Features - Technical
OCTOBER 17, 2001
The 32-bit throttle cable
BY PETER WRIGHT
Traction control is permitted in Formula 1 now and it looks as if it is here to stay. Just about all the dire predictions about it ruining the sport have proved unfounded, but the prediction that it would stop all the carping about cheating and unfair advantages has turned out to be absolutely spot-on.
Two of the erroneous predictions are especially interesting: "It will make no difference in Qualifying" -correct; but it will have the most effect in the race, because tyre wear will be less" - wrong - it is greater. The second is: "The drivers will no longer slide the cars". They do. To look for the reasons that these predictions have turned out to be so far off the mark, one must delve into the control laws encoded into the 32-bit processors that have replaced the throttle cable, and look closely at what a racing driver is trying to do.
It has taken the competing tyre companies a few years to develop grooved, 15in wide, rear tyres that can handle an 800+bhp Formula 1 car; the tyres are basically too small. However, this year it seems that they have got it right, ably assisted by their chassis partners, and in the early races in the year the drivers started to display tail-out driving styles again - before traction control was legalised at the fifth race in Spain. Journalists and spectators extolled the exploits of the more adventurous drivers and deplored the coming final demise of oversteer. Two questions puzzled me though: if tail-out oversteer is the quickest way through a corner, why would traction control be set up to prevent it? And, if it was not the quickest technique and drivers only drive that way for fun, then why do their employers not take them into their motor homes and explain to them that they are paid to drive fast and not have fun, even if it means less entertainment for spectators?
The answer to this question lies in the term "traction control". The cars may all now be fitted with traction control, but we have almost no idea what the encoded laws are that are used to control traction. Classic traction control compares the speed (i.e. the RPM) of the driven wheels with the speed of the car (i.e. the RPM of the undriven wheels, corrected for differences in tyre diameter). The control laws permit a certain amount of slip (i.e. a percentage that the driven wheels are faster than the undriven wheels), based on the slip ratio at which the tyres provide maximum traction. The next level of sophistication is to determine whether and how hard the car is cornering, and modify the maximum slip ratio accordingly, in order to provide priority to stability and control rather than traction. That may be great for road cars, helping to keep inattentive drivers from spinning when road conditions of rain, ice or snow catch them unawares or with insufficient skill to maintain control. However, most road car manufacturers who supply traction control systems provide an override switch to enable the skilled, sporting or unwise driver to exploit power oversteer.
Such an approach does not make sense for a racing car, as Jean Alesi has so valuably stated, throwing his override switch with a grand Gallic gesture. There are other approaches to traction control that can leave the driver with more control over the car, and bring other advantages, such as more power. When did a driver ever turn down more power? The increased tyre wear experienced by some cars tends to indicate that other options are being explored.
In the "good old days", engine throttles were controlled via a cable and linkages, connected to the accelerator pedal under the driver's right foot. This mechanism modulated the air flowing into the engine, and hence the torque produced by the engine. The relationship between the driver's foot and the torque was complex and highly non-linear; in particular, it depended on the RPM of the engine. If the driver maintained a constant throttle opening while accelerating out of a corner, as the RPM rose the torque delivered to the rear wheels changed according to the torque curve of the engine. If the torque curve had peaks and troughs in it, and the driver was trying to balance the car with torque at the rear tyres, the balance would change and upset the car's trajectory, unless he altered the throttle position. Engine designers tried to produce smooth torque curves, at a cost to maximum power, while the chassis designer tried to design a progressive throttle linkage such that the gain (ratio between pedal and throttles) was lower at small throttle openings, and higher when the driver wanted maximum power and pushed the pedal to the metal.
Turbo-charged engines suffer from sudden increases in torque as the turbo spools up to its design point when the driver accelerates. Ayrton Senna perplexed the Honda engineers with the data traces of his throttle pedal movements. In corners he was continuously on and off the throttle at a high frequency, making the data look as if it was faulty. He explained that that was exactly what he was doing to keep the turbo spinning and overcome the lag, without having excessive power delivered to the rear wheels. In control terms, he was "pulse-width modulating" the throttle i.e. opening and closing it fully, but using the ratio of the time it was open to the time it was closed to control the torque of the engine. When Honda produced their first 3.5l V-10 NA engine, Senna worked closely with their engineers to develop a fantastically complex throttle linkage, mounted at the rear of the engine, to assist giving the engine the progression characteristics he wanted. Later, Honda was the first to go to drive-by wire, making the need for complex linkages redundant. When the FIA banned driver aids for 1994, the engine manufacturers fought hard to keep drive-by-wire, as it so simplifies the task of making the engine progressive.
It was the further efforts of some teams to assist the driver in controlling the ever-increasing power of the latest engines that led to the accusations of rule bending. Every trick in the book, using fuel injection and ignition timing to modulate torque, was explored in the quest for progression. Whether the slip ratio of the rear wheels was calculated in some way or other, and used to control the engine output, is the issue that determines whether the techniques contravened the regulations. Now traction control is permitted, and indeed all methods of controlling the engine are free, there is no need to compromise and the ideal technique can be researched and developed.
