Gripping yarns

Towards the end of 1963, the FIA announced that naturally aspirated Formula 1 engines would increase from 1.5 to 3.0 litres in 1966, though supercharged or turbocharged 1.5s would also be eligible. Widely regarded as the sport’s leading design authority, Colin Chapman spoke to Road & Track magazine soon afterwards and made two bold predictions to interviewer Graham Gauld. He felt a forced-induction 1.5-litre engine would be better than the atmospheric alternative – and that four-wheel drive would be essential from 1966 because of the massive increase in engine power. The former was ultimately correct, though it took until 1979 for Renault to prove it. The second prediction was spectacularly disproved, despite the best efforts of the four manufacturers – including Lotus – who built 4WD F1 cars towards the decade’s end.

Defying Chapman’s predictions, neither 4WD nor forced induction appeared in 1966. Jack Brabham won the 1966 world title in his own BT20, using a simple Repco V8 giving little more than 320bhp. The situation changed in 1967, with the introduction of the Cosworth-DFV in Chapman’s Lotus 49. The DFV initially delivered 409bhp – but it had a rather lumpy torque curve, power coming in with a bang at 6500rpm, causing traction and handling difficulties. In Motor Sport, Denis Jenkinson noted, “Grand Prix cars have become very big and fierce.”

During 1968, encouraged by the performance of the 4WD Lotus 56 Indycar, Chapman applied the principle to F1 with the Lotus 63. Matra swiftly followed suit with the MS84 and McLaren joined the party with the Jo Marquart-designed M9A. All three cars used the Ferguson 4WD transmission system. Cosworth hired McLaren chassis designer Robin Herd and pitched in with its own 4WD project, using an in-house transmission design, but the car never raced. Four-wheel-drive Grand Prix cars were not new, but the huge push in 1969 was unprecedented.

In order to understand the decision to pursue these projects, we will look at the physics using only the tools that designers had at their disposal in 1968 – slide rule, log tables and limited access to unwieldy mainframe computers. We will look at 2WD versus 4WD for straight-line acceleration and cornering. Only when we have done this will we use simulation software to evaluate performance around a lap. Finally, we will look at some 4WD design features that affected performance and driveability, particularly front to rear torque split, differential characteristics and steering scrub radius.

Most of us are familiar with the benefits of 4WD on low-grip surfaces, but for a single-seater on asphalt the advantages are not so clear-cut.

To understand what’s happening in straight-line acceleration, we need at least a rudimentary understanding of the laws of friction as they apply to tyres – and the basic physics of rearward load/weight transfer under acceleration. Your school physics teacher probably told you that friction is independent of area. If that is true – and it pretty much is – then why use wide tyres in motor racing? The answer is in heat generation and dissipation. If we use a small tyre tread area, we are forced to use a hard (low-grip) compound to resist heat and wear. If we increase the contact patch, heat dissipation is dramatically improved and we can use a softer compound with a higher coefficient of friction (Mu), generating extra grip.

F1 tyres in the late 1950s achieved a coefficient of friction between 0.9 and 1, allowing cars to corner at almost 1g lateral acceleration. Between 1963 and 1973, driven by a ‘tyre war’ between Dunlop, Goodyear and Firestone, tread widths widened dramatically, correspondingly softer tread compounds were introduced and coefficients of friction crept upward, perhaps reaching 1.5 by 1973. In early 1964, when Chapman made his 4WD prediction, coefficients of friction were probably just above 1, rising to 1.15 by 1968 (when the 4WD cars were in the design stage) and nearer 1.2 by 1969.

Let’s look at rear-wheel drive and 4WD F1 cars accelerating on a straight. The analysis involves calculating the tractive force at the wheels due to the engine, which reduces as the driver goes up through the gears and speed increases, then calculating the tractive force available from the grip of the tyres. While the available tractive force at the wheels due to the engine is greater than the tractive force due to tyre grip, the car is traction limited and the driver cannot use full throttle.

The analysis was done for a 2WD car (590kg, 62 per cent rear) and a 4WD car (624kg, 55 per cent rear) assuming 430bhp at 9500rpm, a 1.1 coefficient of friction for the tyres and zero aerodynamic downforce. Weights included an 80kg driver. The analysis shows that the 2WD car is traction-limited up to about 145mph, but the 4WD car is only traction-limited up to 100mph.

This is probably the level of analysis that constructors used in late 1967/early 1968 – and it looks promising enough for 4WD projects to be given the green light. However, it is worth looking a little deeper.

