The Steam Engine
L. T. C Rolt & J. S. Allen
Moorland Publishing Company
Science History Publications/ USA
New York ‑-1977
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of Moorland Publishing Company.
Published in Great Britain by
Moorland Publishing Company,
The Market Place, Hartington,
Buxton, Derbys, SK17 OAL
First published in the United States by
Science History Publications/USA
a division of
Neale Watson Academic Publications, Inc.
156 Fifth Avenue, New York 10010
Printed in Great Britain by
Wood Mitchell & Co Ltd, Stoke on Trent.
Ed. The following excerpts from Chapter 5, Pages 89 – 110 give detail not found in other publications. J.McV Oct 2006
Fig. 94 The Hawkesbury Engine
Technical Developments 1712-33
In order to present the story of Newcomen and his engine as clearly as possible a great deal of technical detail has been omitted from the preceding chapters. Such detail is essential, however, to any assessment of Newcomen's achievement because so many conflicting statements have been made as to how, and by whom, certain features of the engine were developed. Quite apart from the need to pay Newcomen his due, the story of how men, who were quite ignorant of the nature of steam or the laws of thermodynamics, groped their way to success by sheer practical ingenuity and tenacity of purpose is a fascinating one.
To appreciate to the full the achievement of Newcomen and his associates, the reader must use his imagination, forget his twentieth‑century background and try to think himself back into those far‑off days. Such an imaginative effort is essential if the error of hindsight is to be avoided, because the solutions of many of the problems with which Newcomen grappled now seem so obvious as to be self‑evident even to those who are not technically minded. Even some of Newcomen's contemporaries and immediate successors made this same mistake, so fatally easy to be wise after the event. The most brilliant inventions are those which seem obvious once they have been explained or demonstrated. 'Why, I could have thought of that', we say with a feeling of envy and mortification, and it is clear from their writings that both Desaguliers and Triewald reacted in this way, convincing themselves that they could have done better than Newcomen had they chosen to bend their minds to his problems.
Next to the principle of injecting water into the cylinder to create a vacuum, the most significant feature of the Newcomen engine was the valve gear, or 'working gear' as it was called, which made it self‑acting. No early writer awards Newcomen any credit for this valve gear and the only legend about his engine that deserves to be called popular would have us believe that the valves were worked by hand until a boy, tiring of this monotonous task, connected their handles by strings to the beam. There may be a grain of misunderstood truth in this legend as we shall see presently.
Farey, in his Treatise on the Steam Engine, summarizes the evolution of the valve gear as follows:
At first the valves were opened and shut by hand, and required the most exact and unremitting care of the attendant, to perform those operations at the precise moment; the least neglect or inadvertence might be ruinous to the machine, by beating out the bottom of the cylinder, or allowing the piston to be wholly drawn out of it. Stops were contrived to prevent both of these accidents; then strings were used, to connect the handles of the cocks with the lever [ie the balance beam] so that they should be turned whenever it reached certain positions. These strings were gradually changed and improved, into detents and catches of different shapes: till at last, in 1718, Mr. Beighton, a very ingenious and well‑informed engineer, simplified the whole of these subordinate movements, and brought the machine into the form in which it has continued, to the present day, without any material change.
Stuart, in his History of the Steam Engine, goes further than this. His plate entitled 'Newcomer's Engine' depicts an engine with hand‑operated valves, while the next figure showing an engine with self‑acting valve gear is headed 'Beighton's Engine'. His accompanying text reads:
The mechanism for opening and shutting the cocks also remained perplexed by catches and strings, until Mr. Henry Beighton, an engineer extensively employed in the construclion of mining machinery, erected an engine at Newcastle‑on‑Tyne in 1718, in which all these "cock‑boys" and complications of cords were superseded by a rod suspended from the beam, which operated on a mechanism invented by him called hand‑gear: a contrivance, with some slight mod)fications, employed in engines of the present day.
Although it is not possible to back the statement with positive proof, it is safe to say that the hand‑operated valves referred to by Farey and illustrated by Stuart belong to the experimental period before 1712. From that date forward Newcomen's engine had the self‑acting valve gear evolved by him and his assistant Calley and differing only in one important respect from that which was later standardized and used with only minor variations down to the end of the eighteenth century. This one important difference accounts for the references to catches and strings as well as for the legends about the ingenious 'cock boy'. The fact that these references and legends are woefully misleading is due to a failure to understand the difficulties— some of them due to lack of knowledge and experience—encountered by Newcomen at the outset and the way he met them. Before these difficulties and Newcomen's solutions to them can be understood, it is essential to consider two things: the adequacy of the boilers used on Newcomen's first engines and this method of calculating the work these engines could perform.
It will be recalled that when Nicholas Ridley proposed to erect a second engine at his Byker colliery with a 33in cylinder the inventors, presumably Newcomen and Calley, refused to undertake the project on the grounds that so large a cylinder could not be supplied with sufficient steam. Commenting on this refusal, Triewald writes:
The cause of this conclusion was the false principles concerning the steam which the inventors harboured in their minds according to which the steam rises or is generated by the boiling water in proportion to the quantity of water in the boiler. In consequence their boilers were made very high, as demonstrated by the Stafford [Dudley Castle] engine, the boiler of which is higher than its width. It is thus evident that the inventors do not know that the boiler must be given a suitable shape. Neither did they know that the flames should be allowed to play all around the side of the boiler as well as on the bottom....
