EAA - Experimental Aircraft Association  

Infinite Menus, Copyright 2006, OpenCube Inc. All Rights Reserved.


Tools:   Bookmark and Share Font Size: default Font Size: medium Font Size: large

EAA Experimenter

[ Home | Subscribe | Issues | Articles | Q&A | How To | Forum Review ]
[ Hints for Homebuilders | Glossary | Polls | Around the Web | Submit an Article]

Electronic Fuel Injection for an O-200

By Justin Mace


Justin Mace and his Dragonfly have appeared in the pages of CONTACT! Magazine multiple times in the past. The first time was in issue No. 6, where he detailed his Subaru Legacy engine installation. In issue No. 14 he gave a 160-hour report. Now Justin reports on some significant changes to his plane.

I installed a Volkswagen conversion in my Viking Dragonfly about 20 years ago. After an in-flight crankshaft failure, I wanted a more reliable engine.

In the fall of 1989, Subaru had just introduced the 1990 Legacy model in the United States with its new engine design, the EJ22. There were folks busy converting the venerable EA81 and EA82 Subaru engines for use in their homebuilt aircraft, so I learned what I could about the “new” engine. This “new” one was a computer-controlled, multipoint fuel-injected, super whiz-bang engine. I just had to have one!

Lou Ross of Ross Aero fame [no longer in business] convinced me that he could make a prop speed reduction unit (PSRU) for the new engine, and that’s all it took. I started the conversion process in January of 1990. I found a wreck in Phoenix with 2,700 miles on it and purchased it for my project.

After 18 months of the conversion process (it took only 30 months to build the plane!), I flew behind the “new” engine.

The initial problem was cooling-related. The engine and PSRU ran fine and never offered any problems. I had thought much about the cooling requirements of the engine and had done quite a bit of research. My initial thoughts: If the engine was designed to operate at 190°F in the car, then it should be capable of operating at 190°F in my plane. If an auto engine normally operates at 20 percent to 30 percent power at freeway speeds and an aircraft engine normally operates at 70 percent or more, there will be a need for a much better cooling system in the plane. After a couple of changes to my original design, I was finally able to cool the engine here in the Arizona desert. It was great to be able to taxi out on a hot day, await clearance to take off, and not have to worry about overheating the engine due to a “lack” of cooling.

The biggest overall problem I had with the Subaru EJ22 engine was not related to the engine proper, but rather to the engine’s requirement for unleaded automotive fuel. I used the stock computer, in which its normal operating mode is “closed loop.” Closed loop means that the computer is taking all the information from the engine sensors and making minute mixture corrections continuously. One of these sensors is the oxygen (O2) sensor, which is normally located in the exhaust pipe. The O2 sensor detects oxygen in the exhaust, which tells the computer the exact condition of the combustion process, either rich or lean, based on the O2 content of the exhaust gas. The computer then makes corrections based on this rich or lean condition as well as a host of other conditions.

Since the stock O2 sensor uses platinum as part of its makeup, the tetraethyl lead found in 100LL aviation fuel will eventually render the O2 sensor readings inaccurate. This will have an effect on the proper operation of the engine. I didn’t want to take the chance that the computer running in closed loop would run the engine lean and cause long term damage. So I was limited to unleaded auto fuel. This was not too bad when flying locally; however, on those occasions when I ventured far from home, I found it increasingly difficult to find a convenient source for mogas. If I had it to do over again, I would have used a different computer and run it in “open loop,” thus eliminating  concerns about using 100LL.

After 480 hours of successful operation, I became disappointed with the mogas requirement and the excess weight of the water-cooled engine, so I decided to change to a certified air-cooled engine. I was very happy with the performance of the engine and in particular the computer-controlled, multipoint fuel injection (MPFI), so I decided that I wanted a computer to control the combustion process on my certified engine as well.

A little research convinced me that I could indeed run a new aftermarket MPFI computer system with dual electronic ignition in open loop and use 100LL. That started another engine quest.

