KurtVerbose
Member Since: 08 Aug 2010
Location: Les Arses
Posts: 5848
|
Here's a very short but excellent animation on how sleeve valves work in a Bristol Hercules.
http://www.youtube.com/watch?v=_vrvep_YOio
I wrote this article a few years ago during an idle evening. Just a little industrial history on sleeve valves and why I find them so interesting. Maybe it's boring for you guys, in which case - sorry for being such a nerd.
Quote:
Internal combustion engines have poppet valves – it’s a given. And because that’s been the case for so long it must be because they’re superior to any other kind of valve?
Well actually, poppet valves have dominated engine design for so long only because of an accident of industrial history, and they are not necessarily the best valve design.
I read a biography of an interesting character called Sir Roy Fedden, who was an aero engine engineer at Bristol from the end of the First World War till the end of the Second. I read it because I have an interest in sleeve valve engines
Feddon had a natural affinity for air cooling and for radial engines, believing them to be lighter and more powerful than inline water cooled engines, and these advantages offset any aerodynamic disadvantages they may have. Yet by the end of the 1920's Feddon reached the same conclusion as Ricardo, that the power of these engines would be restricted by the problem of detonation - the fuel/air charge exploding due to the increases in pressure and temperature as the cylinder charge is compressed, rather than being ignited by the spark plug, and in so doing, destroying the engine very quickly. The four valve head was a help in reducing detonation, but was very difficult to fit onto a radial engine that used push rods to operate the valves, especially so on a radial with two or more rows.
But Feddon believed the four valve head only mitigated the problem and that the best solution was to do away with poppet valve entirely, to give a better combustion chamber shape, provide more reliability, and remove a principal source of detonation - the very hot exhaust valve.
Feddon turned to sleeve valves, which had been experimented with before but never very satisfactorily. In a sleeve valve engine the cylinder wall rather than the cylinder head has inlet and exhaust ports in it. Between the piston and cylinder sits a sleeve that is moved vertically and radially by a series of gears driven directly from the crank to expose the ports in the cylinder wall - allowing gasses to be inducted and exhausted.
Sleeve valves were a huge engineering challenge for the day, and hugely important to Britain’s competitiveness in the aviation industry, and later to defeating Germany. To give an idea of the resources dedicated to the project, ten time more was spent on the development of sleeve valves than the early jet engines that came nearly 10 years later.
The engineering challenge stemmed from the sleeve. Endless experiments were made to find a metal that was both a good cylinder liner, was light and had the same coefficient of expansion as the piston and the cylinder itself. Eventually, one was found and the design worked. Fedden was able to raise the compression ratio significantly compared to a two poppet valves, with a corresponding increase in power and efficiency.
However, making the sleeves to the correct tolerance was impossible unless each sleeve was hand polished to fit each respective cylinder, a process completely unacceptable for the volumes required. Eventually, at a time when the projects detractors were gaining in number and both the government and Bristol were loosing faith in Feddon, a solution was found by accident. A machinist had used a worn tool instead of a new one, and the tool had polished rather than cut. Bristol's factory now began turning out large numbers of sleeve valve radial engines - the two row 14 cylinder Hercules, to be used on Bristol's own Beaufighter, the Short Stirling, the Vickers Wellington, the Handley-Page Halifax, and even on some Avro Lancaster's.
And what superb engines they were. The Hercules and the later 18 cylinder Centaurus, compared to the rival American radials with their two valve heads, were smoother and quieter, in the cruise they used only two thirds of the fuel, and they lasted up to 50% longer between overhauls.
How they compared to the water cooled inline engines with their 4 valve heads is a different matter. By the early 40's poppet valves had broken through the detonation constriction Feddon and Ricardo had thought would hold them back. High octane fuel was available, and the exhaust valves were kept cooler by filling them with a metal of very high heat conductivity (sodium), allowing more heat to be dissipated through the valve stem.
The only empirical comparison of sleeve valves vs. a four valve engine was done towards the end of the war by Stanley Hooker at Rolls Royce. RR was experimenting with a 2-stroke V12 using both direct injection and sleeve valves, which was theoretically both efficient and hugely powerful. This ambitious project was running into trouble and Hooker was brought in to advise the board whether it should be scrapped, which is what his advice was.
Hooker was an aerodynamicist with a big reputation. His early work at RR had noticeably improved the performance of the supercharger on the Merlin, a state of the art design many had thought was already close to maximum practical efficiency. He then introduced two-stage supercharging on the Merlin, doubling power at 30,000ft and allowing the Spitfire to convincingly meet the challenge of the FW190.
He found no advantage in sleeve valves, but his analysis was extraordinarily narrow, consisting of a comparison of air flow through open valves. No proponent of sleeve valves had said air flow was a principal advantage, so why did Hooker only look at this area?
It could be it was because Hooker was an aerodynamicist, but this is belying his broader abilities. A more likely explanation was that he resented being temporarily pulled off jet engines to look at a project he thought was irrelevant. To him the engine could be twice as powerful as the best piston engine, but it still wouldn't beat a jet. He knew jet engines would be the future, and he proved this to the air ministry in a paper about a future fighter.
The requirement was for an aeroplane that could fly at 400mph at low level. Feddon proposed an 18 cylinder sleeve valve engine of even larger dimensions than the huge Centaurus, and producing about 4000bhp, and with a development time of 3-5 years. Hooker pointed out that with some subtle revisions to the compressor and turbine of the existing Nene jet engine, 500 mph at low level was readily available. The Nene was also very cheaply made compared to a huge piston engine, and a single Spitfire was more maintenance hungry than a whole squadron of Meteors.
Soon after, all sleeve valve development stopped, together with pretty much all piston engine development. Bristol replaced Feddon with the newly recruited Hooker (who’d fallen out with his boss at RR) when the former proved too attached to his beloved sleeves.
Piston engine development was now passed to the car manufactures, many of whom hadn't moved from side valves to overhead valves, yet alone 4 valves, and they were far too concerned with cost to think about sleeve valves. It wasn't until the eighties, when turbo charging and lots of money hit Formula One before piston engine development caught up with the state of the art in the late 40's. In fact, Paul Roche at BMW later read a book about World War II engines and found out they went through exactly the same process as he'd done concerning inter-cooling and exotic fuels, never mind turbo compounding. He confessed he could have saved himself a lot of trouble by reading the book before hand!
So by an accident of industrial history this great technology had been forgotten.
That is, apart from by one Keith Duckworth, one half of Cosworth. He remembered them from working at Bristol in his youth, and experimented with them again in the 70's, apparently quite easily getting close to the specific output of contemporary Formula One engines, and concluded that they could be more easily and better made than in the 40's by using modern methods. However, as is often the case with engineering, a superior but undeveloped technology often has a difficult or impossible time dislodging an inferior but developed one.
|