Did Sleeve Valves Have an Advantage at Very High RPM?

[spiralcopter]

It occurs to me that at extremely high RPMs, sleeve valves would have the advantage that they do not suffer from valve float, as the valves are positively driven rather than spring-loaded. Ducati motorcycle engines can be revved to over 10,000 RPM in part because they have positively driven desmodromic valves.

On the other hand, I have it on the authority of someone who has worked on sleeve valve aero engines that the valve timing geartrain will essentially explode if you try and run them too fast. So clearly there are limits.

Was this actually a relevant advantage? Did anyone ever get a sleeve valve to rev high enough that the lack of valve float mattered? It occurs to me that the Sabre, with its displacement divided into over double the usual number of cylinders, should have been able to support high RPMs if only they could keep the valve timing geartrain (and everything else) from having problems.

 

[MiTasol]

I cannot see the drive train exploding but maybe, and I would place the chances as remote, the peg that the drive ball fits on that drives the sleeve may shear from the high accumulated forces. Remember as the sleeve moves up/down the force to cause that motion will go up as a square of the increase in RPM. So double the revs and the load increase fourfold. If the peg is only designed for load x it would be easy to exceed that if the revs go significantly higher than the original design allowed for.

 

[cammerjeff]

I don't know of any piston engine A/C engines of WWII era that rotate at a speed that would induce valve float, remember the limiting rpm factor of propeller tip speed. So even with gear reduction most would not spin more than 5-6,000 rpm. Most were around 3200-3500 rpm max.

 

[pbehn]

It occurs to me that at extremely high RPMs, sleeve valves would have the advantage that they do not suffer from valve float, as the valves are positively driven rather than spring-loaded. Ducati motorcycle engines can be revved to over 10,000 RPM in part because they have positively driven desmodromic valves.

On the other hand, I have it on the authority of someone who has worked on sleeve valve aero engines that the valve timing geartrain will essentially explode if you try and run them too fast. So clearly there are limits.

Was this actually a relevant advantage? Did anyone ever get a sleeve valve to rev high enough that the lack of valve float mattered? It occurs to me that the Sabre, with its displacement divided into over double the usual number of cylinders, should have been able to support high RPMs if only they could keep the valve timing geartrain (and everything else) from having problems.

Its a long time since I read up on Ducatis. When I first started riding bikes Hailwood won on one at the Isle of Man but that was a 864cc V twin. Then what I read was the rev limit was set by piston speed, the desmo valve system used less energy and allowed better valve timing, the valves could open and close more sharply. The modern Moto GP bikes are 1000cc V4s. What is high revving depends on cylinders and displacement, Mike Hailwoods 250cc six cylinder revved to just short of 20,000RPM.

With sleeve valves there is a huge contact area between the sleeve and the cylinder wall and the guide above the cylinder head, calculating those forces isnt my bag but they are usually exponential. Maybe "explodes" is incorrect but it is easy to imagine something breaking. Typhoons were difficult to start in winter and had to be kept war and "run up" periodically overnight in winter, if you let them get cold and the oil thicken they were difficult to turn over and start.

Revving higher isnt always the game changer, it is getting as much "mixture" into the cylinders as possible and getting it out again.

 

[kmccutcheon]

It occurs to me that at extremely high RPMs, sleeve valves would have the advantage that they do not suffer from valve float, as the valves are positively driven rather than spring-loaded. Ducati motorcycle engines can be revved to over 10,000 RPM in part because they have positively driven desmodromic valves.

On the other hand, I have it on the authority of someone who has worked on sleeve valve aero engines that the valve timing geartrain will essentially explode if you try and run them too fast. So clearly there are limits.

Was this actually a relevant advantage? Did anyone ever get a sleeve valve to rev high enough that the lack of valve float mattered? It occurs to me that the Sabre, with its displacement divided into over double the usual number of cylinders, should have been able to support high RPMs if only they could keep the valve timing geartrain (and everything else) from having problems.

One must bear in mind that the sleeves are quite heavy. In the aircraft engines I have studied their weight varies from 10 to 15 lb. That is a lot of reciprocating mass whose inertial forces must be borne by the drive train.
American Sleeve-Valve Aircraft Engines

 

[Simon Thomas]

Any mechanical device run outside the design envelope is operating in the safety margin.
My favourite definition of an engineer is: 'Any idiot can design a bridge that won't fall down, but it takes an engineer to design a bridge that will only just not fall down."
If the sleeve valve engines are "unreliable" when operating over their design speed, it is not the fault of the engine.
My theory is that they would be more reliable if the correct oil with the correct additives is used.

