Why Aviation Needs

the Stirling Engine


Darryl Phillips

"Every advance in aircraft design has been preceded by an advance in powerplant design." That quote is attributed to the late John Thorpe, but may indeed be much older. The Stirling engine, invented in the year 1816 by a Scottish minister, Robert Stirling, can provide aviation with that much-needed advance in powerplant design.

Consider our present powerplants. The Internal Combustion piston engine is a mature product, any advances in the state of the art are likely to be small. We have obtained about all this machine has to offer, and the problems of noise, vibration, fuel consumption, and the need for explosive fuel aren't expected to change. Non-piston (Wankel, et cetera) IC designs may offer some improvement in power/weight ratio, but not much.

The turbine engine has it's own set of problems. It does not scale well to the sizes needed for personal aircraft, and while power/weight ratio is excellent, fuel consumption is terrible, as is initial and hourly cost. Future advances in ceramics offer some promise of cost savings, but low altitude and part throttle efficiencies don't show much hope of improvement.

Internal combustion engines have tremendous variations in torque, which show up in the airframe as vibration and fatigue, and in the propeller as well. Common four cylinder engines such as Lycoming and Continental show torque that varies from a negative 100% to a positive 350% of the nominal torque. In other words, twice each revolution the prop is turning the engine rather than the other way around. This is illustrated in FIGURE #1. The prop is the flywheel and it must be heavy and robust. The constant torque reversals make variable pitch designs more complex and troubleprone. A Stirling with the same number of cylinders and identical horsepower has a torque variation of +/- 5%! If we can make an engine with constant torque like this, it can lead the way to major advances in propeller design.

If personal aviation is to prosper, we need a new powerplant. Particularly, we need an engine that is silent. We make a lot of noise, and we do it over people's homes. People equate noise with danger. That may not be a correct impression, but it doesn't matter, it's the perception of society that airplanes are dangerous machines. The certainly sound dangerous, don't they? Perception drives legislation, and aviation cannot stand much more of that. Waldo Pepper is half a century gone, yet we're making as much noise as ever. If aviation is to be accepted in modern society, we must quiet our machines.

Not all noise is from the engine. It is well known that as tip speeds approach Mach 1, a great amount of noise is created by the propeller. Until we solve the powerplant noise, it will be difficult to work on prop noise. With the Stirling, we can move on to the propeller and not only quiet it down, but improve it's efficiency as well. Every horse that goes into making noise is a horse lost from cruise speed. After both the engine and prop are quiet, we'll begin hearing air burble across flap gaps or whatever is the next noise source. Every improvement will benefit the pilot and society as well.

Next, we need an engine that runs on whatever fuel is available, but particularly on fuel that is not explosive. The needless loss of life due to fire in accidents is tragic. And it is preventable. As 100LL production declines, the price will rise. Someday only turbine fuel will be commonly available at airports. We'd better have an engine ready.

"......Containers of different fuels were provided along the frame of the generating set including crude oil, lubricating oil, olive oil, salad oil, diesel fuel, gasoline, and liquid petroleum gas. The engine would run happily on all these individually or as mixtures. On one visit to Eindhoven in 1966 the author witnessed the machine running, at 3000 revolutions per minute, on a mixture of crude oil containing large bubbles of gasoline and alcohol, with one of the multiple-sided English three-penny pieces standing vertically and motionless on the crankcase." (quoted from the book STIRLING ENGINES by G. Walker, Clarendon Press, 1980)

The Stirling offers many advantages. Silent operation is a benefit to the aircraft occupants as well as society at large. The lack of vibration prevents fatigue, both in the airframe and in the passengers. The Stirling cycle is fuel efficient because of the regenerator, a device that greatly improves specific fuel consumption by saving and reusing the portion of heat energy that is thrown away in the exhaust of present engine designs. Fuel consumption affects not just cost of operation, but the design of the aircraft itself, leading to savings in weight and improvements in performance.

FIGURE #2 compares a conventional engine, a Stirling engine, and a rubber band! If we had planes powered by rubber bands, they wouldn't change power at altitude. Internal combustion engines decrease as a function of air density. Stirlings, on the other hand, put out more shaft horsepower at higher altitudes.

High altitude operation is a tremendous potential advantage of the Stirling powerplant. Since it operates on the difference between combustion temperature and outside ambient temperature, the Stirling puts out more power as the plane climbs into cooler air. Consider this: If a given aircraft could hold constant power, it would cruise twice as fast at 40,000 ft as sea level, due to decreased drag alone! But conventional engines work on volumetric efficiency and as air density decreases so does power. Just when the plane wants to go faster, the engine says no. Not so with the Stirling, which has no volumetric limitations (it is a sealed system like a refrigerator and has no knowledge of the density of the air, it only knows the temperature). Power actually increases as the plane climbs, and cruise at 40,000 will be about 2.2 times sea level cruise speed. Besides the benefit of speed, there is the obvious safety advantage of flying above most weather.

