“Atkinson Gas Engine Animated” by MichaelFrey – Own work. Licensed under CC BY-SA 3.0 via Wikimedia Commons – http://commons.wikimedia.org/wiki/File:Atkinson_Gas_Engine_Animated.gif#mediaviewer/File:Atkinson_Gas_Engine_Animated.gif
(first posted 10/15/2012) Let’s start by stating the obvious: Apparently, the “Atkinson Cycle” and “Miller Cycle” names are not trademarked since neither is true to its original patents when it comes modern Otto engines. Of the two engines, the Atkinson is the more significantly changed. From this animation of the original/genuine Atkinson Cycle engine by Bill Todd at douglas-self.com, it’s quite obvious that this complex mechanical dance is not what’s taking place under the hoods of millions of the Toyota and other hybrid cars claiming to use an Atkinson cycle engine. Indeed, a key aspect of James Atkinson’s 1882 engine design is being employed, but in a much simpler form.
Initially, Atkinson wasn’t searching so much for greater efficiency, but for a way to circumvent patents for four-cycle engines that had been awarded to Otto. His solution was to design an engine whose crankshaft rotated only once once per the four strokes. The very complicated (not to mention mechanically complex and inefficient) “crankshaft” that allowed this also made possible a power stroke longer than the compression stroke, and the key to potentially greater efficiency.
In a typical four-cycle engine, the compression and power strokes are of the same length, making the compression ratio identical to the expansion ratio. When the expansion stroke is longer than the compression stroke, more energy from the expanding gas can be utilized; ideally, the pressure at the end of the expansion stroke would be no greater than atmospheric. With less energy converted to heat, and more energy converted to mechanical force, this cycle is intrinsically more efficient.
But the original Atkinson Cycle engine was built only in very small numbers, obviously mostly due to its mechanical complexity and high cost.
Looking for maximum efficiency for their new Prius, Toyota applied the key aspect of the Atkinson Cycle engine to their conventional 1.5-liter four and then adapted it by delaying the closing of the intake valves during the compression stroke. Since the combustion chamber is proportionately smaller, combustion pressures at ignition are comparable with a conventional four-stroke. Toyota also optimized (delayed) the engine’s exhaust valve timing in order to maximize the expansion stroke.
The large throttle opening produces both lower manifold vacuum and reduced pumping loss. The Atkinson Cycle is very efficient within a defined operating range (around 2500 – 4000 rpm in an automotive engine), but substantially reduced torque makes it largely unsuitable in a conventional drive train. However, full-hybrid systems like Toyota’s HSD fully mask that deficiency, and can achieve near-diesel efficiency when the engine is the primary power source.
The Miller Cycle engine was patented in 1957 by Ralph Miller. It describes “a new and improved method of operating a supercharged, intercooled engine” (including two- and four-strokes) by utilizing a servo-operated compression control valve in the cylinder head.
Today’s Miller Cycle engines actually are supercharged Atkinson Cycle engines, as Toyota and others recognize. The use of a mechanically-driven, positive displacement supercharger overcomes the torque deficiency of the Atkinson engine, and the delayed compression stroke, lowers the charge air temperature well below that of a typically boosted engine, thus allowing a higher compression ratio and greater ignition advance.
Of course, there are mechanical losses associated with superchargers (turbos don’t provide the immediate boost required), which probably explains why the Miller Cycle engine hasn’t enjoyed wider use. The Mazda Millenia (its engine is pictured above) ended up making only 10 more horsepower than Mazda’s conventional 2.5-liter V6, yet did so at considerably more expense. Of course, today’s variable valve technologies makes the possibilities for incorporating various elements of Atkinson/Miller Cycle function much more feasible.
In 2007, Mazda announced a normally aspirated “Miller Cycle engine” for its Japanese-market Mazda 2. Given that Miller’s patent was all about supercharging, it seems that Mazda is really bending the Miller name here; after all, a non-supercharged Miller is simply an Atkinson Cycle engine. But what else is new?
Uh…let’s just call this a day and give me a good old pushrod 283 or any other small block Chevy and call it done.
My apologies to Mr. Miller and Atkinson – and to Mr. Rube Goldberg, for that matter, too.
Much too deep a subject for my shallow mind!
Seems to me this could all be reversed to make a high power racing engine. If the intake stroke is longer than the power stroke, you would pull in extra air/fuel and compress it more. This would essentially do the same thing as a standard otto engine with a supercharger attached. But you don’t have to carry the weight of the supercharger.
