Fairbairn, Treatise on Mills and Millwork, Part I, water wheels (1864)

Books Treatise on Mills and Millwork, Part I

Covers W. Fairbairn, "Treatise on Mills and Millwork, Part I: On the Principles of Mechanism and on Prime Movers" (1864 edition, later than the first; 403 PDF pages), archive.org id treatiseonmills01fairgoog (Google scan of the University of Wisconsin copy, metadata title "Treatise on mills and millwork", date 1864, volume 1). The PDF has no text layer: the text was read from the archive's treatiseonmills01fairgoog_djvu.txt (OCR, noisy digits), and every table and figure number below was checked against page images rendered from the PDF. Another copy of Part I exists as 10304787bsb (1861, Munich, 171 MB); not used. Pages below are the printed folios. Offset folio to PDF page: +27 at pp. 119-123 (PDF 146 = p. 119, PDF 150 = p. 123), +35 from p. 141 (PDF 176 = p. 141, PDF 183-184 = pp. 148-149); Plates I-IV (Catrine, low-breast, Cleator wheels) sit between. The table of proportions (pp. 148-149) is printed sideways; rotate the page 90 degrees clockwise. Read in full: Section III, Chapter III, construction of water wheels (pp. 114-150), the start of Chapter IV on the undershot and Poncelet wheel (pp. 151-156), the efficiency paragraph closing Chapter V (p. 177), the horsepower formula (pp. 112-113), and the steam-engine passages on rim gearing and auxiliary engines (pp. 244-245, 250-251, 254). Skimmed: Chapter II, discharge of water (pp. 94-112, orifices, weirs, pipes). Skipped: weirs and reservoirs, turbines (pp. 158-176), water-pressure engines, the properties of steam, the rest of the engines, windmills, the index. Smeaton's own experiments are cited, not tabulated, in this volume. Purpose the period numbers for the iron breast wheel: how water enters, the breast and sluice, what an iron wheel is built from, sizes with fall and power, rim speed and how the load is held to it, and wheel beside engine. The book is a practitioner's (Fairbairn built the Catrine and Deanston wheels himself), so its figures are his practice, not a survey.

What the book says

Why a breast wheel (pp. 115-116)

  • Classes: overshot (water over the crest), breast (water below the crest, at the side), undershot (near the bottom, by gravity in the improved wheel or by impulse in the ordinary and Poncelet wheels) (p. 115).
  • "the breast wheel has taken precedence of the overshot wheel": water of high falls is laid on at about 30 degrees from the vertical diameter, the pentrough being far cheaper to carry than over the top (p. 116).
  • With a variable head the breast takes the water at the highest level of the day and wastes none; an overshot wheel cannot work the supply down without losing power. Mills draw the reservoir down "three or four feet" before the evening stop and refill overnight (p. 116).
  • The large breast wheel clears back water more easily and is hurt less by floods (p. 116).
  • Old overshot pattern, 70-80 years before: wooden arms bolted to a cast-iron shaft with broad flanges (p. 125). "overshot" now names any wheel laid on near the summit; "high-breast is perhaps a more correct designation" (p. 125).

How water enters, breast, sluice (pp. 125-128, 138-143)