FIGURE #3 is adapted from the Pilot Operating Handbooks of the Cessna 172M and turbocharged 285 HP C-210. It is "normalized", so each plane has 100% of it's sea level cruise speed. Conventionally-aspirated (not turbocharged) engines are usually rated for continuous operation at 75% power. As the plane climbs above S.L., power remains at 75% and airspeed increases because of reduced airframe drag. At about 8000 or 9000 ft, however, full throttle is insufficient to produce 75% power due to decreased air density, and above that level airspeed decreases. The drag goes down, but the power goes down much faster, so the effect is reduced airspeed.

Turbocharging allows the aircraft to climb higher before this effect takes hold, but usually by the upper teens the plane again loses airspeed with increased altitude. FIGURE #3 shows what the aircraft could do with a "rubber band", that is, with constant power that does not change with altitude. The Stirling is affected by altitude in the opposite way from the conventional engine, and actually gains power as altitude increases!

Of course the C-172 or Bonanza of today will not operate at the higher flight levels. And that is the point: The Stirling will revolutionize the way we design light aircraft.

If Stirlings are so good, why don't we have them already?

Old Stirling designs from the late 1800s were made of cast iron and were very heavy, as much as a hundred pounds per horse. So the engine got a reputation for being too heavy to consider. Of course IC engines from the same period were heavy too, but we tend to forget that. By the time high temperature alloys were available, IC technology had outrun the competitors.

".....These imperfections have been in a great measure removed by time and especially by the genius of the distinguished Bessemer. If Bessemer iron or steel had been known thirty five or forty years ago there is scarce doubt that the air engine would have been a great success......It remains for some skilled and ambitious mechanist in a future age to repeat it under more favourable circumstances and with complete success..." (written in the year 1876 by Dr. Robert Stirling [1790-1878])

Much work has been done on Stirlings for auto engines. NASA has displayed a blue Dodge pickup at Oshkosh several years, with a ten million dollar Stirling under the hood. You may have seen it running up and down the flightline during the afternoon airshow, transporting performers in front of the crowd.

Ford, GM, and the European car makers have all run Stirlings. But the public wants a new engine to feel like the old one, and Stirlings are different. It is difficult to achieve good idle fuel economy at a red light for minutes at a time, then produce instant tire-squealing power when the light turns green. Stirlings are not good car engines.

But the mission profile of the aircraft engine is totally different from the car engine. The characteristics that make the Stirling wrong for a car make it right for an aircraft.

The premier publication for Stirling engine development is STIRLING MACHINE WORLD. SMW is a quarterly, distributed worldwide to design engineers, the academic community, and individual Stirling buffs. The following four sections are adapted from a series on Aircraft Stirlings by Darryl Phillips published in 1993 and 1994.

Part #1.....Ten Tips for Stirling Engine Design

  1. ..Always keep in mind that an engine is just a machine that converts energy from one form to another. Energy stored in fuel or from some other source is changed to kinetic energy. The task is to maximize thruput and minimize losses.
  2. ..Minimize temperature losses. The first temperature loss isn't even measurable in an immediate way. It is the difference between the temperature a given fuel could have burned at, and the temperature it actually burns. Having purchased a particular fuel, the user should be able to utilize the maximum potential contained in that fuel. The second loss we encounter is the difference between combustion temperature and heater temp. And so on thru the engine. At each step, strive to eliminate temperature differences. Ideally, the entire temperature spread from combustion to ambient should be seen in the hot to cold temperature swing of the gas itself. Temperature losses are inevitable of course, but they should be minimized to the extent possible.
  3. ..Consider the logarithmic nature of temperature, as opposed to the linear nature of heat energy. A given temperature loss on the cold side hurts much more than the same number of degrees on the hot side. Thinking of temperature in linear degrees distorts the problem, it's helpful to work in decibels of temperature (DBT) rather than degrees. (See Part Two, "The Adventures of Heat and Hot")
  4. ..Minimize aerodynamic losses. This is the biggest problem in many contemporary Stirling designs. Think like an aerodynamicist. Pay close attention to gas velocity. Watch sharp bends, abrupt changes in cross section, anything the high speed molecules will dislike. Think like an electrical engineer. Move the gas in parallel rather than series. Keep the bore big and the stroke small, probably around 10% of bore. Think torque rather than speed. Study the low delta-T designs, they utilize tiny temperature differences to the best advantage, and they point the way to efficient Stirlings in higher delta regions too.
  5. ..Minimize dead volume. Strive to move 100% of the gas from heated to cooled space. An ideal Stirling would see average gas temp equal to heater temp at one end of the cycle, and average gas temp equal to cooler temp at the other end. While this is never fully possible in the real world, try, try, try.
  6. ..Use the correct compression ratio. Better yet, forget compression, and use the correct expansion ratio. These are the same number, but expansion better reflects the conversion of heat to kinetic energy. The ideal expansion ratio is the square root of the ratio of average absolute gas temperatures taken at the hottest and coolest points in the cycle. Note that any gas not fully heated or cooled reduces this average. This is the fundamental reason it's so important to heat/cool as near 100% of the gas as possible.
  7. ..Pressurization. It is tempting to double the power of an engine by doubling the internal pressure. But twice the power means twice the heat energy in and out. If exchanger area remains constant, temperature losses will increase, and twice the power won't be obtained. Plus, the higher pressure fluid will suffer greatly increased flow losses and power output will decline further. As pressure rises, containing it requires more material. That material might have been better utilized providing increased heat transfer area. Like all the other design compromises, pressure is a tool to be used. Just remember that excessive pressure is just as harmful as excessive dead volume or excessive flow loss.
  8. ..Regenerators. No doubt the regenerator is the bait that hooked most of us on the Stirling in the first place. The idea of using the same thermal energy more than once is fascinating and very worthwhile. But don't let the regenerator get in the way of the engine. There are reports of Stirlings that ran OK, then ran better when the regenerator was removed. If the above rules are flaunted, nothing in the regenerator will save the design. To the extent the regenerator increases dead volume, or adds flow loss, it is the enemy of the engine. Simplicate the regenerator.
  9. ..Much of the thermal promise of the regenerator can be achieved with the recuperator. Same idea, applied to the incoming combustion air and the exhaust and coolant streams. If energy is only allowed to leave the machine via the crankshaft, the desired efficiency will be attained. And it's easier to deal with dead volume and tortuous flow paths in the combustion gasses than inside the Stirling itself.
  10. ..Build the engine. No amount of computer modeling takes the place of a real engine. The industrial revolution owes it's success to the fact that the computer hadn't been invented yet. Else we would still be modeling and simulating the cotton gin, the telegraph, the steam engine, the railroad. Build the engine, the marketplace awaits.

