Steering geometry

Are the car wheels parallel when steering?

It happens that geom­etry and mechanics are precisely what you need to define how the wheels of your car must be turned.

If the prolon­ga­tion of the wheel axis passes through the centre of rota­tion, then the wheel, while turning, leave a clear tread. If the prolon­ga­tion of the wheel axis is not directed toward the centre of rota­tion, then the wheel skids as it turns. The tread will be erased from the skid­ding, and, above all, the control of a vehicle with wheels like that will be increas­ingly diffi­cult as the speed increases. So, for a good control of the car the prolon­ga­tions of the axes of all the wheels must be directed toward the centre of rota­tion. But what this means for a car with four wheels?

To begin, let us learn to move along a simple curve, i.e., an arc of a circle.

Since the axle of rear wheel in most cars is fixed, the prolon­ga­tion of this axle should be directed toward the centre of this circle. The front wheels are then turned so that the axis of each is directed toward the same centre. This means that for a good control one should be able to turn the front wheels by different angles, so that these wheels will be non parallel!

You will say that the curves are not always arcs of circles, and in addi­tion that the car does not stop, before steering. This, of course, is true, but it happens that along any curve we can consider that at every time the car moves along an arc of a circle whose radius and centre depend on that moment of time.

Consider an ordi­nary road. In order to be viable by car, it must have no sharp angles, namely, its middle line must be, as math­e­mati­cians say, a smooth curve.

Take any point on this line and mark it in red, and another, a little far from the first, marked in blue. These two points define a unique straight line connecting them, and we draw that line. Now move the red point along the curve toward the blue point. At a time when the points coin­cide, the line defined by them will become the tangent line. It is the linear approx­i­ma­tion of the curve in a small neigh­bour­hood of that point. We observe, however, zooming in, that the curve and the straight line are close only for a small stretch.

Let us now take two red points on the curve, one on the right and one on the left of the blue point. Three points, which do not lie on a straight line, define a unique circle, that we draw. Now we move the two red points towards the blue point. When the three points coin­cide we get a circle, which is called oscu­lating. It is the second order approx­i­ma­tion of the curve, and zooming in we may see that this approx­i­ma­tion is better. We observe that in a stretch of the curve, along which the radius of curva­ture increases or decreases (as in the ascending and descending parts of the curve in the figure) the oscu­lating circle always inter­sects the curve, differ­ently from the tangent, which is situ­ated in the same part of the plane cut off by the curve, in a neigh­bour­hood of the point of tangency.

Since the oscu­lating circle in our case well approx­i­mates the curve and can be constructed at each point, the motion along a road with curves can be consid­ered at every instant of time as the motion along an arc of a circle. The radius and the centre of this circle depend, obvi­ously, from where the car is.

In this way, while moving along a certain curve, one can imagine that at every time the car moves on a small arc of a circle. And our first case, where the road is exactly an arc of a circle, is there­fore essen­tial to study the motion.

But how to get that for any posi­tion of the steering wheels the prolon­ga­tion of the axes are directed toward the instan­ta­neous centre of curva­ture (i.e., the centre of the corre­sponding oscu­lating circle)?

It happens that also here the geom­etry comes to our aid, and specif­i­cally with the isosceles trapezium, that we know from school: a quadri­lat­eral with two parallel sides, called bases, and two other non-parallel sides equal to each other. If we choose the correct lengths of the sides of the trapezium, we obtain the condi­tions neces­sary for the optimal control: the prolon­ga­tions of the axes of the front wheels inter­sect at a point lying on the prolon­ga­tion of the axis of the rear wheels. And this point is the instan­ta­neous centre of curva­ture.

This control system was invented by a French master craftsman, Charles Jeant­land, which applied it to the front wheels of the carts. But for carts, moving at low speeds, this inven­tion was not so essen­tial as for cars, and was forgotten. Only after three quar­ters of a century, two German engi­neers, two fathers of the auto­mo­bile, Gottlieb Daimler and Wilhelm Karl Friedrich Michael Benz, building their cars, returned to the Jeant­land trapezium. Daimler in 1889 obtained a patent for his “method of inde­pen­dent control of front wheels with different radii of curva­ture.” And in1893 Benz obtained a patent for his “construc­tion of a control system with control circles tangent to the wheels.” After solving the problem of control of the front steering wheels and other impor­tant tech­nical issues, Karl Benz built his famous car “Viktoria”, with four inde­pen­dent wheels.

