01Executive summary
Three milestones that made the engine small enough to move, by doing the one thing James Watt refused to do.
Watt knew high pressure offered advantages and would not use it, for fear of explosion. Richard Trevithick in Britain, from about 1798, and Oliver Evans in Philadelphia, from 1801, both built non-condensing engines exhausting to atmosphere at pressures Watt considered reckless. Trevithick then removed the beam entirely, producing the first direct-acting engine. The protective devices that made high pressure survivable — an independent safety valve and a fusible plug — arrive with it.
02What high pressure actually buys
An atmospheric engine can never exert more than atmospheric pressure on its piston, so the only way to more power is a larger cylinder. By 1765 there were engines with cylinders 1.8 metres across, over 3 metres long and weighing nearly seven tonnes. That is the end of the road: the machine grows without limit and becomes progressively more difficult to build, align and keep tight.
Raising the pressure changes the arithmetic. Force is pressure times area, so more pressure means more work from the same cylinder — or the same work from a much smaller one. Three consequences follow, and the third is the one that mattered.
The condenser can go
If steam pushes hard enough on its own, there is no need to create a vacuum on the other side. Exhausting to atmosphere removes the condenser, the air pump, the cooling water supply and everything that goes with them.
Expansion becomes worth using
Watt patented an expansive cut-off and never really used it, because expansion only pays at pressures above those he would allow. High pressure makes the most thermodynamically valuable feature of a steam engine economic.
The engine becomes portable
Smaller cylinder, no condenser, no cooling water, no beam. The whole machine becomes compact and self-contained — which is the precondition for putting it on wheels or in a hull. Every development in the following part depends on this.
Boilers can now kill people
Stored energy in a pressurised vessel is real and the failure is violent. High pressure creates the discipline of boiler safety, and the fatal accidents came before the discipline.
03Evans in Philadelphia
Oliver Evans, apprenticed to a wagon maker around 1772, began thinking about propelling carriages without animals. By his own account the idea of the elastic power of steam came from watching the blacksmith's boys fill a gun barrel with water, ram it tight and put the breech in the fire: the crack, he said, was as if it had been loaded with powder.
He petitioned Pennsylvania in 1786 for exclusive rights to improvements in flour mills and steam wagons. He was granted the flour mill rights in 1787 — for a genuinely remarkable system of continuous grain handling and milling — and the legislators ignored the steam wagons entirely. His representations, he said, made them think him insane.
By 1801 he had built, for $3,700 which was everything he could command, a small non-condensing high-pressure engine: a 150 millimetre cylinder, a 460 millimetre stroke, a wooden flywheel, running at around thirty revolutions a minute at some 3.5 bar and exhausting to the air. To get that pressure he encased a copper boiler in wood bound with iron hoops and ran a horizontal flue lengthwise through the centre so the fire reached the water more effectively.
Evans set the engine upright with the piston rod driving directly to a beam above, and put the beam's fulcrum at one end rather than the centre, supported on a rocking column. The beam then aligns itself with the piston rod, so nothing like Watt's parallel motion linkage is needed. It is a neat piece of kinematic reasoning: rather than adding a mechanism to force straight-line motion, change the geometry so the motion is naturally correct. The arrangement became known as the grasshopper type and stayed in use for a long time.
Benjamin Latrobe reported to the American Philosophical Society in 1803 that there were only five steam engines of any considerable power at work in America, and mentioned Evans's engine almost in passing as a small experimental machine grinding plaster. Evans took him to task for it publicly, and by 1812 was reporting engines of his in Philadelphia, on the Mississippi, at Pittsburgh, in Ohio, Kentucky and Connecticut, driving sawmills, milling grain, rolling iron and running a cloth factory. By his death in 1819 there were fifty in service.
Not everything worked. His 1815 installation for the Fairmount waterworks in Philadelphia — a 500 millimetre cylinder, 1.5 metre stroke and four boilers designed for around 14 bar — was overambitious and so costly to run that it was replaced by water power in 1822. Ambition beyond what the supporting technology could sustain is a recognisable failure mode and it deserves recording alongside the successes.
04Trevithick and the direct-acting engine
Richard Trevithick, a Cornishman with a reputation as a wrestler and strong man, ignored Watt's alarm about pressure and simultaneously wanted a way around Watt's patent. He built a non-condensing "puffer" around 1798, before Evans built his, and then went further than Evans did.
- Cylinder inside the boilerPlacing the cylinder upright within the boiler makes the boiler itself the steam jacket, keeping the cylinder hot with no separate provision.
- Crosshead instead of beamThe piston rod drives up to a transverse crosshead, which acts downward through connecting rods on either side onto a crankshaft beneath the boiler. The rocking beam is gone.
- Self-contained unitThe whole arrangement is compact and, decisively, portable as a single assembly rather than a building containing machinery.
- Direct actingLaying the cylinder horizontal and connecting the piston to the crank by a connecting rod gives the first direct-acting engine in history — the layout of essentially every reciprocating engine since.
From there, the next step was a locomotive, and that belongs to the last part of this series.
An explosion occurred when an operator tied down the safety valve on one of Trevithick's boilers. His response was to fit two valves, one of them beyond the operator's control. That is a design principle rather than a repair: a protective device that can be defeated by the person under production pressure is not a protective device. He also used a fusible plug — a piece of lead set into the boiler plate which melts if the water level falls far enough to expose it, venting the boiler deliberately before the plate fails. It works because water in a boiler cannot exceed its boiling temperature at the given pressure while it remains water; once the plate is dry that protection disappears and the metal temperature climbs. Both devices are direct ancestors of modern pressure relief and trip philosophy, and both were written in response to an accident.
One further detail sets the period. Until about 1848 there were no gauges for higher pressures, so the safety valve was the only means of knowing the pressure at all above a few pounds. Engineers were running machines at pressures they could not measure, protected by a device whose setting was the measurement.
05Takeaways for current practice
- When scaling stops working, change the variable. Atmospheric engines could only grow; pressure let them shrink.
- Removing a subsystem beats improving it. No condenser means no air pump, no cooling water and no fixed installation.
- Change the geometry rather than adding a mechanism. The grasshopper beam aligns itself where Watt needed a linkage.
- A protection the operator can defeat is not a protection. Trevithick's second valve was deliberately placed beyond reach.
- Ambition beyond the supporting technology fails expensively. Fairmount ran for seven years and was replaced by a water wheel.
The modern descendants sit in AS 1210 for pressure vessels, AS 2593 for boiler operation and AS 4343 for pressure equipment hazard levels. Cited by number for orientation only — verify currency.