This article was first published in 2002.
Last week we covered the growing use of non-linear steering systems "The New Breed of Controls - Part 1" - designs that can radically change ratio over just 90 degrees of steering input. This week we turn our focus to variable ratio throttle controls, an area which has the potential to make a huge impact in performance cars - especially, modified performance cars.
Back in the days of carbies, progressive throttle linkages were common. Often they were used to sequentially operate the second throat, with effectively two (or more) butterflies opening to allow entrance of the combustion air.
However, with the advent of EFI, the most common approach now uses a single large butterfly, with its rotational angle directly proportional to the movement of the driver's right foot. For example, if the throttle butterfly opens 80 degrees at maximum (a common figure), and the driver's foot moves the throttle down a maximum of 50mm to achieve that opening, then for each 1mm of throttle pedal deflection, the shaft of the throttle body turns 1.6 degrees.
Of course there are some exceptions to this - eg engines that use multiple throttle bodies triggered in sequence and others that use variable ratio mechanical throttles - but the majority of EFI cars have simple systems set up in this way.
So that's a clear example of a fixed ratio control system, isn't it? Well, no it isn't. And the reasons it's not open up huge possibilities for making high powered and/or modified cars much easier and more rewarding to drive.
Those of you who drive turbo cars with well set-up anti-wastegate creep boost controls will know this one already - the flow of air through a throttle butterfly is not at all proportional to its opening angle. In a turbo car of this sort, it's not uncommon to be able to get full boost - and nearly full performance - on just over half full throttle. The turbo whizzes up, it jams the air past the partial restriction, and the engine goes hard. In fact, when you actually do put your foot right down, there's often not a great change - especially around peak torque.
And turbo cars also make a good example of the other end of the performance spectrum. Before the turbo spins up to speed, it's very easy to have your foot flat to the floor, going pretty well nowhere. Then, when boost arrives, you pull your foot right back lest you go faster than you actually wanted to.
This 'elastic throttle' scenario becomes so commonplace to a turbo car driver that often they're not even aware that their ankle is moving around so much - lots of throttle to get away from a standstill and then lifting it off mightily as boost arrives. But it is a very clear example of how the throttle - despite being connected by a linear mechanism to your right foot - acts in a completely non-linear way.
And while it is not as pronounced, even in a naturally aspirated engine, the airflow past the throttle is not in a linear relationship with its opening angle. This diagram from Bosch shows that the flow varies both with the opening angle and also engine rpm. In this typical case, with the throttle fully open (green line), airflow into the cylinder each intake stroke remains largely constant as rpm increases, dropping away a little as volumetric efficiency starts to decline from about half max rpm onwards. With the throttle valve fully closed, plenty of air is available each intake stroke at low rpm, flattening off as revs rise.
But the most interesting curves are those dotted lines that show in-between throttle positions. For example, the blue line shows the throttle open about two-thirds. As can be seen, in the lower half of the rev range the air-charge admitted the cylinder each intake stroke is very substantial, starting to really only fall away as revs climb. So with the throttle in this position, lower rpm performance isn't much different to full throttle! Furthermore, changing the throttle angle between full throttle and two-thirds throttle makes an increasingly large difference to the air available for the intake stroke as revs rise.
And then, if this airflow variation with different throttle openings isn't complicated enough, the actual torque developed at the tyres for a given throttle opening will be heavily dependent on the torque curve of the engine - as we have seen, even more importantly in a turbo engine. Not to mention that the torque at the tyres will also depend on the gear that you're in, and whether you have a torque converter in there as well...
So if ever there were a case for non-linear control, it has to be in throttle operation.
And before you say, "Yeah, so what - the driver quickly adapts to the non-linearity in the power development, anyway," consider again the case of an engine with a torque curve that's very peaky - and even with a shape (not height, but shape) that depends on throttle position. In a turbo car, if you accelerate at a constant throttle position, the shape of the torque curve that the engine develops will look very different from the one that happens if you ram your foot down hard just off idle, because the turbo will spool up in a different ways.
Not to mention that peak boost will often depend on what gear you are in - and so how heavily loaded-up the engine is...
It's for this reason that many expensive turbo cars these days run electronic throttles - the response can be made so much more linear by the use of a control system that is completely non-linear!
And the advantages are far greater than providing the driver with safer, more pleasant control. A throttle that is variable in its opening relationship to accelerator pedal movement can also provide switchable modes that can variably reduce wheelspin on slippery surfaces, dramatically improve fuel consumption, or give very sharp throttle response. In addition, cruise control and traction control are far better incorporated.
