An 'ideal' radial engine for 1939-45

[tomo pauk]

... for the 1st line combat aircraft, and that can be designed & produced on technology and fuel of the day. No turbo-compounding; turbo-supercharger can feature in if the country is in the 'turbo game' historically. Useful both for fighters and bombers, both for land- and carrier-based A/C. Obviously we need as much power as possible, with manageable size and weight penalties, as good supercharger as possible (again that is contemporary with historical superchargers, even if not from the same country). Ability to be mass-produced, since everyone will want it. Development will need to start by 1935, latest by 1936 so the engine can be had in decent numbers by 1939. Air-cooled radial.
What company is to be trusted most per country? People can offer different engines made in different countries.

 

[cherry blossom]

Not quite answering your question but we could simply choose the best ideas from different countries.

Let us start by choosing to build a 14 or 18 cylinder radial with two poppet valves per cylinder.

There were different ideas on cooling the cylinder head. P&W forged the cooling fins of the R-2800C. The very last BMW 801s may have used vacuum casting to give a similar result. Bristol's last Hercules and Centaurus engines, only in service post WW2, used a copper-chromium alloy for the head with higher thermal conductivity. If we are making the ideal engine, we might apply the forging to cut fins in a copper-chromium alloy.

P&W may be greatest expert on baffles to direct the cooling air as judged by the 28 cylinder Wasp Major. However, BMW's integration of the annular oil cooler might be worth copying even if it required a fan.

There is no question that we should use fuel injection, which was first developed by Germany. This gives us good fuel distribution and allows us to have some valve overlap to improve scavenging.

For a fighter especially, we also need the German Kommandogerät.

When we come to bearings and vibrations, it seems that the USA was ahead. The R-2800 was probably the most reliable of the large radials. Nakajima in Japan did manage to produce the fastest turning large radial with the Homare running at 3,000 rpm compared to 2,700 for the R-2800, so we might study their design for any ideas.

Again with superchargers, we might go with the R-2800 C. It had a two stage supercharger with fluid coupling to the last stage, which should waste less power at low rpm. The second place on designing superchargers in widespread service might go to Mitsubishi of Japan. Certainly the Kasei powered J2M embarrassed Nakajima with its performance at high altitude. The very late turbosupercharger of the BMW 801 would be a contender but did not see much service.

Basically, let's make a R-2800 C with Bristol's alloy, fuel injection and hopefully better scavenging and a Kommandogerät. We can hope that we can raise the compression again and get better fuel economy as that was the main weakness of the R-2800.

 

[fastmongrel]

For Britain I think you have 2 choices. First choice employ a man to whack Roy Fedden on the knuckles with a ruler every time he mentions sleeve valves. Eventually the bruising and scar tissue will persuade him to design a 14 cylinder twin row engine with a chunky three bearing crank and enclosed valve gear.

Second choice is throw a hand grenade into the design office at Armstrong Siddeley. A new batch of engineers come in and design a 14 cylinder twin row engine with a chunky three bearing crank and a modern cylinder head design.

Do either or both in 1932 and you could have two 1350hp Radials available for aircraft production in 1937. Sunderland's, Whitley's, Wellington's and Halifaxes could all use the engine/engines. Merlin's can cover everything else.

 

[tomo pauk]

Me - I'd try to gain the most from a technology the companies are most fluent, in order to combine power, availability and reliability in a single package. Eg. the Germans might indeed look at both direct F.I. and the K.gerraet, while Japanese might start adding F.I in the later war years? For the UK and USA, rather pick the a good pressure carb, while staying away from the float type carb.
The R-2800 C series were mostly outfitted with 1-stage supercharger, only the F4U-4 seems to use a 2-stage supercharged C series. R-2800 was making a lot of power, fuel consumption will be high.

As for the 'power section', I'd go with a 14 cyl type, again for availability early enough after the short development time specified (3-4 years before service use), and to keep the weight down. Engine size and dry weight should ideally be kept in check, so that removes R-2600 (size) and BMW 801 (weight). Granted, BMW was capable for making the best power of the lot, but still - the TANSTAAFL rule applies as ever.
The smallest but still powerful 14 cyl engine early on was the Nakajima Ha 41, making 1260 CV at 3700 m (BMW 801C doing 1380 CV at 4800 m, but at 2/3rd more weight; granted, there was a 2-speed S/C on the 801C). It was modified into the Ha 109 with a 2-speed S/C, bigger impeller (diameter of 12in now), turned better RPM, and gained some weight, now at 720 kg; 1440 HP at 2100m or 1220 HP at 5200m. Not too bad on 91 oct engine and 37.5L. Impeller of
Next Japanese engine in size was the Mitsubishi Kasei, that was about as powerful as the ~30-50% heavier BMW 801 of the day. One of earliest adopter of water-alcohol injection among the aero engines, circumventing the lack of hi-oct fuel by a lot. A bit too wide at 1340 mm? Nice, big S/C, started with 11in impeller, moved to 12.6in quickly.

