The Cosworth Turbo
Keith Duckworth and company talk about their turbocharged V8 USAC engine
Surprising as it may seem in view of Cosworth Engineering’s role in Formula One, Keith Duckworth is all for changing the formula. For some years now he has been advocating that Formula One engines should be freed of all restrictions, save a control over the amount of fuel they consume. How this would be achieved, and why he feels it important to the future of motor racing, was the subject of a fascinating conversation between two of our staff and Duckworth recently. However, the reason D.S.J. and I had faced the first snow of 1978, in his uniquely ventilated E-type [It has a hole in the rear screen.-.-.Ed.], was to see and hear how Cosworth were progressing with the development and construction of the 2.6-litre DFX turbocharged engine.
At the time of our visit 18 of these methanol-burning units were scheduled for delivery to the majority of competitive teams in the American USAC formula, another five anticipated for delivery in Spring to cater for serious Indianapolis 500 contenders. Each engine (seven had been delivered by early January) cost about £14,600, but the strengthening pound and the turbocharger installation —which is the responsibility of the team concerned in America—must take the value of each engine towards £20,000 when installed.
The DFX story is a confusing tale. Not only are a number of American teams entitled to their share of the credit, but the Cosworthbuilt engines were still evolving within the batch currently under construction. The specification has been frozen now, but the engine is still comparatively young in Cosworth development terms, and it would be naive not to expect further change. Already the Cosworth-built DFX shares far less than would first appear to be the case with the Grand Prix 3-litre DFV (Double Four Valve) unit, which has enjoyed such a spectacularly successful and long life since it first appeared at Zandvoort in 1967.
Cosworth sales manager Jack Field joined us in conversation as Duckworth’s strong Northern tones observed, “Fm sure there’s a conspiracy to keep me from knowing how many engines we are building by altering the starting dates! I don’t want to be anything but an engineer.” As you would expect, Duckworth has a good memory which is soon prompted into counter-action as Field recalls some of the early days!
The initiative for the engine came from Roger Penske. In 1974 he wanted Cosworth to become involved in turtx)charging their GP unit. At that time, and for a surprisingly long period afterwards, there were people who really felt it would be as simple as strapping a turbocharger onto a shorter stroke (for reduced capacity) version of the GP engine.
Nobody at Cosworth seemed to be under those illusions and the production of suitable crankshafts, connecting rods and pistons, plus thicker wall liners for the aluminium cylinder block, was finalised with a test session on one of Cosworth’s three water-braked dynamometers. This was early in 1975, but Penske had faded (temporarily) from the picture and Cosworth only tested the unit on petrol, minus turbocharger and running a compression ratio below 8 to 1. When Penske dropped out Cosworth were left with the idea that they should break into USAC racing, some parts sitting on the shelves, and the management sense to know that the last thing they wanted was any kind of big publicity blast to herald their American intentions.
Throughout development of the DFX this low key approach has been essential. They were certain that, should they put together the feasible “good Indy engine” the success they gained would be swiftly rewarded with a ban on the use of the engine. Even with the quiet way in which the engine has infiltrated USAC, to the point where it is now the engine to have, it has not escaped the notice of the establishment. There were moves to outlaw the engine as late as last November, with a final decision deferred until January. By that stage so many of the establishment teams had committed themselves to DFX that the odds were tilted in favour of Cosworth.
Earlier in the life of the DFX George Rignotti, the legendary master mechanic within the halls of USAC: fame, tried a cunning ploy. Now team manager of the biggest rivals to the DFX chances of success in 1978, Bignotti then proposed that USAC should allow boost pressures to escalate for the 1976 season. This, it was felt, would see the British V8 explode under practice host pressures equivalent to 25 psi., while the beloved Offenhauser four would remain reliable. In fact it has been another ingredient in the Cosworth unit’s success!
The Vels Parnelli Jones USAC, team were the first people to take the Cosworth parts, and they took them in a complete engine. The new-fangled British power-pack was delivered around the time of 1975’s Indianapolis 500, but first saw action in the last race of the USAC season. Held at Phoenix, Arizona, the event was not a had debut for the DFX for it finished fifth.
The Americans decided that the engine needed much bigger gudgeon pins, and this decision “took half of our bits out of the engine straight away,” says Duckworth. He continues, “This meant having great big rods. That made the balance factor hopeless so they needed a crankshaft as well. It was a real hybrid with American exhaust valves, Offy water pumps, Weaver oil pumps and Hillborn injection.”
In fact the following season, 1976, saw the DFX under very serious development. The VPJ people used such units for a 13-race programme, taking three wins and finishing in the top five on five occasions. Also during 1976 Cosworth and McLaren-Penske began their separate work on the engine. Penske and Teddy Mayer signed a joint agreement to produce their version of the DFX mmtilising McLaren’s facilities at Detroit.
McLaren naturally Started work using old DFVs and Cosworth conversion parts for the capacity difference, but they swiftly went their own way. Connecting rods, pistons and a new crankshaft design (the same American supplier as VPJ, but a different shaft!) marked a third separate development path for the DFX,
Duckworth says of that period, “We knew the context in which they were using the engine, but—as with everyone else that we have dealt with in America—there is absolutely no feedback. At first we decided just to co-operate and watch it gently mature with the rising interest to keep it accepted. We wanted to make sure that it was not too touch of a threat to the establishment at any particular time, for that could lead to a complete ban.”
