USA 01The American System and the Birth of the Machine Tool

Gun contracts and the idea of "interchangeability" (1785–1815)

The idea of interchangeable parts itself existed in France before America. In 1785 Thomas Jefferson, the American minister in Paris, took apart the locks (the trigger and hammer mechanism) of 50 muskets, mixed the parts, and confirmed with his own hands that they fitted properly however they were picked up and reassembled; he reported this home. They were made by a craftsman named Blanc^[1]. Charles H. Fitch, who wrote the 1880 census report, summed up that France had made attempts in 1717 and 1785, but they never became practical because of prejudice and a lack of system and machinery^[2].

National armories and gun contractors

The newly independent United States needed to make its own guns. A law of April 1794 provided for two national armories, and Springfield (Massachusetts) in the north and Harpers Ferry (in present-day West Virginia) in the south were chosen. In 1795 Springfield began making muskets copied from a French pattern, with 40 workers^[3]. At Harpers Ferry, work on the factory and the water canal began in 1799, and in 1801, when the canal for the waterwheels was completed, 28 gunsmiths made 293 muskets^[4]. Since the national factories alone could not supply enough, the government also ordered guns from private contractors.

In 1798 Eli Whitney, inventor of the cotton gin, received a contract for 10,000 muskets. He had been unable to protect his cotton gin patent and was looking for a new business. In a letter to the Secretary of the Treasury, Whitney promised that water-powered machinery would greatly reduce the labor, and he built a factory at Whitneyville near New Haven. But building the machinery from scratch took nearly two years: in the first year he delivered 500 muskets against a planned 4,000, and it took him eight years to complete the contract^[1]. Even so, the government kept advancing him money.

The next year, 1799, Simeon North, a scythe maker in Berlin, Connecticut, received a contract for 500 cavalry pistols^[1]. In an 1808 letter North wrote that if one worker concentrated on just one part, up to 2,000 pieces, the labor fell by a quarter and the work was better. Dividing the work by part and repeating the same operation became the foundation of the interchangeable system. Then in 1813 North signed a contract for 20,000 pistols that contained a clause requiring that "any part fit any other pistol." It is regarded as the first government contract to promise interchangeability^[1]. That year North moved to Middletown and built a three-story brick factory.

The Whitney "legend"

There is a famous story that Whitney, in Washington, assembled ten muskets from parts picked at random from piles. Jefferson also wrote in an 1801 letter that Whitney's locks could be assembled even after 100 of them had been taken apart and mixed^[1]. Roe's 1916 book made much of this story and concluded that the interchangeable system was born in the factories of Whitney and North. But in 1960 Woodbury of MIT questioned this view in a paper titled "The Legend of Eli Whitney and Interchangeable Parts"^[6]. Later research has increasingly held that interchangeability was actually achieved not at Whitney's factory but at Hall's works at Harpers Ferry, described next.

How interchangeable was "interchangeable" in this era? According to Fitch, early interchangeability relied heavily on filing parts to hardened jigs (templates in the shape of the part). Inspection was still loose, and filing along jigs wore out the files quickly, adding to the cost. Interchangeability was truly maintained only after parts were forged in steel dies and brought to size by milling, in which the cutter itself determines the shape^[2].

National armories, the copying lathe and the first milling machines (1815–1839)

Blanchard's copying lathe

The most complex shape in a musket was the wooden stock. Each one was shaped by hand and fitted to its metal parts, so even a skilled worker could make only one or two a day^[2]. Thomas Blanchard first ran a machine for turning the outside of gun barrels at Springfield in 1818, and in 1819 a machine that traces a pattern to cut identical irregular shapes won him a patent^[8][10]. It worked rather like a locksmith's key-copying machine: an iron model (master) and the wooden blank turned slowly at the same speed, and a rotating cutter wheel copied onto the blank the motion of a guide wheel that traced the model. Because the cutter rotates, the NPS explains, it is more accurately a kind of cutting machine ("shaper") than a true lathe^[8].

