How it Works: Fuel Injection

Send This Page To A Friend
Fade To White
How It Works: Fuel Injection

Fuel Injection

The Wright brothers

The history of fuel injection dates back at least to 1903 when the Wright brothers' first sustained powered flight was propelled by an engine with petrol injection. Herbert Akroyd Stuart developed the first system laid out on modern lines (with a highly accurate 'jerk pump' to meter out fuel oil at high pressure to an injector. This system was used on the hot bulb engine and was adapted and improved by Robert Bosch and Clessie Cummins for use on diesel engines.

Jonas Hesselman

Rudolf Diesel's original system employed a cumbersome 'air-blast' system using highly compressed air. The first use of direct gasoline injection was on the Hesselman engine invented by Swedish engineer Jonas Hesselman in 1925. Hesselman engines used the ultra lean burn principle; fuel was injected toward the end of the compression stroke, then ignited with a spark plug. They were often started on gasoline and then switched to diesel or kerosene.

fuel injection was in widespread commercial use in diesel engines by the mid-1920s. Because of its greater immunity to wildly changing g-forces on the engine, the concept was adapted for use in gasoline-powered aircraft during World War 2, and direct injection was employed in some notable designs like the Junkers Jumo 210, the Daimler-Benz DB 601, the BMW 801, the Shvetsov ASh-82FN (M-82FN) and later versions of the Wright R-3350 used in the B-29 Superfortress.

Alfa Romeo tested one of the very first electric injection systems (Caproni-Fuscaldo) in Alfa Romeo 6C2500 with "Ala spessa" body in 1940 Mille Miglia. The engine had six electrically operated injectors and were fed by a semi-high pressure circulating fuel pump system.For a long time it was oil-engine practice that was followed in experiments with petrol injection, most successfully by Bosch and Daimler-Benz in the development of fully aerobatic aero engines for the Luftwaffe; and a Bosch jerk-pump system was evolved for the BMW 328 sports cars specially built for the 1940 Mille Miglia race, though it was not in fact used.

The Hilborn Continuous Injection System

The jerk-pump contained a number of plungers moved by a camshaft geared to the engine speed; each plunger displacing fuel from its cylinder through small-bore piping to an injector nozzle feeding an inlet port or engine cylinder. A rotating rack and quadrant mechanism turned the plungers in their bores so as to vary the unmasking of plunger - controlled ports which, in turn, governed the quantity of fuel transmitted on each stroke; this mechanism was linked to the throttle pedal and to a pneumatic sensor of inlet manifold vacuum, so as to cater for variations in engine demand.

It was an imperfect system when applied to the petrol engine, but for racing machines exploiting alcohol-based fuels that were not sensitive to mixture strength, it worked well enough, and the 1954/55 Mercedes-Benz Grand Prix cars introduced it most effectively to racing, Already, there had been other attempts; also reliant on the forgiving nature of alcohol fuels: the Hilborn continuous injection system, in which the only variation in supply was by varying the pressure of fuel pumped to the injector nozzles in the ports, was much favoured at Indianapolis and was tried successfully in England by Connaught in their 1953 Formula Two car.

The Lucas Shuttle Metering System

Vanwall went to Bosch for help with their Grand Prix car, and in 1958 were probably the first to run a racing car with petrol injection. Nevertheless, it was the perfection by Lucas of an ingenious shuttle metering system that made possible the more popular approach to the problem, in which an electrically driven pump pressurised the delivery of fuel to an engine-driven distributor that puts each injector in communication with this pressure in its appropriate turn. A cam or lever moved by the accelerator pedal adjusted the abutments that stopped the metering shuttle at the end of each stroke, and thus varied the quantity of fuel displaced by each movement.

For racing this was sufficient, but for a road car there was a need for finer control, provided by a pneumatic transducer sensitive to intake manifold pressure. A similar sensor could adjust the mixture according to ambient atmospheric pressure. This Lucas system provided timed doses of fuel to each cylinder, but it was demonstrated by BRM (who were the first to employ it) that the timing was not critical. Other manufacturers of injection systems took advantage of this to simplify their systems by adopting a continuous spray, the quantity of fuel delivered being varied by alterations in delivery pressure.

Fuel Injected Cosworth Vega engine
A Cosworth Vega engine, with the carby replaced by fuel injection. The injection tubes are each side of the induction pipe, linked at their open ends so that only one throttle was necessary.

