The Tower Spherical Engine.

Updated: 28 Apr 2011
More info & comments
Back to Home PageBack to The Museum

The Tower Spherical engine was a unique contender in the rotary steam engine market. It was not merely cylindrical, like most such engines; it really was spherical. It got a good deal of publicity, being described in detail in journals based in both Britain and France; it was praised by experts; and it did at least one useful job of work driving carriage lighting dynamos mounted on the locomotives of the Great Eastern Railway. The line that was once the GER runs about 70 yards from where I am typing now.

Left: The Tower Spherical Engine.

Now this is a real beauty.

The operation of the engine is not easy to comprehend, but goes something like this: The "cylinder" is spherical, and contains two quarter-spheres, with a thin circular disc between them. The two quarter-spheres rotate and engage rather like a universal joint, creating four cavities in the sphere, two of which are expanding and two contracting at any moment. By suitably timing admission and exhaust, rotational power is generated.
Hopefully the lower diagram goes some way to explaining this.

The prototype was intended for the direct drive of dynamos, and had a speed range of 600 to 1100 rpm. A 10-inch engine (presumably this means one with an internal sphere diameter of 10 inches) was reported to consume 37 pounds of steam per BHP-hour. The worm on the right operates a small reciprocating pump for forced lubrication.
This engine did at least one useful job of work driving carriage lighting dynamos mounted on the locomotives of the Great Eastern Railway; this was reported in 1885 as having been the case for some time.

Inventor: Mr Beauchamp Tower

Sources: Modern Mechanism, Park Benjamin. McMillan, 1892.
Proc. Inst. M.E.,1885

The introduction of the electric dynamo, which was best driven directly by a high-speed engine, gave a new lease of life to rotary steam engine development. If these engines could have been made to work efficiently, they would have been ideal for the purpose. Note that the Hult engine was said in 1902 to have been used for driving lighting generators in both the Russian and Swedish navies. This was clearly shown at the International Inventions Exhibition at South Kensington, London, in May 1885, where several engines were shown that were specially designed for driving dynamos directly. These included The Parsons Epicyclic Engine, and of rather greater import, the Parsons steam turbine. Ironically, the Parsons turbine was illustrated on the same page as the Tower engine.
In the event, most generators of the day were driven by Willans or other relatively conventional high-speed reciprocating engines.

Above: An engraving of the Tower engine, from the French science journal Nature: 1884.

The Nature article adds little to the description of the engine, but does give some performance figures, thus:
"At 500 rpm, a Tower motor with a 10 cm sphere can produce one brake-horsepower with steam at 3 atmospheres (44 psi) and 2.5 brake-horsepower if the pressure reaches 8.5 atmospheres (125 psi). A 20 cm motor produces 7 brake-horsepower at 3 atmospheres, 10 brake-horsepower at 4 atmospheres, 15 brake-horsepower at 6.5 atmospheres and 20 brake-horsepower at 8.5 atmospheres. The biggest model, of 65 cm diameter, produces between 234 and 624 brake-horsepower, depending on the steam pressure. Unfortunately we have no figures on the steam consumption corresponding to these different pressures. It is to be feared that it is much greater than motors with conventional cylinders and pistons."

Left: Animation of the Tower engine

Note that the central disc rotates with the two segments.

This superb animation is kindly provided by Bill Todd

There was no information on why the author thought the steam consumption would be greater with a rotary engine- but almost certainly it would be the old story of leakage due to packing problems and inadequate steam expansion due to the cramped geometries usually employed. He seemed to take it pretty much as a foregone conclusion, however, and inn the case of the average rotary engine, there is no doubt he was justified.
The author concluded that the higher steam consumption was usually justified by the smoothness of rotation. However, potential customers seem not to have agreed, as the Tower engine has left little mark on history.

Left: Illustrations of the Tower engine, from the Engineering report on the International Inventions Exhibition.

Below is the complete text of an article on the Tower engine that appeared the British journal Engineering. The date was March 27th, 1885.

Left: A contemporary article on the Tower engine.

At the bottom of this column R E B Crompton says he has "used a considerable number of these engines for driving dynamo machines with capital results" and also reveals that the Tower engines were used with success by the Great Eastern Railway on its locomotives for driving dynamos. Unfortunately I don't currently have access to the issue of Engineering referred to.

They were also used by the Admiralty to drive lighting generators aboard ship. The absence of vibration would have been a definite advantage.

From Engineering March 27th, 1885

Left: The article continued.

Mr Woodley, is enthusiastic, having used Tower engines to drive Crompton dynamos, and having found them reliable.

At the bottom of the column Professor Kennedy is less impressed. He does not believe the concept is new; he makes a tantalising reference to a "disc engine" driving the printing presses of The Times. This is believed to refer to the Ericsson disc engine.

