, and in particular to the pneumatic that raises the metal shutter that reveals what tune is currently being played. Moreover, if the piano has been retrofitted with a dog race diorama, the rubber tube is teed to also connect with a small valve chest under the race track area.

The Balance-Beam Design (with Coil Spring Balanced Ventil Valve) is the vacuum expression methodology used with Simplex style pneumatic stacks.

Single-Tier Unitype Stack

In the writings of the late Durrell Armstrong it is mentioned that some National automatic pianos were retrofitted with Unitype pneumatic stacks. These were a single-tier stack with a row of narrow unit pneumatics that faced forward, and that had a metal spoon attached to the rear end that lifted a piano wippen when the pneumatic collapsed, causing the piano note to sound. And so, it may be that some National pianos were retrofitted with a Unitype stack, either by the factory or by some route operator. But if such retrofitting did in fact occur it cannot currently be verified by any original factory or other documentation, and so it remains hearsay. Moreover, no extant National pianos fitted with a Unitype stack have been observed, reported, or rumored to date. Nevertheless, an example might someday be discovered, and so to become familiar with Unitype stacks please click on the patent illustration thumbnail at right.

Anyone able to provide information regarding the retrofitting of Unitype pneumatic stacks in National automatic pianos is invited to share their knowledge at:

Expression Control Devices

There are at least two types of pneumatic volume expression control devices observed in National automatic pianos. The first, which we have arbitrarily named “Tilting yoke atmospheric-port bleed type control valve” is the type commonly observed in most of the National coin pianos that survive to this day. Moreover, it is an indispensable companion for the National built 3-tier pneumatic stack. Whether, or not, our terminology for this expression control system bears any resemblance to what the National Piano Manufacturing Company originally called it will probably forever remain a mystery.

The second kind of volume control device used a “balance beam” type of regulated spill valve, which was integrally built into the left-side structural protruding foot of Simplex style pneumatic stacks. A separate board with paired lock and cancel pneumatics (for high vacuum on or off) was butted up against the balance-beam mechanism and was attached by a threaded adjusting screw to one end of the wooden balance-beam. This system is much simpler to regulate, but has no provision for the accent (or Sforzando) hole #1 on the tracker bar.

It is important to keep in mind that the vacuum expression control system is a critical part in any National automatic piano, no matter whether the expression device be made by National, Simplex, or someone else. This is because the expression system not only controls the vacuum soft/loud levels, but it also acts as the safety spill valve for the pump. The vacuum system in National pianos has no customary vacuum reservoir with a traditional spill valve. This means that the expression control’s spill valve must be capable of safely accommodating no less than the maximum airflow passing through the vacuum pump, otherwise damage to the pump bellows and/or connecting rod links may occur due to being overloaded and/or stressed to the breaking point.

To recap, here are the two pneumatic expression control variations observed to date:

• National tilting yoke atmospheric-port bleed type control valve.
• Simplex style balance-beam design (with extension spring balanced spill valve).

National Tilting Yoke Atmospheric-Port Bleed Type Control Valve

This commonly observed expression control system could perhaps better be described as two different but interdependent and equally important components, each quite different in appearance, construction, and purpose. The most obvious component, and usually the first and only part of the system easily seen and recognized, is what I have been referring to as the “tilting yoke atmospheric-port bleed type control valve,” a long and cumbersome name perhaps, but one that describes the unit quite nicely. The name is derived from the obvious and centrally located tilting yoke, the lower tail end of which opens or closes an atmospheric-port that bleeds off and reduces the output vacuum level of the unit. The expression control unit sits at the far left of the “keybed” area, usually mounted on top of the stack’s left-side foot, but occasionally it has been seen mounted on the “Keybed” area in front of the stack foot. The other part of the expression system is the regulated spill valve, which is usually mostly hidden behind a mat of rubber tubing that connects the tracker bar with the 3 tiers of individual unit valve assemblies on the stack.

The so-called “tilting yoke” unit consists of a die-cast block with a wide channel formed into the top (early examples used an unstable pot-metal casting that warped and broke apart over time). Secured to each side of this channel is a narrow pneumatic that faces forward. These two little control pneumatics are used to alternately tilt a thin brass yoke (located on the front of the die-cast block) one way or the other—the yoke pivoting about a machine screw. The tail or lower end of this brass yoke either blocks off or exposes a small 1/8" diameter port or orifice, which has been drilled into the front side of the metal block. This small port connects internally with a similarly sized channel drilled through the block from side to side, for which the vacuum supply is connected to its inlet connection on the left side; the right side being the outlet for the unit. There is an adjusting screw into this main vacuum channel located between the inlet and center air bleed port, whose sole purpose is to restrict the airflow allowed through the main vacuum channel. Another adjusting screw located on the right-hand side of the aluminum block is used to adjust the airflow through the small atmospheric port.

• When the bleed port is closed (tracker bar hole #70) with the upper yoke tilted to the left, the vacuum level is low, and the piano volume is soft.
• When the bleed port is open (tracker bar hole #71) with the upper yoke tilted to the right, the vacuum level is normal or high, and the piano volume is loud.

The main side-to-side vacuum channel delivers at its right-hand side outlet a variable output vacuum level that depends upon (1) the position of the brass yoke, which either closes or exposes the atmospheric port, and (2) the two adjusting screws which regulate the airflow rate within the device. The output is connected via a rubber tube to the regulating pneumatic making up the other half of the expression system, and that is an independently situated regulated spill valve unit. This spill valve component is described as regulated because it acts to both regulate the vacuum level throughout the player system, and to serve as the vacuum dump valve during the rewind and roll changing operations.

The regulated spill value unit consists of a die-cast pedestal that serves not only as a mounting platform for a suspended medium sized control pneumatic, but also as the actual spill valve face itself. Directly over the spill valve face is the suspended pneumatic (using a metal bracket attached to either side of the die-cast pedestal). The top side of the pneumatic is fixed to the metal bracket and therefore stationary, while its bottom movable leaf is free to relax and fall open and rest flatly on the spill valve face. But the actual packing leather that seals against the spill valve face is not glued to the lower movable leaf of the pneumatic itself, but is instead affixed to a thin metal plate, which is, in turn, loosely held in place on the lower movable leaf by two small wood screws positioned on either side about halfway toward the back end of the pneumatic. This metal plate is aligned with the front and sides of the pneumatic, its width being the same as the pneumatic, but its length, or depth, is only about one-half that of the overall length of the pneumatic. The two little wood-screws that keep the metal plate in position are located at the back end of the metal plate (about midway on the pneumatic) and they are purposely left loose. When the lower leaf of the pneumatic is raised the thin metal plate must be able to swivel at the two loose attachment points, so as to peel away from the spill valve face gradually at an angle, opening up first at the back of the plate and then arching upward toward the front. The mechanical leverage gained by this arrangement allows for more precise control over the vacuum spill process that would otherwise be possible.

It is important that the leather covered metal plate attached to the lower leaf of the regulating pneumatic tends to rest firmly seated on the spill valve face. To accomplish this, an internal but adjustable compression type spring is used to push a small plunger downward into a dished-out area inside the movable lower leaf of the pneumatic. To make this internal compression spring easily adjustable, on the top side of the pneumatic toward the front, there is a partially hollowed out brass stem that sticks up. The upper part of the compression spring slips into the hollowed-out area of the stem, and through the top of the stem is a threaded hole for a machine screw, which can be screwed in or out, thereby adjusting the spring tension pushing down on the lower movable leaf of the pneumatic. Occasionally the machine screw is unembellished, bare for all to see, but more often than not it is combined with a knurled brass cap-sleeve, or cuff, that facilitates the easy adjustment of the spring tension upwards or downwards by gripping and then twisting the outer sleeve one way or another. When the piano is in play mode, this spring tension, working in conjunction with the airflow adjusting screws on the “tilting yoke” mechanism, will determine the “at rest” player system vacuum level. Except, of course, during the rewind and roll changing phase, when the regulating pneumatic is mechanically raised by a cam operated lever, so that all vacuum is spilled whenever the tracker bar is raised away from a music roll.

While it may seem counterintuitive at first, the higher the vacuum level in the regulating pneumatic the lower the overall player system vacuum level, which directly affects how soft or loud the piano plays. Because of the restricted airflow on the vacuum inlet side of the “tilting yoke” unit, when the atmospheric bleed port is open the vacuum level at the outlet side, which is applied to the regulating pneumatic, will be relatively low, due to air bleeding into the exposed bleed port. Thus, the bottom movable leaf of the regulating pneumatic will only be gently pulled away from the spill valve face, and so the system will balance out with a relatively high vacuum level. Conversely, however, when the atmospheric bleed port is closed the vacuum level in the regulating pneumatic will rise, because no vacuum is being bled off through the bleed port. This causes the regulating pneumatic to more vigorously work at pulling the movable leaf of the pneumatic upwards and away from the spill valve face, which, once the system reaches a new equilibrium, will result in a lower vacuum level.

On the top side of the regulating pneumatic is a second but smaller brass fitting. It is connected directly to tracker bar hole #1, which is used for accent (or Sforzando) effects. When air rushes into the accent hole through the tracker bar the vacuum level in the regulating pneumatic is almost instantly lowered (as a consequence of those airflow limiting adjusting screws in the “tilting yoke” device), thereby lessening the tug trying to pull open the spill valve. This has the immediate effect of momentarily raising the overall vacuum level throughout the player system, thereby providing an accent to the piano melody.

Out the backside (or sometimes out the bottom) of the die-cast spill valve pedestal is an approximately 1-inch diameter tube fitting to which is attached a rubber hose. This hose is teed into the main vacuum supply hose going between the vacuum pump and the pneumatic stack. There is no vacuum reservoir in National coin pianos, although it is something commonly used by other coin piano manufacturers to smooth out the periodic fluctuations in airflow due to the physics of how bellows type pumps are constructed. Consequently, the spill valve in National pianos must regulate not only the desired vacuum level, but it must also reliably spill and relieve the constant pulsing caused by the rotary pump, along with any excess vacuum that may be generated by the pump, otherwise, if the vacuum level were to go too high the pump could be severely damaged.

But there is more to consider here, because of the compound duties of the regulated spill valve it is imperative that the two airflow limiting adjusting screws in the “tilting yoke” unit, as well as the spring tension in the spill valve pneumatic, be properly adjusted. And, as should be obvious, these three adjustments, which are spread across two separate components, are closely interrelated, with any adjustment of one affecting the other two. This begs the question: How does one go about adjusting these three items for maximum effectiveness? It would be wonderful if we could simply refer to an instructive pamphlet by the National company itself, but no such thing is known to exist. Thus, some deductive reasoning seems appropriate, although it may not achieve the exact same result as originally sought by the National Piano Manufacturing Company.

