Newcomen engine automation
newcomen-steam-engine
Before a young lad by the name of Humphrey Potter [12] came up with a means to fasten strings and catches to the working beam so that the machine's motions could open and close the valves, these activities were carried out by hand. The engine then moved one step closer to being a self-regulating device when more permanent appendages were added to serve the goal. It was called the Atmospheric Engine when it was in this straightforward and effective stage. Around 1712, the Newcomen steam engine was refined to this level, and such engines were erected in various places in England.
Automation of continuous motion
Desanguliers second volume of his book "A Course of Experimental Philosophy," which was published in 1744, a little over 30 years after the first engines were installed, appears to be the first longer description of the Newcomen steam engine. Desanguliers was a practical person as well. For a Newcomen engine, he created a safety valve, for instance. How much first-hand experience he had with the Newcomen steam engine is unclear, though. Desanguliers hardly ever cites his sources, as was customary at the time. He appears to be the major source of inspiration for other publications, which makes them equally unreliable. Despite his academic and engineering credentials, the descriptions in his book are excessively wordy and unconcise, as it seems to be customary at the time. Hence, using mostly the copper plates No. 37 and 38 in Desanguliers book as a guide, I created from scratch a description of the valve gear of the Newcomen engine in the manner of a patent application.
The philosopher John Theophilus Desagulier (12 March 1683 – 29 February 1744) attributed to his friend Henry Beighton the perfection of the controls of the Newcomen engine. Desaguliers describes a steam engine of the type Newcomen that Henry Beighton had constructed some years later in Newcastle on Tyne in 1718. Beighton removed all of the catches, leaving the beam itself to supply everything much more effectively [13]. This is done by fixing another arch to the great beam from which by a chain hangs a perpendicular plug frame Q, also called plug ord or plug tree. This plug frame Q having a slit in it, and several pins, gives motion to several levers, which open and shut the regulator (valve plate for admitting the steam to the cylinder) and Injection-cock at proper times.

Such a device to control input valve, the injection valve, and if present, an exhaust valve, is later called a valve gear.
If a different timing was required, control pins could be inserted in the multiple holes of such a plug frame Q and moved to a different location with ease. The plug frame Q in the steam engine built in Newcastle upon Tyne has four control pins. The opening of the regulator valve is controlled by a first control pin p that is inserted into the side of the plug frame Q opposite the viewer. The closing of the regulator valve is controlled by a second pin that is positioned between the slit of the plug frame Q. The third and fourth control pins, r and s, positioned at the side of the plug frame Q that is facing the viewer, controlling the injector valve's opening and closing.

The regulator RSYZ is composed of a circular lid R that is installed over the boiler's top opening and below the working cylinder. A throat pipe S, attached to the upper side of the lid R, provides a fluid connection between the boiler and the cylinder. A valve plate Y is pivotally mounted with a square shank z to a pivot arrangement v w x below the throat pipe S. The pivot arrangement vwx comprises a square cone v, with a hollow inside that is also a square cone through which the square shank z form-fits and passes through from a side below the lid R to a side above the lid R. Thus the square cone v and the square shank z share a same pivot axis, which is directed vertically upright and perpendicular to the valve plate Y. Compared to a 4 inch diameter cock-valve, which is of a similar size, this regulator valve's construction only caused 1⁄10 of the friction. [Desanguliers].
A regulator spanner PQ has an end with a square opening and an arm that is split in its middle in two branches for receiving a flat end O of a forked rod MON between the two branches. The branches are fitted with several through holes, allowing to pass a fastening pin q through these through holes and a corresponding hole of the flat end O of the forked rod MON. The square shaped closed end P of a regulator spanner PQ engages the squared cone v of the regulators pivot arrangement vwx with its closed end P. When the forked rod MON pushes in the direction of the regulator, the spanner PQ pivots the plate Y anti-clockwise into the opened position opening the fluid connection, allowing steam from the boiler to enter the cylinder. Conversely, when the forked rod MON pulls with the flat end O of the forked rod MON towards the plug rod, the spanner PQ pivots the plate Y clockwise and the fluid connection is interrupted, preventing steam entering the working cylinder. As there is a slight overpressure in the boiler the regulator plate is pressed against the lower side of the throat pipe S and therefore seals the fluid connection sufficiently.

