Converting a Subaru EJ-22
Awesome power for my Q2By Jon Finley, EAA 394580
The propeller attachment is a two-piece affair. On the right is the adapter that bolts to the flywheel end of the stock crankshaft. On the left is the prop hub (which also acts as an extension) that bolts to the adapter once it’s securely installed.
Cool air enters the plenum (through the belly scoop) where it is slowed and expanded before meeting up with the VW Rabbit radiator. After leaving the radiator, it’s compressed and sped up to re-enter the airstream through the rear exit.
This is a side view of the custom fiberglass plenum (designed and built by Jon) shown in the adjacent photo, prior to being installed. It is, of course, shown without the radiator installed.
The Q2 has a removable tail cone, making the retrofit of the cooling plenum much easier than it could be with most designs. This photo is a view of what lives behind the seats of Jon’s Q and shows the plenum and radiator.
In addition to the exhaust header, this photo shows the coolant plumbing running to the underside of the fuselage, as well as more details of the engine mount.
A great view of the prop blades and the underside of the plane, showcasing the P-51 style belly scoop and the external coolant lines.
My engine is an EJ-22 from a 1990 Subaru Legacy. The only internal change to the engine was to install a Delta Camshaft (Tacoma, Washington) “220” grind with the goal of lowering the torque curve to increase torque at lower rpm. A friend that owns a towing business recovered the car from an accident and sold the whole thing to me for a song. It had approximately 160,000 miles, ran great, and had good compression.
My conversion details and photos are on my website, www.FinleyWeb.net. There are also a number of videos on the site so you can see and hear it in operation. In summary, I did the entire conversion myself including construction of intake and exhaust manifolds, motor mount, crankshaft adapter, prop hub, cooling system, all of the little brackets, and mounts. The firewall forward weight of the complete engine with every single accessory and part is 250 pounds. I generally find discussions about weight to be quite annoying; there is too much guessing, hearsay, and flat-out misinformation out there. It seems that the majority of the time the comparisons being made are apples to oranges. What I do know is that I have a 250-pound direct-drive installation that produces enough power at high-density altitudes to be safe and to satisfy me. It cost me less than $3,000—including the engine control unit (ECU)—it’s fuel efficient and inexpensive to maintain and repair.
I used the stock Subaru automotive electronic fuel injection (EFI) system for the initial 20 hours of flight. I was not thrilled with the inability to tune the system or its behavior at idle and high airspeed, so I switched to the Real World Solutions (RWS) EC-2 controller produced by Tracy Crook. (See CONTACT! Magazine issue No. 87 for more information.) The RWS EC-2 has proven to provide absolutely flawless performance, is incredibly flexible/tunable, and is easy to operate. Since I already had an EFI system installed, adding this new controller was quite easy and only required that I change the spark plug coils and some wiring. I cannot say enough positive things about RWS and the EC-2 product, as well as Tracy Crook and his wife, Laura.
One area where I have done things differently than most is with the fuel pressure regulator. I honestly do not know what the mean time between failures (MTBF) is for regulators, but it stood out to me as a potential single point, so I installed two of them in parallel. The first unit is the stock device in the stock location (about 36 psi). The second unit is an aftermarket adjustable regulator, installed just downstream of the first and set to a slightly higher pressure.
My prop hub is a three-piece unit. The adapter bolts to the crankshaft, and the prop hub/extension is bolted to this piece. A modified ring gear from a 1991 Toyota Camry with a V-6 and automatic transmission is captured in between these assemblies and is driven by a Dodge Colt starter engine. All three custom pieces were machined from 2024 aluminum, and “that looks about right” (TLAR) design standards were applied. Of course, this means that they are probably overbuilt and heavier than necessary.
The engine mount is less than traditional and mimics an approach that I spotted once on a Subaru powered VariEze. It is made from 1/8-inch 4130 steel plate that has been cut and bent as required. Engine mount rubber “cones” were machined and welded to the mount. These cones use standard O-200 engine mount isolators. The mount bolts directly to the cylinder heads using the stock head bolts. This is a real pain when it comes to maintenance, but otherwise works quite well. Again, TLAR design standards were closely adhered to.
Proper liquid cooling system design for aircraft is now pretty well understood and no longer the art form it was just a few years ago, thanks in part to the Alternative Engines collection of books written by Mick Myal. However, implementing a good design in a Q2 is not the easiest task due to the layout of the airframe and various components. I elected to install the radiator in the tail cone with a P-51 style air plenum. Being a cheapskate, I used an inexpensive VW Rabbit radiator rather than a custom unit. This necessitated a bit more complex design than would have been required with a custom radiator. However, what I have works quite well, so I am content.
I have the coolant running in 1-inch diameter aluminum tubes on the belly of the aircraft. Initially, I did not want hot coolant in the cockpit. I am now comfortable with that (assuming proper installation) but have very limited space to work with, so they will probably stay on the belly. I have never experienced any cooling problems on the ground or in flight. On the ground I can run the airplane for a very long time at very high power settings without fear of overheating; my longest ground test was 25 minutes at full throttle. The only issues have been the introduction of engine exhaust gases finding their way into the cockpit and hot air escaping from the plenum, causing the cockpit to become quite warm. I have both issues sorted out now.
The initial exhaust system was nothing more than 1.5-inch diameter aluminized exhaust tubing welded to the stock exhaust manifold and pointed down and aft (cantilevered).
After about 50 hours, one of those welds cracked, so it was time for a better plan. I still use the stock exhaust manifold, but that is now welded to a flexible automotive exhaust coupling that then runs to aluminized tubing exiting out the bottom. A bracket to the engine now supports the whole setup. The couplings provide some flexibility and are holding up well after about 50 hours.
