“When it showed 402bhp, we knew then we had an engine to win races”
Mike Costin looks back on the making of a masterpiece: Cosworth’s revolutionary DFV
Look at the Lotus 49 from any angle and it will always seem delightfully sharp and slim. Delicate, even. It was the product of the innovative skills of company founder Colin Chapman and his chassis designer Maurice Phillippe, the engineering wizardry of Mike Costin and Keith Duckworth of Cosworth fame, and the intuitive vision of Ford’s media specialist Walter Hayes.
The roots of the combination stretch as far back as 1955, when the 32-year-old Hayes joined the Daily Mail ’s Sunday Dispatch publication to become Fleet Street’s youngest-ever editor. He immediately commissioned promising Hornsey-based engineer and racer Colin Chapman to write the occasional article.
Hayes joined Ford UK as head of public affairs in 1962, just as Ford US was targeting the youth of America with its Total Performance programme. Similarly attuned to the marketing benefits of motorsport, Hayes approached Chapman to build 1000 Lotus Cortinas at his Cheshunt factory, featuring his race-tuned 1500cc Lotus/ford engine with Harry Mundydesigned twin-overhead-cam cylinder head.
The story of the Lotus 49, meanwhile, had already begun. Chapman sketched out revolutionary plans for its forebear, the Lotus 25, on
a napkin while having lunch with Costin, finance director Fred Bushell and purchasing director John Standen in 1961. Critical to the 25’s concept was a monocoque chassis, which made it three times more rigid than the also new, customer-focused 24 model, but only half the weight. A very narrow frontal area would run back to a bathtub-like cockpit, with the drivetrain installed neatly behind. This ultra-thin appearance was achieved partly by moving the front suspension inboard, while power was provided by a MKII 1496cc Coventry Climax. Fourteen Grand Prix wins and two World Championship titles followed, ending only when the FIA doubled Formula One’s engine size to three litres for 1966. Climax immediately announced its withdrawal from racing, and Lotus opted for the troublesome BRM H16 to power its 43.Though groundbreaking, reliability was poor and its only Grand Prix victory came at Watkins Glen courtesy of Jim Clark.
Something better was evidently required, so Chapman asked Cosworth to calculate costs for an all-new engine; a figure of £100,000 was put on the table. Chapman’s overtures to the Society of Motor Manufacturers & Traders to bankroll the programme drew a blank, so he called Hayes and a meeting was convened between them and Ford’s European vice president of engineering, Harley Copp. A budget of $350,000 (hip-pocket money by American standards even then) was made available for a two-stage programme. “First would be a one-off 1500cc experimental engine,” explains Costin today. “Called the FVB, it produced 200bhp and was later installed in our Brabham BT10 chassis for testing. Then came the all-conquering DFV.” Ford, for its money, would get world publicity; Cosworth received the business, which grew exponentially; and Chapman, without paying a penny, had his F1 engine – but not, as he’d hoped, sole rights to it.
In 1966, and after much wrangling, Cosworth’s Duckworth drew up the three-page contract with Ford, having declared that he couldn’t understand the original.
“For Cosworth,” Costin says, “it wasn’t a very big contract. To put the DFV engine into perspective, Cosworth’s task was to design the engine and we built seven units for the 1967 race season. At the time we were building a number of other engines including the FVA, the Lotus Twin Cam, as well as working for other companies such as Mercedes-benz. We were also markedly increasing the size of the factory.”
But the pressure on those involved with the DFV was intense from the outset. Duckworth was working 16-hour days to hit the agreed deadline of May’s 1967 Monaco Grand Prix.
The parameters for the DFV engine were to be a 90º V8 of 2993cc, with a bore and stroke of 85.67mm and 64.897mm, producing at least 400bhp at 9000 rpm and 270lb ft of torque at 7000rpm. The oversquare unit would feature the latest in Lucas Oscillating Pick Up System (OPUS) electronic ignition and fuel-injection technology. It looked an impressive package: exciting and certainly innovative.