Although Triewald was a conceited young man and apt to be wise after the event, there is no doubt that in this case his criticism was valid. It may seem obvious to us that the steam generating capacity of a boiler depends on its heating surface but this was by no means obvious in 1712. To suppose that it was only necessary to increase the quantity of boiling water in order to obtain a proportionate increase in the volume of steam produced from it was then a perfectly
understandable error. Moreover, although Triewald appreciated the need to increase the heating surface, this conclusion was purely empirical and boilers continued to be built on an empirical basis for more than sixty years. The first Watt engines were under-boilered and this defect led Watt to work out, for the first time, a desirable ratio between boiler heating surface and cylinder volume. He specified four square feet of heating surface per cubic foot of cylinder volume.
The type of boiler used by Newcomen consisted of a cylindrical copper vessel with a concave bottom directly above the furnace. This was surmounted by a hemispherical steam dome of lead. The diameter of this dome was greater than that of the copper cylinder below, the upper portion of the latter being flanged outwards at right‑angles and then turned upwards again in order to form a circumferential seam with the lead dome. It thus resembled a haystack in crosssection and it became know as a ' haystack' or 'flange' boiler, sometimes also called a 'balloon' or 'beehive' boiler. This was certainly the type of boiler supplied initially to the Whitehaven engine, although Spedding stated that it was the first of its type that Newcomen and Calley ever had and they did not approve of them as much as those without flanges. The brickwork enclosing the furnace was carried upwards round the boiler and sealed against the lead dome plates. This arrangement left an annular space beneath the dome and through this the hot gases from the furnace circulated before passing up the chimney stack. The proportions of the boiler were such that at working level there was but little depth of water upon the copper 'roof' of the annular flue and this made it vulnerable in the event of mismanagement. Although at an early date, high and low level try‑cocks were provided so that the water level could be checked, attendants all too frequently allowed the level to fall low with the result that the copper 'roof' overheated and this led to the failure of the copper‑lead seam at its circumference. The Austhorpe engine is said to have burned out four boilers in this way in as many years. The first boiler at Whitehaven suffered the same fate and had to be repaired first with lead and copper patches and then lined all round with lead held on with lead nails. Although the eventual substitution of wrought‑iron plates for copper and lead boiler less liable to suffer from overheating, what was known as the 'tun' boiler came to be favoured by later builders. Instead of the angular construction of the Newcomen 'haystack', the lower portion of the tun boiler was cylindrical, tapering down from the diameter of the steam dome to that of the concave bottom plate immediately above the furnace. This entailed a certain loss of heating surface, but this sacrifice was offset by altering the proportions of the boiler and its furnace and flues. The second boiler at Whitehaven was made from iron plates with a lead top and supplied by Stonier Parrott in 1717. It was reported as being broader at the bottom and narrower at the top than the first boiler. This was to give trouble in service at the junction of the iron and lead and the iron corroded externally in this area. As early as March 1717 Mr Calley had commented that 'he approved very well of iron boilers' indicating their use elsewhere prior to this time.
So ingrained in us is the importance of a large heating surface that modern diagrammatic representations of the early Newcomen engine almost invariably depict a haystack boiler of a diameter exceeding its height. By contrast, Barney's engraving of the Dudley Castle engine shows a brick furnace and boiler casing of considerable height in proportion to its diameter. It is particularly interesting to note that Barney shows the firehole door in an impossible position above an arched access doorway to the ash-pit so high that the fireman depicted in the foreground could walk through it without stooping. Without the aid of a stepladder he would have had to be a gymnast to put coal on the fire and would certainly need the 'little Bench with a Bass to rest when they are weary' which, according to Barney's key, was thoughtfully provided. This odd aspect of the drawing may be due to Barney's method of showing the ash hole and engine‑water pump below the engine‑house floor. The engine man is standing at the firing‑floor level but this part of the drawing is moved forward to permit the lower level to be seen. Barney's accompanying key tells us that the boiler was 6ft 1 in high but that the diameter of the bottom plate was only 4ft 4in. Barney also gives the capacity of the boiler as 'near 13 Hogsheads', this being the equivalent of approximately 680 imperial gallons. Having regard to the srr.all heating surface available, this was an immense volume oi water and Barney’s figures thus corroborate Triewaid's criticism of the Oucley Castle engine. It is clear that: as he says, Newcormen did believe at this time that if he increased the volume of water in a boiler its OUtpilt of steam would increase proportionately . The full significance of this will appear presently.
It is also interesting to note that the von Schonstrom drawings of the Konigsberg engine show a plain boiler without flanges.
The valve or 'regulator' that controlled the admission of steam from the boiler to the bottom of the working cylinder was mounted in the top of the steam dome and it resembled the type of valve used by Savery on his later engines. It consisted of a fan‑shaped brass plate which, moving to and fro horizontally, alternately covered and un‑covered a steam port formed in a second brass plate rivetted into the top of the boiler. On some engines the moving plate was held to its port face by a flat steel spring, bridging the steam orifice below and bearing against a boss formed on the lower surface of the plate. From the fixed brass plate in the top of the boiler the steam pipe extended upwards to meet in a butt joint a pipe of the same diameter passing through the centre of the cylinder bottom. This butt joint was wrapped, first with canvas covered in white lead and oil, then with sheet lead and finally bound tightly with cord. This primitive method of jointing persisted for years despite frequent failures caused by movement of the cylinder while the engine was working under load.
