Electronic Fuel Injection and Ignition: 


 (click here to go to the most recent posting)

I currently have installed the Real World Solutions EC2 electronic fuel injection and ignition.  My decision for choosing this system hinged greatly on it's backup functions.  The EC2 computer is literally two complete systems soldered to one board.  The "A" computer utilizes an intake air temperature sensor, a manifold pressure (MAP) sensor, a cam position sensor and a crank position sensor to operate.  The "B" computer utilizes the cam position sensor and it's own MAP sensor to operate.  Selection of either the "A" or "B" system is done via a toggle switch on the pilots control board mounted on the instrument panel.  At the time I selected this product it appeared to be the best system due to this backup feature.

So far, tuning and operation have been pretty straight forward.  The simplicity of the control panel is very nice.  While performing the initial tuning the controls rapidly become very intuitive.  The EC2 instructions are very clear and make installation and tuning a breeze.

During this initial tuning, my system developed a problem.  Fortunately I wasn't flying or doing any high speed taxiing yet.  The system started cutting out and running rough as it would warm up.  This would happen regardless of which computer was selected.  I chased every wire three or four times without finding a problem.  I decided to replace each sensor with a known good part to eliminate them as the issue.  Replacing the crank sensor made no difference.  When I replaced the cam sensor, the system functioned properly and I was able to continue tuning.

Later that evening, while thinking about the reality that the backup system would serve no real use if the cam sensor failed, I started looking into the failure rate of this sensor.  My research wasn't very scientific.  In fact, it wasn't scientific at all.  It was more like Yellow Pages.  I started calling Subaru dealers and after-market parts suppliers and found that this is one of the most commonly sold sensors for this engine.

That finding was enough to ground my flight plans.  As with every aspect of this project, I weighed the risks.  I came up with the idea that I could just replace that sensor every 50-100 hours to be sure it would not fail.  Of course, because it's an electronic component, preemptive replacement is no guarantee.

What now?  Well, unfortunately, more research.

I'm going back to my initial idea of installing a Subaru computer, sensors and wiring.  As of 2000 the Subaru computers can store the information that they have learned about how a sensor functions.  In the event of a sensor failure or (and this is the best part) partial failure (my cam sensor situation), the malfunction indicator lamp (MIL) comes on and this stored information replaces the sensor until the engine is shut down.  Now that is a neat concept.

Of course it's much easier to type "installing a Subaru computer, sensors and wiring" than it is to do.

These pages will document my research and tests of the Subaru computer installation for my airplane.

Collecting Data:

I decided to use the 2001 Subaru Legacy Outback SW with 2.5L H4 engine and manual transmission.

First I had to collect wire schematics to determine what I was up against.  The Subaru wire diagrams are very easy to read and give a ton of information regarding how the system functions.  I've started a small library to store all of this information for use at the hangar.

The wiring diagrams were only the start. To determine what information the computer looks for and what it will set a code for, I had to gather all the information on the diagnostic trouble codes that pertain to this vehicle.

Generally speaking, the cars have several systems that are of no use in the airplane environment.  I've performed only book research on these systems up until now and have determined that each of them should either be ok to leave as designed, or easily tricked into thinking the system is still functioning to avoid the MIL coming on. More on that below.

Manual Labor:mess.JPG (40975 bytes)

12/22/2004: I purchased a complete vehicle wire harness for the car along with the engine harness, and related sensors.  I received the ECU as a gift.  The next step is sorting through the wire harnesses eliminating what is not used by the ECU.  The engine harness is in the middle of the picture on the left side.  The complete cabin harness takes up a majority of the photo and the fuel tank sensors are scattered in the upper left corner along with the under hood fuse box.  Most of the sensors and relays are still attached to the harnesses.  I left the main power distribution harness out of the picture because it won't be used in the plane.  I just wanted to make sure I had everything I would possibly need.

11/04/2005: I have returned to the task of replacing the current EFI system for a complete Subaru system.  I spent the a lot of my spare time over the past eleven months researching this task. The best assistance I found for my questions concerning the harness conversion was from the guys on the Yahoo group "Subaruvanagon".

The Subaru system has a couple areas that come from the automobile environment that are not used, and frankly, a pain in the butt for aircraft use. These systems include the Vehicle Speed Sensor (VSS), Fuel Tank Pressure (FPS) and fuel vent system.

The fuel vent/pressure system is a series of valves and sensors that the ECU utilizes to determine if the fuel tank on 1996 and newer vehicles is sealing properly. I used the repair manual diagnostic charts to attempt to come up with a way to fool the ECU into thinking this system was intact and functioning properly. Just prior to finalizing the wiring, I found a company, www.smallcar.com, that has already found an inexpensive, single box, solution to these sensors and valves and also the radiator fan relay wiring that the ECU looks for. It's called the SIB - Interface board (at the bottom of the linked page). Again, most of this info and additional supporting documents are in the Library

As for the VSS sensor, Eggenfellner engine manufacturer, Jan, stated at one point that the ECU needs to see a speed at least above 35 mph. This is what I will be shooting for. So, I decided to build a square wave oscillator utilizing a 555 timer IC and a pre-designed kit from www.electronics123.com. This oscillator puts out a square wave signal in several frequencies via jumper; approximately 1Hz, 10Hz, 100Hz, 1KHz, 10KHz & 100KHz. This is what I will experiment with to duplicate the speed sensor signal to the ECU. I was also able to find some additional information on the Subaru VSS signal. While reading a website belonging to a speedometer repair shop, I found that the ECU is programmed to interpret 4,000 pulses as 1 mile. I made this simple Excel sheet to calculate what I needed from the oscillator.

