The X files

Colin Chapman’s reputation as an innovative genius endures, but some of his cars performed less well than others. The imperfect 10 was one such, but modern thinking has eliminated some of its design flaws

While Jaguar monopolised the sports car racing headlines during the 1950s, as a consequence of its Le Mans 24 Hours successes, a youthful Colin Chapman was honing his engineering skills on the aerodynamic, full-bodied Lotus MkVIII and IX sports cars. These were the forerunners to many a successful Lotus – and also the less celebrated, often overlooked MkX.

Introduced in 1954, the Lotus MkVIII had been powered by an 85bhp MG engine while the following year’s MkIX was available with a range of options, though most were delivered with an 1100cc Coventry Climax. These were successful cars and achieved many class wins, but they were not contenders for major honours at international level.

Responding to client requests for more power, to compete in the popular class for cars up to 2.0 litres, Chapman built six 2.0 Bristol-powered cars. Assigned the MkX designation, these were based heavily on the MkVIII but had a large power bulge to clear the relatively tall Bristol engine. The first MkX, registration PCD13, was delivered to Mike Anthony and also broke fresh ground for the marque by featuring disc brakes.

Despite the greater power and light weight, the MkX did not set the world alight. Anthony scored three class wins and three second places in national events during 1955, but international results were disappointing. In the sports car race supporting the 1955 British GP at Aintree, Aston Martin DB3S models took the first three places and Cliff Davis finished 15th in a MkX. Davis also finished 12th in the Goodwood 9 hours and sixth at Crystal Palace. In the 2-litre class, Archie Scott Brown’s Lister Bristol regularly beat the MkX.

In period the best MkX result was Anthony’s third place at Goodwood in May 1955. After 35min he was 27sec behind Duncan Hamilton’s winning D-type – and the MkX always struggled to get on terms with the Jaguar.

Fast forward to May 2012 at Brands Hatch, however, and the one-hour Woodcote Trophy for pre-1956 sports cars. In their ex-Anthony MkX, Malcolm Paul and Rick Bourne held the lead for 11 laps and finished a strong second, splitting the highly developed D-types of Gary Pearson/Carlos Monteverdi and Nick Adams/Robin Ward. Paul and Bourne also took the second-fastest lap of the race, only 0.8sec shy of the winning Jaguar.

In the hands of Paul and Bourne, the MkX has become a regular class winner and is often on the overall podium in the Woodcote Trophy, particularly at the twistier circuits.

So why is a car that wasn’t particularly competitive in period – and certainly didn’t compete seriously with D-types – now able so to do? In order to understand this, we will take a detailed look at the MkX and D-type – and a brief glimpse at the Lister Bristol. We will also look at circuits then and now, to see whether they have a bearing on competitiveness.

THE LOTUS CHASSIS is a fully triangulated spaceframe with semi-stressed aluminium panels around the cockpit area. The chassis is very well designed and quite sophisticated for its time. GSD RaceDyn has run torsional rigidity tests on the chassis, which is surprisingly stiff at 1464Nm/degree (1079lb ft/degree). This is slightly stiffer than the 1962 Lotus 24 F1 spaceframe – but less than half the stiffness of the monocoque Lotus 25.

Rear suspension is by De Dion tube with vertical concentric coil spring/damper units and lateral location by Panhard rod, giving a roll centre height of 263mm – higher than ideal. Four radius rods provide longitudinal location. The De Dion tube keeps the wheels and tyres vertical when cornering and ensures that engine torque is reacted directly through the rigidly mounted final drive casing into the chassis. There is no rear anti-roll bar.

Front suspension is a fairly crude low-pivot swing-axle arrangement, with angled concentric spring/damper units. There is no front anti-roll bar. This swing-axle is the car’s main weakness, formed by cutting a Ford 10 or rigid front axle in half and welding on ‘ears’ to form lowered pivots. Angled compression struts react to braking loads. This gives a very high (196mm) front roll centre, massive camber change with bump (1.04deg negative per 10mm bump) and excessive track change/contact patch deflection at 3.04mm per 10mm bump. The high front roll centre causes a jacking force of more than 70kg when cornering at 0.95g. This causes the chassis to lift by 22mm on the soft springs used in period – and therefore applies more than 2.2deg of positive camber at the front, causing a reduction in grip of more than 23 per cent and consequent understeer.

Steering is by means of a worm-and-peg box operating long track rods via a steering arm, drag link and central idler. Not an ideal arrangement for a racing car... Dunlop hydraulic disc brakes are fitted, outboard at the front, inboard at the rear.

The Bristol straight six is a 1936 BMW design manufactured under licence. Combustion chambers are hemispherical, with angled exhaust and inlet valves opened by rockers operated by pushrods. In 1955, the engine developed 135-140bhp at 5750rpm.

