Development of pumps and steam engines Part 1

Simple suction pumps had been shown in a manuscript called De Re Metallica by Agricola from 1556.

Illustration of two lift suction pump for mine drainage by Agricola c1556

These pumps worked on the principle of a piston in a tight fitting cylinder lifting the water by applying a vacuum to the top of the water column.

In 1640 Torricelli had shown that these pumps were limited by atmospheric pressure. Water could not be raised more that 33 feet at a time.  For deeper mines it was necessary for each pump to discharge it's water into a tank not more that 33 feet above the base of the pump.  From this tank a second pump would pump it another 33 feet.

The limit of 33 feet was an absolute maximum and inefficiencies in the suction pump often meant that a much lower actual pumping head was achieved.

The first steam driven engines were designed by Thomas Savery who was an officer in the British Army. He came from Cornwall and was familiar with the atmospheric pumps used in the copper mines at the end of the 17th century.

In 1698 Savery demonstrated his new improved " fire pump" to King William III at Hampton Court.  The main parts of the of the pump consisted of a boiler ,connected to a tank. At it's base the tank was fitted with two pipes for water inlet and outlet both closed by valves. The boiler and tank could be isolated by a valve.

Diagram of principles of the Savery engine	Diagram in the 1702 patent for the Savery fire engine

The water in the boiler was heated by an external fire to raise steam. This steam entered the boiler where air was displaced via the outlet pipe. Water remaining in the outlet pipe held the outlet flap valve shut preventing steam escaping. Once sufficient steam had entered the tank the connecting valve was shut. Cold water was then poured onto the sides of the tank.  This condensed the steam in the tank and pulled a partial vacuum which opened the inlet valve and allowed water to enter the tank forced in by atmospheric pressure.

When the tank was almost full of water the isolation valve was opened and more steam admitted. This raised the pressure in the tank, closed the inlet valve and eventually forced water out of the outlet pipe. The steam pressure in the tank meant that water could be raised higher than an atmospheric lift pump. The Savery pump had two tanks connected to the boiler which filled and emptied alternatively.

The Savery pump was used in some Cornish copper mines from the start of the 18th century.  It had a practical limitation of raising water around 100 feet in a single lift.  A boiler operator was required to switch over between each boiler. His job was a dangerous occupation as in the absence of pressure gauges or regulators the steam pressure had a tendency to burst the boiler

In around 1711 a new system known as Newcomen's engine was beginning to replace the Savery engine.

The Newcomen engine condensed steam in a cylinder to cause a vacuum which pulled on a piston.  The action of the piston was transferred to another piston in a valved pump arrangement to move water. No high pressure steam was used. The engine was described by it's inventor as a "fire" engine and not a "steam" engine. It's main motive force was from atmospheric pressure.

The Newcomen engine was called an atmospheric engine because the work was performed by the atmosphere reacting against a partial vacuum created by condensing steam. At 100°C (or 212°F as would have been current at the time) steam occupies a volume of approximately 1660 times greater than the same volume of water (both at atmospheric pressure).

The water in the boiler was heated to raise steam.

At the start of the power stroke, the steam valve closes and the injector valve opens. This allows water into the cylinder which condenses the steam, and atmospheric pressure at the top of the piston forces the piston down (and the opposite end of the beam up) so the pump is stroked upwards.

At the end of the power stroke, the injector valve is closed and the steam valve opened. As the steam enters water, from injection and condensed steam, is expelled into the feed tank. The weight of the mechanism at the opposite end of the beam would pull the piston up, so allowing steam into the cylinder. The pressure of the boiler (3-4psi) was not sufficient to do this alone.

When the piston is at the top of its stroke, the steam valve is closed and the injector valve opened, and the cycle repeats itself. This procedure was repeated around  6 or 8 times per minute.

The crossbeam vibrated 12 times/minute and for each stroke, the pump piston sucked 45 litres of water which was then pushed to reach a height of 46 metres. The ends of the cross beam where attached by chain to the piston at one end a pump rods descending to a pump at the bottom of a mine shaft at the other. The crossbeam was balanced on a specially strengthened wall of the engine house.

The engine was a single acting condensing cylinder type and that the power stroke was not the steam admission stroke.

The first pump and engine used to raise water from a mine was installed at Dudley castle coal mine in the midlands in 1712. 

The initial pumps suffered from poor fitting pistons and cylinders. The methods used to control the piston movements and  condensation efficiency improved over the following years. In 1717 Henry Beighton of Newcastle invented a safety valve which prevented boiler failure. By 1720 the engine was much improved and for 50 years remained unaltered giving good service in many mines and other applications where water pumping was required. It was often applied to pump water to a water mill which was then used to drive machinery. However the continual need to heat and then cool the cylinder meant that the fuel usage (coal) was in far excess of the amount of work done.

