I am just quoting from memory of the days when Hailwood and Haslam were racing on the Isle of Man, Haslam on a short stroke four cylinder Honda and Hailwood on the Ducati long stroke but desmodromic V twin. The problems of the Ducati were all explained in terms of maximum not mean piston speed. In the end rider and rideability won.
Highest RPM of any WWII Piston Engine
- Thread starter Zipper730
- Start date
Ad: This forum contains affiliate links to products on Amazon and eBay. More information in Terms and rules
And where does that torque come from? Power is the product of piston speed and BMEP, it is the same equation in different form. Higher crankshaft speeds do not produce more power as the stroke shortens. The reason for a short stroke and a high crankshaft speed is weight. You can make the same horsepower at a given piston speed with a short crank and high rpms as you can with a long crank and low rpms, but the longer stroke engine will have longer cylinders and longer connecting rods, and a heavier crankshaft due to the longer throws. The heavier weight is an anathema to high vehicular performance, whether it be in a military aircraft or a racing automobile.
Holding materials, technology, and processes the same between the two, you will always get less torque from a high-revving short-stroke engine than a slower, long-stroke one. However, once geared to the same output or final drive rpm, torque will be about the same. If both have the same piston speed, they will both draw in the same amount of air (and oxygen) per unit time. The bottleneck for getting more oxygen into the cylinder is (all things being equal) the diameter of the cylinder. As the stroke gets shorter and the rpms get bigger for the same piston speed, this breaks down because it is difficult to create valve gear that function at extreme high rpm. Also, the internal shape of the combustion space changes, from a long tube to a short "pie plate". At the extreme upper end, you have to do away with things like valve springs and go to pneumatic valves.
Internal stresses are rarely the limiting factor, unless one is modifying an existing engine. You cannot rev the Chevy in your garage to 10,000 rpm and expect it to hold together! But aftermarket parts are readily available for it to do so, if you can get the cylinder heads to breathe at those stratospheric rpms. Internal stresses rarely limit racing designs. Take a look at a connecting rod for a top fuel dragster to see why.
https://i.redd.it/0i30so5fa7bz.jpg
So why do racing motors break? Because everything is made as light as possible, and stressed to an inch of its life. Also, the shorter the life, the more stress it can withstand for that short duration, and if you can afford to tear down and rebuild after a couple of hours of use (or even a few seconds of use, for drag racing), why not? The connecting rod in the photo is much more highly stressed than the one in your car. If you have an V8 that produces 400 hp (pretty hefty for the family flivver), it's making 50 hp per cylinder. The con rod in the photo makes 1100 hp...for five seconds. It makes 22 times the power. I can assure you that it doesn't weigh 22 times more.
Deleted member 68059
Staff Sergeant
- 1,058
- Dec 28, 2015
"Internal stresses rarely limit racing designs"
- Thats so mindbendingly wrong its hard for me to actually answer, and it doesnt even make any sense with the rest of your own post either - but lets make some examples. If you increase cylinder pressure (upping Comp.R or Boost) the conrod will fail if insufficienly strong in section as a beam in buckling if the "critical length" is too short in. For this you use Eulers beam equation:
...thats called internal stresses limiting your racing engine.
Or the rod can also have compressive failure (which is just about cross sectional area, not the stiffness of its section) if the original rod design was marginal enough, which you can get through simple:
Stress = Force / area ("area" being the connecting rod cross sectional area, and "force" being as per the graphs below, the force down the rod axis)
...thats called internal stresses limiting your racing engine.
This gets very hard to fix because if you increase the cross-section of the rod it normally bangs into the base of the liner at about 90degrees crank angle. If you increase rpm the rod can stretch plastically, which occurs on the exhaust stroke as the gas pressure isnt resisting it, your crank and rod bearings will also start to go once the mean stress approaches 100Megapascals on the bearing shells, which is the typical fatigue limit on modern shell bearings, road cars would be 80Mpa, F1 its about 120Mpa+. Thats basically nearly impossible to counter as you cant increase the bearing area easily so if it happens after youhave designed the engine - you`re stuffed.
....thats called internal stresses limiting your racing engine.
"As the stroke gets shorter and the rpms get bigger for the same piston speed, this breaks down because it is difficult to create valve gear that function at extreme high rpm"
- yes thats why nobody in their right mind since about 1930 has produced engines with pushrods because the rods bend and whip everwhere and the valve motion bears no resemblance at all to the actual camprofile anymore. The japanese motorbike engines are perfectly happy at 15,000rpm with perfectly normal steel valve-springs and
over head cams on finger followers, providing you know your camshaft mathematics and valvespring resonance theory - which they do.
"Higher crankshaft speeds do not produce more power as the stroke shortens."
