Design for cornering
Joseph Lowrey, B.Sc.Eng, writes of some elusive characteristics in reply to certain Editorial remarks
In the September issue of Motor Sport the Editor put forward certain views on the subject of cornering qualities. While agreeing with his Statement that “a whole lot remains to be written” on this matter, I cannot find equal agreement for his other comments.
My first criticism of his interesting, even if controversial, remarks on the subject of cornering qualities concerns the statement that his Lancia “Lambda” “has vintage stiffness about its construction.”
Here, I would suggest, the Editor has gone right off the rails. It may be that the “Lambda,” with its stressed-skin construction, is reasonably rigid in spite of its open body; but to describe rigid construction as a “vintage” quality is grossly misleading. The basic idea of “vintage” designers in their search for good road-holding seems to have been to use stiff springs with heavy friction damping, in conjunction with a frame having considerable flexibility, particularly in torsion.
Provided the weight distribution was reasonably satisfactory, this scheme worked exceedingly well over most road surfaces. Occasional really bad bumps were dealt with by the springs, while normal road irregularities were apparently dealt with by the tyres and by torsional flexing of the frame.
In contrast, the modern car, with a steel saloon body rigidly secured to a box section chassis, is really rigid, in both torsion and bending. Vauxhall Motors, Ltd., have released some interesting figures for their cars, showing that between 1933 and 1938 the torsional rigidity of typical complete cars of their construction was roughly quadrupled; chassis rigidity without body was increased tenfold, and this is a particularly relevant figure, since the open body of the vintage sports car did little to add to chassis rigidity.
In conjunction with this vastly increased chassis rigidity, the “modern” vehicle uses flexible springs, relatively light hydraulic damping, low pressure tyres and probably independent suspension at front and/or rear. The use of extremely large section tyres running at very low pressures is a distinctly questionable step, so far as road-holding and cornering are concerned. It is true that the pneumatic tyre is the only form of suspension having negligible unsprung weight, a considerable aid to good road-holding (taking this in its strict sense as meaning grip between wheel and road), and is desirable as such. But, on the other hand, one can scarcely hope to obtain really accurate steering when the final link between steering wheel and road is 6 in. of wind wrapped up in a rubber bag: when subjected to a sideways load, as in cornering, a modern pneumatic tyre can move in a direction very different to that in which it is pointed, purely owing to flexing of the tyre and without any skidding taking place.
My personal view on the subject of tyre sizes is that we have gone rather too far in recent years. With modern suspension systems it should be possible to obtain comfort and road-holding with quite modest tyre sections, and to improve steering accuracy accordingly. As it happens, a “scorched earth” policy in the rubber plantations will probably result in reduced tyre sizes being forced on designers willy-nilly, and this may well prove a blessing in disguise. The popularisation of independent wheel springing has, in general, been a good thing, particularly as regards the front wheels. It should reduce unsprung weight, reduce or eliminate steering disturbances caused by incorrect geometry or by gyroscopic effects, and give stable steering under conditions of heavy braking. It should also provide good cornering without undue roll. In this paragraph, the important word is “should.” A good i.f.s. system is better than a good front axle, but a good front axle is very much better than a bad i.f.s. system.
As far as reduction in unsprung weight is concerned, most systems do improve on the orthodox rigid axle. But, as a rule, such a large proportion of the unsprung weight is represented by tyres, wheels and brakes that the saving over a tubular axle with quarter elliptic or cantilever springs is quite small. The only way to reduce unsprung weight drastically is to use smaller wheels and tyres, and to mount the brakes on the chassis as on certain front wheel drive cars.
The provision of steady steering with i.f.s. is largely a matter of geometry, in keeping the front wheels upright (to avoid gyroscopic effects) and in providing steering linkages suitable to the layout chosen. In practice, these two objectives are seldom quite achieved, a certain degree of imperfection being accepted in return for such benefits as reduced tyre wear or simpler construction.
If necessary, quite considerable wheel tilting can be “got away with” on cars used at normal road speeds; the best example of this lies in the l.m.b. layout, which works well in spite of gyroscopic effects equal in magnitude to those encountered with a rigid front axle. The more such effects can be reduced, however, the more high geared and frictionless the steering mechanism can be.
The question of correct steering geometry is largely bound up with that of complication, even where “cost is no object.” I have vivid memories of a certain demonstration car, one of the most costly sports cars made in Britain in 1939: the makers had got their geometry right, by using about a dozen ball joints in their steering linkage, and even on a new car the cumulative effect of a little play in each joint was to completely ruin the accuracy of the steering. Steering instability under heavy braking conditions is a fault of many rigid axle layouts: this is generally the result of half elliptic springs twisting and reversing the normal castor action of the steering. Any reasonable i.f.s. layout should be satisfactory in this respect, as are the better rigid-axle cars, although there is scope for cunning design in reducing the tendency for the nose of the car to dip under cornering or braking conditions.