The Tecalemit-Jackson System

The Tecalemit-Jackson system, capable of giving an engine as much power potential as any form of injection or carburetion, was one such. In general, however, it was found that the means of sensing all the different parameters that affected the engine's need for mixture, in terms of quantity and richness, were the aspect of design that needed most development, and numerous injection systems that appeared in the 1950s and 1960s (particularly in the USA but also in Europe and Japan) left something to be desired. The difficulties of translating things such as throttle position and manifold depression into measures of engine load, and relating the answer to engine speed and atmospheric pressure and ambient and engine temperatures (the needs of cold starting and fast idling during warm-up had to be remembered), promoted great variety in experiments.

Bosch, Bendix, Lucas and others tried electronic control, arguing that it enabled any number of factors to be measured and corresponding adjustments to be made: a miniature transistorised computer sorted all these transducer signals and issued a pulse of current of a certain duration to solenoid-controlled injector nozzles supplied with fuel at constant pressure. The longer the pulse, the more petrol would be squirted into each port during each operating cycle of the engine. Other firms, such as Kugelfischer, relied on mechanical refinements, such as a three-dimensional cam controlling the fuel-distributor output.

It was unfortunate that for a long time nobody tried to make the injection apparatus measure what the carburettor measures automatically, which is the mass flow of air into the engine. In 1970, the Tecalemit engineer Jackson produced an electro-pneumatic system that actually measured what was required (instead of inferring it from other measurements), and the idea of mass-flow measurement was enthusiastically taken up thereafter by Bosch, who modified their existing electronic system accordingly, Unlike Bosch, the Tecalemit subsidiary Petrol Injection Ltd did not get their system into production.

The Amal GP

Even during the time when fuel injection systems were proving much more effective, there were carburettors that allowed as free an airflow as the best injection systems - the Amal GP, long used on racing motor cycles, was a good example - and when one of these was used for each cylinder, the power realised was almost identical to when injection was applied to individual inlet tracts. It was during the 1970s that racing teams were trying to overcome the weight and complexity of a corresponding number of carburettors for each cylinder, 8, 12 or 16 in some cases, and each of which had to be adjusted individually.

The centralised control and delivery systems of injection apparatus were by comparison lighter, more compact and easier to adjust. More severe still in racing was the carburettor's sensitivity to surge of fuel in the float chamber as the car cornered hard or brakes heavily: the best Grand Prix cars of the era would corner at as much as 1.7 g, when the fuel inside a carburettor would be flung to one side so forcefully that its surface would adopt an angle 59 degrees away from the horizontal - with what ensuing difficulties in starving or flooding the jets may be imagined.

Fuel Injection on Production Cars

Early versions of fuel injection systems as applied to cars rolling off the production line had an injector nozzle for each inlet port of the engine, but all these ports would draw their air from a manifold controlled by a single upstream throttle assembly and drawing from a single air cleaner and silencer. All the aerodynamic solecisms of the carburettor manifold were thus inherited, and volumetric efficiency may accordingly be no higher, as it is often the case that manifold design handicapped airflow more severely than the carburettor venturi.

Because different quantities of air may reach different cylinders through such a manifold, the mere fact of equal doses of fuel being delivered to each did not ensure equal mixture distribution nor uniform mixture strength. It was consequently unlikely that economy (or freedom from certain noxious exhaust emissions) would be as good as might be expected. On the other hand, if each intake port were fitted not only with its own injector but also with its own inlet pipe, throttle and air filter, results could be achieved that should be superior to any but the very best multi-carburettor systems, with advantages in cost, weight, accessibility, and ease of maintenance.

10 Advantages of Fuel Injection

To our mind Toyota were the first manufacturer to serioulsy standardise fuel injection across their range (excluding the Germans of course), and promote the benefits derived from such a system. Obviously a Toyota did not have a lot of cylinders when compared to a F1 racer, and was never likely to pull 1.7 g around a corner. But they, along with other manufacturers, believed in the advantages and there were at least ten:

  1. Because of the absence of venturi restrictions which ordinary carburettors could not avoid, volumetric (or breathing) efficiency is higher, and so therefore should be the power output;
  2. Mixture distribution is better, each cylinder being given the same dosage;
  3. Mixture strength is uniform for each cylinder;
  4. Fuel economy is better because of 2 and 3;
  5. There is mechanically induced reduction of the liquid fuel to fine droplets, hence no need to heat the inlet air to ensure adequate vaporisation, and therefore no loss of volumetric efficiency due to the lower density of heated air;
  6. Acceleration response is better, the extra fuel needed being injected forthwith instead of flowing only after air flow has changed and depression has drawn it from the jets;
  7. Cold starting is better because of 5 and 6;
  8. Idling is more uniform because of 2 and 3;
  9. Inlet-valve cooling is improved, because latent heat of evaporation need not be taken out of the fuel earlier in its passage (this is not valid in the case of direct injection into the cylinder);
  10. Icing in very cold weather is absent, because of the same reason as in 9.
But it was not only the advantages of injection that ensured it would replace the carby. The requirements of emission-control regulations affected the carburettor severely, making it much more complex by the mid 1970s, and few of these changes helping with performance. Where injection was deemed to be too expensive and complex for deployment on an ordinary family sedan, to comply with pollution regulations the complexity was such as to put the carburettor and injection systems on a more even footing in terms of cost and difficulties of maintenance.

