Electronic fuel injection (EFI) is ubiquitous in cars but still growing in aviation. Rotax is a major user of EFI in its iS line of engines, plus aftermarket EFI systems are commercially available for Experimental aircraft. Lycoming has sold a very limited number of EFI-equipped engines as well.
Note also that some of the Jet A (diesel) burning engines use their own form of EFI, which is a separate topic for a separate article.
As the name suggests, the big deal with EFI is the system is computer (digitally) controlled. This allows all sorts of neat tricks and efficiencies at the price of cost and complexity. Another fundamental of EFI is an electric fuel pump supplies abundant fuel pressure to enable a wide range of precisely metered fuel from idle to full power, along with more finely atomized fuel sprayed from the injector. Above all, it means less engine management work for the pilot. With EFI, there is no mixture knob, and starting means twisting the key or pushing a button.

Understanding electronic fuel injection beyond pushing the button begins by examining a typical system, if even at the basic level. And there is no better system to describe than Rotax’s, as the Austrian company dominates the fuel-injected aviation market.
Fuel is drawn from the aircraft’s tank by an electric fuel pump (there are two pumps for redundancy in the Rotax system). A water/sediment screen (similar to a gascolator) is placed before the fuel pump and a fine fuel filter after the pump. The fine filter uses a metal screen and bowl, again like a gascolator. The multiplicity of filters is your first clue that the close tolerances inside EFI systems are easily clogged by dirt, and water is never welcome in any fuel.
A vapor return loop starting just after the pumps returns any fuel vapors from the churning, pressurized fuel upstream of the pumps.
From the fine filter, the gasoline is plumbed to the engine, where it circulates through a fuel rail—metal piping connecting all the fuel injectors. There are two injectors per cylinder (more on this in a minute), and both fire into the cylinder’s intake port.
At the end of the fuel rail is a fuel regulator controlling the fuel pressure in the rail. Fuel not consumed by the injectors is returned to either a header tank or the fuel tank the fuel was originally drawn from. By restricting the return fuel flow to the tank, the regulator can raise the pressure in the fuel rail.
Rotax’s system maintains fuel pressure at a certain value above the air pressure in the inlet system’s airbox—manifold pressure, in other words. In practice, Rotax fuel pressure is nothing dramatic, roughly 40–100 psi. Obviously, the manifold pressure changes with throttle setting, rising altitude, ambient temperature, and so on, plus whether a turbocharger is boosting manifold pressure.

Unlike the simple, open nozzles in mechanical fuel injection, EFI fuel injectors are more complex electric switches. This allows the computer to open and close the injector for longer or shorter times as conditions demand. This controls how much fuel is delivered to the engine.
Fuel delivery can also be timed with EFI to squirt fuel sequentially—following the engine’s ignition firing order—or all the injectors can be fired simultaneously, called batch firing. It doesn’t make much difference in aviation engines because they run in a narrow, rarely varying rpm range. This is unlike auto engines, which use sequential injection exclusively these days.
Fuel pressure is much higher at the injectors with EFI than the barely there pressure found in mechanical, constant-flow aviation injection systems. This allows blowing the fuel into smaller, easier-to-burn droplets (atomization). It also means the passages in electronic fuel injectors are tiny and easy to clog, hence the emphasis on fuel filtering.
Another way to adjust how much fuel is delivered is to vary the fuel pressure. The higher the fuel pressure, the more fuel will spray from an injector in a given amount of time. Rotax uses both tactics, varying the duration the fuel injector is open along with raising or lowering the fuel pressure. All of this is decided by the engine control unit (ECU) many times per second.
For redundancy, Rotax uses two fuel injectors per cylinder; both fire into the intake port, and both injectors fire during normal engine operation.
There are two injectors because the Rotax fuel system is in many ways two separate fuel systems running in parallel. Rotax calls them Lane A and Lane B. This explains why there are two fuel pumps, two injectors per cylinder, two air pressure sensors in the airbox, and so on. During normal operation, Lane A could be called the primary Lane. Should something go wrong with Lane A—clogged injector, dead fuel pump, or blocked filter—Lane B automatically switches into primary mode. A cockpit display informs the pilot about the trouble, but the engine should run quite well so the aircraft can be safely landed and the system repaired.