Here are some ideas regarding the adjustment options utilized by the vacuum expression control: Please consider any methodology expressed here as experimental, subject to change, and never any kind of official or perfected way of adjusting the vacuum expression system. Moreover, everyone will probably have differing ideas as to what constitutes an ideal level of performance. Nevertheless, let’s start exploring the subject by noting what the three adjustment options are, how they function, and why:

1. The airflow adjustment screw for the vacuum inlet side of the “tilting yoke” unit. If the vacuum source is left unrestricted the atmospheric bleed port will have little, if any, effect when opened or closed. This same issue also applies to the accent (or Sforzando) hole #1 in the tracker bar. Without some degree of airflow restriction to the vacuum supply it would simply overwhelm the significantly smaller air volume capacity of either the bleed port or the accent hole in the tracker bar. Thus, the trick to setting the vacuum source adjustment screw is to restrict the amount of airflow just beyond the point where the spill valve regulating pneumatic is capable of quickly and reliably responding whenever the atmospheric bleed port or the tracker bar accent hole is fully open or closed. Being that the bleed port and tracker bar hole can have somewhat similar airflow capacities, setting the vacuum restriction properly for the bleed port should also satisfy the other. The essential point here is this: When the vacuum restriction is adjusted properly the system vacuum level will quickly and smoothly settle at either the loud or soft setting depending upon whether the bleed port or tracker bar accent hole is open or closed, both of which have a similar effect on the regulating pneumatic. It is also worth keeping in mind that the vacuum side adjustment also has, within limits, the effect of increasing the depth or vacuum range between the minimum and maximum levels. Use the preceding comments as a starting place, keeping in mind that the perfect vacuum inlet side adjustment will best be determined through experimentation and personal preference.
2. The airflow adjustment screw for the bleed port on the front side of the “tilting yoke” unit. Any airflow restriction through the bleed port is only effective when the port is open, which is the only time when air can move through it, whether restricted or not. The adjustment screw for the bleed port can be adjusted to restrict the airflow though it just enough to dampen any adverse spill valve flutter or other equilibrium seeking effects that might be generated when the bleed port suddenly opens. I wonder if this particular adjustment was added as an afterthought during factory testing of the new expression system, in response to occasional instabilities and perhaps noise producing resonances that might sometimes occur when the bleed port was abruptly opened without any flow restriction. It could also be said that this restrictive adjustment can also be used to produce a brief mini-crescendo effect when the bleed port opens, with a similar effect when the bleed port is closed, while the player system seeks to re-balance and adjust itself to a new but lower vacuum level. Consider the bleed port restriction adjustment as a way to finely tune the stabilization of the system, reducing or eliminating unwanted resonances and surges, when suddenly shifting from a low to high vacuum state.
3. The coiled compression spring inside the regulating pneumatic mounted above the die-cast spill valve pedestal. The purpose of the compression spring is to push the lower movable leaf of the regulating pneumatic down against the spill valve surface, attempting to prevent any air-flow through the spill valve. But when the vacuum pump is powered on this compression spring will be met with an opposing force as a partial vacuum in the regulating pneumatic causes an upward lifting force on the lower movable leaf. These two opposing forces seeking equilibrium results is an ongoing balancing act between the compression spring and the lower leaf of the regulating pneumatic. And so the question: How does the compression spring influence the resulting system vacuum level? If the tension on it is increased, thereby applying a greater force attempting to close the spill valve, the resulting vacuum level will go higher. This is because it will take a higher vacuum level in the player system to maintain enough lift in the regulating pneumatic to hold open and maintain the pump spill volume necessarily going through the spill valve. Conversely, when the spring tension is lessened the vacuum level will fall to a lower level, because a lower vacuum level will be required to maintain air-flow equilibrium through the spill valve. As such, the spring tension adjustment inside the regulating pneumatic can be viewed as a way of adjusting, within narrow limits, the “at rest” soft versus loud vacuum range, raising or lowering it as necessary, so as to achieve a soft playing volume level whereby all of the piano notes continue to play reliably with equal strength, and with good repetitive accuracy with no dropouts.

Of course, it almost goes without saying, any adjustments made will probably require some patient back and forth twiddling before the desired perfection in vacuum expression levels will be achieved. And much of that success will depend upon the quality of any restoration work performed on the player mechanisms. For instance, any player system full of leaks and fettered by some poorly working components will never perform satisfactorily, no matter how you might adjust the expression controls. Possible reasons for troubles abound, but are often simply because the pump, regulating valves, and certain air passageways are simply overwhelmed due to the sheer volume of air they are required to process. This overload, in turn, causes other previously operational but now compromised components to act sluggishly or become unresponsive altogether, all because some leak strewn player system forces crucial components well beyond their design capacity.

Simplex Balance-Beam Design (with Spring Balanced Ventil Valve)

The Simplex style balance-beam type expression control is visually simple and intuitively easy to understand and adjust. The ventil or spill valve and its mating metal seat is located on the topside of what casually looks to be an extended foot for the stack’s left side structural support. This foot extension has a vacuum channel partially through it (interconnecting it with the spill valve cavity), which is part of the internally channeled vacuum supply system that feeds each of the stack’s three tiers of unit valve assemblies. The circular metal seat is held fixed in place by two small wood screws, the metal seat being directly above the actual ventil or spill valve, which can freely move up and down within the valve cavity. When a vacuum is applied to the valve cavity the spill valve body will naturally be pushed down and open by atmospheric pressure, unless, and to the extent that an upward balancing force is applied to the valve body. This is where the balance-beam apparatus comes into play. The beam is supported mid-point by a sturdy cast brass stanchion, that also serves as the pivot point for the balance-beam. The brass stanchion is mounted on the stack foot extension near the spill valve, but also located midway between it and a pair of lock and cancel pneumatics mounted side-by-side on an adjoining wooden board that has internal channeling to provide tubing connections for the two control pneumatics.

A vertically oriented hole is drilled near each end of the balance-beam, through which is placed a threaded eyelet that is adjustable up or down using leather nuts. One end of the balance-beam is connected to the spill valve by a coiled extension spring, while the other end is connected to a wide control pneumatic that once activated locks in the down position until unlocked by another but narrower pneumatic at its side. With the wide locking pneumatic fully relaxed, or open, the spring tension on the ventil valve is adjusted to provide an ideal low vacuum level (piano soft). Then when the wide locking pneumatic is activated and latched down it pulls the balance beam down with it, in turn raising the other end, which then increases the spring tension on the ventil valve, thereby causing the vacuum level to rise (piano loud).

The balance-beam spill valve in this type of expression system must have sufficient spill capacity to handle the entire airflow volume generated by the vacuum pump, otherwise excess vacuum could not be relieved, causing the vacuum level to rise beyond safe levels and possibly damage the pump. However, the balance-beam type spill valve used with Simplex stacks does not do double duty, i.e., regulating the vacuum level plus act as a vacuum dump valve during the rewind and roll changing operations. To dump the vacuum a separate, non-affiliated poppet style dump valve is attached to the front cover of the Simplex stack. This large volume capable valve is operated by one of the cams on the Rewind/Play camshaft.

Oddly enough, the balance-beam system offers no provision whatsoever for the accent (or Sforzando) perforation in the music roll (tracker hole #1), which is left unused on the tracker bar. This omission has served to cast doubt on the originality of the balance-beam device. Was it factory installed by National, or was it installed by someone else? The consensus at this time is that the balance-beam vacuum expression system is integral to the use of a Simplex style stack, and was therefore likely factory installed along with the stack.

Adjusting the balance-beam unit is intuitively easy, it being so simple in both theory and construction that correctly optimizing its operation requires no heady instruction sheet or hand-holding to enjoy successful results. It is, however, important to remember that the minimum vacuum level to be allowed must always be a level at least a little above the point whereby the stack unit valves begin to function unreliably, first symptoms often being certain piano play weakly, or have lost their repetitive accuracy, or do not play at all. Moreover, any valve reliability issues will be exacerbated when the player system is full of leaks and malfunctioning but not necessarily crucial components. Suffice it to say, the Simplex style stack associated balance-beam expression control is probably one of the easiest components in a National coin piano to restore, adjust, and then maintain.

Piano Hammer Rail and Sustaining Pedal Control

The hammer rail (soft pedal) is controlled by a pair of interlocking motor pneumatics located in the upper left rear corner of the case, directly above the piano action. Both pneumatics are mounted on the same wood block, which is screwed to the inside of the piano case. The lower pneumatic lifts the hammer rail and when fully collapsed a spring loaded and felted wooden hook falls over a metal blade and latches the pneumatic in the upward closed position. At this point the hammer rail (or soft pedal) is locked in the raised (or soft pedal on) position. The smaller upper pneumatic, when caused to collapse, pushes the wooden hook off of the latching surface, allowing the hammer rail to fall back (of its own weight) to its normal rest (soft pedal off) position.

The sustaining or loud pedal is located on the left side of the “keybed” area behind the left foot of the stack. It is operated by a single pneumatic that is mostly obscured from view by rubber tubing and the stack foot. It functions by pushing upwards against a sturdy lever that is part of a round metal rail running the length of the piano action, which, in turn, lifts the piano dampers from the piano strings. Unlike the hammer rail (or soft pedal), there is no lock and cancel system for it, but simply a single chain perforation on the music roll, which when present causes the sustaining pedal pneumatic to collapse, lifting the piano dampers.

Revolver Magazine (Roll Changer) and Its Control Systems

There are two known but fundamentally different revolver magazine configurations Patented by the National Piano Manufacturing Company, one with a stationary tracker bar and the other with a movable tracker bar. Both types have carriers (or roll stations) for eight music rolls. The first and very early style revolver magazine had a stationary tracker bar coupled with a magazine that slid toward or away from the tracker bar on two horizontal rails. The Patent, No. 1,217,271, titled Automatic Musical Instrument with Magazine, was filed on July 18, 1912, and finally granted on February 27, 1917. Disappointingly, there are no known extant National pianos with this type of revolver magazine arrangement. Consequently, the only reliable information regarding this horizontal sliding type is from the Patent application, and so this type of revolver magazine will not be detailed in this writing. Click here to view Patent Application No. 1,217,271.

The second type of revolver system utilized a movable tracker bar with the revolver magazine axis journaled in stationary bearings (below the “keybed” area) and fastened to vertical structural components of the piano case. Initially the tracker bar moved back and forth horizontally; not the revolver magazine, as in the previously mentioned Patent. The Patent, No. 1,068,217, titled Automatic Piano Player Having Self-Settings Device was filed on August 7, 1912, and granted on July 22, 1913. Click here to view Patent No. 1,068,217. And like the sliding revolver magazine, no examples of a revolver setup with a horizontally moving tracker bar are known to exist. It is a relatively short evolutionary jump for the idea of a horizontal tracker bar to be reoriented so that it moves up and down vertically, a simple change that substantially reduces the depth of the piano case. Now you have a revolver magazine system that could very well be the start of an evolutionary outcome that is universally observed in the typical National automatic piano still enjoyed by collectors today, and that is the configuration that is covered extensively in this writing.

Almost certainly, the revolver (roll changer) magazine and its associated intricate control mechanisms are usually the most eye-catching mechanical feature of the National Automatic Piano. Moreover, when the numerous scattered but closely interrelated components are all lumped together into one whole system it might seem to be an unfathomably complex mechanical nightmare. However, when the apparent complexity is broken down into logical units of basic functionality there is a beautiful simplicity that is easy to see, appreciate, and comprehend. To begin sorting out the various functionally unique sections, the revolver (roll changer) magazine can be categorized as a separate cluster of associated parts whose sole purpose is to hold eight music rolls (each with its own personal take-up spool and spring-loaded rewind gearing) and then, on command, rotate on its axis until stopped when the desired music roll is situated beneath the tracker bar. The revolver assembly reliably performs this function but no other, and its action is coordinated by the two separate but interconnected control systems mounted on either side of it on the flat “keybed” area of the piano.

Temporarily setting the revolver magazine aside, the “keybed” area revolver controls can easily and logically be further divided into (1) the Rewind/Play mode functions located on the left side of the revolver, versus (2) the music roll selection functions located on the right side. Both sides and the functions they carry out are equally important. If either a left or right-side mechanism fails to perform its mandated duties correctly, or in a timely manner, the player system will be rendered inoperative. The logical divisions noted here apply to all National coin pianos, no matter whether they are equipped with a National built or Simplex style stack, and/or have either certain cast iron or die-cast components, such as for the worm gear clutch housing casting.

First, we will deal with the Rewind/Play mode control system, followed by a detailed explanation of the music roll selection mechanisms, which are located over on the right-hand side. Once these two spatially separate but interconnected control systems are understood adding the revolver mechanism to the mix will be an easy next step, with the process building toward a useful comprehension of how the three functionally independent but related systems work, interconnect, and why.

The Rewind/Play Control Sub-System

First off, the so-called “Rewind/Play” term is a convenient simplification, a space saving contraction for the more explanatory “Rewind/Roll Changing mode versus the Play mode” And the frequently used “Rewind/Roll Changing mode” terminology is a reminder that the “rewind” and “roll changing” modes always occur simultaneously, and are essentially one and the same thing. The rewind mode for any music roll automatically occurs when the tracker bar and take-up spool drive gear are lifted away from the revolver magazine, which is the same mode in which roll changing operations are allowed to occur. The player system is either in the Rewind/Roll Changing mode or it is in the Play mode, and it is always powered off and later powered back on while the system is in the middle of a Rewind/Roll Changing mode cycle.

It may be that the Play/Rewind interrelated clutch, gearing, and camshaft segments might look intimidating at first glance, but analyzing the whole thing bit by bit quickly reveals the easy simplicity of the entire unified assembly. To do this, let’s start at the left end and work our way to the right side. All of the Play/Rewind system components to be discussed are mounted on a single underlying metal plate, which keeps all of the attached components precisely aligned, more so than would be the case if the same components were mounted directly onto the underlying “keybed” area wood shelf.

But before going further let’s define what a cam and a cam follower are:

• A cam is meant to change the input motion—in our case is a gear driven rotating shaft—into a reciprocating motion suitable for use by some type of cam follower.
• A cam follower in National coin pianos is either (1) a bluntly rounded pin mounted on some kind of a lever, (2) the flat end of a pivoting lever, or (3) a u-shaped steel bracket that follows the outer surface contour of the cam.