Dragonfly engine

I know the readers of Experimenter are not interested in another engine article about stock, certified engines, so I won’t bother you with the irrelevant details about my Continental O-200. What I have done to the engine, which might be of interest, is the installation of the Simple Digital Systems (SDS) computer system on my new engine. The SDS unit is built as an automobile aftermarket MPFI and electronic ignition system. The fine folks at SDS are homebuilders. They built an RV-6A powered by a turbocharged Subaru EJ22 (http://SDSefi.com/air9.html)! Although SDS’s primary market is automobiles, they are certainly aviation-friendly.

The neat thing about the SDS computer is that you can literally program the engine operation “on the fly.” The unit comes with its own programmer, so you don’t need a laptop computer. Engine operating parameters can be changed with the engine at full power. Take engine ignition for example. Since the computer is controlling the ignition and you have complete control over the computer’s operation, you can tell it what timing you would like at any given rpm. In my current setup I am telling the computer that after engine start I want the engine timing to be 10 degrees before top dead center. As the rpm increases, the timing increases about 2 degrees to 4 degrees every 500 rpm. The fuel delivery is set to increase relative to rpm, throttle position, manifold pressure, barometric pressure, and outside air temperature. The power application is very smooth, and the engine does not sputter or pop upon application of full throttle from an idle.

The SDS computer requires engine rpm information. The way the designers chose to do this (so their system will work on any engine) is through the use of their timing pack. I made an aluminum plate and mounted it between the prop extension and the prop flange. I then constructed a mount to hold the Hall effect sensor pack. The Hall pack and the associated magnets are supplied with the SDS kit, but you will have to decide just how to mount the timing plate and Hall pack. SDS will also sell an extra coil setup if you tell them the installation is going on an aircraft engine and you want dual ignition.

The SDS requires high impedance (12 ohm to 16 ohm) automotive fuel injectors. Mine are specified as being for a 5.0L Ford Mustang. An aircraft conversion will require an intake manifold modification to accommodate the automotive injectors. The modern auto injectors require about 40 psi fuel pressure, so dual (redundant) automotive high-pressure fuel pumps were required, unless you think you can get by with only one. Mine are installed in parallel using a center off switch. During run-up I ensure that each ignition system is operating properly, and I switch to the other fuel pump under load. I start the plane using the fuel pump that was used for the previous flight. This ensures that each pump gets about the same amount of run time. The other sits there waiting to be called to action if the first one fails.
Even though the SDS is designed as an auto aftermarket unit, it comes with a mixture knob for those of us pilots who just have to fool with the mixture at cruise. When properly set up, the mixture will be real close, even at 14,500 feet. Modern auto engines run well at the top of Pikes Peak even if they are down on power from the 7 inches (or so) of manifold pressure drop form the altitude.

Dragonfly engine
With the throttle body located under the engine, it made for a tidy package after installing the injector bosses in the intake runners.

The engine runs very smoothly, and the fuel economy is quite a bit better than the stock O-200 with a carburetor with a set of magnetos. My fuel burn on two trips of 2.5 hours each way was 4.85 gph. This was at 8,500 feet and 9,500 feet using a setting of 22 inches of manifold pressure and 2750 rpm. The book value for the O-200 is about 6 gph.

I know there are many pilots who will never consider flying with a full computer-controlled engine, but some of us are just experimenters. I have significantly more than 700 flight hours behind fully computer-controlled engines. It’s just a matter of time before Continental and Lycoming have FADEC (full authority digital engine control) available for all engines.

Editor’s Note: SDS has a brief story of Justin’s plane on its website: http://SDSefi.com/air28.htm. Justin recently sold his Dragonfly.

Copyright © 2014 EAA Advertise With EAA :: About EAA :: History :: Job Openings :: Annual Report :: Contact Us :: Disclaimer/Privacy :: Site Map