Customer power: the Cosworth DFV story
Author
- Darren Galpin
Date
- 8W Special, June 22, 1999
Related articles
- Engine failures - The worst engines of the 3.5-litre era, by Rainer Nyberg
Who?Jim Clark What?Lotus-Cosworth 49 Where?Spa When?1967 Belgian GP |
Why?
Jim Clark on his way to his first and the DFV's second consecutive pole of the 1967 season, after its stunning debut at the Dutch GP. The new Cosworth engine would come out on top on every qualifying session to follow in 1967, Clark taking four more poles and Graham Hill the remaining two. Just how the all-conquering Lotus team managed to leave its Saturday dominance uncrowned will remain the season's biggest upset. The answer of course is reliability, or rather, lack of it, as far as the Lotus team are concerned. And the wonder engine wasn't to blame, even though it still had to prove itself. Then, as in the many years following, the Ford Cosworth DFV has been a synonym for durability. Instead, Lotus let certain championships slip because of a variety of ignition, diff', clutch and suspension problems. When the 49s lasted the distance, they were usually far ahead. At the Glen, for instance, they walked the race, leading the opposition by a lap or more. But when they failed, there was one team to grab opportunity by its throat, running away with all the points. As it was, the tank-like reliability of the Brabham-Repcos sealed the championship for Denny Hulme and the Brabham team in very much the same way Ferrari picked up McLaren's pieces early this season - although outrun on sheer speed they managed to be there when it mattered, especially in France, Germany and Canada, while collecting useful points when they had to concede to the Lotuses. Never mind Clark's crushing end-of-season victories, the damage had already been done at the start of the season. In South Africa Lotus struggled with its cumbersome BRM H16-powered 43s, while the change-back to underpowered BRM V8 and Climax 33s for Monaco proved equally worthless, allowing Hulme to build enough of a championship lead to stay ahead till the end. Of course, for 1968, history didn't repeat itself, Lotus carrying its form through to Kyalami and a dominant display by Clark. Even the tragedy of Clark's death could not stop the Lotus-Cosworth steamroller as Graham Hill powered away to the 1968 title. The rest of the DFV's total of 155 GP victories is history...
At the beginning of 1966, Colin Chapman was on the lookout for a new engine to challenge those of Ferrari, BRM and Weslake. For an initial estimate of the cost, Chapman approached Keith Duckworth of Cosworth Engineering, who suggested a cost of around £100,000. Chapman approached a number of sources for finance before Ford finally agreed to the proposal, as it was keen to promote a new Total Performance image.
The plan was for Duckworth to first of all build a 1.8 litre, 4 cylinder Formula 2 engine based on the 120E Ford Cortina cylinder block, and then to double it up to make a 3 litre V8 Formula 1 engine. The engine had to have at least 400 bhp, and weigh less than 400 lbs (196 kg). In agreeing to the finance, Ford stipulated that Lotus would only have the rights to the engine's sole use for 1967, and after this, Ford was free to offer the engine to 'acceptable' customers. Thus a golden age of motor racing was ushered in with a contract signed on the 1st March 1966. However, Keith Duckworth didn't actually start the drawings for the engine until the 23rd June.
The Formula 2 engine used 4 valves per cylinder, a relative novelty at the time. BRM and Climax had both used this format with little success, although Honda had managed to exploit the technology more fully. The engine was a crossflow design, with intakes on the right, and exhausts on the left, with a 40-degree angle between the valves and a double overhead cam. A standard Ford camshaft was used in order to drive the water pump.
For the Formula 1 V8, the 400cc per cylinder was reduced to 375cc by reducing the stroke from 2.72 inches (6.91 cm) to 2.55 inches (6.48 cm), and thus it was necessary to increase the con-rod length in order to prevent excessive acceleration of the pistons, thereby reducing wear. It was decided that a V8 rather than a V12 was to be used as it had fewer moving parts, and thereby had less friction. The more friction there is in an engine, the more power has to be used to overcome it, and Duckworth decided that it was absolutely necessary to get as much power as possible. For lightness, an aluminium alloy (LM8WP) was used for the block, which necessitated the use of cast iron cylinder liners sealed with O-rings top and bottom. The cover castings either side of the 90° V were made from magnesium.