Next is another important concept to grasp, and that is to understand that the rotational functions of the camshaft are divided into two distinct 180-degree rotational segments. In other words, each one-half revolution of the camshaft performs a different set of functions—either putting the player system into the Rewind/Roll Changing mode or into the Play mode. And after each one-half revolution the camshaft rotation is automatically and abruptly stopped at a precise point, until triggered again and the next one-half revolution occurs. If and when all is working as designed, whenever a coin is accepted by one of the eight numbered coin slots, and it then falls into the coin switch mechanism, the electric motor is powered on, and with the tracker bar in the raised position and the revolver magazine rotating to position itself so that the selected music roll will be put under the tracker bar. Thus, the camshaft is temporarily positioned in what could be called the Rewind/Roll Changing mode, a state invoked earlier when the rewind pallet dropped into the cutout at the end of a music roll, whereupon the tracker bar was then quickly lifted away from the music roll and the music roll began rewinding.

When the selected music roll is correctly stationed under the tracker bar the camshaft drive clutch is engaged, causing the camshaft to rotate one-half revolution and stop, putting the player system in the Play mode. When this camshaft activity occurs, the tracker bar is lowered onto the music roll, the revolver magazine is locked into position, the vacuum spill valve is allowed to close, and the spur gear that drives the music roll take-up spool is engaged, with some toe-tapping piano music soon to follow. The camshaft sequence is always the same (1) Rewind/Roll Changing mode when the system is powered on, (2) switching to the Play mode when the selected music roll is correctly positioned, and then (3) back to the Rewind/Roll Changing mode at the end of the music roll, when the system is either powered off or the next selected tune is played in succession. (Music roll rewinding is accomplished using a torsion spring system built into each music roll station on the revolver magazine.

Because the Rewind/Roll Changing mode is the rotational position of the camshaft whenever the player system is powered on, let’s define the Rewind/Roll Changing position to always be the starting point when describing the revolver magazine and its controls. To sum it up:

• Rotational position 1 — Rewind/Roll Changing mode: This is the starting place when describing camshaft functions. The tracker bar will be raised while automatically disengaging the take-up spool drive gear, the revolver magazine will be free to rotate, and the spill valve will be forcefully held open to dump all vacuum.
• Rotational position 2 — Play mode: This is the secondary position to be attained once the music roll has been correctly rotated into position below the tracker bar. In the Play mode, the tracker bar has been lowered onto the music roll, the revolver magazine is locked into the precise rotational position required, the spill valve is allowed to close, and the drive gear for the take-up spool has been engaged, pulling the music roll forward.

With some generalities out of the way, here is an enumeration of the individual camshaft components working from the simple left side trigger pneumatic over to right side, with the largest cam that raises and lowers the tracker bar assembly.

1. Rewind/Play trigger pneumatic: At far left is a simple pneumatic that is mounted horizontally atop a metal bracket. The pneumatic is activated by the rewind pallet valve (when it drops into the paper cutout at the end of the music roll). When this pneumatic collapses a metal finger attached to its topside pushes down on a trigger lever, which has an attached cam follower pin that rides on and follows the lower surface contour of the adjacent two lobe cam. When this follower pin is pushed downward, causing it to disengage from the right-angled lobe face, this trigger lever snaps to the rear, in turn causing the worm gear clutch to engage, which causes the camshaft to rotate 180 degrees before it is again stopped when the clutch is forcefully disengaged.
2. Rewind/Play cam: Located to the right of the Rewind/Play trigger pneumatic, it is the left most cam (at the end of the camshaft). This cam has two distinctly terminated lobes that are 180 degrees apart, each lobe having a gradual fall that abruptly ends with a 90-degree angle rise. The resulting vertical lobe face catches on an upright follower pin (during its rotation) and pulls it, and the lever to which it is attached, forward. This forward lever action, in turn, disengages the clutch for the worm gear drive, thereby stopping any further camshaft rotation. By design, whenever the cam follower pin is disengaged from the cam, for whatever reason, the worm gear clutch will be engaged and the camshaft will rotate 180 degrees (or one-half revolution) before it is again stopped. It takes 180 degrees of camshaft rotation to go from Rewind/Roll Changing mode to the Play mode, and another 180 degrees of rotation to go from the Play mode back to the Rewind/Roll Changing mode. As such, the player system is always either in the Rewind/Roll Changing mode or in Play mode.
3. Worm gear clutch drive assembly casting: This casting, which holds the worm and worm gear in place, and that also provides the bearing support for the left end of the camshaft, is what powers the camshaft. This specialized casting comes in at least two distinct forms: (1) a cast iron circular pedestal style with a square base, and (2) the more commonly found smooth die-cast squarish looking casting. Likewise, the separate bearing support for the right end of the camshaft will be of the same material as the worm gear housing, either cast iron or die-cast metal, with both types being functionally identical but differing in appearance. The worm shaft, laying horizontal in the lower part of the casting, has bearings on both the front and backside of worm, with the worm shaft extending to the rear well beyond the backside of the worm gear casting. On this shaft is a short brass shifter sleeve used to control the clutch action, which slips back and forth on the worm shaft and has a groove in the near end for a control yoke and a cone shape on the other. Behind it is a brass block that is tightly clamped to the worm shaft. This block holds the clutch pawl that engages or disengages one of several clutch engagement pins projecting from the hub of the chain sprocket behind it, and that is connected by chain to the vacuum pump. Moreover, the shaft coming out the backside of the clutch sprocket delivers motive power through gears, countershafts, and more sprockets to all other revolver magazine related mechanical systems. The center of the hub is fitted with a long hollow sleeve into which slips a length of the worm shaft, which than acts as a floating bearing for one end of the shaft powered by the chain sprocket. To prevent creeping when the clutch is disengaged, on the shoulder of the worm gear is a hard maple block that acts as a gentle but positive brake. This keeps the camshaft from rotationally creeping or wandering due to friction in the floating bearing surface between the worm shaft and the inside surface of the chain sprocket shaft. This brake should be adjusted only to the point whereby it prevents any adverse creeping by the camshaft.
4. Coin switch reset cam: This second from the left cam is immediately to the right of the worm gear clutch casting. It is a single lobe cam with a fast-rising short hump and then a quick fall at the end of the lobe. Its function is to reset the coin switch mechanism, turning off the power to the electric motor when the camshaft is completing its one-half revolution Rewind/Roll Changing mode cycle, with the cam’s quick fall coinciding with the stopping of the camshaft rotation. The metal cam follower pivots on a little bronze casting and is connected to the coin mechanism by a crank action on the larger hollow shaft that runs across the front of the “keybed” area and into the coin switch housing. This cam has no effect whatsoever during the camshaft’s one-half revolution Play cycle.
5. Spill valve vacuum dump (and play count) cam: The third cam from the left is elliptical in shape. Its sole purpose is to dump the vacuum supply during the camshaft’s one-half revolution Rewind/Roll Changing mode cycle, with the spill valve at its maximum open setting at the point when the camshaft rotation stops. During the Play cycle this cam has no effect on the spill valve. The cam follower is made of hard maple and it pivots on a little bronze casting. Slightly forward of the cam, on the front end of the cam follower lever, and attached by means of a metal wire loosely hooked through a hole in the wood, is a mechanical play counter.
6. Tracker bar lift cam: This cam, the fourth cam on the camshaft, as well as being the last cam on the far right-hand side, lowers and raises the framework from which the tracker bar is suspended, although there are other crucial functions performed by this cam that are not so obvious. This cam is elliptical and the cam follower consists of a steel u-shaped bracket attached to a heavy-duty lever arm. When the camshaft is positioned for the Rewind/Roll Changing mode the tracker bar is raised and the tapered locking hook that engages the revolver magazine’s slotted locking wheel is also raised, allowing the revolver to spin freely on its axis. Two additional functions are dependent on the position of the right-hand side of the tracker bar frame, which, when raised, also lifts (1) the little spur gear pinion and shaft that meshes with the large diameter gear on the music roll take-up spool, and (2) engages the tooth clutch that causes the revolver magazine to rotate. When positioned for the Play mode, the tracker bar is lowered and the locking hook is engaged to keep the revolver magazine properly aligned with the tracker bar. At the same time, on the right-hand side of the tracker bar frame, the tooth clutch that causes the revolver magazine to rotate is disengaged, and the drive pinion for the take-up spool drops down and meshes with the take-up spool gear, pulling the music forward under the tracker bar.

Rotation of the camshaft is triggered in one of two ways. The first method has been detailed above in paragraph #1, which explains the functionality of the Rewind/Play trigger pneumatic. When activated, this lone pneumatic unlatches the cam follower lever under the Rewind/Play cam, which allows the camshaft clutch to engage, rotating the camshaft one-half revolution before it is again stopped. During its momentary activity, the camshaft reconfigures the player system, taking it from the Play mode to the Rewind/Roll Changing mode. However, to reverse this action, i.e., to go from the Rewind/Roll Changing mode to the Play mode, requires a different way of triggering the clutch action—namely because the spill valve will still be fully open, dumping any vacuum, thereby rendering the trip pneumatic useless. And so, to bypass the problem a direct mechanical link to the coin switch mechanism is used. This is accomplished by using an alternate trip lever, which is attached to the smaller diameter shaft that traverses the front of the “keybed” area (and is placed inside the hollow but larger diameter shaft), and that terminates with a small bell-crank located inside the coin switch enclosure.

When the coin mechanism is activated by a coin, and the player system powers up, it will be in the Rewind/Roll Changing mode. Then, as the revolver magazine rotates, the studded selector cylinder in the coin switch enclosure rotates in synchronicity with it. When the rotational position of the studded cylinder signals that the selected music roll is located under the tracker bar, the trip lever (in front of the Rewind/Play cam) alternately pushes down onto and releases the cam follower lever, thereby engaging the clutch, putting the player system into Play mode. (Note that there are two control shafts that traverse the “keybed” area, a smaller diameter “Play mode trip” shaft that goes through the hollow and larger diameter “coin switch reset” shaft.)

It is important that the exact rotational position of the studded selector cylinder in the coin mechanism and the revolver magazine be correctly aligned. If the tracker bar drops too soon the locking hook for the revolver magazine will ride over the top of the locking wheel for some distance and then suddenly drop into a slot. This drops the tracker bar frame with a bang. When the timing is adjusted properly the tracker bar frame will begin to descend just prior to the point whereby the locking hook is approaching the leading edge of a slot in the locking wheel, allowing the tracker bar frame to be lowered onto the music roll in one smooth flowing motion—without any adverse mechanical jarring and cringe-inducing noise.

The Music Roll Selection Control Sub-System

The music roll selection controls to the right of the revolver magazine are less complicated looking than the Rewind/Play camshaft system on the other side. There are only two separate but interrelated major component assemblies that share the workload of getting the selected music roll correctly positioned under the tracker bar. First there is the simple eight tooth clutch mechanism that drives a round leather belt pulley, for which the belt then wraps around a much larger metal pulley affixed to the main axis of the revolver magazine.

• Tooth Clutch: A mechanism for axially engaging or disengaging a drive shaft from another rotating part while the clutch is either rotating or still. In National pianos, an eight-tooth clutch is used, with each tooth tapered to a fine point, so as to minimize and hopefully eliminate the possibility of the clutch hanging on a tooth end point during engagement of the two halves of the clutch.

Operation of the so-called tooth clutch is simple: When the tracker bar frame is raised up (while the Rewind/Play camshaft is in the Rewind/Roll Changing mode position) the tooth clutch is engaged, causing the revolver magazine to rotate whenever the electric motor is powered on. Conversely, when the tracker bar frame is lowered (when the player system is in the Play mode) the tooth clutch is disengaged, during which time the revolver magazine is then rotationally locked in place.

There are two halves to the machined brass tooth clutch: First is the loose or driven one-half of the clutch, which is firmly attached to the belt pulley. The belt pulley can spin freely on the clutch shaft, unless the tooth clutch halves are pushed together, whereupon the clutch is then considered to have been engaged. The other half of the tooth clutch is cut into one end of a brass sleeve (the shifter assembly) that can slip back and forth on the clutch shaft but that is keyed to the shaft, so that it must always rotate along with it. The position of the movable shifter sleeve is determined by a yoke that is attached to a lever that is, in turn, connected by linkage to the tracker bar framework. When the tracker bar frame is raised the tooth clutch is engaged, and when the tracker bar is lowered the jaw is disengaged.

