Moreland’s calculation was not only the most precise estimate for more than a century; it was also critical for building any sort of working steam engine. Knowing that steam, once condensed back into water in a sealed container, leaves behind a vacuum that takes up two thousand times the cubic area of the condensed liquid is very nearly as important as knowing the temperature at which water boils—perhaps more so, since the first working steam engines were built decades before the first accurate thermometers.
Thus, when Savery made the first demonstration of his pumping machine “at a potter’s house in Lambeth,”8 two years before he did so with an identical machine at Gresham, he owed a large debt to his employers at the Royal Office of Ordnance. That machine consisted of a tall cylinder filled with water and connected to a boiler, in which Savery produced steam, which he then introduced into the cylinder at a pressure estimated to be about 120 pounds per square inch. The pressure pushed the water out one end of the cylinder, leaving steam behind; when the steam-filled cylinders were sprayed with cold water, the resulting vacuum pulled water from a chamber below, creating a pumping action.
Fig. 1: Thomas Savery’s pumping machine, as seen in a lithograph from his 1702 book The Miner’s Friend. The image on the right shows the components: When the canister sprayed cold water on the steam filled cylinder, the resulting vacuum pulled the water up. On the left is the machine at work, two-thirds of the way down a mine shaft, since the vacuum could pull the water only a bit more than twenty feet. Science Museum / Science & Society Picture Library
Savery’s machine was a long way from perfect. The use of water as its “piston” meant that the engine couldn’t pump anything else. Unlike Papin’s digester, it lacked any sort of safety valve. Using it even in its (slightly) improved version would have required an operator to open the steam cock and the cold water valve at least four times a minute; to refill the boiler at least once a minute;9 and to stoke the fire under it as needed. The cylinder was soldered together, and the solder had a melting temperature dangerously close to the temperature of the steam at high pressure, which could exceed 350°F/175°C. Worst of all, while the high-pressure steam could, in theory, push the water several hundred feet upward, the machine depended on suction for the first leg of the journey. Any working Savery pump needed to be built not at the top of a mine shaft, but no more than twenty-five feet from the bottom.
The reason for this limit, which had been well known to Galileo, Torricelli, and Papin, was atmosphere. At sea level, the maximum height that water can be lifted inside a tube under perfect conditions is about thirty-four feet, or just a little more than ten meters, calculated by dividing atmospheric pressure at sea level (14.7 pounds per square inch) by the weight of a cubic inch of water, or 0.0361 lbs. At this point, the water inside exerts a pressure equal to the weight of the atmosphere pushing down on the water’s surface. Since this represents a theoretical limit, requiring a perfect vacuum, the practical limit is even lower, usually assumed to be around twenty-five feet, as anyone who has tried to use a suction pump to draw water to the top of a three-story building has learned. This was a fairly serious problem for draining mines that were already more than one hundred feet underground.
Nonetheless, Savery’s machine was a revelation. And not merely to the “gentlemen, free and unconfin’d” of the Royal Society, who reported, with characteristic understatement, “Mr. Savery … entertained the Royal Society10 with shewing a small model of his engine for raising water by the help of fire, which he set to work before them, the experiment succeeded according to expectation, and to their satisfaction,” but to King William III in a private showing at Hampton Court. More modestly, but far more importantly, it also inspired a Devonshire ironmonger and blacksmith named Thomas Newcomen.
FOR CENTURIES, THE LANDED gentry who held the lands around Dudley Castle in the West Midlands of England prospered in direct proportion to the value of the minerals extracted from those lands. Indeed, that prosperity often took precedence over maintaining the land, and by the mid-1660s, the current Baron Dudley had heavily mortgaged the lands—so heavily, in fact, that he was forced to marry his daughter to someone wealthy enough to pull the family out of debt. The priority of succeeding barons was, as a result, the revitalization of the Dudley real estate, the most valuable pieces of which were the Conygree coal mines, lying one mile east of Dudley Castle.