By 1827 Blanchard's gunstock machines had grown to 16 types, for cutting off, planing flat, grooving for the barrel, recessing the lock seat and so on, and they were used at both national armories. Even the early machines are said to have let one worker make five or six stocks a day^[2]. Interestingly, the accuracy of the woodwork got ahead of the metal parts. Because the lock plates still differed from one another, Blanchard even built a device that used the lock plate itself as the pattern and recessed the stock to its shape. Springfield used this method until 1840, by which time the metal parts could be made as accurately as the stocks^[2][1]. The only surviving Blanchard machine (1822) is still at Springfield^[8].

Hall at Harpers Ferry

In 1819 John H. Hall, who held a patent for a breech-loading rifle (loaded from the breech), signed a contract for 1,000 rifles and moved into the Harpers Ferry armory. He turned an old sawmill into a factory and spent years building precision machinery for making interchangeable parts^[7]. Hall built drop hammers, stocking machines, drilling machines, and special machines for straight, lever-shaped and curved cutting; the straight-cutting machine is considered one of the ancestors of today's milling machine^[7]. In December 1822 Hall wrote to the Secretary of War that he had established a method of economically making perfectly identical guns, true to the pattern, by the hands of ordinary workers^[7].

The first 100 rifles, completed in 1824, were fitted so closely that the gap at the breech would "let one sheet of paper slip in, but stop two"^[1][2]. In an age without micrometers, paper served as the feeler gauge. Then in 1827 three commissioners reported that 100 Hall rifles made in 1824 had been taken apart, their metal parts mixed and reassembled with 100 new stocks, and that they fitted well^[2]. Fitch attributed this success not to filing but to machining every fit and to die forging, and wrote that while the clearances were not fine, the interchangeability was real^[2]. In 1832 Colonel Talcott of the Ordnance Department reported that machines did almost all the work, with little dependence on handwork^[7].

The first milling machines

The machine that cuts flat surfaces and grooves with a rotating cutter, the milling machine, took shape in the armories of this period. Who built the first one is not clear. The oldest surviving milling machine is attributed to Whitney's factory; it has a power feed that moves the table by a feed screw, and a device that disengages the feed at the end of the cut^[9][2]. Roe dated it before 1818, but Woodbury pointed out that the only basis for that year was a single sentence in an encyclopedia and dated it "about 1820." Around the same time in Middletown, a small hand-fed milling machine is said to have come into use in 1818 at the shop of Robert Johnson and others, who had come from England, but this account was reconstructed from memory in 1900^[9]. As a firm record, Whitney mentioned milling in a letter to the Secretary of War in March 1823, and by then both national armories were also using milling^[9].

The first milling machine at Springfield arrived in 1834: a machine with a fixed spindle, used to rough out wrenches. In 1835 a mechanic was brought from Middletown to build milling machines^[2]. Around 1830 a method was also devised for cutting lock parts with form cutters whose outline was the shape of the part^[2][9]. Because the shape of the cutter becomes the shape of the part, differences in workers' hands do not enter. This was a big step from file finishing to machine finishing.

Inspecting with gauges

To make parts uniform you need tools that measure them and decide pass or fail. According to Roe, receiving gauges, into which the part is fitted to check it, were used at Middletown in 1829 and were commonplace at Springfield by 1840^[1]. Fitch writes that inspection became stricter from around 1829 and receiving gauges were used. By 1880, limit gauges, which set the limits with two sides, one that must go in and one that must not (go / no-go), had become standard, and the parts of a single Springfield rifle were inspected with 154 gauges^[2]. This is the prototype of today's system of fits (see Fits).

Private armories, milling machines and turret lathes (1840–1850)

Reform at Springfield

In 1840 Thomas Warner at Springfield introduced a system of maintaining interchangeability through milling machines, finishing to jigs and strict inspection. He built five milling machines whose spindle could move up and down along vertical ways, standardizing the machining of bayonets and eliminating grinding work^[2][9]. In 1842 production of a new percussion musket (Model 1842) began, and a full set of model jigs, taps and gauges was assembled^[2][1]. It is the first Springfield arm assembled entirely from machine-made, fully interchangeable parts, according to the NPS catalog^[3]. The ledgers show, however, that the old practice of trial-assembling locks in sets of ten and marking them before hardening survived until 1849^[2]. In the private sector too, Albert Eames of the Ames company assembled a set of gauges and jigs in 1842, and at each inspection took ten guns apart, mixed the parts and reassembled them to check^[2].