- Yes...they DO for two reasons. Race engines use a larger bore and shorter stroke, so the capacity can remain the same but crank speed rises. Increasing bore size is a much
better way of increasing capacity than stroke, because the capacity rises with the square of the bore radius (pi x r^2), but only the linear proportion of the stroke. In fact because
gas pressure is basically independant of crank-speed, if we fix bore, fix boost and shorten the stroke and raise rpm, power does ALSO go up even with less
cylinder capacity, as I have calculated, graphed and proven below. This is because gas pressure is INDEPENDENT of crankspeed, so as speed goes up you make the torque
more quickly, even with less capacity you still win. Any slight loss in torque due to the shorter "moment arm" of the shorter stroke crank-throw is minimal compared to the gain from producing more power cycles per second. Measured power is dependant on time (Joules per second, not per rev!), and cylinder pressure is independant of crankspeed, therefore....!
Note these graphs are from ONE cylinder of a DB601E, so naturally x 12 for he engine total output, although the power shown is indicated power so of course you have to subtract
your mechanical losses to get brake output, call it -15%. (so 1350ish bhp)
(above you are looking at the yellow torque curve, and also the purple line, which shows indicated power for ONE cylinder only - a short stroke DB601 never existed but I altered my software inputs to show what it WOULD have produced in theory. We have gone from 130 to 150hp indicated power at higher revs short stroke, the torque has dropped slightly with the short stroke, but with an extra 700rpm this has been overcome)
"The reason for a short stroke and a high crankshaft speed is weight."
-Nope, its to reduce the force in the rod from piston acceleration. Here are the main bearing forces for a DB601 (two graphs down), I`ve made two versions here of the same
capacity, but with short stroke at high rpm, and long stroke a low rpm. Peak cylinder pressure is essentially independant of crankspeed so if you have a
larger bore and shorter stroke at high rpm you win because the engine makes a lot more power but the inertial forces didnt go up. Here is a
DB601 analysis which I`ve quickly done two versions of, same swept volume, 20% bigger bore and 30% smaller stroke. Makes the same power,
but I had to reduce boost pressure by 30% to equalise them on the short stroke/big bore high rpm! Notice that the combined gas and inertia forces have NOT changed, but we`re making
the same power with much less boost (because the revs have gone up with a larger bore).
(for the observant, the graph on the left is "jagged", this is because running less boost but higher revs (RH Graph) means that the inertial forces dominate, the LH graph is more dominated by gas forces on the piston, the cylinder pressure data is real data from in-cylinder pressure sensor so its "noisy", this causes the ripples in the graph, hence this "noise" becomes less prononced in the RH graph)
....thats called internal stresses limiting your racing engine because the combined gas and interia force is basically what goes down the rod !
"The bottleneck for getting more oxygen into the cylinder is (all things being equal) the diameter of the cylinder"
- Well its actually the valve curtain area, which if you are building archaic 2valve engines is pitifully low, of course that only
applies with a naturally aspirated engine, as if you boost it charge density rises and therefore you can get as much charge into the
cylinder as you like through the same ports, because flow is velocity limited not density limited. Second last engine I was working a designer on was a 1600cc V6 running
over 5Bar absolute manifold pressure; made quite a lot of power....and internal engine stresses had quite a lot of time spent on them in the
drawing office ! - It didnt have a very very long stroke...
- Thats so mindbendingly wrong its hard for me to actually answer, and it doesnt even make any sense with the rest of your own post either - but lets make some examples. If you increase cylinder pressure (upping Comp.R or Boost) the conrod will fail if insufficienly strong in section as a beam in buckling if the "critical length" is too short in. For this you use Eulers beam equation:
...thats called internal stresses limiting your racing engine.
Or the rod can also have compressive failure (which is just about cross sectional area, not the stiffness of its section) if the original rod design was marginal enough, which you can get through simple:
Stress = Force / area ("area" being the connecting rod cross sectional area, and "force" being as per the graphs below, the force down the rod axis)
...thats called internal stresses limiting your racing engine.
This gets very hard to fix because if you increase the cross-section of the rod it normally bangs into the base of the liner at about 90degrees crank angle. If you increase rpm the rod can stretch plastically, which occurs on the exhaust stroke as the gas pressure isnt resisting it, your crank and rod bearings will also start to go once the mean stress approaches 100Megapascals on the bearing shells, which is the typical fatigue limit on modern shell bearings, road cars would be 80Mpa, F1 its about 120Mpa+. Thats basically nearly impossible to counter as you cant increase the bearing area easily so if it happens after youhave designed the engine - you`re stuffed.
....thats called internal stresses limiting your racing engine.
"As the stroke gets shorter and the rpms get bigger for the same piston speed, this breaks down because it is difficult to create valve gear that function at extreme high rpm"
- yes thats why nobody in their right mind since about 1930 has produced engines with pushrods because the rods bend and whip everwhere and the valve motion bears no resemblance at all to the actual camprofile anymore. The japanese motorbike engines are perfectly happy at 15,000rpm with perfectly normal steel valve-springs and
over head cams on finger followers, providing you know your camshaft mathematics and valvespring resonance theory - which they do.
