by Ben Russell
Choreographing all these techniques required careful organization, and each worker came to be an expert in a single, tiny part of the overall process. A commentator remarked how, in preparing the button blanks, ‘A small boy makes the blanks red-hot in a small furnace. Another boy puts them under the punch, one by one. The third picks them out of the punch and greases the upper mould between each punching with a greased brush.’38 A visitor to Soho admired how this approach made the faculties of those involved ‘more expeditious and more to be depended on than when obliged or suffered to pass from one to another’.39 The same visitor recorded how each button ‘passes through fifty hands, and each hand perhaps passes a thousand in a day; likewise, this work becomes so simple that, five times in six, children of six to eight years old do it as well as men’. Although the division of labour was extreme, there was still some latitude for those involved to exercise
considerable artistic skill, or educated art, as in closing, tool making, burnishing, and in the turning of pearl and other buttons, and where so high a degree of workmanship is not necessary, there is still a certain variety of labour and careful attention to be given to it that involve some exercise of the mind as well as the mere physique . . . so that a certain kind of sharpening of the wits goes on, more than exists in many other kinds of labour.40
Some button makers estimated that this way of working was twenty times faster than when each individual item was made by a single worker.41
The button making workforce consisted of men, women and children. Women tended towards specific parts of the process, particularly polishing the finished steel buttons with ‘the combined oxides of a mixture of lead and tin . . . mixed with water or proof spirit’ until a fine black shine was produced.42 Patent agent Thomas Gill went so far as to note that ‘no effectual substitute for the soft skin which is only to be found upon the delicate hands of women, has hitherto been met with.’43 How long their hands remained delicate while being immersed in metal oxides and alcoholic proof spirit is questionable. Children worked largely in the unskilled side of the button trade; they included among their number the ‘nut crackers’, ‘little lads who are engaged in breaking the outer shells off the vegetable ivory nuts [the seeds of palm trees] ready for the workman who saws them up. Every little rascal who is too wild for steady work can be set to do this, their destructive propensities being happily utilised in the manner described.’44
Button making contributed, then, to the industrious atmosphere of Soho that so exhilarated visitors. ‘The front of this house’, wrote Jabez Maud Fisher,
is like a stately palace of some Duke. Within it is divided into hundreds of little apartments, all which like Bee hives are crowded with the Sons of Industry. The whole Scene is a Theatre of Business, all conducted like one piece of Mechanism, Men, Women and Children full of employment according to their Strength and Docility. The very Air buzzes with a Variety of Noises. All seems like one vast Machine.45
We can almost hear the heave and thud of the drop stamps, the tinkle of the button blanks falling from the presses, the whirring treadles of lathes and the thunder of feet as children raced boxes of materials from one workshop to the next. This is how Boulton systematized the button manufactory. From 1774 the challenge for Boulton & Watt was applying these concepts to the business of making steam engines.
The basics of engine manufacturing were well established. To build a Newcomen engine, the customer would contract with an engineer to erect the engine, and with different metal foundries and other suppliers to provide the components and materials needed. The cylinder and pumps would usually be delivered ready for use, and some of the smaller iron parts would be ordered from local suppliers for ease of delivery, but everything else had to be manufactured on site, for which purpose the engineer assembled a team of smiths and plumbers, masons and carpenters. This was broadly the path Boulton & Watt followed, albeit with some significant differences. They were careful to specify particular suppliers of each part to ensure their high standards were met. John Wilkinson of Bersham, Shropshire, became their favoured supplier of engine cylinders and pistons, pumps and condensers. Izon’s foundry in West Bromwich, Wilkinson’s other works at Bradley, near Birmingham, and the city’s Eagle Foundry manufactured many of the smaller metal components, and piston rods were made by Jukes Coulson of Rotherhithe, who had a well-established reputation for making ship’s anchors.46 Importantly, then, beyond the very specialized parts like the ‘nozzles’, which controlled the flow of steam in and out of the cylinder, Boulton & Watt did relatively little of the physical work of engine-building themselves.
This was a canny move: the partners realized that manufacturing Watt’s improved engine and exploiting it commercially were not necessarily the same thing. The former, with the possibility of technical difficulties in getting the engine to operate as required, and needing engineering facilities, foundries and workshops, would require huge investment and carry heavy risks. The latter, obtaining revenues from the engine’s users, was where the potential profits were. So Boulton & Watt, rather than building engines from scratch themselves, designed them, carried out the necessary calculations, produced working drawings and provided an ‘engine erector’ from a pool of appropriately skilled men to oversee the day-to-day execution of each engine’s construction. But the bulk of the work of physically creating the parts for each engine was given to others.
What, then, of assembling engines like buttons? This was a powerful piece of rhetoric, suggesting Boulton’s formidable persuasive abilities and shining a light on the close interconnections between machine making and a wider world of manufacture and consumption. But things would not work out that way in reality; no one had built engines on the scale Boulton & Watt wanted to before, and they needed to establish working practices that minimized the risks involved. As Boulton had bided his time to form his partnership with Watt, now he played a long game to ensure the success of the engine business. ‘Making engines like buttons’ was a statement of intent, perhaps intended to overawe his competitors. But the process of getting there was complex.