Part Two.....The Adventures of Heat and Hot

One of the ten tips listed last time involved the logarithmic nature of temperature. So let's begin there. We'll show why degrees are not the best way to measure how hot something is. Then we will introduce a way to graphically illustrate how much usefulness a given source of thermal energy contains, and how most of the energy escapes unused. Plus some thoughts on the creative process.

Creativity is closely coupled with intuition, that is, with our unconscious grasp of a subject. Thermodynamics isn't very intuitive for a couple of reasons. First, we humans exist in a very narrow temperature range, the spread between what "feels very hot" and what "feels very cold" is tiny compared to the range from cryogenics to combustion. Second, while we can feel quality, we have no sensory ability to feel quantity. Thus our language, our definitions, and our thought processes become confused. We use "heat" and "hot" as though they were forms of the same word, when they actually refer to very different phenomena. Improving our intuition of thermal energy, and thus our creativity, is the goal of this discussion.

Thermal energy is of interest because it can do something useful, such as making a Stirling engine run. This energy has two dimensions, quantity and quality. Heat and Hot. They exist together, but each has distinct properties. Neither heat nor hot can accomplish anything alone. It takes both, just as voltage and current are both necessary to deliver electrical power. Quantity, heat, is a linear concept. A hundred calories will accomplish twice as much as 50. Quality, hot, is not linear at all. This may be the biggest block to grasping thermodynamics at an intuitive level.

Hot has traditionally been measured in degrees. We call it temperature. But why do we use linear units to describe a phenomenon that is nonlinear for our purposes? Let's break out of that tradition, and see where it leads.

Degrees in thermodynamics are analogous to volts in electronics. Yet the linear volt is a cumbersome way to describe gain in an amplifier or loss in a cable. Ratios are involved here, not absolute units, and for this purpose the decibel is better suited. Without the decibel, it's fair to say that electronics could never have made the huge strides we've seen in recent history. It's time to grant the same benefits to thermodynamics.

In FIGURE #4, all the arrows are the same length, that is, each represents an absolute temperature ratio of 2:1. But they don't look the same length, do they? Expressing temperature in degrees is the source of severe distortion that makes some arrows appear much longer than others. This same effect exists mentally, distorting our intuition.

FIGURE #5 shows similar arrows, now with the degrees distorted. Again, each arrow illustrates a 2:1 ratio of absolute temperature. This is a first step, but doesn't do much for human intuition.

FIGURE #6 adds decibels of temperature. Each arrow is 6 dB long. Now it's obvious that 6 dB is a 2:1 absolute temperature ratio, just as 6 dB is a 2:1 voltage ratio in electronics. This temperature relationship is equally true at room ambient, or the cryogenic regions, or at the surface of the sun. It's a factor that can be computed mentally in a flash, something most of us can't do with degrees Celsius. The 0 dB point has been arbitrarily set at 0 degrees C, thus dBT is defined as 20 log T1/T2, with T1 the temperature of interest, and T2 equal to 0 degrees C, both expressed in absolute degrees.

Now we've created a better bridge between thermodynamics and the human mind. The statement that a given engine would run with a 10 degree differential is meaningless unless a reference temperature is cited, but an engine that will run on 0.1 dB will do it anywhere within the limitations of it's material and environment.

A chart for quick conversion from temperature to dBT is shown in FIGURE #7. Plotted to make Celsius a straight line, it provides an interesting and perhaps surprising illustration of the nonlinearity of our common temperature scales.

