Why were all radial engines about the same width?

[Jeff_F_F]

I'm motivated by my interest in the use of aero engines in tank development. A 1m wide engine would be really nice for a low-profile tank though the bump on the rear deck of the Hellcat wasn't a major issue for that design. But most radial of the WWII and pre-war era were about 1.35m-1.40m wide, regardless of the number of cylinders, displacement of each cylinder, etc. Does anyone know why this is? And does anyone know what would be the consequences of using shorter connecting rods, especially with regard to power and durability?

My default assumption would be that the diameter was chosen not based on the needs of the engine since that would change with the number and size of cylinders, but on the typical diameter of a typical fuselage of a plane with a single engine and a single pilot.

 

[pbehn]

It is always best to keep the frontal area as small as possible to reduce drag. Beyond that the bore and stroke dimensions affect how high the engine can rev. The bigger the engine then the bigger the prop and the bigger the prop the more needs to be done to cope with it, like the Corsairs gull wings for example.

 

[Jeff_F_F]

So what I'm getting is that a longer crankshaft means longer stroke which gives more power, but the practical limit is that the engine can't get wider than the fuselage. I've been looking at it wrong thinking that smaller was better and "why were they all so big" and thinking maybe they had to be to get proper cooling in the airflow around the front of the fuselage. Instead it sounds like larger is better mechanically but they are limited to a maximum width by what is practical as an outer limit by the width of the fuselage.

 

[pbehn]

An engine takes years to design and perfect, a fuselage takes weeks/months. The fuselage is designed to fit the engine or its cowling.

 

[Jeff_F_F]

Right, but but if the engine manufacturer can make their engine whatever size they want, why are they all about 1.35m-1.4m in diameter if there isn't some other factor limiting their size? What all single-seat designs have in common is the vertical distance from the feet to the shoulders of the seated pilot (with the exception of the P-47 where the limiting factor was the massive turbocharger and all of its ducting - but nobody is going to make a brand new engine that is as wide as the Thunderbolt when there is the perfectly awesome Double Wasp right off the shelf and being manufactured in staggering quantities). Otherwise, why wouldn't some engine designer have decided they could get more power by making a wider engine if the height of the pilot wasn't an essential limiting factor? Since the height of the seated pilot plus whatever structure goes beneath their seat is the smallest fuselage with the least frontal area that can be made, that seems to be also the maximum diameter of radial engine anyone is going to design so all of the engines end up being that diameter instead of growing larger in the quest for more power.

 

[mikewint]

Back in the early days of aviation radial engines were very popular, it seemed like all of the big planes of the 20s, 30s and early 40s used them. But some time during WWII (or so it seems to me) aircraft manufacturers switched from radial engines over to inline engines.

Now there are only so many cylinders you can put on a radial before the prop becomes a useless prop!

One limitation of radial engines is that they have to have an odd number of cylinders (without heroic design measures). It's due to the 4-stroke timing pattern. Now you can easily get an even number of cylinders using TWO rows of cylinders. So there can be 18 (2x9), 14 (2x7), 28 (4x7) arrangements and more.

The big advantage of radials at the beginning was their large frontal area, which meant they could be air cooled. An inline air cooled engine can run pretty hot on the rear cylinder. The bigger the engine, the more of an issue cooling becomes. There are some huge air cooled radials, in sizes that would be unthinkable for an air cooled inline engine.

As technology advanced, more complex water cooled engines became more popular. However air cooling was still a big advantage in military planes, due to the absence of a delicate cooling system that could be damaged, enabling radials to persist in this arena. Of course these days no-one would dream of putting a piston engine in a fighter.

Nowadays liquid cooling is almost universal for piston engines in general applications, with small aircraft being one of the few areas where air cooling has managed to hold out to some extent. In the absence of a military application that prefers air cooling, the large frontal area has become the radial engine's downfall, due to the large aerodynamic drag it produces. The other problem is that the valve gear is quite complicated, which means overhead camshafts and multiple inlet/exhaust valves per cylinder are impractical. This makes efficiency vs power to weight ratio a big issue as follows:

For an engine to have good power to weight ratio it must run at high RPM. To run at high RPM it must breathe effectively. To do this with just two valves they have to open wide. To stop them colliding with the piston, the compression ratio must be low, which limits efficiency. Hence most modern engines have more than two valves per cylinder to get around this problem.

Modern air cooled aviation engines are only available in much smaller sizes than typical WWII fighter engines. Typically they have 4-6 cylinders, too few for a radial, but enough to cause cooling problems on the rear cylinder in an inline.

The compromise is a flat (boxer) engine with 2 or 3 cylinders on each side. This also overcomes two other problems with radials: upper cylinders obstructing pilot view and oil collecting in the lower cylinder when stopped. You also have at least one of these problems with an inline, depending if you mount it upright or inverted.