This complexity dogged Boulton & Watt from the start because of the location of most of their likely customers. They were not in the more easily reached new manufacturing towns or on coalfields, because these generally had sufficient supplies of water power, or enough cheap coal to run inefficient but adequate atmospheric engines. A key market for the pumping engine was to be Cornwall, where metal mines extracting valuable tin or copper ores needed more power as they dug deeper underground, but were reliant on coal supplies shipped at considerable expense from South Wales. Both these factors meant the Cornish miners sought the most efficient engine available; they needed to replace 75 atmospheric engines, which were struggling to cope with pumping huge volumes of floodwater just as their owners’ finances were stretched by their prodigious coal consumption.47 Boulton & Watt offered an ideal solution to the difficulties faced.
While we have considered the actions of steam and heat inside the engine, we have not yet explored the engine as a physical artefact that had to safely and efficiently exploit those properties, the construction of which was the major challenge that now confronted Boulton & Watt. The engines they made were big ‘house-built’ machines, working inside a building that provided support to the major components, and mainly used to pump water for canals, mines and water supply. The heart of the engine was its cylinder, which was mounted vertically on a strong foundation and measured 4 feet or more in diameter. Within was the piston, a large disc which filled as closely as possible the cylinder bore. The piston was attached to a piston rod that passed through the top of the cylinder via a ‘stuffing box’, which kept the steam inside while allowing the rod to go up and down. Steam was only injected into the cylinder on top of the piston, driving it downwards on its working stroke and making it a ‘single-acting’ engine. The piston rod was attached to a beam, which usually pivoted on top of one of the engine house walls, and the outer end, protruding outdoors, w
as connected to the pump rods that might extend hundreds of feet underground to drive pumps that moved water wherever it was needed. Inside the engine house a plug rod hung from the beam: this moved synchronously up and down with the piston, and pegs on it hit levers that opened and shut the engine’s ‘nozzles’. Engines on this scale, in the form of Newcomen’s atmospheric engine, had been built since about 1710. Now Boulton, Watt and their workforce faced the demands of making a similar-sized machine requiring the precision of instrument making.
Watt’s single-acting pumping engine for draining mines, 1788, as drawn in John Farey’s Treatise on the Steam Engine (1827).
Much of each engine was made of timber. The engine beams were huge logs of oak or deal, several feet square, more than 15 feet long, and reinforced with braces and iron straps to withstand the strains placed upon them. Inside each engine house the engine’s moving parts were supported by a robust wooden frame. Engine builders worried constantly about the supply of large pieces of timber, particularly as the needs of the Royal Navy became increasingly pressing during the wars of the last quarter of the eighteenth century, and its cost rose. And the relative ease of working timber, following long-established techniques, had to be balanced against the difficulties of using it in a machine, like Watt’s engine, that required high precision. Daniel Treadwell wrote how ‘A machine constituted of wood, subject to constant swelling and shrinking, and warping with every change of the atmosphere, is always liable to derangement. Indeed it can be said to be hardly capable of preserving its identity!’48
So, there was some impetus to use different or new materials. Watt’s Kinneil engine had a cylinder made of block-tin, solid tin blocks heated to melting point and then poured into a mould to form the correct shape upon cooling. Tin wasn’t strong enough to withstand the vacuum within and the cylinder collapsed, to be replaced by one cast in iron.49 And it was iron that became the engineer’s main material. Giant blast furnaces in South Wales or Shropshire produced cast iron ‘pigs’, rough iron ingots named after the moulds the molten metal was poured into from the furnace, resembling a sow with piglets suckling at her side. The pigs were transported to Soho and other engineering centres and melted again to be turned into finished products. Alongside cast iron came wrought iron, supplied in bars that had been heated in a specially designed ‘puddling’ furnace to remove the carbon and any impurities, then shaped by massive hammers and squeezed between rotating pairs of rollers until it was very fibrous in composition, which quality helped give it great strength. Iron was durable, strong and capable of being precisely finished. Building steam engines required new parts to be made from iron in two ways: by casting and by forging. And the means of carrying out these processes ‘were rude, and the machinery imperfect . . . Nearly all depended on the individual workman’s skill.’50 Those skills were in employing a range of hand techniques.