"Higher crankshaft speeds do not produce more power as the stroke shortens."
- Yes...they DO for two reasons. Race engines use a larger bore and shorter stroke, so the capacity can remain the same but crank speed rises. Increasing bore size is a much
better way of increasing capacity than stroke, because the capacity rises with the square of the bore radius (pi x r^2), but only the linear proportion of the stroke. In fact because
gas pressure is basically independant of crank-speed, if we fix bore, fix boost and shorten the stroke and raise rpm, power does ALSO go up even with less
cylinder capacity, as I have calculated, graphed and proven below. This is because gas pressure is INDEPENDENT of crankspeed, so as speed goes up you make the torque
more quickly, even with less capacity you still win. Any slight loss in torque due to the shorter "moment arm" of the shorter stroke crank-throw is minimal compared to the gain from producing more power cycles per second. Measured power is dependant on time (Joules per second, not per rev!), and cylinder pressure is independant of crankspeed, therefore....!
Note these graphs are from ONE cylinder of a DB601E, so naturally x 12 for he engine total output, although the power shown is indicated power so of course you have to subtract
your mechanical losses to get brake output, call it -15%. (so 1350ish bhp)
(above you are looking at the yellow torque curve, and also the purple line, which shows indicated power for ONE cylinder only - a short stroke DB601 never existed but I altered my software inputs to show what it WOULD have produced in theory. We have gone from 130 to 150hp indicated power at higher revs short stroke, the torque has dropped slightly with the short stroke, but with an extra 700rpm this has been overcome)
"The reason for a short stroke and a high crankshaft speed is weight."
-Nope, its to reduce the force in the rod from piston acceleration. Here are the main bearing forces for a DB601 (two graphs down), I`ve made two versions here of the same
capacity, but with short stroke at high rpm, and long stroke a low rpm. Peak cylinder pressure is essentially independant of crankspeed so if you have a
larger bore and shorter stroke at high rpm you win because the engine makes a lot more power but the inertial forces didnt go up. Here is a
DB601 analysis which I`ve quickly done two versions of, same swept volume, 20% bigger bore and 30% smaller stroke. Makes the same power,
but I had to reduce boost pressure by 30% to equalise them on the short stroke/big bore high rpm! Notice that the combined gas and inertia forces have NOT changed, but we`re making
the same power with much less boost (because the revs have gone up with a larger bore).
(for the observant, the graph on the left is "jagged", this is because running less boost but higher revs (RH Graph) means that the inertial forces dominate, the LH graph is more dominated by gas forces on the piston, the cylinder pressure data is real data from in-cylinder pressure sensor so its "noisy", this causes the ripples in the graph, hence this "noise" becomes less prononced in the RH graph)
....thats called internal stresses limiting your racing engine because the combined gas and interia force is basically what goes down the rod !
"The bottleneck for getting more oxygen into the cylinder is (all things being equal) the diameter of the cylinder"
- Well its actually the valve curtain area, which if you are building archaic 2valve engines is pitifully low, of course that only
applies with a naturally aspirated engine, as if you boost it charge density rises and therefore you can get as much charge into the
cylinder as you like through the same ports, because flow is velocity limited not density limited. Second last engine I was working a designer on was a 1600cc V6 running
over 5Bar absolute manifold pressure; made quite a lot of power....and internal engine stresses had quite a lot of time spent on them in the
drawing office ! - It didnt have a very very long stroke...
Attachments
Last edited:
I think I need to read that a few times. Great stuff.
Shortround6
Major General
Stress was definitely a limiting factor in many WW II aircraft engines. Since every designer was striving for the best power to weight ratio engines were built as lightly as they could be.
They also wanted engines that had, in commercial service, an economically viable life and even the military, even if they would accept a shorter engine life still had logistic problems (number of spare engines needed per hours flown).
Even back in the 1930s race car engines often had different limits depending on use. Drag racing as we know it today, barely existed but hill climbs were fairly popular. the amount of revs some engines were run at for a 2-3 minute hill climb (or shorter) was often several hundred rpm higher than even 1-2 hour track races, and 12-24 hour endurance races called for an even lower rev limit. During the 60s and 70s it was much the same thing, Nascar or road race engines were limited to less rpm than drag engines and often used different intake manifolds and cams to suit them to the intended duty.
one case is the Hispano V-12 which seemed to hit a wall at just about 1000hp before the war, Post war there were French, Spanish Swiss and russian developments (some were quite different) that went 1300-1600hp (?) but all had gained weight, sometimes well over 200lbs in order to make and withstand more power. One of these engines (Swiss I believe but could be wrong) used a crankshaft over 100lbs(?) heavier than the original Hispano.
They also wanted engines that had, in commercial service, an economically viable life and even the military, even if they would accept a shorter engine life still had logistic problems (number of spare engines needed per hours flown).
