What If: Pratt Whitney build/develop the Sabre under licence?

[wuzak]

It seems to me that the Bristol Sleave valve engines offered no discernable advantage over poppet valve rradials. I've looked for performance advantages but none over radials of equal weight, power, displacement at all. That includes the centaurus The Sabre at least promised advantages in power.

The Sabres trick was reving at high RPM, I suspect that this was a result of the horisozontall opposed configuration as much as the sleeve valve. Junkers was already running the Jumo 213 at 3250 rpm and were planning to productionise the the 3750 rpm jumo 213J. Part of the trick of the Jumo was the twin overhead cams. Overhead cams might allow higher revening but cams increase the dimensions of the rocker cover, hence a tendancy to use pushrods on radials.

It would seem to me a similar configuration to the Sabre but with DOHC poppet valves might produce the same result, perhaps with somewhat more width and frontal area. A back to back arrangment of two 120 degree V's might overcome even this. Two 90 degree V-12 might also achieve this. The Sabre avoided shaft issues by useing two shafts and gearing them together. The RR Vulture didn't, but retained only one shaft. I rather suspect that had RR had the resources to continue with the Vulture it would have turned into an engine abl,.e to match the Sabre and probably entered service as a reliable unit years earlier. Assuming the same volumetric efficiency as the Merlin then the 42 litre Vulture should produce 1.62 x the power of the Merlin: ie 2750hp at 18psig and maybe 3300 at the higher boost levels.

Sleeve valves may have had little to do with it.

The Sabre's ability to rev was due to its (relatively) short stroke. It's stroke was 4.75"compared to the Merlin's 6".

The Sabre would definitely have been bigger with poppet valves and OHCs. Or even if it had pushrod overhead valves (like most radials did).

The advantage of the sleeve valve was claimed to allow more boost for a certain compression ratio, or conversely more compression ratio for a given boost. The Sabre used 7:1 CR compared to Rolls-Royce's usual 6:1, and yet was able to take similar amounts of boost.

I agree about the Vulture - many don't. It's a pity that Rolls-Royce did a version of the Sabre rather than a liquid cooled sleeve valve X engine. The Pennine was an air cooled X, and was more compact than the Eagle.

 

[Shortround6]

Wuzak is correct about the short stroke allowing the high rpm.

The big push for sleeve valves started back in the 20s and it had a lot of theoretical advantages. Many of these advantages did prove out only when compared to poppet valve engines of the 1920s. Both camps were moving targets. By the time the sleeve valve was perfected many of the problems with the poppet valve had been solved.
The sleeve valve solved the problem of the exhaust valve being a "hot spot" that caused detonation. But in the 1920s few exhaust valves used a liquid coolant inside to transfer heat up the stem. By 1940 some people were even starting to use sodium salt cooled valves on intakes as well as the exhausts. More had been learned about supporting the valve stem and providing long valve guides to transfer heat from the valve to the head and cooling medium, air or liquid. Valve spring breakage had been more than common in the 1920s, by 1940 valve spring breakage had been reduced to much lower levels due to improved spring steels, different methods of drawing the spring wire, heat treatment and shot peening.
Much is made of the reduced number of parts in the valve train, which is true, but little is said about what kind of parts. Find a cut away of a Hercules engine that shows the gear train for the sleeve valves. It makes a Swiss watch look like a sun dial. Push rods and rocker arms are much easier to make by the hundred than gears.
The claim that the sleeve valve cylinder head is much smaller also needs looking at. It is in theory but in para rice, at least on air-cooled engines the need for adequate finning requires a head almost as large as the poppet valve engines, at least no vast difference can be found between radials of both types that use a similar stroke.

 

[Siegfried]

There is another aspect to this. The teething troubles and lack of reliabillity of the early Sabres is often attributed to the sleeve valves along with quality coontrol issues at Napier and its suppliers.

However it took 6 years to get a piston engine to production, the R-2800 first ran in 1937 however according to Wiki Halford didn't start design of the Sabre till 1935 and it didn't run till Jan 1938 so its possible that the PW-2800 is running up to 1 year ahead of the Sabre. It still doesn't explain the full extent of the delays.

Wiki gives the Sabre as being 40 inches (1016mm wide) that's fairly wide. The R-2800 was 52 inches wide. The R-3350 was 55.78.

Assuming the stroke of 4.75 inches (ie 120mm) the addition of overhead cams and rocker covers might add twice that so a twin flat 12 engine using DOGC engine would be 49.5 inches wide. 240mm (9.5 inches) to the width. However the spark plugs of the Sabre are clearly deeply recessed and spark plugs on poppet valve engines don't protrude, there is clearly a great deal of material on the sleeve valve head. I would estimate that a DOHC poppet valve flat engine would only be 10% wider.

