Curiosity question on WW2 piston engine designs

[imipak]

I know the Merlin was approaching the fundamental limits of its design, by the time of the Merlin 130 and the RM17SM (which reached 2600hp in branch tests).

Back of the envelope calculations suggest that if you were to build such an engine today using modern materials and modern fuels, you'd probably only squeeze 2800-3000hp if you wanted reliability, and Voodoo (which used an earlier design and upgraded a few things but not everything) topped out at 3100hp under racing conditions.

This puts the Merlin in 1945 as being respectably close to what the design was physically capable of, and respectably powerful for a piston engine even back them.

So, my question: Just how close did WW2 piston engine designs get to their theoretical upper limit?

 

[Sisu]

Much depends on the time period you are looking at. After 1964 some very good minds and well-financed development programs produced Merlins that would live for a while at 3600 and more horsepower. This took Allison connecting rods, custom pistons and such, and much more. We are talking 160 octane fuel, 3600 rpm at 140 inches, and an airport runway available right under you, so far from operational WW2 conditions.So what limits are we talking about, and when?

 

[imipak]

Much depends on the time period you are looking at. After 1964 some very good minds and well-financed development programs produced Merlins that would live for a while at 3600 and more horsepower. This took Allison connecting rods, custom pistons and such, and much more. We are talking 160 octane fuel, 3600 rpm at 140 inches, and an airport runway available right under you, so far from operational WW2 conditions.So what limits are we talking about, and when?

Thanks, that's an excellent question, apologies as this is a little complex

The TLDR version: I'm looking for the absolute upper limit of what they could have built in 1944/45 with the engine design they had but not necessarily the same materials, merely materials that were available.

To me, the absolute theoretical limits of a design are the best that engine design can do, regardless of the materials used. You're going to get declining returns at some point, as you obviously can't get more power out than the fuel has - you'd need to change the design, such as increasing the size of the engine, at some point.

There's then a second maximum, the theoretical maximum you could achieve with the materials available in the first half of the 1940s, assuming metallurgy remained fairly constant over such a short time - an assumption that might not be valid, but is necessary to make it the question can't be answered.

I think I'd be happy with either theoretical maximum (the first tells me how good the design itself is, the second tells me how good the combination of design and materials is), but I'm really aiming more for the 1944/1945 conditions.

 

[Tony Kambic]

Try purchasing this book and reading it. Should have some answers for you: The Secret Horsepower Race: Western Front Fighter Engine Development by Calum Douglas.

 

[pbehn]

I know the Merlin was approaching the fundamental limits of its design, by the time of the Merlin 130 and the RM17SM (which reached 2600hp in branch tests).

Back of the envelope calculations suggest that if you were to build such an engine today using modern materials and modern fuels, you'd probably only squeeze 2800-3000hp if you wanted reliability, and Voodoo (which used an earlier design and upgraded a few things but not everything) topped out at 3100hp under racing conditions.

This puts the Merlin in 1945 as being respectably close to what the design was physically capable of, and respectably powerful for a piston engine even back them.

So, my question: Just how close did WW2 piston engine designs get to their theoretical upper limit?

I would say that the Merlin exceeded its original perceived upper limits by a long way. The status quo in 1935 decreed that for a 2000 BHP engine you needed much more displacement than the Merlin had, hence the Vulture and Sabre. In fact the Griffon made as much power as anyone needed and as far as I know much of the development of the Merlin was not to produce more power, which is the easy part it was to make the Merlin as reliable with extra power enabled by high octane fuels as it was before. The de rated Vulture produced just a little more power than the two stage Merlin/ Packard type fitted to P-51s and Spitfire Mk !X,

 

[MIflyer]

The curious thing about the Merlin was that by 1941 everyone knew it was too small in displacement to power a single engine front line fighter plane. At 1653 cu in it was smaller than the V-1710 and even smaller than the A6M3's 1700 cu in radial. It was almost 200 cu in smaller than the R-1830 and R-1820 that powered inadequate fighters such as the F3F, F4F/Martlet, P-36/Hawk 75, Brewster F2A, P-66. The demand of air combat known by 1941 demanded fighters of more range, better firepower, better altitude performance, better armor protection, the capability to carry bombs, That demanded more power and more power at altitude, and that of course all that made the longer range problem even worse.

The Germans did not start WW2 with engines as puny as the Merlin. The DB601 was 2069 cu in, the DB 603 2715, the Jumo 211 was 2136, and the BMW 801 was 2562. Even the Hisso 12Y was 2196 cu in. Everybody knew that a single engined fighter needed at least 2000 cu in to be at least competitive, and that was the bare minimum.