No official starting date for work on the Cosworth DFX engines was given, but the results of the work in 1976 were first assessed on a complete Cosworth-built DFX in January 1977. That engine illustrated to Duckworth that there were problems still to overcome with the piston rocking on the big godgerm pin American connecting rods. Duckworth commented, “Those huge pins, they were great one-inch horses to do the job.”
Since then all kinds of combinations seem to have been tried within. Current engines contain more Cosworth DFV philosophy (on clearances and sizes, on both sides the Americans going for “biggest is best”) an original DFX crankshaft mating to new cross-section connecting rods, though the rods retain DFV caps. Some of the earlier engines in the current batch are to have American-made Tuftrided crankshafts and connecting rods.
A Cosworth piston was specially developed for the engine and is now installed. To cope with the extra heat a thicker crown is featured as well as their own thoughts on skirting and the provision for the in. pin sizing. As with all Cosworth pistons, the -alloy forgings are in RR58 material.
Duckworth says of the reciprocating parts (a new flywheel is also part of the recipe), “They are more in line with our thinking now. Some of them are big by our standards, but the clearances and so on are now much more Cosworth than before. The bigger small end and stronger rod were a direct result of our research, which showed that there were far greater tilting forces than before. Reverting to the original crank and using the new rod has brought the weight within reason, and we finally seem to have found some transverse bending strength at the small end.” In discussing all these components it is as well to remember that a slightly smaller Version of their GP engine is being asked to reliably produce almost double the existing power : more if the driver turns boost up in a final sprint for the line. No USAC restriction is placed on boost used during racing, so outputs of 900 to 1,000 b.h.p. can he reached momentarily, all from 2,642.7 c.c. The components providing a 57.3 mm. stroke and 85.67 mm. bore certainly do have some stress to bear.
Water and oil circulation within the engine is improved, but Duckworth is careful to emphasise that such increases are small. “We don’t flow a lot more oil than the DIN,” he Comments, “hut there is increased blow-by on turbo engines and the lubrication of the turbo itself to consider: the blower needs to he supplied with about half a gallon a minute” (oil consumption is rated as a gallon for 500 racing miles). Current engines have a tidy new combined scavenge and magnesium development DIN pressure oil pump, the latter now standard on 1978 Grand Prix engines. The oil pump increases pressures by 20 psi. compared with older DFVs, so maximum pressure would be about 110 psi.
This figure is quoted for normal running speeds which go up to 9,500 r.p.m. so far as Cosworth are concerned, but most engines seem to be revved to over 10,000 r.p.m. at some stage, baffling those at Cosworth. As Project Engineer Paul Morgan commented later, “The power drops off very sharply over nine-five.”
Cosworth have produced their own improved water pump for this engine (not shared with current DFVs) but more stress was laid on the revised cylinder-head water passages. More cooling is provided around the exhaust valves and a new (unspecified) metal is used for the valve seats: the exhaust valves are as standard DFV, although this is the side that really takes the terrific heat punishment, which is the penalty of a responsible turbocharged racing unit. The cylinder bloek carries thicker cast iron
liners at present. McLaren use steel liners and research is still going on into this subject at Cosworth. Another change to the bottom end of the engine is the use of a cross-dowelling technique to reinforce main bearing efficiency under constant stress. Full throttle running, for half an hour, is a recent part of Cosworth’s tests on the effectiveness of this modification and the new connecting rods.
As ever, camshafts are the subject of ‘great debate. Interestingly Cosworth use their own BDA Escort RS rallying profile (the BD3) on the exhaust side and a new DA2 profile for the inlet camshaft. Duckworth became truly animated at this point, leaving Jenks almost goggling and the writer with a strained wrist as he described his theories_ Disarmingly Duckworth always looked puzzled when anyone used the words “Of course” or “it is generally accepted that”. His business has been built on conventional principles, but he does not accept theory without proving it for himself. Even as he had apparently proved a point to the onlooker’s satisfaction he was ObviouSly mulling over further questions provoked by that apparent proof. On the camshaft timing for the turbo engine he said: “It makes sense to have different profiles side to side. The pressures are different. You can get more into the engine in a shorter time with blowing, so you can have narrower valve timing. In order to improve response, you have to have surplus exhaust turbine work available
“Another turbo point is that you can’t test on a dyno, you know?” Duckworth beamed happily looking for the return of confusion in his audience. We looked ‘suitably wise, or baffled on the non-Cosworth side of this extraordinarily jovial discussion. Seeing that we were not going to disobey the Duckworth Maxim that covers opening one’s mouth to confirm you are a fool he said, “You simply cannot reproduce the racing circumstances_ of a Wiggly circuit. On such a track there :are transient periods that you cannot explore on a. test bed. In the States they have got to the point of choosing the right -size of turbocharger for each course. We rim up to the USAC limit (equivalent to 5(1 in. -of mercury) using the Air Research TO6 unit on the dyno. For customers’ engines we do not run beyond this point, or the rev limit.