The vertical spindle of the moving sector valve was tapered like the barrel of a plug cock where it passed through the fixed brass plate in the boiler top and it was ground into a taper seating formed in the latter in order to keep it steam‑tight. To the squared end of this spindle a spanner was attached which could be worked to and fro by a linkage arranged as follows. First there was pin‑jointed to the spanner a horizontal link called the 'stirrup' because it was so shaped at its opposite end. The forked ends of this slirrup were supported by two vertical links, harlging loosely from a horizontal arbor which they shared with two levers called the i' lever and the 'little Y'. Both were fixed at their common axis. The Y lever was in the shape of that letter inverted and it carried a weight or tumbling bob on its upper end. As it rocked on its arbor and the weight was thrown over centre, its two lower arms alternately struck the crosspiece of the stirrup and thus opened and closed the valve, the action of the weight ensuring rapid and positive action. The plug rod which hung from the beam imparted this rocking motion to the arbor by means of the little Y lever. Pegs in the plug rod alternately raised and depressed the arms of this little Y as the rod rose and fell. The valve spanner itself was provided with positive stops and there were two methods of adjusting the motion. The pegs in the plug rod could be moved into alternative holes, while a series of holes in the arm of the stirrup enabled the pin‑joint by which it moved the spanner to be adjusted similarly. This type of steam valve and its operating linkage persisted almost unchanged for many years withonly detail refinements and it is safe to credit its inception, not to Henry Beighton, but to Newcomen and Calley. The design of the valve itself may have been derived from Savery, but the ingenious self‑acting gear was entirely original.
The admission of injection water to the cylinder was controlled by a simple plug cock. To the designer of the steam admission gear it would have been a simple matter to develop a similar linkage to actuate this cock from the plug rod, but Newcomen did not do this for two closely associated reasons: the limited steam generating capacity of his boiler and the excessive load he imposed on the engine.
Desaguliers, who was doubtless supplied with the information by Henry Beighton, has this to say about Newcomen's method of estimating the power of his engines: Mr Newcomen's Way of finding it was this: From the diameter [of the cylinder] squar'd he cut off the last Figure, calling the Figure on the left Hand long Hundreds, and writing a Cypher on the right Hand, call'd the Number on that Side, Pounds; and this he reckon'd pretty exact as a Mean, or rather when the Barometer stood at 30 and the Air was heavy. N.B. This makes between 11 and 12 pounds upon every superficial round Inch. Then he allow'd between 1/3 and 14 Part for what is lost in the Friction of the several Parts and for Accidents: and th~is will agree pretty well with the work at Griff Engine, there being lifted at every stroke between 2/3 and 3/4 of the weight of the atmospherical Column pressing on the Piston.
This works out at a mean effective pressure of 9-1/2 1b/in2, an optimistic figure bearing in mind that the atmospheric pressure employed could not exceed 14 lb/in2. Seventy years later the cautious Watt, although he was using steam at a pressure slightly exceeding that of the atmosphere, allowed a mean effective pressure of only 5 lb/in2 on his first rotative engines. The figure of 9-1/2 lb/in2 adopted by Newcomen meant that his first engines were so loaded that their working depended on the creation of a high degree of vacuum below the piston.
When the engine had made its power stroke, or 'come into the house' to use the current engineman's expression, the piston was returned to the top of the cylinder by the weight of the descending pump rods at the other end of the beam. So far from being asssisted in this motion by the incoming steam, the pressure of the steam above atmosphere being negligible, the piston in its ascent actually helped to draw the steam from the boiler into the cylinders. The steam-generating capacity of the first boilers was so inadequate that the sudden extraction of so great a volume of steam might literally have the effect of sending the boiler 'off the boil'. If, therefore, water was immediately injected into the steamfilled cylinder so that the engine made another power stroke, the piston would once again be returned by the pump rods but little or no steam would follow it because the boiler had not had time to recoup. This, of course, would bring the engine to rest because in the absence of steam no vacuum could be created. Faced with this problem Newcomen decided that the solution was to regulate the working capacity of the engine in accordance with the steaming capacity of the boiler.
This solution took the form of a contrivance called a buoy. This was a small buoy floating upon the surface of the water in the boiler and enclosed in a vertical tube—the 'buoy pipe'—which protruded through the dome of the boiler and carried within it a rod attached to the buoy. By means of this rod the buoy controlled the opening of the injection cock in the following way. This cock carried a shaped lever which became known as the 'F' because of its shape. With the cock in the closed position, this F‑lever lay at an angle of about 30°to the horizontal, with its foot nearest to the plug rod and the letter inclining upwards therefrom with the two short arms pointing downwards. The shorter of these two arms was pierced with the square hole fitting over the spindle of the injection cock, while the longer carried a weight at its extremity. A short prolongation of the lever beyond the junction of this upper arm engaged with a pivoted catch or detent lever known as the 'scoggan'. This detent retained the lever in its closed position against the reaction of the weighted arm which would otherwise fall and cause the injection cock to open.
Towards the end of the engine's power stroke a peg on the descending plug rod depressed the foot of the F‑lever until it reached the closed position where it was retained by the detent. It was at this point in the cycle that the buoy came into play. The steam admission valve having opened and refilled the cylinder with steam as the piston 'went out of the house', the buoy ensured that another power stroke was not made until the boiler had recouped itself sufficiently to cope with the next demand for steam which must follow the power stroke immediately. When the dome of the boiler had again filled with steam, the slight pressure so created was aufficient to cause the buoy to rise in its pipe, whereupon the vertical rod attached to it raised the detent lever, thus enabling the F lever to fall by the action of its weight and so open the injection cock. It is important to note that in devising this ingenious cycle of operations Newcomen took it for granted that the steam demand of the cylinder would temporarily exhaust the boiler and so allow the buoy to fall. For if the buoy did not fall it would prevent the detent lever from returning to position in order to retain the F lever when the descending plug rod again restored the latter to the closed position.
It follows logically that so long as the engine was controlled by the buoy its action would be extremely slow. Whereas the word 'engine' to us implies continuous motion, on the first Newcomen’s each return of the pump rods would be followed by a prolonged pause while the boiler regained the necessary strength to perpetuate the cycle of operations. It seems clear, however, that. partly as a result of this extremely slow action, Newcomen did achieve a high degree of vacuum in his cylinder which enahled the engine to work a load as high as 9 lb'in2 on the piston. This compensated to some extent for the very slow rate of pumping.
