The Dawn of Innovation (10 page)

 

D.
Using geometrics mehods, an artisan could mark reasonably-accurate divisions on a large disc, and then capture those same proportions on a much smaler workpiece, as shown. Note that the index wheel includes several choices of toothh arrangemenr and that the index pointer fixes the index wheel and the workpiece in position for each cut.
E.
A pocket-sized thousandth of an inch micrometer first appeared in the catalog of the Providence firm of brown and Sharpe in 1877, and marked a high point of convenient recision in the nineteenth century. The micrometers are still in wide use. The numbered divisions on the barrel signify tenths (0.1) of an inch. Each smaller division is a fortieth (0.025) of an inch, while each small numbered marking on the handle is a thousandth (0.001) of inch In the illustration the readout on the barrel is 0.125 inches, plus an additional .001 on the handle = 0.126 inches.

The great figure in screw threads was Henry Maudslay, one of the greatest machinists of all time. Although he ran a large establishment in his later years, he was at heart a shop-floor machinist. He also appears to be have been a man of immense calm and good humor, certainly comfortably fat in his later years. His workers adored him, almost as much as they admired his technical skills. One recounted fondly, “It was a pleasure to see him handle a tool of any kind, but he was quite splendid with an eighteen-inch file.”
¹⁶
Maudslay's permanent contribution was to stabilize machining at very high levels of precision, hardly surpassed to the present day.

When Maudslay began his career, screw threads were in a state of disarray with respect to pitch (thread count), shape, angle, and uniformity, and they became his pet project. While he did not invent the modern screw-cutting lathe—Ramsden anticipated much of his work—his first versions achieved such a high pitch of perfection that they became the standard for all such instruments.

In a traditional lathe, the worker held the cutting tool against the rotating workpiece. In the Maudslay screw-cutting lathe, first produced between 1797 and 1800, before he was thirty, the workpiece was positioned on a slide by a long lead screw, the cutting depth of the tool was set by a screw-driven micrometer, and a gear-set controlled the thread count of the new screw by varying the speed of the workpiece rotation relative to its lateral motion on the slide. Different gear settings allowed reliable production of a variety of thread pitches. The gearing on one early machine accommodated twenty-eight different thread pitches. One Maudslay-produced screw, created for a dividing engine to be used in the production of large astronomical instruments, was five feet long and two inches in diameter, with 50 threads per inch, or 3,000 threads in all; it came with a foot-long nut with 600 threads. No one before Maudslay could have produced such an instrument.

Maudslay's obsession with accuracy pervaded every aspect of his machinery, since vibrations, misalignments, or slightly loose fittings make a
mockery of ultraprecise tool settings. Maudslay's constructions set new standards for solidity, stability, and perfection with respect to planes, angles, and uniformity of motion. He built a bench micrometer to measure deviances of a workpiece from a pattern to a ten-thousandth of an inch. He called it the “Lord Chancellor,” the final arbiter of any dispute.

Maudslay also insisted on absolutely flat, smooth planes on every surface, and every machinist in his shop was equipped with a plate that met that standard. His famous protégé, James Nasmyth, wrote that they were used to test “the surfaces of slide valves, or wherever absolute true plane surfaces were essential to the attainment of the best results.” When absolutely true surfaces were placed on each other, Nasmyth went on, they “would float upon the thin stratum of air between them until dislodged.... When they adhered closely to each other they could only be separated by sliding each off each.”
¹⁷

The method of creating perfect planes was to start with three plates machined as perfectly flat as they could be. One plate, Plate A, was then coated with a colored powder, and Plate B placed precisely on it. When the two plates were separated, the color marks on B would mark its high spots relative to A. B was then scraped by a very hard hand scraper (both machine and hand grinding were far too coarse) to remove the anomalies. The process was then reversed—coloring B and scraping the discrepancies from A—and repeated as often as necessary until each plate color-matched across its entire surface. But a perfect match between two plates did not yet prove a perfect plane, because they might embody complementary deviances. Therefore the whole process was repeated twice more, first matching A to C and then C to B, at which point, one could be confident that the three approached “absolute truth.”

With some ingenuity, the method may be extended to produce perfect right angles and perfectly parallel rectangular bars with perfectly aligned plane ends. (Hint: each of those processes requires
four
plates and
two
bars.) A nineteenth-century textbook warns merely that “the only thing to be dreaded is the discovery of a hollow portion, which may compel a repetition of the procedure from the commencement.”
¹⁸

Maudslay protégés dominated the British tool industry for decades. Nasmyth invented the steam hammer in 1838. It operated on the same principle as a drop hammer: a giant hammerhead fell on a forging target. The difference was that the steam hammer's blow was piston-driven for an even more powerful impact. Nasmyth's hammer weighed two and a half tons but was under such precise control that it could rattle a whole factory or break an egg in a wine glass without disturbing the glass.
¹⁹

Of all the Maudslay disciples, the greatest may have been Sir Joseph Whitworth, who unified British screw designs under the “Whitworth standard,” specifying radii, pitches, angles, and depths that for many years served as nearly a world standard. He also carried the quest for absolute mechanical precision about as far as it could go before the age of electronics.

