Mike Costin tells the story of the engine on which the success of Cosworth was founded, both technically and financially
By Keith Howard
Formula 1’s most famous engine – the Cosworth DFV – might never have been created but for the development battle that its designer Keith Duckworth fought, and won, with an altogether less adventurous project. It was 1959, eight years before the DFV’s sensational first appearance in the back of the Lotus 49, and Cosworth was a fledgling company yet to make its name. Co-founder Mike Costin was still employed at Lotus while the new business grew sufficiently to justify his full-time involvement.
The engine that was to make or break Cosworth was the new 997cc Ford 105E in-line four, introduced for the Anglia. Duckworth saw in it the makings of a fine racing engine for Formula Junior, where it would be a natural successor to the ageing BMC A-series. So he set about its development while Costin pressed Colin Chapman to put it into the back of the new FJ Lotus 18 for 1960.
Duckworth’s plan almost foundered. A major valve train problem proved intractable to conventional solutions and very nearly drove Cosworth to an early bankruptcy. As Costin recalls, it was only by jettisoning established cam design theory that Duckworth saved the day and set Cosworth on a sound financial footing.
“I argued to use the 105E engine in the Lotus 18 on the score that Graham Hill was lobbying Colin to use the BMC A-series, produced by Speedwell, of which he was a director. My argument was that the long-stroke A-series was long past its prime and reaching the end of its development whereas the 105E was short-stroke, oversquare, lightweight, compact and obviously at the beginning of its development.
“Colin did finally see that this was the sensible course. I have it in mind, though I can’t find any evidence of it, that we actually built an 18 with a BMC engine, but perhaps that’s a flight of fancy.
“Colin was unaware of the development problems that Keith was having, which we hoped would soon be overcome as they were solely down to the camshaft. Keith had been reading Bishop, the acknowledged reference on camshaft design, but eventually threw the book away and decided to do it his own way. Things were getting pretty desperate by that stage because of the lack of durability of the cam and tappets, and the engine wouldn’t rev through 6000rpm because of valve spring surge.
“Keith overcame the problem by designing the cams his way, and that first cam stood us in good stead for many a year. We even sold them to other tuners, we didn’t keep them to ourselves. The A2 cam got us off to a good start and the A3 was even better. (The A1 was the one that didn’t work!) What Keith learned about cam design then carried us through the next 40 years.
“At Lotus we used to buy engines in batches of 10 from the local distributor and ship them down to Keith. He then stripped them down, modified them as necessary and tested them, and the Lotus van would then pick them up, presenting him with a cheque for £145 for each engine. Although when we began building the 18 we decided to make only 25 cars (and we’d only sold three by February 1960), by October we’d produced 125. So Keith had made 125 engines plus spares for Lotus, plus the engines he built for other Formula Junior manufacturers like Gemini and Elva. That really set the ball rolling for Cosworth.
“The Cosworth 105E originally developed 75bhp, then 85bhp and so on. For the best 105E we ever built – the one that Brian Hart and I used to race in our Lotus 22 – we bored through the top of the cylinder head to the back of the inlet valve and brazed in inlet pipes. We eventually got 123bhp out of that – 112bhp per litre; not bad for a 1098cc pushrod engine with two valves per cylinder and a bathtub head.
“We tipped Jimmy Clark’s 18 on its side at the Easter Monday meeting at Goodwood because the big end bearings had run. Lotus was still using a wet sump then, simply because we hadn’t yet developed a dry sump system, and there were oil surge problems. We modified the sump as best we could and replaced the bearings, but baffling sumps is a funny thing – something that is done more often badly than well. Even if you do understand what’s happening to the oil – and many don’t – it is still very difficult to control. A sump doesn’t work like a fuel tank, where you can put in a flap valve to keep fuel in the centre, because oil is coming in and being taken out at a dizzying rate, and it returns to the sump all over the place. When we did switch to dry sump we had to do a lot of development to get even the small amount of baffling there to do its job. Getting air out of the oil is a major factor, too.
“One reason why Jimmy’s car gave problems and Trevor Taylor’s didn’t – apart from the inherent variability of the surge problem itself – might have been because of the way Jimmy drove the 18. Although he could drive very smoothly, he revelled in cars that he could throw around and he loved that little FJ. When he went off once at the Goodwood meeting, he told me he was trying to go all the way round, except the chicane, without lifting off! That meant that at St Mary’s and Lavant he had to fling the car sideways, with the back end fairly well hung out. That might well have had an effect on the oil surge – but it’s only supposition.”