It is well documented that finishing the working surfaces of gears and their root fillet regions to a very low roughness can result in a considerable increase in surface durability, as well as reducing friction and operating temperatures. Achieving such finishes using traditional methods such as surface grinding and honing is very time-consuming though, and carries considerable risk that the profile of the gear teeth can be irreparably changed. The use of chemically accelerated vibratory finishing methods, generically referred to as superfinishing, can avoid such problems while also resulting in a much smoother surface finish. The average roughness achievable with grinding techniques is about 6.0-12 µin, whereas superfinishing can produce a roughness in the region of 1-3 µin.

Superfinishing is undertaken in vibratory finishing tubs, of the same type that have been used for many years in other abrasive media finishing processes. The media used in these tubs is a high density, non-abrasive ceramic material, with the shape and size of the media selected to match the geometry of the gears being finished. The media does not itself remove any material from the gears, it is only once a reactive agent is introduced into the finishing tub that changes in the surface finish of the gears occur.

The reactive agent produces a stable, soft conversion coating across the asperities of the gear surfaces. As the media in the finishing tub rubs across this coating, the peaks and valleys of the material are gradually smoothed out, until the surface is, to all intents and purposes, free of asperities. The reactive agent is mildly acidic and, depending on the concentration used, stock removal occurs at 0.00005-00040 in/h. This removal is beneficial, allowing gears with an initial surface roughness of around 60 µin to be finished to a final roughness of 3 µin.

Interestingly, a consequence of using hard ceramic media in the process is improved wear resistance in use. Although the material is non-abrasive, it still leaves a micro-textured surface on the material being finished. Testing undertaken by the University of Cardiff, Wales, to look at scuffing performance of test discs treated with ceramic and plastic media respectively showed that the ceramic media-treated parts had a much higher resistance to scuffing than those treated in plastic media. This was surprising, given that the plastic media leaves a smoother surface. It therefore follows that there is an optimum surface roughness that needs to be achieved in order to reap the greatest benefits from the process.

Overall then, superfinishing can improve durability and performance in transmissions compared with traditional ground finishes. As a result, many high-performance transmission manufacturers now use the process to ensure their gears are better able to survive the rigours of competition use.

Written by Lawrence Butcher

Materials for gears in general can be divided in two categories: metallic and non-metallic. Unsurprisingly, gears in racing transmissions fall into the first group. Although there is ongoing research into the potential uses of composite materials for gear construction, their application is as yet not truly feasible. The requirements for transmission gear materials are high surface hardness, good core toughness, fatigue strength for tooth bending, rolling contact fatigue resistance, and high density to resist pitting and sub-surface spalling during heavy use. The most common metals used here are wrought surface-hardening and through-hardening carbon steels.

The needs of motorsport and mass-production transmission manufacturers are broadly similar, but the very high loadings and quest for the smallest, lightest gears possible (not to mention greater freedom from cost and complexity constraints) means that the materials used for motorsport gear construction are subtly different from those found in the average high-performance roadcar.  

Surface-hardening steels are hardened to a relatively thin case depth and the various types include carburising, nitriding, and carbonitriding steels. Surface-hardening steels include plain carbon and alloy steels with a carbon content generally not exceeding 0.25% C.

Through-hardening steels can be comparatively shallow hardening or deep hardening, depending on their chemical composition and method of hardening. They include plain carbon and alloy steels with a carbon content ranging from 0.30 to about 0.55%. The mechanical properties can be tailored by varying the quantities of the allying elements and the heat treatment process: for example, the bending and surface hardness can be improved. Another example of such tailoring would be the addition of molybdenum to help reduce rolling contact friction. 

The highest grade of carburising steels tend to be found in motorsport gear applications, although such materials were often initially developed for other industrial sectors such as aerospace. Taking the gears used in a helicopter’s transmission as an example, these need to be exceptionally strong and resistant to fatigue, with the consequences of failure being far more serious than in a racecar. Also, the budgets available for aerospace development dwarf even the biggest racing operations and thus industries such as defence, aerospace and even OEM vehicle manufacturers have the resources to develop new high-performance materials that race teams do not. There are of course exceptions, and several manufacturers of racing gearboxes have commissioned bespoke steel blends to suit their specific applications.

One area in particular where steels used for motorsport and other high-performance applications differ from more run-of-the-mill materials is their ‘cleanliness’. The number and type of inclusions in steel billets destined for use as gears can have a major impact on ultimate fatigue resistance and strength, so the very best steels are smelted in a vacuum, where there is minimal opportunity for contamination of the metal. It is not unusual for very ‘clean’ steels to be vacuum melted up to three times, giving them a very consistent molecular structure but making them anything but cheap!

This is only a brief overview of the materials and requirements that need to be accounted for when specifying gears for use in racing transmissions, but in future articles we will revisit the subject to cover specific areas in greater detail.

Written by Lawrence Butcher

The impact of such coatings on transmission efficiency and durability can be considerable, netting useful gains in both areas. For the purposes of this article, we will look at the Honda team’s transmission development during its previous foray into Formula One, which ended in 2008.

The main losses found in a transmission relate to the sliding motion of the gear teeth as they rotate. It is well known that DLC coatings have good low-friction characteristics, so coating the gears should help reduce overall losses. When Honda first looked at doing this it realised that existing coatings, used for components such as valvetrain parts, would not have sufficient durability for a transmission because of the high surface pressure loads experienced between gears. To counter this, it therefore set about developing a new coating formulation that could better handle these loads.

There were two key requirements for the coating:

• It must be able to withstand surface pressures of 2.2 GPa for a duration of four races (the then current regulation life of a transmission)
• The application method must allow for a uniform coating over the complex gear tooth shape.

There are some clear differences in requirements between gear and valvetrain coatings. As already mentioned, the surface pressures are higher, but also the lubrication regime differs considerably. In valvetrain applications, the oil film breaks down as valve motion is reversed, which can lead to scuffing. To this end, the coating Honda developed for its valvetrain featured a hard bonding layer under the DLC top coat to prevent scuffing. In a gear train though the oil film does not break down, so Honda decided that the hard coating could be removed, with the development focus being on a top coat composition that had improved wear resistance.

The higher surface pressure led Honda to revise the surface finish on the gears to aid adhesion of the coating film, settling on a roughness of 0.1 Ra combined with enhanced cleaning of the parts to remove contaminants that could potentially compromise the coating. Honda then developed a coating composition to match the requirements of the gears. This consisted of a chromium adhesion layer, on top of which was a layer of metal-impregnated carbon coating, the combined thickness of these layers being 0.6 µm. This was followed by a DLC top coat 1 µm thick, with all the layers being applied using a sputtering process.

The results obtained after coating were impressive. The coating itself, when compared to Honda’s standard DLC as applied to valvetrain parts, showed an improvement in seizing pressure of 40%, as well as a considerable reduction in frictional losses. With the coating applied to all ratios in the box as well as the final drive and bevel gears, overall frictional losses were reduced by 3.3 kW – a very useful gain.

This work shows just how effective the use of coatings can be in increasing transmission efficiency. Better still, with commercial suppliers of such coatings now offering products at ever-decreasing prices, it has even become feasible for teams and constructors outside of the rarefied atmosphere of Formula One to take advantage of such benefits.

Written by Lawrence Butcher

For example, one electric drag car I have studied features a pair of astoundingly powerful electric motors, pushing out in the region of 700 bhp, which would have more than enough power to propel it down the strip at great speed. However, it also has a planetary geared transmission, which helps make the most efficient use of the car’s motors.

As the owner and builder of the car explains, “The power band of our motors gives us peak horsepower between 2400 and 3400 rpm, so it is a narrow band. What we are able to do with the use of a transmission is go through that peak four times in a run. Our goal, after studying the data we gained from our first car, was to keep the car in that power band for as much of the run as possible.”

The result is blistering acceleration all the way down the track. By using an under-driven ratio for first gear, the car achieves very low 60 ft times, leveraging the instantaneous torque delivery of the motors to the maximum, while the additional three ratios help it reach a speed of more than 150 mph by the end of a run.

