Bad vibrations

Cosworth was F1’s buzzword – but in sportscars its DFV was a flop, despite the efforts of a famously successful team. Four members of JWA explain why – and how – Le Mans was eventually won

It made perfect sense, to the layman: plop the era’s dominant Formula One engine in a two-seat racer, pop over to Le Mans – and win. To the creator of the Cosworth DFV, however, it made no sense whatsoever. Depending on his mood, he’d laugh or snort at the idea. If Ford had wanted an endurance engine, he would have designed an endurance engine. 

Keith Duckworth had, as usual, hit the valve on the head. Unusually, he’d be proved wrong – but only after four years of head-scratching and heartache for the big-name/okay-budget team that had to balance expectation against the imbalances resonating from this raucous V8’s flat-plane crank. From the moment its first DFV-powered Group 6 sports-racer, the Mirage M3 of 1969, rattled its gauges’ needles off their spindles at a Silverstone shakedown (shake-up?), JW Automotive, of GT40 and Le Mans fame, knew it was up against an insidious force: torsional vibration. It was able to put the problem on hold when Porsche came calling with an unrefusable offer – to run two 917s for it in 1970-71 – but it couldn’t be shaken off entirely.


The 5-litre behemoths have been outlawed and svelte ‘two-seater GP cars’ are the near future of sportscar racing. Ferrari and Alfa Romeo have proven their new, nimble machines in 1971. Indeed, Alfa’s T33/3 V8 already has two wins to its credit: Targa Florio and Watkins Glen. Back at 714 Banbury Avenue, Slough Industrial Estate, there’s a new name above JWA’s door: Gulf Racing Research Co Ltd. There have been other changes, too: founder and talisman John Wyer has taken a part-time role and team manager David Yorke has left. John Horsman is now MD and team manager, as well as chief engineer. He’s got a lot on his plate.

DEREK BELL (Driver, 1972-75): They were building a car from scratch. They hadn’t had to do that with any previous product, and it’s a bloody great difference.

HORSMAN: Maurice Gomm did the chassis, the engine was from Cosworth, the gearbox from Hewland and FKS did the bodywork. We sub-contracted a lot. Yes, we rebuilt our engines and transmissions, but we’d always done that. I don’t think we were operating very differently.

The M6, designed by Ford UK-paid Len Bailey, features a riveted-and-bonded aluminium tub reinforced by steel bulkheads. Front suspension is by lower wishbone/top link, rear by lower parallel links/top link. Its DFV is bolted directly to the frame. The car has its first run at Silverstone on March 14 and completes 25 laps before the mechanical fuel pump fails. With 169 miles under its wheels, it’s shipped to Sebring (March 25), where it retires after 48 laps (250 miles) and a litany of ills. Early niggles are to be expected and can (usually) be cured. There’s something more worrying.

BELL: The car wasn’t dramatic enough, didn’t have enough flair. We needed the input of a Newey-type. We did lots to improve it over the years, but it never gave me the feeling of, ‘Wow! That’s a second per lap.’ The basic car always felt the same.

JOHN WATSON (Driver, 1973): There were a lot of extremely good people at Gulf, but I sensed an extension of John Wyer’s 1950s philosophy. The early 1970s marked a fundamental swing in the philosophy of racing and I don’t think Gulf followed that. They’d learned a lot through their past successes, but perhaps they needed an influx of new design blood to mix with that experience. The Mirage felt like a heavy single-seater – not bad, but not good. 

HORSMAN: I agree that Len’s design was a bit conservative. Perhaps it was a reaction to some of the [reliability and aerodynamic] criticisms of his Ford F3L [of 1968]. That car had been quick straight away, the M6 wasn’t. It was a little beefier. Something might have been said to Len about keeping in mind that a V12 might be used at some point but, as far as I was concerned, the first M6 was designed for the V8. 

At Spa, JWA’s favourite circuit, the M6 loses 16min with a broken throttle linkage, but still finishes fourth. At the Nürburgring it gives the Ferrari 312PBs a scare – until its DFV lets go with two laps remaining. That Le Mans will be 18 hours too long for its engine is clear, so the car’s next outing is at Österreichring, where Bell takes pole but retires due to a misfire caused by a porous head. It’s a struggle.

BELL: [That vibration] was horrible. You knew it was beating hell out of the car. At the Nürburgring, the foam fill-in of my seat slipped from underneath me and left me braced directly against the chassis. I eventually threw up.