The engine was to be a structural member of the car, providing extra rigidity for the chassis, and it was unusual in being wider than it was long, being 26.5 x 21.6 inches (67.3 x 54.9 cm). At one end of the engine were suspension attachment points, and at the other were car attachment points at the top and bottom of the engine. The lower bolts were sited 9 inches (22.86 cm) apart, as the upcoming Lotus 49 was to have side pontoons sighted 9 inches apart in order to fit the body of driver Jim Clark perfectly! As the engine expands when it gets hot, the mountings were made so that the sheer forces were focused onto the bottom bolts. This way of mounting the engine survived all the way through to 1985, and was generally liked by the designers.
Due to the engine's width, it was difficult to put fuel piping around the sides of the engine, so it was dictated that the fuel tank had to be sited in front of the engine, inside the monocoque, a practice used today. Although this was forced onto the designer, it did have the benefit of locating a major source of weight close to the car's centre of gravity, and it enabled a constant weight distribution in the car as the fuel load went down. Due to the location of the fuel tank in the car, it was necessary to make the engine as short as possible in order to minimise the wheelbase of the car, thus avoiding a compromise in the car's handling. The flywheel on the engine was also made as small as possible, in order to maintain the low centre of gravity. This necessitated the starter motor being reversed and bracketed onto the gearbox casing.
The first dynomometer run gave a power output of 408 bhp, thus meeting Chapman's specification, but the engine did manage to pump out all of its oil. This was due to the cylinder heads filling up with oil, and not allowing the oil to drain into the sump. This was initially solve by cutting holes in the cam-covers, and piping the oil into special collector boxes on the side of the engine, before allowing the oil back into the sump. Lotus were forced to use this system for their first few races, but it was solved by fitting an air-pump to the engine, forcing the oil (and any gas which leaked around the cylinder heads and rings) back to the heads, and thus back into the oil scavenger. This caused the engine to run below atmospheric pressure, and the breather pipe drew air into the engine rather than letting the over-run out. The system fed so much air into the oil tank that the oil/air separators would frequently fail, causing engine failure. This was solved by using a larger scavenge pump fitted with its own air separator, drawing air through the heads, and pressurising it to force the oil to the sump. Separation systems then split the air and oil, passing the oil to the cooler and onwards. As the 70s progressed, ground effect meant that ever larger scavenge pumps had to be used, as the oil tended to accumulate on one side of the sump due to the lateral G-forces involved.
The major early problem was torsional vibration, which hit its natural frequency at 8600 rpm (with a harmonic at 9600 rpm). Although the test engine ran perfectly happily for many hours, a strip-down of Clark's engine after its first race showed that the vibration had caused Clark's cam-drive to strip a gear tooth. A new cam-shaft was designed to minimise vibration, and this profile was continually used throughout the engine's lifespan.
The initial track testing was performed by Mike Costin in Lotus 49 R1 at Hethel Aerodrome, before moving on to Snetterton. The engine manufacturers Cosworth wanted to make sure that everything was running OK before letting the drivers loose with it. Once Costin was satisfied, the testing duties were performed by Graham Hill, who actually raced the prototype chassis (designed to be as simple as possible in order to make engine development easier). Hill was quietly impressed with the engine, saying "Got some poke - not a bad old tool... it's certainly got some go!" after coming in after his first five-lap run.
The first race of the engine was at the Dutch GP in Zandvoort, the engine not being ready for the first two races in South Africa and Monaco. The result is now part of racing legend, Hill taking pole position 3.5s faster than the previous years time (before dropping out of 1st place on lap 12 with timing gear problems), and Clark winning the race from 8th position on the grid, setting fastest lap along the way. The was the first of 155 World Championship wins for the engine, all of which are summarised below.
Gear train problems still persisted as the the season wore on, and this forced Keith Duckworth to design a new hub for the gears. His new system featured splines which could wind up under load in order to cushion the gears and bearings against the torque reversal stresses. This was first fitted in 1971.