Well, let’s look at it that way : if you just wanted to have a longer intake stroke, you’d need a larger engine anyway ; why would you ever want to restrict the expansion stroke ? By doing this, you’re ruining efficiency for literally absolutely no benefit whatsoever (and added complexity of course).
What you want is just a larger engine with a very high compression ratio : not a bad idea per se (racers have done this from the dawn of the automobile), but not exactly groundbreaking either.
Wonder if an electrically driven supercharger added to Toyota’s HSD could work?
It undoubtedly would. But why? The Prius’ engine is powerful enough, especially in conjunction with the electric motor. And that would add cost, which is the last thing Toyota would want to increase on the HSD.
For Toyota’s take on a “hybrid sports car.”
“Powerful enough” depends on how one defines “enough”
The 420hp Mustang GT is certainly powerful enough. Yet there is the 662hp Shelby GT500 …
Toyota GT86h
2.0 liter or 2.5 liter “Electric Miller Cycle” boxer engine, with electric blower. The 2.5 liter Atkinson-cycle HSD in the Camry Hybrid makes 200hp combined. So 270 to 300hp combined with the blower seems reasonable.
Delete the rear seat and use the space for the battery pack. Could even put the HSD eCVT in the back (a la Porsche 924) next to the battery to make a rear-biased weight distribution and minimize the length of the thick high-current “orange cable.”
The 86’s famed low center of gravity is preserved by countering the high-placed blower and intercooler with the low-placed battery pack.
35 MPG city (where the supercharger kicks in more), 40 MPG hwy on 87-octane gasoline …
Paul, superb reading. Check out what Coates Engineering , a local firm here in Wall NJ,has been doing for sometime. Their spherical valve technology is unique.
A electrically driven supercharger would work but it would reduce the efficiency significantly.
Why? It’s electric. It doesn’t have to spin when it’s not needed. And there is no parasitic loss in the belt and pulleys. The only real penalty is weight.
Because the efficiency of the Atkinson cycle hinges on the fact that the cylinder is only half filled.
The motor/generators are the “super charger” in the HSD system. If you are going to turn a super charger electrically you’ll need a lot of watts so do you drain the battery and not use the electric motors for power or use the motor/generator to drive the super charger and trade their thrust for a drag?
If you’re going to electrically spin a blower at low RPM, better to do it with an exhaust-driven turbo where you can also generate electricity under high RPM, like the current F1 engines do.
High on my list of peeves are companies which distort the accepted meaning of specific terms in order to create an artificial distinction they can advertise. I’m looking at you Daimler-Benz. “Coupes” have 4 doors.
I never really noticed that the same thing was taking place in drivetrains. Mostly because I never really knew the details of the Miller and Atkinson cycles.
You never knew that the Crown Victoria wasn’t really a Victoria? Or that most Broughams weren’t actually broughams? Accepted meanings of terms change over time. `Creative’ advertising has been going on for a long time. 99.4% anyone?
I’m with you Steve. MB calling their 4-door sedans “coupes” just because they have a swoopy roofline is marketing nonsense.
It *is* marketing nonsense. But it isn’t new, or unusual.
But it is also a pet peeve of mine. Obviously, if I (and Steve) take offense with the practice, then we are aware it is going on. I shouldn’t have needed to state that explicitly to make it clear.
The technical definition for coupe doesn’t have anything to do with the number of doors. It originally meant a vehicle with a short distance between the driver’s hip line and the rear axle, while I think the SAE definition was based on the volume of the rear seat (the cutoff being, as I recall, 33 cubic feet). By either standard, a fair number of two-door cars — including the Coupe de Ville — would qualify as two-door sedans, not coupes. So…
I didn’t know that. Thanks.
Started by Rover in the 60s with their P5B V8 fourdoor coup’e.
An entire web site every bit as big as Curbside Classic could be built devoted to pointing out “Marketing Nonsense” in the automotive world and inviting comments.
Interesting point, and answered well by Ate Up With Motor. I always wondered why the P5 Rover Coupe was named a coupe when it had four doors. But then it had a substantially lower roof line than the P5 sedan, so perhaps that was to reduce the rear seat volume?
Kudos, Mr. Niedermeyer – you have explained the Atkinson and Miller cycles, and provided a nifty animated graphic besides. I had never paid much attention to these before, but once explained, they make for a fascinating variation on the standard Otto cycle engine.