Sir Joseph's portrait shows him long-faced, heavy-lidded, and skeptical. His father was a minister and schoolmaster, so he was better educated than most craftsmen and did not easily tolerate fools. Pompously ignorant officialdom was among his particular bêtes noires. As a leading machinery and metals fabricator, Whitworth was necessarily a force in armament procurement. A rigorous experimenter and prolific inventor, he produced a great flood of weapons designs and experimental new weapons, which were routinely rejected by the military authorities, although a number were later quietly adopted. Those misadventures are documented in a coldly sarcastic little book he published in 1873, when he was seventy.
²⁰

Whitworth was prompted to develop his measuring machine by the mid-century bumblings of a parliamentary Committee on Standards, which labored for eleven years to create a bar equal in length to a standard yard. The stumbling blocks lay in the physics of ordinary materials at microscopic dimensions. Tiny increases in ambient temperatures changed the length of a bar, the minutest forces caused nearly undetectable sags or flexes, and so forth.

But the committee painstakingly worked its way through all such obstacles to the point where the stage was set for the climactic measurements. Ambient temperature was controlled by a thermometer accurate to within 1/100°F. The bar itself was suspended in a tub of mercury to equalize ambient pressures, while the tub was shielded within a tank of water to muffle the slightest external disturbances. The readings were taken by microscopes on platforms at each end of the suspended bar.
²¹

There was no difficulty in producing an initial standard bar—a yard, after all, was whatever the bar said it was. The challenge was to produce
additional
bars of exactly the same length—enough to be distributed among the universities and science establishments to serve as the reference point for high-precision undertakings of all kinds. The Royal Astronomical Society, an unofficial kibitzer in chief, expected that standard bars should be executed in the primary metal types—“copper, brass, cast-iron, Low-Moor iron, Swedish iron and cast steel.”
²²

With the elaborate arrangements all in place, the great day finally came. The committee duly selected one bar, “bronze 28,” as the standard, and proceeded to measure six other bars against it. That there were differences was no surprise: everyone agreed that “bronze 19” was not exactly the same length as bronze 28. The surprise was that committee members differed on whether it was longer or shorter, although the observations had been made sequentially under completely identical conditions. Disconcerted, the committee recruited volunteer observers, all men with professional backgrounds in related fields. Some 200,000 measurements later, the committee reluctantly concluded that no consensus could be reached on any of the bars. The differences, moreover, were not random: some individuals were consistently on the short side, others consistently on the long.

Whitworth's reaction was a jeering
of course!
The measurements were taken by comparing the position of a cross-hair in the microscope with a tiny etched gradation on the bar. But the crosshair, a mere spider-silk to the naked eye, was magnified to a thick, fuzzy, line that, in the blurring vision of a committeeman, danced from one side to the other of an equally
gross and irregular gradation line. The Astronomical Society put the best face on it, remarking that the uncertainties were “not likely to affect any useful observation” and that “a limit seems to be shown which, in any
optical measures
, no amount of observation in our current state of knowledge can overpass.”
²³

Whitworth thought it all surpassing idiocy. The committee's measurements seemed to randomize at resolutions of about 1/30,000 of an inch. His own workshops frequently worked at tolerances of 1/50,000 of an inch, and most experienced machinists could accurately distinguish differences of 1/10,000 of an inch by feel.
^m
Where the committee had gone astray, he argued in multiple venues, was in attempting visual assessments of length as opposed to merely determining the
difference
between two bars, which could be done mechanically with great precision. Given any standard bar with perfect plane ends, Whitworth claimed, he could determine the difference between it and any other like it at resolutions of a millionth of an inch, which he then proceeded to demonstrate. (Maudslay's micrometer measured to the ten-thousandth of an inch, a hundred times larger.)

The measuring machine Whitworth constructed to fulfill his boast comprised two parallel steel bars resting on a heavy cast-iron stand—the bars, the stand, and their respective borings all corrected to near-absolute truth. The working portion of the machine was a grip made by two opposed, perfectly trued small circular plugs. One of the plugs was driven by a leadscrew with a pitch of 20 threads to the inch; that screw was driven by a wheel with 200 teeth; that wheel in turn was driven by a “division wheel” with 250 marked divisions: 250 × 200 × 20 = 1,000,000. Turning the division wheel by one notch advanced or withdrew the plug by 1/1,000,000 of an inch. Equivalently, a 1/1,000,000 of an inch movement in the lead screw generated a visible movement of about .04 inches in the division wheel, a 40,000:1 magnification.

 

Whitworth's famous measuring machine was designed to measure the
difference
between two apparently identical objects to the millionth of an inch. It takes 200 turns of the 250-division “division wheel” to turn both the 200-tooth rear wheel and the interior lead screw one full turn. Twenty full turns of the lead screw will advance the plug by a full inch—250 × 200 × 20 = 1,000,000. The bottom drawing shows the placement of the “feeling piece,” a very thin sliver of metal with the shape as shown. The experimenter placed the first piece in the machine with the feeling piece, then gently relaxed the pressure until the feeling piece moved, and noted that place on the division wheel. He repeated the procedure for the second piece and compared the two markings.