From the math­e­mat­ical point of view, the trapezium does not allow to create the condi­tions neces­sary to ensure that the inter­sec­tion point of the prolon­ga­tions of the axes of the front wheels will always lie on the prolon­ga­tion of the axis of the rear wheels. Using the trapezium, this point will lie a little bit far from the line of the rear axle. But why we have so long spoken of the trapezium, you ask? We worried too soon: we simply must not apply the math­e­mat­ical rigor to tech­nical prob­lems without using the common sense. In order the prolon­ga­tion of the axes of the front wheels will always inter­sect at a point on the prolon­ga­tion of the axis of the rear wheels is neces­sary that the length of the smaller base of the trapezium change a little. For the usual length of its base, which is more than one meter long, the neces­sary changes would consist of the length of a centimetre, which is lower than the inter­stices between the joints and the toler­ance in manu­fac­turing.

From the time of the creation of the first car the speed is increased dramat­i­cally. So also grew the need for a better control of the front wheels. In addi­tion, the trapezium is a flat geometric figure: this method of control of the front wheels can be used only with non-inde­pen­dent front suspen­sion, i.e., when the wheels are rigidly connected to each other and the straight line joining their centres is always parallel to the plane of the trapezium. Now this system can be found in the trucks. In modern cars the wheels suspen­sions are inde­pen­dent, the wheels can move in vertical direc­tion with respect to one another. To control the steering of these wheels, more complex mech­a­nisms are used, no longer planar, often with a central element with the shape of a rack. Even in this case it is a problem for math­e­mati­cians and engi­neers to calcu­late the different elements. But for histor­ical reasons these mech­a­nisms continue to be called steering trapezium.

With the steering of the car another ques­tion is emerging, still related to the geom­etry. The length of a circle of radius R is equal, you will recall, to 2R, and the length of an arc of ampli­tude  is equal to  R. When the car moves along an arc of a circle, the outside front wheel, furthest from the centre of curva­ture, moves along an arc of radius greater than the inside. The same holds for the rear wheels: the outer wheel describes an arc of length greater than the inner wheel. And since the rays are different, so the lengths of the paths of the wheels connected by the same axle must be different. In the oppo­site case one of the two wheels should skid, and this should decrease the control of the car.

In the case when the axis of the wheels is not the primary, i.e., the wheels are not driving — in our case the front wheels- it is simple: each wheel turns with its speed needed to go through the desired curve without skid­ding.

But how to do so that the driving wheels — in our case the rear wheels — on the one hand constantly move the car forward and on the other rotate at different speeds?

This problem is solved by the differ­en­tial, a typical ’plan­e­tary mech­a­nism’. A mech­a­nism is called ’plan­e­tary’ if it has satel­lite mech­a­nisms revolving around axes which in turn rotate.

The drive shaft through the trans­mis­sion, gives the rota­tion to the primary axle, namely, through the plan­e­tary mech­a­nism, to the primary right and left shaft-axles. What­ever the speed at which the rear wheels spin, the speed of the mech­a­nism will always be equal to the average speed of the drive shafts. Each wheel from the differ­en­tial takes a lot of rota­tion, propor­tional to the length of its path, so that in total all the energy is converted into the forward move­ment of the car.

If the car moves on a straight road, and under the driving wheels there is the same soil type, i.e., with the same coef­fi­cient of fric­tion, the wheels take from the mech­a­nism the same amount of rota­tion, so that the drive shafts rotate (and the wheels together with them) at the same speed.

But if the fric­tion coef­fi­cients are different, for example, while one side of the car passes over a patch of oil or over a layer of ice, then the wheels will move if this happens? For the non driving wheels it is quite simple, since they are inde­pen­dent of each other, and do not push the car, so if one of them ends up on the ice, it stops spin­ning, and the control becomes very low.

But now also the driving wheel on the left rear ends up on the ice. For the right wheel on the asphalt, the control is high, while for the left wheel on the ice, it is almost absent. Conse­quently, the left wheel turns more easily, and it begins to take at its side all the rota­tion that the mech­a­nism splits to the two shafts. Since the amount of speed, as we have said, is constant, if an axle shaft turns very quickly, the other will not rotate. To start moving in this situ­a­tion, when one of the driving wheels has lost grip with the ground (for example, on the ice), while the other has not, it is impos­sible.

This would appear as a defect in the differ­en­tial. The engi­neers, in fact, seek to improve the differ­en­tial, while retaining its basic prop­erty, which is that of the optimal control in steering, and by reducing the unpleasant effects, such as not allowing one axle to spin when the differ­ence of angular speed is high. But, it seems, up to now nothing has changed, as well as the laws of geom­etry are always the same.