Currently reading: Throwback Thursday: Servicing Cosworth DFV Formula 1 engines, 1975
The then-ubiquitous DFV was a brilliantly effective design, finding success over nearly two decades. We found out how often it needed attention
6 mins read
22 June 2017

Never before had there been, and in all likelihood never again will there be, an engine like the DFV.

Designed in 1965 by the youthful Cosworth company, its birth was down to Lotus’s Colin Chapman, who both commissioned the project and brought in the crucial backing from Ford.

The DFV made a winning debut at the 1967 Dutch Grand Prix and, continually evolving, took 12 Formula 1 drivers’ championships and 10 constructors’ championships, taking its last win in 1983 and making its last appearance in 1985. At times during the 1970s, nearly the entire grid had the cheap, reliable and competitive DFV bolted in the back.

And it's that reliability part we’re looking back on today, because on 2 August 1975, Autocar’s Michael Scarlett paid a visit to see how the DFVs were maintained at Nicholson McLaren Engines Ltd, run by racer-cum-engine preparer John Nicholson.

At that time, the DFV was a 3.0-litre, overhead-camshaft, 32-valve lightweight V8 outputting around 475bhp at 10,250rpm.

“One is so blasé about engine speeds... that one forgets how incredible it is that such a device should work so powerfully,” Autocar began.

“The power figures quoted for production engines rarely exceed 66bhp/litre. The specific output of the DFV amount to nearly thrice that, at 158bhp/litre.

Back to top

“Basically it achieves that by persuading enough fuel-air mixture into cylinders and burning it efficiently enough at the highest possible crankshaft speed. At that moment, from what John told us, with the latest cylinder heads and larger exhaust pipes and recently introduced taper intake trumpets (plain cylindrical for most of their length), it still doesn’t reach peak power at 10,750rpm, proving that it isn’t breathing that restricts but the mechanical limits of the unit.”

“The normal working range is between 8500 and 10,500rpm. An electronic ignition cut-out is fitted which nominally stops the engine being revved beyond about 10,600rpm. Firstly because even these relatively sophisticated rev-limiters gradually lower the limit during their life, and secondly a rev-limiter cut-out switch is provided because a few hundred more rpm can be allowed to the driver desperate for that little bit extra. Drivers do go to 11,000rpm, but it isn’t recommended.

“At that speed, at the bottom of one of the eight pistons’ stroke, it is stationary. It has to be accelerated from that brief rest to near-enough 75mph in 1¼in, then slowed down again in the same distance. And it has to do that each way every three-thousandth of a second. The acceleration involved is around 230g.”

So, how long between rebuilds?

Back to top

Back in the days where truly tiny independent teams were still common in Formula 1, this varied a lot. McLaren tended to rebuild engines after 600 miles – or one weekend, practice and race.

Showing the remarkable evolution of the DFV, Nicholson said: “When it first came out in 1967, we were probably getting 1000 miles out of them. OK, we were running to 10,000rpm then, but they were producing say 420 to 440hp. Now they’re running to 11,000rpm, really, and that extra is not good for them.”


What wears?

As providers and maintainers of race-worthy engines, Nicholson Ltd’s job primarily was not to measure the wear of each part, but to avoid the results of a crucial breakage.

“Pistons get replaced every 1000 miles,” Nicholson said. “They won’t do two race weekends. If you had a failure because of a piston, which in turn broke the engine – even a secondhand one that cost £4000 – that would have bought you a set of pistons for every race in the season, so it isn’t worth economising here.”

Valves and valve springs, meanwhile, would last two race weekends.

“Crankshafts, with a bit of luck, will run a long time, depending on how they wear,” Nicholson said. “I’ve only seen one or two cranks break in the four to five years I’ve been doing DFVs. They wear on the thrust faces of the webs – the end of the journal where the side of the rod rubs. There are different thoughts on why this happens. The end float on a pair of rods in place is set between six and 12thou, but it’ll increase to 15 or 20thou very quickly – in one run. I think it’s something to do with the nitridising of the crank – it’s too hard, it splinters, and the splinters dig into the con rod, which isn’t particularly hard compared with the crank. Then the con rod works like a little grinding wheel.”

Back to top

Con rods themselves, Nicholson said, he could never get hold of – although fortunately they didn’t fatigue too quickly. Camshafts, while prone to wear a bit at the lobes, would last an entire season. He'd never seen cam followers break.