  • Ordinary overshot: "the useful effect amounts to about 60 per cent."; improved iron high-breast "75 per cent." (p. 126). Spilling is cut by curved buckets, by filling them only part way, and by a close breast (p. 126).
  • Buckets "never more than two-thirds filled" on the average supply (pp. 126-127); the design rule is one-third to one-half (p. 147). Catrine ran one-third full at start and one-half full at full power (p. 131).
  • Close-fitting curved breast: of "considerable importance" at low falls, "of no value" on large wheels for high falls with small bucket openings, where it "is invariably dispensed with" (p. 127). Repeated: high-breast wheels of 25 ft diameter and up need no breast, narrow openings and lips close to the next bucket hold the water (p. 139).
  • Best breasts are masonry "and allow little or no space between them and the wheels"; warning that stones or rubbish jamming between bucket and curb "might cause disaster" (p. 139).
  • Low-breast ventilated wheel (Plate III, pp. 138-139): the wheel stands above the tail water, not down to the race floor as in old wheels; the stone breast "should break off about ten inches from the extremity of a vertical diameter". Its example is 16 ft diameter, 17 ft 6 in between shrouds, fall 5 to 8 ft (p. 152).
  • Falls: high-breast ventilated wheel is for falls "of from one-half to three-fourths of the diameter" (p. 140).
  • Tail race: floor sunk 2 to 2.5 ft below the wheel bottom and kept so to the river; bottom of the wheel 8 or 9 inches above the ordinary tail water. Tail water may leave against, across or with the rotation (p. 128).
  • Tail-race tunnel example: a wheel 24 ft diameter on a 22 ft fall was replaced by one of 32 ft on a 28 ft fall, the tail water taken a quarter mile in a tunnel (pp. 128-129).
  • The shuttle (p. 143): the pentrough front is a cast-iron arc "closely fitting the periphery of the wheel", with an opening from side to side; inclined guide plates in it keep the entering water off the sole plate and the back of the next bucket; over them a door slides up or down with the pentrough level so that "a thin sheet of water" flows over its upper edge. Racks on the back of the sluice with pinions set it. A folding sluice keeps the entry speed the same at any head (p. 136).
  • Catrine pentrough: cast-iron plates bolted together on beams, wall on one side, columns on the other (p. 143); 6 ft deep, reservoir sluice sill 16 in above the lowest overflow of the wheel sluice (p. 131). Deanston: a wrought-iron pentrough on iron columns feeding wheels on both sides (p. 135). Galleries give access (p. 130).
  • Bucket opening, measured square to the entering water: 5.5 to 8 in for high-breast, 9 to 12 in for low-breast, extremes 4 in (very high breast) and 20 in (undershot) (p. 141).
  • Opening area for 25 cu ft of bucket capacity: 5 sq ft for high falls; 8 sq ft where water enters not more than 10 degrees above the horizontal diameter. Shroud depth 2 to 2.5 times the opening (p. 142).
  • Bucket spacing on the rim 1 ft to 1 ft 6 in, low-breast wider. Counts (p. 142): 10 ft diameter 20-30; 20 ft 40-60; 30 ft 60-90; 40 ft 80-120; 50 ft 120-160; 60 ft up to 180 (lower figure illegible).

Ventilated buckets (pp. 136-140)

  • 1828, Linwood wheel for Andrew Brown: in flood each bucket "acted as a water blast", throwing spray 6 to 8 ft above the entry. Openings cut in the sole plate with small inner buckets passed the air out and gave "an increase of power of nearly one-third" (p. 136). Robison's case: wheel 14 ft by 12 ft in 3 ft of back water, three 1 in holes per bucket, power up one-fourth (pp. 136-137). Air passes a hole 30 times faster than water (p. 137).
  • First purpose-built ventilated wheel: Wilmslow, Cheshire, 1828 (p. 137). The sole plate is then dropped and the bucket plates bent and carried up to overlap, leaving a slot for the air; tubular ferrules with a rivet through tie the plates (p. 139). Back water otherwise fills the rising buckets and defeats discharge (p. 138).
  • Ventilated breast wheel efficiency "nearly 78 per cent." in large-scale trials; Smeaton's overshots 76 (p. 177).

Iron construction (pp. 116-124)