“I’m Sure that’s why Renault have all this trouble with their turbo; they just didn’t hit those circumstances before. You got problems even just when the driver lifts-off sharply. At Ontario Penske had a really disastrous evening session. Because it was cooler, and the air mass rather denser than usual, the car was running faster than ever before. As the sun set it got in the driver’s eyes and he started to back off the throttle suddenly and earlier . Another hearty chuckle as Duckworth reported.: “And that was two engines up the slot : they get through ’em you know!” Patti Morgan (the son of the gentleman who rebuilt Neil Corner’s GP Opel) was also interesting on the subject of mixtures. “You’ll have seen that the two authentic American features of the engines are Mallory magneto ignition and the Hillborn injection. Now Hillborn fuel injection has very Om* provision for mixture control, but the whole problem with USAC racingis that we have to run about 20% rich. If you do not do this you can seriously datnage the .erigine. WhereaS a
El petrol engine will survive a degree of testbed advance and retard cif the ignition, and traverse the whole mixture spectrum from full rich to full weak. The problem is that the fuel system delivery does not match the engine power curve. To keep the engine in one piece we have to run 20% rich at MOO r.p.m. Unfortunately this means the engine runs 30% rich at 7,000 r.p.m., and this wastes fuel.” Naturally, one wondered, why use the Hillborn system?
That was pungently -answered by Duckworth earlier on when be Said, “The HillbOrn system has to deliver two gallons of methanol a minute. There is no other injection system .Capable Of handling that sort of flow, in fact you’d have a job to get it out of a cold water tap!” In OCtOber 1977 CoSI,Vortb’S first complete engine of the new breed was behind Bobby Unser as he qualified on pole position for the Phoenix USAC round. This was the first engine that they were sure contained their thinking and workmanship. Derivatives of the DFX from the VPj, Penske and McLaren teams accounted for a total of five cars (the fifth was an irregular mount for Andretti) and enjoyed a good year. There were eight
engines and spares . .. in advance! This late order had to be turned down for capacity reasons. Of course it was a difficult decision to make, because some of that money may go to the development of the Offenhauser Vg.
Incidentally some Of the Cosworth DFXs will go direct to Goodyear for tyre testing sessions, and it’s possible that hard-pressed teams might field one of these units in a race.
Since November 1976 Cosworth have retained the services of one Larry Stutter. An American engine man who is familiar with that country’s racing equipment, he has proved the ideal link between Northampton and North America. Stutter :spends the season in the USA and ha-s a base at Cosworth Engineering in Torrance. California, right on the doorstep of Parnelli Jones. This arm of Cosworth is a separate company to the British parent, its business to look after much of the routine rebuilds (every 600 miles) and Cosworth’s considerable four-cylinder engine business in America. Fed on straight methanol and operating a nominal 7.4-to-1 Cr., these 90-degree V8s are in another world to their OP cousins. Instead of 11,000 plus r.p.m. and 475 b.h.p., the conservative (corrected) horSepower figure
victories, Tom Sneva’s USAC Championship for Penske and pole position at Indianapolis to record. Of course Foyt won Indianapolis for the fOurth time, but the DFX-powered cars were second and third.
Next year the cream of American racing seems lined up to go either directly ?vith Cosworth, or other developments of the DFX. Receiving Cosworth engines will be VPJ (a car for Danny Ongais); Penske (two cars) who will cover all the rounds for Sneva and the 500-milers for Andretti in the second car. The Cobre ‘rice Racing Team are expected to still run a car for the injured Pancho Carter with Cosworth engines, and also to become increasingly involved in the routine rebuilding of other engines as the season progresses. All American Racers, or whatever title Dan Gurney settles upon for a car from Santa Ana, will be with Cosworth as well, Bobby Unser driving. Brother Al Unser is to drive for Jim Hall in a new Lola chassis and he will also use units of Cosworth origin. There are also three cars for privateers on the books: that they are getting their engines at all is a reflection of the conipany’s integrity. Cosworth can only build so many units; one of the bigger teams came along a bit late and offered a quarter Of a million dollars for six
at 25 psi. boost is -850 on 9,500 r.p.m. Maximum torque is -500 lb. ft.. according to Cosworth. It is worth remembering that under race conditions the driver can literally dial up the boost at which he thinks the engine will survive, though the most likely cause of retirement will be running out of fuel. While subsequently walking around the two factories (which are run in the straightforward non-union manner one might expect, with an impressive investment in machine tooling) I reflected how little we could really understand about the work that -went into such -a DFV conversion job. For example the whole helm drive system for the neW oil pump, including pulleys. Then there are separately sealed cylinders, used instead of the normal INNhead sealing, and all the smaller bits of detail .engineering that we have not been able to explore in this brief account. That is why I referred to horsepower and torque at the end of our story: any fool can get a lot Of power from such an engine, The point about Cosworth is how they managed to .engineer the engine into a commercial proposition for Such a demanding market. If men like Duckworth and Chapman were a more prized breed in Britain our future prosperity would not be SO dependent on finite undersea Oil.—JeW,