I am running a Warp Drive three-blade, ground-adjustable prop. I had Gary Hunter (of Pushy Galore Fame) modify the airfoil on the outer third of each blade for high-rpm use. I get 3,500 rpm on the initial takeoff roll and climb, get 3,800-4,100 during cruise, and have seen as much as 4,400 rpm when really pushing hard. The faster I run the engine, the smoother it is.
After a fair amount of prop design research (as well as trial and error) I’m convinced that this blade profile and blunt-tip design is not ideal for high-speed efficiency, and I am working on a couple of other ideas for a more efficient solution. Until then, the Warp Drive works.
Initially, the “old” pilot in me really didn’t want to run the engine over about 3,500 rpm, but time has solved that. I am certain that a direct-drive auto engine conversion, like mine, would benefit from an in-flight adjustable prop. However, that is out of my budget and Santa hasn’t dropped one off yet. So, I wait….
I live at Mid Valley Airpark (E98) in Los Lunas, New Mexico, which is in the high desert, with a field elevation of 4,830 feet and a 4,300-foot by 37-foot asphalt runway. At this location, the airplane performs fairly well even on warm summer days. I typically see about 600-700 fpm climb immediately after takeoff with full fuel (and an empty right seat) in the cooler part of the day. I recently made a trip to 13,000 feet mean sea level (MSL) and found that I still had a 700 fpm climb at 10,000 feet and a 600 fpm from 12,500 to 13,000. A recent trip to Kanab, Utah (365 statute miles each way, nonstop), showed fuel burn to be 4.4 gph while leaned to about a 100 rpm drop (I don’t have an exhaust gas temperature gauge). During this trip I was at 10,500 feet to 12,500 feet MSL and full throttle (between 3,900 and 4,050 rpm).
PERFORMANCE AND ECONOMY
How much different is it to operate than a traditional aircraft engine? Quite a bit as it is basically a single-lever operation (throttle). The EFI system automatically controls mixture based on my settings, which are based on inlet air temperature, rpm, and manifold pressure. The RWS EC-2 unit also controls timing. There is no carburetor heat or cowl flap. Pretty simple— push to go up, pull to go down.
I flew 1,472 statute miles round-trip from home (Mid Valley Airpark, E98) to Beatrice, Nebraska (BIE) for the Quickie/Dragonfly Tandem Wing Fly-in. Total time for the trip was 12.3 hours. The outbound trip with some local flying while there put seven hours on the Hobbs while burning 33 gallons in the process. This comes out to 4.7 gph while performing eight takeoffs and landings, including giving a couple of rides and a little “demonstration” flying. The return trip was five hours, and total fuel used was 29 gallons.
During the first return trip leg I did some testing and did not lean at all and was showing about stoichiometricon the oxygen sensor/gauge—fuel burn was just a hair over 6.6 gph. On the second leg, I went to 10,500 feet and leaned as normal. Fuel burn was 4.8 gph. These fuel burn numbers are a bit higher than I previously thought (4.4 gph at 10,000 feet), and I think this is due to a combination of running at lower altitudes over the Midwest and a very small fuel leak that I discovered after returning home. No oil used/lost, no coolant lost, and zero maintenance anywhere along the trip.
So, what’s next? At the current time I really enjoy just flying. I do have a very aerodynamically dirty Q2 and have lots of cleanup planned—the problem seems to be actually getting to any of it. I have found this to be a real significant problem over the course of the last years as I am doing so much flying that I hate to take the airplane down for aerodynamic optimization.
One engine option that looks interesting is an EJ-25 block paired to the EJ-22 heads, which yields more displacement (torque) for basically no change in weight.
As an aside, a gentlemen in the local area flies a Subaru EA-81 powered Pietenpol. He has 1,300 hours on his engine, and those hours are after pulling it from a car with 160,000 miles (no rebuild).
Jon Finley has been involved in experimental aviation since 1988. He purchased a partially built Quickie (Q1) while living in Montana and finished that project with a Cuyuna engine. It flew twice before being converted to the more powerful Rotax 503 power. After flying for 250 hours behind the Rotax, Jon made another change. A four-cylinder 1835 cc VW was installed and transformed the Quickie into a heavy but fantastic performer. After moving to Minneapolis in 1995, Jon ceased flying the VW-Q1 because in 1997, a hangarmate made Jon an offer (that couldn’t be turned down) to sell him his flying Q2. Jon gladly bought it and flew behind the original 2100 cc Revmaster (VW) for a couple of years before it was time for a change. A carbureted direct-drive turbo Subaru EA-81 was installed and flown for a short time. With this experience in hand, Jon decided that he really wanted the benefit of electronic fuel injection with a turbocharger. The work to adapt these changes to the EA-81 was more than switching to a normally aspirated Subaru EJ-22 Legacy engine so that direction was taken. However, while Jon was still in Minneapolis, the conversion took a back seat to some of life’s common interruptions. For the past two years, Jon and family have lived at an airpark south of Albuquerque, New Mexico, and the Subaru/Q2 progress has been revived. Jon has flown 100 hours in the past year with more than 120 hours total on the conversion (600 hours total on the Q2) and is now feeling that the bugs have been worked out. Jon will be the first to admit that he is a scrounger and do-it-yourself kind of person.
Jon will be a guest speaker at the Alternative Engine Round-Up (fly-in) at the Jean Sport Aviation Center, Jean, Nevada, on March 28, 2009. For more information visit www.ContactMagazine.com/roundup.html