The notorious Centaurus

I´m aware that it´s common practise, at least in the USA, to substitute a Wright Cyclone R-3350 for the original Bristol Centaurus in Sea Fury rebuilds. This is said to be due to the unreliability of the BC and, though realizing that it´s a complex design, I´ve always wondered what causes the...

www.key.aero

 

[pbehn]

Any mechanical device run outside the design envelope is operating in the safety margin.
My favourite definition of an engineer is: 'Any idiot can design a bridge that won't fall down, but it takes an engineer to design a bridge that will only just not fall down."
If the sleeve valve engines are "unreliable" when operating over their design speed, it is not the fault of the engine.
My theory is that they would be more reliable if the correct oil with the correct additives is used.

The notorious Centaurus

I´m aware that it´s common practise, at least in the USA, to substitute a Wright Cyclone R-3350 for the original Bristol Centaurus in Sea Fury rebuilds. This is said to be due to the unreliability of the BC and, though realizing that it´s a complex design, I´ve always wondered what causes the...

www.key.aero

I think there was another issue with Bristol and Napier, they couldnt mass produce the sleeves to the tolerances and quality required.

 

[Builder 2010]

Good question. Sleeves added another large moving piece that needed lubrication both inside for the piston and outside for itself. It was an extra metallic layer between the hot zone and the cooling jacket. I agree that the first weak link would probably be the ball pin followed by lubrication failure and overheating. If the piston clearance got too tight the whole deal would come to a grinding halt. At that point if the pin didn't break, probably one of the myriad of gears in the train would lose a tooth. I read that the reason the Centaurus is not used in Sea Fury replacements is there are no spares available for them, whereas the 3350s are plentiful.

 

[Simon Thomas]

To state they couldn't mass produce them isn't a fair reflection of the whole story.

Bristol has no issues with tolerance until they had to move to mass up production. The issue took 4 years to solve (by accident as it turned out).
After late 1939, it was other issues that held them up. By 1941, production was in full swing.

Napier struggled until Banks forced Bristol to give them a Taurus sleeve and the technology to make them - late 1942/early 1943.

Within a few years of the end of the war, the civilian Hercules was being used successfully and was still being sold until 1968. They only stopped when RR took over and refused to make any more.

I think there was another issue with Bristol and Napier, they couldnt mass produce the sleeves to the tolerances and quality required.

 

[SaparotRob]

To state they couldn't mass produce them isn't a fair reflection of the whole story.

Bristol has no issues with tolerance until they had to move to mass up production. The issue took 4 years to solve (by accident as it turned out).
After late 1939, it was other issues that held them up. By 1941, production was in full swing.

Napier struggled until Banks forced Bristol to give them a Taurus sleeve and the technology to make them - late 1942/early 1943.

Within a few years of the end of the war, the civilian Hercules was being used successfully and was still being sold until 1968. They only stopped when RR took over and refused to make any more.

I'm curious as to what that accident was.

 

[Simon Thomas]

The machine operator accidently used a worn grinding wheel rather than new on the final pass, and the sleeve was within tolerance.
Fedden was also trying many different materials as well in this time frame.

My reference book has vanished, so I am going by memory. If anyone has more detail - feel free to add it.

 

[tomo pauk]

We can recall that the fastest-revving aero engine was outfitted with popet valves.
Napier Dagger

 

[Deleted member 68059]

In terms of airflow potential, yes - BUT; Supercharging/Turbocharging renders this a moot point, even if true in terms of the cross sectional flow area, because flow is velocity not density limited,
(the limit is when the flow reaches the speed of sound in the port) so if you boost, you can have as much MASS flow as you want to, providing your detonation limits are high enough.

Sleeves always had lower detonation limits than poppet valves after about... 1935 (ish) due to advances in valve alloys (avoiding high temperature scaling which
promotes pre-ignition and detonation).

Even ignoring that, its a non-issue as poppet valve engines have been quite happily run normally aspirated above 100% volumentric efficiency at over 20,000rpm,
something you cannot do with a sleeve due to the huge mass of the sleeves (several kilos vs about 100 grams), which renders the dynamic forces untenable,
in addition to the mass, you have gigantic frictional losses due to the immense "wetted" area of oil surface per sleeve.

 

[Trilisser]

Calum, then please explain why the post-war Centaurus had substatially higher power on 100/130 than the poppet-valve R-3350? I love your book but on this issue I side with Harry Ricardo.

 

[Deleted member 68059]

Calum, then please explain why the post-war Centaurus had substatially higher power on 100/130 than the poppet-valve R-3350? I love your book but on this issue I side with Harry Ricardo.

As you would already know from pages 443 and 445 of the book, none of the wartime sleeve valves engines managed to exceeed the boost level
which the simple old Merlin had from December 1941 onwards, even with water injection, not even by the end of the war, nearly four years later.

The R-3350 is a somewhat sneaky example to use to decry poppet valves, what with it being a 2-valve per cylinder, the most athsmatic configuration
possible, and because all the later versions were designed to use 115/145. Even allowing for that, I`m uncovinced there is any "substantial"
difference.