Programming (software) in the Rotax ECU automatically provides the correct fuel (and ignition spark) for all engine conditions from starting to shutdown, so there is no mixture control knob. As with all EFI systems, the Rotax engines thus automatically provide either rich-of-peak or lean-of-peak exhaust gas temperature (EGT) operation depending on how much power the pilot demands—which he does by where he sets the throttle. And if Lane A has an issue, Lane B takes over using a rich-of-peak mixture to better avoid detonation and promote cooler cylinder temperatures. This is akin to the “limp home” mode in automotive engine management systems, but unlike the dramatic power loss of automotive limp home mode, the Rotax strategy delivers a likely imperceptible power loss.
Interestingly, Rotax uses EGT sensors as a sort of double-check on engine operation, but the system runs fine without them.
Other aviation EFI systems, not to mention their automotive big brothers, differ in detail. Two aftermarket systems—Fly EFii and Simple Digital System EFI—are popular in Experimental aircraft. They use a single fuel injector per cylinder and two separate ECUs (computers) for redundancy. Of course, their big advantage is they are available to retrofit Lycoming and Continental engines (among others), while the Rotax system comes only on Rotax engines.

For pilots, the big advantage of any EFI system is that the computer and its software eliminate the need for human control over fueling the engine. The intricacies of starting hot and cold engines, plus closely monitoring the air/fuel ratio or selecting rich- or lean-of-peak operation, are all computer controlled. The pilot’s fuel concerns are thus reduced to ensuring enough fuel is in the tanks and managing which tank he’s drawing from. Better fuel efficiency comes naturally as well because the computer does a better job of monitoring and delivering the correct amount of fuel than a distracted pilot fiddling with knobs. Plus, the EFI system can control the engine cylinder by cylinder, something a pilot can’t do. There are data acquisition possibilities with digital engine control as well; these help share fuel consumption information with flight management software, plus with trend monitoring for maintenance purposes.
If the EFI pilot no longer needs to twiddle with the now-missing mixture control knob, as an aircraft owner and maintainer he does need increased vigilance over the electrical system. That’s because any EFI quits instantly the moment the incorrect number of electrons are present. A very robust electrical system is required, it must be maintained scrupulously, and it must incorporate some form of backup alternator, battery, etc. This is even more important when EFI and electronic ignition are combined on the same engine, which is nearly always the case if EFI is fitted. Such an aircraft is electrically dependent and demands robust primary and secondary electrical systems.

Then there is the complexity of EFI. Carburetors may be old school, but they are eyewitnesses to the laws of physics. You don’t have to tell a carburetor much of anything, as it natively senses things such as density altitude. But EFI is inherently blind to the outside world and must be fitted with an array of sensors to provide the necessary information. Thus, there are crankshaft position sensors along with any number of temperature and pressure sensors, all connected by wires to the computer. It seems the computers are very robust and essentially never fail, but the wiring can get compromised and sensors sometimes go silent. EFI issues are rare—they’re as reliable as any other fuel system—but if issues arise, they’re likely going to be dead sensors or injured wiring.
Oxygen sensors would be nice, but can’t survive the lead in 100LL gasoline, so they are not regular aviation practice. Furthermore, the Rotax specialists at Motive Aero underscore the importance of running quality gasoline of the correct octane rating Rotax calls for. Low-octane gas is especially bad at altitude or high ambient temperatures. The system can compensate somewhat for poor gasoline, but can’t work miracles.
Finally, for no practical application other than the amusement of the tech heads in the audience, aviation EFI practice is to use speed density engine management systems, while automotive practice follows mass air principles. Rotax, Fly EFii, and SDS EFI are all speed density, so they have no mass air meter, meaning less weight, less cost, easier packaging, and lower air restriction. They use simple “look-up” fuel and spark tables so they are fast reacting, but must infer air density from throttle position, manifold pressure sensors, and the like. Automotive systems read air density directly with a mass air meter and require more processing power to run algorithms to arrive at fuel and spark values, but handle tricky rpm transitions and engine conditions a little better. All this doesn’t matter to the huge majority of pilots and A&P techs, but it’s a reminder to tinkering Experimental types that fine-tuning aftermarket EFI systems can be part of the plan when hot-rodding engines.