The regulator spanner PQ is connected to a rocking mechanism by a forked rod M O N. The rocking mechanism, as will be described below in more detail, toggles between two positions and by pulling and pushing on the forked rod M O N unlocks and shuts the regulator instantaneously. The forked rod M O N has a single end O, that is movably connected to the end of the arm Q of the regulator spanner PQ, and two fork arms M, N on its other end. The two fork arms M, N, are suspended by a stirrup I K from an arbor tree AB.

A first spanner G4 with a first squared opening G at one end of the first spanner G4 and a curved first arm 4 on its other end, a second spanner H5 with a second squared opening H at one end of the second spanner H5 and a second curved arm on its other end 5, and a three-armed spanner CDE with a third squared opening e where the three arms meet, are lined up with their square openings on an arbor tree AB. Round pegs located at the ends of the arbor tree AB allow the arbor tree AB to be mounted rotatably. As all three squared openings form are fit on the arbor tree AB, when one lever is pushed by a pin the other levers are forced to follow this movement as the pins for proper function of the valve gear are always arranged not to block the movements of the levers with one another.
The three armed lever is known as a Y-lever because it resembles the letter "Y" when inverted. When mounted on the arbor tree AB, two arms of the Y-lever are hanging down and one arm is pointing upwards. The arm hanging down, stretching to the right (towards the plug frame Q), is called in the following the unlocking arm D and the arm hanging down, stretching to the right (towards the regulator) is called in the following the shutting arm E. The arm pointing upwards is named in the following tumbling arm C as it is carrying a weight F, called in the following a tumbling bob. The height of the tumbling bob F can be adapted by a key.
The tumbling bob's motion is constrained by a leather strap. By fastening the middle of the leather strap to the tumbling bob F the leather strap is divided into a right halve with a right end n and a left halve with a left end m. The right end n of the leather strap is fixed right of the tumbling bob and the left end m of the the leather strap is fixed to the left of the tumbling bob F. Thus the rocking mechanism is a bistable dynamical system having two stable equilibrium states. The first stable state is when the tumbling bob is on the left side, pulling on the right halve of the leather strap, and the second stable state is when the tumbling bob is on the right side, pulling on the left halve of the leather strap. In both stable states the potential energy of the tumbling bob is zero (unless the leather strap breaks).The tumbling bob attains its maximum potential energy when it is sitting above the tumble arm, i.e. the tumble arm is vertically orientated. As soon as the tumbling bob F crosses the vertical line, its torque switches to the opposite side, causing the tumbling bob F to fall to that side. By falling, the tumbling bob's potential energy is transformed into kinematic energy, causing the arbor tree to accelerate. When the opposite leg strikes the stirrup, the linkage of the fork and the spanner opens or closes the regulator valve in a rapid action..

An injection cock, embodied as a stopcock, is used to interrupt the flow of water from an injection cistern into the cylinder. The plug of the cock has a narrow, long, upright hole. A squared cone serves as the plug's stem and can accommodate a toothed wheel with a corresponding square opening at its pivot axis. A third lever, commonly known as the F lever has at its pivot axis a segment of a second toothed wheel. The pivot axis of the second toothed wheel is perpendicular to the pivot axis of the first toothed wheel and arranged so that the first and the second toothed wheel engage. The first toothed wheel lays in a horizontal axis, the first pivot axis thus orientated vertically. The second toothed wheel thus is in a vertical plane. When the F lever is lifted or pulled down the first toothed wheel is rotated in the horizontal plane and turns the plug thereby opening or closing the fluid communication between the cold water cistern and the inside of the cylinder.
Let's refer to the side of the beam that is used to drive the pump as the pump side, and the side of the beam that the piston is suspended from by a chain as the cylinder side. We define the start of a cycle as the point when the piston changes it's direction, immediately following the condensing of the steam in the cylinder. I am proposing this definition as in few cases steam engines were erected with the volume to be condensed being above the piston. However, to keep it simple, for the following description we use the geometry as depicted in the pictures.
At the beginning of a cycle the tumbling bob F hangs on the side of the beam rod and therefore the regulator valve is opened. Conversely, the injection-cock is closed. The cylinder side of the beam is in its lowest position, pulled down by the vacuum that was formed at the end of a previous cycle.
As the regulator valve is open the pressures above and below the piston are essentialy the same. Without a pulling force on the cylinder side of the beam the counterweight force prevails and lowers the pump side of the beam. Conversely, the pin rod Q rises along with the cylinder side of the beam. The pin located in the slit of the pin rod Q is the first pin to make contact with a lever. It raises the second lever H, which in turn rotates counterclockwise the arbor tree. By being pushed out of its stable position, the tumbling bob F accelerates the rotation of the arbor tree once it has moved from the rod beam side to the cylinder side. As a result, the control leg D of the Y-lever CDE that is on the cylinder side is pushed by the stirrup L toward the pin rod Q. When the forked lever is abruptly pulled by the stirrup L, it pulls abruptly at the regulator lever Q, closing the regulator valve instantaneously and interrupting the steam flow from the boiler to enter the cylinder.