The USAAF planned to handle this problem by using two V-1710 or larger or the R-2800 or larger, combined with its Secret Weapon: Turbosupercharging. The British thought that Griffon and Centaurus were their answer; they never even bothered something along the lines of a P-38 with two Merlins, like a souped up Whirlwind.

But the Merlin, by adding a two staged supercharger combined with a liquid cooled inter/aftercooler proved everybody wrong. It was the Little Engine That Could Kick Ass.

 

[imipak]

The curious thing about the Merlin was that by 1941 everyone knew it was too small in displacement to power a single engine front line fighter plane. At 1653 cu in it was smaller than the V-1710 and even smaller than the A6M3's 1700 cu in radial. It was almost 200 cu in smaller than the R-1830 and R-1820 that powered inadequate fighters such as the F3F, F4F/Martlet, P-36/Hawk 75, Brewster F2A, P-66. The demand of air combat known by 1941 demanded fighters of more range, better firepower, better altitude performance, better armor protection, the capability to carry bombs, That demanded more power and more power at altitude, and that of course all that made the longer range problem even worse.

The Germans did not start WW2 with engines as puny as the Merlin. The DB601 was 2069 cu in, the DB 603 2715, the Jumo 211 was 2136, and the BMW 801 was 2562. Even the Hisso 12Y was 2196 cu in. Everybody knew that a single engined fighter needed at least 2000 cu in to be at least competitive, and that was the bare minimum.

The USAAF planned to handle this problem by using two V-1710 or larger or the R-2800 or larger, combined with its Secret Weapon: Turbosupercharging. The British thought that Griffon and Centaurus were their answer; they never even bothered something along the lines of a P-38 with two Merlins, like a souped up Whirlwind.

But the Merlin, by adding a two staged supercharger combined with a liquid cooled inter/aftercooler proved everybody wrong. It was the Little Engine That Could Kick Ass.

Interestingly, the Me 209 was powered by a Daimler-Benz DB 601 in race trim, and was incredibly fast. The DB 601 was 30L, which isn't massively more than the Merlin's 27L.

The max hp output I can find for the DB 601 is 1332 hp. I'm not finding dates, but I'm going to guess that it came out close to the hp rating of contemporary Merlins, despite being larger, although I'm open to correction.

It's only a shade over half of what the RM17SM, the last of the high-power Merlin research machines, was outputting, although it's not possible to say what a flyable Merlin based off it at that time could have done, or where the DB 601 could have been, if it had similar levels of effort and timeframe.

My gut feeling is that you could plot a curve of where you could plausibly get the engine through keeping the architecture and just updating materials and fixing flaws, with a steep diminishing of returns, but that the top of the curve is substantially higher than people give credit.

The architecture limits just how much energy you can squeeze out of a given amount of fuel, and of that how much you must lose through design inefficiencies, but the more I've tried digging around, the more convinced I am that nobody knows what those limits actually are, outside of a general ballpark figure.

Which surprises me, as that would seem like an obvious thing for engineers to want to know, because (beyond knowing what breaks), you obviously want to improve efficiency, but that means knowing, for each section, the ratio of what you waste to what you have to waste. There's no point putting effort in to a section that's wasting a lot but can't be usefully improved on, when there's another bit that's wasting less but can be improved on a lot.

 

[Shortround6]

There is/was a big difference in the air cooled engines and the liquid cooled engines.
The Air cooled engines were very often cooling limited. If you made more power you made more heat and you cooked the engine (cooked the oil in the cylinder bores) which lead to failure very quickly. They also were operating closer to the detonation limits which means they could not use as much boost in general use.
This means that comparison of displacement between air cooled and liquid cooled are always going to be off.

The liquid cooled engines, mostly, were not cooling limited or tweaks to the cooling system (larger radiator, more airflow through the radiator, change in cooling pump capacity) could usually fix the cooling problem/allow for more power.

Merlin stayed much the same from 1000hp to nearly 2000hp, that is the same compression ratio, the same size valves, the same valve timing and same valve lift and the same intake and exhaust ports. Basically all they did was just raise the manifold pressure from the supercharger as better fuel became available. Merlin used the same RPM in the service versions. Racing engines used more RPM.
The improved supercharger/s allowed the engine to go from 20.7lbs (14.7lbs + 6lbs) to 39.7lbs (14.7lbs + 25lbs) of MAP or almost double the amount of air flowing through the engine per minute/hour.
The fuel had the same BTUs per gallon/pound of fuel.

from 1939 to 1945 the Merlin (and the Allison and some other engines) gained around/over 50% in engine life (time between overhauls) and this is were a large part of the improved materials went.