The reduction in pumping losses from the modern version of the Atkinson cycle engine is not due to a larger throttle opening. It is due the late intake valve closing that actually causes the engine to work against that throttle with the low opening to assist in filling the cylinder with the next intake stroke.
Its due to larger intake valve opening and throttle passage. The late closing serves to reduce the amount of air in the cylinder at start of compression. Intake pumping losses have already occurred by this time.
No because some of that air was “recycled” from the previous cylinder. Only approximately half of the intake air to fill the cylinder needs to be drawn past the throttle for each intake cycle once the first cylinder has filled when starting the engine.
When cyl #1 is starting it’s compression stroke cyl #3 is starting it’s intake stroke so the air that is escaping cyl #1 goes into cyl #3. With the VVT-i the previous cylinder is filling the next cylinder for up to 2/3 of it’s intake stroke. That is how it significantly reduces pumping losses and why it needs a low throttle opening. Because the cylinder filling isn’t creating vacuum for much of it’s stroke the pumping loss is significantly reduced and because the preceding cylinder isn’t compressing for much of it’s stroke the pumping loss is significantly reduced there too.
I said pumping losses were decreased by larger intake valve opening and throttle passage. You’re right about losses at the throttle opening, of course. The article is ambiguous on this point. Thanks for your clear reply!
Good technical article Mr. N. You neglected to mention that variable cycle engines became practical only due to the modern invention of variable valve timing. We probably have to thank Honda for popularising it (the vtec just kicked in yo!) but it is doubtful that this technique was not being used a long time ago. All major companies’ engines now have computer controlled VVT/VTEC/VVT-i/MiVEC, etc. So implementing an `Atkinson cycle’ or a `Curbside Cycle’ is a much simpler reprogramming matter, given a suitable engine. It would at most require a different cam profile. It would be fun to study the valve timing profiles of `normal’ Otto-cycle engines and determine how they deviate from the Otto-standard, but companies don’t release that information to the public. Bah, humbug, is what I say.
The Atikinson cycle engine does not need variable valve timing to be functional. In the case of the Prius engine it always operates in an Atkinson cycle.
Valve timing
Intake Open –30~10° BTDC
Close 120~80° ABDC
Exhaust Open 32° BBDC
Close 2° ATDC
As you can see the valve timing is commonly available for almost all engines that have made it to production either in the factory or aftermarket service manual/online service.
Hey, if the intake and exhaust timings are different, that requires variable timing. Prius can get away with fixed cam profiles because the engine never has to take full load. And no, the valve timing charts for newer computer-controlled iVTEC etc are not available from the manufacturers (nor are ignition charts, injection amounts/pressures, you name it). They are reverse engineered (and sold for good money) in the aftermarket. The fools think its some top secret proprietary information. As if their competitors can’t reverse engineer their stuff.
> Hey, if the intake and exhaust timings are different, that requires variable timing.
No it doesn’t. Intake and exhaust valves are actuated from different cam lobes, even on non-OHC engines.
A lot of conventional OHV V8 cams have a split profile which holds the exhaust valve open for longer duration than the intake valve to enhance the scavenging effect. There sure isn’t VVT in my old Chryslers to achieve that!
OK! Completely forgot about that. Thanks for the correction.
Prius’ Atkinson cycle could work without variable timing, but for the record their engines have always used Toyota’s VVT-i system. Surely it lets them tweak the engine’s performance beyond a fixed cam profile’s capability.
Honda’s iVTEC can almost simulate an Atkinson cycle. So can Suzuki’s. In my mileage-obsessed country, this is the raison d’etre for variable valve timing. Better efficiency==Better mileage. Especially on highways, Honda’s 1.5L mill easily gives 20kmpl or more if driven in the sweet rpm spot. In city driving, engine shifts to high-torque, low-mileage mode. I’m sure Toyota is pulling something similar off. Of course Prius’ cam profiles will be Atkinson type, but VVT-i is a big part of it.
but isn’t valve timing only part of the equation? higher lift = more CFMs?
Excellent! That’s a brilliant and mesmerizing animation. I like the Atkinson story as a case of patents stimulating innovation. To work around Otto’s patent the wacky linkage wouldn’t have been enough by itself, since it didn’t actually improve anything. So he comes up with the Atkinson cycle improvement.
Funny, I always thought the Miller cycle came from engine builder Harry Miller. Thanks for setting me straight on that.
Fine job of explaining a difficult subject. In engineering school, when I saw those partial differential equations I dropped Thermodynamics like a hot rock.
