Water wheels

A mill's water wheel lowers water from the head race to the tail race and takes the work out on the way down. The iron breast wheel is the form that machine had reached by the middle of the nineteenth century, and the game's first prime mover for iron shafting. How its shaft network behaves in the game is in the shaft power guide. This page is about the real wheel, why it ended up in iron with the water laid on at the side, and how the game's follows it.

An iron wheel by E. and H. Roberts at Deanshanger Iron Works: the arms have shrunk to thin rods. A wheel like this hangs from its axle rather than standing on its arms, and each rod only pulls.
An iron wheel by E. and H. Roberts at Deanshanger Iron Works: the arms have shrunk to thin rods. A wheel like this hangs from its axle rather than standing on its arms, and each rod only pulls.
The wheel at Helmshore Mills, looking up at the rim: a toothed ring runs round the rim, and tension rods run between the heavier arms. The power leaves at the teeth, not through the axle.
The wheel at Helmshore Mills, looking up at the rim: a toothed ring runs round the rim, and tension rods run between the heavier arms. The power leaves at the teeth, not through the axle.
A breast wheel in section, from a paper of Fairbairn's (Rankine pp. 178-179, Fig. 60). It is 50 ft across on a fall of 48 ft. Water passes over the top edge of the sliding shuttle A into the buckets, is held in them by the masonry breast from C to F, and drops away into the tail race at F, its fall spent. The cogs of ring B take the power off at the rim.
A breast wheel in section, from a paper of Fairbairn's (Rankine pp. 178-179, Fig. 60). It is 50 ft across on a fall of 48 ft. Water passes over the top edge of the sliding shuttle A into the buckets, is held in them by the masonry breast from C to F, and drops away into the tail race at F, its fall spent. The cogs of ring B take the power off at the rim.
The game's normal wheel in its planned drawing, 5 m across: cast shrouds in segments, thin wrought arms and braces on a cast axle, the masonry breast curving under the wheel, and the pentrough at axle height with the handwheel that works its sluice. The toothed ring on the shroud's inner face drives a pinion in the slot in the near wall., iso
The game's normal wheel in its planned drawing, 5 m across: cast shrouds in segments, thin wrought arms and braces on a cast axle, the masonry breast curving under the wheel, and the pentrough at axle height with the handwheel that works its sluice. The toothed ring on the shroud's inner face drives a pinion in the slot in the near wall.
The large wheel in its planned drawing: 7 m across and three blocks broad, its pentrough a row above the axle, a higher fall and room for up to three streams. A broad iron wheel needs a middle ring and bucket stays to stay square, as Fairbairn's did (Fairbairn Mills I pp. 122-124)., iso
The large wheel in its planned drawing: 7 m across and three blocks broad, its pentrough a row above the axle, a higher fall and room for up to three streams. A broad iron wheel needs a middle ring and bucket stays to stay square, as Fairbairn's did (Fairbairn Mills I pp. 122-124).

How the real one worked, and why it was built so

Before iron, undershot wheels were frequently employed, with straight floats struck by water rushing under a shuttle. This "was the invariable practice down to Smeaton's time even", the principle being to use the impulse of the stream and not its weight (Fairbairn Mills I p. 151).

The old undershot wheel, Fig. 123 (Fairbairn Mills I p. 151). Straight radial floats, the water let in low under the shuttle on the left. The work is done by the blow of the stream, not by its weight.
The old undershot wheel, Fig. 123 (Fairbairn Mills I p. 151). Straight radial floats, the water let in low under the shuttle on the left. The work is done by the blow of the stream, not by its weight.

An overshot wheel works the other way, by weight. The stream is led over the summit, and the water retained in the buckets "continuously depresses the loaded side of the wheel" (Fairbairn Mills I p. 124).

The overshot wheel as built seventy to eighty years before Fairbairn wrote, Fig. 112 (Fairbairn Mills I pp. 124-125). Water comes from the dam along trough c, past sluice d, and over the top. Wooden arms b are bolted to a flanged cast-iron shaft a, and earlier wheels had a shaft of wood as well (Fairbairn Mills I p. 125). Buckets begin to empty well above the bottom, so the last part of the fall does nothing.
The overshot wheel as built seventy to eighty years before Fairbairn wrote, Fig. 112 (Fairbairn Mills I pp. 124-125). Water comes from the dam along trough c, past sluice d, and over the top. Wooden arms b are bolted to a flanged cast-iron shaft a, and earlier wheels had a shaft of wood as well (Fairbairn Mills I p. 125). Buckets begin to empty well above the bottom, so the last part of the fall does nothing.

