Back to the future
Ford-financed Cosworth-built three-litre engines dominated Formula One racing in the decade before the arrival of the turbos, finally winning an unprecendented 155 Grands Prix. This year the same alliance has produced a brand new power unit for the 3.5-litre formula. Few details have escaped from the veil of secrecy which pervades racing at the top level, but David Delison Hebb gleaned the philosophy behind the Ford V8 from Cosworth’s chief designer.
Surprisingly for an engine designer Geoff Goddard orders the factors determining the competitiveness of today’s Grand Prix cars as follows: (i) tyres; (ii) chassis; (iii) driver; (iv) engine. Of these, only chassis considerations influence engine design significantly.
Chassis designers want the smallest, lightest engine possibIe. However, they also want as much power as possible. Engine designers, therefore, are faced with trying to reconcile the conflicting demands of compactness and power. How successfully they resolve this dilemma is determined by their intelligence, their experience, and the resources at their command. The Cosworth Ford design team is amply supplied with the first two, and Ford has the financial muscle to compete with any company in the world, though it is unclear whether it is prepared to match Honda’s enormous technical commitment.
Work on the new V8 began with an analytical assessment by the Cosworth design team of the crushing capacity (and therefore potential power) of various layouts. Engineers developed a mathematical model to calculate the power of hypothetical 8-, 10-, and 12-cylinder engines. These studies suggested that a 12-cylinder engine should rev higher than the others and, therefore, have the greatest power potential. However, friction and pumping losses are relatively greater with more cylinders; consequently, the theoretical power advantage of a twelve appeared to be marginal. By Cosworth’s calculations, only about 20hp would separate 8-, 10-, and 12-cylinder engines designed to the same standard; differences in size and weight would be more significant.
The W12 layout was quickly eliminated by Cosworth on the grounds of excessive width. A V12 had more appeal since it would be narrow, but it would also be long and heavy. A V10 would be shorter and lighter than a V12, but not as much so as a V8. However, an orthodox 90° V8 would be prohibitively wide.
With years of experience building V8s, Cosworth naturally was attracted to this configuration, especially when the engineers realised that by using an included angle of only 75° the width problem of a 90° engine could be overcome.
Narrow-angle V8s are not all that unusual. In the past others, Lancia in particular, have built engines with included angles of 22° or less. And in the 1950s Fiat made a 2-litre 70° V8. An included angle of 75° still provides enough room between the cylinder banks for inlet tracts of the right size and shape. Both Yamaha and Judd reached the same conclusion, producing V8s of 75° and 76° respectively. Two additional features attracted Cosworth to the V8: weight and fuel consumption. A V8 should be lighter than 10 or 12-cylinder engines built to similar standards. Mechanical and pumping losses should also be less, there by producing better fuel consumption. A light engine makes for a lighter car and more rapid acceleration, and every additional gallon of petrol adds more than 6lb which have to be accelerated and braked throughout a race. The extra weight of engine and fuel also results in more wear on tyres — the most influential factor, remember, in Goddard’s list of factors influencing the competitiveness of a racing car.
Weight was a primary concern, but Cosworth has not sought lightness by using exotic materials: some carbon-fibre is used for the injection trumpets and the cam covers are cast in magnesium alloy, but in general the use of super-lightweight materials is minimal. Instead, the accent is on simplicity, a word that crops up again and again in conversations with Cosworth officials. Complex solutions are eschewed because they tend to weigh more, reduce reliability, and cost more.
For example, when asked if Cosworth had considered the possibility of using variable lift and timing cams or variable-length inlet trumpets, Goddard replied that while such mechanisms might bring small improvements to the power curve of the engine, the additional weight and complexity did not make them worthwhile. Given the competition, one must wonder whether Cosworth can afford to pass up even marginal power increments in the name of simplicity, but this an issue that can only be settled by racing.
Engine design proper did not get under way until May 12, 1988, but the project moved ahead with astonishing rapidity considering the small sum er of people involved. Goddard completed the overall layout drawings by June 1 and then set his select team of six to detail particular engine groups or components. One drew up the crankshaft assembly, another the pumps, while a third, Stewart Grove, was entrusted with the detail design of the cylinder head.
According to Cosworth chief, Keith Duckworth, head of design is “the most important part of an engine as far as power output is concerned”. The more fuel/air mixture that can be brought in, burned quickly and completely, and expelled efficiently, the more power an engine will produce. The starting point of cylinder head design is the combustion chamber and intake system. The shape of the inlet tracts and the amount of the valve opening area will determine the size of charge that can be brought into the chamber. For years Cosworth has been synonymous with four-valve narrow-angle pent-roof chambers. Over the last twenty years, the valve-included angle of atmospheric engines designed by Cosworth has diminished from 40° in FVA to 32° for the DFV, and down to 22.5. in the DFY. Goddard would not reveal the angle used on the new V8, but indicated that further reductions brought diminishing returns. From this it would seem that the angle is probably between 22. and 32°.