Even back in the 1930s race car engines often had different limits depending on use. Drag racing as we know it today, barely existed but hill climbs were fairly popular. the amount of revs some engines were run at for a 2-3 minute hill climb (or shorter) was often several hundred rpm higher than even 1-2 hour track races, and 12-24 hour endurance races called for an even lower rev limit. During the 60s and 70s it was much the same thing, Nascar or road race engines were limited to less rpm than drag engines and often used different intake manifolds and cams to suit them to the intended duty.
one case is the Hispano V-12 which seemed to hit a wall at just about 1000hp before the war, Post war there were French, Spanish Swiss and russian developments (some were quite different) that went 1300-1600hp (?) but all had gained weight, sometimes well over 200lbs in order to make and withstand more power. One of these engines (Swiss I believe but could be wrong) used a crankshaft over 100lbs(?) heavier than the original Hispano.
And you use Euler's beam equation to design a stouter conrod. Problem solved. Internal stress does not limit the engine.
And you use that equation to design a conrod with greater cross sectional area, just like the one I showed for the engine that develops 9,000 hp. Problem solved. Internal stress does not limit that engine.
Except that just doesn't happen with real extreme engines using modern materials.
Look, you've done a great job of summarizing the things a designer has to take into consideration when designing an engine of any capability. Whether it develops 20 hp or 2000 hp, the conrods must meet Euler's criteria for bending, and must handle compressive stresses, and must have large enough bearings to handle the stresses of the intended rpms.
But in practice, these things rarely limit designs. They are a factor that must be considered and accounted for. As I said, try to run your pushrod V8 at 10,000 rpm, it's going to blow up. Rebuild it with high quality forged conrods designed for racing, and low mass forged pistons designed for racing, and make sure you have one of the blocks with the 4-bolt mains instead of the 2-bolt mains, etc, and it'll handle all that just fine. (Of course, I'm leaving out a few details, but you get the idea.) The point is, none of these things limit a clean-sheet design. They may limit what is possible in improving an "as-built" design.
Nobody? Nobody? Yeah, nobody except Chevy, who will sell you a 1,020 hp crate engine with pushrods. Or the guys at NASCAR who are required to run pushrod engines by the rules, who spin them up to 9,000 rpm for hours at a time. Or the guy who holds the current record for piston speed, in a 2100 hp naturally aspirated engine...with pushrods.
I say this as a guy who has raced cars that use those 15,000 rpm motorcycle engines (at power levels about 30% over stock), I love OHV designs, but despite repeated pronouncements, pushrods ain't dead yet.
I will also say that if your design turns much over 10,000 rpm, you probably shouldn't be using pushrods.
I think we're talking past each other here. I was referring to the case where you hold bore size constant and "destroke", using a shorter throw crankshaft and higher rpm. This is contrary to conventional wisdom that "stroker motors make more power". At the same piston speed, no they don't. They make more torque, but since power = torque x rpm, you can get the same torgue at the final drive from the "destroked" motor so long as you change the gearing to the correct ratio.
On the other hand, you're talking about the case where displacement is constant, and the cylinders grow fatter as the stroke grows shorter. Of course I agree with you there, but outside of racing rules and tax regulations, displacement is usually "free" to the designer. And as your cylinders grow fatter and shorter and your rpms climb, eventually you run into breathing problems with your "pie pan" cylinder heads. You can solve that by either going to exotic technology, or by simply increasing the number of cylinders for the same displacement. This is why Honda developed a V5 engine for its MotoGP program some years back. More recently, MotoGP had a 1000 cc limit, with 2 to 6 cylinders, and different weight breaks for different numbers of cylinders (V4s seemed to dominate under those rules, but all the major manufacturers agreed that V6s would have produced more power).
And all this is why I say piston speed is a better, fairer method of comparison than RPM. For fixed cylinder size, HP = BMEP X Piston Speed. If cylinder size is free, then HP = Piston Area X BMEP X Piston Speed.
Tell me about the "before" and "after" stroke of your modified cylinder. I'm not convinced we're comparing apples to apples here.
Again, you're playing the constant-displacement game. Aircraft designers don't design to capacity, they design to weight.
Provided that you can get the thing to breathe correctly with shorter, squatter cylinders, you will aways make more power from the same displacement from cylinders with a total greater cross-sectional area.
Provided that you can get the thing to breathe correctly with shorter, squatter cylinders, you will always have less mass from the same displacement with shorter cylinders.
More power and less weight...what's not to love? You're telling me that inertial forces haven't changed? Why, you must be running at the same piston speed! (Actually, piston speed has increased about 1% for the 2nd case because you picked a nice, round number for RPM, and I'll bet that inertial forces also went up by a similar, inconsequential amount.)
Why all the emphasis on breathing for the shorter, squatter cylinders? Because getting that to work is a function of technology, and the Germans probably didn't have the technology to make a 3900 RPM DB601 to your specs in the 1930s. It would have been a monster if they could.
How so? Weren't the internal stresses the same between the two different designs that had the same piston speed?