The Conygree mines, like all excavations, were only workable when dry, or at least free of standing water. This, of course, is why Savery called his pump the “Miner’s Friend” in an eponymous 1702 book. The book, and the invention, demonstrated how Torricelli’s (and von Guericke’s) vacuum could be economically created using the two-thousandfold difference in volume between water in its liquid and gaseous state, and showed how such a vacuum could pull water out of any mine.
So long as the pump could be built no more than twenty-five feet from the mine’s floor.
After two hundred years of excavating, however, the mines at Conygree were more than six times deeper than the working distance of a vacuum pump, which meant a Savery-style engine would need to be built (and operated) more than one hundred and twenty-five feet below ground level. What they needed was an entirely new machine. Even more, they needed an act of genius, and this time the word, so frequently devalued by overuse, is appropriate. One historian of science calls the machine that made history near Dudley Castle in 1712 “one of the great original synthetic inventions11 of all time.”
The synthesis in question tied together two intellectual threads whose history dates from Heron’s first-century Alexandria. The first was man-made vacuum: the concept that was studied by Torricelli and pursued by Giambattista della Porta and Salomon de Caus, and that reached its culmination in Savery’s “new Invention for Raiseing of Water.” The other was the realization that a functional piston could be driven by atmospheric pressure, which was investigated by Huygens and described, though not built, by Papin in 1690. The 1712 engine of Thomas Newcomen,12 probably the first working engine built by this enigmatic man, was certainly the first to connect the two threads; and if any single invention can be said to have inaugurated the steam revolution, this was it.
Much more is known about the machine than its inventor. He was born in 1664 in Dartmouth, to a family that may have been in the shipbuilding trade. In his teens he was in all likelihood apprenticed to an ironmonger—part smith, part hardware salesman—since he was practicing the trade as a journeyman by the age of twenty-one, but his name doesn’t appear in records of indentured freemen, likely because his family was Baptist in a land that recognized only the Church of England.
Newcomen’s religion had consequences greater than absence from a local census. Dissenters, including Baptists, Presbyterians, and others, were, as a class, excluded from universities after 1660, and either apprenticed, or learned their science from dissenting academies.
Bad luck for the universities, good luck for the nation. Only decades after a tidal wave of scientific knowledge started washing over Britain—the first English translation of Galileo’s Dialogue Concerning Two New Sciences was published in the 1660s, nearly seventy years before an English edition of the Principia of Isaac Newton (Latin edition, 1687)—some of the nation’s most ambitious and practical young were excluded from Oxford and Cambridge. At the same time that he chartered the world’s first scientific society, Charles II had created an entire generation of dissenting intellectuals uncontrolled by his kingdom’s ever more technophobic universities. Some attended so-called dissenting academies, which mimicked an Oxbridge classical education with notably less arrogance about the teaching of science and modern languages. Many more learned their science in the most practical way: as apprentices to artisans who were more likely to be literate than ever before in history.
Newcomen may have been unable to translate Horace, but that did not mean he was, in any important sense, uneducated. He could perform calculations rapidly, knew a fair bit of geometry, could calculate the stren
gth and velocity of moving parts, could draw clearly, and—obviously—could read all that was available on subjects that interested him.
With books to read, and tools to practice his trade, Newcomen might still have lacked sufficient resources to travel all the way to Conygree but for one more unanticipated consequence of the Restoration: a package of laws that prohibited pastors who refused to conform to their dictates—using the Book of Common Prayer, for example—from teaching or preaching anywhere within five miles of their former “livings.” As a result, Dartmouth’s Baptists hired as their pastor the Reverend John Flavel, a well-known Presbyterian who not only led secret community services (another law forbade religious gatherings of more than five people) but organized secret community banks as a method of pooling their resources. One of them funded Newcomen’s first experiments.
This is worth underlining. The significance of the Dartmouth “bank” in the history of steam power is real, but modest. However, it is also a reminder of what we might call the British advantage in the development of the Industrial Revolution. Compare, for example, the experience of being a Baptist in Restoration England with that of being any sort of Protestant in seventeenth-century France. Newcomen may have been invisible to his local census, but at least he was not, like Papin, exiled from his country.