The importance of dBT is hard to overstate. Each dB is exactly as useful and important and valuable as any other. Now it becomes easier to see how much capability exists in the spread between T(hot) and T(ambient), and to see where it may be going astray. Question: Is it a better design compromise to accept a 200 degree loss in a heater, or a 40 degree loss in a cooler? There is just no way to answer that without referring to the specific temperatures and doing the math. But in the dBT world, a loss in one place of 0.4 dB can be directly compared to a loss somewhere else of, say, 1.3 dB. Intuitive.

Carnot faced the same questions of temperature ratio long ago. Using the decibel just gives us another viewpoint of that ratio, hopefully providing a tool more suited to human intuition. But temperature is only a single dimension of the problem, we must look further.

Heat and hot. Heat (calories) is linear, and by using dBT we can express "hot" in a linear manner. Again, it's necessary to remember that they are as distinct as current is from voltage. And just as we can multiply electrical voltage and current to find power, we can multiply heat and hot to find an "area" of energy. This simply wasn't practical with temperature measured in degrees. Each square in FIGURE #8 is exactly as valuable as every other square. Now we can see the relationship graphically between x axis quantity (heat) and y axis quality (temperature). To produce an area both the x and y axes are needed, neither heat nor hot can do it alone. They can, however, be traded one for the other. That is, a given number of area units will do the same job whether that area is composed of many calories at a small temperature spread, or fewer calories with a greater delta T. The "area" can be square, or tall, or short and squat and if it has the same number of squares, it can do the same amount of work.

For an example, suppose we are designing a Stirling engine to operate at an ambient of 33 degrees C (91 F, or 1 dBT). This is shown as point A in FIGURE #8. The fuel we have chosen has a capability, under ideal conditions, of burning at 2450 C (20 dBT), point B. We plan on burning at a fuel flow rate equivalent to point C. Note that the x axis can be scaled to any units of heat and time that may be convenient, such as 10,000 calories per second, or whatever. Thus, A-B-C-D defines an area of 19 dBT by 15 quantity units, or 285 squares. And remember that each of these 285 squares contain precisely the same amount of capability or usefulness.

Now let's examine where this energy is going. See FIGURE #9.

Loss #1 in this hypothetical design represents fuel that passes out the exhaust without burning. Perhaps due to a bad atomizer design, poor mixture control, or other mechanical flaw. This is fuel purchased and consumed without yielding any benefit. In this example, it represents 10% of total fuel, or 28.5 squares.

The second loss involves the difference between the temperature this fuel could have produced under ideal conditions, and the temperature of combustion actually realized. Perhaps fuel/air mixture is to blame here too. 37.8 squares is 13.2% lost.

Conduction and other thermal losses involving energy that does not flow through the Stirling cycle are covered in loss #3. These might include loss due to simple metallic heat conduction, radiant losses, exhaust temperature rise above ambient, et cetera. Here we're seeing 40.5 squares, or 14.2%.

We've already lost over 37%, about 62% is still available to feed into the engine. Loss #4 is the difference between the combustion temperature (17.2 dBT) we managed to achieve above, and the 14.6 dBT at the inner surface of the heater. This stems from thermal conduction loss within the metallic heater, but also from inefficient transfer from the burner gasses to the heater structure. Here we see a loss of 2.6 dB, which represents 28.6 squares or another 10%.

A similar loss exists in the cold side, between ambient temperature and the actual temp we manage to achieve at the inner surface of the cooler, this is loss #5. Chalk up another 30.8 squares representing 10.8% of the 285 we purchased from the fuel supplier.

Now, finally, we're into the engine itself! Loss #6 is the difference between the heater surface and the mean temperature of the working fluid, taken at the hottest point in the cycle. To find this mean, we must account for all the gas in the system including any dead volume, not just the gas residing in the hot chamber. This points out the strong need to minimize dead volume since any gas not exposed to the heater won't be heated. Loss #6 represents 48.4 squares, or 17%.

Loss #7 is the mirror image of #6, taken at the coolest time in the cycle. Again, dead volume or incompletely-swept gasses won't be cooled, and will contribute to the difference between mean gas temperature and the cooler surface itself. Tally another 26.4 squares, amounting to 9.3%.

Of the 285 squares of capability, we've managed to lose 241 of them along the way.

The remaining 44 (15.4%) are available to run the engine. They represent the thermal energy that produces the actual rise and fall in pressure that makes things go around. They must supply energy to overcome all the internal aerodynamic drag, all the friction in bearings and other materials, and if we're lucky leave a little to come out the shaft as rotary power. This message bears emphasis. The last thing is the output power, it only exists as the residual after all the losses are satisfied.

The above is not discouragement. To the contrary, it shows the extreme promise of the Stirling once we identify and minimize each individual loss. Now we have a way to visualize the relative value of one design choice over another, taking into account the relationship of heat and hot.

Of course the numbers above are hypothetical for the purpose of illustration. A particular real-world engine will have different numbers in every department. But the losses are real, and the goal in each design must be to minimize all losses and thus maximize the remainder available as output power.