The Electronic fuel injection System

The first commercial electronic fuel injection (EFI) system was Electrojector, developed by the Bendix Corporation and was to be offered by American Motors (AMC) in 1957. A special muscle car model, the Rambler Rebel, showcased AMC's new 327 cu in (5.4 litre) engine. The Electrojector was an option and rated at 288 bhp (214.8 kW). With no Venturi effect or heated carburettor (to help vaporize the gasoline) AMC's EFI equipped engine breathed easier with denser cold air to pack more power sooner, reaching peak torque at 500 rpm lower than the equivalent no-fuel injection engine.

The Rebel Owners Manual described the design and operation of the new system. Initial press information about the Bendix system in December 1956 was followed in March 1957 by a price bulletin that pegged the option at US$395, but due to supplier difficulties, fuel-injected Rebels would only be available after June 15. This was to have been the first production EFI engine, but Electrojector's teething problems meant only pre-production cars were so equipped: thus, very few cars so equipped were ever sold and none were made available to the public. The EFI system in the Rambler was a far more-advanced setup than the mechanical types then appearing on the market and the engines ran fine in warm weather, but suffered hard starting in cooler temperatures.

Chrysler offered Electrojector on the 1958 Chrysler 300D, Dodge D500, Plymouth Fury, and DeSoto Adventurer, arguably the first series-production cars equipped with an EFI system. It was jointly engineered by Chrysler and Bendix. The early electronic components were not equal to the rigors of underhood service, however, and were too slow to keep up with the demands of "on the fly" engine control. Most of the 35 vehicles originally so equipped were field-retrofitted with 4-barrel carburettors. The Electrojector patents were subsequently sold to Bosch.

Bosch developed an electronic fuel injection system, called D-Jetronic (D for Druck, German for "pressure"), which was first used on the VW 1600TL/E in 1967. This was a speed/density system, using engine speed and intake manifold air density to calculate "air mass" flow rate and thus fuel requirements. This system was adopted by VW, Mercedes-Benz, Porsche, Citroën, Saab, and Volvo. Lucas licensed the system for production with Jaguar. Bosch superseded the D-Jetronic system with the K-Jetronic and L-Jetronic systems for 1974, though some cars (such as the Volvo 164) continued using D-Jetronic for the following several years.

Bosch's L-Jetronic

The Cadillac Seville was introduced in 1975 with an EFI system made by Bendix and modelled very closely on Bosch's D-Jetronic. L-Jetronic first appeared on the 1974 Porsche 914, and used a mechanical airflow meter (L for Luft, German for "air") that produces a signal that is proportional to "air volume". This approach required additional sensors to measure the atmospheric pressure and temperature, to ultimately calculate "air mass". L-Jetronic was widely adopted on European cars of that period, and a few Japanese models a short time later. The limited production Chevrolet Cosworth Vega was introduced in March 1975 using a Bendix EFI system with pulse-time manifold injection, four injector valves, an electronic control unit (ECU), five independent sensors and two fuel pumps.

5000 hand-built Cosworth Vega engines were produced but only 3508 cars were sold through 1976. A major milestone was reached in 1980 when Motorola Corporation introduced the first engine computer with microprocessor (digital) control, the EEC III module, which is now the standard approach. The advent of the digital microprocessor permitted the integration of all powertrain sub-systems into a single control module. In 1981 Chrysler Corporation introduced an EFI system featuring a sensor that directly measures the air mass flow into the engine, on the Imperial automobile (5.2 litre V8) as standard equipment. The mass air sensor utilized a heated platinum wire placed in the incoming air flow. The rate of the wire's cooling is proportional to the air mass flowing across the wire. Since the hot wire sensor directly measures air mass, the need for additional temperature and pressure sensors was eliminated. This system was independently developed and engineered in Highland Park, Michigan and manufactured at Chrysler's Electronics division in Huntsville, Alabama, USA.
Latest Classic Car Classifieds