His comments on another aspect of valve design were more forthcoming. When quizzed about the merits of five valves per cylinder, he replied that Cosworth had examined such layouts. He agreed that they had some attractions, but pointed out that they had negative features as well, adding, “it’s what people don’t tell you that matters as much as what they do.” Goddard would say little except that five valves were not used, and that it is possible to gain from a four-valve design without the additional complexity.
In comparison to its radical early years, Cosworth appears to have become a conservative firm. Innovations are actively examined, but they must provide proof positive before the firm’s engineers will adopt them. An example of this attitude came when we discussed the combustion requirements of the new Ford engine, which operates at a 12:1 compression ratio. While Goddard was loath to talk about such elementary details as the size of the bore, outward dimensions indicate very generous bores, similar to the 100mm ones used on the GAA 3.4 V6 of a few years ago. When asked whether more than one spark plug would be needed to achieve rapid and complete ignition in the short time available in such a very high-revving large-bore engine, Goddard got up and pointed to a cutaway drawing of the GAA. The cylinder head of that engine was designed to take three spark-plugs, as the casting reveals, but lengthy development tests showed that only one plug was necessary.
Rejection of the latest fashion is evident in the camshaft drive of the new engine, which takes its power from the front of the crankshaft, rather than from the rear like the Judd and 2.65-litre Indy engine designed by Ilmor Engineering (exCosworth people).
Taking the power from the rear, so its advocates claim, has the advantage of reducing the destructive shock loads fed into the camshaft drive system. The crankshaft flexes less at this point because the mass of the flywheel (found at the rear) dampens torsional variations from which all crankshafts suffer. Goddard accepts the theory but is not convinced that the deflections are great enough to disturb a properly supported crankshaft and well-designed drive system. Moreover, as he points out, taking the drive off the rear adds to the engine’s size at the very place where chassis designers want it to be as small as possible for aerodynamic reasons. Putting the camshaft drive at the front allows the gearbox to be placed snugly against the engine; it also makes maintenance of the camshaft drive system easier. In the past Cosworth had designed engines to use shafts, belts, gears and chains to drive the camshafts. Each can be made to work, if designed properly. The 1.5-litre turbo (the last Formula One engine designed by Cosworth) could rev to 14,000 rpm without the chain-drive system malfunctioning. It used to be thought that chains would fail at very high rpm because the small sprocket required to achieve a 2:1 reduction would induce a variable, jerky movement to the drive. However, Cosworth has found that if the drive is not taken directly from the crankshaft but from a 2:1 reduction gear, the problem is eliminated.
In its favour, a chain-drive system has several attractions. It is light, it is simple, and it is economical: one can see why Cosworth might be attracted to it for the new engine. Late-model DFVs and DFYs used a gear-drive system dependent on ingenious quill shafts to take up torisional variations and shock loads. It worked well but required many parts, much machining, careful assembly, and was costly.
Economy has always been a watchword at Cosworth, but not false economy. A cursory trip around the factory leaves one in no doubt that large amounts of money are spent on machine tools and equipment — where it matters. Even the most brilliant design is worthless if it cannot be made accurately, assembled properly, and at an acceptable cost.
Manufacturing ability is Cosworth’s forte, its “special ingredient” if you like. This capacity made Cosworth the foremost manufacturer of racing engines. The Heenan and Froude dynamometers at Cosworth are calibrated frequently to an accuracy of plus or minus 1%, and the accuracy of design and manufacturing tools is calculated to seven decimal points. Without such facilities, no firm should consider competing with Honda or Renault.
All major castings of the new Ford engine are also done in-house using a special technique developed and patented by Cosworth. Casting is usually accomplished by pouring molten metal from a cauldron into a sand mould. Two problems may arise: the molten metal can cool and solidify unevenly; or its surface can be exposed to air allowing oxidization to occur; when the metal is poured, this metal oxide goes into the mould, creating bubbles which weaken the casting.
Until the late-1970s, Cosworth farmed-out casting work which was done by traditional methods. The quality of the castings received from suppliers was inconsistent: some displayed poor dimensional control; in particular, small water passages in four-valve heads did not match design criteria, others were weak due to porosity from oxidized metal. When raced these castings cracked or failed catastrophically.
In Cosworth’s new process, the molten metal is contained in an airless vessel (an inert gas such as Argon fills the void), and is then drawn from the centre of the vessel to avoid any impurities which sink to the bottom. It is then pumped at constant temperature up into the bottom of the mould so that any part of the metal exposed to air is eventually pushed out the top. As a result the cast component is solid and strong as designed. Tolerances from true position are also held to an astonishing minimum: 0.2mm for combustion chambers; 0.3mm for ports within the heads.