Once in general use by the racing fraternity, it soon became apparant that the sudden surge in power between 6000rpm and 6800rpm was difficult to handle, many an inexperienced driver being forced into a spin. To counteract this, Cosworth attempted to design a progressive throttle which would smooth out the powerband. However, increases in downforce and tyre width negated this problem, and the special throttle was never raced.
Airboxes on top of the engines became common during the 70s, and this aggravated the already high cylinder liner wear rate by scooping in debris with the air, forcing this mixture into the engines air intakes, and from there on into the cylinders. This then scored and marked the liners, causing some to catastrophically fail. Molybdenum-sprayed graphite iron liners were tried, but these cost power, so Cosworth reverted to cast iron. After a number of years Nikasil Aluminium liners appeared, and were subsequently fitted to all engines, as they also gave a 8 lb (3.9 kg) weight advantage.
In 1970, the celebrated 'bad batch' of crankshafts struck Cosworth and its DFV. The first crankshafts were machined from metal bars, but later developments saw them forged, and then forged from better materials in much more clean environments. By 1970 the crankshafts had a toughened nitrided surface, which required minor machining. An overenthusiastic machinist managed to grind right through this toughened surface, and repeated this with the entire batch of crankshafts, this batch being due to last Cosworth for the entire year. This fault was uncovered when crankshaft after crankshaft was destroyed when running in engines on the dyno, the crankshafts first cracking and then breaking. As it would take many months for Cosworth to get a replacement batch, Cosworth was forced to scavenge cranks from old engines, and mend them wherever possible.
As the engine aged, the inlet ports were enlarged and the trumpets shortened as revs and power were increased. This brought an increase in stress on the parts, and every part of the engine had to be looked at, the pistons, con-rods and big-end bolts all being revised. The shear number of competitors using the DFV meant that Cosworth subcontracted work to licenced engine builders. Of these, Nicholson-McLaren, Engine Developments (Judd) and Hesketh did their own engine modifications. These modifications were minor, and Cosworth's own work often superceded what these builders did themselves.
1982 was the final year that a DFV-engined car took the World Championship, and in 1983, a new short-stroke version of the engine was produced, known as the DFY. The cylinder heads were recast, and the inlet and exhaust ports were enlarged, resulting in a shorter exhaust system and intake trumpets. A new magnesium inlet manifold was also introduced. These changes reduced the engine's weight to 307 lb (150 kg), producing 520 bhp. These were first delivered to Williams and McLaren for the 1983 French GP. A second DFY was produced with a new cylinder head design, a narrower V, and a further 15 lb (7.3 kg) weight saving. A new crankshaft was produced, using four counterweights instead of eight, helping with the weight saving. Enlarged aluminium pistons were also used. The power curve of the engine had now been completely changed, with usable power being available from 6000 to 11,300 rpm, and no torque surge. The piston speed and acceleration had been reduced, minimising valve gear inertia, which meant that the engine's efficiency was much greater, allowing greater fuel efficiency relative to torque and power gains. Seven of these new engines were built for Tyrrell.
The DFY won the 1983 Detroit GP, the 155th and final World Championship win for the DFV lineage. Other derivatives of the engine were still raced, the 2.65 litre Turbocharged V8 DFX being used for Indycars, the 3.3 litre DFL V8 used for stockcars, and larger capacity versions being used in sportscars. The engine didn't totally die - once the Turbo Years were over, Cosworth produced the 3.5 litre DFZ for customer use, and a 3.5 litre works DFR. The DFV even continues on to this day - DFVs had found their way into Formula 3000 in the mid-80s, and can still be heard in the FIA Thoroughbred Grand Prix series.
