The Why's of Efi
The Why's of EFI
TDC
Kevin Cameron
TODAY’S SOPHISTICATED ELECTRONIC fuel-injection traces its origin to the Bendix Electrojector system of 50 years ago. During WWII Bendix made thousands of injection carburetors for aircraft engines, then built mechanical direct fuel-injection systems for late-model B-29s. By the 1950s the booming civilian economy tempted them. Could Bendix design a mass-market fuel-injection system for autos?
To ensure smooth cold-starting and warm-up, it would have to measure many more variables than were considered in wartime metering units. The best way to sum the effects of many variables was via the fast-evolving field of electronics. Sensors would detect engine rpm, intake manifold pressure, air temperature and the like. Circuits would proportion and add these signals together to result in a sin gle fueling signal. In what form should it be implemented? The decision was to use electromagnetic injection valves connected to a constant fuel pressure. A single variable-the “on-time” of the injection valve-would determine the amount of fuel delivered per cycle.
Bendix employed analog control rather than digital, but the basics of modem EFI were defined in Bendix patents. Only a very few Electrojector fuel systems were produced-back in 1958 early transistor electronics lacked reliability.
When in the 1960s air pollution abatement became a U.S. priority, VW turned to Bosch for solutions. The pioneering Bosch D-Jetronic system that appeared on certain American-market VWs in 1967 was the mother of all that was to follow-and it rested upon Bendix patents.
Since then, digital systems have displaced analog, as it is easier to keep track of “yes/no” information than it is to measure degrees of “maybe.”
Most automotive fuel-injection systems measure the mass airflow entering the intake system, then proportion fuel delivery to it. This enables such systems to accommodate engine modifications without reprogramming. Early systems used a swinging air-door measuring element, but that has been replaced by hot-wire anemometer systems. A current-carrying wire or wire array is placed in the intake airflow, and the cooling of the wire by the airflow is measured as a change in the wire’s resistance. Suitably calibrated, this is a measure of airflow.
Motorcycle systems are usually of the “mapped,” or N-alpha variety, employing no air meter. A sample engine is run on the dyno, and the correct fuel quantity is determined for many rpm and throttle positions. This information, called a fuel map, is stored in digital form in the engine-control computer of the production vehicle. As the engine runs, the computer measures rpm (N) and throttle angle (alpha). Using this information, it looks up the correct injection valve on-time on the fuel map. This information may be modified according to other variables, such as air and engine temperature, barometric pressure or speed of throttle movement. Finally, pulses of the correct width are sent to the injection valves.
Dynamic range can be a problem. If the injector is sized such that its smallest repeatable cycle supplies correct fueling for idle, can it supply enough fuel for peak torque? In some cases, so much ontime is required per pulse to supply adequate fuel that the last part of the fuel injected has too little time in which to evaporate. In this case, if wider-range injectors are not available, the designer may equip each intake pipe with two or more injectors-a small vernier injector to handle idle-to-midrange, plus a larger higher-flow injector to handle the rest of the range.
Range can also be extended by varying fuel pressure supplied to the injectors-a low pressure being supplied from idle to just above, at which point the fuel pressure jumps up to a higher value to push adequate fuel through the injectors at peak torque. These 30-60-psi pumps are electric, mounted inside the fuel tank where they are cooled by the fuel itself.
Two-stroke direct injectors squirt very fast because events occur in double-time in two-strokes. This technology has been adopted in some four-stroke applications to deliver fuel more rapidly, allowing more time for evaporation. In Fonnula One, not only are the injectors very fast and located far upstream, above the intake bellmouths, but fuel of very high volatility is used. This is life at 20,000 rpm.
You might expect injectors to spray only during the intake stroke, but this is often not the case. In some engines, fuel sprays against closed intake valves, and in others it is half one way, half the other. To equalize ; differences, some designs inject half the fuel against the closed valve and the other half during the intake stroke. The purpose of these variations is simplicity-to allow all injectors to be “fired” together.
Some early motorcycle EFI systems evaporated fuel poorly, causing oil dilution or impaired throttle response. New technologies have corrected this. One of these mimics the action of a CV carburetor by adding a second throttle butterfly whose rate of opening is limited. This prevents air delivery from outrunning fuel delivery. Multi-hole injectors have replaced earlier single-hole types, improving the breakup of fuel into more numerous, smaller and faster-evaporating droplets.
Intensive auto industry development has made fuel-injection components extremely reliable. For those who have spent hours fiddling with jetting, balky chokes and fast-warmup devices on carburetors, electronic fuel-injection is a turnkey marvel. Because fuel-injection requires no restrictive venturis for its fuel metering, much larger, freer-breathing intake systems have become the norm. EFI not only continuously optimizes mixture for atmospheric or elevation changes, it can also hold mixture in the “sweet spot” required for best operation of threeway exhaust catalyst systems, if required.
At present, highly fuel-efficient, stratified-charge direct-injection gasoline engines are entering automotive service. They may one day rival the efficiency of diesels, but that will require fundamental changes in the combustion process.