Another interesting EV transmission design draws its technology directly from that used in the current generation of MotoGP ‘seamless-shift’ transmissions. It eliminates the dog rings found in a regular sequential transmission, and instead uses ‘bullet rings’ to engage the gears. A gear is selected when one ring is moved until its bullets hook onto drive teeth on the side of that gear. A second bullet ring moves in the same direction, with its bullets filling the gap between the teeth, eliminating any slack between the gears.

Eliminating this slack is what creates seamless upshifts and downshifts. By integrating the dampers inside the gear hubs, the manufacturer claims that the need for a clutch is eliminated – and this, combined with a seamless change of ratio, means the motor’s torque delivery can remain constant.

On a simpler level (although in terms of overall system complexity, far higher), hybrid vehicles that use electric drive motors on one or both axles can also benefit from a transmission. Taking the latest generation of hypercar road-going hybrids and both Porsche’s and Audi’s LMP cars as an example, the use of clutch systems and reduction gears helps increase motor efficiency. For instance, Porsche’s 918 roadcar uses a system of clutches on the front drive motors to disengage them above 146 mph, with the clutches being electronically controlled to provide the function of an active differential.

Meanwhile, other cars use reduction boxes to maximise the already high torque output of their electric motors in order to assist acceleration. This is a logical approach when thinking of an optimised racecar package. The boost provided by an electric motor should be far more beneficial to lap times if it can increase corner exit speed, rather than being used at high speed where the extra power has far less impact.

The age of electric and hybrid racers is still young, but rest assured that the rise of electric motors does not mean the end of multi-speed transmissions.

Written by Lawrence Butcher

The usual method of measuring such losses is by using a dedicated transmission dynamometer. In general, these use a drive motor and an absorption device such as a water brake. The motor drives the transmission through the input shaft, and the absorption device is connected to the output. Torque sensors on both the input and output shafts allow the losses through the transmission to be deduced. Much as with an engine dyno, transmissions that have an external oil lubrication system with pumps and coolers can usually be accommodated and, with the correct instrumentation, lubricant temperatures and flow can be monitored. 

One interesting area relating to transmission testing is measuring the operation of torque converters, such as those used in drag cars in classes such as Pro Mod. In such applications, the characteristics of the converter – for example its torque multiplication effect and stall speed – have a significant effect on the way a car launches. To clarify, the stall speed is the rpm at which a torque converter’s impeller (known as the stator) has to spin for it to overcome a given amount of load and begin moving its turbine element (the component that transfers drive to the gearbox).

This is a vital consideration when specifying a converter for racing use, as the stall speed is the rpm at which the converter needs to be spinning to create enough fluid force to overcome vehicle inertia at wide-open throttle. The torque multiplication effect occurs when the converter is in ’stall mode’ and during initial vehicle acceleration. As the vehicle accelerates, the torque multiplication decreases until it reaches a ratio of 1:1 with the crankshaft torque. The design of the stator will influence the torque multiplication characteristics of a converter, with a typical figure being in the region of 2.5:1.

Therefore, being able to measure such parameters accurately can be very beneficial in informing other areas of car set-up. While some aspects of a converter’s operation can be analysed on a regular transmission dynamometer, for more detailed investigation using a dedicated machine is required. With such a device, input and output torque of a converter can be measured, allowing for the torque multiplication effect throughout the rpm range to be assed, while other important factors such as fluid flow rate and temperature can also be monitored. A key aspect of a torque converter dynamometer is a motor powerful enough to induce stall in converters designed for very high power applications; while electrical or hydraulic power is often used, the author has seen a converter test bed powered by a Big Block Chevy motor!

Such dynamometers are expensive pieces of equipment, so are not found in the average transmission shop. However, some competition torque converter manufacturers do have such capabilities in-house. As a result, they can accurately assess the performance of each converter produced, meaning they can provide customers with much useful data on a particular unit. In turn, this data can help end-users get more consistent car set-ups, something that is of vital importance in drag racing.

Written by Lawrence Butcher

In recent years, our understanding of the way composites behave under load has improved immeasurably, and the tools needed to predict this behaviour accurately are becoming more widely available. But the cost of materials and the manufacturing processes required for their construction – notably the use of pre-preg fabrics and the need for an autoclave – remains very high. However, with the increasing use of hybrid powertrains, which currently have a tendency to increase overall vehicle weight, it is likely that composite transmissions will start to appear in more and more racing machines as engineers strive to keep weight under control. That could see the adoption of alternative manufacturing processes that allow for more cost-effective production of such parts.

One such process, resin transfer moulding (RTM), is seeing growing use in other areas of industry, notably aviation. RTM uses a two-part mould, with dry carbon fibre weaves laid in the mould halves. The mould is closed, and epoxy resin is injected into it under pressure, while a vacuum is applied at strategically placed vents, ensuring the resin is distributed throughout the fibres. Depending on the precise manufacturing requirements, the part can then be cured at elevated temperatures. The fabrics used are laid up and orientated in a similar way to pre-preg construction methods, with different orientations and tows used to tailor the structural properties. For example, unidirectional tapes may be added in areas that need extra reinforcement.

The advantage of RTM is that parts can be produced with a very high degree of repeatability, and the exact level of resin distribution in the material is closely controlled. Also, the cost of the basic fabrics is much lower than with pre-preg, and it is not subject to the same storage and shelf-life issues (pre-preg needs to be kept refrigerated to prevent the resin curing).  

While no-one (to the best of my knowledge) has yet produced an automotive transmission casing using this process, there have been a number of applications developed for rotary winged aircraft transmissions, which need to withstand similarly extreme loadings as a racing gearbox.

As is often the case, aerospace leads the way in the development of such techniques, the focus in that industry being the same as racing in that parts must be as light as possible; however, the production requirements differ owing to the fact that they also need to be produced in volume. There is also the matter of funding for research into making techniques such as RTM viable. While teams in Formula One undoubtedly have large budgets, they pale into insignificance when compared to those of companies such as Boeing or Bell Aerospace (who have produced resin transfer moulded transmission casings).

Composite racing transmission casings are currently produced in only very small numbers, so traditional hand lay-up and autoclaving are still the most viable manufacturing method. The cost and resources needed to produce moulds for RTM are considerable, and any design changes would require the production of a new mould. However, it would not be inconceivable that RTM could be a viable approach for producing casings cost-effectively for a spec series, where a lot of units are needed and the design does not evolve over a season. As the technology continues to mature it will be interesting to see if any manufacturers adopt this approach.

Written by Lawrence Butcher

Carbon-carbon clutches have been the norm in Formula One since the mid-1980s, providing high levels of  bite and heat resistance coupled with low inertia. Before regulations dictated a minimum crankshaft height, there was a drive by clutch manufacturers towards ever-smaller diameter clutch discs. The smaller the clutch, the lower the engine could be placed, and thus centre of gravity would also be lowered.

These clutches, some as little as 87 mm in diameter, still contained the same amount of material as a larger clutch, to absorb the thermal load on launch, but putting them in a smaller package reduced inertia. These days, most Formula One clutches are about 97 mm in diameter and feature four carbon friction discs, housed in a titanium basket. The assembly must be robust enough to survive the forces exerted by a 750 bhp engine while also remaining dimensionally stable when rotating at 18,000 rpm.

Clutch release is controlled by an hydraulic release bearing, the movement of which is actuated by a high-pressure electro-hydraulic servo controlled by signals from the ECU, which in turn are interpreted from driver inputs to a steering wheel paddle. With launch control systems now banned, drivers are responsible for controlling the clutch on launch. A Formula One car has a narrow power band, and getting the balance between bogging down off the start and lighting up the tyres is a fine one. Drivers therefore need a clutch that provides a consistent biting point, both at the start line and when leaving the pits after a stop.

To achieve this consistency, clutch manufacturers invest considerable research time into ensuring that the interfaces between the clutch discs and the clutch basket and pressure plates are optimised. For example, if there is any degree of friction between the clutch drive plates and the arms of the clutch basket, the point at which the clutch bites can vary. Other factors such as the ratio between clutch paddle movement and the movement of the release bearing also need to be considered, with mechanical systems within the clutch helping to ‘soften’ the initial bite as the clutch is released.