The 155 DFV Victories
Date | Grand Prix | Driver | Car |
18 Jun 1967 | Dutch | Jim Clark | Lotus 49 |
15 Jul 1967 | British | Jim Clark | Lotus 49 |
1 Oct 1967 | US | Jim Clark | Lotus 49 |
22 Oct 1967 | Mexican | Jim Clark | Lotus 49 |
1 Jan 1968 | South African | Jim Clark | Lotus 49 |
12 May 1968 | Spanish | Graham Hill | Lotus 49 |
26 May 1968 | Monaco | Graham Hill | Lotus 49B |
9 Jun 1968 | Belgian | Bruce McLaren | McLaren M7A |
23 Jun 1968 | Dutch | Jackie Stewart | Matra MS10 |
20 Jul 1968 | British | Jo Siffert | Lotus 49B |
4 Aug 1968 | German | Jackie Stewart | Matra MS10 |
8 Sep 1968 | Italian | Denny Hulme | McLaren M7A |
22 Sep 1968 | Canadian | Denny Hulme | McLaren M7A |
6 Oct 1968 | US | Jackie Stewart | Matra MS10 |
3 Nov 1968 | Mexican | Graham Hill | Lotus 49B |
1 Mar 1969 | South African | Jackie Stewart | Matra MS10 |
4 May 1969 | Spanish | Jackie Stewart | Matra MS80 |
18 May 1969 | Monaco | Graham Hill | Lotus 49B |
21 Jun 1969 | Dutch | Jackie Stewart | Matra MS80 |
6 Jul 1969 | French | Jackie Stewart | Matra MS80 |
19 Jul 1969 | British | Jackie Stewart | Matra MS80 |
3 Aug 1969 | German | Jacky Ickx | Brabham BT26A |
7 Sep 1969 | Italian | Jackie Stewart | Matra MS80 |
20 Sep 1969 | Canadian | Jacky Ickx | Brabham BT26A |
5 Oct 1969 | US | Jochen Rindt | Lotus 49B |
19 Oct 1969 | Mexican | Denny Hulme | McLaren M7A |
7 Mar 1970 | South African | Jack Brabham | Brabham BT33 |
19 Mar 1970 | Spanish | Jackie Stewart | March 701 |
10 May 1970 | Monaco | Jochen Rindt | Lotus 49C |
21 Jun 1970 | Dutch | Jochen Rindt | Lotus 72C |
5 Jul 1970 | French | Jochen Rindt | Lotus 72C |
18 Jul 1970 | British | Jochen Rindt | Lotus 72C |
2 Aug 1970 | German | Jochen Rindt | Lotus 72C |
4 Oct 1970 | US | Emerson Fittipaldi | Lotus 72C |
18 Apr 1971 | Spanish | Jackie Stewart | Tyrrell 003 |
23 May 1971 | Monaco | Jackie Stewart | Tyrrell 003 |
4 Jul 1971 | French | Jackie Stewart | Tyrrell 003 |
17 Jul 1971 | British | Jackie Stewart | Tyrrell 003 |
1 Aug 1971 | German | Jackie Stewart | Tyrrell 003 |
19 Sep 1971 | Canadian | Jackie Stewart | Tyrrell 003 |
3 Oct 1971 | US | François Cevert | Tyrrell 002 |
23 Jan 1972 | Argentine | Jackie Stewart | Tyrrell 003 |
4 Mar 1972 | South African | Denny Hulme | McLaren M19A |
1 May 1972 | Spanish | Emerson Fittipaldi | Lotus 72D |
4 Jun 1972 | Belgian | Emerson Fittipaldi | Lotus 72D |
2 Jul 1972 | French | Jackie Stewart | Tyrrell 003 |
15 Jul 1972 | British | Emerson Fittipaldi | Lotus 72D |
13 Aug 1972 | Austrian | Emerson Fittipaldi | Lotus 72D |
10 Sep 1972 | Italian | Emerson Fittipaldi | Lotus 72D |
24 Sep 1972 | Canadian | Jackie Stewart | Tyrrell 005 |
8 Oct 1972 | US | Jackie Stewart | Tyrrell 005 |
28 Jan 1973 | Argentine | Emerson Fittipaldi | Lotus 72D |