Small aero engines can be either water cooled or air cooled. The designer has to ask: given the issue of reliability in aviation, does the weekend pilot really want an engine as complicated as the one in his car, with water cooling, cambelts and engine management, all of which could go wrong

 

[Jeff_F_F]

Thanks for the detailed answer. My specific question was prompted by the apparent convergence of a large number of radial engine designs in both Germany and America to a very similar diameter in spite of a wide variety of different numbers of cylinders. There are upgrades from 7 to 9 cylinders per row, and from 1 to 2 rows. There was even the crazy 4-row BMW 803 where they couldn't air-cool the rear rows and had to water cool it, and the 28 cylinder Wasp Major that managed to air-cool all of its cylinders and even managed to work unlike the BMW. But nobody tried to increase power by making their engines wider than 1.4m until the Wright Cyclone 22 and that was only a bit wider at 1.5m wide. At the same time I wasn't seeing any engines significantly smaller than this either which puzzled me even more.

I'm guessing that the upper limit is due to the height of a seated pilot. As for the lower limit, I'm realizing I had neglected to look at Japanese and British radial engines, where there are a number of radials up to 2-row 14 cylinder models that are much smaller than this, such as the Mitsubishi Zuisei and Bristol Taurus, though there were others from both nations in the same diameter range as U.S. and German engines.

I'm starting to think that part of what I was seeing in WWII engines was that so many U.S. radial engines and all German radial engines in that era were evolutionary descendants of the Pratt and Whitney Hornet which BMW produced under license. Meanwhile the Wright engines were similar in size to the P&W engines since the two companies were competitors and neither was motivated to make smaller engines when the quest at that time was for ever more power. And since I wasn't looking at other nations I got a false perspective on the field.

 

[pbehn]

Jeff as an engine increases in radius the frontal area increases by pi x radius squared inches. There is a convergence in benefit/cost with engine diameter (frontal area) bore stroke. I have no idea what you mean by vertical distance of feet to shoulders of seated pilot. I have seen the footwells of most WW2 fighters modelled here and there is lots of space. The Germans in particular were keen on lifting the feet upwards purely to cope with high G forces.

 

[mikewint]

While it is true that pilot visibility may be poorer due to the width of the engine on single-engine aircraft, tight fitting cowlings helped reduce this factor somewhat. Equivalent inline engines often resulted in overly long noses, which similarly impaired visibility directly forward.
As to drag, having the cylinders exposed to the airflow did indeed increase drag considerably. The surprising answer was the addition of specially designed cowlings with baffles to force the air between the cylinders. The first effective drag reducing cowling that didn't impair engine cooling was the British Townend ring or "drag ring" which formed a narrow band around the engine covering the cylinder heads, reducing drag. The National Advisory Committee for Aeronautics studied the problem, developing the NACA cowling which further reduced drag and improved cooling. Nearly all aircraft radial engines since have used NACA-type cowlings. Because radial engines are often wider than similar inlines or vees, it is more difficult to design an aircraft to minimize cross sectional area, a major cause of drag, although by the beginning of the Second World War, this disadvantage had largely disappeared as aircraft sizes increased, and multi-row radials increased the power produced in relation to the cross sectional area.

 

[pbehn]

There are other simpler issues, for every inch increase in diameter the axis of the propeller moves 1/2 inch closer to the ground unless you eliminate the pilots forward view completely.

 

[Shortround6]

Basically the problem with a radial is you have a bunch of almost opposing cylinders.

Now in order to make power each cylinder has to a be a certain size, many companies built 3,5, 7 and 9 cylinder radials all using a common cylinder size. use the appropriate number of cylinders for desired power. Due the length/height of the cylinders, the length of the connecting rods and the throw of the crankshaft there was very little, if any difference in overall diameter. at least until you get to 11 cylinders in each row. Because of the number of cylinders acting on one crankpin trying for high rpm was rather difficult.

At any given time in history you were only going to get so much power from a single cylinder ( most aircraft engines, even liquid cooled V-12s started as one or two cylinder test rigs) due to limits on fuel, limits on cooling (number of sq in of cooling fins per cylinder on air cooled engines) and limits on rpm (limits on materials). Basically anybody trying to build a 500hp radial in the late 20s was going to come up with a 9 cylinder engine of between 1340 (or a bit more) and 1750 cubic inches.
Several companies did try to build small diameter 14 cylinder radials, The P&W R-1535, the Bristol Taurus and the Gnome-Rhone 14M.being the most common. Germans used it on the Hs 129.
The last was the champion of small diameter at just under 1 meter but it achieved that small diameter by using a small displacement, 1159 cu in which is a result of using small cylinders 122mm bore and 116mm stroke. Few large aircraft engines were over square but this helped hold down the piston speed of this high rpm (for a radial) engine. 3,030 rpm. Production engines were limited to 700hp and one announced but not flown (?) engine was rated at 820hp.
However a two row 14 is longer and heavier than a single row 9 of the same power and has a lot more parts. Two row 14s became really popular when using cylinders of similar size (usually just a bit smaller) to the big 9 cylinder engines in order to provide much more power than the 9s could. (or to avoid cooling problems and vibration of the rally big 9 cylinder engines).
The smaller diameter of the small 14 never really paid off in streamlining as the more powerful 1520-1820 cu in 9s just bulled their way through the air.

