Simple horizontal engine

A horizontal engine lays the cylinder on its side and joins the piston to the crank by one straight rod, with a single sliding valve to work the steam. How to build and run one in the game is in the steam engines guide, and its parts are in the catalogue. This page is about the real one and how the game's follows it.

A horizontal mill engine at Quarry Bank Mill, running. Everything sits on one long cast bed: the cylinder at the far end, the crosshead and its guides along the bed, the crank in a disc, and the flywheel turning in its pit against the wall. No beam, no tower and no wall of masonry to carry a pivot.
A horizontal mill engine at Quarry Bank Mill, running. Everything sits on one long cast bed: the cylinder at the far end, the crosshead and its guides along the bed, the crank in a disc, and the flywheel turning in its pit against the wall. No beam, no tower and no wall of masonry to carry a pivot.
The form as Thurston's History shows it, an engine by Tangye, with its parts named over the plate (Thurston History p. 312, Fig. 95). Thurston calls it "simple, strong, and efficient". Frame, front cylinder head, crosshead guides and crank bearing are one casting, and the cylinder is bolted against the end of it (Thurston History p. 311). The governor is belt-driven from the shaft. Such an engine with a 10 in bore and 20 in stroke was rated by its makers at about 25 horse-power, and one of 30 in at 225 to 250 (Thurston History pp. 311, 313).
The form as Thurston's History shows it, an engine by Tangye, with its parts named over the plate (Thurston History p. 312, Fig. 95). Thurston calls it "simple, strong, and efficient". Frame, front cylinder head, crosshead guides and crank bearing are one casting, and the cylinder is bolted against the end of it (Thurston History p. 311). The governor is belt-driven from the shaft. Such an engine with a 10 in bore and 20 in stroke was rated by its makers at about 25 horse-power, and one of 30 in at 225 to 250 (Thurston History pp. 311, 313).
The game's horizontal engine in its planned drawing: the cast bed with its two box girders, the cylinder and valve chest at one end, the crosshead between its bars, the crank at the height of a line shaft, and the governor beside it., iso
The game's horizontal engine in its planned drawing: the cast bed with its two box girders, the cylinder and valve chest at one end, the crosshead between its bars, the crank at the height of a line shaft, and the governor beside it.

Why the beam went

The Watt engine needs a beam because it began as a pump: a rocking lever over a wall, chains to the rods in the shaft. Put a crank on it and the beam stays as a heavy way of turning a push into a turn. Laying the cylinder flat removes it. The piston rod ends in a crosshead that slides in straight guides, the connecting rod runs from there to the crank pin, and nothing else stands between the steam and the shaft.

The reasons the period gave were about money and speed. Fairbairn, writing of iron works, says the horizontal non-condensing engine "from its compact form and convenience of handling, is admirably adapted for giving motion to the machinery of iron works", that it is "superior to the beam engine" there, and "much cheaper in its original cost"; and "all its parts being fixed upon a large bed plate", it "requires a comparatively small amount of masonry" (Fairbairn pt. II pp. 257-258). In 1870 an article on mill engines puts it the same way: less first cost, less expensive foundations and engine house, "a greatly higher speed" and, if properly built, "at least equally durable". It adds that "mill-owners had a strong prejudice against horizontal engines", now "rapidly dying out" (Van Nostrand p. 34).

A horizontal rolling-mill engine of about 1890 from the Atlas Forge, Bolton, where wrought iron was last made in 1975, now at Blists Hill: it drives a mill that rolls hot blooms into bar. Cylinder cover, guides and bearings are all bolted to the one painted bed. Thurston says that a forward-running engine, which turns the top of the wheel away from the cylinder, is the one to prefer, because the thrust of the connecting rod then falls on the lower guides, "always stiffer and better lubricated than the upper" (Thurston Manual p. 92).
A horizontal rolling-mill engine of about 1890 from the Atlas Forge, Bolton, where wrought iron was last made in 1975, now at Blists Hill: it drives a mill that rolls hot blooms into bar. Cylinder cover, guides and bearings are all bolted to the one painted bed. Thurston says that a forward-running engine, which turns the top of the wheel away from the cylinder, is the one to prefer, because the thrust of the connecting rod then falls on the lower guides, "always stiffer and better lubricated than the upper" (Thurston Manual p. 92).