You want to impress me, figure out a way to increase the piston speed. Try to imagine what the internal stresses are on a 1000 cu in V8 turning 7700 RPM.
Annnnd...the valve curtain area scales as the cylinder area, for the same geometry. Double bore size, get 4x the cylinder area and also 4x the valve curtain area.
I don't understand what you mean by "density limited". Flow in an optimized NA engine is velocity limited. This doesn't change when you add boost, although a lot of other things inside the combustion chamber do. And I wasn't aware that pressure changed the speed of sound in air very much, although temperature does, and you uusually pick up a fair amount of temperature along with that boost.
None of that has changed in the 70-odd years since then. There's a curve for stress vs. life, and it is exponential. You pay very dearly in terms of life for each extra HP at the far end of the curve!
In the 1980s, BMW raced with highly modifed stock 4 cylinder engines in F1. In stock form, they would produce between 80-130 hp, depending on the model, and do that for 60,000 miles without a fuss. Prepped for the race and running much turbo boost, more like 800 hp. They would last for one race, a few hours. They also had qualifying versions of the same engine boosted up to 1500 hp (nobody really knows for sure, since the dyno only went to 1280). These would last for...a few laps. Then junk.
All other things being equal, stress is equal to piston speed. Double the piston speed, 4x the stress. During wartime development, piston speeds were climbing higher and higher for the same point on the stress-vs-life curve...which makes piston speed a very good way to compare different engine programs by different nations.
Deleted member 68059
Staff Sergeant
- 1,058
- Dec 28, 2015
I ask you a very serious question, without the answer to which I am not going to contribute to this thread further.
Please tell us how many engines you have designed. I`ll patiently await your list.
Please tell us how many engines you have designed. I`ll patiently await your list.
Just a basic question Snowy. Do you take compressive and tensile yield strength to be the same value, did they have equipment for accurately measuring both in the 1930s?
Deleted member 68059
Staff Sergeant
- 1,058
- Dec 28, 2015
Entirely material dependant, for example cast-iron is specifically VERY different in tension and comression as when its formed its full of micro-cracks, which is one reason its very good as a bore material because the cracks retain tiny oil particles. Under compression these cracks are simply forced shut, and so basically dont matter but under tension mean the value with be VERY different to that under compression.
Wrougt materials will typically be similar in tension and compression.
In the 30`s they had reasonably advanced test machinery even for fatigue and torsional testing and suchlike. So yes I think 100% all good engine designers from the late 20`s onwards would have certainly had a very good grasp of all that stuff. I think the chief disadvantage then was that materials standarisation almost didnt exist in the very early days so you`d
have to do tests on the exact stuff you got from your chosen mill, as it was probably specific to them and not something you could order from other mills. This changed slowly between WW1 and WW2. Much of the work of the Royal Aircraft Factory in WW1 was on materials standarisation and testing, so the designers would be able to make better parts.
Wrougt materials will typically be similar in tension and compression.
In the 30`s they had reasonably advanced test machinery even for fatigue and torsional testing and suchlike. So yes I think 100% all good engine designers from the late 20`s onwards would have certainly had a very good grasp of all that stuff. I think the chief disadvantage then was that materials standarisation almost didnt exist in the very early days so you`d
have to do tests on the exact stuff you got from your chosen mill, as it was probably specific to them and not something you could order from other mills. This changed slowly between WW1 and WW2. Much of the work of the Royal Aircraft Factory in WW1 was on materials standarisation and testing, so the designers would be able to make better parts.
KiwiBiggles
Senior Airman
Please don't make your contributions dependent on someone else's posts. I for one greatly welcome your posts, and read them with great interest. What Biff is to the air combat threads, you are to the engine design threads.
Some engine designs that shouldn't work do and others that should be wonderful are disasters. It comes down to fixing the unexpected problems by been able to recognise what they are.
If you can find the book by Harry Ricardo of Rolls Royce "(1953) The High-Speed Internal Combustion Engine (4th ed.)) it covers lots of interesting stuff as he was there from 1915 to 1964 as all the major piston engine developments took place (and he was responsible for alot of it). His development story to get the Merlin engine reliable shows lots of interesting "outside of the box" hacks to make it work, Rolls Royce Aero engines earned its name for reliability.
If you can find the book by Harry Ricardo of Rolls Royce "(1953) The High-Speed Internal Combustion Engine (4th ed.)) it covers lots of interesting stuff as he was there from 1915 to 1964 as all the major piston engine developments took place (and he was responsible for alot of it). His development story to get the Merlin engine reliable shows lots of interesting "outside of the box" hacks to make it work, Rolls Royce Aero engines earned its name for reliability.
I spent my life in mechanical testing. In pipeline engineering the material is tested in a tensile test although most loads are actually compressive. On one project this became a major issue and so there were tests of both compressive and tensile loads using test samples and full sized pipes, while similar they were not the same.