Robbins & Lawrence in Windsor

Another center of this era was Windsor, a small town in Vermont. In 1843 the skilled gunsmith Richard S. Lawrence and N. Kendall opened a gun shop, and the businessman S. E. Robbins brought them a bid for a government contract for 10,000 rifles^[1]. The three won the contract, built a four-story brick armory in 1846, designed new machines and rebuilt old ones, and refined the techniques of making interchangeable parts^[13]. They finished the first contract 18 months early and received another for 15,000^[1]. ASME (the American Society of Mechanical Engineers) has designated the shop a historic landmark as "the first machine shop to achieve interchangeable manufacture on a practical scale"^[14].

Windsor had three outstanding mechanics: Lawrence, the designer Frederick W. Howe, and Henry D. Stone, who would later develop the turret lathe. In 1848 Howe designed the company's milling machine and a profile milling machine (profiler), which traces a template to cut the outline of a part^[9][1]. The milling machine Howe designed in 1850 appears to have been the first milling machine built for sale to others, and it had power feed on the cross feed as well^[9]. It marks the passage from the era when armories built the machines they used themselves to the era of selling machines as products.

The beginnings of the turret lathe

When a single part needs many tools applied in turn, such as drilling, boring, threading and parting off, an ordinary lathe needs realignment every time the tool is changed. So the idea arose of mounting several tools on a rotating holder (a turret), so that simply turning it brought the next tool into the same position. According to American Machinist magazine in 1900, at a pistol factory in Middlefield, Connecticut, Stephen Fitch built at his own expense a turret machine carrying eight tools that indexed automatically, to make 30,000 percussion nipples, and used it from 1845 to 1853^[12][1]. Fitch's turret had a horizontal axis, and Root at the Colt factory also used horizontal-axis turrets. The vertical-axis turret lathe, which later became widespread, was first built as a product for sale by Robbins & Lawrence in 1854; Stone drew the drawings, and Howe and Lawrence were also involved in the design^[1]. In 1850 the company also built a box tool that held two cutters and a steady rest in a single box^[1]. The later story of the turret lathe (automatic screw machines, flat turrets) is covered in Chapter 6.

Colt and Root

Samuel Colt patented the revolver (a pistol with a revolving cylinder) in 1836, but his first company, in Paterson, closed in 1842^[1][11]. When the Mexican War brought two government orders of 1,000 revolvers each, Colt had the first batch made at Whitney's factory in New Haven and learned the details of manufacturing. He moved to Hartford in 1848, and in 1849 hired, at a high salary, Elisha K. Root, who had modernized the Collins company's ax factory^[1]. Root created drop hammers for die forging, machines for drilling, rifling and stocking, and a whole system of jigs, fixtures, tools and gauges. Roe summed up that the credit for the revolver itself belongs to Colt, but the credit for the way it was made belongs mainly to Root^[1].

World's fairs, the British inspections and machine exports (1851–1860)

London, 1851

At the Great Exhibition in London in 1851 (the Crystal Palace), Robbins & Lawrence exhibited Mississippi rifles made for the government on the interchangeable system and received a medal^[2][1]. Colt also exhibited revolvers^[13]. American guns that could be assembled from mixed parts caught the eye of British engineers. In 1853 the Englishman Joseph Whitworth, who had come to inspect the New York exhibition, visited Springfield Armory, timed each step of making gunstocks on Blanchard-type machines with a watch, and published the results in his report (1854)^[15]. What Whitworth recorded were the minutes and seconds of each operation. This very approach of measuring and comparing the working time of each machine leads on to later production management.

The British committee and Enfield

Next the British government sent a committee to study American machinery (the Committee on the Machinery of the United States). The committee toured Colt, Springfield, Robbins & Lawrence in Windsor and others. At Springfield the superintendent, Major Ripley, took apart ten muskets, one from each year from 1843 to 1853, mixed the parts and reassembled them in front of the committee^[2][1]. Fitch and Roe give this as 1853, but the committee's report is dated 1855 and was presented to Parliament^[16]. The visit is often dated 1854, so the timeline on this page gives it as "1853–54."

