The purpose of this article is to set straight some common misconceptions many people have about automotive engines as applied to aircraft propulsion and answer some equally common questions about package performance and weight. This article is based upon my experience as a professional performance/ race engine builder for over 25 years, my consulting experience in the field of automotive engines in aircraft for more than 15 years and my experience building and flying a Van's RV6A with turbo Subaru power. I'm currently constructing an RV10 with turbo Subaru power again.
Automotive Engines Will Not Take Continuous High Rpm Use
This is the most common misconception put forth by lay, anti-auto power people and is utter nonsense. They often go on to say that auto engines were designed for low rpm operation and 15-30 hp as required to cruise a car at 70 mph. This simplistic, flawed reasoning is completely unsupported by facts. When asked to supply facts to support their contention on various forums since 2003, not one person has ever done so.
Let's examine the facts at hand which completely contradict this notion:
1. Auto engines today are designed and routinely tested to higher standards than certified aircraft engine requirements. The FAA only requires 100 hours of full throttle, full rpm for certified engines and another 50 hours at 75-100% power, 50 hours of which are required to be at redline oil and cylinder head temperatures. Most auto engine manufacturers today do a minimum validation of 200 hours of WOT at rated hp rpm and some as much as 1200 hours. In addition to this test, they perform cold weather testing to the tune of 1000+ cycles of cold soaking the engine to 0F and immediately taking the engine to WOT and high rpm until coolant reaches 240F. While the engine is still hot, 0F coolant is pumped into the engine until the block achieves 0F and the test is repeated- over 1000 times. Additional tests often include idle testing to 2000 hours with oil temperatures of 260F+ and transmission validation where the engine is cycled from low rpm to shift point rpm at WOT while the transmission is shifted up and down for up to 1600 hours. Not just one engine is put through these tests- dozens are. Wear rates are noted and obviously failures are not acceptable before release of the design.
2. The EJ series Subaru has proven itself capable in the real world of sustained high rpm/ WOT operation with the 1989 world speed record of 100,000 km by three Legacy RS turbos over 447 hours of 138.78 mph and an estimated 250,000 flight hours by over 1000 Subaru EJ powered aircraft including gyroplanes from RAF and Groen, fixed wing conversions such as those from Eggenfellner Aircraft and hundreds of other private conversion worldwide. One high time gyro operator in Australia reported 3800 hours on an EJ engine without overhaul!
3. The shorter stroke of the EJ22/ EG33 Subaru engines gives these engines the same piston speed at 4000 rpm as a Lycoming O-360 has at 2700 rpm. Much tighter, ring, piston and bearing clearances on the Subaru combined with more stable dimensional control through liquid cooling and superior lubrication through synthetic oils result is much lower wear rates compared to air cooled aircraft engines. My EG33 with an estimated 2100 hours on it showed NO measurable wear on the pistons, valve stems, guides, cylinders, camshafts, bearings and crankshaft journals. One high mileage EJ25 in Scandinavia accumulated an estimated 11,000 hours until a head gasket had to be replaced! Liquid cooled engines keep the aluminum alloy parts well below the typical 350-400F levels that aircraft engines often run at. Aluminum alloys lose about half their tensile strength at 400F which causes relatively frequent cracking in these parts. The EJ20/22/25 engines have 5 main bearings compared to only 3 for a Lycoming four cylinder. The short stroke also gives rod journal overlap to the main journals of the crankshaft. These features combined with an integrated closed block/ cylinder design and head bolts threaded into the main bearing saddle rather than the base bolted design of the Lycoming, make the EJ engines far stiffer than the Lycoming design. EJ/EG engines have a forged, heat treated crankshaft and forged rods from the factory and have had far better crankshaft reliability than Lycoming engines.
4. Specific output is the measure of hp produced per unit displacement of an engine. It is a measure of relative strength and efficiency of an engine design. One 2.5L turbocharged Subaru drag engine has been dynoed at 960hp and completed over 60 passes at this hp level. This equates to 384hp/ liter and all with stock crankshaft and block. This would be 2300hp on a Lycoming! How long would a Lycoming O-360 last at even half this power level? Probably less than 10 seconds. The Subaru has gone over 600 seconds at this power level. This proves the tremendous strength of the Subaru boxer design compared to aircraft engines. A detuned version of the 2.5liter STI engine operating at 200-250 hp and 4500 rpm is obviously nowhere near any critical strength limits.