As with the discussion of steering weight last week, actually driving a car where the throttle/butterfly relationship can be varied by a dashboard control very clearly shows how effective such an approach is in changing the whole character of the car. In my own 1998 Lexus LS400, setting the electronic throttle control to 'Snow' mode immediately alters the car to being an astonishingly economical, slow and smooth machine. Flick back to 'Power' mode and the petrol consumption can literally rise by 30 percent - but so does the fun quotient. I'll come back to the Lexus again shortly.
In heavily modified, peaky cars, variable ratio throttle control presents the possibility of as great a change in driveability as was achieved in turbo cars with the move from carbies to electronic fuel injection. (If you're not old enough to remember, basically in all pre-EFI modified turbo cars making decent power, you had to feather the throttle if loading up the engine in higher gears, or you'd inevitably bog as the engine ran too rich or too lean. Flat spots were pretty well unavoidable - the only exceptions were the handful of well-developed blow-through carby models.)
Electronic Throttle Control
The easiest way in a new car to gain a non-linear variable relationship between accelerator and throttle blade position is to adopt electronic throttle control. We have already covered electronic throttle control in two major stories - the most recent at "Electronic Throttle Control Advances" - so here we'll make just a broadbrush outline of one system.
The Lexus electronic throttle control (ETC) system used in my car is shown here. It varies from some other ETC systems in that a throttle cable still connects the throttle pedal to the throttle body. However, the blade itself is operated by an electric motor unless a fault develops, whereupon a magnetic clutch disconnects the motor and a "limp mode lever" allows direct driver control.
The construction of the throttle actuator motor and its relationship to the throttle blade and throttle position sensor can be seen here.
But what actually happens when using the throttle, especially in this 4-litre V8 engine which has a relatively flat torque curve? Or does it? With the infinitely variable valve timing of the inlet camshaft and a changeover manifold, the torque curve is actually peakier than you might expect. So one of the functions of the ETC is to smooth out acceleration.
As can be seen in these diagrams, if the driver pushes down fully on the throttle, the actual throttle blade opening rate slows for a moment before it reaches fully wide open (and the ignition timing is also pulled back a little in its rate of advance), decreasing the jerk that would otherwise occur. At the other end of things, when the accelerator pedal is released suddenly, the transition to deceleration is smoothed.
The latter is particularly noticeable when I switch off the cruise control at high speed without my foot being on the throttle - there is almost no jerk. In this situation it's almost like a dash-pot function - although in more normal lift-offs (for say a corner) thankfully it can't be felt to be operating.
As indicated above, probably the most intriguing action of the ETC throttle is in Snow mode. Designed - as the name suggests - for slippery surfaces, the throttle mode limits the instantaneous torque that can be applied. So if you jerkily depress the throttle, there's no sudden acceleration; instead the throttle blade smoothly and slowly opens. Simply, longitudinal acceleration is always limited to a predetermined value. And while for many performance enthusiasts that wouldn't be much fun, exactly the same approach can be taken to putting down maximum power in a car that would otherwise wheelspin on dry roads.
Ultimate Performance Parameters
We've discussed the non-linear relationship between a conventional throttle linkage and the torque that is being developed by the engine. But, given that literally any sort of relationship can be gained by electronic throttle control, what should be strived for? Part of the Lexus approach we've just seen - but what about in an overtly sporting car?
The importance of this area of performance car development can be seen when a tale about Ayrten Senna is told. In the turbo F1 era, the racecar driver caused much concern to his Honda engineers with his throttle log trace - he was continuously on and off the pedal when cornering. However, there was a simple explanation - Senna was pulsing the throttle in order to keep the engine on boost without developing excessive power for the cornering situation. Even more fascinatingly, when the first naturally aspirated Honda V10 3.5-litre was produced, what has been termed "a fantastically complex throttle linkage" was used to give the engine the progressive throttle characteristics that he desired.
So those of you who think that we have been overstating the driving gains available in this area of high performance car control should take on board the critical importance attached to the relationship between the throttle and blade opening by one of the best drivers ever.
For a sporting car there are four main potential options that can be taken in determining the relationship between throttle pedal movement and the car's performance.