Take casting iron, for instance. First, a wooden model or ‘pattern’ of the required component had to be made. The pattern maker’s tools were similar to those used in fine joinery or cabinetmaking, and he worked in deal, pine or mahogany, carefully dried so that it would not warp, screwing the pieces of wood together so they could be easily altered if need be and using a plane for fine adjustments.51 The pattern maker had to understand the entire casting process: he had to avoid sharp internal corners, which could lead to cracks in the cooling metal, and surfaces were always tapered, not quite parallel or square; if they were, there was a worse chance that they would stick in the sand mould, damaging its surfaces. Some allowance also had to be made for the metal contracting as it cooled, and the thickness of the metal was kept as constant as possible, so it cooled evenly.52
The patterns would be placed in huge beds of sand, which was packed around them so that, when they were lifted out, a perfect negative copy was left behind for molten metal to be poured into. The tools for this were simple trowels and shovels, and mauls to ram the sand in closely to the pattern. But the composition of the sand was critical, and depended on the moulder’s skill: it needed to be damp, or ‘green’, to form a good impression, but this carried the risk of an explosion if the molten metal contacted too much moisture, or, if gases could not escape, a ‘blown’ casting with areas of spongy, porous metal that would be much weaker than the surrounding material.53 And prior to the molten metal being poured, the details of the mould would be carefully prepared with tools: copper spoons for removing sand, trowels to smooth surfaces and round-ended tools to make sharp corners smooth. ‘Runners’, the channels in the mould to convey the molten metal as quickly as possible to where it was needed before it began to cool and lose its liquidity, were also carefully laid out.54 Finally, with the metal poured and cooled down over a period of days, the casting would be taken from its mould, cleaned up and prepared for further work.
Alongside casting, forging the iron to shape was an important part of steam engine construction. Wrought or ‘malleable’ iron was supplied to the Soho Engine Manufactory by Isaac Spooner of Birmingham, and Watt also specified use of ‘the best tough scrap iron from Wednesbury’ or ‘the best gun barrel iron’.55 Often the fibres of this high-quality metal were treated the same way as the filaments of spun cotton, and twisted as they were worked to avoid the outer surface becoming marred by dirty, longitudinal seams or ‘spills’, which could spoil inferior ironwork.56 Forging these bars to the required shapes needed the use of a smith’s furnace to raise them to the right temperature, from a black-red just visible in daylight, to a bright red at which most work was performed, and white-hot and burning with vivid sparks for welding pieces of iron together.57 From the furnace the pieces of iron were hastily transferred to the blacksmith’s anvil and there, with tongs, hammer and a range of other tools, they would be worked to the required shape: ‘drawing down’ reduced the thickness and increased the length of a piece of iron; ‘upsetting’ it made it thicker and shorter; and ‘building up’ saw pieces welded together.
‘The Iron Founder’, from The Book of Trades (1824). He is pouring molten iron from a ladle into a box mould for making small castings.
These three main processes were central to the manufacture of many parts of the engine, and would have comprised much of the work carried out at Soho. The centrepiece of the Engine Manufactory was a blacksmith’s shop with two great hearths for heating the iron, and at least one lathe – probably a great lathe with a hand capstan to drive it, given the size of some of the components that might be turned there: the largest of these might be the piston rod that attached the engine piston to the beam. This was made by taking a bundle of smaller rods, the central one round in section, the outer ones comprising angular ‘mitre iron’, all heated until they began to emit sparks, and then welding them into a single mass under the blows of a team of men wielding sledgehammers under the guidance of the foreman, indicating where he wanted the hammer strokes to fall with a long wooden wand.58 At the other end of the scale of forge-work came the making of smaller items like nuts and bolts, demonstrating the versatility of the smith engaged in engine manufacture.
With foundry and forge work highly dependent on the hand skills of the foundryman, pattern maker or smith, the only part of the engine that required a dedicated, highly accurate machine tool was the cylinder. This reflects that much of Watt’s frustration with the engine arose from being unable to make a piston and cylinder that were a close fit to each other; if there was any gap between the two, steam would leak through and the engine would come to a halt. Finding a suitably elastic but durable ‘packing’ material to fill the gap between piston and cylinder while still allowing the former to move up and down caused major delays to the engine’s development; Watt experimented with cork, pasteboard, leather and linseed oil. Later he even resorted to ‘horse-muck’ and ‘paper pap mixed with flour paste’.59 Finally, the piston had rope wrapped around its circumference, compressed tight by a ‘junk ring’, and this worked well.60 But the best strategy was to make the cylinder as accurately as possible and, for this, Watt
depended on the ingenuity of John Wilkinson who, in 1775, made a cylinder-boring machine. A cylinder, cast from iron as a tube open at both ends, was secured on its side upon the machine and a boring bar was placed along its axis, also supported at both ends. The bar supported a cutter that was advanced along the bore by a rack, rotating as it went and cutting the cylinder straight and perfectly true. The degree of accuracy Wilkinson acquired using this machine was unequalled in its time; Watt boasted that ‘Mr Wilkinson has bored us several cylinders almost without error, that of 50 in. diameter . . . does not err the thickness of an old shilling at any part.’61 The cylinders on Watt’s experimental engines, at 18 inches’ diameter, might vary in diameter by around 1/8 of an inch. With Wilkinson’s machine, a cylinder almost three times larger could be machined to a maximum error of around 4 millimetres.62 Here was precision manufacturing scaled up from scientific instruments to steam engines.