The hydraulic shock absorber, which has completely supplanted the friction type on modern racing machines, has done much to improve road-holding. With the rigid chassis necessary when i.f.s. is used, free spring movement is necessary, to accommodate the slow rise and fall of the wheels over what may best be described as long wavelength road irregularities. Where the “vintage” machine dealt with these movements mainly by weaving of the frame, the “modern,” with hydraulically damped flexible suspension, does the obvious thing and deal with them by movement of the springs. With friction shock absorbers any attempt to provide this free spring action would result in hopelessly inadequate damping at speed over rough surfaces: a suitable hydraulic damper will provide free spring action at low speeds, stiffening up for large or rapid wheel movements as is desirable.
I seem to have wandered some way from the original subject of cornering, ancient and modern, but all the detail differences in chassis design which have been adopted over the last decade, in the interests of either comfort or roadworthiness, have their effect on cornering.
Fundamentally, the speed at which a car can negotiate a corner depends simply on the available coefficient of friction between the tyres and the road, so long as the centre of gravity of the vehicle is low enough to rule out any risk of overturning.
On this basis, modern developments in suspension systems, which when properly applied can definitely improve wheel adhesion on imperfect road surfaces, should result in improved cornering. In some cases this improved cornering is realised, but in others it is not, since “complications set in” and prevent all the available wheel-grip being used.
The trouble is that, on the road, a car must not only be capable of taking a curve of a particular radius at a certain maximum speed, but it must also be readily controllable by the driver when doing so. Driving skill enters into the matter, of course, but if the driver is unable to direct the car accurately along the best path into and around the corner, cornering will suffer. Any divergence from the “best” path inevitably involves turning more sharply than is really necessary, with an appropriate reduction in speed.
It is on this question of controllability that quite a number of “moderns” seem to fail: they have good wheel adhesion, yet on the road they cannot negotiate a particular corner as fast as a “vintage” car with more accurate steering.
Faults such as steering having excessive lost movement, or too low a ratio, are purely mechanical and may be common to vehicles of any age or style. The more subtle vices which can affect cornering are “over-steering,” “under-steering” and sluggishness in response to sudden movements of the steering wheel; or, of course, combinations of these vices. (This matter of under- and over-steering is nicely clarified in Pitman’s “Elements of-Automobile Engineering,” by Maurice Platt.)
Over-steering is a fairly accurate description of a tendency for a car, when steered into the corner, to turn ever more sharply; you feel that the tail is drifting outwards and, after steering into the corner, you must turn the wheel back towards or even beyond the straight ahead position, in order to maintain a steady course.
Under-steering is the opposite effect, since it represents a tendency for the car to go straight on rather than to go round the corner, so that exaggerated steering wheel movement is necessary.
The best known example of an oversteering car is probably the Austin Seven, in which the characteristic is inherent as a result of the combination of transverse front spring and torque tube rear axle. Aircraft seem to show the same tendency, in that it is generally necessary to “steer” into the turn and then centralise the controls.
Under-steering is a common characteristic of modern touring cars, such as the Vauxhall, which generally have to be held forcibly into the corner.
Just whether a car over-steers, understeers, or has a neutral characteristic, depends on front and rear tyre pressures and loads, and on steering and springing layout. The main deciding factor, with normal suspension systems, is whether the front or rear tyres “creep” most, but with some suspensions rolling of the car “steers” the front and/or rear axles bodily, affecting the resulting behaviour of the car.
The view of American (and some British) designers is that, if a car is to be stable on a straight road, it should be made to under-steer. An over-steering car is said to be unstable and to need constant attention from the driver if a steady course is to be maintained; the Austin Seven certainly provides support for this belief.
From the cornering aspect, however, I think there is a lot in favour of an oversteering characteristic. A sports car is, or should be, an enjoyable car to drive, and an under-steering car always feels as if it is being cornered by brute force, whereas an over-steering car virtually corners itself and can be a real joy on winding roads. I should be interested to hear other people’s views, but in my experience most cars which are really enjoyable to drive show a very slight over-steering tendency.
Sluggish response to the steering is sometimes confused with under-steering, whereas it is actually quite a separate phenomenon. It does not affect cornering under steady conditions, but rather the ease and precision with which sudden swerves can be dealt with. Soft tyres and flexible steering connections contribute to it, but I think weight distribution has a big effect.