No attempt has been made to include all the possible loss mechanisms. It may be helpful to break the above categories down further. The intention here is only to demonstrate a way to see how much capability the fuel contains, and where it goes.

Finally, let us look at one improvement in the above Stirling design. Again refer to FIGURE #9. Mentally make just one change. Drop the ambient temperature by 2 dBT, to around -30 degrees C. This moves losses #1, 3, 5, and 7 down 2 dB, and increases the gross engine power from 44 to 66 units, a 50% improvement. (I'm ignoring a number of lesser factors here for simplicity.) This 50% improvement in gross power, that is, power prior to supplying the internal aerodynamic and friction losses, might equate closer to 100% improvement in net shaft horsepower. But where would we find such a cold ambient to operate this engine? Answer: The higher you go, the colder it gets. We'll use this engine to power an aircraft.

And that takes us to the subject next time, the amazing match between the Stirling and the lightplane. Stay tuned.

Part Three.....Stirling Engine - The Perfect Aircraft Powerplant

Designers of light aircraft are searching for a better powerplant, something that could revolutionize flight. Designers of Stirlings are searching for a better marketplace, a field not already saturated with acceptable engine technology. Could it be that these groups are looking for each other? The purpose of this discussion is to compare the abilities and requirements of these two fields and to illustrate how perfectly they match.

What is a lightplane? Is there a market here? What engines are needed? Is it possible that the Stirling can compete? Questions and more questions. So let's dive right in.


We're not talking about airliners or military aircraft. Light aircraft usually have 2 or 4 seats and one engine, are mostly flown for pleasure by their owners, and frequently are based at outlying airports. They are factory-built and usually fly at 100 - 200 miles per hour. In the entire United States there are about 180,000 lightplanes. While this is a modest number, it is much larger than the 7000 or so airliners.

A subset of lightplanes are owner-built craft. In past decades these were usually one-of-a-kind airplanes, and the Federal Aviation Administration established an "experimental" category for them. Today, most are constructed from kits. This is the fastest growing market in aviation. About 14,000 owner-built airplanes are licensed and flying today, with thousands more under construction. Some are formed from aluminum, others use even older technology, but the majority are made of advanced composites including fiberglass, kevlar, graphite, and the newer ceramic fibers. With exceedingly low drag coefficients and high strength to weight ratios, these "experimentals" achieve performance far beyond any store-bought aircraft. A few models are pressurized and operate near 20,000 ft. There are no designs capable of operation at 40,000 because there is no suitable powerplant. Yet.

And yet these fantastic new craft run old engines. Aircraft powerplant designs can be traced directly back to the 1930s, in some cases the parts are still interchangeable!

Experimental aircraft enjoy relative freedom in selection of a powerplant. While most use certified airplane engines, the quest for better propulsion has included aircraft powered by car engines of every sort, motorcycle and outboard marine engines, as well as original designs and powerplants from snowmobiles and chainsaws. Diesels, Wankels, you name it. And the search continues.


Unlike automobiles or lawnmowers, personal aviation is a niche market. Unit quantities are very low and unit prices are very high, exactly what the Stirling producer needs today. A used, overhauled aircraft engine frequently costs $10,000 and up! And it is operated by a pilot who is trained to monitor and manage many operating parameters. Certainly a very different market from the automobile.

There are few fields where such intimate cooperation could exist between the engine developer and the customer, and this fact alone makes the experimental aircraft an attractive market for the small manufacturer.

Since we are talking about marketing, it is well to remember the old admonition that you can buy in any language, but only sell in the language of the customer. For that reason, I'm listing speed in knots, engine output in horsepower, and so on. The average pilot is not likely to be impressed by a 100 KW engine, but 134 HP is another matter altogether!


In a word, everything. Besides the ancient technology, the internal combustion engine is poorly suited to the needs of the aircraft. Noise and vibration are difficult to control in lightweight structures. Pulsating torque places severe constraints on variable-pitch propeller design. The need to keep propeller tips sub-sonic dictates engines with large displacement and relatively low speeds, not conducive to efficient operation. And slow, high displacement engines are heavy.

To the non-flying public, noise is a major concern. People perceive noise as danger, resulting in increased levels of governmental control. Aviation needs a quiet powerplant so it can co-exist peacefully within society. Even if the Stirling held no operational advantages, it would be worthwhile for this reason alone.

Next, there is the fuel. Most planes burn 100 octane gasoline, a specialty product that is priced accordingly. Availability is decreasing, in another decade the only fuel available at most airports will be turbine fuel. This is essentially kerosene, not usable in piston engines. Plus, the combination of a highly explosive fuel and a lightweight (thus fragile) containment leads to fatalities in accidents where occupants might otherwise walk away unharmed.

Besides the concerns of safety and availability, fuel economy is a critical aspect of aircraft design. Every pound of fuel requires a couple of pounds of structure to lift and carry it, which requires a bigger engine, which calls for still more fuel. A vicious circle. Every pound of fuel is a lost pound of payload. Obviously, fuel consumption in an aircraft has consequences that reach far beyond the monetary cost of the gasoline itself.