There will be a number of new regulation changes in 2014 that directly and indirectly affect the design of transmissions. The most important is a change from seven to eight gears, in order to account for the reduction in maximum engine revs from 18,000 to 15,000. The number of forward gear ratios will also be increased from seven to eight, the benefits of which will be offset by restrictions on gear ratios.

At the start of the season, each team must nominate a set of ratios for the entire season – a considerable challenge given the wildly differing nature of tracks. If a team picks an unsuitable set though, they have the chance to change their ratios once within the season. The demands placed on the transmission by the new turbocharged engines and energy recovery systems will also be considerable. The power units will produce more torque than the current high-revving V8s, which will have to be accounted for in the gearbox and clutch designs.

Written by Lawrence Butcher

One of the most popular set-ups is a transmission using a planetary gear design, relying on a central ‘sun’ gear to provide a 1:1 direct-drive ratio once it has shifted from the launch ratio. To achieve underdrive in first gear, ‘planet’ gears rotate around the sun gear, with each set of sun and planet gears providing two speeds – a low ratio and a direct, 1:1 locked-up drive.

To create multiple speeds, the system is modular, with extra segments being attached together to provide up to five forward speeds. This layout means multiple underdrive ratios are available, as each segment of the system contains a direct and an underdriven ratio. This modular design can make life complicated though when it comes to creating a shift system, as each box requires a dedicated shift lever, but this can be simplified by adding a pneumatic shift system that removes the need for complex linkages.

In addition, automatic shift systems are available that use a control box and pneumatic system to select each gear automatically. The set-up of the system is exceptionally simple compared to many automatic-manual shifters found in race or production cars. It uses an rpm activation box that is controlled by interchangeable ‘chips’ that govern the shift point. To set a shift point for each gear the relevant chip is simply inserted into the box, and at a predetermined rpm the activation box triggers a shift. The system is not particularly smooth, but when running a 6.25 s, it doesn’t need to be. 

The internal action of these planetary boxes is such that when the transmission is in first gear the dedicated planet gear assembly provides the launch ratio. The planet gears are driven by an encompassing ring gear, and when the transmission is shifted a clutch pack compresses the rotating sun gear assembly to the direct-drive central sun gear, achieving a 1:1 ratio. Each case is shifted in succession, with its particular underdrive ratio determining the immediate drive ratio. Once all cases have been shifted then the entire transmission assembly is locked in the final 1:1 direct-drive ratio.

Another popular manual set-up uses a twin gear cluster arrangement as opposed to a planetary gear system. The idea behind this is to halve the load on the gear clusters, with each having to deal with only half of the input power from the engine. That means each gear has twice the number of effective teeth as a single-shaft set-up, with double the number of support bearings, increasing the torque capacity. Another benefit is that each rotating assembly can be made lighter than a similar single-shaft set-up, producing a lighter overall design. Once the power is applied on this system, each gearshift automatically overrides the previous gear, disengaging it and providing a near-seamless shift.

Both these transmission solutions are very different from those found anywhere else in racing, and highlight the lateral thinking needed to harness the power of engines capable of producing more than 2500 bhp. 

Written by Lawrence Butcher

One example of this optimisation is the shift mechanism developed by a supplier of transmissions to several leading Cup teams. Until recently, all its transmissions featured a ‘three-rail’ shift system that relied on three separate guide rails for the shifter mechanism. In a quest to reduce the complexity of its transmissions, the company developed a new single-rail shift system. Internally, the transmissions are similar but the single-rail system, introduced in 2010, uses just one external rail to rotate and move two internal shift forks to engage the four forward gears. The result is both a cleaner transmission casing, aiding packaging within the car, and a reduced number of moving parts, which ultimately reduces the potential for component

The same manufacturer’s gearboxes also have a novel ‘floating’ input shaft. Instead of using a solid shaft going into the clutch and the back of the engine, the design uses a splined hub that flexes with the car’s torsional movement. When the engine and transmission flex together, with the bellhousing absorbing the movement, the floating shaft will flex, eliminating any chance that the internal gears will bind up. Any such binding will increase both frictional power losses and increase transmission temperatures, with dynamometer testing showing that the floating-shaft system provides a useful increase in efficiency.

Following a similar line, teams will also pay great attention the construction of a car's driveshaft, to ensure that it is optimised from both a geometric and dynamic perspective. Any avoidable misalignment between the transmission output and the differential will increase frictional losses, as will any dynamic imbalance in the shaft (which can also cause unwanted vibrations, potentially impacting component durability).

Due to the regulatory constraints, finding ways of increasing the contribution of the transmission to overall car performance is no easy task, and involves looking at each area of the transmission in order to assess where developments may be possible. One area in which gains can often be found is mass reduction. Despite the regulations ruling out the use of exotic lightweight materials, advances in CAD and simulation systems have allowed manufacturers to further optimise the design of existing components.

The shift fork pictured below is a result of such developments. Fig. 1 shows the fork in its first iteration, while Fig. 2 is an improved version. Through extensive finite element analysis of the fork’s design, the manufacturer was able to remove excess material while also optimising the component’s geometry to increase its stiffness, improving the accuracy of gear changes.

Given the shift in focus among Cup teams back to weight saving with the current Gen 6 cars, combined with the incredibly tight nature of NASCAR racing, it is inevitable that developments such as these, that cumulatively can provide a performance edge, will continue to be a feature of transmission development.

Fig. 1 - The first iteration of a shift for designed for a four-speed NASCAR Cup spec gearbox

Fig. 2 - A later iteration of the same fork, optimised to reduce weight and increase stiffness 

Written by Lawrence Butcher

Although often housed in very different packages, the gears used in transmissions destined for racing use are essentially the same whether they are for a Formula One car or a WRC racer. Generally, racing transmission manufacturers produce gear sets from high-quality steel billets, and it would be fair to say that many manufacturers use very similar materials. This is because there is a limited number of steel producers worldwide who can refine high-alloy steels to the level required for producing reliable components. Some companies, particularly those involved in the construction of Formula One transmissions, have their own in-house metallurgy formulae, often developed in conjunction with mainstream steel suppliers.

There are notable differences in the manufacturing process used to produce gears intended for mass production and those designed specifically for the demands of competition use. Traditionally, high-volume gears tend to be made from forged blanks, whereas competition components are usually CNC-machined from billets. By using billets, the manufacturer can better ensure that material consistency is maintained from one batch of gears to another.

For all but the most severe applications, cutting will usually be done using specially designed and precisely ground tooling, in order to achieve the very close tolerances needed on complex tooth profiles. Lower quality gears are generally produced using off-the-shelf tools to save on costs, and these do not provide as fine a surface finish, which compromises tooth strength. Also, any misalignment due to poor machining has an impact on the performance and wear characteristics of dog engagement gears. Mass-production transmissions generally feature synchro engagement systems, which are much more forgiving in terms of tolerances, due to the slipping of the brass cones used in the synchro mechanism, making it practical to use less precise manufacturing processes.

Such levels of accuracy in manufacturing are vital because tooth design and profile is critical to the power-holding capability of a gear set. This requires that a number of factors be taken into account, including (but not limited to) gear ratio, tooth count and load paths through the gear shafts. For this reason, in the most highly stressed racing transmissions applications, even traditional machine cutting is not sufficiently accurate to produce gears to the tolerance required. Instead, these gears are ground to shape, a process involving the use of shaped grinding wheels that form a ‘negative’ of the gear tooth profile. Throughout the grinding process, the profile of the grinding bit will be constantly monitored to ensure it remains within tolerances.

Traditionally machined gears must be heat-treated after cutting, which can introduce measurable distortion into the piece, negating the precision of the tooth profiles. Grinding, however, takes place after heat treatment and thus gives a finished part with much closer tolerances. A second benefit of grinding is that it allows for gears and shafts to be made as a single component, which is not always practical due to tool path constraints using traditional milling methods. Normally, a cut gear will be welded to its shaft, introducing another potential source of distortion and reducing its structural integrity. With a ground gear this is not an issue, and again gives a better fitting and stronger component.

The final stage of the gear production process will often involve surface treatment of the gears. In some applications this will involve the use of DLC (diamond-like carbon) or similar coatings, but it is more common is for gears simply to be ‘superfinished’ using a number of proprietary abrasive or chemical processes. Depending on the process, such finishing techniques can improve by several percentage points factors such as component durability or result in reduced friction and thus increased efficiency.