11 Feb 1973 | Brazilian | Emerson Fittipaldi | Lotus 72D |
3 Mar 1973 | South African | Jackie Stewart | Tyrrell 006 |
29 Apr 1973 | Spanish | Emerson Fittipaldi | Lotus 72E |
20 May 1973 | Belgian | Jackie Stewart | Tyrrell 006 |
3 Jun 1973 | Monaco | Jackie Stewart | Tyrrell 006 |
17 Jun 1973 | Swedish | Denny Hulme | McLaren M23 |
1 Jul 1973 | French | Ronnie Peterson | Lotus 72E |
14 Jul 1973 | British | Peter Revson | McLaren M23 |
29 Jul 1973 | Dutch | Jackie Stewart | Tyrrell 006 |
5 Aug 1973 | German | Jackie Stewart | Tyrrell 006 |
19 Aug 1973 | Austrian | Ronnie Peterson | Lotus 72E |
9 Sep 1973 | Italian | Ronnie Peterson | Lotus 72E |
23 Sep 1973 | Canadian | Peter Revson | McLaren M23 |
7 Oct 1973 | US | Ronnie Peterson | Lotus 72E |
13 Jan 1974 | Argentine | Denny Hulme | McLaren M23 |
27 Jan 1974 | Brazilian | Emerson Fittipaldi | McLaren M23 |
30 Mar 1974 | South African | Carlos Reutemann | Brabham BT44 |
12 May 1974 | Belgian | Emerson Fittipaldi | McLaren M23 |
26 May 1974 | Monaco | Ronnie Peterson | Lotus 72E |
9 Jun 1974 | Swedish | Jody Scheckter | Tyrrell 007 |
7 Jul 1974 | French | Ronnie Peterson | Lotus 72E |
20 Jul 1974 | British | Jody Scheckter | Tyrrell 007 |
18 Aug 1974 | Austrian | Carlos Reutemann | Brabham BT44 |
8 Sep 1974 | Italian | Ronnie Peterson | Lotus 72E |
22 Sep 1974 | Canadian | Emerson Fittipaldi | McLaren M23 |
6 Oct 1974 | US | Carlos Reutemann | Brabham BT44 |
12 Jan 1975 | Argentine | Emerson Fittipaldi | McLaren M23 |
26 Jan 1975 | Brazilian | Carlos Pace | Brabham BT44B |
1 Mar 1975 | South African | Jody Scheckter | Tyrrell 007 |
27 Apr 1975 | Spanish | Jochen Mass | McLaren M23 |
22 Jun 1975 | Dutch | James Hunt | Hesketh 308 |
19 Jul 1975 | British | Emerson Fittipaldi | McLaren M23 |
3 Aug 1975 | German | Carlos Reutemann | Brabham BT44B |
17 Aug 1975 | Austrian | Vittorio Brambilla | March 751 |
2 May 1976 | Spanish | James Hunt | McLaren M23 |
13 Jun 1976 | Swedish | Jody Scheckter | Tyrrell P34 |
4 Jul 1976 | French | James Hunt | McLaren M23 |
1 Aug 1976 | German | James Hunt | McLaren M23 |
15 Aug 1976 | Austrian | John Watson | Penske PC4 |
29 Aug 1976 | Dutch | James Hunt | McLaren M23 |
12 Sep 1976 | Italian | Ronnie Peterson | March 761 |
3 Oct 1976 | Canadian | James Hunt | McLaren M23 |
10 Oct 1976 | USA East | James Hunt | McLaren M23 |
24 Oct 1976 | Japanese | Mario Andretti | Lotus 77 |
9 Jan 1977 | Argentine | Jody Scheckter | Wolf WR1 |
3 Apr 1977 | USA West | Mario Andetti | Lotus 78 |
8 May 1977 | Spanish | Mario Andretti | Lotus 78 |
22 May 1977 | Monaco | Jody Scheckter | Wolf WR1 |
5 Jun 1977 | Belgian | Gunnar Nilsson | Lotus 78 |
3 Jul 1977 | French | Mario Andretti | Lotus 78 |
16 Jul 1977 | British | James Hunt | McLaren M26 |
14 Aug 1977 | Austrian | Alan Jones | Shadow DN8 |