Deleted member 68059
Staff Sergeant
- 1,058
- Dec 28, 2015
A selection of torsional fatigue testers is visible in this aviation industry torsional testing laboratory in Germany, 1941 at the Schenck company. Caption on the wall reads: "We are inescapably bound to the great eternal laws of nature and must obey them, whether we like it or not"
I`ll say this is still JUST on-topic as it maybe illustrates that the means to develop higher speed engines was certainly not limited by materials testing know-how.
I used to have to visit the Salzgitter Mannesmann Vallourec "Vorschung" in Duisburg, it is like a place owned by a Bond villain, hundreds of machines nodding and whirring constantly, but hardly a human to be seen lol. They managed to prepare the test pieces to the wrong standard, so no one is perfect lol.A selection of torsional fatigue testers is visible in this aviation industry torsional testing laboratory in Germany, 1941 at the Schenck company. Caption on the wall reads: "We are inescapably bound to the great eternal laws of nature and must obey them, whether we like it or not"
I`ll say this is still JUST on-topic as it maybe illustrates that the means to develop higher speed engines was certainly not limited by materials testing know-how.
View attachment 525335
Deleted member 68059
Staff Sergeant
- 1,058
- Dec 28, 2015
After some cajouling I`ll try to respond to this post in a useful way.
I think the first point to make is that detail is getting in the way here, as there is obviously a fundamental difference of viewpoint which me posting endless graphs wont address.
What you are essentially trying to say is something like "with modern materials the limits experienced by old designers dont exist and modern designers can basically do what they want without creative boundaries imposed from stress issues, if they have enough money to use the very most advanced materials"
This is about half right, the limits for what can be done have of course moved very significantly since 1945 in reciprocating engines, but - they have NOT become unrestrained by them on any level, in any component of an engine, right up to F1 or NASCAR (incidentally nether of which actually make very good studies for the state of the art from a materials OR engine layout perspective as both racing series are extremely heavily restricted legally by the regulations imposed by the race regulators). So its very dangerous to say things like "yea but in NASCAR..." these guys are fighting a very unique battle, in that they are trying to make power from hopelessly archaic engine layouts (2valve heads, an injection system thats far more primative than most WW2 systems etc etc. Their challenge is to make the engine work ANYWAY, regardless - its a real mistake to look at a NASCAR engine and say thats what you`d aim for if you had a clean sheet of paper, NASCAR engines are a triumph of development over layout, nobody designing them would argue (I know two personally). Similarly an F1 engine is useless in almost any other application, and are also heavily regulated (bore, stroke, bank angle, materials... all fixed, no ceramics, no cermets (a hybrid metal with ceramic particles) etc etc).
Now to try to make a useful comment on your actual points:
1,2,3,4) All basically broadly answered by the first part of my post above ^
5,6,7) I think you probably typed a response after reading the first line of each of my paragraphs, because you`re basically asking me to repeat myself. Anyway I better allow for the 2nd possibility that I didnt explain well the first time. I did a few different models, the graphs are showing if you keep capacity constant but increase RPM, this showed that the reduction in stroke (and the associated reduction in piston acelerations) actually made it possible to keep stresses in-check despite the higher speeds. In order to keep the power "matched" between both engines I had to actually reduce the boost on the high rpm, big-bore short-stroke engine. I though that was an interesting way to present the info given the title of this thread, which is at least in part about increasing the RPM of aero-engines.
The power can also go UP even fixing the bore size and reducing the stroke, and increasing rpm with set boost. Its just because the increase in bangs per second overcome the loss in torque. Nobody is saying high RPM is so simple as to just rev it more, but if you can take care of the mechanical stress, its perfectly possible to make more power from a
smaller engine which has a higher rpm. There will obviously be limits to that, if you reduce the stroke by HALF but increase the revs just 20%, NO of course this will
result in less power, but if you allow for even a small increase in mean piston speed (using it as a guide to how much to lower stroke), and then increase RPM, power goes UP.
Lets just put in two sets of numbers which you can go and test to satisfy if you wish. We`ll use a DB601 cylinder of standard bore and stroke to start.
(Con-rod length 259mm in both cases, cylinder pressure kept same)
Standard DB601E, 2700rpm, mean piston speed = 14.4m/s
:Bore = 150mm - Stroke = 160mm - Cylinder Swept Volume = 2.827 Litres (33.9L total)
Power (assuming 85% mechanical efficiency) = 1358bhp
De-Stroked DB601, higher rpm, 4200rpm, mean piston speed = 15.4m/s (only 7% higher because the stroke has gone down).
:Bore 150mm - Stroke = 110mm - Cylinder Swept Volume =1.944 Litres (23.32L total, 68% of the size of the standard engine)
Power (assuming 85% mechanical efficiency) = 1400bhp
(The crossover point is about 4000rpm, when the power is identical even with 10 litres swept volume missing.)