Written by Lawrence Butcher

The internals of a Prototype or GT gearbox are very similar to those found in most other high-end racecars, be they single-seaters or saloons. However, they are subject to many small modifications to enable them to withstand the pounding meted out at circuits like Le Mans or Sebring. Some of the biggest challenges are presented by circuit topography, notably the bumps and kerbs.

This pummelling effects the transmission in two ways. First, it can create issues with lubricant distribution, and second it can cause shock loading of the driveline.

Extreme loadings are experienced under these conditions because shifts take place when the car’s wheels are not on the floor. This means the wheels are overspeeding and, as the car hits the ground and traction is regained, the driveline is shock-loaded. Clearly the way to ensure reliability is to engineer the transmissions to be tough enough to accept such abuse, but it is also imperative that the performance compromise this entails is as small as possible.

One manufacturer has invested heavily in test equipment to simulate such occurrences, allowing it to accurately model the demands placed on a transmission at any circuit. The system consists of a gimbal-mounted test rig that can reproduce the g-loadings a transmission encounters in use, based on actual track data. This rig, combined with an active transmission dynamometer to simulate the loadings on internals, has allowed the manufacturer to produce transmissions with a level or reliability that would have been unheard of even a decade ago.

The company also has another test rig that aids race reliability by allowing engineers to gain a greater understanding of transient differential operation. This knowledge then allows them to better adapt their differential settings during a race, for example moving from wet to dry running, something often required to account for the ever-changing weather conditions at Le Mans.

Previously, the only way to measure the operating characteristics of a differential was to use an entire gearbox ‘three-motor’ rig test facility, a sophisticated piece of equipment that characterises the whole driveline. Not only is this an expensive procedure, however, it’s also very difficult to isolate and characterise just the differential. Until now, the only other alternative was to measure the characteristics in the vehicle and on the track, which is even more costly

The differential rig features a single 45 kW motor that drives both sides of the differential under evaluation, thereby eliminating the complexity and cost of synchronising two motors. Despite its simplicity, the computer-controlled rig can operate in the ‘quasi-transient’ test mode necessary to characterise the dynamic behaviour of ramp- and plate-type differentials. In this test mode the rig can perform a combination of closed- and open-loop tests in order to analyse differential response to quasi-transient inputs. 

Other modes of operation permit torque values to be ‘dialled in’ to the rig to evaluate steady-state characteristics of the differential with respect to torque, speed and temperature sensing. The control system allows measurements to be sequenced manually or automatically for both 2D and 3D mapping, and can be fed with simulated or actual cornering data. It was also written to provide measurements with high rates of data logging.

To simulate a corner the relative motion across the differential can be varied from 0 to 500 rpm, and the rig has an input torque and locking capability of 1000 Nm (737 lb-ft). An oil system has also been incorporated to simulate the differential oil system, while the test environment can be pre-heated and is fully temperature controlled.

Overall, the development of test systems such as these go to show that, although transmission reliability is considered a given in modern endurance racing, it is only kept this way through relentless r&d.

Written by Lawrence Butcher

The dual-clutch transmission is by no means a new idea, having been conceived by engineer Adolphe Kegresse in the late 1930s (incidentally, Kegresse also developed the half-track drive system for off-road vehicles). He hoped to use the dual-clutch system in Citroen’s Traction Avant, but World War II intervened. There then followed a 40-year hiatus until Porsche picked up on the idea.

Porsche used the PDK gearbox in its 956 and 962 Le Mans cars. Its first success with the transmission was a win at the Monza 1000 km race in 1986, although the system’s reception among the team’s drivers was mixed. Derek Bell, who won the 1986 World Sportscar Championship, liked the ease of use of the system but was less than impressed with the extra weight it added over a standard five-speed sequential, going so far as to describe the cars as feeling like they had “a trailer on the back”.

The biggest benefit the system delivered from an endurance perspective was reliability – not so much in terms of the physical durability of the transmission components but in terms of saving the driver from having to operate the clutch. In a 1000 km race, one mis-shift on a manual transmission of the period could destroy it; the PDK removed this possibility. With improvements in sequential transmissions, however, and the introduction of various pneumatic and hydraulic shift systems, the need for the PDK disappeared, but Porsche did not forget the technology and continued to develop the system for roadcar use. Other companies, including transmission manufacturer Borg Warner, also developed dual-clutch systems, and since the turn of the century they have become common in many roadcars.

At the heart of a dual-clutch system is a two-piece main shaft, with one shaft section running inside the other. Each shaft carries three gears, with odd-numbered gears on one and even numbers on the other. Attached to each shaft is a multi-plate wet clutch, one running inside the other, with the engagement of each clutch being controlled by a hydraulic circuit. Gear selection is also controlled by hydraulic servo motors, and the vehicle’s control electronics govern overall operation, the theory being that the ECU determines which gear is likely to be needed next and pre-engages it. As soon as the driver initiates a shift, the clutches engage and disengage, selecting the next gear with minimal lag.

Porsche claims a shift speed of less than 100 ms for its PDK system, and given that Seat falls under the same VAG Group banner as the Stuttgart manufacturer, there is good reason to believe that the Seat system operates with similar rapidity.

But how effective is the transmission in racing use? Because of its extra weight and complexity over traditional sequential transmissions (and often regulations outlawing the system), the upper echelons of racing have not seen the adoption of dual-clutch gearboxes. However, anecdotal evidence from those who have used them in club motorsport, particularly in the US, have found they provide very short shift times without the expense of a dedicated motorsport sequential transmission. The main issues experienced have related to keeping the clutch oil temperatures under control, but the addition of extra cooling capacity has minimised such problems in all but the most punishing endurance races.

Whether the systems will see more widespread adoption in racing remains to be seen. It is telling though that one motorsport outfit, which prepares cars for SCCA competition, has found the control software on PDK transmissions so effective that it encourage its customers to let the system operate in fully automatic mode – under competition conditions!

This video provides a very good illustration of a dual-clutch transmission in operation.

Written by Lawrence Butcher

‘Roller’ chains have been the universally favoured means of transmitting drive on motorcycles for many years. While belt and shaft drives occasionally feature on road-going machinery, the chain is still the most practical means of transferring drive.

Both belts and shafts have their advantages. Belt drives are very clean, quiet and relatively inexpensive to produce; however, the belts are not nearly as strong as chains. To make them as strong, they are made wider, and a belt running on a modern-day superbike would need to be many inches wide, making it completely impractical from a packaging perspective. Shaft drives also provide a very ‘clean’ package, but are far too heavy and complex and with high parasitic losses to be considered for modern racing machinery. The chain though has the advantage of being lightweight, easy to package and the ability to withstand the power levels produced by the most powerful bike engines.

Roller chains [Fig. 1] used in motorcycle applications consists of a series of short cylindrical rollers held together by side links. There are two types of link alternating in the roller chain – inner links and outer links. The inner links have two inner plates held together by two sleeves or bushings upon which rotate two rollers, while the outer links consist of two outer plates held together by pins passing through the bushings of the inner links.

Despite appearing to be a very simple mechanism, the latest generation of racing chains have seen each of these parts optimised to the nth degree to ensure they provide the greatest possible strength for minimum weight. To put these improvements into perspective, a 1970s vintage GP racer produced in the region of 130 hp and would generally use a 530-size chain weighing about 2.5 lb per yard. Modern MotoGP racers produce more than 200 hp but run a 520-size chain that weighs only 2.3 lb per yard. The chains have also become more mechanically efficient, sapping less power through frictional losses.

While much of this improvement has been down to gradual development of parts such as the bushing materials, there have been a few key advances in recent years. One such advance is in the sealing mechanism used between the inner and outer links. Before riveting in the factory, the internal parts of the chain are filled with chain grease by vacuum. The sealing rings then have two purposes – to keep the internal lubrication in, and the dirt out, thus vastly improving the durability of the chain.