The wacky linkage *is* the Atkinson cycle. Newer engines simply replicate that with wacky valve timing, which Atkinson was too dumb to think of. The effect of increased efficiency is a happy serendipity (totally cancelled, and then some, by mechanical losses in the original). Of course, patents caused development of such a markedly stupid design is more a commentary on the patent system than on Atkinson. Its not as if he wouldn’t have realised it was stupid, but he had to do something!
He undoubtedly could have made the crankshaft linkage to still rotate the crank once every four strokes (to get around the Otto patent) without the variable stroke. I think he was quite aware of the benefits of that, and saw an opportunity to incorporate it in his design. Or so I assume.
I doubt efficiency was on his mind for this engine, but of course can’t tell for sure. He apparently developed other engines too. It would make a nice patent history project for someone into that sort of thing.
From his patent : http://www.google.com/patents/US6526935 which makes it quite clear that the longer expansion stroke was a specific aspect and benefit:
4. An apparatus as described in claim 1 which, when operated as an internal combustion engine, is capable of being configured by said drive means and ratio of the offset of the input/output shaft crank to the offset of the second crankshaft crank, to cause the expansion stroke to be substantially longer than the intake or compression stroke.
5. An apparatus as described in claim 1, when operated as an internal combustion engine, which will produce both a moment about the output shaft axis which is greater throughout the expansion stroke than that of a conventional engine of equivalent intake stroke and an expansion stroke which is longer than that of said conventional engine.
Thanks for digging this up! I wasn’t expecting *you’d* do it, but thanks anyway. So he *was* aware of the efficiency benefit of his asymmetrical stroke design. Don’t know why he chose this complicated way to implement it instead of using different valve timings, but then, as MikePDX says, cam grinding may not have been very reliable back then, and that would still have infringed Otto’s patents, so he had to make do with what he had. Very interesting.
Took about 30 seconds….
Easy when you’re not in a sleep deprived state. CC is so good I can’t even go to sleep without at least glancing through. But I’m currently in professional hell.
Ditto re having to read CC before bed; and being in professional hell…
Such a study has been published. C. Lyle Cummins did a book on this, Internal Fire on the history of I.C. engines. There’s a chapter on Atkinson as well as others working around the Otto patents (6 cycle engines were common, and fairly efficient because of the better scavenging. Just don’t ask for more power or torque.) Atkinson also did a “differential” engine with opposed pistons and more linkage that got much the same cycle.
Lyle did several other books, including a biographies of Rudolf Diesel and his father, Clessie Cummins. Yep, that Cummins. (Clessie is also responsible[if that’s the word :-)] for the Jacobs engine brake found on big diesels.)
Thanks for the rerun, Paul, I never get tired of a mesmerizing animation. In your reply to CarCounter above I think you meant to use this link, https://www.google.com/patents/US367496, which is Atkinson’s patent. Shaw’s 6,526,935 cites Atkinson’s 367,496.
James Atkinson was awarded the John Scott Medal by The Franklin Institute in 1889, so it’s safe to say he wasn’t a complete dolt.
An invention’s utility depends on countless factors at the time. What was the state of the camshaft grinding art in 1882? Linkage lengths in the crank could have been a more manufacturable way to do it then. Agreed, it would be an interesting study in the History of Technology.
> The Mazda Millenia (engine pictured above) ended up making only 10 more horsepower than the conventional 2.5 L Mazda V6, but with considerable more expense.
I think you got your horsepower and torque numbers mixed-up, Paul. 210-170hp = 40hp advantage for the 2.3L Miller engine. That’s a 23.5% horsepower improvement. Torque was also up 31%.
Also, look at the complete specs for both engines:
2.5L naturally aspirated: 170hp @ 5800, 160ft.lb @ 4800
2.3L Miller cycle: 210hp @ 5300, 210ft.lb @ 3500
The 2.3L has +40hp and +50ft.lb on the 2.5L and the torque peak has dropped by over 1000 RPM, pushing it into the useable range of normal driving. The 2.3L engine should pull like a freight train compared to the standard one because it doesn’t have to wind-up to get into its power band. (As anecdotal proof, I read that the Milennia S had better highway fuel economy than the regular version.)
I omitted to say “ by the end of its life cycle , the most powerful Mazda 2.5 V6 was making only ten hp less than the 2.3 L miller engine”.
And yes, the supercharger certainly filled out the torque curve.
