Engine Control Theory

This is part one of a series I will be writing on engine control theory, tuning and hardware. Feel free to e-mail me questions about your bike or anything you would like to know through the MSR Google group e-mail and I will do my best to answer.

History

Back in the good (bad?) old days your bike engine was fed fuel through a sort of straw that worked based off the vacuum created by a venturi. You had a few different holes that would let more or less fuel through into the engine and the flow was roughly proportional to the airflow into the engine. All appeared well, even if the physics of venturi flow wasn’t understood by most. The idea of bigger jets equals more fuel, and blocked jets equal rough running, and so on, was easy to understand. Carburettors were never very precise, so in general bikes always ran a bit rich (excess fuel) to cover production variation and various temperature and altitude changes. This was good for throttle response but not so good for the environment or fuel economy, throwing raw hydrocarbons into the atmosphere.

As emissions regulations tightened for bikes, manufacturers were forced to adopt fuel injection. This also had the significant benefits of improving fuel consumption and increasing power to the engine through less intake restriction. But what happens inside the black box? And why do manufacturers tune the engines the way that they do? I hope to answer some of these questions in this article.

Note: I will cover the concepts of OEM ECUs here. Some aftermarket ECUs and Power Commander type devices aren’t very clever and often give misleading information on how engines work.

Basic Operating Concept:

Pretty much all fuel injection systems split up in the following way.

Part of the software calculates how much air is going to be trapped in the cylinder. For simple ECU’s (like motorbikes) this is often a table of engine speeds and loads (defined by throttle position, or manifold pressure, or both). And cells in the table are filled out with units of air per cylinder. The air per cylinder is predicted for the next engine cycle, since the fuel for that cycle must be delivered before the air is sucked into the cylinder. On cars, which mostly have a common manifold, (bikes have individual throttles), an airflow meter can be used and some complicated modelling of the airflow into and out of the intake manifold is performed. Okay, so now we know how much air is going in.

There is a table of ‘Target air to fuel ratio’. Often people speak of AFR or Lambda. In most manufacturer ECUs we express air to fuel ratio as EQR or Equivalence Ratio. This is much easier to understand as an EQR of 1 is always stochiometric (AFR 14.7:1 for petrol). Stochiometric means that the correct mass of fuel is injected so that in theory all the fuel will burn with all the air. In practice this doesn’t happen but let’s say on a warmed up engine, greater than 99% of all the fuel burns with the air.

An EQR of 1.2 would be 20% rich and 0.8 20% lean. The nice thing about EQR is that even if your fuel changes (running LPG, for example) the EQR remains the same. So you don’t have to change all your tables. The EQR is multiplied by just one number in the ECU that defines the stochiometric ratio for the fuel you are using (a duel fuel car might have two points or a blend table – like say petrol to E85). Say we had 0.3g of air predicted to be in a cylinder, then we need 0.3/14.7 = 0.0204g of fuel at EQR 1.

Now we know how much air is going in and how what air to fuel ratio we want. Next, we need to find out how long to open the injectors for. The fuel injectors have a table which will be the amount of fuel versus injector opening time and pressure across the injector. Older cars had a fuel pressure regulator at the fuel rail which referenced the engine manifold pressure and kept fuel pressure across the injector constant. Modern ECUs just measure the manifold pressure, compare it to atmospheric pressure, and then adjust the injector opening time appropriately. The fuel pressure regulator is in the tank and outputs a constant pressure. There is no requirement for a return line (saves dollars). These systems are called “returnless”.

The ECU manages the fuel injection timing. The fuel for the next cycle is injected before the intake valve opens. The idea is that the fuel hit’s the hot valve and vaporises, thus cooling the intake charge (thus making more power) and has time to mix evenly with the air in the intake runner. The timing is adjusted to maximise the cooling and mixing while minimising ‘puddling’ of the fuel in the intake where the drops of fuel stick to the walls.

At very high speeds there may not be enough time to inject all the fuel before the valve opens. Also there may not be enough resolution in the injector opening time to control both the low fuel flow required at idle, and the high fuel flow at peak power. For this reason some bikes have a second set of injectors above first set that work at high engine speed and load to deliver the required amount of fuel. These are often located further away from the valve to maximise the cooling effect at these high gas flow operating points.

Finally, the spark is triggered at the appropriate time to initiate the burning of the air fuel mixture. The spark is triggered as the cylinder is moving upwards. The fuel burns outward from the spark plug, usually in a swirl created by the airflow of the incoming mixture. The aim is to get peak pressure just after the cylinder has passed top dead centre. As long as the cylinder pressure and temperature do not get too high, knock, which is uncontrolled burning of the air fuel mixture will be avoided. Knock is more of a problem in large bore engines so racebikes with their generally small bores (except for cruisers) are not so susceptible. This allows for an increased compression ratio which helps gas exchange, efficiency and power.