The drawing on the right illustrates the situation when the thumbling bob F has crossed the vertical line and is beginning to fall in a circular motion toward the regulator side just before the leg E engages with the stirrup L.
Once the next pin comes into contact with the cock lever and opens the injection cock, the cylinder side of the beam is still rising.e cylinder and starts to condense the steam. As a result an increasing vacuum is created in the cylinder. As a result, the air pressure experiences no counterpressure and pushes the piston downward, thereby lifting the counterweight and the water in the pump.
A pulley on the first pin pulls the first lever down as the pin rod Q descends, forcing the arbor tree AB to rotate in a clockwise direction and shifting the tumbling bob from its stable position to the unstable vertical position. The lever leg D pushes the stirrup L toward the regulator and instantly closes the regulator valve once the tumbling bob F has left the vertical position and is falling downward in a circular route toward the pin rod Q. This completes a cycle and a new cycle can begin.
What is the right speed?
Newcomen and all the other people who initially built "fire engines" appear to have tweaked their engines until they were running while being ignorant of the fundamentals. They appear to have neglected to recognize that power is defined as work per unit of time and set up their engines to deliver the most lifting power possible instead of optimizing the power delivered over a given time. When Smeaton optimized already installed Newcomen engines his modifications increased the engine speed by little more than 50%. An atmospheric engine can be made to run faster by preventing the condensation process from completing, for instance by opening the regulator valve earlier. As a result, less work is produced per stroke the earlier the condensation process is stopped. The increased speed, however, initially outweighs this loss of work.

Due to the lack of performance information for atmospheric engines operating at different speeds, or strokes per minute, I developed a small model based on the hypothesis that, after a certain point, the amount of work produced by each stroke decreases exponentially. I further assumed that the maximum power for a Newcomen engine operating at 8 strokes per minute is attained at roughly twice this speed. As a reference point, 8 strokes per minute are used. The generated work corresponds to the maximum amount of work that can be done up to this point or at this speed, respectively. As a result, the relationship between power and strokes per minute up to this point is linear. The power produced by four maximum-power strokes per minute is exactly equal to half the power produced by eight maximum-power strokes per minute. A stroke no longer reaches its full power after this point, although the increased speed still generates more work per minute. The work done each stroke now decreases faster than the increase in speed can make up for it, indicating that the maximum fraction of work per minute has finally been reached.
Due to the lack of a second or third reference point, the curve may be flatter or steeper near the maximum. The graphic does, however, accurately depict that Newcomen engines would operate more effectively at greater speeds. The roadblock in the mind of Newcomen and contemporary engine erectors was presumably to realize that, in order to enhance speed, the weight of the water column to be lifted with each pump stroke needed to be decreased in order to match the lower lifting power of an uncomplete condensation. The idea that a smaller diameter for the pump tubes would allow for improved performance was beyond the comprehension of Newcomen & Co.
-----------------------
[6] Thomas Tredgold, The Steam Engine, Vol. I, London, 1838
[7] H.W. Dickinson, page 29/30
[8] Ibid, page 33, referring to Transaction Newcomen Society, XVII, page 6.
[9] H.W. Dickinson, page 33.
[10] Ibid.
[11] J.T. Desaguliers, "A Course of Experimental Philosophy", London, 1744, page 474
[12] J.T. Desaguliers, "A Course of Experimental Philosophy", London, 1744, page 533
[13] Ibid.
[15] James Greener, Thomas Newcomen and his Great Work, October 2015, available on ResearchGate
[20] Mårten Triewald, Mårten Triewald's Short description of the atmospheric engine: published at Stockholm, 1734, Translated from the Swedish [by Are Waerland] with foreword [by Carl Sahlin], introduction [by Rhys Jenkins] and notes [by Are Waerland] https://books.google.ie/books?id=2IWFAAAAIAAJ
[21] Ibid.
[22] Ibid.
[25] Switzer, Stephen (1729) ‘An introduction to a general system of hydrostaticks and hydraulicks’, page 342 available Internet Archive
[30] John Farey, page 132
[31] John Farey, page 131