tl;dr - I'd probably want the Kasei-lookalike with a bit shorter stroke for a smaller diameter and better RPM, and then add country-specific bits & pieces. Dry weight of probably 750-800 kg. For the USA, a 2-stage S/C or turbo. Germans - fuel injection, K.gerraet, better exhausts early on. Hi oct fuel when available, ditto for ADI. Historically, both Japanese and Soviets were getting on the F.I. for the radials by late war.

 

[tomo pauk]

For Britain I think you have 2 choices. First choice employ a man to whack Roy Fedden on the knuckles with a ruler every time he mentions sleeve valves. Eventually the bruising and scar tissue will persuade him to design a 14 cylinder twin row engine with a chunky three bearing crank and enclosed valve gear.

Second choice is throw a hand grenade into the design office at Armstrong Siddeley. A new batch of engineers come in and design a 14 cylinder twin row engine with a chunky three bearing crank and a modern cylinder head design.

Do either or both in 1932 and you could have two 1350hp Radials available for aircraft production in 1937. Sunderland's, Whitley's, Wellington's and Halifaxes could all use the engine/engines. Merlin's can cover everything else.

I think we'd quickly agree that a 14-cyl engine based around the Mercury's bore and stroke would've been useful. We'd probably see 1250+ HP at ~15000 ft early on, or 1350+ HP down low, all on 87 oct fuel.
Better S/C is needed past 1940, though.

 

[Shortround6]

For perspective, not claiming the R-2800 was the gold standard or could not be improved.

Specification released April 21st 1937.

18 cylinder two row radial, 5 3/4 in bore, 6 in stroke , 2804 cu in estimated weight 2100lbs.

ratings on 100 octane fuel (the 100 octane that existed in 1937)
take-off power 1650hp at 2500rpm
Normal power (max continuous) 1300hp at 2300rpm to 10,000ft
Max cruise....................952hp
Normal Cruise.............850hp
Fuel consumption
max cruise................0.45lbs/bhp/hr
normal.......................0.42lbs/bhp/hr

Overall length 69 in
Diameter 52 1/2 in.

getting the R-2800 into service in 1940 required tens of thousands of man hours and testing, there were several test engines/prototypes starting with the X-77.
The X-80 was a nine cylinder test rig that racked up 695.92 hours by March of 1939 and countless rebuilds.

from the start of the project in March of 1937 until the A series engine passed it's type test (July 1st 1939) 325,000 engineering man hours and 3300 hours of ground running testing.

This was roughly (by numbers, not accounts) 40 men doing 10 hour work days every single day for the time period in question. This was peacetime.
And this was for the 1850hp take-off A series engine with single speed supercharger.

 

[tomo pauk]

Moving to 18 cylinder radials already by 1935 was not something that was within the scope for many engine companies, even if they were already making radials. Even the 14 cyl radials left a lot to be desired, be it reliability, power - whether at altitude or overall, or size. The 18 cyl idea adds another level of complexity to the development, with early 18 cyl engines (by Alfa-Romeo and Fiat) being more trouble than worth.
Then we have installation issues, increased weight and/or size will drive the size of airframe up. Total price and complexity will go up, ditto for manufacturing time. Not a good thing in the time nations were trying to quickly rearm.

 

[Shortround6]

Just pointing out the time and effort needed.

Now work backwards. The 14 cylinder P & W R-2180 used the same size cylinders (and P & W never went larger after the early 30s).

The Wright R-2600 was first used on the Boeing 314 Clipper which first flew in the summer of 1938 and used 6.125in bore by 6.3125 in stroke cylinders, same bore as the R-1820 radial but slightly shorter stroke (15mm) Same cylinders as the R-3350 (at least bore and stroke).

2600cu in or 42.7 liters is going to be near the upper limit on displacement.
Larger diameter cylinders give more cooling problems and start to give fuel burn problems.
Longer stroke cylinders tend to have rpm problems (high piston speed) the Pegasus got around part of that by using a small bore so the pistons were light.
A 14 cylinder Pegasus would be about 2726 cu in. Assuming you could get the same power per cu in as a Pegasus XVIII you get 1555 hp at 3,000ft and 1376hp at 15,500ft.

Just about anybody else's 14 cylinder radial is going to be smaller than the Wright R-2600 or Pegasus 14 cylinder. Which leaves RPM as the only real answer but few countries could match the US and Britain in Bearing technology (and lubrication)

The Mitsubishi Kasei engine was 2567 cu in.