The R-3350 appears to be fairly happy at 2880hp on 100/130

Turbo Compounds

Having looked through all my centaurus data-sheets, THE most powerful version I can find on 100/130 was the
"373" which had direct fuel injection (2980hp) - hardly "substantial", and indeed all non-direct injected Centaurus
engines appear to top-out at 2730hp (some have more but only with Methanol/Water).

Harry Ricardos interest in Sleeves was from an era before poppet valves had reached the high degree of metallurgical refinement,
and I`m afraid that the reason nobody makes them today, is not some sort of missed opportunity, or conspiracy, but because
they are simply not advantagous anymore. (I think there was talk of using them in an opposed piston diesel, where they`re use
is essentially mandatory because you cant physically install poppet valves into that configuration).

As I`m sure you`ve noticed, and as have other reviewers, the book "does not cheerlead for any particular engine", and I`m not here
to make fun of sleeves, I`m simply saying there is no good evidence they have had any real advantage since about the mid 30`s,
yes, you can put in "bigger holes" to let air in with a sleeve, but if you cannot burn it without it detonating, or run it faster without
the sleeve drives falling to bits, there is no way of taking advantage of that fact.

 

[wuzak]

Harry Ricardos interest in Sleeves was from an era before poppet valves had reached the high degree of metallurgical refinement

And before the sodium cooled exhaust valve was invented/widely employed?

 

[Trilisser]

One should not compare a turbocompounded variant to a non-compound one. The Centaurus 661 maximum is 2700 hp at 2800 rpm on 100/130 while the R-3350 as fitted in the AD-4 is 2700 hp at 2900 rpm on 115/145. To me the same power from a slight smaller size on fuel with 15 points lower PN is a massive advantage.

And Ricardo advocates sleeve-valves in his book written in the 1950s, after those metallurgical advantages.

That sleeve-valves are not used today means nothing. After all, how many decades it took for direct injection to be universally adopted in cars? VHS won over Beta even though the latter was much superior.

 

[Deleted member 68059]

One should not compare a turbocompounded variant to a non-compound one. The Centaurus 661 maximum is 2700 hp at 2800 rpm on 100/130 while the R-3350 as fitted in the AD-4 is 2700 hp at 2900 rpm on 115/145. To me the same power from a slight smaller size on fuel with 15 points lower PN is a massive advantage.

And Ricardo advocates sleeve-valves in his book written in the 1950s, after those metallurgical advantages.

That sleeve-valves are not used today means nothing. After all, how many decades it took for direct injection to be universally adopted in cars? VHS won over Beta even though the latter was much superior.

using a completely different engine to say sleeves are better whilst saying "don't compare it with THAT" one.

I think you've discovered the hole in your methodology !

Richardo are still going, why do YOU think they have not made one in three quarters of a century?

You've also ignored the data driven point about boost levels, still waiting for you to post a single piece of actual data to support your assertion?

Why didn't the Centaurus or sabre exceed 15psi boost in the whole war? Why didn't the crecy work?

We already know, and it's all in the book, with references.

Sleeves give you a considerably lower knock limit. This is not a discussion point I'm afraid.

If you think with the pressure today to develop better engines that the fact nobody is making them quote "means nothing" then I'm not sure I can help you any further since your proclivity for the technology is obviously ideological in root,

 

[Shortround6]

Calum, then please explain why the post-war Centaurus had substatially higher power on 100/130 than the poppet-valve R-3350? I love your book but on this issue I side with Harry Ricardo.

What did they have to do to get that power?

A quick peek at some old books has the 165 series and above Centaurus engines weighing around 3400lbs. The older Centaurus engines were under 3000lbs.
See Calum's picture.
R-3350s went about 2900lbs or under without the Turbo compounding, they went about 3450lbs with it.

Airplane engines are different than car engines.
Horsepower per pound of engine weight was more important than horsepower per unit of displacement.

Adding 500lbs of engine weight to get 100rpm more max revs and another 3-4lbs of boost is not crying out the superiority of sleeve valves.

 

[MiTasol]

And then you must consider overhaul costs and on that metric the sleeve valve engine fails because the valve train has so many more parts that must be assembled in exactly the right order and at exactly the right angle plus the sleeves and barrels require far more work than poppet valves do. Yes the poppet valve has more parts after the cam drum but the vast majority do not require the same level of precision fit. Likewise the honing of sleeves is a precision item and takes hours. Lapping poppet valves take much less time.

On the poppet valve engine only the cam drum requires critical positioning and everything else follows automatically and clearances are adjusted at the rockers. It is not possible to fit an exhaust rocker to the inlet valve position on radials but it is very easy to fit a sleeve drive crank one tooth out and that can spell disaster in some cases.