The load-bearing pins and bushings that enable a chain to bend over a sprocket have precious little oil to keep them happy. As if that wasn’t enough, high centrifugal forces that occur when the chain turns around the drive sprocket throws the oil out. The single biggest factor in chain wear is loss of lubricant, and the advent of O-ring sealed chains enabled the chain to keep its oil inside and stay lubricated where it counts. For most chains, O-rings are still the favoured sealing medium, but in high-performance applications a new X-ring design has taken over.

The biggest problem with O-rings is that they create a large amount of friction between the sealing surfaces, a less than ideal situation. An X-ring seal is exactly as it sounds: a seal with an X-shaped cross-section. When clamped between the two side plates, only two points of the ring make contact, providing a good seal but with greatly reduced friction. The improvement is significant, with some manufacturers claiming a reduction of up to 40% in friction over O-ring sealed chains.

Clearly this is only one improvement to what is a surprisingly complex component, and many factors – not least of which is correct installation and lubrication – are of critical importance to ensuring a chain lasts a race distance. But it goes to show that even the most humble of components still presents engineers with a neverending scope for performance improvement.

Fig. 1 - Layout of a roller chain

1. Outer plate
2. Inner plate
3. Pin
4. Bushing
5. Roller

Written by Lawrence Butcher

Organic

So-called ‘organic’ clutches have been around for the best part of 50 years. Organic is something of an outmoded term now, as it harks back to when asbestos was used in the clutches’ construction; a more accurate description now might be ‘mineral’. Asbestos made a good friction material due to its high heat resistance, good strength and high Mµ (coefficient of friction) when subjected to high temperatures. Unfortunately it is also carcinogenic and is therefore now banned in most industrial applications.

Organic friction material is now generally made from cellulose reinforced with materials such as fibreglass and mineral wool encased in a thermosetting phenolic resin base. The cellulose provides the initial bite while the mineral wool and fibreglass provide burst strength. While organic friction materials provide very good ‘feel’ and excellent initial bite though, they are not very effective in high-temperature applications. It is very easy to overheat the organic material under high torque loads, making them impractical for racing use except with small-capacity, low-torque engines, such as a bike-engined sportscar.

Kevlar

The next stage up from an organic friction material is one that uses chopped Kevlar fibres. Kevlar offers good burst strength and wearing characteristics but has a relatively low coefficient of friction. That makes for smooth engagement characteristics but requires the use of very high clamping pressures to provide sufficient friction to prevent slippage.

Kevlar provides far greater temperature resistance than organic clutches, and much lower wear rates, but it can still be ‘burnt out’ if subjected to excess heat. This is because the friction material does not return to its original state after exposure to high temperatures, greatly reducing its frictional properties.

Semi-metallic

Semi-metallic clutch materials look similar in construction to a regular organic clutch material but can withstand much higher levels of heat and are thus more suitable for high-torque applications. The clutches can still feature a woven structure, but instead of organic materials they will contain stands of brass or copper to improve the burst strength of the material while also increasing the resistance to high temperatures.

Semi-metallic clutch compounds can also contain powdered ceramic, copper, bronze, carbon or even iron mixed in with the organic material to further increase friction at elevated temperatures. Semi-metallic discs that contain high levels of iron or ceramic material can have somewhat reduced ‘feel’ though, with the discs tending to bite suddenly.

Sintered metals

Sintered items are made by filling a mould with powdered material and then fusing it under heat and pressure. For clutch discs, it is usually a mixture of metallic compounds designed to provide the optimum coefficient of friction and wear resistance, and these ingredients include (but are not limited to) copper, bronze, iron and carbon.

This is where so-called carbon-metallic materials come into play. Mixing carbon and ceramic into the compound gives the self-lubricating benefits of a copper or bronze base material – providing smooth engagement – as well as the high bite and temperature resistance of carbon and ceramic. Under extreme conditions though, the copper base material can melt and coat the friction-modifying materials, massively reducing the friction level.

In applications where extreme temperature is an issue, for example an 8000 hp Top Fuel dragster, the favoured option is sintered iron. These are also produced from a powdered base stock and can withstand very high temperatures – in fact the friction increases with temperature. Due to the very aggressive nature of the material, however, engagement is very harsh and sintered iron discs tend to be used only in drag racing applications.

Carbon-carbon

The most recent development in clutch material has been the introduction of carbon-carbon. In this type of clutch all the friction surfaces, including the flywheel mating and floater discs, are made from amorphous carbon material. These parts are made by heating preformed discs of white polyacrylonitrile (PAN) fibres until they turn to a black, pre-oxidised state. PAN is a synthetic, semi-crystalline organic polymer resin, which is used as the basis for high-quality carbon fibres. Once pre-oxidised, the fibres are layered together before being oxidised and then cut to a rough shape.

These rough blanks are subjected to two densification heat cycles at more than 1000 C before being machined to a finished shape. It is these densification cycles that make the manufacturing process so lengthy, with each cycle taking several hundred hours. During the process, hydrocarbon-rich gases are injected into the ovens used to heat the blanks, allowing the layers of material to fuse together and form a solid disc.

Friction modifiers can be added to the mix to alter the material’s characteristics, the result being a clutch that is highly resistant to temperature, and with friction increasing as the clutch heats up; the discs are also very light, reducing drivetrain inertia. The very long manufacturing process means carbon-carbon clutches are very expensive though, and thus limited to the top echelons of motorsport, although lower cost carbon-steel options are available. These feature carbon friction discs combined with steel floaters and flywheel mating surfaces.

Conclusion

Ultimately, carbon-carbon is the ideal solution for a racing clutch friction material. Unfortunately its high cost makes it inaccessible to most racers. Manufacturers are working to produce cheaper varieties of the material, and no doubt these will filter down from the upper reaches of the sport over time. Until then, however, other clutch manufacturers will continue to improve the performance of their organic and metallic compounds.

Written by Lawrence Butcher

In essence, the clutch found on a Top Fuel dragster or Funny Car is a centrifugal unit, augmented by a pneumatic release bearing. It operates according to a process of controlled slippage to feed in the engine power gradually; if it were to lock immediately then the dragster would simply break traction and light up the rear wheels.

The basic components that make up the Top Fuel clutch are as follows. At the engine end is the flywheel, which forms an integral part of the system, while at the rear is a pressure plate or ‘cap’, with the rest of the components clamped between these parts. The components consist of up to five friction discs, separated by ‘floater plates’; these transfer the drive between the friction discs. The floaters are held in place by a number of stands, with studs running through their centres, which also connect the cap and the flywheel together. The cap also doubles as a mounting for the clutch fingers, which as explained below are the key to the clutch’s operation.

The clutch’s friction discs are generally considered disposable items on a race weekend, and sintered iron is the favoured material for their construction owing to its excellent friction properties. The material can be engineered to provide excellent bite, and is available in varying Rockwell hardness ratings to govern the level of bite and slippage.

Sometimes all the discs will be replaced after each run; however, combinations of fresh and worn discs can be used to provide a particular level of slippage. Sometimes the discs and floaters can weld together during the run, and even after a car returns to the pits the components can be so hot that the mechanics need asbestos gloves to handle them. Most clutches now feature titanium pillars for the floaters and titanium cover plates, while the floaters themselves are made from steel.

The clutch slip is primarily controlled by levers attached to the clutch cover. As engine rpm increases, the levers are forced out against the pressure plate, compressing the clutch and locking the friction discs and floaters together, thus allowing more power to be gradually transmitted to the wheels. Most clutches have 18 levers, which can be adjusted individually to control locking of the plates. The addition of weights, in the form of nuts, is used to fine-tune the rate at which the clutch engages.

Gauging how much weigh to add or remove is one of the black arts of dragster set-up. Adding weight to the fingers means the clutch will engage at a lower rpm on the initial hit of the throttle, whereas taking weight off will allow the motor to reach higher rpm before engaging. If a car’s crew chief determines that a track has plenty of grip he will add weight to the clutch, meaning the power will be deployed quicker, resulting in a faster run. Conversely, if a track is low grip he will remove weight to provide a softer ‘hit’. It is a very fine line between pulling a fast run and vaporising the tyres, and the winner of most events will invariably be the one to have best judged their clutch set-up.

