STEAM ENGINE VALVE GEAR.

[Stanley]

STEAM ENGINE VALVE GEAR.

The earliest steam engine valves, on both Savery’s pump and the Atmospheric beam engine, were simple plug cocks and the ‘valve gear’ was a lad who opened and closed them in the right order. This was soon superseded by an arrangement of rods and catches taking their motion from the beam of the engine which operated the valves automatically. The same gear can be seen on any of the preserved Cornish Beam engines used as pumps, the rotative engines tended to use standard slide valves, either flat or piston, and a simple eccentric driven linkage from the flywheel.

In 1799 William Murdock patented the slide valve in both flat ‘D’ pattern and the cylindrical form. The sliding valve was not new, very ancient forms of the principle were used for regulating the flow of water out of a trough by a sliding board in the bottom obstructing a hole. I made such a valve in 2000 for a dolly scouring machine at Helmshore and I am sure I didn’t invent it! The valves of the original Trevithick engine were flat plates oscillating on a pivot and this, with a cavity, was used by Watt who called them ‘sliding valves’. Examples of drop, slide, piston and ‘D’ valves have been found in Watt drawings dated before 1783. This was a time of experimentation and various methods of controlling steam ingress and exhaust were being tried. In 1786 Murdock’s steam carriage used a piston valve which Watt said was similar to one used on a 12” engine at Soho Foundry. Towards the end of 1798 or 1799 Murdock proposed the slide valve but Watt and Southern rejected the idea on the grounds that the faces would ‘become fluted’ and be difficult to repair. They favoured drop valves. Be that as it may, Murdock patented the ‘D’ valve and within 2 years the slide valve was being used at the Soho Foundry. The first slide valves had to be faced by hand but a few years later the advent of a planing machine at Soho made the manufacture much easier.

For almost a hundred years the slide valve reigned supreme in British engine building, at first in the flat ‘D’ valve form but later as a semi-rotating cylindrical valve or a piston valve. Whatever configuration was used these were all at that time fixed event valves and any governing of the engine had to be done by throttling the steam supply.

It has always amused me to hear people talk about ‘the simple slide valve’. To the uninformed eye it looks dead simple, a metal plate that slides over a hole and opens or closes it. Believe me, it is far from simple. One of the great railway chief engineers (I think it was Churchward) said in his retirement speech that it was a pity he had to go as he thought he was just beginning to get to grips with the basic principles of the slide valve. I am not going to delve into this pit here but I know exactly what he meant. Apart from problems with angularity of rods and settings, the valve’s biggest problem is that both steam admission and exhaust use the same port and therefore the valve is cooled by the exhaust.

In the 19th century engineers were aware of the limitations of the simple valve and understood that if they could leave the throttle fully open but alter the length of stroke during which steam was admitted at full pressure they could gain the benefits of expansive working. This would be far more efficient and save a lot of fuel. The first serious attempt to address this problem to my knowledge was the Meyer system of varying the configuration of the ‘D’ valve whilst it was running so as to alter the length of admission. This system was used by some manufacturers and had considerable success.

Then along came George Henry Corliss. The Corliss Steam Engine Company was originally known as Fairbanks, Clark & Co. in the 1830s. In 1843 it was renamed Fairbanks, Bancroft & Co. when Edward Bancroft joined the company. In 1846 it was renamed Bancroft, Nightingale & Co. when George H. Corliss joined the company, and in 1847 it was renamed Corliss, Nightingale and Co. In 1848 the company moved to the Charles Street Railroad Crossing in Providence, Rhode Island. In 1857 the company was renamed for the last time to Corliss Steam Engine Company. By 1864 Corliss bought out his partners and was the sole owner of the company. In 1900 the Corliss Steam Engine Company was purchased by the International Power Company. In 1905 it was purchased by the American and British Manufacturing Company. In 1925 the company merged into Franklin Machine Company. By then Franklin Machine Company already owned the William A. Harris Steam Engine Company. During his time with the company Corliss developed a completely new system of valve operation and took the first of his patents for it out in 1849.

Corliss had one idea in mind, he wanted to improve the efficiency of the steam engine. The first thing he did was create separate admission and exhaust valves using cylindrical valves operated from a wrist plate driven by eccentric rods whose motion originated from the rotation of the flywheel. This meant he could separate admission and exhaust events and keep his steam port hot. He realised that if he could make a linkage to the steam admission valves capable of adjustment while the engine was running he could vary the cut off point of steam admission and achieve expansive working. He did this by incorporating a catch in the steam valve linkage which could be controlled by the governor to trip at any given point in the stroke allowing the spring-loaded cylindrical valves to snap shut. Once all these ideas had come together the Corliss gear was born. Notice that the valves are nothing to do with the operation of the gear, they are not Corliss valves, it is the control gear and separate ports that made the breakthrough.