The result was a large order. In 1855 American companies supplied Britain with 20,000 interchangeable Enfield rifles and a complete set of arms-making machinery. The first 157 machines cost $44,360 and included 57 milling machines, 8 universal milling machines, 12 four-spindle drilling machines, and special machines for threading and grooving^[2]. Roe noted that this list included almost no lathes. Most gun parts were not round, and were cut on milling machines and special machines^[1]. The machines were built mainly by Robbins & Lawrence and the Ames company, and James H. Burton, who had worked at Harpers Ferry and at Ames, went to England to direct their installation and operation at the government factory at Enfield^[1][2]. Fitch writes that these machines were widely copied in Britain and Germany^[2].

Interestingly, Robbins & Lawrence, which exported the 157 machines, did not think of itself as a machine tool builder, since it was normal for a factory to build the machines it used. Even so, Ashburn writes that this order was a turning point for the business of building and selling machine tools^[21]. Robbins & Lawrence later ran into trouble after venturing into railroad cars and taking a large rifle contract that fell through, and went bankrupt in October 1856^[17]. In 1859 the Windsor shop passed to Lamson, Goodnow & Yale. The designer Stone stayed on, which eventually led to Jones & Lamson^[1].

The Lincoln miller and Colt's new armory

In 1855 George S. Lincoln & Co. of Hartford built milling machines designed by F. A. Pratt for Colt's new factory. This was the Lincoln miller. It was based on the Robbins & Lawrence machine. Pratt changed the table feed from a rack and pinion to a feed screw and nut, because the former tended to chatter at the start of a heavy cut, and made the whole machine sturdy and compact^[9][1]. The machine itself was made in quantity with interchangeable parts from the start, and Fitch writes that about 100,000 had been made, including copies, by 1880 (Roe says more than 150,000)^[2][1]. The year before, in 1854, Root had fitted the milling machine with an overarm that supports the end of the spindle from above^[9]. How the structure of the milling machine developed is covered in detail in Chapter 7.

Colt's Hartford armory (the East Armory) was completed in 1855. Designed by Colt, Pomeroy and Root, with large windows, ventilation and gas lighting, it was for a time the largest private armory in the world, the NPS explains^[11]. According to Roe, it held more than 1,400 machines, most designed and built in-house, and as much was spent on fixtures and tools as on the machines^[1]. By 1857 it was finishing 150 revolvers a day^[11]. This factory trained Pratt and Amos Whitney, who later founded Pratt & Whitney, Spencer of the automatic screw machine, Billings of forging, Gleason of bevel gears and others (Chapter 2)^[1].

Why America? In an 1812 letter Whitney wrote that the aim of his system was to replace with the precise operations of machinery the skill of the craftsman, which is acquired only by long practice and which this country hardly possessed^[1]. Fitch summed up the conditions for the interchangeable system as "an assured large demand" and "a need for fine work," and explained that they first came together in arms making under government contracts^[2]. Skilled craftsmen were scarce, yet quantity was needed in bulk, and the customer, the government, paid advances and waited for years. This combination made investment in special machines, jigs and gauges pay off.

★The Civil War and beyond (1861–1865)

When the Civil War began in April 1861, the Harpers Ferry armory was burned and Springfield took over making rifles for the Union. Production, 1,000 a month at the start of the war, reached 1,000 a day at its peak^[3]. In the private sector, Colt built the West Armory, doubling the plant, and took orders for Army and Navy revolvers^[1][11]. Colt died in January 1862, and Root led the factory as president. In February 1864 the East Armory burned down^[11].

The war gave birth to new machine tools outside the gun factories. In 1861 Brown & Sharpe built a turret screw machine for the Providence Tool Company (Providence), which had a contract for Springfield-pattern muskets. It appears to have been the first machine tool the company built for sale^[1]. Furthermore, to eliminate the labor of filing by hand the helical grooves of the drills that bored musket nipples, J. R. Brown built the universal milling machine and sold the first one to that company in March 1862^[20][1]. This machine, able to cut helical grooves, gears and irregular shapes, became the star of the next era (Chapter 2). The repeating rifle of Christopher M. Spencer, who had worked at Colt, was bought by the government in some 200,000 units^[1]. Spencer later built a cam-driven automatic screw machine.