The driver has no control over the clutch’s engagement. In addition to the centrifugal weights, the level of locking is also governed by a pneumatic release bearing. As the driver hits the throttle, a microswitch initiates a pneumatic timer that controls the operation of the clutch release bearing, often called the ‘cannon’. This bearing limits the movement of the clutch fingers and the level of locking in the clutch, and as the clutch timer counts down, pressure is released from the hydraulic system, moving the release bearing and allowing the fingers to lock the clutch up.

It would be reasonable to assume that this timed engagement is controlled by a microprocessor system; however, only a very basic mechanical system is allowed. The clutch will typically be set so that when the driver hits the throttle to launch the car, the engine will be spinning in the region of 5000 rpm before it engages enough to actually launch the car off the starting line. If for some reason the driver needs to get off the throttle to prevent wheel spin, or any other factor interrupts the run, the timer sequence cannot be adjusted.

Overall, the Top Fuel clutch is a fascinating piece of engineering, tailored to the unique demands and regulations of a brutally powerful form of motorsport.

Fig. 1 - A Top Fuel team will use many clutch plates over a weekend, with combinations of worn and new plates used to tailor the level of clutch lock-up

Written by Lawrence Butcher

What makes torque converters unique is the transfer of power from the engine to the transmission, with fluid being the only connecting factor, allowing the engine to idle without the threat of stalling. Along with preventing engine stall, the converter helps improve performance by nearly doubling the torque of the engine, offering a higher degree of acceleration - a key factor for a quarter-mile racer.

One of the most important issues with a torque converter in racing applications is the stall speed. This is the rpm at which a given torque converter (impeller) has to spin for it to overcome a given amount of load and begin moving its turbine element (the component that transfers drive to the gearbox). This is a vital consideration when speccing a converter for racing use, as the stall speed is the rpm at which the converter needs to be spinning to create enough fluid force to overcome vehicle inertia at wide-open throttle.

Clearly the inertia is governed by vehicle weight, gear ratios, the rolling resistance of the tyre, suspension set-up and chassis stiffness. All of this has to be factored in when specifying a converter for a particular car, as the stall speed will in essence govern the launch rpm of a vehicle. If the stall speed is too high, the engine will produce too much power and the driver will smoke the tyres; too low and he/she will bog down off the line. The units used in classes like Pro Mod are adjustable at the trackside for stall speed, allowing the crew chief to fine-tune the launch dynamics of the car based on track and environmental conditions. Whereas a standard-style converter is welded together as a sealed unit, these race-specific components can be fully dismantled in order to change the characteristics of the stator and turbine.

The torque multiplication through the stator occurs when the converter is in ’stall mode’ and during initial vehicle acceleration. As the vehicle accelerates, the torque multiplication decreases until it reaches a ratio of 1:1 with the crankshaft torque. The design of the stator will influence the torque multiplication characteristics of a converter, with a typical figure being in the region of 2.5:1. The more drastic the change in fluid path caused by the stator from its ‘natural’ return path, the higher the torque multiplication ratio a given converter will have. Torque multiplication does not occur with a manual transmission clutch and pressure plate, hence the need for heavy flywheels, very high numerical gear ratios, and high launch rpm.

This is where the relationship between converter stall speed, engine output and chassis and tyres becomes very important. If an engine producing 200 lb-ft of torque at 3000 rpm is matched with a converter with a torque multiplication of 2.5:1 and a stall speed of 3000 rpm, 500 lb-ft will be available at launch. However, if that engine produces 300 lb-ft at 4000 rpm, a converter with a higher stall speed would give a bigger ‘hit’ of torque at launch.

Whether this is beneficial depends on if the chassis and tyre set-up can handle the extra torque without breaking traction. If not, and the torque overcomes the chassis, a converter with a conversion factor of 2:1 and a stall speed of 4000 rpm could be a better option. Although the initial torque hit will be lower, the disadvantage will be outweighed by a clean start and a resulting lower elapsed time. Clearly the different permutations of this scenario are almost endless, but it gives an insight into the complexity of using a torque converter to best effect.

Written by Lawrence Butcher

This does not necessarily mean increasing engine power though; efficiency gains in the transmission can be just as useful, allowing engine power to be better exploited. While the internals of the gearbox and differential are tightly controlled, engineers have been able to put coating technology to good use here.

There are many different proprietary coatings on the market, but they all are aimed at reducing wear and friction between rotating assemblies. Most manufacturers and teams remain understandably tight-lipped about the exact make-up and performance of various coatings, although one manufacturer has published research to show that its coating technology reduced parasitic power losses in a ring and pinion gear set by between 0.15 and 0.21%, which on a 500 hp engine equates to almost 1 hp. However, to gain a better understanding of the impact coating technology can have on gear longevity it is necessary to look a little further a field - into outer space.

To the casual observer, NASCAR could not be further removed from NASA (apart from similar acronyms) but both industries attract some of the best and brightest engineers and present them with an extremely harsh working environment for their designs. Engineers at the NASA Glenn Research Center undertook a research study* to investigate wear characteristics on both uncoated and coated spur gears.

For the experiments, a lot of spur test gears made from AISI 9310 gear steel were case-carburised and ground to aerospace specifications. The geometries of the 28-tooth, eight-pitch gears were verified as meeting American Gear Manufacturing Association (AGMA) quality class 12, after which one-half of the gears were randomly selected for coating.

The method of coating was selected to achieve the desired adhesion, toughness, hardness and low-friction characteristics. First the gears to be coated were prepared by blasting (vapour honing) with aluminium oxide particles, then they were given a thin adhesion layer of elemental chromium followed by magnetron sputtering of the outer coating consisting of carbon (70%), hydrogen (15%), tungsten (12%) and nickel (3%). In total, the coating thickness was about 2.5-3 µm. As compared with the steel substrate, the coated surface was harder by a factor of about two and had a lower elastic modulus.

All gears were tested using a control synthetic oil, at 10,000 rpm and with a Hertzian contact stress of at least 1.7 GPa (250 ksi). Tests were run until either surface fatigue occurred or 300 million stress cycles were completed, using either a pair of uncoated gears or a pair of coated gears (coated gears mated with uncoated gears were not evaluated). The results showed that the coating extended the surface fatigue lives of the gears by a factor of about five compared to the uncoated gears. When combined with the potential frictional gains, it is clear to see why coatings technology has proved so attractive to transmission developers.

*Krantz, Timothy L., et al, "Increased Surface Fatigue Lives of Spur Gears by Application of a Coating", NASA, 2003

Written by Lawrence Butcher

Composites have been used to great effect in racecar construction since the 1980s, but it is only in the past decade that materials technology and understanding of composites have made their use in transmissions possible. When first introduced, most carbon fibre parts were constructed using the 'wet lay-up' process, where dry carbon fibre weave is coated with an epoxy resin, in essentially the same manner as glass-reinforced plastic had been constructed.

The advent of pre-preg materials, where the carbon weave is pre-impregnated with thermo-curing resin, and the use of autoclaves, dramatically increased what was possible with composites. Unsurprisingly, Formula One was the first place composite gearboxes appeared, and with the exception of Audi's R18 e-tron quattro and ultra sportscars, this remains their sole preserve. The main reasoning behind this is cost, in terms of development, production and durability.

The biggest challenge in creating a composite gearbox casing is ensuring that it has the correct structural properties while still allowing for the compact packaging of the internals. Carbon structures can be made very substantial, but until recently, FEA simulation of laminate structures was not as advanced as for metallic structures, so engineers had to work very much on knowledge gained through prior experience. Modern FEA packages, however, are capable of accurately modelling the properties of composites, allowing structures that require strength and stiffness in multiple planes (such as a gearbox) to be analysed.

If resources are sufficient then the performance gains that a composite casing affords make their adoption viable. Honda's 2006 RA106 Formula One car featured a composite transmission casing, which the team said was 30% lighter with 14% greater torsional stiffness than a similar aluminium unit. A titanium casing would come close to these gains, but the cost implications of producing one are just as onerous as a composite unit.

The RA106's casing was made from a monolithic, uncored laminate that varied in thickness from 2 to 4 mm (0.08-0.16 in). The casing also featured two internal 'bulkheads' consisting of composites with titanium inserts, which supported the rotating assemblies. Similar to the rest of the car's monocoque, the external mounting points for suspension assemblies were made from titanium, and chemically treated to improve bonding to the carbon structure.