More on spark in a future article.

Tuning Fuel:

Okay, so why are engines tuned the way they are? Is peak power restricted for emissions? Why do manufactures run their engines ‘lean’ at part throttle and why are they too rich at full throttle? There are answers to all these questions.

The first is that engineers don’t give away power and torque for ‘free’. There are always compromises. It’s helpful to understand the relationship between engine torque and EQR.

The graph on the left is taken from an actual engine at 100% throttle and 6000rpm.

You could make a maximum of 4% more engine torque at EQR about 1.15 or 15% rich compared to EQR 1. Note how rapidly the torque drops away as fuel goes lean and how even with 40% extra fuel you still make the same power as EQR 1.

Assuming the graph is also true for other engine speeds and loads (and it pretty much is). We should run the engine at 15% rich all the time right? Well there are some compromises.

At high speeds and loads the exhaust valves, spark plugs and catalyst can get very hot (>1000degC). Not only does this lead to valve recession, melted valves, damaged plugs and knock. The heat can trigger catalyst meltdown which is the main cause of ‘blocked cats’ you hear so much about on forums.

One way to cool combustion is to add extra fuel in, often as much as 45% rich. The vaporising fuel helps keep temperatures low. Spark can also be retarded but this shifts the heat from inside the cylinder thus preventing knock and protecting the spark plug to the exhaust where it melts the valves and catalyst. Titanium and sodium filled exhaust valves can handle higher temperatures and thus allow leaner EQR and more power.

Some more advanced ECUs calculate the temperature of the exhaust valve and catalyst and only add fuel when it is required for cooling. This allows maximum power for say 2 or 3 seconds which all you ever use on the street while still protecting the engine if you ever took the bike to the track. This feature also confuses dyne operators using Power Commanders to adjust fuel on the dyne and leads to a lot of engine failures.

Secondly, we will be using 15% more fuel to get 4% more torque. So straight away on say, a highway cruise, we will be using 11% more fuel than we need to. So we want to run EQR 1 on the highway.

Also, if we run 15% rich, our engine will be throwing out hydrocarbons (unburnt fuel) out the back which isn’t going to be good for emissions. Also, for reasons I will explain another time the catalyst only really works around EQR 1. So we need to run at EQR 1 to get best emissions. However, we only need to do this in the area where they test on the emissions cycle. (typically low loads, and speeds up to 120km/h). The graph below shows the Euro 3 cycle we use in Australia. Key Point: Emissions are not tested at 100% throttle or high engine speeds, so no compromise required at these points!

Euro 3 Emissions Cycle.

Finally, to save fuel and increase engine braking, we might cut the fuel off altogether when the throttle is closed. This can also help emissions as the engine can have poor combustion at very load loads and engine speeds. This shows up as popping on deceleration from the exhaust. Some manufactures actually engineer this popping in! (like in my Triumph and it’s off road exhaust and tune).

There is one other problem with running EQR close to 1: when the throttle is opened rapidly the bike can run lean as the fuel for the next cylinder is calculated from the event before. Also, there may be errors in the fuel delivery due to production variation and especially modifications (such as a pod filter). Even small lean dips drop a lot of torque so you can get hesitation, stumbles and jerks. Riders of early FI systems will know what I mean. For this reason, outside the emissions window, fuel is often added to help response.

The picture below shows what a typical AFR map might be for a motorcycle.

Maintaining EQR 1 is important for emissions but going lower than 1 has a big penalty for rideability. What we need is a system where the actual EQR can be measured and then the fuelling tables adjusted so that they remain close to 1 (good for emissions and economy) without going too lean (loss of torque). To do this we use an Oxygen (O2) sensor and closed loop control.

An O2 sensor measures the oxygen content in the exhaust. If the engine is running lean the unburned oxygen will be coming out the exhaust. When running rich only a very small amount of oxygen remains unburnt so the sensor voltage changes; see the picture below (lambda is 1/EQR).

When entering closed loop mode the engine uses the normal table’s described above to calculate how much fuel to put in and this value is trimmed by the O2 sensor. There is normally a quick learning control loop that rapidly adjusts fuel as the engine is running and in the background long term fuel trims are recorded and saved from one key cycle to another. This helps maintain the correct fuel trim even if the throttle is opened quickly and the engine jumps from one load point to another.

So there you have it.

Next month I will discuss why all fuels aren’t created equal and discuss ignition timing and torque in more detail.

Henry Wright (who works as an engine calibrator for GM Holden)