For certain pianos, and possibly all of them with a Simplex stack, the tooth clutch shaft is supported on each end by cast iron bearing posts, with another bearing post to support the lever and yoke that engages and disengages the clutch. And there is yet another (or forth) adjacent cast iron bearing post, this one used to support the near end of the long shaft that synchronizes the revolver magazine with the studded selector cylinder located inside the coin switch enclosure. However, for most pianos associated with National built stacks, a single die-casting takes the place of the four separate cast iron bearing posts, one casting supporting (1) the clutch shaft (and everything on it), (2) the activating lever and yoke for working the tooth clutch shifter sleeve, and (3) the bearing support for one end of the shaft used to synchronize the revolver magazine with the studded selector cylinder within the coin switch enclosure. This multi-purpose die-casting is mounted directly on the wooden shelf that makes up the “keybed” area. No underlying spacer or metal mounting plate is necessary in this instance, because any critical alignment necessary between the various working parts are suitably maintained by the single die-casting.

The second crucial component in the roll positioning process is the studded selector cylinder located in the coin switch enclosure, which must be precisely synchronized with the rotational position of the revolver magazine, something accomplished by means of a non-slip chain and sprocket arrangement. Tune selection for a National piano is accomplished by dropping a coin into one of eight different coin slots, one for each of the eight music rolls mounted on the revolver magazine. Correspondingly, there are eight rugged studs (which are adjustable in height) mounted on the studded cylinder, with each stud located on a different track spread across the length of the cylinder, and each track assigned to a particular music roll. As such, each stud represents the exact rotational position the revolver magazine must attain for the tracker bar to be perfectly lowered over the selected music roll. So, when a tripped cam finger for a selected music roll contacts the corresponding stud, and is pushed downward, the Rewind/Play camshaft clutch is engaged and the player system immediately goes into Play mode. In contrast, when the end of the music roll is reached, and the rewind pallet valve drops into the rewind cutout, the player system switches from Play mode to the Rewind/Roll Changing mode, raising the tracker bar and engaging the tooth clutch. When this occurs the revolver magazine will begin rotating, and, if no other tunes have been selected, it will immediately begin slowing down to a stop, following the electric motor which has been powered off and stopping. Now the player system is set for the next coin drop, with the tooth clutch engaged and ready to continue rotating the revolver magazine while seeking to position the newly selected music roll.

The Revolver Magazine

The revolver magazine is, of and by itself, a heavy piece of complicated looking machinery housed and anchored in its own centrally located cabinet space down below the “keybed” area of the piano. It is completely independent of all other player system mechanisms, except for three simple interconnections: (1) The leather belt that rotates the revolver magazine, (2) the metal rod that pushes down on the locking hook that locks the revolver magazine into position when the Play mode is initiated, and (3) the little spur gear that meshes with the large diameter gear on a take-up spool. Curiously, it is possibly the only contemporary coin piano whereby the roll mechanism can be disconnected from a mechanical power source and still happily carry out rewinding the music roll unattended.

However, before going further, it need be mentioned that the revolver magazines normally encountered can usually be divided into two easily recognized sub-categories (although minor variations using some evolutionary combination of both categories probably exist): However, confusing the issue, two other variations of significance have recently been reported, one of them having a spring-barrel with a fan type governor for the rewind system, but no solid details are currently available. What follows accounts for those rewind system variations for which photographic evidence and verifiable details are available.

• Those early revolver magazines that have and/or make use of (1) spindly looking cast iron side-plates with straight spokes, (2) a spring-barrel (reminiscent of a wind-up spring motor in a music box or phonograph) to power the rewind process, (3) a single feed spool follower that brakes the take-up spool, and (4) a wooden belt pulley; or,
• Those later revolver magazines that have and/or make use of (1) rugged looking cast iron side-plates with curved spokes, (2) a long, exposed, coiled torsion spring to power the rewind process, (3) two spool followers to separately brake the feed and take-spools at the end of the rewind cycle, (4) a 12" outside diameter sheet metal belt pulley.

The brass spring-barrel motor type rewind system detailed here is comparatively quite rare, and is thought to be the original or earliest spring-barrel version employed—as well as it being a relatively short-lived design. Patent No. 1,070,698, filed May 4, 1910, describes this early mechanism in great detail. Very notably, the feed spool is not removable, nor is the take-up spool, unless the revolver magazine is purposely disassembled. As such, this design suffers by making it rather challenging to change music rolls. In most coin pianos changing a music roll was a quick and simple thing to do, but not so for this particular type of National revolver magazine. The old music roll had to first be manually unspooled from the feed spool, and then the tabbed tail end of a new music roll had to be inserted into a thin slot on the feed spool, and finally the new roll had to be hand spooled onto the then empty feed spool. This meant that National music rolls had to be double ended, i.e., each end of the music roll had its own leader paper with a tapered end and a centered paper tab. Fortunately, when the feed spool is in its fully rewound and stopped position, the task of putting on new music is ameliorated by pushing the feed spool to the left, which disengages the feed spool from the intermediary gearing, allowing the feed spool to rotate freely. Moreover, by placing a flat shim (with a slot to accommodate the feed spool shaft) of sufficient thickness between the feed spool gear and the carrier side plate, the shim can mechanically hold the feed spool in the left disengaged position, freeing the hands to manually spin the feed spool to load the new music roll. This would make it relatively easy to unspool an old music roll and then spool up the new one with much less time and effort.

But there is more to the spring-barrel story. There are several known examples of a more advanced version of the brass spring-barrel type of rewind system, which does incorporate the use of interchangeable feed spools, greatly facilitating the exchange of music rolls. And it makes use of the same type of left side adjustable spring-loaded feed spool chuck as is found in the late revolver magazines with the common coiled torsion spring rewind system. This single improvement—the ability to easily remove and exchange feed spools—probably made roll changing duties much quicker, and no longer a dreaded hassle.

The spring-barrel for both the early or later intermediate variation is similar, but the early version uses a brass spring-barrel, while the improved version makes use of a steel spring-barrel. For both, the actual coiled flat spring is safely housed in the geared barrel, which is mounted with its axis of rotation in line with that of the music roll feed spool, although the two are only indirectly connected by means of intermediary gears. The coarsely square threaded shaft in the early style rewind system is gone. In its place is a finely threaded shaft with a die-cast feed spool follower that has bifurcated extensions on either side, these two extensions fitting over and guided along by a pair of rods located on opposite sides of the follower. Both guide rods have stiff compression springs aligned with each other and located to the right of center, which—when the feed spool follower bumps up against these springs—act in tandem to make up a rewind stop position point. Maintaining paper tension during rewind is the job of a fan type governor geared by means of a small pinion to the large drive gear affixed to the take-up spool shaft. The take-up spool also has a rewind stop point, which is tracked by means of a Geneva drive arrangement.

The late coiled torsion spring rewind system is the type most collectors are familiar with, and very likely the only National rewind methodology they have ever experienced—or heard about. Patent No. 1,647,112, detailing the mechanics of the torsion spring system, was filed on October 3, 1923. This suggests the possibility that National pianos built using the torsion spring rewind system were likely made during or later than 1923. Ironically, the majority of surviving National pianos recently surveyed favor what may well be a misleading illusion, one that implies that those pianos utilizing the late torsion spring rewind system are the most likely to have survived, but, yet, those same late model pianos represent only a small minority of the total number of National pianos manufactured since the company’s inception, most of which probably used some variation of the spring-barrel rewind system. Although there is far from enough statistical information to draw any definite conclusions, it does seem as though the survival rate for pianos built circa 1923 and later may be significantly greater than for earlier pianos made between 1909 all the way up to 1923, and that featured a spring-barrel-powered rewind system.

The great advantage over the earliest spring-barrel rewind system is that with the torsion spring method the music roll feed spool is interchangeable, and can be easily removed, and then replaced by a new already spooled music roll. This was probably a wonderful boon for route operators fed up with the cumbersome toil necessary for the advertised weekly change of music. The downside of the newer rewind system, if any, was that it was more complicated, and some might say temperamental. Both the feed spool and the take-up spool had their own independent spool follower and braking mechanism, which needed to be reasonably well synchronized to simultaneously play their part in terminating the rewind phase. Moreover, the coiled torsion springs did not always age well (compared to the more durable flat springs used for the spring-barrel mechanism), eventually losing enough strength to cause incomplete rewind of music rolls. It is common practice today to replace the torsion springs when restoring a National piano.

The revolver magazine is characterized by the eight independent music roll stations (referred to as carriers in early National Patents) that are equally spaced around the outer circumference of its two side plates. Each of the individual stations is self-contained, with its own music roll feed spool, a take-up spool, and a spring-motor based rewind mechanism equipped with the necessary gearing and braking to safely rewind a music roll, and do so in such a way that the rewound roll is ready to be played over and over again. When the player system is switched into the Play mode, and the tracker bar is lowered onto the music roll, a small spur gear meshes with the large diameter gear on the take-up spool, pulling the music roll paper forward under the tracker bar. As the paper advances from the music roll onto the take-up spool, the rewind system spring-motor geared to the music roll feed spool is slowly wound ever tighter. It is this spring-motor that will be used to power the roll rewinding phase. Then, at the end of the music roll, when the tracker bar is raised and the spur gear meshing with the take-up spool gear disengages, the music roll is suddenly free to instantly begin its speedy rewind phase, eventually stopping short and leaving a few layers of music roll leader wrapped around the take-up spool, ready for the roll to be played once again.

Loading and Unloading Music Rolls

When the National service agent arrived to take care of the weekly change of music he would have been carrying something akin to a suitcase. Inside would have been a small assortment of commonly needed spare parts, some hand tools, a selection of wound up new music rolls (without spools), and a hand-cranked apparatus used to quickly unspool the old music from a removable feed spool, and then wind new music back onto the emptied spool. Regrettably, collectors today do not have the luxury of such handy luggage, and must find a source of now rare music rolls and then learn to innovate while dealing with whatever variation of revolver magazine might confront them.

There are at least two very distinct types of spring-motor powered rewind systems employed by National, and there may be more yet to be discovered. Nonetheless, of the two documented methodologies—the early spring-barrel powered and the late exposed torsion spring powered types—both have been extensively studied and the principles of operation are well understood. However, each rewind system is quite different when it comes to exchanging music rolls, replacing old music for new tunes. And things are complicated by the fact that for the spring-barrel powered rewind systems the earliest design had a non-removable feed spool, whereas the improved design allowed the feed spools to be easily removed and exchanged for a relatively easy roll replacement experience. The differences and a suggested method for dealing with the various revolver magazine configurations are detailed in the following paragraphs:

Early Spring-Barrel Powered Rewind System
(with Non-Removable Feed Spool)

Each of the eight roll stations is equipped with both a music roll feed spool and a take-up spool, both of which are non-removable. Thus, music rolls cannot be easily exchanged in the quick and easy fashion normal for most rewind type coin pianos, which is to lift out the old roll with its own individual spool, and merely pop in a new already spooled roll. And so how does someone change music rolls when the spools are permanently fixed and cannot be removed short of disassembling the revolver magazine? There are a couple of ways that come to mind, neither as ominous as might be imagined, with the former requiring no special tools or equipment, while the latter needs one simple tool that almost anyone could rather easily fashion.

The suggested first method comes from Dick Hack, a professional restorer of things relating to mechanical music. His idea is presented below in eight simple steps:

1. Manually unspool any leader paper wrapped around the take-up spool, which is can be done by rotating the spool. This will require overcoming any braking force applied by the rewind stop brake and the paper tensioning brake.
2. Manually pull the old roll paper off the feed spool, and when the end of the roll is reached grab a flange on the feed spool to keep it from rotating.
3. Pull the roll’s tail end leader tab free of the narrow slot in the feed spool.
4. While continuing to hold onto the feed spool flange, insert the tail end leader tab of the new music roll into the narrow slot in the feed spool.
5. Now let the wound tight spring motor do its job by allowing it to “rewind” the new music roll onto the feed spool, all the while keeping your thumb pressed lightly against a spool flange to meter the rate at which the new music roll is wound onto the spool.
6. Once the new roll is rewound and the spool follower prevents any further “rewind” action (because it is at its stop position), push the feed spool to the left to disengage the rewind gears and then freely rotate it one way or the other, so that there is enough leader paper exposed to wrap around the take-up spool several times.
7. Push the new music roll paper leader tab into the narrow slot in the take-up spool.
8. Then manually rotate the take-up spool until the leader paper is snugly wrapped onto the spool. This will require overcoming any braking force applied by the rewind stop brake and the paper tensioning brake.