The Allison shows some interesting changes. Like the friction losses almost doubled from the early 1040hp engines to the late 1700-1800hp engines. They needed a lot more piston ring tension to keep the combustion pressures in the combustion chambers. If you have too much blow by past the rings you can turn the oil into carbon grit. They also needed stronger valve springs to keep the intake valves closed against the much higher pressure in the intake manifold which meant more friction on the camshafts.
Everybody (or most people) were turning engine assemblies over with electric motors to measure the friction of different parts (like take the pistons out to measure the friction on the main bearings).
Most of the major companies had a very good idea of what was going on in their engines. They may not have had modern flow meters but they did a lot testing.
They also were not designing racing engines. Aircraft engines needed to be compact (low drag) and light.
The 1939 Mercedes Benz Formula 1 engine made 480hp from 3 liters, at sea level. It also weighed about 603lbs or 1.26lb per hp, or very close to what some RR Kestrel engines could do (800hp for take-off for just under 1000lbs). The MB racing engine was using 2.31 ata (about 19lbs of boost) and fuel that was 86% methanol, 4.4% nitrobenzol, 8.8% acetone and 0.8% sulphuric ether and a two stage supercharger.
Yes, even race car engines from the 1970s could do much better but high flow cylinder heads are larger and if you want more rpm, you need better springs.
Many WW II V-12s used single overhead cams, not dual, for a more compact and lighter head design.

The other big difference between racing car engines and airplanes engines is that the race car driver can walk back to the pits if he blows up the engine.

 

[Sisu]

Here is a genuine 3800 hp Merlin.

A stock Mustang will do Reno at about 330-350 mph at 61 inches. Dan's Sparrowhawk motor had Ridge Runner lapping in the 480 mph range, for a while.

 

[GrauGeist]

The Germans did not start WW2 with engines as puny as the Merlin. The DB601 was 2069 cu in, the DB 603 2715, the Jumo 211 was 2136, and the BMW 801 was 2562. Even the Hisso 12Y was 2196 cu in. Everybody knew that a single engined fighter needed at least 2000 cu in to be at least competitive, and that was the bare minimum.

Yes, they did.

The Jumo210 was 1,201 cubic inches and was the engine used for the early Bf109, Bf110, He112 and the Fw187 prototypes.

The BMW132 radial was a bit larger than the Merlin (but less than the Allison).

 

[Shortround6]

We value displacement a bit too much.

The Germans did not start WW2 with engines as puny as the Merlin. The DB601 was 2069 cu in, the DB 603 2715, the Jumo 211 was 2136, and the BMW 801 was 2562. Even the Hisso 12Y was 2196 cu in. Everybody knew that a single engined fighter needed at least 2000 cu in to be at least competitive, and that was the bare minimum.

What is not listed here is weight, which was much more important to the aircraft designers than displacement.

Engine.............................Displacement/cu................RPM.........................weight...........................power/altitude.

Allison.................................1710...................................3000...........................1325lb.........................1090hp/13,200ft
DB601A..............................2069....................................2400...........................1367lb.........................1000hp/14,760ft
Hispano 12Y-51..............2196....................................2500...........................1085lb.........................1000hp/10696ft
Jumo 211...........................2136...................................2300............................1356lb...........................975hp/15,585ft
Merlin III.............................1647...................................3000............................1375lb.......................1030hp/16,250ft

This in early 1940(?) things changed fast but the basic engines were there. Note that except for the Hispano the other 3 engines are within 50lbs of each other. They make similar power, within 100hp (10%) and similar altitudes (3000ft?) with the highest power engine (Allison having the lowest altitude).
Germans often used an extra 100-200rpm for take-off.

The Hispano is the odd man out. Lightest weight for greatest displacement and the lowest altitude by 2500ft. It needed major changes for either higher RPM or higher boost pressures (higher cylinder pressures). This is problem for the French in 1939-40 and 'what if' scenarios. The Basic engine was old and weak. While it was OK back in 1932 at 650hp the block and crankshaft/con rods were not up to the job of 1000hp +. Soviets added around 170-180lbs of weight and the Soviet engines were not noted for durability.

Germans increased RPM while Allison and RR increased boost pressure using better fuels.
Germans tried using their 96 octane fuel but could not make enough and had to resort to a lot more 'tricks' and finally the DB 605 of 2167cu in and 2600rpm (later versions could use 2800rpm for emergency power) but weight went to around 1670lbs.