Most “vintage” machines have a fairly long wheelbase, with the weight concentrated amidships; modern touring vehicles, in contrast, are usually of fairly compact wheelbase, with the weights of engine, spare wheels, etc., mounted far apart. This modern tendency has inevitably increased the moment of inertia of the complete vehicle, about all axes, and although pitching may be reduced one can hardly hope to be able to execute sudden swerves.
My references to under-steering and over-steering have dealt with cases when there is no actual skidding; when skidding occurs conditions can be very different from those prevailing, when there is only tyre creep, at lower speeds. On the road the majority of corners are taken at speeds below the skid point, feel and accuracy of control under these conditions being important accordingly. But, for a sports or racing car which will fairly frequently be cornered at the highest possible speed, importance also attaches to what occurs when a corner is taken at and beyond the skidding speed.
According to most of the text-book laws, the coefficient of friction between tyre and road should drop when sliding begins, and skidding on a corner should, therefore, result in a waste of time. In practice, some skidding is essential to the fastest cornering, for two reasons. First, the most perfect of drivers cannot be certain that he is cornering on the extreme limit of available wheel adhesion unless he occasionally goes beyond that limit by a small fraction. Secondly, on certain surfaces, particularly those encountered when racing on cinder tracks or sometimes on sand, the grip between tyre and track obeys the laws of viscous rather than dry friction; in other words, a certain amount of sliding, on a loose surface, can provide the effect of increased friction.
The actual behaviour of a car when skidding a corner depends, of course, on whether the corner is taken at a steady speed or with the car accelerating or braking. Cornering tactics are very much a personal matter, but myself, on the road, I prefer to enter a corner at a speed just below the highest I think possible and to use as much acceleration as possible while still rounding the corner. My preference for this technique, which others may dislike, inevitably affects my views on the desirable characteristics of a car, and this must be borne in mind in considering my further remarks.
Different types of vehicles vary enormously in their skidding characteristics, particularly noticeable being tendencies to either front or rear wheel skidding.
Most British sports cars seem prone to skid their tails outwards on corners before the front wheels lose their grip. I have never driven one, but the S.S. 100 always appeared an outstanding example of this tendency, when it was used for racing. Provided skidding does not occur too readily, of course, it is probably best for the rear wheels to slide first, so far as the average driver is concerned. Within reason, a rear wheel skid is usually quite controllable, but the same cannot always be said of a front wheel skid.
A point which is brought home by study of racing, and of action photographs, however, is that most of the very best drivers, on purely racing machines, tend to corner in a gentle front wheel skid. Where the lesser lights may be doing spectacular broadsides, the true experts corner relatively inconspicuously, usually with the front wheels pointing straight at the kerb on the inside of the corner.
There appears to be little doubt that this style of cornering can give about the best possible results, but its adoption seems to depend on a correct combination of driving technique and car characteristics. A very nice balance between front and rear skidding tendencies is necessary, as far as the car is concerned: also, the driver aprarently has to enter the corner at a speed extremely near the limit, since if he approaches more slowly and then accelerates, wheelspin will start a rear wheel skid and put him off his course. Effective cornering by this method seems to call for extreme judgment and delicacy of control, such as few drivers are capable of (I certainly am not), and although a really well-balanced car should be amenable to this technique, most folk will not actually use it.
Another point which affects the behaviour of a sliding car is weight distribution. Apart from its effect on a machine’s tendency to front or rear wheel skidding, the positions of heavy items such as engine in the chassis, either amidships or near the front or rear end, will affect the rapidity with which a skid develops. A car with a large moment of inertia, apart from being liable to sluggishness in its response to the steering, will often corner in a fairly slow broadside skid, whereas a more compact vehicle may, when it slides, flick its tail round extremely quickly. Which type of skid is the easier to control I can hardly say.
Looking over what I have written, I seem to have talked around the subject of cornering, and the comparative behaviour of vintage and modern types of machinery, without producing any definite conclusions. I see little else that one can do. The more we can get different people’s views on this complex subject, the better we may be able to assess the merits of different points in design and driving technique.
Meanwhile, cornering remains very definitely an art. By obeying fairly clearly defined rules, any sports-car designer is able to produce a machine which handles better than its touring counterpart; by obeying rather similar rules, the ordinary sports-car driver can develop greater cornering skill than the average saloon car motorist displays. But there still remains just that little extra something, which makes a Bugatti or a Nuvolari, and which cannot be put on paper…. [Still no one has cleared up the phenomena of tyre howl under varying conditions and in differing circumstances, around which my ramblings in “Rumblings” mainly centred. – Ed.]