By the end of the period the question was closed. "Where much power is required, the stationary engine is usually an horizontal direct-acting engine" (Thurston History p. 311). The modern mill engine, Thurston says in his manual, "is commonly a direct-acting, horizontal engine", at least for moderate and large powers (Thurston Manual p. 82). Horizontal direct action had also gone to sea early: the "Charlotte Dundas" of 1801-2 had a paddle wheel near the stern driven by "a direct acting horizontal engine, with a connecting rod and crank" (Rankine pp. xxiv-xxv).

The independent frame came from smaller work. Murdoch and Southern, meeting a demand for engines of 2 to 8 horse-power, began to build ones "all the parts of which are supported one by another", standing "quite independently of the building" (Farey p. 677). A bed plate is that idea made into one iron casting.

The slide valve

A slide-valve cylinder, with the parts named over the plate of 1913; the engine here stands upright, and a horizontal cylinder is this drawing laid on its side. The chest is full of boiler steam. The valve slides up and down on a flat face over three openings: the port to one end of the cylinder, the exhaust in the middle, and the port to the other end. Its hollow underside joins whichever port it covers to the exhaust, so steam goes into one end of the cylinder while the other end empties.
A slide-valve cylinder, with the parts named over the plate of 1913; the engine here stands upright, and a horizontal cylinder is this drawing laid on its side. The chest is full of boiler steam. The valve slides up and down on a flat face over three openings: the port to one end of the cylinder, the exhaust in the middle, and the port to the other end. Its hollow underside joins whichever port it covers to the exhaust, so steam goes into one end of the cylinder while the other end empties.

Watt's engine has four valves and the working gear to open them in pairs. The slide does it with one moving part driven by an eccentric on the shaft. Farey credits Murdoch's patent of 1799: the sliding valve "produces the distribution of the steam, to and from the cylinder, in a more simple manner than the four valves and working gear of Mr. Watt's original engine", and "is now universally adopted" (Farey p. 677). Thurston names the common three-ported slide "invented by Murdoch, while with Watt, about 1799" (Thurston Manual p. 41).

A plain slide opens and shuts the port at set moments. To stop the steam before the end of the stroke, so that it goes on pushing by its own expansion, the valve is made wider than the port: the "extension of the face is called lap or cover" (Bourne p. 184). Bourne gives the proportions: with an 18-inch valve stroke, 4 1/2 inches of lap on the steam side cuts the steam off with a quarter of the stroke still to go (Bourne p. 189). Boulton and Watt put enough lap on their valves to cut off at two-thirds of the stroke as the least expansion (Bourne p. 376).

Expansion is why this matters. Steam stopped at half stroke does 1.7 times the work of the same steam let in for the whole stroke, at a quarter stroke 2.4 times (Bourne p. 183). A fixed lap, though, cannot follow the load, and what the governor does about that is the plain engine's weak point. On the small horizontal engine with a separate cut-off valve, Thurston counts it a disadvantage that setting the cut-off by the governor means "the heavy work which it throws upon the governor"; the best examples set the expansion valve by a screw and regulated through the throttle (Thurston History p. 311). Rankine prefers the other course, cutting off earlier rather than narrowing the opening and lowering the pressure by wire-drawing: "The former method makes the engine more economical, the latter less" (Rankine p. 417).

Why the numbers are what they are, drawn for this site from the game's planned engine and Bourne's rule (Bourne p. 183): steam at 64.7 psi absolute, 50 on the gauge, in a cylinder of 0.3068 square metres by 0.625 m at 73 turns a minute. The area under each curve above the back-pressure line is the work of one stroke. Cut off at 1/4 and the steam does most of its work on expansion, for 0.267 kg/s of steam against 0.712 at 2/3. Blowing it into a jet condenser instead of the air lowers the floor of the diagram.
Why the numbers are what they are, drawn for this site from the game's planned engine and Bourne's rule (Bourne p. 183): steam at 64.7 psi absolute, 50 on the gauge, in a cylinder of 0.3068 square metres by 0.625 m at 73 turns a minute. The area under each curve above the back-pressure line is the work of one stroke. Cut off at 1/4 and the steam does most of its work on expansion, for 0.267 kg/s of steam against 0.712 at 2/3. Blowing it into a jet condenser instead of the air lowers the floor of the diagram.