When the war ended, surplus arms and armory machinery went overseas. The 1880 report says that the governments of Britain, Russia, Prussia, Spain, Turkey, Sweden, Denmark, Egypt and others bought American arms-making machinery, and that Japan, Argentina, Chile, Peru, Mexico and others also obtained much from the United States^[2]. In the 1870s Pratt & Whitney supplied machinery to the Prussian arsenals at Spandau, Erfurt and Danzig, and the Prussian government wrote in a letter of thanks that it had become "largely free of the skill and strength of the workman" and that wages had been cut in half^[2]. The interchangeable system born in the armories spread to sewing machines, clocks and watches, bicycles and farm machinery.

What it means on the shop floor. Many of the tools born in this era are still in factories today in changed form: form milling cutters whose shape determines the part, copying by tracing a pattern (the ancestor of today's NC tracing and digitizing), turrets that bring tools into position in turn, go/no-go limit gauges, and jigs and fixtures that locate parts in a set position. Writing tolerances on the drawing, deciding pass or fail with gauges, and leaving the setup to fixtures: the starting point of today's mass-production thinking is here.

⛩Links with Japan

While the American System was growing, Japan was still in the era before it had machine tools. In 1860 (Man'en 1), the Japanese embassy sent to the United States to ratify the Treaty of Amity and Commerce landed at the Washington Navy Yard, and later toured its furnaces and other facilities. The envoys also visited Philadelphia and New York^[19]. However, the earliest machine tools to arrive in late Edo-period Japan were the machines of the Nagasaki Ironworks (dated 1856), probably Dutch, and the French machines bought by the Yokosuka shipyard in 1867 (marked Mulhouse 1866), not American ones (Japanese Archive, Chapter 1)^[22].

As a primary source on the relationship between American arms-making machinery and Japan, the 1880 census report says that Japan was one of the countries that obtained much from the United States^[2]. However, it is unclear whether this refers to the guns themselves or to the machines for making them, and we could not confirm which machines went to which arsenal. Japanese machine tool building took a big turn toward American methods from the end of the Meiji era, when the Tokyo Higher Technical School invited the American engineer Charles A. Francis and Ikegai Iron Works began mass-producing American-type lathes. This is covered in Chapter 2 and Japanese Archive, Chapter 1, "The Lineage of Japanese Lathes".

FAQ

What is the "American System"?

It is a way of mass-producing without relying on the hand fitting of skilled craftsmen: parts are made identical with special machines, jigs and gauges so that any part can be assembled (interchangeability). It was developed by American armories and arms contractors in the first half of the 19th century, and came to be called this in Europe from the 1850s, when Britain adopted it for its arms factory at Enfield.

Did Eli Whitney invent interchangeable parts?

That used to be said, but the view has been revised. The idea had precedents in France (reported by Jefferson in 1785), and in 1960 Woodbury questioned whether Whitney's musket parts were really interchangeable. An early recorded demonstration of interchangeability at a national armory is the 1827 inspection in which 100 of John H. Hall's rifles at Harpers Ferry were taken apart and reassembled.

Who built the first milling machine?

It is not clear. The oldest surviving one is said to be the milling machine from Whitney's factory (about 1820); there is also the story of a hand-fed machine at Middletown around 1818, and Hall's "straight-cutting machine." The firm record is Whitney's letter of 1823, by which time milling was also in use at the national armories at Springfield and Harpers Ferry. Milling machines were first widely built for sale with the Lincoln miller of 1855.

How does Blanchard's copying lathe connect to today's machining?

The idea of tracing a model (master) to copy the same shape led to later copying lathes and copy milling machines, and to digitizing, which measures shapes and turns them into NC data. Its other significance is that it showed that the wooden gunstock, the part that "most needed a craftsman's skill," could be made by machine, which helped extend the interchangeable system to metal parts.