Once this technique for changing music rolls is commenced is it critical to understand that the feed spool must never be let go to run free and “rewind” itself with no roll paper attached to the feed spool. Failure to heed this requirement will definitely complicate the roll changing process, and will require the new music roll to be tediously wound onto the feed spool by hand.

The other method requires the use of a simple tool, consisting of a flat spacer (or shim) a few inches long and with a short slot in one end. This spacing shim needs to be wide enough (3/4” overall width should be more than sufficient) so that the slot can slide over the right carrier plate bearing pin on which the feed spool rides. The thickness of the shim need be sufficient so that when it is slipped between the carrier side plate and the little gear on the feed spool, the spool will be shifted to the left enough to keep the rewind gear disengaged. And so, with your spacer shim tool in hand, follow these ten steps:

1. Nudge the feed spool to the left and slip the spacing shim between the right-side carrier plate and the small gear on the end of the feed spool. This will keep the rewind gear disengaged.
2. Manually unspool any leader paper wrapped around the take-up spool. This will require overcoming any braking force applied by the rewind stop brake and the paper tensioning brake.
3. Manually pull the old roll paper off the feed spool. With the rewind gearing disengaged this will be easy, and as the paper comes off the feed spool it can be wound onto some kind of core for storage and future use.
4. Detach the tail end leader tab for the old roll free of the narrow slot in the feed spool.
5. Insert the tail end leader tab for the new music roll into the narrow slot in the feed spool.
6. Manually “rewind” the new music roll onto the feed spool. By brushing the fingertips quickly across the rim of a feed spool flange the spool can be spun up and the roll loaded onto it without much effort.
7. Once the new roll is fully wound onto the feed spool, rotate the feed spool one way or the other so that there is enough leader paper exposed to wrap around the take-up spool several times.
8. Push the paper leader tab for the new music roll into the narrow slot in the take-up spool, and then manually turn the take-up spool until the leader paper is snugly wrapped around it.
9. Manually turn the take-up spool until the leader paper is snugly wrapped around it. This will require overcoming any braking force applied by the rewind stop brake and the paper tensioning brake.
10. Remove the spacer shim tool, so that the rewind gears reengage, and you should be ready to play new music.

While the above roll replacement methods are perhaps not convenient, they are not so arduous as to be avoided. National actively encouraged route operators to simply destroy old music rolls, and only keep fresh new music on hand, no doubt to maximize the profitable flow of nickels. As such, the built-in inconvenience of this early spring-barrel rewind system, with non-removable feed spools, may have aided in the process of getting rid of old music, the route operator ripping the old music roll off the non-removable feed spool in a fit of unbridled frustration. Whatever the case, the company policy of destroying used rolls is surly a contributing factor for the rarity of early National music rolls.

Improved Spring-Barrel Powered Rewind System
(with Removable Feed Spool)

Each of the eight roll stations is equipped with both a music roll feed spool and a take-up spool, but in this improved design the feed spool is removable. This single improvement greatly simplifies the music roll replacement experience, turning a dreaded hassle into a more or less pleasant routine. To proceed, assuming that the music roll has fully rewound and the feed spool follower cam is in its stop position, follow these five simple steps:

1. Manually unspool any leader paper wrapped around the take-up spool. This will require overcoming any braking force applied by the rewind brake and the paper tensioning brake.
2. Manually remove the feed spool and replace it with a spool containing the new music.
3. Leave loose enough music roll leader paper to wrap around the take-up spool several times.
4. Insert the roll’s front leader tab into the narrow slot in the take-up spool.
5. Manually rotate the take-up spool until the loose leader paper is snugly wrapped around it. This will require overcoming any braking force applied by the rewind brake and the paper tensioning brake.

Late Exposed Torsion Spring Rewind System

It is important to understand that for each of the eight music roll stations on the revolver magazine there are two separate spool followers, one for the feed spool and yet another for the take-up spool, each mechanically independent of the other. The sole purpose of a spool follower is to keep track of (from a predetermined starting point) the number of spool rotations made while the music plays, and then, once rewind is initiated, to count backwards and precisely decelerate and then stop the movement of its associated spool the moment it reaches the starting point. Thus, when the rewind process is halted, there should be several wraps of music roll leader paper still safely wound around the take-up spool, with only roll leader paper exposed between the feed and take-up spools—all perforated paper should be fully wound back onto the music roll feed spool.

If a music roll was correctly loaded onto the revolver magazine (with both spools synchronized with their respective spool followers, when it is eventually removed the two spool followers will each be similarly seated at their own stop position. What is a stop position? This is the exact position on a threaded rod (geared to a spool’s axis) where a spool follower normally brakes the spool, stopping the rewind process. This occurs when the spool follower bumps up against a stiff compression spring that serves as a stop bumper. During play, the feed spool follower essentially counts the number of rotations of the music roll feed spool, beginning from the stop bumper starting point. Likewise, the take-up spool follower also counts the number of rotations of the take-up spool, but using a 4-to-1 gear ratio, which means that the take-up spool follower moves along its threaded shaft much more quickly than does the slower feed spool follower. This increased speed basically amplifies the accuracy and reaction time during rewind when the take-up spool follower comes up against the stop bumper spring.

When a music roll is finished playing and the tracker bar frame is lifted, the music roll will immediately commence rewinding, with the paper accumulating on the music roll spool tugging against a light braking action provided by a friction pad riding on the left side hub of the take-up spool. This light tensioning of the paper will prevent the paper from looping and help keep it snugly wound on the music roll feed spool. Now it is the job of the two spool followers to stop the two spools from spinning precisely when the roll paper reaches the stop bumper starting point. And, as should be obvious, it is imperative that the two spool followers be synchronized properly so that they, more or less, brake the feed and take-up spools simultaneously. However, being that the take-up spool follower moves some four times more quickly it will tend to take precedence and apply a more accurately timed and greater stopping force at the end of the rewind process, which suggests that accurately synchronizing the music roll with the take-up spool follower is probably a bit more important that it is with the feed spool follower.

Release Lever for the Take-up Spool Follower

Whenever you remove or insert a music roll into the revolver magazine you will need to disengage the gearing for the take-up spool follower. But where is the so-called release lever for the take-up spool follower? Facing the revolver magazine, on the outer left-hand side of each roll station is a long brass lever toward the back of the station. It is parallel with the metal side plate. This lever pivots toward the lower end and extends upward to rest against a shaft that runs behind the take-up spool. When the upper end of the lever is pushed to the right it disengages the spur gears (located on the opposite side of the revolver magazine) for the take-up spool follower. This allows the take-up spool to be rotated freely one way or the other without affecting the position of the take-up spool follower.

Removing Music Rolls

To remove a music roll, follow these four simple steps:

1. Push (toward the right) and hold the release lever for the take-up spool follower.
2. Unspool any leader paper still wrapped around the take-up spool.
3. Let go of the release lever for the take-up spool follower.
4. Without rotating the feed spool, grasp the spooled music roll, and gently push it toward the left. When it disengages from the drive chuck (on the right-hand side) swing the roll away from the revolver magazine and then let it fall free of the pinned left side chuck.

Synchronizing the Spool Followers

Once the music roll is removed the spool follower for the feed spool will be automatically and correctly set against its stop bumper. This possible reset will occur because of the torsion spring that powers the feed spool during the rewind process, which will try to “rewind” the drive chuck to its full stop position once the music roll has been removed. However, be aware that the take-up spool will not try to “reset” itself against its stop bumper. Consequently, it may be prudent to make certain that the take-up spool follower is properly seated against its stop bumper. To do this, turn the take-up spool gently in the rewind direction until a braking force can be felt. Keep in mind that there is a friction pad that rides on the off-side take-up spool hub, whose purpose is to apply a slight but constant tension on the roll paper during the rewind process to keep the paper snuggly wound onto the feed spool. Thus, when turning the take-up spool you will need to be aware of this light but constant braking effect, and be watchful of an abrupt increase in braking force due to the take-up spool follower coming up against its stop bumper.

Loading Music Rolls onto the Revolver Magazine

To load a new music roll, follow these six simple steps:

1. Unspool enough leader paper from the new music roll to wrap around the take-up spool several times.
2. Place the cupped (left) end of the feed spool onto the spring loaded pinned chuck on the left side of the roll station. Push the spooled music roll to the left and then engage the drive chuck on the right side. You may need to hold the drive chuck stationary while gently rotating the music roll, so that the drive chuck fully seats into the feed spool flange.
3. Push (toward the right) and hold the release lever for the take-up spool follower.
4. Rotate the take-up spool so that the narrow slot for the tapered roll leader is facing forward, and then insert the roll leader into the slot.
5. Spool the loose leader paper onto the take-up spool (without turning the feed spool).
6. Let go of the release lever for the take-up spool follower.

During normal operation, there should always be several wraps of leader paper remaining on the take-up spool when the rewind phase terminates. The excess leader paper helps to compensate for any spool follower tracking inaccuracies and/or rewind overshoot due to momentum. However, if no or too little paper is wrapped around the take-up spool (when inserting a new music roll) the rewind process will likely decelerate too late, probably about the time the roll tears loose from the take-up spool. The goal here is to end up with multiple wraps of leader paper still on the take-up spool when the rewind brakes are applied, which is when both spools come to their “at rest” stopped position.

Vacuum Pumps

There are two fundamentally different types of vacuum pump that has been observed in National automatic pianos—the reciprocating type and the rotary type. Because the sampling at this point is still statistically almost insignificant, it is unknown whether another type of pump might have been used. Complicating our studies, photographs of extant instruments usually omit showing the interior of the pump or motor compartments. Moreover, the majority of the serial numbers recorded in the database study are in the 7,000 range, with a but a slim few in the 3-digit range, leaving us to wonder about the construction characteristics of the earliest National automatic pianos, circa 1909 and how they might have evolved onward toward the 1920s.

One of the oddities of all known National automatic pianos, unlike most other coin piano brands, is the total lack of a vacuum reservoir, which normally would also include a spill valve to regulate and keep constant the vacuum level throughout the entire player system. (Nelson-Wiggen was the only other major brand that normally had no vacuum reservoir.) The vacuum pump in National coin pianos is connected directly to the stack, which, in turn, provides a vacuum source for a vacuum regulation valve, which must not only maintain a fairly constant vacuum level, but that additionally controls the piano expression, the exact setup depending upon the type of stack installed.

Early Wood Frame Reciprocating Pump

What is believed to be the earliest National vacuum pump observed to date is what we have arbitrarily termed a wood frame reciprocating pump. This is because it has a four-throw crankshaft, with individual connecting rods that attach to bearing cleats fastened to each of four bellows situated on the pump framework below the crankshaft, a layout reminiscent of how a multi-cylinder reciprocating automobile engine might be constructed.

• Reciprocating: To move back and forth alternately (such as a crankshaft driven pump that has bellows that move up and down in a periodic or cyclic fashion).

The pump consists of a strong wooden rectangular framework with four vacuum bellows mounted on the bottom floor of the framework. There are two pairs of bellows arranged side-by-side, with the two pairs facing each other. On the outer edge of each movable bellows board there is a die-cast cleat, to which a die-cast connecting rod is connected by means of a bearing shaft going though the flanges of the cleat. The uppermost horizontal part of the framework supports a 4-throw crankshaft, and to connect each of the four bellows (located on the bottom floor of the framework) there are individual die-cast connecting rods with a matching bearing cap for the crankshaft end. This type of pump is usually associated with a sheet metal drive pulley on its rear-facing side, and the usual metal chain sprocket on its front-facing side.

Two distinct sheet metal pulley variations have been noted, one with circular reinforcement around the hub and the more commonly observed type with ribbed reinforcement radiating outward between the hub and the outer rim. Wooden drive pulleys have also been noted on the reciprocating type pumps, but they may or may not have been original to the pump. The reciprocating pump also made use of a belt idler to help keep the belt tension constant, and that was fastened to the outside of the pump framework near the bottom floor. Overall, the reciprocating type pump is well-built, but probably more difficult to maintain and repair than would be what is thought to be a later and relatively simple rotary style unit with a single crankshaft throw and bearing.

Late Aluminum Frame Rotary Box Pump

What is thought to be the latest of the National pump types is what we have termed the aluminum frame rotary box pump. It basically utilizes the same box-frame and central single-throw crank idea that many other coin piano manufacturers used throughout the 1920s. It is an efficient design and is easy to maintain in good repair.