From Engineering March 27th, 1885

Left: The article continued.

The reference to "Heenan and Ford" just below the diagram is a misprinting of "Heenan and Froude", a Manchester firm that manufactured the Tower engine. (They also built the 518ft high Blackpool Tower) "Heenan" refers to Mr R Hammersley Heenan, who wrote a paper on the engine in the Proceedings of the Institution of Mechanical Engineers (1885) on the Tower engine. See below

It is known that the firm put out illustrated advertisments for the engine, as the University of Glasgow has a copy; regrettably you can't see it on line.

The descendant of this company is still very much alive, and making dynamometers: see www.froudehofmann.com (external link)

From Engineering March 27th, 1885

Left: The article concluded.

As noted elsewhere, in the rotary engines gallery, George Stephenson was no fan of rotary steam engines.

From Engineering March 27th, 1885

Left: The start of a paper by R Hammersley Heenan on the Tower engine

This confirms that Beauchamp Tower got the idea for his engine from observing a universal joint in action.

From Proceedings of the Institution of Mechanical Engineers
Volume 36, March 1885

Students of the rotary steam engine galleries will have noted the close relationship between rotary engines and rotary pumps. The Tower engine appears to have resurfaced as a pump in the late 1950's, though what, if anything, happened in the intervening period is unknown.
To quote the source:
"When the rotor revolves, four sets of flat surfaces open and close for a pumping action of four discharges per revolution."

Left: The Tower engine reborn: the universal-joint pump.

There appears to be some kind of spring-loaded poppet valve at the top of the casing; its function is so far unknown, but it may be a pressure-relief valve. Note that the stub axle is now pointing upwards rather than downwards."

From Pump Selection & Application by Tyler Hicks, pub McGraw-Hill 1957


MR BEAUCHAMP TOWER (1845-1904)

Left: Mr Beauchamp Tower.

This is one of the very few occasions when the biographical details of a rotary-engine inventor can be supplied.

Beauchamp Tower was the son of Robert Beauchamp Tower, rector of Moreton, Essex. He was educated at Uppingham School in Rutland. He decided at the age of 16 to be an engineer and trained at the Armstrong Works at Elswick, where he stayed on for a few months as a draughtsman after completing a four year pupillage.

His major claim to fame was not his engine, but the immensely important discovery of full-film or hydrodynamic lubrication. He was appointed in 1882 by the Institute of Mechanical Engineers to study high-speed bearings and showed that with a suitable supply of oil the bearing surfaces never actually touched but were kept apart by a continuous film of lubricant; this is fundamental to the design of high-speed machinery. He continued this work until 1891. It is therefore not surprising that the Tower Engine had a pump for forced lubrication.

Beauchamp Tower is a major figure in the world of tribology (the study of friction and bearings) but is almost unknown as a steam-engine designer.

Beauchamp Tower conducted trials of motors in 1887 for electric lighting, with John Hopkinson and Kennedy, for the Society of Arts (Journal of the Society of Arts, February 1889). He invented an extended slide-rule with logarithmic scales on metallic tapes, unwound from one roller to another.

I was planning to open this page with a remark that the Tower Engine was nothing to do with Tower Bridge in London. Fortunately I didn't, as there does appear to be some sort of link.

Henry Marc Brunel (born 27 June 1842, London) was the second son of the famous Isambard Kingdom Brunel. Throughout 1886 Beauchamp Tower and Henry Brunel were involved in working on Tower Bridge. From 1880 Henry Brunel was a member of a company promoting Tower's Spherical Engine; another member of the syndicate was Sir Frederick Bramwell. Tower had to demonstrate that the spherical engine could run satisfactorily for three million revolutions before the syndicate would back it. This was achieved and it is said that a considerable number of the engines were then built, though the actual number has so far escaped detection.

One of the strangest things about Mr Tower was that his name echoes part of the Tower of London, which is of course very close to Tower Bridge. There is a "Beauchamp Tower" built around 1275-81, and named after Thomas Beauchamp, Earl of Warwick. I wrote here earlier "I can only assume that Mr Tower's parents had a rather peculiar sense of humour."
In fact the situation is a bit more complicated. His grandfather, Christopher Thomas Tower (1775-1867) married Harriet, daughter of Sir Thomas Beauchamp-Proctor in 1803, so this is where "Beauchamp" enters the Tower family. It still, however, seems a bit strange to take half of a double-barrelled surname and use it as a first name, and it is hard to see that the Tower of London reference is purely a coincidence.

Never let it be said that this site fails to bring you the important details.

Back to Home PageBack to The Museum EntranceTop of this page