Thus, between 1700 and 1705, while Papin was wandering across central Europe trying to secure a pension, Newcomen, and his partner, John Calley (sometimes Cawley), a glazier (sometimes a plumber),13 set up a workshop in Newcomen’s basement, financed by Flavel’s bank, and started experimenting. During the next decade, more or less, most of their time was spent on trying to improve one or the other of the two seemingly independent threads of steam engine development: Savery’s vacuum, and Papin’s piston.
Frustratingly, we know little of just how and when the knowledge of the two came to Newcomen. Relatively detailed descriptions of Savery’s “Miner’s Friend” were, of course, available to anyone who could read once it appeared in the Royal Society’s Philosophical Transactions in 1699, and certainly after the onetime military engineer published his book cum sales brochure in 1702. Intriguingly, Savery was then living in Modbury, only fifteen miles from Newcomen’s Dartmouth home, and we know that he was regularly hiring artisans to build models and parts for his engines. Given Newcomen’s reputation as an ironmonger and wheelwright, it isn’t a huge leap to imagine some contact between the two.
Details about Papin’s piston-driven engine weren’t quite as public, but they were scarcely secret. Newcomen maintained active correspondences with a number of contemporaries, none more important than that with Robert Hooke, one of the most wide-ranging intelligences of the entire century. Hooke corresponded not only with Newcomen but with Papin as well, and he very likely kept the former apprised of the latter’s progress.14 At some point before his death in 1703, Hooke even talked Newcomen out of Papin’s idea of driving the pump’s pistons by air pressure, urging him instead to pursue the idea of creating a vacuum under the piston, writing “could he [Papin] make a speedy vacuum15 under your piston, your work is done….”
Well, not quite done. Newcomen’s real conceptual breakthrough came when he finally combined the strongest features of the two different approaches—or, at least, discarded their weaknesses. His brilliant synthesis lay in forgoing Savery’s dependence on vacuum to raise water, and Papin’s use of a piston operated by expanding steam. And in adding one critical element: the beam.
The most conspicuous mechanical element in Newcomen’s 1712 engine—for that matter, the most conspicuous element in virtually every steam engine for the next century and a half, including the Crofton pump station’s engine 42B—was its horizontal working beam. It looks a bit like an unbalanced seesaw, with the underside of one end attached to a piston and the other to a pump rod holding a bucket, which made the pump end much heavier. When at rest, therefore, the beam angled down toward the bucket at the bottom of the mine shaft, which forced the piston, inside a cylinder filled only with air, up to its highest point. Since the bucket on the end of the pump rod could be hundreds of feet below the guts of the engine, the critical problem with Savery’s engine—the need to place it near the bottom of the mine shaft—was solved. In fact, the only limitation on the depth at which it could work was the weight of the cable holding the bucket, which was relatively insignificant.
Newcomen’s first brilliant innovation—to lift water by seesawing a horizontal beam—was entirely dependent on his feel for the machine’s geometry; the beam, however, didn’t do anything about the need to renew the cycle of vacuum in a “speedy” manner, as advised by Hooke. This was critical for a working machine, which had to do more than impress German princelings or even the Royal Society; it had to return to room temperature after being heated well past the boiling temperature of water, and it had to so a dozen times a minute.
Savery’s method for producing condensation—spraying cold water on the outside of the cylinder—was simply too slow. Calley and Newcomen had designed a lead envelope to surround the cylinder, into which cold water could be poured, which improved the speed for heating and cooling the cylinder, but not a lot. Enter luck: The cylinder was essentially a flat piece of tin wrapped into a cylinder shape, its ends held together with a strip of solder. At one point, the solder was imperfectly applied, and the heat of steam in the cylinder melted it, opening a hole. When Newcomen poured cold water into the lead envelope wrapped around the cylinder, a stream of it found the hole, rushed through it, and condensed the steam immediately, with powerful results. For purposes of the experiment,16 Newcomen had attached a weight to the end of the beam to represent the weight of water; when the steam condensed, it pulled the beam down so violently that it broke the chain, the bottom of the cylinder, and even the lid of the boiler underneath.