If we take it a bit further and go towards (but NOT at) the piston speeds Jumo were looking at in 1944 with the 213J.
De-Stroked DB601, very high rpm, 4500rpm, mean piston speed = 16.5m/s
:Bore 150mm - Stroke = 110mm - Cylinder Swept Volume =1.944 Litres (23.32L total, 68% of the size of the standard engine)
Power (assuming 85% mechanical efficiency) = 1490bhp
If you are not interested in believing this, all I can say is that you must go and build your own program (can be done even in EXCEL) to model this, and prove it to yourself. You can
find all the equations needed online if you search, or buy the book: "Internal combustion engine fundamentals" by John B. Heywood, which is to be found on the bookcase of
anyone who works seriously on engines.
8)
I was (along with the point about curtain area), demonstrating that this is pretty meaningless because since flow through the valve IS velocity limited, NOT
density limited, you can shove almost any amount of air you want through any set of valves by increasing boost, as its just higher density. So more
energy is contained within the same volume when it passes through the valve. Hence bore diameter really doesnt come into it. I think you just
didnt read that properly as you seem to be throwing the exact point I made back at me.
Your point covering that base by stating "all things being equal" is sadly not very useful to make as its never possible to do, for example if you change the bore size the only way
to fit bigger valves without altering all sorts of other things is to go to a shallower included valve angle, and so on etc etc. Sometimes its a useful phrase,
but not when discussing bore sizes, too many variables which just make it meaningless, as there are just too many knock-on factors.
Hope that clarifies and helps in some way, this is advice from someone who`s job it is to actually do this, hence why I happen to have a load of pre-prepared
software to do all these calculations.
wuzak
Captain
If your destroke engine ran at the same piston speed it would have the same power? ie at 3,927rpm?
Using the same rod length would, I assume, lead to a small improvement in efficiency because it would have less extreme angles in the short stroke version.
Deleted member 68059
Staff Sergeant
- 1,058
- Dec 28, 2015
1) Correct, if you put them both at 14.4m/s mean piston speed (2700 vs about 3950rpm)
2) Yes because the piston thrust loads will be reduced slightly, but in reality you`d almost certainly want to shorten the rod to lighten it and allow the deck height
to be reduced, which would give a good reduction in overall engine weight. The limit to this would be the interference between the side of the rod and the bottom
of the liner, which will bang into eachother at some point as you reduce the rod length.
See, that statement just baffles me. I've never encountered a team that said, "We'd turn more RPMs, but our conrods would break". Every single time, it's always been, "If we could just get a little more air into the cylinder, we could turn more RPMs and make more power". Every single time.
I'm aware of a motor that has a mean piston speed of 7540 fpm, 38.3 m/s. EN30B crankshaft, 7075 aluminum connecting rods, 2618 aluminum pistons. Now, those are quality, expensive materials. They represent a special order from Timken or Alcoa. But they are not state of the art materials nor are they exotic materials. 7075 in various grades is heavily used in the aircraft industry today.
For those following along, 38.3 m/s represents about 13,900 rpm in NASCAR, almost 29,000 rpm (!) in F1, or 7,180 rpm in a DB 601 (not destroked).
If your argument is that TBO is also a performance spec, and that the above combination of materials meets the power spec but not the TBO spec, then I agree with you. I doubt it would last 2,000 seconds between overhauls, let alone 2,000 hours. If that's what you mean by "limited by materials", then I don't think we have any disagreement. We can then proceed to a discussion of how long a fighter engine needs to last between overhauls when it is subject to be shot down every sortie.
We got to this point because in comparing different aircraft engines, we are essentially comparing apples to oranges. What is truly important here? Power to weight? Power to displacement? RPMs? Piston speed? Mean cylinder pressure?
The original poster asked for an RPM comparison, and I pointed out that piston speed might be a more fair way to compare dissimilar designs. Looking at your DB601E numbers, I think they just back up what I'm saying. You say that power is equal at roughly 4000 rpm for the destroked version. Well, without a fancy combustion-chamber simiulation, I can observe that the piston speeds exactly match at 3927 rpm...and that mechanical stress should be about the same between the original at 2700 and the destroked design at 3927.
So, why didn't they try to build a destroked version of the DB601? It couldn't have been materials technology, since at about 4000 rpm the materials stresses are the same. You'd want a casting that had cylinders 50mm shorter so you could save some weight, but what's wrong about that? Same power for less weight is a good thing in an aircraft.
I will throw out for your consideration the suggestion that they may not have known how to make cylinder heads that would flow the same amount of air at 4000 rpm. Didn't they have to go to 4 valves per cylinder to get the 213J up to 3700 rpm? The DB601 started with 4 valves per cylinder, so I'm not sure what the path to higher-flowing heads would be.
I think I understand you perfectly, and of course your simulations are perfectly believable to me. I just think we're talking past each other.