The engine drives through a Bristol four-speed gearbox via a short propshaft to a Salisbury 3HA final drive and differential unit in a lightweight Lotus aluminium casing, bolted rigidly to the chassis. The differential is free – there is no limited slip.

Designed by aerodynamicist Frank Costin, the aluminium bodywork is quite efficient. Drag is approximately 116kg at 120mph, giving a theoretical maximum speed of 136mph. The car weighs circa 740kg with driver and distribution is surprisingly good at 47.8 per cent front, 52.2 rear.

GSD’s vehicle dynamics analysis shows that, as configured in period, the car suffered from excessive pitch under braking, relatively poor transient response at turn-in, then strong but inconsistent understeer at the apex, caused by the jacking and camber change previously described. Under power in the corner exit phase, the lack of front roll stiffness caused oversteer, partially countered by understeer due to nose lift under acceleration.

MOVING ON to the D-type, this car was designed specifically to win Le Mans, which meant that straight-line speed on the pre-chicane Mulsanne was an absolute priority. Aeronautical engineering practice was used in both body and chassis design, led by Malcolm Sayer. His low-drag body had a very low frontal area, but later tests show that it was not quite as efficient as it appeared, yielding a drag coefficient of 0.45-0.49. He was successful in minimising frontal area, though, achieving very low drag levels and high top-end speeds. Analysis suggests that drag at 120mph is 121kg and the car’s theoretical maximum speed with 250bhp is 169mph.

The chassis used an advanced aluminium monocoque cockpit section, and a square-tube spaceframe to carry the engine and front suspension. GSD has not measured torsional stiffness, but assesses it to be significantly stiffer than the Lotus.

The car weighs approximately 1071kg with driver and weight distribution is poor at 53.2 per cent front, 46.8 rear.

Front suspension is by dual unequal length wishbones with interposed telescopic dampers. The lower wishbones operate torsion bars, which are the sole springing medium. A front anti-roll bar is fitted. Suspension geometry and kinematics are generally good, giving minimal track change/deflection, sensibly small camber change with bump and an appropriately low roll centre. Roll camber is less than ideal at 26 per cent, but not really a problem. Rack-and-pinion steering is fitted.

Rear suspension is by live axle with dual trailing links/radius arms per side, in the form of steel blades. Lateral location is by an unusual double A-bracket arrangement. This is quite good, giving a 184mm roll centre height – higher than ideal, but much lower (better) than Watts Linkage or sliding-peg systems. This also avoids the large differential case loadings imposed by single A-bracket systems. Geometry and kinematics are generally good, except for the high roll centre.

The lower trailing blades operate short transverse torsion bars (the sole springing medium), concentric with the trailing arm pivots. Two different rear anti-roll bar designs are used, ‘customer’ and ‘factory’ versions.

While the rear suspension looks reasonably good, it is actually, together with the forward weight distribution, the limiting factor in the D-type’s cornering performance, particularly under power in the exit phase. The live axle brings two key problems – large unsprung mass and engine torque reaction through the rear tyres. The latter is much more significant. Whereas a De Dion system transfers engine torque loadings directly into the chassis via the rigidly mounted final drive casing, the live axle relies on tyre loadings to react to engine torque loads. This is true of all live-axle cars, but particularly problematic in the case of the D-type because of its very narrow track and the torquey XK engine: when full throttle is applied in second gear, vertical load on the left rear wheel increases by 32kg while that on the right rear wheel reduces by 32kg – quite enough seriously to affect handling balance and traction.

The second issue relates to the four trailing arm blades. When the car rolls, it rotates around the dual A-bracket pivot, which is 184mm above ground. The blades are significantly higher than this and are mounted in rigid bushings, not spherical bearings, so as the car rolls, the blades bend into an S shape and twist, causing a non-linear increase in roll stiffness. This, again, has a significant effect on handling balance and traction.

The D-type has outboard Dunlop disc brakes all round. Aircraft practice was again followed in the use of Marston ‘bag’ fuel tank bladders.

No introduction is required for the legendary 3.4-litre twin overhead camshaft six-cylinder Jaguar XK engine. In 1955, this delivered 250bhp at 6000rpm. Drive was transmitted through a four-speed Moss gearbox via a short propshaft to the rear axle. A ramp-and-plate limited slip differential was also used.

SUMMARISING, the D-type was immensely quick in a straight line, but suffered when the road turned due to the forward weight distribution, live rear axle and blade trailing arms. These caused understeer at turn-in and apex, followed by unpredictable power oversteer in the corner exit phase. To compound the problem, torque reaction meant power oversteer was much worse in right-hand corners. In his book Jim Clark at the Wheel, the Scot described the D-type as a handful – a very difficult car to drive.

Making direct comparisons, the Jaguar’s power-to-weight ratio is 25 per cent better than that of the MkX and its power-to-aerodynamic drag ratio a staggering 74 per cent better. So the Jaguar is much quicker in a straight line – and the faster the circuit, the bigger its advantage.