Where this engine stands against the Watt

The planned rule is the same arithmetic period engineers used: boiler pressure, Bourne's law of expansion, a real back pressure, a stroke at the piston speed Fairbairn gives. Lying down adds no efficiency by itself. The gain over the Watt comes from using the boiler's pressure and cutting the steam off early, and the design rates it at about 1.85 times the Watt's work for each kilogram of steam.

In the plan a single Cornish boiler open makes 0.264 kg/s, and the engine at its rated cut-off wants 0.267. It carries the engine to within 1% of full working, at 55.7 kW vented. Most of what sets this engine apart in play is that it starts without a condenser. A jet condenser on its exhaust is a choice the player can make later: it lifts the same engine to 94.8 kW.

Period basis, and what is the game's own

  • Period: the horizontal direct-acting engine on a large iron bed, cylinder at one end and crank at the other, with a slide valve and a governor (Fairbairn pt. II p. 258, Thurston History p. 311). The game draws a cast bed, cylinder, valve chest with a slide valve, crosshead on bars, crank, flywheel and governor.
  • Period: Fairbairn's iron-works engine is non-condensing (Fairbairn pt. II p. 257). The game plans it to run vented, and a condenser on the exhaust is optional.
  • Period: the best small horizontal engines regulated through the throttle and set the expansion by hand (Thurston History p. 311), and a fixed lap gives a fixed cut-off (Bourne p. 189). The game plans the governor to set the cut-off between 1/4 and 2/3, and draws the governor working a throttle. A plain slide valve could not do that: the cut-off range is the game's own, for the sake of one rating rule.
  • Period: cylinder condensation cost a good part of the steam, near 37% in one engine (Bourne p. 367). The game's own: the plan leaves it out.
  • The game's own: today the block can be placed and shows its shape, but does not run, and it has no recipe. Every rating here is planned.
  • The game's own: the base game has no steam engine. See vanilla against reality.
  • The game's own: an overloaded engine labours and stalls and does not cut out, and a governor droop of 4% sets how fast it falls off above its rated speed.

The numbers

Engines keep period figures (see the scaling page), drawn at true size (scale and units). Every value below is planned: the engine does not run today. Fuller context for all the engines is on the ratings page.

In the gamePeriod figureHow it was scaledGame value
Piston speed300 ft a minute (Fairbairn pt. II p. 101)The true value on the drawn 10-voxel strokePlanned 73 rpm, 7.7 rad/s
Bore and stroke10 in by 20 in to 30 in bore (Thurston History pp. 311, 313)Whole voxels at true sizePlanned 0.625 m bore, 0.625 m stroke
Steam49 lb on Allen's engine (Bourne p. 356)Round 50 psi, 3.4 atm gaugePlanned 50 psi; it breaks above 100 psi, the game's own limit
Cut-off1/4 gives 2.4 times the steam's work (Bourne p. 183)Period rule, the rated settingPlanned rated 1/4, latest 2/3
Back pressure18 psi absolute open, 4 psi condensing (Rankine p. 409)True valuePlanned the same
Friction12 to 15% of indicated power (Bourne p. 367)15%Planned indicated 66.4 kW at 1/4
Useful power, cut-off 1/4 ventedNo test of this engine; the rule aboveMean pressure x piston x stroke x strokes, less frictionPlanned 56.4 kW, 75.7 hp, 0.267 kg/s
Useful power, cut-off 1/4 condensingSameSamePlanned 94.8 kW
Useful power, cut-off 2/3 ventedSameSamePlanned 117 kW, 0.712 kg/s
TorqueNone givenPower over speedPlanned 7.3, 12.3 and 15.2 kN*m for the three settings above
Governor droopNone givenThe game's ownPlanned 4%

Open, and what you can tell us

The books at hand do not say why mill-owners distrusted the horizontal cylinder, only that they did (Van Nostrand p. 34). They also do not name the first stationary horizontal engine or date its adoption. And they give two dates for the sliding valve: Thurston has Murdoch using "D-slide valves" on his oscillating engine of 1785 (Thurston History p. 133), and his manual and Farey both date the three-ported valve to 1799, Farey by its patent. If you know a source that says what the objection was, or who first built one for a mill, we would like to read it.

Picture credits