This casting process contributes significantly to the small size and light weight of the Ford engine. Water passages can be less than 2.5mrn wide and major walls only 3.5mm thick. This means very large bores can be placed close together but still with enough room for coolant to flow easily around the cylinders. For the some reason, valve seats and spark-plug bosses may reside in close proximity without fear of cracking from thermal fatigue.
Perhaps the most obvious feature of the Ford V8 is its small size compared to the DFR. This is because the new engine was designed to benefit from the Cosworth casting process, and because it was designed as a tight 3.5-litre motor, whereas, in Goddard’s words, “the DFV (and its descendant the DFR) is a 3-litre engine in a 4-litre package”. The new motor can more easily be reduced to 2.65 litres if desired, for CART racing, than increased in size. Cosworth continues to use traditional metals such as a very pure LM25 for major castings. In the past the firm has been criticised, rightly or wrongly, for not speedily adopting the latest super-materials. The new Formula One engine does use some special alloys developed by Ford Scientific in America. A range of new high-strength single-crystal steels and lighter aluminium alloys (some containing lithium) are becoming available. Moreover, the strength weight ratios of traditional materials can be increased substantially when the metals are reinforced with silicon carbide whiskers.
Whether Cosworth-Ford has taken full advantage of such new, improved materials has not been revealed. In the past its ability to machine or manufacture exotic materials reliably has largely determined their use. Titanium alloy valves have reached that stage and are probably used in the new engine, but one wonders whether newer materials, such as metal matrix composites, are used for parts such as connecting rods.
Honda’s R&D Chief, Kawamoto, is on record as saying that he does not expect exotic materials to play a crucial part initially in the development of engines under this formula. However, the pressure of competition may precipitate early adoption. One would also like to know how much production engineering considerations have influenced the selection of materials and design choices made by Cosworth. If production of a series of racing engines (such as with the DFV) was envisioned from the start, rather than just a few engines for one team, then it is possible that special materials might not have been used even though they may offer some performance advantages.
Though Ford may have assisted Cosworth in the selection of materials, the main contribution of the American company (other than money) has been in the development of an integrated, computer-controlled ignition and fuel-injection system for the new engine. Ford has the necessary computer facilities and long experience in this area of design, which has become crucial to the development of efficient, economical racing engines.
To have an engine running, let along meeting its design targets, after only six months is a major accomplishment, and Cosworth has achieved this without throwing men and money at the project. Compared to Honda or Renault, the Cosworth design team is tiny and so has had to work long and hard; every member put in on average 60-65 hours every week, without a penny of overtime promised or expected.
At first the new engine was run for hundreds of hours on the test beds, but good results there are one thing, carrying them over to the race track is another. Early road testing in the new Benetton F189, revealed a crankshaft defect which has since been cured. Reliability clearly is a major concern and much has been achieved as is evident in the finishing record of Nannini’s Ford powered car. Company officials will say little about the engine or its performance. Basic dimensions are: length, 595mm (23.4in); width, 591mm (23.3in); height, 521mm (20.5in). Weight is undisclosed; all Goddard would say is that it is less than the 140kg of the DFR. It probably weighs less than any competing engine (except possibly the Judd), but even so, its advantage over the Honda and Renault engines does not appear to be very great — about 10-12kg, if published figures are accurate.
Without doubt the Cosworth engineers have made a compact motor. It is approximately 6in shorter than the Subaru, 5in shorter than the Lamborghini and 4in shorter than the Renault, but only 1in shorter than the Honda. On width the comparison is less favourable. As one might expect it is considerably narrower than the Subaru (nearly 6in) and about as wide as the 80° Lamborghini V12, but still over an inch wider than its main competition the Honda and Renault V10s and the Ferrari V12.
But weight and size are only partial measures of an engine’s effectiveness: if it is to succeed, it must be competitive in terms of power and reliability.
Before the racing season began we were treated to a succession of claims and rumours about engine power. At that time Cosworth thought there would be very little difference in power between the major new engines — a spread of about 20hp — and that its V8 would beat the others on torque. The old DFR developed slightly less than 600hp, according to Goddard, thus it would seem an expensive exercise in futility if the new engine did not produce at least 10% more power initially. Honda expected its V10 to match the 650hp produced by last year’s 1.5-litre turbo.
Lap times suggest that Honda has met its target. The V10 has more power than any competing engine, and is reliable and abstemious as well. The Renault V10, the Ferrari V12 and the Ford V8 are running second and seem to have similar power. Has Cosworth underestimated the opposition or failed to anticipate the speed with which the others are developing their new engines?
Honda is now racing the third, if not fourth, major version of its V10, and has revealed the existence of a new V12 to be raced in 1991. Under the last formula, the Cosworth-Ford alliance was surprised by the speed with which rivals developed their turbocharged engines; it took almost two years of hard development work to draw even. If only for the sake of exciting racing, one hopes that the new Ford can be developed quickly. Already it has become clear that the competition is formidable, far more so than when the DFV was introduced; matching the record of its illustrious predecessor will certainly prove a challenge. DDH