• Rotary: A device or having an important part that revolves or rotates around an axis (such as a pump with a centrally located single-throw crank connected to individual bellows clustered around the crankshaft axis).

A slightly modified version of the National rotary pump used in its coin pianos—one that had a wooden platform mounted on its top to accommodate an electric motor—was used as the vacuum source for the National built Welte pipe organ roll changer mechanisms. These tall and elaborate roll changers had 10 music roll stations attached to a long endless chain that wrapped around hefty chain sprockets located at the bottom and top of a steel framework. These impressively constructed pipe organ roll changers were housed in a special cabinet, or built into a convenient closet.

A similar rotary pump, with some relatively minor modifications to the basic cast aluminum framework design, was later used in Auto-typist roll-operated duplicating typewriter machines. The Auto-typist was reportedly introduced in 1927, and commercially manufactured in the 1930s by the M. Schultz Piano Company, and then later by the American Automatic Typewriter Company, both Chicago, Illinois, based enterprises. The Auto-typist basically consisted of two machines. A perforator with a hole spacing of 5 per inch was used to punch an 11-1/4” wide paper roll. The paper roll was then transferred to and deciphered by another machine that was housed in a special typewriter desk, which controlled the keys of a standard typewriter through a system of vacuum powered control valves, pneumatic bellows, and a maze of interconnecting tubing alongside various other mechanical linkages. It was used by letter duplicating services, which created mass-produced “hand-typed” letters for businesses that wanted their mailings to appear to be personalized. The most elaborate models had a console with a selector device, based on the Western Electric “Selectra” patent (which Otto Schulz purchased at Western Electric Piano Company’s bankruptcy sale). The master roll included a number of boilerplate paragraphs; the “Selectra” fast-forward and rewind were used for printing these paragraphs in a specified sequence. Larger setups had a master console containing the roll and selector mechanisms, connected to a group of typewriters, each having a pneumatic stack that pulled its keys down.

The National rotary pump, however, is visually quite different than the rotary type pumps used in most American built coin pianos. It makes use of a hollow aluminum framework (instead of the common wood framework), which is both structural and it channels the vacuum generated by each of the four bellows. Small mounting feet protrude from each side of the aluminum casting at its bottom edge. Attached to the rugged exterior framework, at its vertical mid-point on both the front and back sides, is a horizontal aluminum cross member with bronze bearings to support the single throw crankshaft. The cast metal housing surrounding the crank bearing has four flanges protruding outward from it. To each of these flanges is attached (by means of two heavy duty machine screws) a length of rubberized canvas belting, which is, in turn, fastened by wood screws to a wood block mounted about midway on the movable board for each of the four bellows.

These flexible “connecting straps” eliminate the need for the four die-cast connecting rods with lubricated end bearings, which are necessary with the earlier wood frame reciprocating style pump. They also have the advantage that as long as the very durable rubberized canvas material remains flexible, they will never knock like most wood or metal connecting rods eventually end up doing. Although the rotary style pumps have been seen with a sheet metal drive pulley, the rotary pumps are usually observed with a wooden drive pulley on its rear-facing side, and the normal metal chain sprocket on its forward-facing side. It is likely that the rotary pumps used in National pianos are an adaptation of an off-the-shelf pump offered by some unidentified supplier.

Coin Switch and Related Items

The coin switch mechanism could be said to include a compendium of closely related items, all of which need do their part effectively for the coin handling process to function as intended. It is, however, the complicated looking mechanisms housed in the coin switch enclosure that usually grabs most, if not all the attention. Here, however, we will deal with the entire coin handling system, from top to bottom, which can logically be divided into three main interrelated and functionally diverse components. One perhaps ancillary item has been added to the lot, and that is the device that indicates which tune is currently playing. The following enumeration highlights the generalized properties of each logical division:

1. At top is an intricately cast metal faceplate with raised lettering, eight separate coin slots, push buttons for each slot, a coin return for rejected coins, and a window for displaying tune title strips for each of the eight coin slots. This is where the coin handling and tune selection process is initiated.
2. Next in line is an elaborately constructed and contoured wooden multi-coin chute connector that delivers the coin dropped into one of the eight slots to the correct reception point on the coin switch mechanism below it.
3. In the middle is the most complicated part of the coin handling process, the coin switch mechanism. This intricate device not only switches the electric motor on or off, but, just as importantly, it is also plays an indispensable role in the music roll playing process. Central in the coin mechanism enclosure is a studded selector cylinder that is precisely synchronized with the rotation of the revolver magazine, and that triggers the Play mode cycle when the correct music roll has been positioned under the tracker bar.
4. At the tail end of the coin handling process is the lockable metal coin collection box located in the bottom right-hand side of the piano (above the electric motor).
5. Ancillary, but important nonetheless, is the illuminated tune indicator wheel that is synchronized with the rotation of the revolver magazine, and that displays through a circular aperture a printed number representing what musical selection is currently being played.

Coin Slot Faceplate

The coin slot faceplate is an artfully designed and crafted casting, with a series of clever coin-handling mechanisms mounted firmly on its backside. At the front side top are eight different coin slots, with a push button directly below each of the eight slots. Below the row of buttons is the cast inscription: “Play Nickel in One or More Slots; Push Button Each Nickel.” Further below is a display window that lists, by number, each of the eight tunes available for play. Above the tune display window is a simple key-lock used to gain access to the tune title strips inside the display window.

Each coin slot, numbered 1 through 8, correspond to a specific music roll loaded onto the revolver magazine. The slots are accurately sized to accommodate only the coin denomination to be accepted, usually a U.S. 5-cent nickel. Any coins that are too thick or that have a diameter greater than the required coin size will not fit into the slot, and so are rejected outright without further ado. However, coins inserted that are too small in diameter will simply fall into the slot’s coin trap and be rejected as under-sized, and then further drop unceremoniously into the coin return tray where the customer can easily recover any rejected coins.

When a coin is dropped into a slot, it rolls through a short zig-zag chute and into a slotted die-cast coin trap—there is a separate, independent trap for each coin slot. At the bottom of each slot (in the coin trap die-casting) there is a spring loaded blunt ended lever, which blocks just enough of the slot to catch and prevent a properly sized coin diameter from falling through the slot. However, any coin smaller than the required coin diameter will fall freely through the trap and be guided by a metal chute to ring a small bell on its way into the coin return tray. To dispense a correctly sized coin still held in the coin trap—so that it can fall into and trip the coin switch mechanism—the push-button below the coin slot must be pushed fully inward to its stopping point. When this is done, the button pushes against a lever arm attached to the corresponding coin trap, causing the trap to swing back away from the coin slot faceplate and over the appropriate chute in the wooden multi-slot connector block below it. When the coin trap moves over the connector chute a dog on the blunt lever that retains the coin within the trap bumps up against a fixed metal bar. On contact with the bar, the blunt lever arm keeping the coin in the coin trap swivels downward, thereby exposing the full width of the slot and releasing the coin. It is important to understand that the coin trap must be moved back away from the faceplate casting until it is directly over the appropriate chute in the multi-slot connector block, and reach this position before the coin is released from the coin trap, otherwise the coin would end up in the coin return tray.

The coin trip mechanism does not function like other contemporary combination coin switch and play accumulator mechanisms, such as commonly used in many single-slot setups. Some of these single-slot mechanisms can accumulate up to 20 or so plays, and do this quite independently of what the player system might be doing. But in National coin pianos the coin switch mechanism is highly integrated with the player system, in fact guiding certain crucial operations, like determining when the revolver magazine is correctly positioned and then triggering the player system into the Play mode. These are not functions normally associated with a single-slot coin switch accumulator.

For the National piano coin switch device, only one coin per slot deposited during any one Play mode cycle will be acknowledged, i.e., once the player system is put into Play mode, multiple coins dropped into any one slot will only count as a single coin. This rule of thumb does not, however, perfectly apply when the player system is powered off and the first coin is dropped, causing the player system to be powered up. This is because when the electric motor initially starts up the player system will be in the Rewind/Roll Changing mode, not the Play mode. It is when the system switches to the Play mode that the trip lever for the currently playing tune is reset. This is done when a stud on the studded selector cylinder in the coin switch mechanism engages a tripped cam lever, setting in motion these actions: (1) the player system is triggered into Play mode, and (2) the coin trip lever for the selection being played is reset. Once this reset action for the currently selected tune occurs, another coin can be dropped into the same slot and it will register, allowing the same tune to eventually replay, although if other tunes have also been selected during the current Play mode the system will cycle through and play them in succession.

Multi-Slot Coin Transfer Connector

Made of many carefully cut and fitted together bits and pieces of wood, this multi-slot and gracefully contoured connector delivers coins from the upper and much wider coin slot faceplate unit to the lower and much narrower lineup of coin chutes down in the coin switch mechanism. The transfer connector sits below and to the rear of the row of individual coin traps located on the backside of the coin slot faceplate. It is only when a coin slot push-button is pressed that the corresponding coin trap is moved rearward, so that it is rightly lined up over the waiting connector chute below it, otherwise the coin would miss the opening in the transfer connector. To alleviate any minor alignment issues, the topside wooden chute entry openings are wide and spacious, with the upper ends of the wood separators tapered to a point. This wide-open spaciousness allows the individual wooden chutes enough leeway to reliably catch and collect coins dropped from the coin trap directly above. On the bottom end of the connector, the woodwork fits snugly above the multi-chute die-casting located at the back of the coin switch mechanism.

Coin Switch Mechanism

The National coin switch mechanism, with its eight separate coin slot layout, is unique amongst all of its competitors, all of which utilized a much simpler single coin slot design, which in many cases used standard off-the-shelf components. But not National. The company choose a very different, independent methodology. Nonetheless, both the National way and the somewhat standardized industry wide coin processing methods proved to be durable, and could usually be counted on for many years of mostly trouble-free service. And so, within its enclosure, the National style coin switch mechanism, although perhaps complicated looking, can logically be divided into three functionally unique but interconnected sections. These are enumerated below, going from front to back:

1. At front is the electrical switch and its associated trip and reset linkages.
2. In the middle is the studded selector cylinder that is rotationally synchronized with the revolver magazine.
3. At the very back is the wide die-casting containing eight separate coin chutes, each with its own corresponding coin trip lever mechanism immediately to its front.

What follows next is a brief and simplified description of how the coin switch mechanism works:

When a coin falls through one of the eight coin chutes (located at the back of the coin switch mechanism) it impinges upon a trigger pawl near the bottom of the chute. When this happens a vertically oriented extension on the trigger pawl rotates toward the rear, which then allows the end of a lever arm resting on top of the pawl extension to slip free and drop. The other end of this lever arm is enlarged and rounded where it swivels around its pivot point, Below the pivot point is a tension spring that works to rotate the lever's trip arm downward, while above the pivot point there is a short projection sticking up that is forced in a rearward direction. This projection has two functions. Firstly, it is drilled with a hole to accept a long hooked wire that connects with corresponding linkage that interacts with the studded selector cylinder in the middle of the coin switch enclosure. When this wire is pulled it causes a finger to pop forward toward the studded cylinder, putting it in the path of one of the cylinder studs. Secondly, as the upward sticking projection is rotated to the rear it pushes against a suspended metal bar that is free to swing to and fro (within limits). This bar runs the entire width of the coin switch enclosure, so that any coin chute pawl that has been tripped will cause an affected lever arm to impact this swinging metal bar, pushing it toward the rear of the enclosure. When it is pushed back it comes up against a vertical lever that, in turn, trips the electrical switch at the front of the enclosure, which then makes electrical contact. If this metal bar is kept in the rearward position, such as when multiple tunes have been selected, the coin switch will not open (break the electrical circuit) until all of the coin trip levers have been reset, one by one, as any additionally selected music rolls have been played.

Moving on to the center of the enclosure, is the studded selector cylinder and associated linkages. When one of the wires connected to a coin trip lever pulls the corresponding linkage associated with the studded cylinder, a metal finger is caused to move closer to the studded cylinder (each finger is akin to a vertical lever with a projecting tip or finger at the top end that is facing toward the cylinder). With the finger moved closer to the cylinder surface, when the corresponding stud on the rotating cylinder touches the finger it will be pushed downward, and at some point, after sufficient travel distance, it will be disengaged from the stud. When any finger is pushed downward it both resets the corresponding coin trip lever, and it presses down against a heavy-duty metal bar attached at both ends to a steel shaft, which has a simple lever arm on its front facing right end. When this shaft is rotated (perhaps a quarter revolution) a wire linkage between it and a cast bronze bell-crank (near the front of the coin switch enclosure) twists the small inner shaft traversing the front of the “keybed” area. Momentarily rotating this long inner shaft triggers the Rewind/Play mode clutch to cycle the camshaft into the Play mode position. When multiple tunes have been selected, the corresponding fingers next to the studded selector cylinder will be moved closer, so that when the current music roll trips the rewind pallet, and the player system goes into the Rewind/Roll Changing mode, the studded cylinder will again rotate and a different stud will push down the next successive finger to be encountered. This process will repeat until all the selected tunes have been played through and any associated coin trip levers have been reset, whereupon the electric switch contacts will finally be pulled open, shutting down the player system.