Notwithstanding the fact that instead of positive evidence we have a fog of confusing and contradictory statements; the subsequent development of the engine can be deduced with the assurance that, although it is necessarily conjectural, it cannot be far from the truth. One explanation of the legend of the ingenious boy with the piece of string is that it arose as a result of a misunderstanding of the fact that the ‘buoy' used by Newcomen was not of the two‑legged species. Although such a misunderstanding may have added to the confusion, it seems far more probable that such a boy did exist but that those who mentioned him did not understand what he was set to do or what it was that he achieved with his piece of string.
It is obvious that, as soon as a Newcomen engine was provided with a boiler of adequate steam‑generating capacity, it became no longer necessary for the boiler to govern the working rate. Moreover, the old control gear could no longer operate because the buoy would keep the detent or scoggan permanently raised so that there would be nothing to retain the injection cock in the closed position during steam admission. When this first occurred there was only one way in which the engine concerned could be kept at work. This was by putting the buoy out of action by wedging its rod in the pipe and then lifting the detent by hand each time the piston reached the top of its stroke. The engine would then work at a much faster rate provided the boiler could continue to maintain the steam supply. I~o perform this monotonous operation, a two‑legged boy was pressed into service instead of Newcomen's similarly named device. Standing close beside the rising and falling plug rod, it would not take this boy long to realise that the plug rod could very easily be made to do his repetitive job for him. A suitably positioned nail in the plug rod and a length of cord from the nail to the detent lever were all that was necessary. How delighted he must have been with his crude but effective improvisation as each time the plug rod neared the top of its stroke the cord tautened, raised the cletent and so allowed the injection cock to open !
John O'Kelly tells us of the boy's invention:
At the beginning, they only made 6 to 8 and 10 strokes per minute and it was as a result of the invention of a youth who watched over the machine rhat they managed 15 and 16 strokes in the same period of time. This boy was called ilumphrey Potter, but this invention made the machine very complicated.
Young Humphrey Potter was most probably the brother of Isaac and John who were the sons of Stephen Potter, brother to Humprey, Sen, and clearly a family trio who had a tremendous influence on the introduction and development of the engine.
Desaguliers however gives a most muddled account of the affair as follows:
They used to work with a buoy in the cylinder, enclosed in a pipe: which buoy rose when the steam was strong, and opened the injection and made a stroke: thereby they were only able from this imperfect mechanism to make six, or eight, or ten strokes in a minute; till a boy named Humphry Potter, who attended the engine, added what he called a scoggan ‑ a catch, that the beam (or lever) always opened; and then it would go fifteen or sixteen strokes in a minute.
It is evident from this that the worthy scientist totally failed to grasp both the object and the working principle of Newcomen's buoy gear. He begins by placing the buoy in the cylinder, then blames the 'imperfect mechanism' for the fact that the engine worked so slowly and finally credits the boy Potter with the introduction of the detent lever or scoggan; all this to the infinite bemusement of subsequent writers.
A good example of the way in which nonsense begets nonsense appeared long afterwards in the Mechanic's Magazine and was quoted by Stuart. Accepting Desaguliers's account, the writer announced that the word scoggan was derived from a north Yorkshire verb 'to scog', meaning to skulk or idle and that it was obviously applied to the detent lever because it enabled its youthful inventor to idle instead of attending to his monotonous job. In fact the word is of Cornish origin and was doubtless first applied to the detent lever by Newcomen or his associates during the course of pre‑1712 experiments in Cornwall. The term is still applied in Cornwall to the Cornish engine valve gear, but it will soon become an archaism now that this ultimate development of the non-rotative beam engine has become a museum piece like its predecessors.
It should be emphasized at this point that in designing his sorely misunderstood method of operating his injection cock. Newcomen was the first to adopt the principle of opening the valve by the fall of a weight and closing it against gravity, a principle which has proved as long‑lived as the word scoggan, since it persisted through the Watt era down to the last Cornish engine to be built.
The acceleration of the engine from a mere six strokes a minute to twelve of fifteen strokes by the use of a more efficient boiler and the substitution of the 'Potter cord' for the buoy to regulate the opening of the injection valve was not achieved without loss in another direction. The faster rate of working allowed less time for the cylinder alternately to gain and lose its heat with the result that the degree of vacuum achieved below the piston was reduced. The mean effective pressure of 9-1/2 lb/in2 originally adopted by Newcomen was reduced to 7 lb/in2 or in other words half atmospheric pressure, to counteract this loss, but it was immediately found that the less heavily loaded engine working at the faster rate pumped more water per hour than it did before. Nor was any more fuel consumed in proportion to the work done because the more efficient boiler wasted less of the heat generated by the furnace.
If contemporary engravings are to be relied upon, it would appear that the earliest Newcomen engines were soon fitted with a refined version of the 'Potter cord' but that the buoy gear was retained. Beighton (1717), Barney (1719), and, most surprisingly, Triewald's engraving of the Dannemora engine (1734) all show both. No doubt when the boiler was steaming badly and could no longer sustain the rate of working imposed by the Potter cord, it was disconnected so that the buoy could take control. Meanwhile, however, the provision of more adequate boilers was accompanied by the rapid accumulation of experience among the men concerned with the erection and management of the engines with the effect that the time soon came when it was felt that the buoy gear could be safely discarded. By far the most likely explanation of the claim that it was Beighton who "invented' the self‑acting valve gear in 1718 is that it was he who first took this step by dispensing with both the buoy gear and the Potter cord on the engine which he erected at Oxclose Colliery, Washington Fell. This would also explain the statement that until this time the engine had remained 'perplexed by catches and strings'.