Task hours: 6.0: Over the past week, I modified the factory wire harness for aircraft use. The documents I used to determine which wires could go away and which had to stay are in this section of the Library.

The next two pictures are the results after 6 hours of disassembly and re-wrapping the harnesses to fit the aircraft. You can click on the thumbnail to enlarge them, or click on the title to view them with component location markings

Harness Main.JPG (114296 bytes)                        Harness Engine.JPG (115571 bytes) 

main harness  photo with markup           engine harness  photo with markup

11/05/2005: Task Hours: 2.0: Today I was going to fit the harness into the airplane, however, I was in a nosey mood instead. All along the Subaru harness exist these little tape-wrapped splices (pic #1 below). I was initially going to just leave them alone, except I'm unable to really do that with anything mechanically oriented. Instead, I removed the tape from one of the splices. I'm really glad I did. These splices are merely a small crimped-on piece of metal that resembles the internal, crimp portion, of a cheap electrical connector (pic #2 below). I decided that the best way to deal with these and avoid creating further risk of failure was to simply solder the crimp (pic #3 below) and reseal it with some new tape (pic #4 below). Of course, I had to do all twenty-three of them. I feel better now.

Harness Splice1.JPG (64604 bytes)  Harness Splice2.JPG (59984 bytes)  Harness Splice3.JPG (59787 bytes)  Harness Splice4.JPG (59016 bytes)

11/25/2006: Task Hours: 2.0: I'm back after a long break. I went to Carr Subaru in Beaverton Oregon today to get some help with the sensors that I didn't have. I took my ECU and wiring harness to help match the parts as best possible. The gentleman working this Saturday helped more than anyone ever has. I told him I was using a 2001 Legacy Outback H4 with manual transmission to convert to a "sand rail" and that I needed several of the engine sensors. I handed him the ECU to use as a reference and he went to work on his parts catalog.  After several minutes he returned with the list of sensors and an ECU cross-reference sheet from the catalog. He said he had several of the sensors in stock and could get the remainder of them by Monday. Then he proceeded to deliver the bad news. He pointed out on the cross-reference sheet that the ECU in my possession was, in fact, for an automatic transmission. He handed me the sheet and I verified for myself that he was correct. This ECU was supplied to me by a long-time friend at a large, national chain, auto recycling company. I'm very happy that I found this out before I had everything installed and was beating my head against the wall because the engine wouldn't run. I'm including a copy of the Subaru ECU cross-reference chart in my library and throwing away the listing that my friend provided. It covers the years 1999 to 2002 which includes 23 different ECUs for the Legacy 2.5L and 3.0L engines.

1/8/2007: Task Hours: 20.0: I spent a week on vacation and a couple weekends removing the EC2 system and installing the modified Subaru harness in the airplane. I also spent several evenings making new drawings of the modified Subaru harness. I've posted them in the Library in the Airframe Electrical section. I will post more information along with pictures as I have time.

7/2/2007: Task Hours: 2.0: Now that the engine is running I'm turning my attention to the weaknesses of the factory fuel injection. The first issue was discovered when an Eggenfellner engine quit running and the pilot was forced into an off-field landing due to vapor-lock. The Subaru system, as with most modern car-based EFI systems, is not equipped to handle vapor-lock. Subaru instead mounts their fuel pumps submerged inside the fuel tank to avoid vapor-lock altogether. It would have been almost impossible to mount my fuel pumps inside the tanks so I settled for under the footrest just ahead of and below the wing leading edge. This places them in an area where they will have positive head pressure most of the time during normal flight. For the marginal times like takeoff and landing I equipped my plane with the optional header tank. However, I still wanted a way to handle a vapor-lock situation minimizing the forced landing outcome.

Eggenfellner came up with the idea of a small, controlled leak, at the fuel pressure regulator. The problem is that when vapor-lock happens vapor will collect behind the fuel pressure regulator and cause it to close due to pressure loss when the fuel pumps can not pressurize the air beyond the minimum required to open it. The controlled leak is an .015" orifice placed between the high pressure fuel rail before the regulator and the return line after the regulator. The leak isn't large enough that the fuel pumps can not keep up. If the regulator vapor-locks, the engine may still die, but, the air will bleed out of the high pressure rail and into the return line with greater probability of causing fuel to replace it and the pressure to build back up opening the regulator. Of course the situation that caused the vapor-lock needs to be remedied in the interim, ie; descend or switch tanks.. etc.

As brilliant as this was, personally, I never liked the somewhat cheap looking way that Eggenfellner implemented the bleed orifice. He drilled out a piece of brass round stock and stuffed it into a rubber fuel hose. I prefer jic fittings and racing hose. Here's my solution... First, I installed the fuel rails and injectors off of a 1996 2.5L on my 2001 2.5L. This took a little filing to make the mounting holes line up, but once they did, the 1996 style seals worked perfectly with the new style manifold bosses. Next, I had a shop CNC a chunk of aluminum that gets sandwiched between the regulator and the fuel rail and has an 1/8" NPT T-d into it with the bleed orifice drilled through the aluminum chunk and into the high pressure fuel passage. Click the picture above to view a three-view of the part.

Below are photos of my system installed. The first photo is the block as it is mounted between the fuel rail and regulator. The second photo shows the bypass hose coming out at a 45deg angle (ya, ya... cheap brass, I was testing) heading under the manifold to the T-fitting in the return line. The third photo is the T-fitting in the return line. the fourth photo is the overall system view.