Starting again from scratch
The banning of turbocharged engines and the introduction of a new 31/2-litre formula for Grand Prix racing gave manufacturers a clean sheet of paper on which to design new racing engines, but they have adopted a variety of approaches. Historian David Hebb considers the relative merits of eight, ten and twelve cylinders, and the motives behind their selection. Indications so far suggest that, under the 1989 Formula One regulations, design priorities will be different from prevailing ones. Even though little information has been released so far, one can deduce certain trends.
The emphasis on compactness and lightness, evident in both the V10s and the broad-arrow, “W12s”, reveals a growing desire on the part of designers to achieve speed through a combination of superior acceleration, handling and braking. The advent of the V10 also implies that aerodynamics will still be an important factor in engine design. The narrow included angles of the engines recently introduced point to continued interest in airflow management over and around the rear of the car, even at some cost in terms of centre of gravity.
Not all designers agree with these priorities; for some time, power has become all-important. Engine configurations which promise high specific output are gaining favour, even though these layouts might entail some aerodynamic cost or weight disadvantage. The path to power is easily recognised. So much is now known about combustion chamber and port shapes and so on that there is very little hope that much more power can be obtained from improvements in this area. Therefore designers seeking more power are following the traditional route.
Put simply, the more fuel and air an engine can burn in a given time, the more power it will produce. Therefore, the faster an engine runs, the more fuel and air it will take in and burn, producing more power. The traditional way of making an engine run faster is to make one with more, smaller cylinders. Smaller cylinders mean less reciprocating mass, and therefore less inertia stress and higher revs. More cylinders also results in greater total piston area and potential valve area, better breathing, and therefore more power. This relationship can be illustrated by comparing engines of the same swept volume (3.5-litres) and stroke/bore ratio (0.6/1). If 4350 ft/min is taken as a maximum safe piston-speed, then an engine of 12 cylinders would have a maximum piston area of 105.5 sq in and could spin safely to 12,999 rpm. For a 10-cylinder engine of the same cubic capacity and stroke/bore ratio, the figures would be 99.7 sq in and 12,277 rpm. For an 8-cylinder engine, the corresponding figures are: 92.6 sq in and 11,340 rpm. Thus, the 12-cylinder engine should turn faster and produce more power.
Ferrari appears to accept this conclusion, for the Italians have chosen an engine of 12 cylinders, the maximum number permitted by the regulations. Furthermore, characteristics of the Ferrari engine indicate that priority has been given to high specific output. Reportedly, this engine (84mm x 52.8mm) is long and heavy with a power-band of only 1500 rpm. These features lead one to conclude that maximum power has taken precedence over other considerations; Ferrari appears willing to accept a very narrow power-band, extra weight, and greater mechanical complication to achieve this end.
The 80° V12 Chrysler-Lamborghini engine represents a similar approach, as does the 180° V12 which Subaru and Minardi are developing. The latter engine’s wide-angle was, of course, used by Ferrari, Tecno and Alfa Romeo under the 3-litre formula. With a “flat”12 shortness, balance and high rpm can be obtained. Unfortunately, the width of such an engine poses aerodynamic problems. Despite this, Subaru may have chosen this configuration because the strength of the company’s engineering tradition (and product-line) lies in “flat” engines.
The broad-arrow units built by MGN and Life also give evidence of the priorities of these firms. Both engines are fairly wide and tall compared to a V10 or V12, but the “W” configuration makes for a very short engine and low weight, thereby enabling the designer to house all the fuel centrally without producing a long and heavy car.
The choice of 12 cylinders (as opposed to nine) by both companies also implies a search for power through high cyclic speed. The designers of all these 12-cylinder engines are following LH Pomeroy’s dictum that “Revolutions are of an abstract nature; they cost nothing, weigh nothing, have no shape or substance. If one can get more revolutions than another, it is hard to find a reason why he should not benefit thereby.”
Revolutions may be highly desirable and cost nothing, in the abstract; but in the real world, they are not acquired quite so easily or cheaply. Extra weight, friction, and complexity may reduce the theoretical advantage. Obviously, not every designer agrees that the 12-cylinder engine (maximum number of cylinders = high rpm = power) is a race-winner.
Designers of the V10s clearly think differently. As we have seen, these engines show signs of compromise: greater priority has been placed on weight, compactness and aerodynamic considerations than on absolute specific power. However, other features of both the Renault and Honda engines lead one to conclude that they, at least, believe they have found ways of matching the power of the twelve-cylinder units while still retaining the advantages in weight and compactness that a V10 offers. On the basis of Fritz lndra’s calculations (Motor Sport, December 1988), a V10 should rev 6% lower than a V12 and produce less power (but only 1.8% less bhp). In reality, even this slight deficiency may not be present as the experience of the Cosworth V8 shows. According to the same theory, a V8 such as the Cosworth should have revved 13% lower and produced 4%, less power than a V12. In fact, its deficiency in revs was no more than 8% while its power usually matched or exceeded that of the V12s. Moreover, from the little that has been revealed so far of the V10s designed by Renault and Honda, it seems certain that both companies do not expect to suffer any deficiency in rpm or power.