It's also worth cosidering that the R-3350 was no wider than the 3370 cubic inch centaurus and that the R-335 ran at 2800 rpm. I see no problem with a twin flat 12 horizontally opposed engine like the Sabre running at over 3000RPM given the small dimensions.

Hence, the sleave valves were bypassable and it is possible that a poppet valve engine could have matched all but the final versions of the Sabre.

Fairey H24 Monarch with much greater dispalcement was a 2000hp class engine that was only 43 inches wide.

 

[wuzak]

Fairey H24 Monarch with much greater dispalcement was a 2000hp class engine that was only 43 inches wide.

It may have been only 43 in wide, but it was 52.5 inches tall - being a vertical H instead of the Sabre's horizontal alignment. It was also 86.25 inches long.

Most Sabres were 40"wide x 46" high x 82.25" long. The Sabre VII was 47.75"high x 83"long, but still 40" wide.

The Dagger was a vertical H-24 with geared together crankshafts. It's bore and stroke were 3.813" x 3.75", giving a capacity of 1027ci/16.8l. Its dimnsions were 80" long x 22.5" wide and 45.125" high. So, despite having a stroke 1" less than the Sabre's the dimension in the direction of the stroke was over 5" longer.

It's also worth cosidering that the R-3350 was no wider than the 3370 cubic inch centaurus and that the R-335 ran at 2800 rpm. I see no problem with a twin flat 12 horizontally opposed engine like the Sabre running at over 3000RPM given the small dimensions.

The R-3350's stroke was 6.312" compared to 7" for the Centaurus. The bore of the R-3350 was greater, giving much the same capacity (6.15" vs 5.75"). Thus if we change the R-3350's stroke to be the same as the Centaurus its diameter would grow by at least 0.7" - probably more, as the con rods would have a stroke to length ratio of greater than 1 - so the diameter would grow by 2.1".

No doubt a poppet valve H-24 of the same capacity as the Sabre running the same bore/stroke and revving to 3750rpm would match the Sabre's output. But it would be physically larger.

btw, if reliable contra-rotating prop systems were available early in the war a Sabre version may have been designed with contra-rotating cranks, rather than same direction of rotation for the cranks, with each half driving one half of the contra-prop, and dispensing with the heavy and complicated reduction gear.

However it took 6 years to get a piston engine to production, the R-2800 first ran in 1937 however according to Wiki Halford didn't start design of the Sabre till 1935 and it didn't run till Jan 1938 so its possible that the PW-2800 is running up to 1 year ahead of the Sabre. It still doesn't explain the full extent of the delays.

The P&W R-2800 was first run in 1937, but when did design commence? It also didn't reach full production until 1940/41. The Sabre was also in production from 1941.

The Centaurus also first ran in 1937, but didn't get into production until late in the war ('43 or '44).

Given that it took 6 years to get an engine into production, and Fairey had no engine production experience/facilities and the P24 was not ready for production in 1941, how long before the P24 becomes available?

 

[Shortround6]

It's also worth cosidering that the R-3350 was no wider than the 3370 cubic inch centaurus and that the R-335 ran at 2800 rpm. I see no problem with a twin flat 12 horizontally opposed engine like the Sabre running at over 3000RPM given the small dimensions.

The limitation on engine speed had more to do with the crankshaft and crankcase than on the valve train by 1940. the general rule of thumb was a 3000ft per minute piston speed. All those pistons starting, accelerating, stopping and then accelerating in the opposite direction create a lot of strain. and the forces go up with the square of the speed so a 10% increase in RPM or piston speed means a 21% increase in the forces acting on the crankshaft and crankcase. A closer approximation can be found using the concept of corrected piston speed was devised by Frederick Lanchester, and defined as the average piston speed divided by the square root of the stroke/bore ratio.

Fairey H24 Monarch with much greater dispalcement was a 2000hp class engine that was only 43 inches wide.

Yes but it's cylinder banks were arranged vertically and not horizontally and its height was 52.5 inches. Try turning it sideways

 

[Shortround6]

I believe P&W started work on the R-2800 in 1936 as an R-2600 and bumped it's size when they learned of Wrights R-2600.

I believe it is a mistake to compare the size of air-cooled engines to liquid cooled engines as the air-cooled engines have to be larger in order to accommodate the massive amounts of fining a high-powered engine needs. This can include more space between the cylinders making an inline (or V) engine longer in addition to the several in deep fins on the cylinder head. Bristol cylinders are dis-proportionally tall as the top several inches provide horizontal cooling fins for the junk head that is sunk down into the open top. The Sleeve valve has a more compact head area but it may be a more pronounced advantage on a liquid cooled engine.