By Fairbairn's day "the breast wheel has taken precedence of the overshot wheel" (Fairbairn Mills I p. 116). Water laid on below the crest, at about 30 degrees from the vertical, needs a far cheaper pentrough than one carried over the top. A mill drew its reservoir down three or four feet in the day and refilled it overnight; a breast wheel takes the water at whatever height it stands, while an overshot wheel cannot work a falling supply without loss of power (Fairbairn Mills I p. 116). The name was already slipping by then, and "high-breast is perhaps a more correct designation" for most wheels called overshot (Fairbairn Mills I p. 125).

The breast itself is a curved trough of masonry, close to the wheel, which keeps water from spilling out of buckets as they tilt: "little or no space" between it and the wheel (Fairbairn Mills I p. 139), about 0.4 in in Rankine's description (Rankine p. 179). It is what makes the difference between the two names (Rankine p. 177). The tail race is sunk two to two and a half feet beneath the wheel, so the buckets can clear (Fairbairn Mills I p. 128), since a wheel standing in its own back water loses about a quarter of its efficiency (Rankine p. 176).

What each form returned, from Fairbairn's own figures (pages on the chart). A well-made iron wheel is near the top of what any prime mover of the time could do, and Fairbairn doubted that turbines beat "a well-constructed water wheel" (Fairbairn Mills I p. 177). Rankine takes the part of the fall that acts by weight at 0.8 (Rankine p. 177). The game's figure, in red, is the 78% trial times its gearing loss.
What each form returned, from Fairbairn's own figures (pages on the chart). A well-made iron wheel is near the top of what any prime mover of the time could do, and Fairbairn doubted that turbines beat "a well-constructed water wheel" (Fairbairn Mills I p. 177). Rankine takes the part of the fall that acts by weight at 0.8 (Rankine p. 177). The game's figure, in red, is the 78% trial times its gearing loss.

Efficiency was only half of why mills turned to iron. The other half was how much one wheel had to do. Some mills had first had a separate wheel for every machine; "one large water-wheel" with the machines around it concentrated the power, led to the factory building, and needed shafting (Fairbairn Mills II p. 50). Three changes made the large iron wheel work.

The first was to hang the wheel. Instead of stiff arms that act as columns, T. C. Hewes of Manchester supported the rim from the axle by thin rods of wrought iron, from a suggestion of William Strutt (Fairbairn Mills I p. 120, Fairbairn Useful p. 229). Arms were 2 to 2.5 in across on almost all wheels (Fairbairn Mills I p. 122). The rods only pull, those pointing down at the moment carrying the weight, and Rankine says the change "has greatly diminished the weight, cost, and friction of large water wheels" (Rankine p. 178).

The second was to take the power off at the rim. Driven from the axle, the wheel's whole torque goes through the arms and shaft. With a toothed ring cast on the shroud meshing a pinion, "the arms, braces, and axle have only to sustain the weight of the wheel" (Fairbairn Mills I p. 141). The pinion must sit on the loaded side, below the end of the horizontal diameter, or a wheel hung on two-inch rods "would twist round upon the axle" (Fairbairn Mills I p. 140). Rim gearing also raises the speed at once: Catrine's ring of 48 ft 6 in drove a pinion of 5 ft 6 in (Fairbairn Mills I p. 132).

A 40 ft iron wheel, 20 ft broad, built by Fairbairn for Gefle in Sweden (Fairbairn Useful pp. 229-230, Fig. 35). The arms c and braces b hang the wheel from the cast axle a, and the internal ring d, cast in segments, drives a pinion. The spur wheel f and pinion g, then bevels k and h, carry the power on to the upright shaft i and the floors of the mill.
A 40 ft iron wheel, 20 ft broad, built by Fairbairn for Gefle in Sweden (Fairbairn Useful pp. 229-230, Fig. 35). The arms c and braces b hang the wheel from the cast axle a, and the internal ring d, cast in segments, drives a pinion. The spur wheel f and pinion g, then bevels k and h, carry the power on to the upright shaft i and the floors of the mill.