If we look at the tractive force required from the front wheels, the optimum torque split is only 24 per cent front at 100mph, falling to zero at 140mph.

Moving from straight-line acceleration to cornering, we need to look at the tyre’s traction circle. Very simply, this means that if we use grip for acceleration, grip available for cornering is reduced. With rear-wheel drive, the throttle can be used to adjust the car’s handling balance in the corner exit phase – known as ‘steering the car on the throttle’. However, with a 4WD car, the application of power to the front wheels reduces front grip available for cornering – causing understeer. This lessens the driver’s ability to steer with the throttle. A relatively simple calculation shows that the 4WD car, cornering at 100mph with 24 per cent front torque split, loses 22 per cent of front cornering force when 50 per cent power is applied in the corner exit phase – but it only loses 8 per cent of rear cornering force. Therefore the car understeers and its exit speed is restricted by available front grip. This is a consequence of the traction circle and rearward weight transfer under power.

So the 4WD car performs better than the 2WD under straight-line acceleration, but worse when cornering. Simple hand calculations cannot give us the combined effect on lap time.

WE MUST NOW return to the 21st century and use lap simulation software to determine whether 4WD gives a significant advantage over a full lap. GSD RaceDyn ran simulations at Spa, Brands Hatch GP and Monaco for 4WD and 2WD cars, again using a tyre coefficient of friction of 1.1 and zero aerodynamic downforce.

The 2WD car was faster at all three circuits, by 1.768sec per lap at Spa, 1.02sec at Brands and 0.746sec at Monaco.

This leads to some interesting conclusions. Lap simulation shows that the 1969 4WD cars would be slower over a lap than comparable 2WD cars, even with relatively low-grip tyres and zero aerodynamic downforce. With the calculation tools available in period, the constructors could quantify the straight-line acceleration benefit. They probably knew that 4WD would cause understeer, but they could not quantify the effect on lap time because the computing power and simulation software was simply not available. The constructors’ decision to pursue 4WD was justifiable, but they certainly wouldn’t have done it if they had been able to run a full simulation. Conventional wisdom suggests that the 4WD cars were rendered unnecessary by tyre development and aerodynamics. That is not strictly true. Analysis has shown that, even with relatively low grip and without aerodynamic downforce, the 4WD cars were fundamentally slower than 2WD.

Looking at actual performance, Jochen Rindt took pole position for the 1969 British GP in the 2WD Lotus 49. John Miles’s 4WD Lotus 63 was more than 4sec slower. Jackie Stewart tried the 4WD Matra MS84 in practice, but was 3.3sec off Rindt’s pole time. Graham Hill and Jochen Rindt refused to drive the Lotus 63 and severe understeer plagued all the 4WD cars. Reducing the torque split to the front wheels from 40 per cent to 20 per cent was found to be an improvement, so Matra disconnected the drive to the MS84’s front wheels. The drivers preferred it that way, but now it was just an overweight rear-wheel-drive car with the weight distribution too far forward.

The performance of the 4WD cars relative to 2WD was a little worse than analysis suggested. There are three reasons for this. Firstly, tyre development increased coefficients of friction to more than 1.2 by 1969, reducing the straight-line acceleration advantage of 4WD. Secondly, the advent of wings and aerodynamic downforce in mid-1968 further diminished the advantage of 4WD in straight-line acceleration. Third, the detail design of the 4WD cars, particularly the transmission, rendered them difficult to drive.

Let’s take a look at the designs, starting with the Lotus 63. Penned by Maurice Philippe, the chassis was an aluminium monocoque with tubular sub-frames carrying front and rear suspension. The DFV was fitted ‘back to front’, so that the gearbox, Ferguson 4WD transmission system and centre differential could be mounted amidships. This pushed the driver forward so that his feet were ahead of the front axle centre line and his lower legs slotted beneath the front drive assembly. The Ferguson transmission used sprag clutches and epicyclic gears to allow differences in front and rear axle speed and adjustment of front to rear torque split.

Front torque could be reduced to as little as 18 per cent. Short propeller shafts ran past the driver’s left side to the front and rear final drives, which used cam and pawl limited-slip differentials. Brakes were inboard front and rear. Suspension was by lower wishbones and upper rockers front and rear, the robust, beautifully fabricated rockers operating inboard spring/damper units. Great care was taken to reduce the steering scrub radius, minimising torque steer. Steering was by angled rack and pinion, operating an unusual arrangement of bell cranks and steeply angled track rods, probably needed to clear the front-drive arrangement. Despite the front-mounted radiator, it had wedge bodywork with single-plane front wings and, initially, a flat rear deck. A rear wing was an early addition.

