Fuel Injection

by Kevin ash

 

It used to be the case that fuel injection was mostly the choice of smaller manufacturers who couldn't get the big carburettor manufacturers to produce small runs of carbs tailored to their own bikes at an economical rate. Companies such as Moto Guzzi, Ducati, Bimota and to some extent BMW found it more cost effective to buy in proprietary fuel injection components, then program the systems to suit the requirements of each particular type of engine.
These days, lagging many years behind the car industry, more and more factories are turning to fuel injection as a means of meeting increasingly tough exhaust emission laws without losing too much in terms of power and driveability. Injection is still more expensive than carburettors where large production runs are involved, and for outright, maximum horsepower makes little difference to an engine, but often the only way to keep an engine as responsive and smooth running as modern riders demand while minimising pollutants is to take advantage of the very fine control offered by fuel injection.
The various systems all operate on the same principles. Controlling everything is an ECU (electronic control unit) which is constantly fed a whole raft of information about the engine and the ambient atmospheric conditions. Typically, there will be sensors reading throttle position, engine rpm, engine temperature, crankshaft position, camshaft position (so the ECU knows if an engine is on its power stroke or intake stroke, for example) intake air pressure, intake air temperature and atmospheric pressure.


 

The engine management computer is fed information from a whole array of sensors - this is the Suzuki Hayabusa

This data is constantly changing, sometimes hundreds of times a second. To make sense of it, the ECU has information stored in the form of injection maps on which every foreseeable combination of the more important sensor inputs is recorded. The ECU goes to the point on the map which corresponds to the particular combination of sensor inputs it's currently being fed, then reads off a pre-determined amount of fuel to squirt into the inlet tract.
The Suzuki Hayabusa's two-stage system is quite typical. Under light load conditions (partial throttle openings and low rpm) the injection is controlled primarily by a map based on air intake pressure and engine rpm. At high revs and heavy loads (large throttle openings) it switches to a second map based on throttle position and engine rpm. The reason for the switch is that Suzuki's engineers found they were getting better throttle response this way.
Each of the Hayabusa's four cylinders has its own version of these two basic map types, to allow for the slightly different operating temperatures, which means the engine depends on eight different maps across its rev and load ranges.
The information on these maps is then modified by additional information. For example, during cold engine starts the engine coolant temperature and intake air temperature are used to correct the amount of fuel injected (making it richer), and at high altitudes the atmospheric pressure information is used to lean off the mixture.
All of this information is painstakingly accumulated by the engine design teams during the development stage. First of all, the new engine is run on a dynamometer and a basic ECU calibration is worked out to get it running reasonably well. For example, the engine will be run at 3000rpm first with a very small throttle opening, and the optimum amount of fuel noted for the best combination of power and exhaust emissions. Then the throttle is opened slightly (and the load on the engine increased to keep it at 3000rpm) and once again, the optimum amount of fuel determined. This will be repeated at 3000rpm with increasing throttle openings right up to full throttle, before the whole process is repeated at, say, 3100rpm. Yes, it takes a very long time, especially with modern generation engines running up to as high as 14,000rpm or more!


 

A typical fuel injection map - the height represents the amount of fuel injected, the x-axis the engine rpm and the y-axis the throttle position. This is for one cylinder only, and for high loads.

This is only the beginning. Now that the engine runs well enough to fit to a bike, the whole system has to be fine-tuned, constantly balancing the demands of exhaust emissions, power and driveability (which is mainly throttle response). Ask an injection engineer what his reaction is when half way through this process a decision is made to change the airbox shape, alter the camshaft timing or something similar, and you stand a good chance of getting hurt!
The ECU controls the amount of fuel being injected by altering the amount of time the injectors are open. The fuel lines are held at a fixed pressure by the fuel pump and pressure valves, usually about 3 bar in car systems and at slightly higher pressures on motorcycles, so if an injector is open for a certain time, the amount of fuel which will flow through it will always be the same.
Or it is until the limits of the injectors start to be reached. At very small throttle openings the time taken for the fuel in the fuel line to accelerate from being stationary starts to have an influence and you can no longer assume that halving the opening time will give you half the fuel delivery - it will be slightly less. This non-linear relationship is usually dealt with by the injection map, but a bigger problem can occur at the other end of the scale.
The flow rate through a typical injector is around 150 grammes per minute, but some high performance machines with large cylinders (Ducati V-twins, for example) might demand a lot more fuel, even as high as 300 grammes per minute. Purpose-built high flow injectors would be prohibitively expensive, even for factory race teams, so the solution is to use more than one injector per cylinder.
This introduces other possibilities. One is that low rev, small throttle opening mixture control can be more accurate, as one injector can be switched off or set to a controlled minimum while the other varies the flow rate. This is of no use on race bikes of course, but is already used in some cars so will come to road bikes eventually. Another variation used by Ducati is to have two main injectors positioned conventionally, close to the throttle butterfly, with a third, smaller flow injector further up the inlet tract whose job is to provide a finer control of the mixture.
Most of the more refined fuel injection systems are sequential, which means the injection pulses are timed to coincide with a particular engine position (so, for example, the fuel is squirted into the inlet tract just as the inlet valve is opening). This sequential control also allows extra injection pulses on occasion, particularly when the throttle is suddenly snapped open, so extra fuel can be added to prevent a sudden leaning of the mixture (the same job done by accelerator pumps, where fitted, on carburettors).
The alternative is non-sequential injection, often called banked injection where all the injectors in a multi-cylinder engine fire together. This works well enough in many applications but doesn't offer as much control.