Newcomen and Calley had, in broad strokes, the design for a working engine. They had enjoyed some luck, though it was anything but dumb luck. This didn’t seem to convince the self-named experimental philosopher J. T. Desaguliers, a Huguenot refugee like Papin, who became one of Isaac Newton’s assistants and (later) a priest in the Church of England. Desaguliers wrote, just before his death in 1744, that the two men had made their engine work, but “not being either philosophers17 to understand the reason, or mathematicians enough to calculate the powers and to proportion the parts, very luckily by accident found what they sought for.”
Fig. 2: The engine that Thomas Newcomen and John Calley erected at Dudley Castle in 1712, as seen in a 1719 engraving, used its vacuum to drive not water, but a piston attached to a beam. Science Museum / Science & Society Picture Library
The notion of Newcomen’s scientific ignorance persists to this day. One of its expressions is the legend that the original engine was made to cycle automatically by the insight of a boy named Humphrey Potter, who built a mazelike network of catches and strings from the plug rod to open the valves and close them. It is almost as if a Dartmouth ironmonger simply had to have an inordinate amount of luck to succeed where so many had failed.
The discovery of the power of injected water was luck; understanding and exploiting it was anything but. Newcomen and Calley replaced18 the accidental hole in the cylinder with an injection valve, and, ingeniously, attached it to the piston itself. When the piston reached the bottom of the cylinder, it automatically closed the injection valve and opened another valve, permitting the water to flow out.
Indeed, the valves, one between the boiler and the piston, and the self-acting valve, with a flap that closed once the condensed water was let out of the bottom of the cylinder, demanded quite as much ingenuity as the horizontal beam itself. All of the discoveries since Torricelli had underlined the potential power of vacuum combined with atmospheric pressure, but the power remained potential so long as the vacuum was unstable; losing the vacuum in the middle of a cycle was functionally equivalent to getting off one end of the seesaw while the other kid is still up in the air. Maintaining it, which meant in practice keeping
any air out of the cylinder, was therefore critical, and to do so Newcomen invented what he called a “Snifting Clark” (so called because, in the words of a contemporary observer, “the Air makes a Noise19 every time it blows thro’ it like a Man snifting with a Cold”), a valve carefully designed—not too heavy, not too light—to blow the air out of the chamber without letting any steam escape. Another vacuum preserver, probably the simplest, was the layer of water Newcomen added at the top of the piston, which served to seal the chamber from any air, which would compromise the vacuum.
Newcomen spent ten years experimenting with solutions to the problem of maintaining a regular and stable motion in his engine. None of his solutions was more innovative than his so-called plug rod. Since the machine depended on regular injections of water to condense the steam, it required an equally regular water supply. In Newcomen’s machine, this water was held in an overhead tank; gravity could be relied upon to move water from the tank into the cylinder, but to feed the tank itself, another pump was necessary. Newcomen suspended the plug rod from the horizontal beam itself; this rod, in turn, operated the cylinder valves, thus connecting the flow of water from one chamber into another. As water was pumped into the overhead tank, it also lifted the plug rod and thereby opened the valves of the cylinder, giving the beam a continuous (though jerky) motion. The ingenious F-shaped lever that opened the catch and operated the injection valve may have been a primitive design, but when the University of Manchester Institute of Science and Technology built a scale model of the original 1712 engine in 1968, “the valve still functioned perfectly,20 and was another amazing case of Newcomen arriving at the correct answer.”
Even more elegantly, the onetime ironmonger designed a Y-shaped lever to control the steam entering the engine itself. The lever stayed balanced on the trunk of the “Y” until the piston reached the bottom of its stroke, where an attached peg pushed one of the arms, overbalancing and opening the steam valve, simultaneously destroying the vacuum and pushing the air out through the snifting valve. As the piston rose, the valve stayed open until the top of the stroke, when another peg pushed the other arm, shutting the valve during the complete working stroke.
The Most Powerful Idea in the World Page 5