The Corliss gear became the standard valve gear on just about every stationary steam engine in the world with enormous benefits in efficiency. The linkages varied, makers tended to fit their own version, but all adhered to the same principle of fixed exhaust events, variable admission and governing on the cut-off with the steam valve to the boiler wide open. The amazing thing about this simple and efficient principle was the way many ‘experts’ and engineers managed to reduce the possible efficiency. A very common fault was giving the cylindrical valves too much cover on the port opening. If you go looking you will find extensive tables for calculating how much cover a given size of valve needed for a certain duty. Without exception, all the ones I have seen are wrong. I can only assume they were written by theorists who had never run a Corliss engine in their lives and were adding on a spurious safety factor to make the engines ‘safer’. It’s quite common in these tables to find covers of as much as half an inch recommended.

Here’s the SCG way of setting a Corliss valve. First of all, ignore any markings on the end of the valve when you take the back cover off. These are supposed to give you the opening point and the recommended setting for cover. With the engine stopped take up the adjustment in the linkage until you hear steam whispering through the valve, wind the adjustment back to give as little cover as you think you can get away with. On a single ported valve this will be about a sixteenth of an inch, on a double ported valve such as makers like Burnley Ironworks fitted, Newton tells me he has gone to below one thirty second of an inch. By doing this you ensure that the valve responds immediately when the linkage pulls it off its seat. It also takes a lot of strain off the linkage and bonnets because a large valve with full steam pressure on it is virtually locked on its seat and requires a great deal of force to shift it. The less effort required the less strain on linkage, bearings and castings. Newton told me that it was quite common for bonnets to fracture because of this repetitive strain. They always replaced these with heavier castings.

The beauty of the Corliss gear is the sharp cut-off which enables exact governing. This desirable attribute is commonly destroyed by engineers setting the dampers on the dash-pots to over-cushion the force of the spring returning the valve to the closed position to make the engine run ‘sweeter’. The damper is there to cushion the last half inch of return, not slow down the whole of it. In effect, an over-cushioned dashpot acts to alter the characteristics of the gear and make it little better than a slide valve. You can tell when an engine has minimum cover and sharp cut-off, it will run steadily on full boiler pressure under any condition of load. You find this out when you weave a shed out and end up running only twenty looms perhaps. Many engines were very dangerous running in this condition and I always like to hear the pistons slamming into the bottom of the dashpots. On an old engine, running can be improved by either fitting new springs or installing a spacer to increase the tension. Be careful if you do this, you need to be sure that the valve gear has enough travel to cope with the increased length in the pot. If you get it wrong it will pull the bonnet off the valve the first time you turn it over.

Continental engines were the first to use drop valves and some UK manufacturers followed them, particularly on that abomination, the Uniflow. They were an efficient valve but I am told needed more attention that a standard Corliss gear with cylindrical valves.

Notice that I have been talking exclusively about stationary engines. Locomotives were a different kettle of fish altogether because of the duty demanded from the locomotives under extremely varied conditions. Like stationary engines, the most efficient running was obtained by running with the regulator wide open and shortening the cut-off. Linking up in railway parlance. Conversely, when starting under a heavy load full cut-off was needed controlled by the regulator to minimise wheel slip. Add to this the need for a reversing gear in the linkage and I think you can begin to see that loco valve linkages offered tremendous scope for the designers. Two gears in particular probably accounted for most of the UK locos, Stephenson’s and Walchaert’s but there were dozens of others each with their own champions. The last improvement was the Caprotti gear which could possibly have eclipsed the others eventually but came in to late to get a foothold.

Funnily enough, one of the best places to look to find out about the various types of loco valve gear is in the back numbers of the Model Engineer. For many years a man writing under the pseudonym “LBSC” (Lillian “Curly” Lawrence) gave wonderfully clear descriptions of different gears and how to build them. If you want to dig further there are other good books; Youngson’s Slide Valves and Valve Gearing, 1948. Slide valves and Link Motions, Auchinloss, 1872. Valves and Valve Gears for Locomotives, Lake and Reidinger, 1940. I’ve never had any experience of loco gears so there is nothing more I can add. One thing is certain, the loco gears are far more complicated than stationary engine gears and it would take a lifetime to fully understand even the simplest of them.

SCG/22 January 2007

[Stanley]

Bumped. If you were looking for it, this is vital information!

[Stanley]

A fascinating subject if you are that way inclined!