01Executive summary
Five milestones across eighty years that turn a laboratory observation about atmospheric pressure into a general source of industrial power.
Savery's engine of 1702 was the first steam pump offered commercially. Newcomen completed his in 1712 and it revolutionised mining despite wasting more than ninety-nine per cent of its fuel. Smeaton, the first man to call himself a civil engineer, took the atmospheric engine to its practical limit. Watt's separate condenser of 1769 saved three-quarters of the fuel, and his double-acting rotative engine of 1782 to 1784 turned a mine pump into the prime mover of an industrial economy.
02Newcomen: a good machine made of familiar parts
Thomas Newcomen, an ironmonger who also preached, completed his engine in 1712 after years of experiment, with John Calley providing the manual skill. Its first installation pumped water from a mine at Dudley Castle. Newcomen's contribution was an ingenious combination of existing things rather than a new principle.
- Raise steamA brewer's copper and furnace supply low-pressure steam — no new technology at all.
- Fill the cylinderSteam enters beneath a piston in a vertical cylinder of the kind used for lifting water since Roman times.
- CondenseWater is sprayed directly into the cylinder, condensing the steam and creating a partial vacuum. This is Newcomen's improvement on Savery, who cooled the vessel from outside.
- Atmospheric power strokeAtmospheric pressure above forces the piston down. The work is done by the air, not by the steam.
- ReturnFresh steam breaks the vacuum and the weight at the far end of the rocking beam raises the piston again.
The beam is chain-connected at both ends, running over circular arcs so the rods rise and fall vertically as it rocks. A valve gear driven from the engine's own motion replaced hand operation and raised the cycle rate from perhaps six or eight strokes a minute to fifteen or sixteen.
The tale that Newcomen's valve boy Humphrey Potter rigged a cord so the beam would work the valves while he went fishing is diverting and has no foundation. It survives because it is a good story, which is a reasonable description of a great deal of engineering folklore. The automatic valve gear was designed.
What it cost and what it was worth
The Newcomen engine was slow, ran at or below atmospheric pressure, made no use of the expansive force of steam, and wasted well over ninety-nine per cent of the energy in its fuel. Critics said it took an iron mine to build one and a coal mine to keep it going. Its cylinder had to be cooled to condense and immediately reheated by the next charge of steam, within a few seconds, which is where almost all the heat went.
And it was still transformative. At Fresnes in France by 1739 an engine with a 760 millimetre cylinder and a 2.7 metre stroke, working forty-eight hours a week with very little attendance, replaced thirty men and twenty horses working in shifts around the clock. A Swedish observer recorded a machine at Dannemora doing the work of 528 horses in a day. Engines went to Hungary in 1722 and to Passy in 1726 to raise water from the Seine for Paris.
An engine wasting ninety-nine per cent of its fuel revolutionised an industry, because the comparison was not against a better engine but against thirty men, twenty horses and mines that could not be worked at all below the water table. Thermal efficiency became the governing metric only once engines competed with each other. Judging a technology by the metric that matters later rather than the one that matters now is a persistent error, and Newcomen is the clearest case of it in this set.
03Smeaton: the first civil engineer
John Smeaton was five when Newcomen died. He made a working model of the engine as a boy, abandoned his father's profession of law, set out as an instrument maker, and by twenty-nine had established himself broadly in engineering. He was the first man to describe himself as a civil engineer, as distinct from a military one — which is the moment the modern profession acquires its name.
He is better known for the Eddystone Lighthouse, his canals and his investigations of cements and water power, but he also built Newcomen engines at the Carron ironworks and improved their mechanical efficiency substantially through more precise cylinder boring, better proportioning of parts and general quality of shopwork. He is generally credited with taking the atmospheric engine to the best practical performance of which the type was capable.
Systematic experiment on power
Smeaton investigated water power experimentally rather than by precedent, and estimated the mechanical effect a horse could produce at about 22,900 foot-pounds per minute. Quantifying performance is his real contribution.
Improvement without invention
He added no new principle to the Newcomen engine. Precision, proportion and workmanship alone produced a large gain — a reminder that execution quality is a legitimate and often underrated source of performance.