Lastly, at the front of the coin switch enclosure, is the insulated electrical switch block, and in front of it the operational linkages that control the knife switch action. This insulated block uses copper strips to connect the electrical wiring terminals to the two fixed switch blade contacts. The knife blade itself is electrically neutral, and is connected to but electrically insulated from the steel lever arm that raises and lowers the switch blade. When the player system is off the knife switch blade is raised, breaking the electrical circuit, and it is kept latched in that upward raised position until specifically released. When a coin falls through one of the coin chutes at the back end of the coin switch enclosure, and trips a coin trip lever, the vertical oriented latching lever (at the front of the enclosure but connected by a long metal rod to the coin trip mechanism) is pushed forward. This action forces the latching lever to swivel slightly, which releases the latch pin projecting outward from the knife blade lever arm, allowing the electrical contacts to close. Cleverly, the bearing pin, on which the latching lever swivels, is attached to a horizontal lever that can be raised and lowered by a bell-crank, which is clamped onto the larger hollow shaft that traverses the front of the “keybed” area, and that is further attached to the cam follower lever for the “coin switch reset cam.” When the vertical latching lever is raised, by the action of the reset cam, the latch pin on the switch blade arm can re-engage. Then, when the vertical latching lever is lowered, it pushes down on the latch pin located on the forward end of the knife switch arm, in turn swiveling the knife switch on the opposite end of the lever arm into the open position, breaking the electrical circuit. However, when several tunes have been selected the rod controlled by the coin chute trip levers will keep the vertical latching lever pushed forward. This will prevent the latch pin on the knife blade lever arm from re-engaging the vertical latching lever, until all selected tunes have been successively played and their associated coin trip levers have been reset, whereupon the player system will then be shut down.

Coin Collection Box

The coin collection box is located underneath the "keybed" area shelf and on the right-hand side of the piano case. It and the electric motor share the same compartment. Directly above the black painted coin box is a sheet metal coin collection chute that funnels any coins passing through the coin switch mechanism into the large and lockable steel coin collector box. Inside the top of the metal box are a pair of brackets that are used to hold a stiffened hoop sewn into the top edge of a suspended cloth bag used for the collection of coins, and that can be quickly pulled out and replaced with an empty bag. This makes transporting the day’s receipts both quick and easy.

Retrieving the coin piano receipts may seem to be a trivial occurrence, but I am reminded of a story told to me in the late 1960s by the late Jim DeRoin. Jim had some number of coin pianos out on location in the San Francisco and Oakland areas. One day he was down on his hands and knees collecting coins from the bottom of a piano when he felt a strange tickling sensation across his throat. Puzzled, he looked up only to see a man standing over him with a switchblade knife, having just attempted to slit his throat, but not cutting deep enough to do major damage. Jim, a strong and burly fellow by nature, grabbed a piano stool next to him and broke it over the guy’s head, which, needless to say, aborted the attempted coin piano robbery.

Tune Indicator Wheel

All known National coin pianos have, or at one time had, a tune indicator wheel attached to and located in the upper left-hand side of the piano case. The rotational position of the tune indicator wheel must be perfectly synchronized with the angular position of the revolver magazine, so that the tune number being displayed always exactly coincides with the music roll positioned under the tracker bar. To accomplish this end, a chain around the sprocket on the backside of the tune wheel connects with a similar sprocket underneath the “keybed” area shelf. This lower sprocket is fixed to one end of a countershaft that extends over to the left side of the revolver magazine, where another length of chain and set of chain sprockets connect the countershaft to the revolver magazine. Because all the linkages are chain, there is no slippage resulting in misalignment between the tune wheel and the revolver magazine. However, although functionally identical, two different types of underlying support construction have been observed, as follows:

• Cast iron support structure: All the working components are held in place by a beautifully contoured iron casting, which requires careful machining of various surfaces. When a Simplex stack is installed, it has been observed that the chain-fall from the tune indicator wheel drive sprocket drops down to an intermediate countershaft, which basically moves the chain-fall forward toward the front of the stack, where it then drops down through a cutout hole in the top cover of the Simplex stack. If a National built stack is installed, however, a pair of idler sprockets divert the chain from the tune indicator wheel around the left side of the stack. Whether, or not, there are other chain routing configurations is currently unknown, but no matter what the upper chain routing, the chain sooner or later continues downward through a cutout in the “keybed” area shelf, where is goes still further down to the interconnecting countershaft between the tune wheel and the revolver magazine.
• Steel framework: The working components are held in place by an inexpensive fabricated steel framework. The chain-fall from the drive sprocket is diverted to the left around the National built stack by idler pulleys and then drops straight down through a cutout hole in the “keybed” area shelf, going still further down to the interconnecting countershaft between the tune wheel and the revolver magazine.

The underlying support structure provides (1) a bearing surface on which the numbered tune indicator wheel can revolve, and (2) a mounting platform for a small pneumatic that raises and lowers a metal shield to cover over and conceal the otherwise exposed tune number whenever the player system is powered off. The tune indicator wheel itself is nothing more than a thin sheet metal disk with eight large equidistant holes punched a little inside its outer circumference. In the center of the disk is a small metal hub, to which is attached a short shaft on which the tune indicator wheel rotates, and at its outer end is a small chain sprocket used to synchronize the tune wheel to the revolver magazine. The front surface of the metal disk is painted black. On the backside of this disk is a pre-cut paper circle, with large numbers (1 through 8) printed inside its circumference, which has been pasted onto the backside of the tune wheel. From the front side, the circle of individually printed large numbers show through the large diameter holes punched in the metal disk.

The selected tune number is made all the more apparent by an electric light hanging near the backside of the tune indicator wheel, which illuminates the printed numbers from behind. This electric lamp is located a little above the tune indicator wheel, but yet effectively spaced between it and the piano action, which, when the player system is powered on, illuminates the entire upper interior of the piano. When the case front is in place the tune indicator device is hidden, except for what can be seen through a circular aperture in the case front, which is precisely positioned to be in front of the number representing the currently selected tune. When the player system is powered off a metal shield drops down and covers the otherwise displayed number, but when the player system is powered on the electric light behind the tune indicator wheel is brightly lit and the protective metal shield is lifted out of the way by a small pneumatic connected into the stack vacuum supply. This action allows the number for the currently playing tune to shine brightly through the case opening, announcing to all the number of the currently playing musical selection.

The Electric Motor

For National coin pianos, the electric motor of choice seems to have been a Holtzer-Cabot motor. Thus, having the motor nameplate information can be very helpful in sometimes approximating the manufacturing date of National automatic pianos. This is because original Holtzer-Cabot factory information survives that makes it relatively easy to determine what year the motor was manufactured. Of course, there will be some delay between the manufacturing date and when the motor was eventually installed in a National piano, but so far using the motor to estimate the piano manufactured or shipping date is about the only tool currently on hand that has a reasonable chance of being useful. This situation may change as more pianos are added to the National coin piano database, perhaps providing more clues once the mechanical evolution of the pianos can be more accurately studied and compared.

It should be understood that using the motor as a piano dating guide has its shortcomings. For instance, a troubled motor in a piano bringing in money from a lucrative location might have hastily been swapped out with another motor, perhaps one in an idle piano sitting in storage, or maybe some off brand found in a local store or warehouse was used. Moreover, when a piano was factory refurbished and/or retrofitted with an Automatic Whippet Race diorama, or a “dancing figure” trade stimulator, a fresh new motor might have been installed. Or maybe a “new” motor was temporarily “lost” in a secluded nook in a distributor’s storage warehouse, and then when eventually discovered this “new/old” motor was sold and later inserted into a new National piano, which would cause the motor date to be much earlier than the piano shipping date. Anyone with a good imagination could come up with a myriad of other scenarios that would upset any scheme using a motor to determine when a piano might have been factory made or shipped.

To determine the approximate manufacturing date for a Holtzer-Cabot motor please refer to the Dating Holtzer-Cabot Motors section on the Electric Motors in Coin Pianos page.

But the motor dating situation gets worse. Toward the very end of the 1920s and up into the early 1930s there is evidence that suggests the company may have begun replacing old and troubled Holtzer-Cabot motors with Emerson motors. There are several instances of this popular brand of electric motor having been observed in late refurbishments, where the case had been refinished in green with decorative pinstriping and an Automatic Whippet Race had been installed. As of this writing there is no known simple serial number to date conversion table for Emerson motors. Nor is there any way to know one way or another whether the company was turning away from Holtzer-Cabot, and favoring the Emerson motor.

Late Retrofit Automatic Whippet Race Diorama

The Automatic Whippet Race (alternately known, in lieu of its true name, as a Dog Race Diorama) was a popular trade stimulator retrofitted by National in many of its pianos in a last heroic but vain attempt to extend the commercial viability of the company’s ever-growing stock of idle coin pianos. The diorama’s mechanisms are both intricate and beautifully constructed, but that effort did little to prolong the coin piano business, which faced ever more aggressive competition from radio, the automatic phonograph, and talking pictures, all entertainment venues finding enormous popularity with the public sector. The hand-writing was clearly on the wall, and even National (reorganized as AMI in 1925) itself recognized this, and had been busy developing its own version of the automatic phonograph—The National Selective Phonograph, Introduced in 1927.

• From Wikipedia: The Whippet is a sighthound breed that originated in England, where they descended from the larger greyhound. The name was derived from an early 17th-century word, now obsolete, meaning to "move briskly."

The whippet race diorama, and all of its many parts, contribute to what amounts to a substantial piece of ancillary equipment. It is bulky, it makes use of mechanical sub-assemblies, it has its own pneumatic valve chest for control and certain randomizing functions, and it requires a steel framework in which to contain and hold the various components. Consequently, it is weighty. Moreover, the external Geneva drive operated randomizer module needs its own source of motive power, as does the input shaft for the whippet race mechanism itself. To accommodate this additional need for rotational motive power, more chains, sprockets, and countershafts were added to the already existing maze of gears, sprockets, and drive shafts. Unfortunately, for pianos with the retrofitted diorama, it became much more difficult to tune the piano or work on anything requiring access to the stack or associated mechanisms.

The most obvious feature of the race diorama is the colorful hand-painted artwork background, signed by the artist, T. Koster. And although with a quick glance the background in various Whippet Race units may appear to be identical, each one differs slightly, because each has been individually painted. Another feature that catches the eye are the four little Whippets, either waiting impatiently to begin a new race, or perhaps they are randomly spaced along the raceway after finishing a race, still waiting to be reset, and made ready for the next event. When a Whippet’s nose touches the finish gate the race is over, and the underlying mechanism that moves the hounds is instantly disengaged—forever, unless someone intervenes and manually resets the race mechanism. On the front of the piano case, and to the lower right of the race diorama window, is a control lever with a Reset position to the left, neutral in the middle, and with a Race position on the right side. Starting a new race, or resetting a finished race, i.e., lining up the Whippets at the starting line, are manually initiated operations, and can only take place when the player system is powered on and rotational power is applied to the input shaft of the whippet race machinery.

There are two equally spaced Whippets mounted on each of four endless thin steel bands (copper bands reported in pianos #240 and #794), each band of which moves one of the hounds along the race track, with each metal band threading around a large diameter pulley at each end of the diorama’s short but pretty race track. The individual pulleys on the left side are merely idlers that maintain sufficient tension on the thin metal bands, while the pulleys on the right side do the work of pulling the hounds forward. For each metal band, only one Whippet is visible and on the race track at any given time; the other one being upside down and underneath the race track until it is moved into a starting position for the next race. Oddly enough, the real starting gate is a metal plate that stops the movement of the hidden Whippets underneath the race track, near the working pulleys. When the noses of all four upside down Whippets are touching the starting plate, the visible Whippets at the beginning of the race track will also be properly lined up and ready to go.