Without the higher octane fuels Allison and RR had a bit of problem. Perhaps they could have increase RPM a bit. Perhaps (they did) increase the capability of the supercharger a bit even using 87 octane fuel. Perhaps they could had developed fuel injection.
But the two allied engines were robust, high rpm engines. The extra RPM allowed them to flow 25% more air per minute than a 2400rpm engine and that made up for a lot of the displacement difference. Allison went through 4 different crankshaft models from 1940 to 1944(?) with 3 of them being identical except for heat treatment, much greater fatigue life.
The last version got 27lbs more counterweight and was rated at 3200rpm. They had the space for it in the crankcase and you could put the new crankshaft in an older engine.

The DB 603 was 2,717cu in, it was also around 2030lbs.
The Vulture was 2592cu in but it was 2450lbs(?)

The Merlin XX and 45 getting around 10% more power 2,500-3,000ft higher than Merlin III tided the British over (and making 300-400hp more down lower) also helped cover over the failure of the 3 big British engines. The two stage Merlin 60 series and the full potential of 100/130 fuel propelled the Merlin to legendary status.

The 100/130 fuel allowed for cylinder pressures almost double what 87 octane allowed. And the Merlin was strong enough to take it without too much modification. The 2nd stage took over 200hp to drive and that power came out of the cylinders. Power being generated was lot more than the power to the crankshaft.

 

[tomo pauk]

Interestingly, the Me 209 was powered by a Daimler-Benz DB 601 in race trim, and was incredibly fast. The DB 601 was 30L, which isn't massively more than the Merlin's 27L.

The max hp output I can find for the DB 601 is 1332 hp. I'm not finding dates, but I'm going to guess that it came out close to the hp rating of contemporary Merlins, despite being larger, although I'm open to correction.

Db 601 was at 33.9L, not 30L. More than 25% more than a Merlin or a V-1710.
The DB 601s used on the racing/'rekord' aircraft were making more than 2600 HP - see here, for example.

It's only a shade over half of what the RM17SM, the last of the high-power Merlin research machines, was outputting, although it's not possible to say what a flyable Merlin based off it at that time could have done, or where the DB 601 could have been, if it had similar levels of effort and timeframe.

My gut feeling is that you could plot a curve of where you could plausibly get the engine through keeping the architecture and just updating materials and fixing flaws, with a steep diminishing of returns, but that the top of the curve is substantially higher than people give credit.

The best late-war DB 605s were supposed to make 2000 HP in actual, combat aircraft; see here, for example.

 

[Bretoal2]

The Hispano is the odd man out. Lightest weight for greatest displacement and the lowest altitude by 2500ft. It needed major changes for either higher RPM or higher boost pressures (higher cylinder pressures). This is problem for the French in 1939-40 and 'what if' scenarios. The Basic engine was old and weak. While it was OK back in 1932 at 650hp the block and crankshaft/con rods were not up to the job of 1000hp +. Soviets added around 170-180lbs of weight and the Soviet engines were not noted for durability.

Hispano wasn't the Odd Man, but was the LATE MAN.

Indeed, in 1932 the HS 12Y, with 760 hp, was very good. Seven years later, it had gained barely 200 hp - roughly 100 hp due to a better fuel octane rating, and another 100 hp thanks to a new supercharger, designed outside the company, incidentally (Planiol-Szydlowski/Turboméca).

In 1939, the 12Z should have been available long ago. Unfortunately, the development of French engines from 1935 to 1940 was completely ruined by the political context, especially for Hispano Suiza, whose founder and technical director, Marc Birkigt, had been ousted in favor of Robert Blum, son of the socialist leader of the Popular Front government.

The 12Z aimed for 1,200-1,300 hp with an unchanged displacement of 36 liters (150 x 170 mm bore/stroke). Heavily reinforced, it weighed around 1,350 pounds and finally had four valves. But in a clear sign of the slump reigning in the Hispano-Suiza design office, it was fitted with the in-house supercharger and not the Turboméca P/S. Hence, recovery altitude was uncompetitive.

Same situation for the latest version of the 12Y, the 12Y51, which featured a reinforced connecting rods/crankshaft system, enlarged intake valves, and increased engine speed (Klimov did the same with the VK102/105, but instead of enlarging the inlet valves, he installed two per cylinder!). But the 12Y51 was introduced very late, with still a Hispano S/C, giving a rated altitude of only 3,250 m. Very late, too late....