One installation makes the point about scale. A Smeaton engine sent to Kronstadt in 1775 to pump out Catherine II's dry docks replaced two windmills 30 metres high installed by Dutch engineers in 1719. The windmills were said to need a year for the job. The engine did it in about two weeks.
04Watt: four ideas, then rotation
James Watt, like Smeaton, began as an instrument maker. The idea came to him while repairing a model Newcomen engine in the laboratory at Glasgow where Joseph Black was lecturing on heat. Watt saw that the waste was structural: the cylinder walls had to be cooled to condense and reheated every cycle.
His answer was to condense somewhere else. A separate chamber, connected to but apart from the cylinder, does the condensing; the cylinder itself is kept hot inside a steam jacket. That alone saved three-quarters of the fuel. Between 1765 and his first patent in 1769 he added two more things: an air pump to maintain the vacuum in the condenser by removing water and air, and a closed upper cylinder end with a stuffing box around the piston rod, using steam rather than air to push the piston down.
| Contribution | Date | Problem addressed |
|---|---|---|
| Separate condenser | 1769 patent | Heat lost cooling and reheating the cylinder every cycle |
| Steam jacket | 1769 | Cylinder wall temperature falling between strokes |
| Air pump | 1769 | Condensate and air accumulating and spoiling the vacuum |
| Steam above the piston | 1769 | Reliance on atmospheric pressure alone as the working force |
| Double-acting engine | 1782 | Only half the cycle doing useful work |
| Parallel motion | 1784 | A rod now in compression as well as tension could no longer use a chain |
| Centrifugal governor | — | Speed varying with load; adapted from flour mills |
| Indicator diagram | — | No way to see cylinder pressure through the stroke |
| Expansive cut-off | 1782 patent | Steam admitted for the whole stroke rather than expanded |
Two entries deserve comment. The parallel motion — an arrangement of links keeping the piston rod in near-straight alignment as it drives the beam — was the invention Watt said he was proudest of, and it exists only because double action put the rod in compression, so the chain that Newcomen used would buckle. The expansive cut-off he patented and never really exploited, because it pays only at pressures higher than he was willing to run.
Matthew Boulton, who understood the market, pushed Watt toward a rotative engine, observing that London, Manchester and Birmingham were "steam mill mad". The obvious route was a crank and flywheel. But James Pickard had patented the crank as a method of applying steam engines to turning wheels, and Watt — whose own patents were exceedingly broad and whose monopoly he did not want tested — chose to avoid the fight, using an awkward and noisy sun-and-planet gear until Pickard's patent expired in 1794. Watt's own comment was that applying a patent to the crank was "like taking a knife to cut cheese which had been made to cut bread". Notably, nobody at the time understood that the flywheel smooths the engine's rate by storing and releasing energy, because the concept of energy itself did not exist until the middle of the nineteenth century.
05What the engine did and did not do for people
It is worth resisting a comfortable story here. Newcomen's engine freed thousands of men and horses from pumping water out of mines. Those men were not released into leisure; they moved to other work, often no more pleasant, and some were out of work entirely for a period, even though reopening flooded mines soon increased demand for labour.
Over a longer span, with population growth and the advance of medicine, demand for labour rose and living standards rose markedly, and engineers who devised the improved instruments of production deserve real credit for that. But the immediate profits of the period accrued not to the workers but to the relatively few who accumulated and risked the capital and owned the tools. Manufactured goods became more plentiful and cheaper; it does not follow that those making them received a larger share of what their effort produced. Both halves of that are true and an account giving only one of them is not honest.
06Takeaways for current practice
- Judge a technology against what it replaces, not against its successor. A one-per-cent-efficient engine beat thirty men and twenty horses.
- Look for a loss that is structural rather than incidental. Watt did not make the cylinder better; he moved the condensation out of it.
- Execution quality is a legitimate source of performance. Smeaton added no principle and gained a great deal.
- A change in one place forces a change elsewhere. Double action put the rod in compression, which is the entire reason parallel motion exists.
- Patent position shapes engineering decisions. The sun-and-planet gear was worse than a crank and was used for a decade regardless.