“Cylinder blocks have a fair old life. If a main bearing tunnel or cylinder liner register gets out of shape – which they do – we rebore to suit a 15 or 20thou oversize bearing or liner. This lets you use the block, if all goes well, for a couple of years.” The cylinder heads lasted longer still, although Nicholson did have to replace the valve guides once a year.

Bore wear wasn’t an issue, due to the teams’ eventual acceptance that an air-cleaning device didn’t have to siphon horsepower.

Originally, DFVs’ throttle controls would cause a lot of trouble, especially their metering unit rods, but, said Nicholson: “now we have our own little rod-end jointed linkages instead of the ball ones, and we haven’t had one break since”. Metering units were changed regularly.

Back to top

A standard rebuild and its cost

A full Nicholson rebuild of a DFV cost £800, (around £6200 today), with the engine itself costing £10,000 new (£77,000). Today’s Formula 1 V6 hybrid engine costs into the millions.

A standard rebuild in 1975 involved two of Nicholson’s fitter-mechanics, “who start by stripping the engine down and cleaning it for inspection. The user should have supplied mileage and maximum revs used.

“Nicholson has a Magnaflux crack-testing rig, on which crankshaft and connecting rods are tested. Pistons (at £25 each) and possibly valves (£7 each, and there are 32 of them), and valve springs (£4 each) are replaced.”

Nicholson explained: “We have a kit of gaskets, oil seals, bearings for oil pumps, water pumps and all those sort of things – we automatically put them in. Maybe I overdo things a bit, but it’s the way McLaren works and it’s the way I’ve been taught. I probably spend £50 more on seals and gaskets and bearings than I have to – but I think that’s right. A team has half a dozen mechanics, thousands of pounds of transporter and equipment, and out goes the car and brm-brm-brm that’s it, all for the sake of a £2.50 bearing.”

Back to top

Still, though, Autocar wrote: “It cannot help being an expensive business, when parts are not made in mass production quantities, and when they are made much more carefully.

“A crankshaft costs £518; connecting rod £55; camshaft £78; cam follower £3; flywheel including ring gear £60; cylinder liner £19; scavenge oil pump £380; water pump £75. A kit of gaskets, seals and minor bearings is around £89, and labour costs amount to at least £400. And if things do go badly wrong, there’s always a set of pistons £239; a camshaft carrier £234; a cylinder head £590; or a cylinder block £1850.

“A normal 6000-mile service for a Ford Capri 3-litre is around £17. Doesn’t seem quite so bad now, does it?”

Nicholson McLaren is still a thriving business. It still rebuilds engines, both for racing cars and now for aeronautical purposes, as well as precision machine engineering of components and one-off reverse-engineered parts.

Join the debate

Add a comment…
bomb 22 June 2017

Really great article, thanks.

Really great article, thanks. Seeing the costs broken down was fascinating reading and lays bare the immense cost of motorsport even when it was 'cheaper' in the 70s.

Reading up on the Mirage effort in the 70s recently and how long it took to make a suitably reliable DFV to work for sportscars. Vibration over the longer distances than those seen in F1 was a big problem and many decent finishes were lost through engine failures of one kind or another. Of course they managed to win Le Mans in 1975 with a detuned DFV but much Gulf money had been spent getting there.

The Tyrrell team using the short-stroke DFY development engine in 1983 just before they switched to Renault turbos. They were higher revving and put out about 515bhp, can't imagine the rebuild costs on those.

johnfaganwilliams 22 June 2017

RE costs

pretty sure that Brian Hart charged about £27k for a re-build of a sports car DFV. Sounds absurd so maybe my brain has gone but "Nosher" certainly knew how to charge.
Straff 22 June 2017

Lovely engine

The sound of a DFV really is wonderful, too. Motorsport has lost its way over the last thirty years; too many Manufacturers spending a fortune, then whining about the cost and getting out, and far too much corporate hospitality. Rallying has been ruined in this way.

Oh, and thanks for using the word 'thrice'. I haven't heard that in ages!

MarkII 22 June 2017

Nice to reminisce about the glory days...

Enjoyed reading this article. Nice to reminisce about the DFV, which was without doubt a landmark piece of british engineering. I agree with @johnfaganwilliams though - sadly Formula 1's glory days are long gone. I stopped travelling to watch F1 about 15 years ago, it had all become too boring and too corporate, with access limited to VIP guests and those with only a passing interest in motorsport but with deep pockets and a desire to be seen in the right places. I sometimes watch a few races on TV now but I'd much rather watch GT, touring cars or rallying, which are more accessible and more real.