Beighton is also said to have been the first to fit a weighted safety valve, or 'puppet clack' as it was called, on the boiler of this engine at the suggestion of Desaguliers. It lifted at a pressure of 12 lblin2. This sounds logical since the elimination of the buoy would make a safety valve the more necessary, but the fact remains that a weight safety valve features in Beighton's engraving of 1717 along with the buoy and the Potter cord.
Stuart's engraving of what he calls 'Beighton's Hand‑gear' shows a very simple method of working the injection cock by means of two toothed sectors engaging in each other at rightangles, the driving sector being mounted on the horizontal axis of a lever which was alternately moved up and down by two pegs in the plug rod. One of the early engravings featuring a gear similar to this is that showing the engine at Passy (1726) which illustrates the French account of that engine. If this was Beighton's gear, it was soon discarded in favour of a return to the weighted lever released by a detent. Desaguliers describes and illustrates the toothed sector method of opening the injection cock, but also illustrates the weighted lever and detent, which he says is 'more used, and I think a great deal better; because it moves with a Jerk, which is the best way to overcome Friction'. The method by which this lever turned the cock might vary and, instead of the Potter cord, the detent was henceforth tripped by a striker on the plug rod, but the principle of opening by falling weight became firmly established. Experience has shown that the more rapidly and positively the injection cock could be opened, the better the result and it was for this reason that the simple gear credited to Beighton failed to supplant the weighted lever and detent first developed by Newcomen.
It must be emphasised that early pictures of Newcomen engines are not an infallible guide to chronological development. Artists either failed to understand the principle of the valve gear and drew it indistinctly or inaccurately, or else they copied their predecessor's work. Thus the Sutton Nicholls engraving of what purports to be the York Buildings engine (1725) shows the buoy gear only, which is certainly incorrect, while it seems most unlikely that the Dannemora engine would have had the buoy gear unless it was that Triewald doubted the ability of his boiler to supply so large a cylinder. Such vagaries become excusable when we realise that with the sole exception of clock‑work, no other self‑acting mechanism existed in those days.
Besides the alternative admission of steam and injection water to the cylinder, provision had to be made for the exhaustion from the cylinder of the hot condensate and of any air brought in by the steam. When steam was first admitted, the hot water, which had accumulated in the cylinder bottom during the previous stroke, was expelled down a pipe into a hot well. A leather non‑return valve at the foot of this eduction pipe prevented the water being drawn back up the pipe when a vacuum was created in the cylinder. In the earliest engines the hot well was at ash‑pit level, but later it was located above the boiler steam dome so that the hot water it contained could be fed by gravity into the boiler. Beighton is said to have been responsible for this innovation. This first form of feed water heating brought about a sign)ficant saving in fuel.
Apparently Newcomen did not at first appreciate that a certain amount of air would be carried into the cylinder with the steam and he was accordingly mystified by the fact that his engine would gradually lose power until it finally stopped, air having accumulated to such an extent that an effective vacuum could no longer be created. When the cause of this malady, which became known to enginemen as 'wind‑logging', was recognized, Newcomen cured it by fitting a small outlet pipe to the lower part of the cylinder through which the incoming steam could expel any air. Like the eduction pipe, and for the same reason, it was fitted with a non‑return valve. Because of the noise it made, this air outlet pipe became known as the sniffing valve, 'sniff' being then the equivalent of our'sniff'. The pipe was led into a small tank of water which Barney calls a 'sniffing bason'. There any steam which passed was condensed while the air bubbled through it. The overflow from this tank, led either to the hot well or directly to the boiler.
A cistern, mounted a little below the axis of the beam in order to provide a head, supplied the water for injection. It was replenished by a small pump worked, like the plug rod which operated the valve gear, by a small auxilliary arch‑head set closer to the axis of the beam than the main archheads so that the stroke was reduced. In Beighton's engraving the arch‑head for this pump is shown outside the house near the pump end of the beam. Barney's engraving, on the other hand, shows the pump in the engine house and driven from the end of the plug rod. The pumps used for this purpose were usually of the type known as 'jack‑head'. They were of the common lifting type, but the top of the working barrel was closed, the bucket rod passing through a leather‑packed gland in the cover. As the bucket ascended, the water above it was forced up a pipe which branched from the top of the working barrel. If Barney's drawing is correct, however, the Dudley Castle engine was fitted with a plunger force pump of the type introduced by Morland. This pumped on the down-stroke and to cope with this the plug rod is shown coupled to the beam by two opposed chains anchored to the midpoint of the arch‑head. Such an arrangement cannot have worked satisfactorily and a jack‑head pump driven by a second auxiliary arch‑head on the pump end of the beam became the rule.
The water seal on the top of the piston was replenished by a branch taken from the injectionwater supply pipe. Barney's drawing omits this supply pipe and shows both the water seal and injection pipes branching from the pump delivery pipe which delivers into the top of the cistern, an arrangement that could not work satisfactorily and makes the cistern a mere ornament. The top of the cylinder was belled out to prevent the water on the top of the piston from spilling over at the top of the stroke. On the first engines the surplus water from the top of the cylinder was led from the bell‑mouth by an overflow pipe directly back to the boiler. The temperature of this water was never very high, however, and it was led to waste after this arrangement had been abandoned in favour of drawing feed‑water from the hot well.