FWIW, the 200 hp KL-ZE was not US emissions compliant, and really was a high-RPM screamer along the lines of Honda DOHC VTEC engines. The most powerful KL-series V6 sold in the US made 170 hp.
I stand corrected. Since Mazda’s K Series topped out at 2.5 Liters, forced induction certainly makes sense.
Correction (or better, update): The Prius XW30 and XW50 models use the 2ZR-FXE, which is 1.8L . It’s based on the US Corolla’s engine, a good hint that it couldn’t use the Atkinson crankshaft. Toyota boasts it has no accessory belts, not even for the water pump, & rates 98hp alone, 134hp with electric.
Where hybrids excel is in city & suburban driving; high-speed Interstate MPG is probably worse than diesels & only a bit better than some recent Otto-cycle models. The Prius Prime is intriguing as I could commute all electric one-way. Now if it was just fun to drive…
This whole discussion of holding the intake valve open during the compression stroke brings back strong memories of Briggs & Stratton’s “E-Z-Start” which they started using in the early 1960s from what I can figure.
As a kid, I remember rebuilding a 3hp vertical-shaft lawn mower engine. I was checking the valve clearances at TDC of the compression stroke, but something wasn’t right. The exhaust valve wasn’t fully closed (at least if memory serves correctly, this was 35 years ago)! I decided that this couldn’t possibly be right, so I ground down the valve stem to achieve some clearance.
Being that the engine was a crank-start model, I quickly discovered that I had defeated the easy-start feature! When warm, releasing the crank start lever would cause the engine to spin until it hit the compression stroke and then suddenly stop! I had to replace the starter with a recoil type and then yank the heck out of it.
Sorry for drifting a bit OT, and thanks for bringing up the explanation – so many auto articles these days mention the Atkinson cycle but so few actually explain it.
I did the same exact thing when i was like 12 years ild :). Never heard of anyone else doing it too. I ran it for years, but the rope would zing your hand ever time you started it. Later i used a welder to build up another lobe on the opposite side (intake too) to make a steam engine out of it.
Too bad we can not tap into the power of that little bird who leans forward to take a sip of water then returns to its original upright position then continuously repeats the deed.
This makes my head hurt so badly that I moved it to my desktop so I can read again until the need for pain is satisfied. One thing nobody seems to have mentioned is the possibility for flexibility with respect to rod length. That’s something constantly preached by Yunick (long rods) but generally ignored. Otherwise I don’t see how it’s better because of the increased number of moving parts.
Don’t bother answering that again unless it’s simpleton friendly. I will read again when my head is better.
The short stroke on the (intake &) compression stroke, then longer stroke on the power (& exhaust) stroke gives you greater a greater expansion ratio compared with the compression – you get more work from the engine, in the sense of the technical term which can be expressed as work = force x distance.
You might make a comparison to a compound-expansion steam engine where the steam goes into a second piston to further expand and extract more work.
Watching that atkison GIF cycle over and over gave me an inspiration: Put another cylinder at about the 11 or 12 o’clock position, with the connecting rod attached to where the first cylinder’s conn rod attaches. You would have a combustion cycle twice as often. This would help with the torque issue without introducing as much rotating mass as a separate connecting rod and (pivoting rod?)
Just put a 24V cylinder head on a slant six, and call it a day. You’d have a powerful, torquey motor thst would run 200,000 miles.
And now I read that next year’s Toyota Yaris for the European market with have a new 1.5 liter motor-to replace the 1.3-that has variable valve timing with the ability to switch between Atkinson and Otto cycle “as needed.” It will be a little more powerful than the 1.3, but a lot more efficient.
Nope. It’s a term that predates cars, going back to the horse-drawn era. And it always meant a closed body cut lengthwise — or short wheelbase, with one seat (for two pax), as opposed to the four-seater berline. Convertible / open top body styles have their own terms, but coupé is usually not one of them. Some British manufacturers do use the term “drophead coupé,” but not the majority. There was also the coupé de ville, a.k.a sedanca, where the driver’s compartment was open and the rear passenger seat was enclosed, very much like the horse-drawn carriages. French, Italian, German (until recently), Japanese and American carmakers always used the term coupé as two-door closed body, not open ones.
Atkinson’s patent seems simpler than what Honda came up with. In production about a decade ago and installed in houses in Japan, a 1Kw cogeneration (power and heat) unit.
https://global.honda/innovation/technology/power/Exlink-picturebook.html
Super interesting technically. I’m excited to see the full transition from ICE to electric drivetrains and the theoretical reduction in maintenance.