It took P & W 4 years to build the R-2800 "C" series engine and the only thing/s that were kept were the bore, stoke and starter dog. Totally new engine with no interchangable parts with the older R-2800s.

edit, what ever you can build in 1939-40 will be obsolete in 1944.

 

[tomo pauk]

Good points. Indeed, by 1944 a good redesign of the power section will be needed, and already by 1942-43 for the supercharger section (as it was the case anyway on many engines).
As one can find out in my post above, I don't intend to go beyond 2500 cu in for a 14 cyl engine. Eg. the 'Mercury 14' will be about 2350 cu in.

 

[Shortround6]

'Mercury 14' will be about 2350 cu in.

It will be 2356 cu in, same as a Hercules or a Gnome-Rhone (or any licensed derivative, like Piaggio P XI or Russian M-88)

So you can play games with cooling (more fins) and RPM (center bearing, stronger crankshaft, etc) but basically you have 90.5% of the displacement of an R-2600 or 92% of the displacement of a BMW 801. You need either higher rpm or higher boost pressure to get the same power.

 

[Admiral Beez]

All the nations of WW2 would be lucky to have the Pratt & Whitney R-2800 Double Wasp. There is no better mass produced engine.

 

[tomo pauk]

So you can play games with cooling (more fins) and RPM (center bearing, stronger crankshaft, etc) but basically you have 90.5% of the displacement of an R-2600 or 92% of the displacement of a BMW 801. You need either higher rpm or higher boost pressure to get the same power.

We can work from there.
R-2600 was making 2400 rpm (for 'base' power of 1600 HP), 2600 rpm (for 1750 HP) and 2800 (for 1900 HP). No ADI, but hi oct fuel. No great shakes in boost, better versions used up to 42-43.5 in Hg (1750 HP versions) or 48 in Hg (1900 HP). At altitude it was around 44 in Hg; for reasons unknown to me, R-2600 was not allowed for max RPM beyond certain altitude (cooling problems?), something shared with a number of engines.
Per US data, Japanese engines we talk about here were using 40-42 in Hg boost already on 91/92 oct fuel - sometimes not confirmed by Japanese data. Good for 1500-1550 HP down low on 2650 rpm (Ha 109) or 2450 (early Kinsei types). On later Kinsei types, with ADI + that fuel, it was 47.7 in Hg + 2600 rpm = well into 1800 HP range.
What set apart the Japanese engines from R-2600 was their far better altitude performance, much due to their bigger impellers. Impeller on R-2600 was IIRC without the inducer vanes, and whole S/C intake was not as streamlined (granted, it was far easier to streamline the intake on the V engines than on radials). It took until 1944 for the R-2600 to be competitive beyond 15000 ft, while still being heavier and bulkier than Kinsei.

As for the Mercury spin-off, the Mercury itself was already making good RPM, up to 2750. Not sure how much the 4 valve head played part there, and how easy will be to engineer a 4 valve head in a 2-row radial. Probably the joint camshafts will be needed for pairs of cylinders.
Bristol's superchargers were big, 13 in impeller on Hercules, but were also suffering from lack of inducer vanes and 'messy' intakes. Same story was on the BMW 801A/C/D. It took both companies until late in 1944 to have better supercharging in service engines, although still no 2-stage S/C.

 

[Shortround6]

R-2600 was making 2400 rpm (for 'base' power of 1600 HP), 2600 rpm (for 1750 HP) and 2800 (for 1900 HP). No ADI, but hi oct fuel. No great shakes in boost, better versions used up to 42-43.5 in Hg (1750 HP versions) or 48 in Hg (1900 HP). At altitude it was around 44 in Hg; for reasons unknown to me, R-2600 was not allowed for max RPM beyond certain altitude (cooling problems?), something shared with a number of engines.

We have been over this before, the 1600hp engine and the 1700hp engine (not 1750) were different engines, different crankcases and crankshafts at the very least. The 1900hp engine was a 3rd engine, another new crankshaft and crankcase and for sure new cylinders and cylinder heads. Even the new cylinders and heads had problems in one installation (Martin PBM Mariner got cooling fans) so cooling problems in high, thin air does make sense.

I don't know the answer to the question about the limited boost at higher altitudes (or lower rpm?) but it may have to do with cooling on two levels? One is the general heat of the engine. The other is the heat of the intake charge from the less than ideal supercharger. Wright engines were generally set up with about 7,0 and 10.0 gear ratios on a two speed supercharger which means the supercharger was taking twice the power in high gear (power to drive the supercharger goes up with square of the speed of the impeller) The Supercharger may have been heating the intake charge up too much near full rpm and come too close or crossed to the detonation limits?
The R-2600 never got water injection.