I would note that P & W actually went to longer connecting rods with a piston pin placed higher in the cylinder on post war "C" series engines. Kept the same over all length of the assembly and same diameter engine but lowered stress and added to engine life (less side thrust?).

Now if you want low powered engines then they made a host of 5, 7, and 9 cylinder radials of 55 to 300hp, the smaller ones being rather small in diameter. For some reason this was national. The US having quite a number of different small radials and very few 4 cylinder inlines before WW II, while Europe and a host 4 and 6 cylinder inlines and very few small radials. Post war the US flat four and six engines quickly dominated the market for light planes.

 

[Jeff_F_F]

I was wondering about the side-thrust issue since the angle between the centerline of the cylinder and the lobe of the crankshaft would be smaller at the middle of the piston's stroke if the crankshaft was longer.

I'm working on a game where players can build their own tanks and specify and improve the various sub-components including engines. My knowledge of radials is way behind my knowledge of v, flat, and boxer engines so I needed a better sense of how these engines scaled in power, weight, and dimensions with different design choices, so this is really helpful.

 

[Shortround6]

A lot of large aircraft radial engines had reduction gears on the front and a supercharger on the back. also the accessories were on the back, any maxillary pumps ( hydraulic, air pumps, generators. starters, etc.

 

[Jeff_F_F]

That was a bit of a dumb question on my part... I pulled up a cutaway pic of a BMW 801 and the supercharger was really obvious even though it was in German.

If a radial was installed as a tank powerplant with a 10-15 degree tilt, would it mess up the flow of oil in the engine?

 

[Shortround6]

The US did install the engines at an angle.

The US used radial aircraft engines because, pre war, they were available and there were no good US made alternatives. The light tanks (Stuart) needed around 250hp and the M-2, M-3 and M-4 Series needed 350hp or more. The M-6 heavy needed over 600hp.
Army tank engines ran on 80 octane fuel, only airplane engines that ran on 80 octane were trainer engines (or old transports).
The 80 octane is from a post war US army manual. tales of US tanks needed av-gas are nonsense even if most trucks ran on 72 octane.
British jumped to the old Liberty V-12 engine.
Trying to use air cooled aircraft engines in tanks calls for really big fans and lots of air ducts.

Trying to use really powerful engines actually requires very large and heavy transmissions and steering gear. Even modern tanks use combination transmissions/steering gear that is often heavier than the supercharged diesel engines driving the tank. and they are similar in volume. Maintenance is also a big problem, pull the engine or change spark plugs and check valve clearance on bottom cylinders how? Extra room for mechanics? lots of hatches in the sides/bottom to reach the engine in place?

 

[swampyankee]

First, to correct an error: stroke is not important for power; since piston speed is roughly constant, increasing stroke reduces rpm.

Because of combustion dynamics, engines with stroke to bore ratios much less to than one tend to be inefficient. For decent dynamics, one ends up with a diameter of about 9 to 10 times stroke.

 

[Jeff_F_F]

Good to know about stroke - thank you!

And that was exactly the application of a tilted engine I had in mind, since engines in U.S. tanks installed horizontally had U-joints to redirect their driveshafts downward like that. I appreciate the pic.

 

[Shortround6]

First, to correct an error: stroke is not important for power; since piston speed is roughly constant, increasing stroke reduces rpm

In theory yes, in practice? in the 1920s and 30s a major limit was the valve gear. in the late 20s and very early 30s just getting an engine to run for several hundred hours without replacing one or more valve springs was a major achievement.
Very few pushrod car engines in the mid 1950s were good for much over 5000rpm and many were less than that. By the late 60s/early 70s the push rod limit was over 6000rpm in street cars and well above that in race cars.
If something other than piston speed affects your rpm limit then bigger cylinders are the way to more power.

 

[swampyankee]

In theory yes, in practice? in the 1920s and 30s a major limit was the valve gear. in the late 20s and very early 30s just getting an engine to run for several hundred hours without replacing one or more valve springs was a major achievement.
Very few pushrod car engines in the mid 1950s were good for much over 5000rpm and many were less than that. By the late 60s/early 70s the push rod limit was over 6000rpm in street cars and well above that in race cars.
If something other than piston speed affects your rpm limit then bigger cylinders are the way to more power.

Very true. Also, because of combustion dynamics, longer strokes tend to be better, and lower rpm means that there is less power required to drive the valve gear. On the other hand, engine weight tends to correlate directly with displacement, so it tends to be better to use a shorter stroke and greater rpm to achieve power than lower rpm and greater stroke. Except for niche (I'm inclined to say toy) applications, like racing cars, very short strokes are not useful (I think Formula 1 engines are have stroke:bore ratios of about 0.4). At the other extreme, as stroke/bore ratios get larger, there comes a point where trunk pistons become impractical, and one needs to go to a crosshead engine.

 

[MiTasol]

Given the large number of tank experts on this thread I am asking a loosely related tank engine question.

Some US WW2 tanks used a pair of Cadillac 346 engines instead of the radial.

I have a friend with one of these engines who is desperate for a set of manuals. Parts for the engine are readily available but he is yet to find any manuals.

Does anyone have, or know of a source of, manuals for the Cadillac 346 tank engines.

Thanks

Mi