The McLaren M9A also used the Ferguson 4WD transmission and was similar in general layout to the Lotus 63, though rather more bulbous in appearance. Instead of a rear wing, the car used two large tray type spoilers and suffered from torque steer and steering kickback, which probably accounts for Bruce McLaren’s comment that “Placing the car was like writing your signature with someone jogging your elbow.”

The Matra MS84 again used the Ferguson transmission and the same basic layout, but the chassis was a tubular spaceframe. Unlike the Lotus and McLaren, the Matra used conventional outboard spring/damper units.

The drivers all reported that the 4WD cars understeered, could not be guided on the throttle and were very difficult to place precisely. The understeer was largely caused by the traction circle effect, reducing front grip when power is applied. The Ferguson 4WD system allowed for adjustment of front to rear torque split, but only in the garage. Once set, the torque split was constant. What is really required is active control of the centre differential – which governs torque split – to give about 24 per cent front up to 100mph in a straight line, then falling to zero at 140mph. When cornering, front split would be close to zero. This would give all the 4WD advantages when accelerating and the ability to balance the car properly when cornering.

The Lotus, Matra and McLaren all used cam and pawl limited-slip differentials at front and rear. These are rarely used today. They are ‘free’ when no power is applied, but as soon as power is applied, the differential locks solid. In conventional rear-wheel-drive cars, this causes strong power understeer. To overcome this, cars were often set up with a very stiff rear anti-roll bar and a lot of castor. With a cam and pawl LSD at both front and rear, and a fairly coarse torque split arrangement, it would be difficult to overcome power understeer in the 1969 4WD cars. Ramp and plate-type (Salisbury) differentials would have been a significant improvement, but the real solution is electronic control of front, rear and centre differentials.

THIS LEADS US TO successful applications of 4WD. Rally cars running on asphalt or low-grip surfaces such as gravel, ice and snow benefit hugely from 4WD. Most current rally cars are based on front-wheel-drive production cars with drive added to the rear. Weight distribution is generally more than 60 per cent front, but under acceleration, weight/load is transferred off the front wheels to the rear, so 4WD significantly improves traction and acceleration. Differential torque transfer maps can be used to avoid power understeer or provoke power oversteer.

Of course, 4WD has been banned in F1 1 since 1982. If that was lifted, could 4WD be used advantageously? It would depend on the regulations, but I think that in a hybrid world the answer is probably yes. Imagine using small electric motor/generators mounted in the front wheels. These could be used to harvest braking energy at the front wheels (it is currently done only at the rear). Battery power could be applied to the front wheels during acceleration from low speeds, when the car is traction limited. When cornering, power could be applied at different levels to the two front wheels to help control handling balance. The downside would be the weight of the electric motors. Something similar has been tried in LMP1, but it seems unlikely that F1 regulations will ever permit it.

FOUR WHEELS GOOD…

A brief history of 4WD racing cars

The first 4WD racing car was the Spyker. Built in Holland in 1903, it was not particularly successful. In 1932, Bugatti built the 4WD Type 53, which won two significant hillclimbs but was not successful in racing – and had a reputation for poor handling. Piero Dusio commissioned Cisitalia to build a rear-engine 4WD 1500cc supercharged Grand Prix car in 1946. The design was very advanced, with strong Auto Union roots, using Ferdinand Porsche’s flat-12 engine design and von Eberhorst’s influence on the chassis. On paper, the car should have been highly competitive against the dominant Alfa Romeo 158s, but unfortunately Dusio ran short of money and the project was abandoned.

The 1961 Ferguson P99 was a front-engined 1.5-litre 4WD F1 car built by Ferguson Research for Rob Walker. Using the advanced Ferguson Formula 4WD torque-splitting transmission system, the car achieved fame as the last front-engined car to score an F1 victory – and the only 4WD car ever so to do – when Stirling Moss won the 1961 Oulton Park Gold Cup in the damp (above). The P99 did not appear particularly competitive in the dry.

BRM also built a 4WD F1 car in 1964, using the Ferguson system. The P67 qualified last for the 1964 British GP and the project was mothballed.

Both the Ferguson P99 and BRM P67 had successful second careers in hillclimbing, where the 4WD gave a significant advantage off the start and out of slow hairpins. Peter Westbury used the Ferguson P99 to win the British Hillclimb Championship in 1964 and Peter Lawson did so with the BRM P67 in 1968.