The third was to let the air out. Water falling into a closed iron bucket traps its air, and in flood a wheel at Linwood in 1828 threw spray six to eight feet above the entry. Vent openings cut in the sole gave "an increase of power of nearly one-third" (Fairbairn Mills I p. 136). The first wheel built for it was at Wilmslow the same year (Fairbairn Mills I p. 137).

Two ways to vent a bucket (Rankine p. 180, Figs. 61 and 62). In Fig. 61 a closed sole plate has an air-passage behind each bucket with a hole into it, in Fig. 62 the sole is gone and each bucket opens behind the next. Rankine credits the present mode to Fairbairn (Rankine p. 180).
Two ways to vent a bucket (Rankine p. 180, Figs. 61 and 62). In Fig. 61 a closed sole plate has an air-passage behind each bucket with a hole into it, in Fig. 62 the sole is gone and each bucket opens behind the next. Rankine credits the present mode to Fairbairn (Rankine p. 180).

How the water was held and the wheel kept to speed

A wheel was not started by opening the sluice wide. The shuttle was a closely fitting cast-iron arc with a door that slides, so "a thin sheet of water" flows over its upper edge (Fairbairn Mills I p. 143), and its motion "should in every case be slow", or the water surges and the wheel runs "first faster and then slower" (Fairbairn Mills I p. 147). A ball governor could work the shuttle (Fairbairn Mills I pp. 143-147), and on the Cleator wheel it raised and lowered it (Fairbairn Mills I p. 140). Buckets ran a third to a half full (Fairbairn Mills I p. 147), never more than two-thirds on the average supply (Fairbairn Mills I pp. 126-127), and Fairbairn kept the rim between 4 and 6 ft/s (Fairbairn Mills I p. 141).

All of this tied a mill to its stream. Catrine's two wheels of 50 ft ran on 48 ft of fall with "auxiliary steam-power" kept for dry seasons (Fairbairn Mills I p. 130). By the time Fairbairn wrote "nineteen-twentieths" of the mills were driven by steam (Fairbairn Useful p. 229), and it was the engine, not a better wheel, that freed mills from the site (Fairbairn Mills II p. 51). In the game the wheel gives way to steam the same way, through the Cornish boiler.

Period basis, and what is the game's own

  • Period: the wheel is laid on at the side, on a fall of a half to three-quarters of its diameter (Fairbairn Mills I p. 140). The game's wheels take water at the axle on the normal and a row higher on the large, 0.6 and 0.71 of the diameter.
  • Period: a cast axle on plummer blocks bolted to the masonry, segmented cast shrouds, wrought sole and buckets, thin wrought arms, a pentrough with a shuttle, a tail race, and the power off the rim (Fairbairn Mills I pp. 116-124, 141-143). The game draws all of these, and builds them from castings, plate and bar in stages.
  • Period: the pinion on the loaded side below the end of the diameter (Fairbairn Mills I p. 140). The game's pinion sits there.
  • Period: a wheel in a rising tail water loses power (Rankine p. 176), and the race has to be kept open (Fairbairn Mills I p. 128). The game keeps only the extreme of this: a tail race with its mouth blocked floods the pit and the wheel takes no water. The partial case is not modelled.
  • Period: the governor on the shuttle (Fairbairn Mills I pp. 143-147). The game has no governor, because a stream that never runs short would give it nothing to save. The shuttle is the player's job, in quarter turns of the handwheel, and the slow opening is kept: on three streams the large wheel has to be brought up on half the sluice or its starting torque passes what the cast main can carry.
  • The game's own: water is counted in whole streams. A stream is vanilla's rapid water, a rivulet running from a single spring, and a spring shared by several wheels gives each its share. There is no flood, drought, mill pond or back water, so a wheel's output depends only on the streams and the fall.
  • The game's own: the torque curve. The books do not give a curve of wheel speed against load. The game holds power constant around the rated speed, doubles the torque at a stall because a slowed wheel can at most fill its buckets to twice their working level, and lets it run out at twice the rated speed, where the rim passes the water. The two-to-one comes from the fill rule above; the shape is the game's.
  • The game's own: the base game's wheel gives a speed with no storage, and it needs rapid water at its paddles. See vanilla against reality. The iron wheel keeps that water source, and adds a fall built by the player, shaft torque, and a coupling that goes iron to iron with no timber axle in the line, as the iron shafting will not run off a wooden one.