The Lotus would have stood no chance whatsoever at Le Mans.

Given the Jaguar’s forward weight distribution and live rear axle – and the Lotus’s 14 per cent advantage in load per unit tyre area (see table), the Lotus should have been significantly faster in the corners, but those advantages were lost by its swing-axle front suspension.

GSD ran simulations at Goodwood and the current Le Mans to compare the Jaguar and Lotus in their 1955 specifications. These showed the Lotus to be 1.76sec per lap (2.1 per cent) slower than the Jaguar at Goodwood, but more than 16sec per lap (5.18 per cent) slower at Le Mans. The Jaguar pulls 156mph before braking for the first Mulsanne chicane, the Lotus 133mph. With its front suspension fixed in the simulation, the Lotus’s deficit is reduced to 9.37sec at Le Mans… and 0.2sec at Goodwood.

Now let’s take a brief look at the Lister Bristol. It is slightly heavier than the MkX, its twin-tube chassis is less sophisticated than the Lotus’s spaceframe and its weight distribution is slightly worse at 50/50. It has proper double wishbone front suspension, however – and while not perfect (it lacks roll camber correction) it is vastly better than the Lotus’s swing axles, giving at least 12 per cent more front grip. The Lister has a De Dion rear, quite similar to that of the Lotus. The Lister’s superiority over the Lotus in 1955 has sometimes been attributed to the undoubted excellence of Archie Scott Brown’s driving, but that is only part of the story. The Lister had better grip, balance and traction than the Lotus due to its wishbone front suspension and front anti-roll bar. With equal drivers, the Lister would still have beaten the Lotus.

Returning to modern times, the MkX now generally beats the Lister Bristol and gives the D-types a very good run for their money. Let’s look at the reasons.

LONG-STANDING FRIEND and client Malcolm Paul phoned GSD in 2012 to report that he had bought PCD 13 and would like the handling to be sorted out. Given what it already knew about the car, GSD’s enthusiasm was firmly under control…

Upon analysis, however, the car was found to be generally very good. The only major problem was the swing-axle front suspension. Steering system play and structural issues with the De Dion tube were resolved through a great deal of hard work and sound engineering from Ian Grimes and, early on, Rob Wells.

Solving the front suspension problem while staying within the regulations and maintaining the car’s originality was a challenge.

GSD could not reduce either the large jacking forces, or the large camber change inherent in the geometry. The solution was to set the static ride height to give the desired static camber (the car has no camber adjustment, short of heating and bending the axles) and using quite stiff front springs and large damping coefficients to contain front suspension movement within reasonable limits. Jacking was reduced from more than 22mm to less than 10mm.

Once the problems at the front had been resolved, rear spring rates and damping force curves were specified to give a reasonable compromise between transient response at turn-in, mid-corner understeer and appropriate balance and traction in the corner exit phase.

Engine development, better materials and fuel and lubricant improvements mean that the Bristol engine now develops 160bhp, the Jaguar 300bhp. Though basically similar to those used in 1955, the Dunlop tyres are also better.

GSD has run simulations incorporating the improvements made to the Lotus’s front suspension and the engine and tyre improvements already mentioned. These suggest that the Jaguar ought to be 3-4sec per lap faster than the Lotus at Spa and 2-3sec per lap faster on the Silverstone Historic GP circuit. At Brands Hatch GP the Jaguar and Lotus should deliver virtually identical times.

During 2017, actual performance has mirrored the simulation – except at Brands Hatch, where the D-type of Gary Pearson, a very accomplished driver, is still about 0.8sec per lap faster than the Lotus.

In conclusion, the Lotus MkX was potentially a good car, marred by very poor front suspension design. With its much better power to weight and power to drag ratios, the Jaguar was – and is – much quicker than the Lotus on fast circuits with long straights. With the front suspension problems under control, the Lotus is now faster than the Jaguar wherever the road turns. Modern circuits are smoother than those of 1955, with far more slow corners and chicanes, placing a premium on handling and traction and allowing the Lotus to close the gap dramatically. The Jaguar’s live rear axle impairs cornering and traction, a disadvantage on modern circuits but no problem at all in 1955, because the car’s raison d’être was Le Mans, with its long straights and (then) very few corners.

Finally, Colin Chapman must have known that the swing-axle front suspension was a major handicap compared with its rivals double-wishbone systems.

The ubiquitous Alford and Alder upright was available, so it should have been cheap and easy to produce double-wishbone front suspension. It is difficult to understand why he didn’t plump for that option, but then nothing is much cheaper than a Ford 10 axle chopped in half.

Nigel Rees is founder of GSD RaceDyn – a vehicle dynamics analysis firm. GSD has analysed more than 560 cars, many notching up victories and championship titles in historic racing.