Power = BMEP X Piston Speed. This is right out of Charles Fayette Taylor. It hasn't changed in decades. In all your examples where HP increases, BMEP presumably stays about the same and piston speed goes up. As you pointed out for the 4000 rpm version, at the same piston speed, the power is equal.
I don't understand your fixation with displacement. Unless you're designing for a racing series, or a tax code, does anyone every hand you a spec that says this engine must have between 2995 and 3000 cc of displacement? Now, if you're using displacement as a proxy for weight, that makes sense to me. If your only goal is to maximize power for a fixed displacement, then you want to maximize total cylinder area at the expense of stroke. There are limits to this; you're not going to have the bore in excess of 3 times the stroke, and you're not going to be designing any 48-cylinder engines.
If area is fixed and stroke is not, you can save weight by shortening the stroke, even at the cost of displacement. For the same piston speed (and higher RPM), you'll make the same power. If your block size is fixed, this little trick doesn't accomplish much of anything. If you can redesign your block to reduce deck height, then you can save a lot of weight.
If you can figure out a way to increase either the BMEP or the piston speed, you can make more power for the same displacement (something known to top racing teams for decades).
I thought the root of our disagreement was this:
But Power = BMEP x piston speed is the same equation! Take two motors, same horsepower but at different RPMs. The one that revs higher will have less torque. Now put them both into a gearbox so that they have the same final drive RPM: Torque at the back end of the gearbox is equal!
Do the same thing with BMEP held constant, different strokes, different RPMs, same mean piston speed: Power is equal. Many of the examples you've posted in this thread show that, one to within 1%.
Ok, now I understand where you're coming from. This is a case of looking at the same thing from a different angle will yield interesting and informative insights. What you say is true but misleading to people (like a lot of amateur engine builders) who don't see the forest for the trees.
I agree that you can increase flow by increasing boost. But there's a lot more to the story!
Many people comprehend "more boost = more flow = more power", and neglect to understand that airflow is just as important for a boosted engine as for an NA one. "Why do I need to spend a lot of money on heads and valves and stuff? I'll just turn up the boost..."
But the boost pressure exists because there is resistance to flow. Improve the airflow into the cylinder, and depending on the source of boost, either the boost pressure will drop for the same power, or the boost pressure will stay the same for more power.
And, for a boosted engine, as flow starts to become orifice limited, bigger valves make for better flow. And larger bore cylinders have room for bigger valves.
I'm not sure I follow. If we're talking about being limited by some constraint, such as boring out an existing block, then yes, there can be a kajillion knock-on effects.
If you're trading off stroke for bore in a clean-sheet design, then the shape of your combustion chamber is changing. For small changes, the knock-on effects are small, and the larger valves, etc, are just pure profit. As the changes become larger, combustion effects become dominant, and you start running into stuff like your larger valve is sitting right where your spark plug needs to be.
That said, I think that most people understand that "all things being equal" doesn't apply over an infinite range. All things are equal...until they're not any more.
Shortround6
Major General
Known to most (if not all) aircraft engine design teams too. Which is why they kept increasing the boost from the superchargers.
"According to A C Lovesey, who was in charge of the Merlin's development, "The impression still prevails that the static capacity known as the swept volume is the basis of comparison of the possible power output for different types of engine, but this is not the case because the output of the engine depends solely on the mass of air it can be made to consume efficiently, and in this respect the supercharger plays the most important role."
This of course assumes you have fuel that will support the increased BMEP in the cylinders without detonating.
I would note that in the last 60-70 or so years a number of racing organizations either limited the amount of supercharging that could be used or banned it out right in attempt to level the playing field.
Increasing the BMEP by 10% increases the stress on the parts by about 10% (peak loads may be a bit higher?). increasing the rpm by 10% increases the stress loads by 21%, which is why they try to get the loads back down by shorting the stroke but the piston (and valves) still have to come to a complete stop and accelerate to full speed either once (piston) or 1/2 times (valve) per revolution and the valves, since they are closed for a larger part of the two revolutions of the crank may have even higher accelerations?
BTW friction also increases with the square of the speed so you have more internal losses with the high speed/ short stroke design.
Just throwing this out there since I never designed an engine but you might want to keep the original deck height (or near it?) and use longer rods?
Post war P & W R-2800 engines used different pistons that had higher wrist pin locations than the older engines that allowed the use of longer rods for less rod angle (this was in the 40s right after the war) and by older engines I mean the C series engines used in the P-47M & N and the F4U-4.
as far as this goes:
In an airplane you can't walk back to the pits (or wait for the tow truck) when an engine blows up.
and a less dramatic level, engine life plays a big part in overall logistic footprint/needs. If you are trying to operate several hundred aircraft (some of the twins) 6000 miles from the engine factory what kind of engine life do you want? Super performance and light weight calling for an engine swap every 4-8 missions or perhaps 5% less performance but swapping out the engines at 20-30 missions? That 5% might mean the difference between life or death for the pilots. Not having planes in the air at all might mean life or death for troops/sailors depending on air cover.