The four working pulleys (that move the Whippets) all ride, side-by-side, on the same gear driven shaft, i.e., they could spin freely on the shaft if it were not for the pair of unseen brake shoes opposite each other inside each pulley. These brake shoes are supported by a framework that is firmly attached to the drive shaft. Between the two shoes are compression springs, so that each brake surface exerts the same amount of resistance against the inner circumference of the pulley’s rim. The spring tension must be just sufficient to provide enough force so that the metal bands (to which the hounds are attached) will pull the Whippets unerringly forward, and yet, at the same time, easily and gracefully slip whenever the working pulley is intentionally or otherwise halted.

There are two normally occurring scenarios that necessarily halt the forward motion of the Whippets, but do so without stopping or interfering with the rotation of the drive shaft on which the four working pulleys ride. By default, a locking pawl is pushed into the toothed flange on the outer edge of the four working pulleys, which causes the brake shoes inside each of the locked pulleys to harmlessly slip, unless the locking pawl is released, during which interval of time a Whippet can race forward. As such, at the beginning of each race all four Whippets are halted, unable to move forward except when intermittently allowed to do so by the so-called randomizing system (consisting of several separate but interconnected components). Each Whippet will move forward in randomized spurts, with the outcome more or less unpredictable.

The other situation (when a Whippet must be halted) is when the whippet race mechanism is being manually reset, and the Whippets are moved forward until, one by one, each Whippet’s nose bumps up against the starting gate (located underneath the race track) and stops it dead in its tracks. Here, again, the brake shoes inside the working pulleys must slip, but this time because the associated Whippet is physically preventing from moving past the starting gate. Thus, no matter whether the Whippet is paused by locking the working pulley or by physically stopping the Whippet itself, the amount of braking applied, or perhaps alternately described as the amount of slip allowed, is a critical adjustment that must be set with care. The racing Whippets need to be reliably moved along when appropriate, but it is imperative that they be safely and gracefully stopped whenever a Whippet’s nose touches some kind of fixed barrier.

The mechanical mechanisms are generally easy to understand through casual visual inspection alone; the randomizing functions not so much. The valve chest and the four control pneumatics that link to the locking pawls for each of the four working pulleys are contained within the diorama underneath the race track. Two important associated but external components are connected by means of rubber tubing. The first of these is a narrow board containing a row of spring metal pallet valves mounted on top of the stack cover. Two different pallet valve board lengths have been observed to date, one with ten pallet valves and the other with fifteen pallet valves. Why this difference is not clear, but John Perschbacher, a collector who has studied National coin pianos, commented that the more pallet valves interacting with piano notes the faster will be the race. Whatever might be the reasoning behind the number of pallet valves, the free end of each spring metal strip rests over the end of a piano wippen, so that when the piano note is played, and the end of the wippen is pushed upward, the spring metal strip is raised slightly, opening a port otherwise sealed shut. On the fifteen pallet valve board approximately every forth valve is teed together, resulting in just four rubber tubes that represent the combined output signals generated by the individual pallet valves. These four tubes are connected to the Geneva drive operated randomizer module mounted on the “keybed” area of the piano.

The piano notes used by the pallet valves are the top ten or fifteen piano notes in the mid or tenor section of the piano. The idea here is that these selected piano notes will consistently be played often and somewhat randomly, or at least differently, for each and every tune. Thus, the resulting signal for each of the four output tubes will be randomized through various combinations of pallet valves being “unpredictably” opened and closed as the music plays. But to further confuse, scramble, and randomize the signals, the four output tubes are attached to the Geneva gear driven randomizer module.

• From Wikipedia: A Geneva drive or Maltese cross is a gear mechanism that translates a continuous rotation into an intermittent rotary motion. The rotating drive wheel has a pin that reaches into a slot of the driven wheel advancing it by one step. The drive wheel also has a raised circular blocking disc that locks the driven wheel in position between steps.

The gear driven randomizing device gets its motive power from an added chain sprocket stuck onto the end of the tooth clutch shaft used to rotate the revolver magazine during the music roll selection cycle. The purpose of the randomizing module is to take the four channels of gibberish coming from the pallet valves and further randomize it, by further breaking up any repeating patterns. Fundamentally, music is not random, but follows very precise rules, although taking part of a melody and screwing up any obvious patterns is probably more than sufficient for the simple requirements of the whippet race mechanism.

The randomizing part of the module consists of two thick circular metal parts mounted on a common shaft, and with a common machined-flat mating surface between the two halves, with each half drilled and channeled with various air-tight passageways. The stationary half has four incoming rubber tube connections (from the pallet valves operated by the piano wippens), and four output connections that go to the valve chest in the diorama, which, in turn, operate the four pneumatics that serve to intermittently and randomly unlock the working pulleys. The rotating half of the randomizer moves in discreet steps, advanced one step at a time by the Geneva drive, followed by a pause, with each step connecting the input and output tubes differently, adding more unpredictability to the action of each individual racing Whippet. The upshot here is this: Each Whippet moves forward unpredictably in spurts, each effectively inching toward the finish gate at varying speeds, all according to the multi-level randomization of signals generated by the piano note operated pallet valves working in combination with the independent stepped Geneva gear driven randomizing module.

And so, remains the question, who built this marvelous piece of novelty machinery? Was it National or some outside supplier? Fortunately, in a booklet published circa 1930, intended for the stockholder and interested member of the organization of the Automatic Music Instrument Company (and made available courtesy of John Perschbacher), it specifically states that the Automatic Whippet Race is a product of the company, as follows:

The product of the company consists of selective automatic phonographs, both coin-controlled and push button controlled, similar instruments newly developed and capable of playing either 10-inch or 12-inch records at will, selecting devices, novelty attractions such as dancing figures in the pianos and automatic whippet races for stimulating business, automatic distant control mechanism for musical instruments and the radio, automatic multiple selecting devices for the selective playing of a series of rolls for piano or pips organ and known as the “Multi Control.”

While the above might not be considered proof positive by some people, it does imply that National and/or AMI routinely manufactured novelty trade stimulators, which were retrofitted into many refurbished pianos during the late 1920s. Moreover, there is ample evidence that the company was both innovative and very capable of manufacturing various kinds of complex mechanical equipment, such as the Automatic Whippet Race.

Late Retrofit Dancing Figure Device

There are no currently known surviving National automatic pianos that have been retrofitted with the “Dancing Figure” trade stimulator. Moreover, until recently it was a mystery as to what it might have looked like. Thus, the idea of a dancing figure remained shrouded in confusion and mystery, with some palpable doubt as to whether such a device ever truly existed. But now a picture of a real “Dancing Figure,” installed in a National automatic piano, has come to light—a simple captioned sidebar picture in a booklet published, circa 1930, by the Automatic Musical Instrument Company, the successor to the National companies. And so a real “Dancing Figure” device did exist, and it is thought that the hand-painted background for the stage was done by the same artist that painted the colorful backgrounds for the Automatic Whippet Race.

But there are, however, some loose parts that were once thought might purportedly make up the “Dancing Figure” apparatus, crude parts that are thought to have come from a junked out National piano with a Simplex style stack. While it was hard to believe that National might indulge in such poorly executed craftsmanship, there was yet nothing to prove one way or another what a real “Dancing Figure” might look like. But, then, once a genuine picture of the real thing surfaced the amateurish and poorly executed kludge was immediately recognized for the counterfeit it was, made by some unknown person. Nonetheless, this homemade device is retained as part of this writing as an example of what might be encountered, something consisting of two main parts:

• One medium sized pneumatic that was attached to the inside of the removable upper front panel, and located above the glass window and to the left side of the “Music Changed Weekly” sign, and,
• A flexible doll, approximately the height of the front glass window, whose head was held suspended by a single wire around its neck, with the hooked end of the wire attached to an overhead control pneumatic (as described above).

The head of the doll is dark-skinned in color and sits atop what appears to be a flexible, jointed body clothed in a simple but attractive Native American costume. The only mechanical attachment was a single stiff wire encircling its neck (and then bent upward at a right angle), which, when yanked up and down by a pneumatic, caused the doll to flop around and “dance” as it bobbed to and fro. There was nothing sophisticated about its movement or how it was controlled.

To accommodate a retrofit “Dancing Figure” device, late National music rolls—possibly beginning with roll #9392, titled “Lady Luck” (Fox Trot)—used perforation #68 to operate the dancing figure attachment, and hopefully keep its flailing antics somewhat in time with the music. Curiously, roll #9392 only animated the dancing figure for the last half of the music roll, leaving the first half with a lifeless dancing figure. Because tracker bar hole #68 was re-purposed some restorers have reportedly disabled this tracker bar position for regular National pianos that do not have a “Dancing Figure” retrofit. Doing so prevents the uppermost playing note from sounding in odd and non-musical ways when playing late music rolls, which, if hole #68 was left connected, would upset the otherwise pleasant musical arrangement.

Using the Information on this Page

This page contains detailed descriptions and photos of various features, which are deemed important for a truly useful understanding of both the database reports and to effectively fill in the Survey Reporting Form. The actual database reports and survey form can be accessed in the Distribution of Database Information section and then clicking on the large Download button at the bottom of this page.

The National Automatic Piano Database

The primary information that inspired the creation of the National Automatic Piano database was gathered by Andrew Barrett, a mechanical music enthusiast located in California. Many other people, listed under Acknowledgments in the Introduction to the Registry, have also contributed information, all of which is presented in an orderly, easy to read format on this web page and in the database report offered below.

By default, current ownership information is not integral to the database project, but a provision exists whereby the current owner's name information can be accommodated and then shown in database reports. However, this will be done only if and when specific written permission is granted to the Mechanical Music Press specifically authorizing us to show and/or distribute individual ownership information. Furthermore, if and when such authorization is granted the Mechanical Music Press and/or its authors shall assume no liability or responsibility of any kind, nor to any extent, regarding any inferred, purported, or actual privacy intrusions, incidents, or claims.

Updating the Database and Reporting Errors

We cordially invite and solicit additional information for instruments currently not listed, and additional details for those already listed but that have little recorded information—every little bit more information helps.

To ADD ANOTHER ITEM TO THE DATABASE or to facilitate the reporting of errors please send an e-mail message to and include the following information:

1. Click here to download the Survey Form and then copy (press Ctrl + A to select all text) and paste the text into the body of an email. Then, at your leisure, you can answer the survey questions pasted into your e-mail message. You can also easily "Save As" the file to your computer or print the Survey Form and use it as a reference guide.
2. Attach a series of photographs showing the instrument's exterior and interior mechanisms to the e-mail message (for answering many technical questions often overlooked or misunderstood when answering questions on the Survey Form).

Thank you for any assistance you may provide. If we have any questions or need further information regarding your e-mail submittal someone will contact you.

Distribution of Database Information
Last Updated on November 28, 2017

The database PDF report files require Adobe Acrobat Reader (or its equivalent) to view, use, or print their contents. The free Acrobat Reader can be downloaded from Adobe by clicking on the icon at left.

Download the current database report as a PDF file by clicking on the button below.

6 pages.

All database report information is offered "as is," without any guarantee or warranty whatsoever of any kind, neither stated, implied, nor inferred, as to the accuracy, correctness, exactness, suitability, or usefulness of any content.

Credits:

John H. Perschbacher, Q. David Bowers, Art Reblitz, Dick Hack, Michael Montgomery (deceased), Willard E. Burkhardt (deceased), Durrell Armstrong (deceased), Terry Hathaway, Dana Johnson, Andrew Barrett, Bob Gilson, Jody Kravitz, and David Krall.

Graphics:

Dana Johnson, Dick Hack, John H. Perschbacher, David Anderson, Bob Gilson, Dan Zelinsky, David Krall, and Terry Hathaway.

References (courtesy of John H. Perschbacher.):

Foldout: The Nimble Nickel, The Coin That Makes Our Millionaires, National Automatic Music Company, circa 1912.

16" x 20" broadsheet: "Multi-Control" Heralds a Great Forward Step in the Reproducing Piano Industry, circa 1925

16" x 20" broadsheet: Nickels Piled Up into Millions of Dollars Under the Management of this Organization, circa 1926.

16" x 20" broadsheet: "The Aristocrat" Reproducing Piano, Peer of All Coin-Operated Musical Instruments, circa 1925-26

Booklet: History, Organization and Personnel of the Automatic Musical Instrument Co., Grand Rapids, Mich., circa 1930.

The Grand Rapids Press: News articles relating to Rowe/AMI dated July 11, 2006, and February 27, 2009.

Copyright © 2003-2024 Mechanical Music Press, et al. - All rights reserved