It is the driver's use of the throttle that causes him to spin the wheels or power oversteer, but the engine control system does not use the throttles to control the torque, unless the driver persists in calling for too much for a significant length of time. Instead, fuel injection and ignition are used for instant modulation, being capable of controlling the torque even as the car rides bumps that may cause the wheels to break traction - hence the awful sound they make as traction control cuts cylinders in turn. Misfiring is pretty destructive to a racing engine, as one or two crankshafts left in the road have testified, and preventing too much slip in the first place is a priority. In control system terms, the driver makes an open-loop demand of torque, and the fuel injection/ignition system makes closed-loop adjustments to wheel slip. The driver may or may not make a change to his input, based on feedback of the sound the engine is making. If not, the throttles may be altered by the system to reduce misfiring. If the driver were able to make a better open-loop, initial input, the misfiring could be minimised.
There are a number of options, the four main ones being:
throttle pedal as an engine torque demand;
throttle pedal as a rear wheel torque demand;
throttle pedal as a speed demand;
throttle pedal as a slip ratio command.
As we have seen, modulating the airflow into the engine is not the ideal way of controlling torque. If the fully mapped characteristics of the engine are known, taking into account RPM, throttle opening, air density (including ram effects), engine temperature, fuel temperature etc., the throttles can be controlled by the engine computer to deliver torque proportional to accelerator pedal position. If the engine cannot deliver as much as the driver is demanding (due for instance to too low RPM) it just goes to maximum throttle. This is the most progressive engine, as it takes out all the dips and bumps in the torque curve, and means that the engine designer can tune for maximum power. The second option is a variation on this, with the gear ratio being used to modify the throttle setting so that the rear wheel torque is controlled directly. With the ratio selection now also under full computer control, the optimum ratio for acceleration can be selected instead of a higher than optimum ratio, which drivers often use to reduce the response of the engine to throttle changes and thus facilitate control of the car.
The latter two approaches give the driver a smooth change in rear wheel torque as he moves the accelerator, greatly facilitating control of the car coming out of a corner. However the maximum torque that the rear wheels will accept without wheel spin or oversteer, changes as both the cornering load reduces and as the speed and aerodynamic load increases. Wheel spin or flick-oversteer will also be induced over bumps as constant torque is applied to the driven wheels. The third option gets round this problem by maintaining a constant wheel speed (and hence car speed) selected by the accelerator pedal position, with the car accelerating up to the selected speed at a rate proportional to the error between the actual speed and the demanded speed, the rate depending on the torque delivered. Full pedal gives full power, unless maximum speed is limited, as in the pit lane. In a corner, pushing on the accelerator pedal from a constant speed setting would increase power to increase speed, but would filter out engine characteristics and be reasonably benign over bumps. Drivers would have to adapt to this rather different approach. Better still would be the fourth option, which combines the best of the others.
Controlling rear wheel slip ratio provides the driver with the means of either demanding the maximum acceleration the car can deliver (full pedal deflection), either limited by the traction available or the power available, or a means of steering the car under power by varying the slip ratio, and thus the location on the friction ellipse of the tyres at which they are operating. The slip ratio will be set as required by the driver's pedal, irrespective of engine characteristics, speed, or rear axle load (aerodynamic, load transfer or bump induced). He can still try and negotiate a corner too fast, and lose control, but under all conditions in which power is required, he will be operating at the maximum and still be able to steer the car on the "throttle", though now he has virtually no control over the engine throttles. Because he will nearly always be at the maximum slip ratio, tyre wear will become greater.
With any of these approaches to controlling the power, the driver can still slide the car and hang the tail out, but he can now do it with far greater impunity, without worrying about peaky engines, bumps etc. The engine manufacturers can tune for power, and the chassis engineers can perhaps set up the cars for more grip without worrying so much about stability. But, "What about the skill?" I hear cried. It does not seem that traction control has changed the pecking order in Formula 1. Perhaps some drivers have adapted more quickly, or it better suits their driving style. Personally, I do not believe it will make any difference other than to encourage drivers to drive their cars nearer the limit for more of the time, and that cannot be bad for racing. Take the example of fly-by-wire fighter aircraft.
When Sukhoi produced their fly-by-wire, relaxed stability fighter, the Su27, (fly-by-wire equals "care-free flying", equals "no skill required"....) their test pilots discovered that the fly-by-wire control system allowed them to use parts of the flight envelope where they had previously not dared to venture. They developed super-manoeuvres to enable combat pilots to escape from pursuers, including the Tail-slide, the Cobra and the Hook. Western fighter pilots and manufacturers were stunned when Viktor Pugachov demonstrated these manoeuvres at low level, at Farnborough in 1989, flying the Su27. Using computers to fly the aircraft did not spoil the entertainment one iota. Military pilots, just like racing drivers want the maximum performance from their mounts.
With numerous sensors, connectors and lines of code between the drivers right foot and the engine throttles, it is inevitable that reliability will suffer. Throttle cables were never the most reliable components of a racing car, and a broken or stuck throttle used to be a common cause of both retirements and crashes. Some teams fitted two throttle cables so that provided the car would still run, the spare could be quickly connected up in the pits and the car continue. Software engineers can ensure that all known failure modes result in a closed throttle, and even in the event of a stuck open throttle (usually caused by a hydro-mechanical malfunction) the computer is able to detect that full throttle and full brakes means something has gone wrong and shuts down the engine. Drivers still have a switch to disable traction control and revert to a fixed, albeit electronic link between throttle pedal and engine.
The "art" of synchronising throttle, clutch and gear lever position to change gear is a thing of the past. Also, Formula 1 drivers do not have to delicately tiptoe up and down the bumps and hollows of the torque curve as they accelerate out of corners. No points have ever been awarded for crispness and speed of gear changing, nor the artistry of controlling oversteer. Instead we have drivers using all the technology permitted to explore new and greater limits of performance, more of the time. Points are awarded to those who do that best.