Beyond the obvious loading that a gearbox is subjected to, thanks to powertrain and cornering forces, the environment it lives in is a hostile one, especially in a single seater. The biggest enemy is heat - both the engine and exhaust systems are in very close proximity, and the ambient temperature in this area is likely be considerably higher than regular composites can tolerate.

Advances in the epoxy resins used in the laminating process have therefore been key to making composite gearboxes a viable proposition. Most epoxy resins begin to fail above 150 C, but new high-temperature epoxy resins have become available that are stable to over 300 C which, combined with technologies such as heat-resistant coatings, allow composites to survive in areas where previously this was not possible.
These days, composite casings are still far from commonplace. The front-running teams in the Formula One paddock all run them, but it is still likely to be some time before they can be found on racers in the lower series.

Fig. 1 - Honda's 2006 RA106 featured a composite gearbox casing (Courtesy of Honda Racing)

Written by Lawrence Butcher

One area that has seen considerable development in recent years is the material choice for the transmission casings, with new techniques allowing for lighter units that still have the durability to survive the punishment of Le Mans. This has been especially apparent in the efforts of the leading works teams culminating in Audi's use of a full composite gearbox housing in the 2012 R18 e-tron quattro and ultra LM P1s. This will be covered in more detail later, but first it is worth looking at developments slightly lower down the scale.

Traditionally, motorsport gearboxes have been cast from magnesium or aluminium alloys, and these are still the predominant materials used. Cast magnesium alloy is lightweight and highly fluid when cast; it allows for thin wall sections to be achieved and is more resistant to fatigue than aluminium; however, aluminium is ultimately stronger. For example, RZ5 magnesium, which is used in a number of sportscar gear casings, has a tensile strength of about 200 MPa, compared to BSL169 aluminium alloy, which has a tensile strength of 240 MPa when sand cast.

MSR/EQ21 magnesium is also seeing increased use due to its improved properties at higher operating temperatures, thanks to the incorporation of a higher percentage of rare earths in its chemical composition, while new casting methods that allow for very thin wall sections to be produced has seen aluminium gaining more widespread use. Advances in FEA analysis have allowed for more accurate placement of material, resulting in aluminium casings being marginally lighter than an equivalent magnesium component.

Most of the front-running LMP cars use thinwall aluminium casings, including the Toyota TS030, as did the Peugeot 908 HDi FAPs until 2011 (when the team switched back to a magnesium unit). Thinwall alloy cases are considerably more expensive than magnesium, so the cost-performance benefit is not always clear-cut, especially when the fatigue life of the two materials is taken into account, with aluminium components having a shorter service life than more shock-resistant magnesium parts. There is also a middle ground between aluminium and magnesium in the form of hybrid aluminium-magnesium alloys that provide some of the benefits of both materials.

Over the past two years, Audi has rather upped the ante; the R18 featured an investment cast titanium casing, housed within a carbon 'superstructure' that carried the suspension mounting points. For the R18 e-tron quattro and ultra it went a step further, introducing a fully composite casing, part of an extensive weight-saving programme that looked at all areas of the car.

Fully composite casings first appeared in Formula One in 2002, running in the Arrows A23, with McLaren soon following suite with the MP4-18; however, it has taken most of the decade for the technology to filter down to Le Mans.

The biggest challenge to producing a composite case is in accommodating the complex internal shapes required within a casing, as well as ensuring that the composites remain stable when exposed to potentially high operating temperatures and high loadings on several planes. Advances in FEA analysis software and specific software for ascertaining the optimum lay-up arrangements for composites means that the construction of complex structural parts is now feasible.

Producing a composite transmission is still a daunting prospect, and requires significant resources and expenditure, but as hybrid systems become more prevalent at Le Mans, and teams look to offset the weight they add, it is likely that composite boxes will become more commonplace.

Fig. 1 - The full carbon case from Audi's 2012 R18 e-tron quattro and ultra LM P1s

Written by Lawrence Butcher

With the simplification of the technology allowed now in the WRC, teams are limited to using purely mechanical systems, but this was not always the case. Until 2006, active differentials were permitted, and it would be fair to say that they were some of the most advanced units ever to grace a racecar. Their use enabled a flexibility of set-up previously impossible with a mechanical diff, and more important they allowed every ounce of available grip to be exploited.

Active differentials are usually composed of a classic, free (that is, non-self-locking) device that holds what can best be described as a clutch. The clutch progressively operates the differential's locking and is commanded by an ECU and activated by hydraulic pumps or electric current.

At its most basic level an active differential can control the level of slip in the diff using information gathered from various sensors, but those found on WRC cars were far more complex than this. Through the use of front, rear and centre diffs, linked through a main ECU, the level of wheel slip could be controlled on each corner of the car, with torque being transferred both laterally and from front to rear.

Each differential would be mapped, much like an engine's ECU, and drivers would often have access to several different maps from a dashboard controller. The ECU would draw data from multiple sensors - typically those for throttle, rpm, speed, brake pedal, handbrake, switch position and so on - and interpolate the information onto a 3D map that then establishes the appropriate amount of differential lock for the particular driving circumstances.

Throughout a corner, drivers would be looking for the car to behave in a certain way, so the relationship between throttle position and the state of the diff was very important. For example, under braking, the diffs would initially lock fully, providing a very stable platform, but as steering angle was applied (still under braking) the locking had to be eased back to prevent understeer.

As the driver began to get back on the throttle, the state of the diffs would change again, with a balance having to be struck between the front, centre and rear settings. This would see the front diff locking considerably to help pull the driver into the apex, with the centre diff helping to keep the car's attitude neutral while the rear diff steadily increased its pressure to aid traction, but not so much as to produce oversteer.

Given the huge variety of surfaces experienced during a rally, this capability was evidently a great advantage, and unsurprisingly the development of differential maps was an ongoing process for engineers. Unfortunately such open-ended technical development cost money, and with the WRC looking to reduce spending in the sport the active diff became a casualty of new regulations in 2006.

Fig. 1 - A Mitsubishi active rear differential; note the clutch set-up on the left-hand side

Written by Lawrence Butcher

The forces the shaft is subjected to are substantial, so for many years the only option was to use metallic materials - usually steel or in some cases aluminium - for their construction. Despite providing the strength required to achieve the durability needed with high-output race engines, especially when used for endurance racing applications, metallic prop shafts have some disadvantages. Notably they are heavy, which adds to vehicle weight and increases parasitic power losses, but more important the material choice severely limits the maximum shaft length used in high-rpm applications.

This is due to the bending resonance experienced by a tube as it reaches its critical speed. This resonance is described by the following equation:

where:
Nc is the critical shaft speed
L is the shaft length
I is the tube's second moment of area
A is the cross-sectional area of tube
C is the constant vibration
E/ñ is the specific modulus of the shaft material

The resonance will destroy a prop shaft, and while that's not a huge problem on low-revving mainstream automobiles, it is a serious issue in high-speed race applications. The only ways to increase the critical speed are to increase the diameter of the tubing or increase the specific modulus of the material. Space constraints invariably set a limit the diameter of the tube, while it is an interesting characteristic of metals that the specific modulus is approximately constant, despite big differences in density.

One driveshaft manufacturer discovered that at a speed of 8000 rpm the longest steel shaft that could be used before resonance became a problem was only 1250 mm. This is too short for most applications, so the general approach is to use a two-piece shaft with a supported joint in the middle. Evidently from a racing perspective, that is not ideal, as it introduces both extra weight and complexity into the system.

The solution to this problem was the introduction of composite prop shafts, the use of which has been made possible by advances in materials technology. By using fibre-reinforced composites, it is possible to orientate the fibres in a tube's structure so that the bending modulus has a high value (above 100 GPa) while the specific gravity is low (below 1.6). This leads to a favourable specific bending modulus and enhanced critical speed.

This is not a particularly new idea, and the first composite shafts began to appear in production cars in the late 1980s, although these used an aluminium core reinforced with a carbon fibre outer sleeve. Composite prop shafts were experimented with in the 1990s in rallying and circuit racing, but there were many failures. However, the past decade has seen great improvements in both composite materials technology and, more important, the ability to simulate and assess different construction techniques. The result is that nearly all competition GT cars now feature composite shafts, which have proved to be more than up to the required tasks.