  • Parts: main axis on masonry, arms and braces, shrouding, sole plate and buckets, pentrough with sluice, tail race, internal or external spur ring and pinion (p. 116).
  • Main axle: a large cast-iron shaft on plummer blocks bolted to the wheel-house masonry, cast "with deep ribs or wings" (four ribs), a square end for a screw-jack, recesses for arm and brace ends keyed by gibs and cotters. Shown for a wheel 20 ft diameter by 22 ft wide (pp. 117-118).
  • Power from the axle puts torsion on axle and arms; with a suspension wheel taking power at the rim the axle carries weight only, "trifling" by comparison (p. 118).
  • Journal diameter, inches, for wheel diameter by breadth (p. 119; reading of the printed fractions, +-0.5 in): 20 ft wheel 7/8/10/12 for 5/10/15/20 ft broad; 30 ft 8/9/11.5/13; 40 ft 8.5/10/12.5/14.5; 50 ft 9.5/11/14/16; 15 ft 6/7/8.5/10. Bearings one and a half journal diameters long. Tredgold: d = (l w)^(1/2)/9 with w half the wheel weight. Worked example: a wheel 18 ft diameter by 20 ft broad "weighs 34 tons" and wants an 8 in gudgeon (p. 119).
  • Rigid wheels: cast-iron arms as columns, fitted by chipping strips to a cast "main centre" and bolted; or flat wrought arms riveted. Suspension wheels hang from light wrought-iron rods; T. C. Hewes of Manchester, about fifty years before, from a suggestion of William Strutt; first tried with chains, then "round bars of malleable iron" (p. 120). Early screw-and-nut ends loosened with vibration; in 1824 Fairbairn used gibs and cotters, arm ends forged square into sockets in the centre (p. 121).
  • Arms are 2 to 2.5 in diameter on almost all wheels, braces 1.25 to 2 in (p. 122; 2.5 in on the 20 ft wheel, p. 118; 2 in arms at Cleator, p. 139). Braces run from the centre on one side to the shroud on the other, alternating with arms; broad wheels take them to a cast middle ring in segments riveted to the sole plates (p. 122). Fig. 110: middle ring for 20 ft by 22 ft (p. 123).
  • Shrouds: cast-iron plates in segments with curved flanges taking the bucket plates, bolted by overlap joints with the overlap on the bucket side to keep a smooth outer face; arms dovetailed into T-head recesses cast on the shroud, boss tapered to avoid casting strain (p. 121-122). Shroud 15 in deep in the detail (p. 121).
  • Sole plates are wrought iron, "No. 10 Wire Gauge", lap-riveted; buckets riveted to them along their whole length; bucket stays every 2 ft of breadth, turned wrought iron with two collars or cast iron, forming rings of columns (p. 123-124).
  • Power off the rim: internal segmental spur wheel cast on a shroud, meshing a pinion; "the arms, braces, and axle have only to sustain the weight of the wheel" (p. 141). The pinion must lie below the centre of gravity of the water, below the end of the horizontal diameter, on the loaded side; a wheel hung on 2 in rods "could not resist the strain" if driven from the unloaded arc, but "would twist round upon the axle" (p. 140). Old wheels drove from the axle through the arms to a pit wheel (pp. 140-141). Fairbairn introduced rim gearing "nearly twenty-five years ago" (p. 254).

Named wheels (pp. 129-136, 140, 148-149, 152, 156)