And the engines are fuel-cooled! Aircraft powerplants are rated for continuous operation at 75% of maximum output. During takeoff and climbout, the engine is operated at 100% power by running an overrich mixture to provide fuel cooling to the cylinders. This results in very poor fuel specifics, and adds unburned hydrocarbons to the atmosphere.

There are a dozen other good reasons why the present lightplane powerplant needs to be scrapped, but the point is made. Aviation needs a quiet, smooth torque, multifuel, efficient engine. Isn't that the Stirling?


As altitude increases, air density and temperature decrease. FIGURE #10 shows, for a typical experimental aircraft, the true air speed that would be realized if engine power could be held constant. This increase in speed is wholly due to decreased airframe drag at altitude. Between sea level and 40,000 ft, airspeed would double if the engine would only provide constant power. But the Otto cycle can't.

FIGURE #11 indicates typical aircraft engine performance. In the non-turbocharged engine, power decreases at any altitude above sea level, and by the mid-teens has declined unacceptably. As thinner air is reached that would otherwise allow the plane to fly faster, the engine says no. The Stirling, on the other hand, will increase power at altitude.

In FIGURE #12, the combination of increased speed due to drag reduction and decreased speed due to power reduction is shown. Two production aircraft, one turbocharged and one not, are depicted. Owner-built planes of similar power typically cruise about 50% faster.

In naturally-aspirated aircraft, maximum speed is usually found somewhere near 8000 ft. since this is where the 75% power curve meets the full throttle curve. Turbocharged engines in light aircraft typically do not gain any power rating from the charging. Instead, turbocharging (more correctly termed "turbo-normalizing") is used to achieve sea level performance to a specified altitude. Above that point both suffer equally. And turbocharged engines are very expensive to purchase and maintain.

With output determined by ambient temperature rather than volumetric considerations, the Stirling will increase power at higher altitude. Thus the effects of airframe drag and engine power will work together rather than in opposition, and the aircraft depicted in figure 1 will fly even faster as shown by the dotted line. There are operational reasons why aircraft must fly slower at low altitude. These include turbulence, traffic, and legal restrictions. But once above these constraints, a Stirling powered plane will exhibit performance undreamed of today.

There is a rule in aviation which dates back to the beginnings of aviation. "Every advance in aircraft design is preceded by an advance in powerplant design." The Stirling can provide that next great step forward. We have milked all the Internal Combustion engine has to offer. We have inflicted noise on society long enough. It is time for a new powerplant.


Unlike automobiles, aircraft spend most of their time at a constant power setting. This greatly simplifies the matter of control. In most cases power changes are gradual enough to be accomplished solely by burner control. There are a few instances where a sudden increase of power is potentially needed, such as during the approach to landing when gusty wind or unexpected traffic might require application of power. During this phase of flight, it is appropriate to run the burner at full temperature and regulate power via a simple shunt or bypass. While this technique is wasteful of fuel, it only occurs during a small part of the trip and therefore adds little to overall consumption. This method allows mechanical simplicity and virtually instantaneous response.


The largest supplier of aircraft engines is Textron Lycoming, which lists 44 models in their current catalog. They range from the O-235 at 108 HP and 213 lbs, to the IO-720 at 400 HP and 568 lbs. Other manufacturers have similar offerings. These engines are air cooled, 4 or 6 cylinder opposed designs with direct drive to the propeller, and are fuel injected except for a few of the smaller models. Rated speed is between 2500 and 2800 RPM.

Dry engine weight is about 2 lbs/HP in the smaller units and only improves to 1.5 lbs/HP on a few of the largest models. These power figures are maximum takeoff power, continuous power is typically 75% of rating. Specific fuel consumption is usually about 0.5 to 0.7 lb/HP/hr.

Weight and balance in aircraft is critical. For a Stirling to directly replace an existing engine, physical shape and mass must closely match. However, the experimental aircraft builders have shown an amazing ability to adapt engines of all stripes to their needs. They will welcome the Stirling. In writing on this subject in the aviation press, the response I've received has been 100% positive. The recurring question is "Where can I buy one?"


YES! In the final installment of this series, a proposed engine will be presented in some detail. It embodies the ideas presented in part one, "Ten Tips". It is designed from scratch to be an aircraft powerplant. By stressing internal aerodynamics, keeping dead (non-swept) volume near zero, paying strict attention to delta T losses, and designing to the control needs of the aircraft, it is possible to build the powerplant aviation has been waiting for.

The Stirling cycle has been simulated enough. It's time to produce and sell engines.

Part Four.....Putting the Aircraft Stirling Together

In the first part of this series, "Ten Tips for Stirling Engine Design" (SMW March 1993), we discussed some basic concepts that are fundamental but sometimes overlooked. These include such things as minimizing dead space and using the proper expansion ratio, as well as watching aerodynamic and temperature losses.

In part two, "The Adventures of Heat and Hot" (SMW, June 1993), a potent tool was introduced, the decibel of temperature. The dBt gives the user a better way to look at the quality of thermal energy, and allows quality and quantity to be combined into one number. This thermodynamic tool allows the Stirling designer to see the nature of losses in a new light.