 

The injector is angled into the air stream, usually at about 30-40 degrees, just downstream of the throttle butterfly (here fully open)

While the injectors themselves are off-the-shelf items produced by the specialist industry, the angle they're positioned at in the inlet tract comes down as much to experience and experiment as pure science. The problem is that the injector has to squirt fuel across the air flow at an angle, from the side of the inlet tract. With a fast air flow it is rapidly carried downstream, atomising from a fine spray into a partial vapour (ideal for efficient combustion) on the way. But at lower throttle openings the air flow is reduced, leading to a problem known as 'wall-wetting' where some of the fuel is squirted across to the other side of the inlet tract where it forms a puddle. This starts to vaporise while more fuel is being injected, upsetting the carefully calculated mixture values.
Several manufacturers deal with this by using flapper valves in the airboxes, which restrict the cross-sectional area the air can flow through, in turn speeding it up through the inlet tract. Suzuki has taken this a stage further with its new GSX-R750 by adding a second electronically controlled butterfly to each inlet tract, which closes down as the airflow falls, maintaining the speed of air past the injectors. As a result, it has been able to increase the angle of the injectors from 32 degrees to 60 degrees (which improves atomisation) without inducing wall-wetting problems.


 

A typical bank of four fuel injector intakes, these on a Triumph TT600. Although more compact than carburettors, they demand more peripheral equipment and high electrical loads




So why no fuel-injected grand prix bikes?

Grand prix racing machines - the pinnacle of motorcycle high technology, massive power outputs from featherweight, incredibly compact engines, exotic space age materials, the Blue Riband class, a showcase for motorcycle factories' ultimate engineering abilities. And every one of them uses plain old carburettors.
So why not fuel injection, with its computer-controlled ultra-precise fuelling, full engine management control, infinitely adjustable mixture mapping systems and all the advanced technology you'd expect to see on the world's fastest racing bikes?
Most World Superbike machines after all are fuel-injected, so clearly it works for them, and they're only based on road bikes...
Well, there's the clue. The common perception of fuel injection is that it's a performance enhancing feature - it costs more money, computers are involved, therefore it should make an engine produce more power.
But it's not as simple as that. In fact, fuel injection on its own is rarely responsible for any significant increase in an engine's power. What it does instead is allow an engine to produce the same power as a carburettor version, but with reduced emissions. Or to put this another way, where the power of an engine with carburettors might have to be artificially reduced in order to meet emissions regulations, fuel injection allows it to reach its full potential while remaining within the regulations, making it seem as if it's responsible for extra power.
World Superbikes use fuel injection (except for the ageing Kawasaki ZX-7R) because the rules specify that their fuelling systems must be essentially the same as the road bikes they're based on. And the road bikes have fuel injection because they need it to pass the increasingly strict exhaust emissions laws around the world.
Grand prix bikes have no such artificial limitations, and there are plenty of technical reasons why there's little incentive for them to switch to fuel injection. For a start, carburettors are extremely good at what they do anyway. The fuel they supply is determined by the air flow through them sucking it up through the various jets, and at times that air flow can display some very subtle and complex changes. So much so that matching the fuel to the air flow via a computer map as in fuel injection can sometimes be beyond even the most sophisticated system, whereas in a carburettor the those tiny changes in air flow automatically mean tiny changes in the amount of fuel sucked into it.
Carburettors also have a happy habit of dumping extra fuel into the air flow when the throttle is snapped open, which by chance is exactly what's needed for a sharp but controllable throttle response. Fuel injection manufacturers and programmers still have difficulty in mimicking this behaviour (or at least, doing so without ruining their clean emissions), so much so that a general characteristic of most fuel injection systems is a throttle response which riders find too sharp and sudden, especially when opening the throttle again as they're about to exit a high speed corner.
Nevertheless, Honda experimented with fuel injection on its 500cc grand prix bikes for several years since 1993, but more than anything it was this problem with throttle response which finally caused the Japanese to give up and return to carburettors.
Another point is that the emissions factor with racing bikes is irrelevant, and as one fuel injection manufacturer commented, it's a lost cause with two-strokes anyway. New technology such as Aprilia's Orbital-patented DITECH direct injection system promises to overcome this, but that's still in its early stages and operates on a quite different principle to the conventional fuel injection systems we're familiar with. We might see DITECH-fitted Aprilia GP bikes, but there's still a way to go with this.
There's no reason why a two-stroke can't be fuel injected in the same way as a four-stroke, but the inherent high exhaust emissions aren't significantly reduced. Also, because the fuel-air mixture in a two-stroke has a much longer inlet path its atomisation (where the fuel droplets separate out for better efficiency) is better than a four-stroke's, so the improved atomisation which comes with fuel injection is much less noticeable.
Some companies have experimented with placing the fuel injectors in the transfer ports instead of where the carburettors would normally sit in the intake tract, but they found little advantage, and they also had to cope with lubricating the bottom end of the engine without mixing oil in with the fuel.
There are additional physical problems. A fuel injection system needs a high pressure pump and relatively large computer to operate, and these add weight to a machine compared with its carburettor-equipped equivalent. And this is important on a grand prix bike, where every gramme counts. Fuel injection systems also have high electrical demands, typically of the order of 4 amps or so with higher peaks. Grand prix bikes have no alternators and rely on total loss battery-powered electrical systems which aren't capable of running fuel injection. Either unacceptably bulky and heavy batteries would be needed, or the engines would have to be redesigned with alternators, which mean more weight and slower throttle response, as well as additional loads on already enormously stressed crankshafts.
In the light of all this, the only surprise is that Honda ever bothered to try out fuel injected grand prix two-strokes in the first place!