 

[Koopernic]

Me - I'd try to gain the most from a technology the companies are most fluent, in order to combine power, availability and reliability in a single package. Eg. the Germans might indeed look at both direct F.I. and the K.gerraet, while Japanese might start adding F.I in the later war years? For the UK and USA, rather pick the a good pressure carb, while staying away from the float type carb.
The R-2800 C series were mostly outfitted with 1-stage supercharger, only the F4U-4 seems to use a 2-stage supercharged C series. R-2800 was making a lot of power, fuel consumption will be high.

As for the 'power section', I'd go with a 14 cyl type, again for availability early enough after the short development time specified (3-4 years before service use), and to keep the weight down. Engine size and dry weight should ideally be kept in check, so that removes R-2600 (size) and BMW 801 (weight). Granted, BMW was capable for making the best power of the lot, but still - the TANSTAAFL rule applies as ever.
The smallest but still powerful 14 cyl engine early on was the Nakajima Ha 41, making 1260 CV at 3700 m (BMW 801C doing 1380 CV at 4800 m, but at 2/3rd more weight; granted, there was a 2-speed S/C on the 801C). It was modified into the Ha 109 with a 2-speed S/C, bigger impeller (diameter of 12in now), turned better RPM, and gained some weight, now at 720 kg; 1440 HP at 2100m or 1220 HP at 5200m. Not too bad on 91 oct engine and 37.5L. Impeller of
Next Japanese engine in size was the Mitsubishi Kasei, that was about as powerful as the ~30-50% heavier BMW 801 of the day. One of earliest adopter of water-alcohol injection among the aero engines, circumventing the lack of hi-oct fuel by a lot. A bit too wide at 1340 mm? Nice, big S/C, started with 11in impeller, moved to 12.6in quickly.

tl;dr - I'd probably want the Kasei-lookalike with a bit shorter stroke for a smaller diameter and better RPM, and then add country-specific bits & pieces. Dry weight of probably 750-800 kg. For the USA, a 2-stage S/C or turbo. Germans - fuel injection, K.gerraet, better exhausts early on. Hi oct fuel when available, ditto for ADI. Historically, both Japanese and Soviets were getting on the F.I. for the radials by late war.

Was the BMW 801 really that heavy? The Smithsonian quotes 880kg dry for their example but wiki quotes 1070kg Dry.
BMW 801, Radial 14 Engine | National Air and Space Museum
I suggest the oil cooler is being included.

 

[tomo pauk]

Was the BMW 801 really that heavy? The Smithsonian quotes 880kg dry for their example but wiki quotes 1070kg Dry.
BMW 801, Radial 14 Engine | National Air and Space Museum
I suggest the oil cooler is being included.

Look here: link

 

[Koopernic]

Look here: link

The PW R-2800-54 (single stage two speed 2804.5 cu inch) weighs 1070kg, the BMW 801 2550 cubic inch) 1010kg so the 8% lower displacement is reasonably matched by the 6% lower weight. This is dry weight without oil cooler.

 

[Koopernic]

There is nothing wrong with the BMW 801 apart from two thing:
1 it is not in service earlier.
2 a two stage 4 speed (independent drives) inter cooled version not becoming available by early mid 1943. Such an engine would have been the BMW 801R but this was thwarted by successfully allied bombing in 1944. Corsairs and Hellcats both had two stage Super Chargers. The BMW 801R would have transformed the performance of the Ju 88, Do 217 and Ju 290

 

[tomo pauk]

There is nothing wrong with the BMW 801 apart from two thjngs
1 it not being in service earlier.
2 a two stage 4 speed (independent drives) inter cooled version not becoming available by early mid 1943. Such an engine would have been the BMW 801R but this was thwarted by successfully allied bombing in 1944. Corsairs and Hellcats both had two stage Super Chargers. The BMW 801R would have transformed the performance of the Ju 88, Do 217 and Ju 290

3 - Reliability was bad for the 1st 15+ months of service.
4 - It took almost twice the manhours to produce when compared with R-2600 or Hercules, per German data, as can be found in the Calum's book pg. 263 (1st edition). Most of the extra time required was due it being with fuel injection vs. a single carb on the Western engines, as well as the time to furnish the other ancillaries and drives for those; it was also more time-consuming to assemble the parts in a complete engine.
5 - It took eternity (when looked from the piot's point of view) to introduce a better S/C & tubing/piping from the 801E for better high-alt performance.
6 - Fuel consumption was very high, especially with the A and C versions. That mattered in German situation.