Part three, "Stirling, the Perfect Aircraft Engine" (SMW, Sept 1993), outlines the advantages offered by the Stirling over any other aircraft engine type. Ambient temperature decreases at altitude, leading the way to powerplants that increase output as the plane climbs. Nothing available provides that needed benefit.

Which brings us to this final part. (First printed in the March 1994 STIRLING MACHINE WORLD) Here we'll explore some ways to combine all of the above into an engine concept that has a chance of powering practical private aircraft. Of course there are many other applications for high power/weight Stirlings as well.

FIGURE #13 shows the classic Beta configuration Stirling. There is much dead space, all the volume of the heater, regenerator, cooler, and the connecting passages is non-swept volume which reduces efficiency. Since the mean temperature of all the gas at the hottest and coolest points of the cycle determine the mean pressure excursions, any gas that doesn't make the whole trip from hot space to cold space does not contribute to power output as it should. And in many engine designs there is no particle of gas that makes the whole trip.

Classic attempts to reduce dead volume hurt us in two ways. First, if passages are made smaller (for instance, if cooler tube diameter is reduced), then gas velocity must increase, leading to huge aerodynamic losses. Or, if the heater, regenerator, and/or cooler are simply made smaller, heat transfer is impaired and thermal losses are the result. We need to get rid of that excess volume without compromising heat transfer or creating additional drag.

FIGURE #14 is the same Beta configuration, but there are numerous changes. For starters, consider the regenerative displacer. Suppose you had a common hex-sided wooden pencil. Further suppose it was sharpened on both ends, with only a small (perhaps 1 cm) length of hex body remaining. Next, imagine this pencil with no lead, but rather a hole extending through the pencil where the lead should be. Finally, imagine hundreds of such pencils, joined into a disk shape with their hex sides mating in a honeycomb fashion, as shown in FIGURE #15.

This is the concept of our regenerative displacer. It is one piece, molded of ceramic, and modest in cost. The outer surfaces are "displacer", while the hole bores are the regenerative surface. No matrix is added in the regenerator bores, this would raise aerodynamic losses to an unacceptable level.

The cylinder head has become the heater. It's internal shape exactly matches the form of the displacer. This might be machined, or it might be a rather thick electroform. It has a surface area about 10 times the area of the disk itself, permitting reasonable heat transfer.

The power piston is likewise shaped with hundreds of matching "pencil points". In a real engine this might be thousands.

FIGURE #16 shows that when the displacer nests into the power piston, the volume of cold space (and also what was formerly cooler volume and connecting passageways) becomes virtually zero. Likewise, when the displacer nests into the heater, the hot space and heater volume are negligible, but their surface areas are still high. This achieves the objective of reducing dead volume, and permits almost all of the gas to be swept from the cold to the hot spaces.

Further, look at what happens when the displacer is in close contact with the heater. It's surface tends to achieve heater temperature, and when gas flow resumes both the heater surface and the displacer surface act to heat the fluid. The same thing happens on the cold side.

There are several problems with this design as it stands, an obvious one is the matter of removing heat from the cooler/piston. We'll get to that later.

Aerodynamic losses accumulate as approximately the third power of gas velocity, so it's imperative to keep fluid velocity low. At the same time, reasonable engine size requires that we keep speed high. How do we reconcile these contrary requirements? Notice that gas flow is just a matter of stroke, it is not influenced by cylinder bore. So we keep stroke very small, thus keeping gas velocity low, and we make bore as large as necessary to achieve the desired power.

There are engines today that have a regenerative displacer and a bore 10 or 20 times stroke. The beautiful low delta T designs by Jim Senft and Rob McConaghey come to mind. Working with a very low temperature difference, they can't give any efficiency away and still expect the engine to run. The question is: If we know how to achieve such good efficiency, why don't we apply it to the high delta T designs as well? I believe we should.

In each of the illustrations, stroke is greatly exaggerated for clarity.

Next, let's look at the need to mechanically move the regenerator. If we eliminate the drive linkage and use the Ringbom configuration we gain several benefits. Obviously the cost and weight and mechanical complexity of a phase-shifted drive are gone. That's good. And we no longer have the problem of displacer-drive lubricants leaking into the working gas. Thermodynamic benefits accrue as well. Ringbom displacers naturally vary phase according to operating conditions, providing some added efficiency. And the Ringbom tends to "over-stroke", that is, it tends to travel in a discontinuous motion with some time spent stationary at both extremes of travel. This allows the engine to come closer to the Schmidt cycle, where even higher efficiencies are possible. And finally, the extra time spent in close contact with the heater permits a greater fraction of energy to transfer from the heater to the surface of the displacer. The same effect is present at the other end of the stroke, where more energy can be transferred from the cold side of the displacer to the cooler/piston. In both cases this results in better energy transfer to and from the working gas.

FIGURE #17 shows the regenerative displacer attached to an annular Ringbom piston. The simplicity is apparent, and indeed a small demonstration engine should work if built as shown. But in higher-powered designs there is still the matter of removing waste heat from the piston.