There was no means of machining the brass cylinder internally: its bore had to be laboriously fettled and smoothed by hand. The piston was usually of cast‑iron and, according to Desaguliers, Newcomen first used as a seal a disc of leather above the piston with its periphery upturned so that it became, in effect, a gigantic cup washer. This leather very speedily wore away in such a manner that the upturned portion broke away to leave only edge contact between the leather and the cylinder wall. Desaguliers goes on to say that Newcomen was delighted to find that this made an effective seal. The use of such seals is confirmed in the description of the engraving by Sutton Nicholls of 1725. The arrangement adopted and standardized on some of the early commercial engines was to cast the piston with an annular flange on its upper side, the outside diameter of this flange being three inches less than that of the piston head below which fitted the cylinder as closely as the techniques of the day permitted. With the piston in the cylinder, soft hemp packing was then coiled and rammed into the space between the face of the flange and the cylinder wall and finally segmental weights were added to hold this hemp packing tight and in place. The function of the water seal was to make this packing effective by keeping the hemp soft and pliable. It could not, as is sometimes supposed, seal a seriously defective or irregular cylinder bore because in such circumstances so much water would pass the piston that creation of a vacuum would be seriously impaired or prevented by excessive condensation of steam on admission.
The beam or 'Great Lever' consisted of either a single massive oak timber or of two such timbers secured together. The timber arch‑heads were mortised into the ends of the beam and securely braced above and below by timber diagonals. The upper‑portions of the arch‑heads were additionally braced by a stout iron rod passing right through the arch‑head and so anchoring the chain by which piston and pump rods were suspended. In all early engines, the trunnions were placed at the mid point of the beam so that the strokes of piston and pump bucket were the same.
Some latter‑day observers have wondered why, throughout the long history of the beam pumping engine, the pump end of the beam and its attendant gear should have been exposed to the elements outside the house. Newcomen rightly decided that the load on the beam trunnion bearings was such that they must be supported by the main wall of the engine house. It was called the lever wall and was more massively built. To enclose the whole would thus have involved the construction of additional walls and roofing which would have proved obstructive when it became necessary to withdraw the pump rods from the mine shaft. Very often these early mine shafts were used for access into the workings and for raising coal and minerals as well as pumping, so it was clearly impractical to enclose the top of the shaft with an engine house. Examples of totally enclosed engines with beam trunnion bearings supported on massive columns or 'A' frames may be seen, particularly in waterworks installations, but in the case of the Cornish mine pumpingengine Newcomen's practice of using wall support persisted right down to the early years of this century when the last Cornish engines were built.
On a non‑rotative beam engine the length of stroke is not positively determined as it is by the crank of any type of rotative steam engine. When the engine was started by the hand manipulation of the levers controlling the steam and injection valves, the length of stroke it made depended on the skill of the engineman. When the self‑acting gear took over control, the length of stroke was still determined by his judgement because successful transition from manual to automatic control depended on the correct placing of the pegs in the plug rod which actuated the valve levers. It was therefore necessary to provide a form of positive stop to prevent the piston coming out of the top of the cylinder or, through mismanagement, knocking the bottom out of the cylinder on its descent.
52 William Pryce in Mineralogia Cornubiens~s, 1778, includes a drawing of a Fire Engine as used in Cornwall. The drawing shows the use of a balance bob to counterweight the weight of the long pump rods.
When first admitted to the cylinder, the steam might exert a small power on the piston, but this diminished rapidly as the piston ascended until, as it neared the top of the stroke, it was literally drawing the steam out of the boiler. But when a new engine had been completed it had to be most carefully balanced to prevent the great weight of the descending pump rods drawing the piston up with a violence which would damage both the engine and the pumps. Weights were added to the beam immediately behind the piston arch‑head and to the piston itself until the beam was in perfect equipoise. A weight equivalent to 1 Ib/in2, of piston area was then removed from the piston to give the pump end of the beam the necessary advantage. Naturally, if the mine was deepened and additional pump rods were added the engine would have to be rebalanced. In calculating the correct balance in favour of the pump end, account was taken of the resistance of the pump bucket as it descended through the water in the pump barrel, the passage of the water past the bucket being restricted by the orifice of the nonreturn valve in the bucket. The danger here occurred if the level of water in the mine sump was so reduced that the pump drew air instead of water. In that event the resistance of the air below the bucket might be insufficient to open the nonreturn valve when it began its descent. If this happened the weight of the whole column of water above the bucket would drive it down with great and damaging violence before compression of the air became aufficient to take effect. To guard against this the engineman was provided with a float‑operated indicator which recorded the depth of water in the mine sump. Even so, some form of positive stop was essential to arrest the descent of the pump bucket and this took the form of sprung timbers known as 'spring beams' which provided a limited cushioning effect before becoming a positive stop. The pictures of the Dudley Castle and Dannemora engines show such spring beams mounted on stout timber stages above the mine shaft. The moving stop connecting with them consisted of a strong iron rod which passed through the top of the archhead and projected on each side of it. This became known as the 'sword'. The alternative, which soon superseded this arrangement, was to mount the spring beams on a staging within the upper part of the mine shaft, the sword being then carried by the first length of pump rod. Presumably this arrangement was used in the engine portrayed by Beighton since he shows no overhead spring beam staging on the outdoor end.
Inside the engine house a precisely similar arrangement was used to prevent the piston from damaging the cylinder bottom if it 'came into the house' too violently. In this case the spring beam stagings were carried on two beams extending from wall to wall of the house as shown by Barney and Beighton. The artist of the Dannemora engine has omitted them in error because the supporting beams are shown. Below these supporting beams two even more massive timbers bridged the engine house in order to carry the cylinder, the latter having a cast flange which rested upon them.
Most early pictures of Newcomen engines show single forged‑link chains coupling piston and pump rods to the arch‑heads. These were soon discarded in favour of pin‑jointed chains with flat plate links which were easier to repair or renew, lay more snugly on the arch‑heads and were more readily coupled to cross‑heads when two or more sets of pump rods were used. Duplex or multiple chains took the place of single chains on both arch‑heads as mines deepened and engines increased in power.