The numbers

These are period ratios and formulas applied at the real wheel's size (the scaling page, scale and units). Every power and speed figure is a planned rating; each planned value is marked. The wheel's size follows from the blocks, and the water from one stream.

In the gamePeriod figureHow it was scaledGame value
DiameterFixed from the fall (Fairbairn Mills I p. 147)Whole blocks at true size5 m (16 ft 5 in); large 7 m (23 ft)
FallHalf to three-fourths of the diameter (Fairbairn Mills I p. 140)Whole blocks3 m; large 5 m
One stream of water2,150 cu ft a minute at the start at Catrine, about 1,015 L/s (Fairbairn Mills I p. 131)A rivulet a block wide, 0.3 m deep at 0.8 m/s, about a quarter of Catrine'sPlanned 250 L/s
Water's power1 hp for 8.88 cu ft/s falling 1 ft (Fairbairn Mills I p. 113), the same as weight times fallApplied as it standsPlanned 7.36 kW normal; large 12.3 kW on one stream, 36.8 kW on three
Efficiency75%, iron high-breast (Fairbairn Mills I p. 126); "nearly 78" ventilated, large trials (Fairbairn Mills I p. 177)The trial value, 0.98 taken off for the rim gearPlanned 76.4%
Power at the shaftCatrine pair, 240 hp from a 48 ft fall (Fairbairn Mills I pp. 129-130)Water power x efficiencyPlanned 5.62 kW (7.5 hp); large 9.37 kW on one stream, 28.1 kW (37.7 hp) on three
Rim speed4 to 6 ft/s (Fairbairn Mills I p. 141)Taken as a design speedPlanned 1.4 m/s (4.6 ft/s)
Bucket fillingA third to a half (Fairbairn Mills I p. 147)Width and depth chosen to suit one streamAbout 0.34 full at rated speed, planned
Buckets40 to 60 on a 20 ft wheel (Fairbairn Mills I p. 142)Counted from the drawing36; large 48, planned
Shroud depth22 in on Cleator's 20 ft wheel (Fairbairn Mills I p. 140)Taken from the book for the large; the normal's is the game's own drawing25 in; large 22 in, planned
Rim gear ratio8.8 : 1 at Catrine, 48 ft 6 in to 5 ft 6 in (Fairbairn Mills I p. 132)Teeth set by the drawing4 : 1; large 5.6 : 1, planned
Stall torqueBuckets at most two-thirds, working at a third (Fairbairn Mills I pp. 126-127, 147)Twice the rated torquePlanned 5.02 kNm; large on three streams 25.1 kNm, so it starts on the half sluice
Rotating ironwork34 tons on a wheel 18 ft by 20 ft broad (Fairbairn Mills I p. 119)Each part's drawn volume at iron's density18.8 t; large 47.8 t

Open, and what you can tell us

Fairbairn gives no curve of a wheel's speed against its load. A heavy wheel held its speed to a governor, so what it would have done without one, slowed by a sudden load, is reconstructed in the game from how full its buckets could get. If you know a period trial that logged speed and torque, or a modern measurement of a preserved wheel, we would like to read it. The game also leaves back water short of a blocked race, and a frozen race in winter, as things not modelled; tell us if a real mill's account shows how often either stopped the work.

Picture credits

  • Water Wheel by E and H Roberts, Deanshanger Iron Works: Chris, CC BY-SA 2.0, cropped.
  • Water wheel, Helmshore: Clem Rutter, Rochester, Kent, CC BY 3.0, cropped.
  • Fig. 60, and Figs. 61 and 62: Rankine, A Manual of the Steam Engine and Other Prime Movers, pp. 179 and 180, Library page. Fig. 60 labelled for this site.
  • Figs. 112 and 123: Fairbairn, Treatise on Mills and Millwork, Part I, pp. 124 and 151, Library page. Fig. 112 labelled for this site.
  • Fig. 35: Fairbairn, Useful Information for Engineers, p. 230, Library page. Labelled for this site.
  • Share of the water's power: drawn for this page from Fairbairn Mills I pp. 126, 154-156 and 177.
  • The game's wheels: the mod's planned drawings.