With a large bore/stroke ratio, mechanical difficulties quickly arise. The forces may be light in the low delta T engines, but there are monstrous problems trying to apply the same techniques at high power levels. Crankshaft and connecting rod bearing loads quickly reach levels beyond any available technology. And with forces this high, mechanical strength would require massive components that would make the power/weight ratio completely unacceptable.

FIGURE #18 shows the next step, the introduction of hydraulics. Instead of making a power piston that has sufficient mechanical strength, what if we made a very thin electroform instead, and backed it up with a suitable oil? The piston becomes just a "skin" to keep the oil the proper shape! Since there is no pressure difference between the gas and the oil, this piston can be very thin. And since it is thin, it has negligible thermal gradient. Now the oil can be the same temperature as the cold space. Or more properly, the cold space can be brought to the temperature of the oil.

For our purposes the oil is incompressible, and there are many hydraulic techniques that can be employed. In this schematicized illustration, a common piston is driven by the pressure variations in the working gas, as transmitted through the hydraulic fluid. The ratio of diameters between the two pistons can be adjusted as desired to provide an acceptable force/stroke ratio to the connecting rod and crankshaft. I don't believe this version has much promise as shown, it's just a step along the evolution.

A new feature seen in FIGURE #18 is the bellows separating the oil from the working gas. The bellows are subject to the gas pressure variations and must be robust enough to withstand both positive and negative excursions from the ambient gas pressure found in the buffer space. This buffer space must be sufficiently large to maintain a reasonably steady pressure as the Ringbom annular piston strokes up and down, but since the buffer space is not a part of the Stirling cycle it can be as large as desired. In a non-pressurized design the buffer can be the open atmosphere. The design of the bellows is greatly simplified by the fact that the stroke is quite small compared to the diameter, much smaller than these exaggerated illustrations would suggest.

As shown, the working gas is hermetically sealed and is not subject to any contamination from the oil. The gas could be hydrogen or helium, and since it is sealed a charge could be expected to remain indefinitely. But in this illustration the oil can leak past the piston, not an acceptable state of affairs. We need to get rid of that piston.

And so we do! FIGURE #19 shows a practical engine. The crankshaft/connecting rod/piston arrangement has been replaced with a very simple eccentric rotor on the output shaft. The vanes are somewhat schematicized, there are more appropriate ways to seal the flow, but they serve our illustrative needs. There is only one seal, on the power shaft, and it is exposed only to oil.

This entire engine can be pressurized as desired, and with the good rotary shaft seals commonly available (similar in design to those used in automotive air conditioners) both the oil and the hermetically sealed gas charge can be expected to last for many years.

There is one problem remaining, extracting the low grade heat. FIGURE #19 takes care of that with a secondary cooling loop, rejecting the heat to the ambient environment. Again, there are no cyclic constraints on the volume of oil or size of the secondary exchanger, they are not subject to the limits we found in the original Beta configuration cooler. The hydraulic fluid is acting as primary coolant, lubricant, and force multiplier simultaneously.

Note that the two Ringbom displacers could be mechanically connected if this proves to be advantageous.

It appears that a two cylinder design is the minimum practical arrangement. FIGURE #20 and FIGURE #21 depict three and four cylinder designs, respectively. Obviously more cylinders are possible, and with appropriate arrangement it's possible to produce an engine with virtually 100% static and dynamic balancing.

By utilizing the hydraulic drive, there is no limit to the shapes and form these engines can take. In an aircraft application, for example, it's worthwhile to put the output shaft as high as possible. This permits the maximum diameter propeller while allowing sufficient ground clearance. The bigger the prop, the slower it can turn, which yields benefits in both efficiency and low noise. A design can be visualized where the horizontal shaft is at the very top, and cylinders or pairs of cylinders extend down. This would be somewhat similar in basic shape to the old Ranger engine, for example. In other applications, the shaft can exit the engine in any location desired.

There is one more factor needed for a practical aircraft powerplant. The matter of instantaneous power control. And that is a part I'm not giving away! It is simple, light weight, inexpensive, and potentially reliable. If you are interested, contact me.

Most of the techniques described have been incorporated in some previous design or other. It is not my intention to claim them, only to present them in a combination that offers advantages not yet realized. To the extent that any concepts or devices are novel, they are hereby placed in the public domain.

The Stirling engine community needs markets for it's creations. Most markets, such as automobiles and lawnmowers and the like, are now served by reliable engines produced in such quantities that competition is very difficult. But the personal aircraft engine is unique. It is a small market, and engine costs are absurdly high. Plus the existing engines do not serve the needs of the aircraft. The purpose of this series has been to bring together the engine producers and the marketplace.

I welcome inquiries from anyone in either the Stirling or the aviation communities. Contact me at this address:

Darryl Phillips
c/o AirSport Corporation
1100 West Cherokee
Sallisaw OK 74955 USA
Phone: (800) 343-6690
Fax: (918) 775-4000
Email: darryl@airsport-corp.com

© 1994 Darryl Phillips

Back to AirSport Home Page