HOWDEN GANLEY (Driver, 1973): When I first tested the Mirage my legs became affected by the vibration and I couldn’t drive it late in the day. It was incredible. But you got used to it. Of course, the car was supposed to have a V12, but it didn’t work out like that.

While half the team is away at Watkins Glen, where Bell/ Carlos Pace finish third, the rest are back home preparing a chassis to accept Harry Weslake’s Ford-supported V12.

WATSON: In sportscars, the DFV loses a lot of its packaging gains. Enveloping bodywork means it matters less how long, wide and tall an engine is, plus there’s space to put the ancillaries where you want them rather than hiding them away as best you can. Also, in the search for reliability we had to knock back its maximum revs. That reduced power [420bhp at 9000rpm, 440 at 9700] and further narrowed an already peaky torque curve. The V12s were thirstier and heavier, but smoother and easier to drive, a great help over longer distances. You’d have thought the DFV’s weight and mpg would have been a boon in this form of racing, but Mirage’s package simply wasn’t as good as Matra’s. 

GANLEY: The Weslake had top-end grunt, was faster at the end of the straights, but it wasn’t as agile, or as quick over a lap, as the DFV. Matra, with all its resources, had been developing its V12 since 1968; Weslake couldn’t match that.

HORSMAN: The season finished early and we spent quite a bit of time doing back-to-back tests of the V8 and V12. We were pinning our hopes on the V12, but, really, we wasted that part of the year and it was a massive scramble to get ready for Daytona in January. We should have been further ahead with the DFV car. 


Bell qualifies on pole at Daytona, 2sec ahead of Matra, but both M6s retire. Horsman is confident his team has the V12s covered, but a new cylinder head puts Matra on top at Vallelunga in March and its MS670s spend the rest of the season vying for victory with Ferrari. Except at Spa, where Mirage scores a one-two, Bell/Mike Hailwood leading home Ganley/Vern Schuppan. 

GANLEY: The car was fantastic at Spa. It was an inherent understeerer, but in the quick stuff you could turn in early and it was fine. In the slow/medium stuff, however, you’d turn in and there’d be a slight delay. Disconcerting. I found a way around it: a quick flick to establish oversteer. We were constantly making changes to the nose in search of more downforce. We weren’t standing still.

HORSMAN: We gradually extended the nose as far as we could without it wrecking itself on the track. Don’t forget, we had suspension movement and body roll in those days. We used the wind tunnel at MIRA quite a lot. The Matra never looked that aerodynamic to me: very plain, blunt nose, slab sides. But it worked. Perhaps we should have copied it.

WATSON: I always felt that Matra, with its aerospace background, had a better aerodynamic package.

BELL: I didn’t think the Mirage was particularly understeery. By the time we finished the project we had a pretty neutral car. A bit of understeer early in a stint was perfect because, as things degraded, it would become a bit of oversteer. You have to be adaptable in long-distance racing. But when you had to push the car extra hard to keep up with the Matras and Ferraris, it struggled with that extra weight. 

GANLEY: I’d been with Matra in 1972 and knew how beautifully built those cars were. They were down near the weight limit [650kg]. At Vallelunga, I went over to chat with the Matra boys and, when they weren’t looking, picked up a nose-cone. It was much lighter than ours. Our car had to be built strongly to survive the DFV’s vibrations, but we could have done with hacking 200lb (90kg) off – and a lot of that was in the bodywork. But there wasn’t much lightweight technology in the UK at that time and, had we found it, it would have cost a fortune. I’m not sure we had the budget. 

HORSMAN: Gulf paid us whatever it took. It required trust on both sides, but it was a wonderful working relationship. We’d build the car, tell them the cost, and they’d send a cheque – as long as those costs were reasonable. The only thing they baulked at was carbon fibre. They didn’t want to get involved in that: too expensive and too time-consuming. Matra probably spent three or four times more than us.

It’s little wonder the French marque cruises to its second Le Mans win in a row – and that both Mirages retired. It’s been a useful learning exercise for the Gulf team, however. Its ZF transmission, in place of the usual Hewland DG300, suffers a sheared input shaft (the German firm’s engineers would consequently reduce its diameter to absorb the DFV’s buzz), and the long Mulsanne blast highlights a problem with the DFV itself: cavitation in the cross tubes linking the right-hand water pump to the one on the left. (A direct feed from the radiator to the latter will cure this.)