5. In Europe, where speed limits are higher (or none), cars are often cruised at very high speeds for a good portion of their lives. This demands high rpm, high hp operation. We just don't see a high percentage of smoking wrecks on the side of the Autobahn or AutoStrada. Engines today must be designed to withstand this kind of use or there would be a lot of warranty claims. Many showroom stock race cars also prove every day how robust OE engines are, spending their lives between 5000 rpm and the rev limiter for many, many hours.
Weight
Are auto engines competitive on a hp/ weight basis with popular air cooled aircraft engines? The short answer is maybe. Naturally aspirated auto engines with PSRUs and radiators from firewall forward suppliers have generally not able to achieve the power to weight ratios of a typical Lycoming engine to date. Turbocharged auto engines can be more competitive. Often hindering auto conversions has been the mismatching of PSRU ratios which has not allowed these engines to achieve power peak rpm. This is critically important with naturally aspirated engines. Detailed analysis in weights though shows that it is quite possible to be close in weight vs. hp using modern Subaru EJ/ EG designs.
If we take a modern EJ25 rated at 175hp and add the basic engine weight to the radiator, coolant, heavier engine mount, backup battery and typical PSRU, we get right around 300 lbs. This is about the same as a Lycoming O-320 with carb, mags, baffles, fuel pump and ignition harness. If we derate the EJ25 to 160hp by reducing takeoff rpm to about 5200 rpm, we can get theoretically similar performance to an O-320 Lycoming. These weights do not include alternator, starter or exhaust system for either engine which we assume are close to the same.
Comparing the modern 2.5 Legacy turbo engine to a 200hp IO-360 Lycoming we get weights of around 330-340 lbs. for the Lycoming and something very similar for the Subaru with turbocharger, intercooler and ducting. The turbo only needs to turn about 4400 rpm to achieve 200hp. The Subaru can of course maintain this hp level to the turbocharger's critical altitude. We can also spin the Subaru down to 3500 rpm in cruise for 150-160 hp to get better fuel specifics than a naturally aspirated Subaru. Where the 6 cylinder Subarus are used to replace 180 and 200hp Lycomings, they are somewhat heavier to the tune of 40-80 lbs. depending on models compared.
Comparing higher hp engine like the 230hp EG33 six cylinder or later EZ30R to the 235-260hp Lycoming 540 series, we again see competitive possibilities. In fact, both Subarus will have lower installed weights than the big Lycoming.
It should be stressed than auto engine hp can be lower than factory ratings on naturally aspirated versions employing variable induction tract tuning and variable valve timing and lift if these features have been disabled or modified. Careful modification of the induction and camshaft systems should carried out to negate these possible effects and ensure full rated hp potential.
Engines like the Chevrolet LS-1, LS-2 and LS-6 are quite a bit heavier than an IO-540, to the tune of 50-100 lbs. depending on PSRU and radiators. These engines are easily capable of producing power far in excess of what the Lycoming can but C of G concerns for a given airframe design will take careful consideration.
Cooling and Drag
Air cooled advocates often like to say that liquid cooled engines exhibit higher cooling drag than air cooled installations. This is again another "feeling" they have with no facts to back up the statement. The simplistic reasoning usually offered is that because air cooled engines operate at higher temperatures, Delta T is higher so less mass flow is required. What they fail to consider is the mechanism of heat transfer and the pressure (energy) loss is very different between air cooled cylinders and coolant radiators. A modern aluminum radiator is many times more efficient at heat transfer than aicraft cooling fins per unit area or unit volume. Radiators also can be placed in a very efficient divergent/ convergent duct to maximize heat transfer, minimize core drag and exit the air near free stream velocity for minimal drag with no turning of the cooling air. Air cooled fins by comparison on opposed engines, force the air to turn sharply over 180 degrees past many obstructions to flow, creating turbulence and drag.
My study of several identical airframes fitted with both air and liquid cooled engines clearly show the advantage of well designed liquid cooled installations:
Bristol Beaufighter. 1280hp Merlin 330 mph, 1590 hp Bristol Hercules 323 mph.
Hawker Tempest I. 2240hp Napier Sabre 466 mph. Tempest II 2520hp Bristol Centaurus 440 mph.
Reggiane RE 2001. 1175hp Alfa Romeo 337 mph. RE 2002. 1175hp Piaggio 329 mph.
DC-4/ Northstar. 353mph with 1760hp Merlins, 280 mph with 1450hp Pratt R2000s.
It must be said that many experimental aircraft have had relatively inefficient radiator setups and consequent cooling problems. Today, we understand what is required to get good cooling performance with a minimal drag penalty. Designs which continue to place rads in the cowling cheeks with short ducts and poor pressure recovery mixed with inefficient exit paths likely do have higher drag than a well designed air cooled installation. Dedicated ducts mounting proper radiators with oval tubes and proper fin density are required for best performance.