54 The use of quadrants and horizontal connecting rods is shown in this drawing of 'The Slide Engine at Mill Close 1756', one of the London Lead Company's lead mines in Derbyshire. This is an early example of the use of flat rods, widely used in Cornwall in the next century, to pump from a shaft some distance from the engine [Image Deleted]
Little has been said so far about the arrangement and construction of the pumps in the mine. It seems most unlikely that Newcomen was responsible for any notable innovation in this department, but his engine, by bringing much more power to bear, necessitated the rapid development of techniques which were first applied in mines a little earlier. The first mechanical means of raising water from mines consisted of chain‑and‑bucket pumps powered by waterwheels, or by horse gins in situations where water power could not be harnessed. Rodoperated lifting pumps began to supersede bucket pumps in the first decade of the eighteenth century, such pumps being driven by crank from a waterwheel. George Sorocold, an English engineer noted for his water‑powered water supply installations, is said to have been responsible for introducing lifting pumps to the Scottish collieries in 1710 and their use soon spread to Tyneside through the agency of the Earl of Mar who owned property in both districts. It seems probable, too, that lifting pumps were beginning to supersede chains and buckets in Cornwall at the time when Newcomen was making his first experiments there.
While the means available were so inadequate, miners were prepared to go to immense pains to avoid the necessity of pumping. Adit tunnels were driven to carry the water away to low ground and in some districts, notably in Derbyshire, these tunnels were often of great length. The introduction of Newcomen's engine enabled mines to be worked below the level of these edits, but they were still used since it was obviously uneconomic to raise water farther than was absolutely necessary. If the edit level was far below the top of the mine shaft, it meant that the pump end of the engine beam had to carry a great length and weight of dry rods, or 'dry spears' as they were called, which necessitated the use of very heavy counterweights. To avoid loading the beam to such an extent, the practice of using an auxilliary beam or 'balance bob' which was mounted at the mouth of the shaft or as a separate beam above the engine main beam, was initiated. One end of this beam was coupled to the pump rod, while the other end carried a box filled with blocks of stone or scrap metal to serve as a counterweight to the rods. This not only relieved the load on the beam but greatly simplified the task of rebalancing when alterations in the mine made this necessary.
In order to save weight the pump rods or 'spears' were of fir or mast timber in lengths of from 40 to 60ft, the joints between them being scarfed and held together by crossbolts and fishplates. In the case of the 'wet spears' that passed down the pipe through which the water was drawn, the buoyancy of the wood in the water relieved the weight on the engine. To provid access to the mine there was a series of timber stages interconnected by ladders in the mine shaft and on each of these stages the pump rods were guided by rollers to prevent them from bowing as they descended.
Forty‑five yards was considered the maximum lift for a single pump and in practice the single lift was usually substantially less than this. In deep mines this necessitated a series of pumps placed at different stages down the shaft, each drawing from a wooden cistern filled by the pump next below. Such pumps would be driven by rods of smaller section coupled by crossheads to the main rod. Such a series of pumps would be identical except that the lowest, which drew from the mine sump, would have a different suction pipe or 'wind bore'.
As we know from the Edmonstone and Whitehaven engine accounts, the earliest pipes through which the water was lifted were of elm banded with iron, but these wooden pipes were soon superseded by iron pipes with flange joints. The lowest length of pipe next to the working barrel of the pump was of slightly larger diameter than the rest and was provided with an access door. Into this pipe the bucket of the pump could be drawn when it, or the valve within it, required repair or replacement and on this account it was called the 'bucket piece'. Below this, the working barrel of the pump was generally of brass or bell metal to resist corrosion and would be 9ft long if the working stroke was 7ft. Immediately below the pump barrel was a length of pipe called the 'clack piece'. Its expanded upper end formed a conical seating for the clack or foot valve, an access door being provided similar to that in the bucket piece. The foot valve was fitted with an iron loop which could be grappled from above after the pump bucket had been withdrawn in case it required attention when the mine was flooded above the level of the access door. Finally, below the clack piece was the suction pipe or wind‑bore, the combined lengths of these two sections being so determined that the suction lift did not exceed 25ft when the pump bucket was at the top of its stroke.
Unlike those fitted to the intermediate pumps, the lower end of the wind‑bore in the mine sump was deliberately constricted to produce a powerful suction and so reduce the inequality of load in the eventuality of air being drawn in. It terminated in an elongated bulb pierced with holes and the suction could be varied according to need, either by stopping some of the holes with wooden plugs or by adjusting a sleeve of leather over the bulb. The sound produced by a large beam pump as it drew through these holes was tremendous. To miners who depended for their livelihood on the reliability of the pumps it may have been a familiar and reassuring noise, but to the uninitiated it was awe‑inspiring to hear, reverberating through the cavernous darkness of the mine galleries, a sound like the stentorian breathing of some sleeping giant. It was because of this characteristic sound that these holes were aptly named 'snore holes'. In deep mines the lowest pump was generally given a short lift. Th1s was an additional precaution because it minimized the reduction in load which would result if the pump drew air.
The only purpose other than mine pumping to which Newcomen's engine was applied during the inventor's lifetime was that of supplying water to towns, the engines at Passy, near Paris, and York Buildings being the first examples. Here, instead of being lifted from a great depth, the water had to be forced upwards and this necessitated a totally different arrangement of pumps. Two pumps were installed side by side, the one a jack‑head lifting pump of the type used to raise water to the injection water cistern and the other a plunger forcing pump as patented by Sir Samuel Morland and first produced commercially in Great Russell Street, London, by Isaac Thompson. The crosshead linking the two pump rods to the arch‑head chains was guided by grooves cut into two upright timbers. The jackhead pump rod was of round section iron, turned and polished where it passed through the leather pump gland. The middle portion of the force pump rod consisted of two wooden planks holding between them weights of lead or pig‑iron. The engine raised the jack‑head pump bucket on its power stroke, while the descending weights actuated the force pump and returned the engine piston to the top of the cylinder. The delivery pipes from the two pumps were led into a closed cistern or receiver in which the compression of air acted as a balancer, forcing the water in a steady stream up a rising main to a storage reservoir placed at a sufficient height to provide enough head for the town supply pipes. According to Triewald the York Buildings engine pumped to a reservoir containing 'several thousand hogsheads' from which water was conducted through lead pipes to every floor of 'over 500 many‑storied houses' in the vicinity of Hanover Square, now Hanover Place off Long Acre. On each floor, he tells us, two taps were provided, one for domestic use and the other a'fire tap' with a threaded outlet to which a leather hose could be attached. Such a precaution reminds us that the great fire of London was then still recent history.