 

[GregP]

Here is a genuine 3800 hp Merlin.

A stock Mustang will do Reno at about 330-350 mph at 61 inches. Dan's Sparrowhawk motor had Ridge Runner lapping in the 480 mph range, for a while.


View attachment 836948




View attachment 836949

Seriously doubt 3800 hp.

 

[GrauGeist]

Seriously doubt 3800 hp.

Greg, I believe he's referring to Frank Taylor's Dago Red (P-51D).

 

[Sisu]

Definitive horsepower figures for Merlin (and Allison) engines are impossible to come by -- unlike big radials with torque pressure guages that can be used to calculate accurate power levels. The 3800 figure is probably a good guess, judging from how fast Dan Martin went in a basically stock airframe that could be back in airshow trim in a couple of weeks. It even still had the gun blisters. That motor was producing prodigious power -- it beat Dreadnought by a lot.

Besides lots and lots of raw power that motor had some cool (and expensive) tricks like revised supercharger gearing (between -3 and -7 gearing) and custom nose gears to give about a .400 final gear ratio vs the stock .479 or so.

Greg -- if you run across young Steven at the Museum ask him if the 3800 figure is plausible or likely. He has experimented extensively with Merlins in this power range and probably has ideas.

There is no denying that this Mustang, configured like this, lapped Reno 40% faster than a stock Mustang. 480mph takes a lot of power...

 

[imipak]

I've not been able to validate 3800hp, best I can find is 3100hp for Voodoo in races and 3400hp reported on the bench.

Having said that, I've been doing a lot of number crunching in the background. If you update all the materials and cooling, you can get a racing Merlin to 4806hp, but that's the very best the architecture can support.

So 3800hp isn't a physical impossibility, but Voodoo was hitting compression limits at 3100hp, so I'd find something as dangerous as 700hp more puzzling. It's hard to see what it would buy you, except more expensive rebuilds and a greater chance of things going wrong.

However, just because I find something odd doesn't mean someone else didn't think it absolutely the thing to do. Me finding something odd isn't proof of very much.

 

[GrauGeist]

I've not been able to validate 3800hp, best I can find is 3100hp for Voodoo in races and 3400hp reported on the bench.

See my post above - #15

 

[GregP]

I know the Merlin was approaching the fundamental limits of its design, by the time of the Merlin 130 and the RM17SM (which reached 2600hp in branch tests).

Back of the envelope calculations suggest that if you were to build such an engine today using modern materials and modern fuels, you'd probably only squeeze 2800-3000hp if you wanted reliability, and Voodoo (which used an earlier design and upgraded a few things but not everything) topped out at 3100hp under racing conditions.

This puts the Merlin in 1945 as being respectably close to what the design was physically capable of, and respectably powerful for a piston engine even back them.

So, my question: Just how close did WW2 piston engine designs get to their theoretical upper limit?

The Merlin was never reliable at 2,600 hp. The stock Merlin rods are fine up until about 2,100 - 2,200 hp, and then they get pretty fragile. ALL the racing Merlins that are making north of 3,000 hp are using Allison G-series rods. Merlins making 2,100 hp or less can get away with Merlin rods, but they won;t win races with higher hp Merlin airplanes.

Methinks Voodoo made more than 3,100 hp. The only two airplanes I saw race that I think made 3,800+ hp were Strega (with the only Dwight Thorne racing cam) and Voodoo ( used a few more rpm to get it) in the last 2 - 3 seasons. Not too sure Dago Red made that number, but it was potent, for sure. Could have.

 

[GregP]

Nobody dynos a racing Merlin. The engine life is too short.

They do the caculation using a torque meter reading IF they do one at all. You can easily get hp from torque:

hp = (torque [lb-ft] * rpm) / 5252.

Some big radials, like R-3350s in a Lockheed Constellation, for instance, have a BMEP readout and you can back into hp from that, too:

For a 4-stroke engine:

BMEP = (150.8 * Torque [lb-ft]) / (Displacement [ cubic inches]).

For a 2-stroke engine, the 150.8 becomes 75.4 .

The 4-stroke only gets one power pulse every 2 revolutions while the 2-stroke gets one per revolution. Hence, the difference between the 150.8 and the 75.4.

If they don't do a torque meter calculation (many times not), the estimate comes from speed and drag estimates by engineers very familiar with the airplane and it's mods. If you know the areas and coefficients, you can back into the power required to go thath fast. Aerodynamics can get you 30 - 40 mph out of perhaps 500 mph when stock is 360 mph. The rest is pure horsepower.