There were limitations to the Newcomen engine The areas of concern, which we see nowadays, were fires and their shape, the pressure which the boilers could stand, the inefficiency of the single stroke engine, the inefficiency of cooling and the excessive coal consumption. The fires were several feet deep, which meant that combustion was inefficient. There was insufficient understanding of chemistry at that time to allow for the fact that oxygen had to get to the coal to allow it to burn. The chemistry of combustion is not a simple

C + O₂ = CO₂

but is a combination of two reactions which occur at different times, producing an intermediate product of carbon monoxide.

C + ½ O₂ = CO   and  CO + ½ O₂ = CO₂

Both reactions are exothermic (heat producing), otherwise they'd be no good for making a fire! But what you do need is an oxygen supply for both of the above reactions to work, and be forced to the right. What happens in a deep fire is that you have oxygen starvation, the fire at best will glow a dullish red rather than a bright orange and the heat transfer will be limited in efficiency. The speed of heat transfer relies on the difference in temperature, so a much slower rate of heat transfer will occur if you have a dull red fire rather than a bright orange one. It wasn't to be until the end of the eighteenth century that changes in technology and knowledge of heat transfer processes would bring a difference in fire shape and hence efficiency, about.

Another limitation to this technology was the construction of the boilers. Copper was the favoured material, but this meant that pressures had to limited to 3-4psi.

Once wrought iron plated became available and the technology for riveting them together, then boilers could be made to withstand higher pressures. However, in the absence of secure riveting technology, another method was needed. What happened was that when you'd joined to plates as best you could you packed any gaps with iron filings and rammed them home. Then you expanded the iron filings by oxidising them, i.e. you made them rust. In order to do this you sealed the outlets with clay to prevent leakage and made a reservoir of the same material to contain your fluid to produce the final reaction:

2Fe + 3O₂ = 2Fe₂O₃

It speeds the reaction if you make your fluid acid and apart from peaty water (a very dilute acid), gone off beer (which you'd use as vinegar) the most readily available material for free was urine. When your reaction had proceeded, and your joint allowed no escapes, then you could proceed to the next section. It might be expected that boiler explosions were not uncommon!

The third factor to any limits of efficiency was the sealing mechanism at the top of the cylinder. Cylinders had to be large - the active force was a maximum of only 14.7 psi - the atmosphere and obviously less because of the impossibility of producing an absolute vacuum and 100% efficiency. Technology did not exist for producing a perfectly circular cylinder or piston, so the seal was water, available and topped up occasionally.

Coal consumption was a problem. A large engine could consume twelve tons a day. Pumping engines had to run seven days a week 24 hours a day. They couldn't have a day off on Sunday otherwise on Monday morning you couldn't actually start work because of the water. So just consider what is involved:

• 1 packhorse load is 3cwt

• engine consumes 12 tons of coal per day

• therefore (12×20 /3) packhorse load per day = 80 packhorse loads per day

• 1 packhorse train - 40-50 horses, so 2 packhorse trains per day.

• 1 50 ton ship could carry perhaps 20 tons of coal

• You need 12/20×7 = 4.2 shiploads a week just to keep one big engine going.

Add to this the manhandling, the unloading, the horse forage, the food for the men and it was a very large undertaking to have just one pumping engine on your mine.

And of course, the Newcomen engine was not efficient by today's standards - it was the best that could have been done at the time. You could not heat up and cool the cylinder completely every stroke, the system would get warmer and so less efficient as the vacuum would not be created at every stroke. The vast coal consumption was prohibitively expensive. It only acted on one stroke. The pump was however used in many mines until the development of the Watt engine.

Old mines have a lot of bends and twists in them. The advent of steam engines meant that something had to be straightened out somewhere.

You need to transfer the pumping mechanism to where the water was, i.e. in the sump at the bottom of the mine. You also needed to transfer the power down there as well. In the following discussions about engines, we will use reciprocative motion as supplied by the beam engine.

It was most convenient to have your power source on the surface and the pump in the sump. Alternatives are having a boiler in the mine and having to bring coal and air down. Burning a fire at the bottom of a shaft was practised in Midlands coal mines to ventilate by convection. Transference of power between them was by coupling them with a long rod which is attached to the end of the beam. However if you allow the long rod to hang from the beam it is going to be pretty heavy, and will break your beam, if not the rod itself. Bear in mind that we're taking about timber and cast or wrought iron for the materials that were able to be used. If, however, you arranged counterweights down the rod, then you would only be overcoming the momentum of the mechanical system rather than supporting a heavy load. You did have to have some weight acting on the beam, because it was this weight which cause the admission of steam into the cylinder on the Newcomen style engine. As the pumping requirement increased as the depth of mining increased, then you extended the pumprod, increased the counterweights and did not change the mechanics of the systems much.

The next great step forward in steam pumps came when a young instrument repairer called James Watt was asked to repair a model of a Newcomen engine. 

Part 2