  • Catrine, Ayrshire, 1825-27, fall 48 ft, two of four planned; each pair "working to 240 horses' power", assisted by auxiliary steam in very dry seasons (pp. 129-130). Wheel 50 ft diameter, 10 ft 6 in between shrouds, 15 in deep, 120 buckets, 6 in opening, bucket capacity about 17 cu ft; internal spur segments 48 ft 6 in, 3.25 in pitch, 15 in broad, 560 teeth; pinions 5 ft 6 in; main journals 14 in; first-motion shafts 13.5 in (OCR) by about 19 ft between couplings (pp. 131-132). Thirty years later "required little or no repairs" (p. 129).
  • Catrine water: 2150 cu ft/min (about 60 tons) at start with rim speed 4 ft/s, 182 buckets per minute, bucket one third full; at 240 hp with 75 per cent efficiency, 98.2 tons/min at 5 ft/s, 7.7 cu ft per bucket, one half full (p. 131).
  • Catrine speed train (p. 132): wheel 1.6 rpm at 4 ft/s; segment 48 ft 6 in to pinion 5 ft 6 in gives 13.3 rpm; 18 ft 3.5 in wheel to 5 ft 6 in pinion gives 44 rpm at the mill main shaft; 7 ft bevels carry the power of all four.
  • Deanston, Perthshire, River Teith, fall raised from 18 to 33 ft by weir and a three-quarter-mile canal; 8 wheels planned for 800 hp, 4 built, supply 260 cu ft/s ordinary (pp. 134-135). Wheels 36 ft diameter, as Catrine; one side internal spur gearing, the other external; 4.15 ft/s rim, 2.208 rpm; segments 33 ft 8.5 in to 6 ft 6 in pinion 13.59 rpm; 18 ft 2 in wheel into 6 ft 6 in pinion 45.38 rpm at the main shaft (p. 135).
  • Cleator, Whitehaven, for T. Ainsworth: ventilated high-breast, 20 ft diameter, 22 ft between shrouds, 22 in shroud, close riveted sole of No. 10 gauge, fall 17 ft, circular shuttle raised and lowered by a governor, two pinions into internal segments, one each side, feeding two mills (p. 140; Plate IV).
  • Table of proportions, about 68 wheels, printed pp. 148-149 (diameter of shrouds, fall, shroud depth, buckets, opening, rim speed, segment diameter, teeth, pitch, breadth of teeth). It has no wheel breadth and no power column. Rows: 60 ft 4 in on 56 ft 4 in fall, 4.0 ft/s; 50 ft on 48 ft 4.08; 40 ft on 36 ft 4.17; 30 ft 6 in on 27 ft 4.33; 24 ft on 21 ft 4.5; 20 ft on 17 ft 4.5; 16 ft on 11 ft 4 in 6.0; 14 ft on 8 ft 6 in 5.83; 10 ft on 8 ft 4.5; 7 ft on 5 ft 5.0. Rim speed is 4 to 4.7 ft/s on most wheels, 5 to 6 ft/s on several of 14-16 ft. Ventilated is marked on the wheels of 18 ft and under; "floats, ventilated" on one 16 ft wheel. Tooth pitch runs 3 to 3.5 in on big wheels, 1 to 2.5 in on small.
  • Design sequence (p. 147): fix the wheel diameter from the fall and outlay; choose the bucket; find water per second, which fixes bucket capacity and breadth; bucket one-third to one-half full; revolution rate sets buckets passing the shuttle per second. Worked example (p. 150): 5 ft/s, 3000 cu ft/min, shroud 14 in, spacing 14 in, section 1 sq ft: 5 buckets/s, each holding 10 cu ft of water, capacity 20 cu ft, so breadth 20 ft.

Rim speed and the load (pp. 141, 143-147)

  • "I have usually made breast wheels ... with a velocity between 4 and 6 feet per second at the periphery"; minimum 3 ft 6 in for falls of 40-45 ft, maximum 7 ft for falls of 5-6 ft. 5 to 6 ft/s is "very generally adopted" not for economy of water "but for the purpose of obtaining more easily the requisite speed under the variable conditions of supply"; 4 to 4.5 ft/s with a larger wheel would be better (p. 141).
  • The book gives no curve of wheel speed against load and does not say the wheel slows as load rises. It treats the speed as held by the governor working the sluice.
  • Governor (pp. 143-147, figs. 119-122): two heavy balls on a vertical shaft driven by bevels, a cam and clutch box between two counter-rotating bevels; a slow wheel drops the balls and the clutch is thrown to lower the shuttle (more water, on this reading, as the sheet flows over the sluice's upper edge), a fast wheel raises it; between those the clutch is out and the gear stands still. Governor shaft 36 rpm; rack shaft 0.0314 to 0.058 rpm; a worm and wheel pair twice reduces; an 8 in pinion drives the sluice rack. The sluice motion "should in every case be slow", or the supply surges and the wheel runs "first faster and then slower" (p. 147). Computed: 8 in pinion at 0.0314-0.058 rpm moves the sluice 0.8 to 1.5 in per minute.
  • Overshot: the sluice or shuttle at the end of the pentrough "regulate[s] the discharge" (p. 125).