“The idea to incorporate the powertrain as an integral part of the chassis frame was Keith’s,” recalls Costin. “This demanded a close working relationship between him and Lotus’ chassis engineer Maurice Phillippe. The first sketches, in June 1966, were for the front and rear faces of the engine, which Phillippe and Lotus would use to produce the layout for the chassis and the rear suspension pick-up points.
“Paper plans were the order of the day,” he smiles. “Keith produced the drawings for the cylinder head and block himself. For every other part he made sketches with the relevant dimensions, and these were passed to draughtsman Roy Jones, who created the engineering drawings for each part to be manufactured.”
Meanwhile, EX-BRM engineer Mike Hall undertook the detail design work for the ancillary units, such as the oil and water pumps, and the fuel metering units located on both sides of the engine, again from detailed schemes produced by Duckworth.
In front of Costin today, reunited at the Hethel base of Classic Team Lotus, is the lastbut-one chassis built: 49/BR10, from 1968. “It’s a little different to the early cars,” he says,
scrutinising the Lotus. “My first thought is the huge width of the wheels and tyres; when the car first raced it ran on 6in-wide wheels on the front and 9in rims on the rear.”
He begins to reacquaint himself from the front and its sleekly elegant nose, which takes the triangulated oil tank. This feeds lubricant through tubes let into the chassis to the engine, while the radiator behind also utilises similar pipes to transfer coolant to the power unit. On later cars such as this one, air scoops were added to allow hot air to escape. Behind, the inboard coil-sprung dampers are linked to the upper wishbone rockers with conventional tubular arms below locating the suspension uprights. Ventilated brake discs were used initially, but they proved too efficient at the 1967 German GP and were later swapped for solid versions.
Moving along the slender body he reaches the cockpit. “In designing the 49,” Costin explains, “Cosworth’s thoughts were that it should be large enough only for accommodating Jim Clark. However, Ford wanted Graham Hill to drive as well, so it had to be bigger. In the event, even I was able to get in and I’m larger than Hill. Notice the elbow scoops and the gearshift mounted by the driver’s right knee.” Over the cockpit, the specially shaped Venturi windscreen moves airflow, which reduces both buffeting for a given height and drag.
Further towards the rear, Costin points out the essential arrangement for the rear suspension pick-up points and the interface between the drivetrain and the chassis frame.
Paramount to the unified chassis and drivetrain structure were Duckworth’s calculations for linking the cylinder block and chassis monocoque: “He determined that the top bolts, which hold the engine to the chassis, had to be capable of withstanding a 3g bump-force diagonally on the front and rear, resulting in a hypothetical load of 4000lb. This was peanuts when you consider that at 9000rpm our big-end bolts had to absorb 10,500lb. The cylinder block was designed to act as a very stiff structure and this created a very torsionally rigid interface.”
Two chassis mountings are on the cam covers; on the bottom, two bolts spaced nine inches apart locate with the chassis via an aluminium bracket mounted on the front face of the sump.
It’s clear to see how the drivetrain also provides the location for the suspension coilspring/damper units, and how the trailing-link radius arms attached to the uprights with their driveshafts and disc brakes. Within the V8 engine bay sits the alloy casing, holding the Lucas OPUS electronic ignition and fuel-injection system with its vertical intake trumpets.
Recalling Cosworth’s testing of the DFV, Costin says: “When we put the first engine on the dynamometer it showed over 402bhp at 9000rpm [this rose gradually to 480bhp at 10,500rpm], which was the designed output. We knew then that we had an engine on our hands that would win races. However, reliability was always the main consideration.”
A decision taken early on concerned sourcing a suitable gearbox and differential. The choice lay between either Hewland or ZF. Chapman and Hayes both preferred the latter, especially in view of Ford’s links with the German company through the GT40 programme.
When it came to the choice of which of their gearboxes to fit, it was simply a matter of sending ZF a DFV specification sheet of power and torque outputs. ZF was then able to match its most suitable ’box and gear ratios. “However, the gearbox had a major shortfall,” Costin says. “As part of the final-drive housing they’d designed the end plate as a flat surface, which you should never do because it would be doomed to failure. You always fit a cone-shaped plate.”
“I was accelerating hard, when the rear wheels left the ground at the join. They spun, then bit as they landed on the track”