Aerosance / TCM rolled out a full-FADEC ignition and mixture system on the IOF240B (F for Fadec) some 22 years ago now – it was installed almost exclusively in the Liberty XL2, a nimble two seat training / personal touring airplane. (Think of it as a less powerful, certificated Vans.) I bought one, in part because we were ‘about to’ switch to unleaded fuel, and to modify the engine/plane for that incipient event simply would have meant bolting on two new computer boxes – a 45 minute job.
The system used a variety of sensors – crankshaft position (to time the spark), RPM, MAP, MAT, EGT, Voltage and Fuel Pressure – and even included firing a second spark during the exhaust phase to help clean the plugs. Computer-analyzed test runs could be performed, with reams of data captured on a laptop for analysis by the owner / mechanic and subsequent upload to TCM for review.
I flew behind three of these engines for nearly 5000 hours over the next 20 years. While not without the occasional maintenance issue, I was very pleased. The engine ran smoothly, started easily even when hot, and consumed about 1 GPH less than a similar non-FADEC motor, since it automatically leaned to LOP when at cruise RPM. This considerably extended range on my small plane, and of course saved money – though the amount of money saved on fuel was nearly identical to the increase in cost when it was time to overhaul the more complex system…
But, as the article points out, such systems are inherently more complex than the stone-simple carb or even a standard Fuel Injection system. Rather than use the ~120 Liberty aircraft as a test fleet in real-world conditions, TCM abruptly decided to stop supporting the engine, and neither parts (sensors, injectors, computer boxes, wiring harness etc.) nor technical support are now available, orphaning the engine and aircraft. Indeed, a broken wire in the harness between one of the MAP sensors and the computer boxes was the coup de grace for my plane – this despite pleas to TCM for help and countless hours spent with my mechanic trying to repair the defect.
So – by all means adopt a cutting-edge modern technology, but ensure that you can retrofit to the basic configuration if technical or hardware support should disappear.
KISS !!!
Yup, A carb only uses air velocity to deliver fuel. And in most high wing Cessnas, the fuel flows by gravity…no fuel pump. Stone simple.
Yup #2 … and in MY 172M, the fuel tanks are aluminum vs. bladder types, too.
The only faux pas in it is the fuel line from each tank running down the door post behind a plastic overlay a few inches forward of the rear seat ash trays !! (sic) What genius thought of THAT ??
My ’61 C175B has aluminum door post covers, no ash trays in the rear area, and the front ash trays were replaced with recessed headset jacks.!!
The benefits of a properly installed and tuned EFI are enormous. First and foremost it eliminates the need for carb heat. Second, it practically eliminates water in fuel issues, not because of a “water screen” but because it pumps so much fuel past the injectors, (80 percent of the fuel is returned typically) that any water present is so mixed in or widely distributed that it does not effect anything. 5 micron fuel filtration is required on all FI systems, mechanical or electronic. EFI systems can be optimally tuned but rarely are. The most common ones now in use in aircraft don’t have the sophistication to do a good job. This is now evident with some Rotax IS engines grounded with melted exhaust pipes and worse.
As Tom points out, the electrical requirements are critical. This is because the constantly running high pressure pumps require a lot of power ~7A for one pump alone. Add 1/2A per injector and 5-6A for a dual (inductive) ignition system and your power demand is 15A minimum in cruise, a little less at idle. These loads are beyond the rating of a toggle switch and need to be handled with relays and those need to have a back-up relay and so on. It typically requires a larger alternator and a much larger battery or two.
With proper tuning, great improvements in efficiency and engine reliability can be achieved (see the automotive world) but the likelihood of a fire from a high pressure fuel leak is omnipresent especially with experimentally cobbled together systems.
Flying with a very custom EFI since 2003.
I have to say I am a little disappointed in the accuracy of the article. As someone whose carreer has spanned the entirety of the EFI developments in the automotive world and who spends his spare time in aviation, I see this as being written by someone who should have asked the experts more questions.
To start at the end – Airflow sensing is not the prevalent methodology in the automotive industry but is used when the maker is struggling to match certain emission rules. Alpha-N is the most common which means the primarty information is throttle position and RPM, manifold pressure and air temp is at the second level of look up tables. The same system can be MAP, RPM with throttle position as the second lewvel table but in cars that can have a slightly spongy response to rapid driver demand.
Rotax – the system built by rotax is one of the messiest, most cumbersome and overly complex I have experienced. The connectors are below the standard most would apply in this environment. Deutsche Connectors would be considering the minimum. Dual pumps and filters is standard in motorsport as they are the things experiencing rubbish in the fuel, so that is good. Modern systems mostly sense the fuel pressure and use PWM to regulate the fuel pump speed which does away with the regulator and return lines etc.The plenum, manifolds and injector systems are pretty poor in regards layout and airflow needs of the cylinders. Dual injectors are not required as they are the least likely to fail and if they do, only one cylinder is affected.
My 40+ years of experience shows me that well atomised fuel, placed at the valve will allow an engines with a proper electronic ( not flywheel CDI ) igntion to run leaner and smoother than anything else. They can also be very reliable and fault tolerant. The Rotax system in the lunchboxes is horrible and I would have sent the apprentice back to try again if he had presented me with that!!
Phil Kennedy,
Thank you for your input, I am similarly disappointed by the FI systems out there.
I am surprised by your statement “not flywheel CDI”. I would love to hear your concerns, maybe Ross can give you my email for further discussion.
At this point I cruise at over 200ktas on <5gph for 48mpg with my 0-360 in a loaded two seat airplane using crank triggered Dual CDI, Lambda 1.2. I am not aware anyone else is even close. If you think I could improve on that with another ignition type, I would love to hear about it.
A mixture knob can be added to an EFI system and is highly desirable. You typically want max power (somewhat rich) during takeoff and initial climb. Then leaner in shallow climb to cruise altitude and finally lean to best economy at cruise (unless you want full power to get there fast). This is not something that can be controlled by throttle only.