Propeller Speed Reduction Units (PSRUs)
PRSUs are required in most cases to allow the engine to spin up to its power peak rpm while allowing the prop to spin at its most efficient rpm. This is the other part of liquid cooled installations often attacked by air cooled advocates and sometimes rightly so. Many PSRU designs have had less than stellar longevity. This is generally the result of inadequate design and especially testing. Of considerable concern is torsional vibration (TV) problems causing failures of the gears, bearings and shafts of the gearbox and sometimes propeller. Few PSRUs have had proper TV analysis or instrumented testing done on them. Some later designs such as the SPG-2, EPI and Autoflight using better coupling devices between the crankshaft and gearset have shown better reliability. Lightweight flywheels often used have aggravated TV problems. There is no reason that PSRUs with proper design and validation should not be as reliable as modern manual transmissions on cars. Gearboxes have been used on hundreds of thousands of aero engines and are very reliable when properly designed. PSRUs do add weight to the installation. Most units weigh between 25 and 50 pounds for engines in the 150-300 hp range.
PSRUs have losses inherent in their gearsets or belts. Lay people have often speculated that these losses are up to 40 hp in the case of a 200 hp class drive. This is absurd as it would represent about 30,000 watts being dissipated as heat. If this was in fact true, the case or belts would melt in just a few minutes. Typical losses for single mesh spur and helical gears is around 2-2.5%. HTD belts run at 3-4%. Twin mesh helical gearsets would then have perhaps a 6% loss as worst case including bearing losses.
Cost
Costs vary widely depending on whether the package is DIY or from a commercial vendor. In the case of DIY installations, costs may be 1/6 to 1/2 of what a typical air cooled engine costs. This of course does not take into account the many hours spent designing and building the various components needed nor the testing and modification which invariably follows. Commercial packages generally are about the same cost as an air cooled engine. Where the savings come in is at overhaul time. The core auto engine is a fraction of the cost of the aircraft engine and rebuild parts are much cheaper. In some cases, a complete new engine is under $5000, low time used engines are under $1000 as are generally the parts and labor required to do an overhaul. This experience has been born out but many high time gyro operators using auto engine for training where total operating costs for fuel and overhaul reserve is under $25/hr.
Performance
The big question. Do auto engines perform comparably to traditional air cooled aircraft engines? In a word, no. To date, few if any have achieved the overall light weight, speed and fuel efficiency of a typical Lycoming. A handful of forced induction and Wankel powered aircraft have shown greater overall speed but always at a higher fuel flow and package weight. My view is that auto engines should be lightly turbocharged to allow lower rpms for lower frictional losses and better fuel flow while retaining power at altitude for higher cruise speeds. I believe that a properly executed turbocharged Subaru installation will exceed a comparable Lycoming installation for speed, weight and climb performance while burning close to comparable amounts of fuel. To date few if any people have combined all the best technology and tested it in flight. Our EJ22 turbo powered RV6A demonstrated roughly equal performance to an O-320 powered RV below 8000 feet at a weight penalty of around 50-60 lbs. and a fuel flow penalty of around 1 gallon/ hr. This was a first attempt and many improvements can be made to lower weight, drag and fuel flow from the lessons learned.
Package Reliability
This is the grey area. While the reliability of the popular Subaru engine core in proper repair has been shown to be as good or better than the Lycoming core, inflight reliability of auto engine conversons is dependent on several supporting systems like fuel, PSRU and electrical. Most power losses are attributed to problems with the supporting systems, not the core engine. This is where careful design decisions pay benefits and poor practices often cause engine shutdowns. Until the systems used are proven to be reliable through extensive flight testing, reliability on par with a typical Lycoming is often in question. Shared information on various forums and learning from both good and bad experiences will result in better designs which should improve overall reliability. Auto engines are not for everyone. Some people have had excellent results with them, others dismal. Some owners just want to fly while other like the experimental experience and difference an auto engine provides them. Auto engine conversions open up a second choice for those so inclined.
Good News
In the first part of 2009, Randy Crothers in Washington State completed a series of test flights in his 2.5 Subaru turbo powered RV7A. He was able to easily exceed Vne in level flight at 8000 feet, going 213 knots TAS while the engine also delivered lower fuel flows than the best Lycoming powered RVs achieve at the same speeds. Randy now has the fastest RV7A in the world by a large margin. Thankfully we no longer have to put up with the nonsense sprouted by Lycoming owners that auto engines cannot match the speed or fuel efficiency of their old tech engines.