- Linearising as much as possible the throttle response in respect to engine torque, so that a given movement of the throttle results in a proportionate increase in engine torque (ie irrespective of engine rpm, boost, and the shape of the engine's torque curve). This potentially gives the driving sensation of having an engine with a completely flat torque curve - a little like driving an electric car - although of course such an approach cannot produce torque where even at full throttle, the engine has little. However, in taking this approach, the throttle/acceleration relationship will still vary in different gears.
- Linearising as much as possible the throttle response in respect to wheel torque, so that a given movement of the throttle results in a proportionate increase in torque at the tyres (in addition to the engine torque development, this also takes into account the gear ratios and action of the torque converter, if one is fitted). In this case, the relationship between the throttle movement and the acceleration of the car is as constant as possible.
- Throttle opening in proportion to the requested speed differential. Here the torque delivered is dependent on the speed at which the car is travelling, in addition to the throttle angle requested. This relationship can be programmed to allow, for example, larger throttle openings at high speeds, given that the torque actually passing through the tyres is reduced due to the higher gear selected.
- Throttle opening in proportion to the tyre slip ratio, where the amount of torque resulting from a given throttle movement is determined by the slip angle of the powered tyres. This is similar to traction control, where a degree of wheelspin in a straightline is permitted in order that the fastest acceleration be realised. However, in this system, the cornering slip angles are also used as part of the relationship between requested and actual throttle opening.
Mechanical & Electro-Mechanical Linkages
The above four approaches can all be accomplished using sophisticated Electronic Throttle Control strategies. But what is sometimes forgotten is that a variable throttle relationship can also be gained, even when there is a direct mechanical connection.
Some EFI cars use systems of levers and pins travelling within curved slots, so that the throttle opens more rapidly as greater pedal movement is used. This has a twofold advantage:
A sensitive throttle pedal can be retained at small openings when lots of control is needed (the driver is less likely to be using sensitive modulation of the throttle at 70 or 80 per cent opening) while not making the pedal travel excessive.
The last part of the throttle opening - where often only a small amount of airflow variation occurs anyway - is reduced in sensitivity.
The 2.5-litre Mazda/Ford V6 pictured above is an engine that - at least in some forms - used this sort of sophisticated throttle linkage.
We have also seen variable throttle linkages used on mechanically injected Mercedes V8 engines. Some other engines use a curved bracket that - as the butterfly opens - continuously changes the pulling distance of the throttle cable from the shaft. This approach gives an increasing rate of throttle opening.
There is enormous untapped aftermarket potential in this area. A turbocharged car that is very sluggish off boost could relatively easily have a throttle linkage constructed for it that used a moving pivot point, with the pivot altered in position by a miniature pneumatic cylinder. As the car came up on boost - pressurising the cylinder - the pivot could be moved, reducing the throttle sensitivity. In this type of arrangement, pushing down on the accelerator pedal to (say) the halfway point in an off-boost situation could in fact open the throttle to three-quarters, while the same throttle movement on boost would open it only half way. This would transform the driveability of the car.
As opposed to a purely electronic throttle control system, such a mechanically variable system could be easily engineered and also retain the straightforward safety spring return of conventional throttles.
Alternatively, it would also be a relatively simple engineering exercise to use an electric motor to turn a leadscrew, again moving a pivot point within the linkage system to alter the throttle ratio. Even if relatively slow in action, this could still be operated from a dashboard control to give, say, excellent wet weather driveability in a high-powered two wheel drive turbo car.
The aftermarket is crying out for throttle systems that start to make use of the technology that's been used in OE cars for more than 15 years.
Apart from seeing different ratio throttle lever arms sold for a Porsche model, we have not seen any specific aftermarket products designed to improve the throttle control of EFI vehicles. (Incidentally, one reason that many people talk about a response improvement after swapping to a larger throttle body is that effectively they have changed the airflow relationship of the throttle body with accelerator pedal position - more air can get through at a smaller throttle opening, and so the car seems to get up and go faster.)
The market is wide open for products to improve throttle control - from rising rate throttle brackets that increase the opening speed of the throttle in turbo cars equipped with larger turbos (and also in NA engines with modified camshafts), right through to a fully-developed aftermarket programmable Electronic Throttle Control System. The technology of the latter is now well known for OE cars, and the magnitude of this breakthrough in the aftermarket would literally allow cars so heavily modified that they are currently nearly undriveable to be pedalled with relative ease.
A mid-point in technological terms would be a mechanical system with electrical or pneumatic ratio adjustment - again the benefits in driveability and control are likely to be very substantial.