A while ago in a reply to another blog post, I wrote about the current Lancer Evolution that:

“The Evo should use far improved throttle mapping where blade angle is mapped against foot position and the calculated instantaneous tractive effort value. It should also use a smaller turbo. ”

At least one reader was so excited by this notion that he wished to “quietly roll up into a foetal position and rock back and forth on the floor”. However, leaving aside bizarre responses, it’s a concept sure to interest some.

I won’t discuss the ‘smaller turbo’ bit because most of you will have a good understanding of this idea. But what about the throttle mapping?

In electronic throttle cars, the relationship between the accelerator pedal position and the throttle blade opening no longer needs to be linear. In a linear system, the throttle blade would be half open at 50 per cent accelerator pedal travel, three-quarters open at 75 per cent accelerator pedal travel, and so on.

(In fact, even many mechanically operated throttles don’t take this approach, using cams and/or levers to open the throttle more quickly as full accelerator pedal travel is reached. And if you want to add another complexity, airflow past a throttle blade is not proprtonal to its angle of opening!)

In an electronic throttle car, any type of relationship between accelerator pedal position and throttle blade angle can be used. In other words, the software can dictate what actually occurs when you put your foot down.

However, having just a two parameter table (accelerator position versus throttle blade opening angle) is much less effective than taking other options.

For example, what about mapping also against engine rpm?

By taking into account engine speed, the actual torque output of the engine can be largely linearised. After all, putting your foot down 10mm at peak torque (eg 3000 rpm) gives a different response from putting your foot down 10mm at (say) 1200 rpm.

By altering the amount the throttle blade opens (at all openings except when full power is requested), the system can make the response to the throttle much more as the driver expects.

But then again, the actual result (eg in terms of acceleration) also depends on what gear the car is in. So the system really needs to know the gear ratio being used. The easiest way of expressing the engine’s torque output and the gear ratio is to say ‘tractive effort’, that is, the backwards push the tyres are making on the road.

So using “throttle mapping where blade angle is mapped against foot position and the calculated instantaneous tractive effort value” is clearly a good way of doing things. But how do you calculate the instantaneous torque output of the engine – especially if it’s a turbo engine where torque also depends on developed boost? The easiest way is to have a map that relates intake mass airflow to torque – the map would be specific to each engine.

These ideas are not new – quite a lot of engines use torque modelling as a major input into electronic throttle control. But it was in driving the Lancer that it struck me anew how such an approach could really radically improve that car.

And if this all seems too complex, a very simple way of gaining a great ‘linearising’ improvement in turbo engines like this is to control turbo boost pressure by accelerator position. The Independent Electronic Boost Control (Part 1 and Part 2) does this very well.