So concerned was Triewald to 'sell' the new power to his countrymen in Sweden that he made claims for Newcomen's engine the extravagance of which would have shocked the inventor. His most gross exaggeration reads as follows:
As to the durability of this machine it certainly possesses no small advantage over other art)fices; because the noblest part of the machine are made of metal, copper, lead or iron and ought thus, as a matter of course, to be able to defy time, nay, a cylinder, after a hundred and even a thousand years' use, is better, and never can be worse.
So far was this from the truth that some engine cylinders required renewal within Newcomen's lifetime. This might not be due solely to bore wear but to casting defects disclosed either by wear or by the 'working' of the cylinder on its wooden bearers when under load. Of the twenty‑three number and size, from three mines with 47in engines in 1746, to a 70in engine at Herland Mine in 1753. Borlase, writing in 1758 remarks upon the number and size of the engines then at work and refers to engines at North Downs mine, Redruth (2), Pit Louran, Redruth (2), Polgooth, Wheal Reeth, Bullen‑garden, Dolcoath, the Pool, Bosproual and Wheal Rose. In 1769 when the north country list was compiled, John Smeaton collected particulars of eighteen large engines then at work in Cornwall, eight of them having cylinders exceeding 60in diameter. Nine years later it was stated by Pryce in his Mineralogia Cornubiensis, that more than sixty engines had been built in Cornwall since the coal duty was remitted in 1741 (although at least another ten engines built between these dates can be accounted for now) and that many of these had subsequently been rebuilt and enlarged.
The most notable Cornish engine builders at this time were Jonathan Hornblower (son of Joseph, Newcomen's associate), John Nancarrow (who had provided Borlace's list of engines) and John Budge. The engines built in Cornwall were of a higher standard than those in the north of England, for fuel economy was all‑important and provided a great incentive for engine builders to strive for greater efficiency. Moreover, Jonathan Hornblower disseminated a wealth of experience which he had inherited from his father.
It is appropriate that it should have been Josiah, a younger brother of Jonathan Hornblower, who was responsible for introducing the Newcomen engine to the New World. This historic engine was ordered in 1748 or 1749 by Colonel John Schuyler who, with his two brothers, owned a copper mine in what is now North Arlington, New Jersey. Copper had been found on the Schuyler estate in 1715 and was profitably worked by driftways until 1735 when it became necessary to sink a shaft. The ore was exported to the Bristol Copper and Brass Works where it fetched 40 a ton. When the shaft reached a depth at which the water could no longer be cleared by horse power, John Schuyler made the inquiries in London which led to his order.
Josiah Hornblower was chosen, perhaps by his better known elder brother, to erect the engine and on 8 May 1753 he set sail from Falmouth on a coasting ship bound for London where the engine parts, many in duplicate and some in triplicate, had been gathered ready for shipment.
With the engine and its erector on board, the American ship Irene sailed from London on 6 June 1753 and encountered such rough weather and adverse winds in the North Atlantic that she did not reach New York until 9 September. The rigours and perils of the crossing were such that Hornblower swore he would never make an ocean voyage again. Much to the sorrow of his family in Cornwall he kept his vow and never returned.
At New York the engine was trans-shipped to a smaller craft which carried it through Newark Bay and up the Passaic River to an unloading point at Belleville opposite the mouth of the Second River. It was then carted overland for about a mile to the head of the mine shaft which was located near the junction of Belleville and Schuyler Avenues in North Arlington. So arrived the first steam engine in the American continent, an event less celebrated than the landing of the Pilgrim Fathers but no less pregnant with sign)ficance for the future.
Another Cornish engineer who emigrated to America was John Nancarrow, one of the most notable engine erectors of his day. Although still in Cornwall in 1757, by 1786 he was operating a steel furnace in Philadelphia when he was consulted about building an engine for a steamboat.
James Brindley was associated with the Newcomen engine in the mid‑eighteenth century. In 1756 he erected a 36in engine for Thomas Broade at Fenton Vivian in Staffordshire. According to details of this engine which Carlisle Spedding sent to his friend William Brown, Brindley mounted his cylinder in the wall of the engine house opposite the lever wall. It would seem that Brindley did this, not because he was aiming at a more rigid structure, but because his patent boiler could not be conveniently accommodated within the engine house. This boiler consisted of a brick vault 18ft long with a floor of cast‑iron plates over four small furnaces. Each furnace had its own iron flue pipe which passed back through the water in the boiler before entering the chimney. The parts for the engine were supplied by the Coalbrookdale Company, and their accounts show that Brindley was still experimenting with his boiler in 1759. It was not a success. Although Brindley erected a number of engines in North Staffordshire and elsewhere he was relatively little involved in steam power. Compared to the many other engine builders, such as Wise, Nancarrow, Hornblower, Budge, Curr, Brown, Smeaton, Thompson and other who are even less well‑known today, Brindley's contribution was very small indeed, and he is best remembered as a canal engineer.
Rev. March 2010