Powers and efficiencies compared (pp. 112-113, 126, 152-156, 177)

  • Theory: 62.5 lb per cu ft, 550 ft-lb/s per hp: "for every foot of fall 8.88 cubic feet of water per second ... one horse power" (p. 113). Efficiency m = useful effect over water power; "efficiency or modulus" (p. 113).
  • Overshot ordinary: 12 cu ft/s per ft of fall per hp at 60 per cent.; improved iron high-breast 10.8 cu ft/s at 75 per cent. (p. 126). Computed: against his own 8.88, 60 per cent. needs 14.8 and 75 per cent. needs 11.8; the printed 12 and 10.8 imply about 74 and 82 per cent. The 75 per cent. and 8.88 figures reproduce his Catrine sum; use them.
  • Undershot (radial floats, impulse): 12 to 25 ft diameter usual, 14 to 18 ft "perform the best duty"; float depth 18 in to 2 ft, vein 6-9 in; rim speed 2.4 times the square root of the fall (OCR) is quoted from a second author (p. 153).
  • Poncelet wheel: best rim speed 0.55 of the current (0.5-0.6); 50 to 60 per cent. of the water power; curved iron buckets, no sole, number 8/5 of the diameter plus 16 for 10-20 ft. Example 16 ft 8 in diameter, 30 ft wide, 6 ft 6 in fall, 20,000 cu ft/min, rim 11-12 ft/s, "140 horses-power"; good only to falls of about 6 ft, above that "the low breast wheel is certainly more advantageous and costs less" (pp. 154-156). Poncelet's 11 ft by 28 in wheel 52 per cent.; Morin, falls 3 to 4.5 ft, openings 6-11 in, 52 to 62 per cent. (p. 156). Fairbairn's low-breast ventilated wheel gives about the same, 60 per cent. on a fall of 5 to 8 ft (p. 156 note).
  • Turbines 65 to 78 per cent., "doubtful" that they beat "a well-constructed water wheel of 50 years' duration" (p. 177).

One wheel and engines (pp. 129, 244-245, 250-251, 254)

  • Catrine's wheels carry "auxiliary steam-power" in very dry seasons (p. 130). Small high-pressure engines suit "small establishments and ... auxiliaries to water wheels when the supply of water fluctuates, and a small engine is needed when the supply is deficient" (pp. 250-251).
  • A pair of 60 hp beam engines can be joined on one main shaft with two horizontal high-pressure engines through a spur fly-wheel and pinion (pp. 244-245). The first engine flywheel gear: a 36-ton wheel for a pair of engines of 240 hp convinced sceptics that rim gearing kept the motion regular (p. 254).
  • The same rim-gear method, internal for water wheels, external for engines, drives the first-motion shaft direct, without speed-up gearing (p. 254). No page gives a single wheel and a single engine on one line with their powers.

Smeaton: named only for his 1759 experiments (p. 115) and an undershot report (p. 152); this volume does not attribute the cast-iron axle to him.

Computed figures

Computed rim speed and rpm, from the pitch circle: 20 ft at 4.5 ft/s is 4.3 rpm; 30 ft at 4.5 ft/s is 2.9 rpm; 50 ft at 4 ft/s is 1.53 rpm; 7 ft at 5 ft/s is 13.6 rpm. Computed water for a given power, at 75 per cent., Q = hp x 8.88 / (0.75 x fall): 12 hp on a 10 ft fall takes 14.2 cu ft/s; 12 hp on a 5 ft fall takes 28 cu ft/s; 100 hp on a 33 ft fall takes 36 cu ft/s (Deanston's 260 cu ft/s ordinary supply drives about 725 hp on 33 ft, a computed check on its "nearly 800"). Computed Catrine: 240 hp per pair is 120 hp per 50 ft wheel; rim force at 5 ft/s is 120 x 550 / 5 = 13,200 lb. Computed Deanston: 800 hp over 8 wheels is 100 hp per 36 ft wheel. Computed tooth speed-up, wheel rpm to main shaft: Catrine 1.5 to 44 (about 29 to 1), Deanston 2.21 to 45.4 (about 20 to 1). Not in the book: an iron wheel's mass beyond the single 34 ton figure, wheel breadth against power, rpm under load.