Recent experience suggests that the major restraint to higher cyclic rates (and therefore higher power) lies in the limits set by valve-gear. Mean piston-speed and maximum acceleration, ring and bearing loads are within acceptable bounds. Valves, or more precisely the coil-springs which usually close them, are the problem. To overcome this limitation, Renault plans to avoid metal springs altogether by using pneumatic valve-springs, as it has done successfully in its 1.5-litre V6 turbocharged engine which was capable of revving to 12,500 rpm without difficulty. Honda on the other hand, believes pneumatic valves take up too much power; according to engineer Nobuhiko Kawamoto, the company will seek a different solution.
Torsion-bar springs are a possibility, since they were used very successfully on some earlier high-revving Honda engines. Unlike coil-springs, which are liable to surge with destructive effect and are subject to bending stress, torsion springs are stressed only by twisting forces.
Another solution might lie in the use of new and lighter materials (such as ceramic/metallic-matrix composites) on order to reduce valve component mass and inertia-stress. Following aero-engine manufacturers, who increasingly are moving away from homogeneous metallic materials, Honda might try to use non-homogeneous composite materials to reduce weight and increase valvegear stiffness so that its V10 engine will rev higher.
What of the V8? Under the 3-litre formula this was the dominant engine type, and currently it is the only one used by naturally aspirated Grand Prix cars. Judd will continue to offer a V8 engine, though next year’s model will be a thoroughly revised shorter-stroke version of the current engine. This revision implies a search for more power through higher rpm, possibly with larger valves. There is little doubt that incremental gains can be obtained by such development work; however in Honda’s view, such gains will be largely offset by the need to make the engine stronger, and therefore heavier, to deal with increased stress — thereby losing any initial weight advantage a V8 might have. Honda’s old rival from the motorcyling world does not agree; Yamaha intends to offer a newly-designed V8, the 0X88, which possesses two distinctive features: the 75° angle between the cylinder-banks is narrower than usual, and each cylinder contains three inlet and two exhaust valves. However, this engine does not mark a radical departure for Yamaha; each of these features has been used before, either in the company’s motorcycle or automobile racing engines.
With the resources of Ford behind it, Cosworth was free to choose any type of engine, but decided to design and develop a new V8. This might seem surprising at first. However, the choice of a V8 fits in with the company’s engineering tradition. Cosworth knows this layout exceedingly well, and in the past (and contrary to theory) its V8s have been able to match any 12-cylinder engine. A Cosworth V8 is likely to be superior to the V10s and V12s on length and (possibly) weight (the 3-litre DFY came in at only 292 lb), and its mechanical losses are likely to be less. One must assume Cosworth believes it knows how to make a V8 equal the power of the new 10 and 12-cylinder engines.
In the past, the power potential of Cosworth V8s has been limited by their valvegear, and the new engine is likely to have to rev higher and therefore will put more stress on valve components. How Cosworth will cope with this problem is not yet known, but it should be recalled that the company spent considerable time and effort experimenting with desmodromic valves (ie, valves closed as well as opened by cams) for a development of the DFV.
This work stopped when Keith Duckworth concluded that a new test-rig and another two or three man/years of development would be needed to bring the programme to a conclusion. Since getting the backing from Ford, Cosworth has had the time to complete such research and development, though Duckworth may eschew the complication and manufacturing difficulties of desmodromics and prefer simpler, more traditional solutions. Precise control over materials and manufacturing and the application of new materials with their improved weight/ strength ratios might offer a solution.
The beginning of a new formula is an exciting time. Everything seems up for grabs; everyone thinks he knows best how to meet the problems thrown up by the new rules. Nobody knows for certain, of course. Given human nature, different engineering traditions and resources, a range of solutions is bound to be put forward. Moreover, as we have seen, designers start with different sets of presumptions, priorities, and prejudices. The validity of the choices they make will be discovered by the test of the track.
In the next few years we can enjoy watching the process work itself out. Will eight, ten or twelve cylinders be the optimum number? Will “V” or “W” engines reign supreme? Has the time of the rotary valve come at last, or will torsion-bars, pneumatic springs or desmodromics sweep the field?
No doubt diversity will continue for a time, if only because each engine design represents great financial, intellectual, and personal commitment. But after a few years, some configurations or features will be dropped, perhaps because they were genuinely inferior, perhaps because they were just unlucky. Imitation will begin and there will be growing conformity, at least until the next change of formula or until some new (or revived) feature becomes fashionable. DDH