11 Sep 1977 | Italian | Mario Andretti | Lotus 78 |
2 Oct 1977 | USA East | James Hunt | McLaren M26 |
9 Oct 1977 | Canadian | Jody Scheckter | Wolf WR1 |
23 Oct 1977 | Japanese | James Hunt | McLaren M26 |
15 Jan 1978 | Argentine | Mario Andretti | Lotus 78 |
4 Mar 1978 | South African | Ronnie Peterson | Lotus 78 |
7 May 1978 | Monaco | Patrick Depailler | Tyrrell 008 |
21 May 1978 | Belgian | Mario Andretti | Lotus 79 |
4 Jun 1978 | Spanish | Mario Andretti | Lotus 79 |
2 Jul 1978 | French | Mario Andretti | Lotus 79 |
30 Jul 1978 | German | Mario Andretti | Lotus 79 |
13 Aug 1978 | Austrian | Ronnie Peterson | Lotus 79 |
27 Aug 1978 | Dutch | Mario Andretti | Lotus 79 |
21 Jan 1979 | Argentine | Jacques Laffite | Ligier JS11 |
4 Feb 1979 | Brazilian | Jacques Laffite | Ligier JS11 |
29 Apr 1979 | Spanish | Patrick Depailler | Ligier JS11 |
14 Jul 1979 | British | Clay Regazzoni | Williams FW07 |
29 Jul 1979 | German | Alan Jones | Williams FW07 |
12 Aug 1979 | Austrian | Alan Jones | Williams FW07 |
26 Aug 1979 | Dutch | Alan Jones | Williams FW07 |
30 Sep 1979 | Canadian | Alan Jones | Williams FW07 |
13 Jan 1980 | Argentine | Alan Jones | Williams FW07 |
30 Mar 1980 | USA West | Nelson Piquet | Brabham BT49 |
4 May 1980 | Belgian | Didier Pironi | Ligier JS11/15 |
18 May 1980 | Monaco | Carlos Reutemann | Williams FW07B |
29 Jun 1980 | French | Alan Jones | Williams FW07B |
13 Jul 1980 | Britain | Alan Jones | Williams FW07B |
10 Aug 1980 | German | Jacques Laffite | Ligier JS11/15 |
31 Aug 1980 | Dutch | Nelson Piquet | Brabham BT49 |
14 Sep 1980 | Italian | Nelson Piquet | Brabham BT49 |
28 Sep 1980 | Canadian | Alan Jones | Williams FW07B |
5 Oct 1980 | USA East | Alan Jones | Williams FW07B |
15 Mar 1981 | USA West | Alan Jones | Williams FW07C |
29 Mar 1981 | Brazilian | Carlos Reutemann | Williams FW07C |
12 Apr 1981 | Argentine | Nelson Piquet | Brabham BT49C |
3 May 1981 | San Marno | Nelson Piquet | Brabham BT49C |
17 May 1981 | Belgian | Carlos Reutemann | Williams FW07C |
18 Jul 1981 | British | John Watson | McLaren MP4 |
2 Aug 1981 | German | Nelson Piquet | Brabham BT49C |
17 Oct 1981 | Caesars Palace | Alan Jones | Williams FW07C |
4 Apr 1982 | USA West | Niki Lauda | McLaren MP4B |
9 May 1982 | Belgian | John Watson | McLaren MP4B |
23 May 1982 | Monaco | Riccardo Patrese | Brabham BT49D |
6 Jun 1982 | USA | John Watson | McLaren MP4B |
18 Jul 1982 | British | Niki Lauda | McLaren MP4B |
15 Aug 1982 | Austrian | Elio de Angelis | Lotus 91 |
29 Aug 1982 | Swiss | Keke Rosberg | Williams FW08 |
25 Sep 1982 | Caesars Palace | Michele Alboreto | Tyrrell 011 |
27 Mar 1983 | USA West | John Watson | McLaren MP4/1C |
15 May 1983 | Monaco | Keke Rosberg | Williams FW08C |
5 Jun 1983 | USA | Michele Alboreto | Tyrrell 011 |