Ro dive values
Two Cosworth e gines, 40 years and 350bhp apart: Mike Costin, one half of the design team, relates the DFV’s origins and how much technology has changed i are problem solvers, and that gives them a different perspective. While the world at large remembers the DFV as the most successful racing engine of all time, Mike Costin, co-founder of Cosworth, remembers things differently. The hype
is unimportant to him. What he recalls is the engineering process — problems encountered, solutions found.
And the DFV had its problems, the most severe of which nearly spoiled that fairy-tale debut win in Holland. Solving that timing gear mystery took Cosworth into new territory, and culminated in an invention still widely employed today.
Improved materials, enhanced design and manufacturing techniques, pneumatic valve closure — all contribute to making today’s Fl V8s very different from the 40-year-old DFV. On its first dyno test Keith Duckworth’s 3-litre achieved 408bhp at 9000rpm, rising to 480bhp (160bhp/litre) at 10,500rpm near the end of its pre-eminence. Today’s Fl engines, from just 2.4 litres, generate around 750bhp (313bhp/litre) revving to a governed 19,000rpm. But while the engineering challenges are even sterner now than in 1967, Costin isn’t a fan of what almost limitless budgets have done to the design process. He’s content to have worked during racing’s most enjoyable era, and to have had studied at “the University of Duckworth”. DYNO PROBLEMS
MC “We had two initial problems with the DFV. The first was piston cracking. On each side of the piston we had four small holes for gas and oil to pass through, and they were acting as stress-raisers. A small modification to the forging cured that.
“The second problem was crankcase breathing. We weren’t getting oil out of the heads quickly enough, so it accumulated there. Initially, we had to add ugly external piping to get over this. We’d tackled the problem successfully in the FVA/FVB F2 engine and all our FJ engines before that, and carried the same principles over to the DFV. But the V8 had a much higher throughput of oil. Instead of three gallons we were putting 6.5 gal/min just through the bearings. The old rule that you used scavenge pumps of twice the capacity of the pressure pump didn’t work. We initially overcame the problem by fitting a twin-vane auxiliary pump on the end of the existing scavenge pumps. This had a good volume capability for its size and the scavenge system just about worked, but only just. Eventually Keith designed a complete new scavenging system using lobe pumps, like a Roots blower, with a capacity of 55gal/min. These removed so much air that we had negative pressure internally to tell the oil where to go. At the top of the engine we put a small hole to allow air in to give a controlled pressure difference across each section of the engine which guided the oil in and out. We had no more breathing problems.” SOMETHING WORSE
“The really big problem with the DFV was a timing gear torsional issue caused by torque spikes. It didn’t come to light until we started racing the engine. It could never have been foreseen, because its magnitude was just not appreciated at the time. Keith called in a Rolls-Royce engineer as a consultant who took one look at the timing gears and said there was no problem with them or the bearings or the pins in respect of taking the horsepower. So it must be something to do with stab torque — very brief transient torque peaks.
“We’d never done research into this but we put a proximity transducer on a gear pin and looked at its output on an oscilloscope. This allowed us to measure how the pin bent in reaction to the torque it was transmitting. Testing quickly showed that the mean torque was very low, as we expected, equivalent to 8.8hp. But for around 1.5 degrees, every now and again — quite often, in fact — there was a stab torque that went right off our scale. We estimated it to be higher than the mean torque from the crankshaft itself.
“A key problem was finding a fix that would fit into the space, which couldn’t be increased because the engine was an integral part of the car’s structure, so Keith designed the famous quill hub, using 12 tiny quill shafts. It accommodated the enormous stab torque through its springiness and had friction faces for damping, so it didn’t introduce a resonance problem. That cured everything. As far as I know, every F1 engine to this day still uses Keith’s quill hub.” BEHIND THE WHEEL
“I did all the testing on the FVB four-cylinder [effectively half a DFV] in a Brabham chassis — and the first two days of DFV testing, although that’s not in the history books. The first day was on the runway at Hethel, the second at Snetterton. Graham Hill then took over, but can’t have done much testing before Zandvoort otherwise they’d have uncovered the timing gear problem and might never have gone there with the 49.
“We didn’t do much on the first day because the clutch friction discs failed. We took the car back into the workshop, saw Bob Dance who was looking after the Cortina team, and asked him to lend us one of its twin-plate friction assemblies. That was a Cosworth unit — in fact, I’d designed it myself — and it became the standard DFV fitment.
“At Snetterton the next day I did 20 or 30 laps until the bracket bent where the top rear radius arm attached to the monocoque. That was a problem they had for a while with the 49. I remember touching 175mph on the back straight.
“Although I didn’t have any problem with the torque delivery, some drivers complained about it. The engine’s pickup was a bit splashy because we’d designed the fuel injection cam on the dyno and richened the mixture up a bit where you changed down. As each engine came back for service, the fuel cam went on the jig to check it, but one of Jack Brabham’s cams was quite different. He’d modified it. When we put it against the graph, we found it was exactly up the line we’d originally developed without enrichment. So we reverted to that and it improved the engine slightly. We didn’t tell Jack…” A DIFFERENT ETHOS
“I would say a good half of what is designed and manufactured these days is scrapped because of the manner in which things are tackled. We would never manufacture anything until we were pretty certain it would work. Keith, Ben Rood and I would talk for hours about a design while it was still a concept. Now it’s normal to scrap a high proportion of parts made because they will not assemble with other parts. “Budgets are higher and time pressures are greater now but you never know which pushes which. Teams go to sponsors and say, ‘We’ve used up all that money, it’s not enough — we need some more’. I don’t think they put it quite that way. Mind you, I started out as a young cynic and have steadily grown to be an old cynic.” ID