Making these shafts is a complex process, and predominantly uses a filament-wound carbon tube for the main structure. Manufacture involves winding filaments under varying amounts of tension over a male mould, or mandrel. The mandrel rotates while a carriage moves horizontally, laying down fibres in the desired pattern. The exact pattern of the 'lay-up' will be determined using FEA software and specialist composite CAE packages, to ensure the fibres are oriented in the optimum direction to absorb loadings. The connections at the tube's ends are usually a very tight interference fit, and bonded to the tube with high-strength resin.

The overall result is impressive. One manufacturer says a representative 1.5 m prop shaft for GT use weighs only 2 kg, compared to about 10 kg for a similar two-piece steel item. Yet despite the low weight the shaft can still withstand up to 3500 Nm of torque while operating in an environment at 100 C. This represents a major improvement in both overall vehicle weight and rotating mass, and makes it easy to see why nearly all front-engined, rear-wheel-drive GT cars now run composite props.

Fig. 1 - Composite prop shaft for use in endurance racing

Written by Lawrence Butcher

Mandating spec components is the route taken by many touring car championships, so here I want to look at some of the transmission types and technologies used in touring cars around the world.

If you look at different national 'touring car' series, it is clear that the definition of what constitutes a touring car varies somewhat. In the UK, they have always been closely related to their road-going counterparts - the same being true in the World Touring Car Championship - with both front- and rear-wheel-drive models competing. Both series use I4-configuration engines, producing about 300 bhp.

The German national series, the DTM (Deutsche Tourenwagen Masters), is a more serious proposition, using space-framed cars with power in the region of 500 bhp supplied to the rear wheels. Across the Atlantic, the Brazilian TC2000 also uses space-framed cars, which from 2012 will be powered by a 2.7 litre V8 delivering about 430 bhp to the rear wheels.

All of these series, excluding the WTCC, mandate the use of a 'spec' transmission - in the case of the DTM, teams can chose one of two boxes - so what sort of technology does an off-the-shelf, control transmission offer in 2012?

Starting with the BTCC, the gearbox specified under the NGTC regulations is a development of a unit with its roots from the turn of the century. The transmission is a six-speed sequential with two different casing styles to suit front- and rear-wheel-drive applications, both casings being made from aluminium.

An unusual feature of the new gearbox is that the gears on the layshaft are stacked in a non-conventional manner to accommodate an overlapping barrel gearshift system. In a conventional barrel shift mechanism, there is no overlap between the barrel, and the gears are stacked in numerical order - 1-2-3-4-5-6. In the new arrangement, the gears are ordered 1-6-2-4-3-5, allowing overlap on the barrel track and leading to a reduction in gearshift time with no difference in the way the driver operates the gearbox. Other interesting features include the inclusion of an external adjuster for the differential, allowing mechanics to rapidly change the diff characteristics trackside.

A very similar unit is also used for Brazil's TC2000 cars, which goes to show how versatile a modern racing transmission can be. Although the transmission needs to deal with a considerably higher power output, the maximum torque of the V8 in the TC2000 and the force-inducted I4s in the BTCC are very similar, which is the important factor in terms of transmission longevity.

Despite also using a V8, the German DTM cars are closer to sports prototypes than touring cars. They use a rear-mounted transaxle system, very similar to those found in many front-engined GT racers, and also feature a semi-automatic paddle shift system for actuation. DTM also operates at a much higher cost level than other touring car series, so there are fewer concessions to the bottom line. For example, where both the TC2000 and the BTCC use an aluminium alloy casing, DTM units are made from lighter but far more expensive magnesium.

As you can see, despite cost savings and spec parts, manufacturers are still developing new and innovative transmissions outside single-seater and sportscar racing.

Fig.1 - The current 'spec' gearbox of the BTCC. This rear-wheel-drive variant is also the transmission of choice for the Brazilian TC2000 championship

Written by Lawrence Butcher

As a brief overview, a quick shifter allows for full-throttle gearshifts without having to use the clutch. As the rider shifts gear, the system senses the impending change and cuts the ignition, allowing the shift to take place. The key part of this operation is the sensing of the start point of the shift, the accuracy and reliability of which is essential for a smooth gear change.

In the past, this has been accomplished using a pressure sensor or transducer, both of which rely on the mechanical operation of a switching mechanism by the shifter. There are, however, a number of disadvantages to this type of system that can disrupt shifts and lead to unreliability. The most notable of these is their susceptibility to vibration and shock, isolation from which can reduce the sensitivity of the system. The gear lever also has to move a certain amount in order to initiate the shift. This, combined with the fact that a spring or tensioning system is relied upon to hold the switch open - the strength of which dictates the force needed to initiate a shift - can make it difficult to tailor the system to different riders, linkages and transmissions.

In recent years, there has been a move away from this type of sensor towards the use of strain gauges to detect the onset of a shift. By integrating a strain gauge into a load cell attached to the shift lever, the exact point at which the rider wants to shift can be detected. (A load cell is simply a strain gauge with an amplifier that produces a voltage relative to the force being applied. This is typically 2.5 V for no load and increases to 5 V for a positive load or decreases to 0 V for a negative load.)

There are, however, a number of issues that increase the complexity of using a load cell to actuate the throttle cut-off.

The first is complexity. Unlike a simple switch, the load cell will require additional electronics to process the signal and interface the system with the onboard control electronics. Then there is drift from zero. This means that when the force is released, the load cell does not always return to the no-load voltage.

Next comes temperature drift, a serious problem with all load cells. When the cell changes temperature, the output voltage changes radically relative to force. This means that when the bike is cold, not much force is required to initiate a shift, but when the bike is hot then a lot more force is required. And then there's creep. With age and use, the load cell's zero and load values will change. This means the electronics will require regular recalibration.

However, most manufacturers of quick shifts provide a compact digital signal processor unit that is easily integrated with existing controls and compensates for these shortcomings. Despite these issues, load cells still provide a far superior solution to a pressure switch, with the new generation of shifters now dominating among teams.

Fig. 1 - Load cell or strain gauge shifters are easily added to a bike's existing gear change system

Written by Lawrence Butcher

Magnesium:
Magnesium has long been a favoured material for gearbox casings, thanks to its useful mechanical properties. Magnesium alloys are about 35% lighter than aluminium alloys, and certain alloys can be heat-treated to UTS values approaching 43 ksi, making them attractive because of their high strength:weight ratio.

The stiffness of magnesium is generally only about 63% of aluminium alloys, however, so components being switched from aluminium to magnesium will need larger cross-sections and section moduli to achieve the same stiffness as the aluminium part, but can still give a weight saving of 20-25% depending on the design. Magnesium can also absorb shock loadings better than aluminium, making it ideal where durability is need. However, in recent years the emergence of new casting techniques for aluminium has toppled magnesium from its position as the material of choice.

Aluminium:
In the past, the major factor governing the weight of aluminium gearbox casings was the casting wall thickness. Traditional casting methods did not allow thin wall sections to be cast reliably, thus limiting weight savings irrespective of the actual component strength. However, new casting methods that allow for very thin wall sections to be created mean engineers can now create casings that are marginally lighter than a magnesium counterpart, while still retaining the same strength characteristics.

As a result of these new casting methods, combined with advances in FEA simulation, the benefits of an aluminium casing now outweigh those of a magnesium unit, although the production costs are considerably higher. As a point of note, the manufacturer of the aluminium gearbox used by the HRT and Virgin Racing teams quotes the weight of the unit as being "in the region of 40 kg".

Titanium:
The first titanium gearbox to appear in Formula One was produced by Ferrari in 1997, and was fabricated as opposed to cast. Although it was exceptionally lightweight compared to a magnesium or aluminium unit, the difficulties of fabricating titanium meant the production costs were astronomical.

It was not until the Minardi team began experimenting with a cast titanium unit in 2000 that the material became a viable option for casing construction. The use of a rapid casting process, with the extensive use of rapid prototyped patterns (created using a selective laser sintering process), meant that not only did the production cost of the transmission fall considerably, but the entire design could be optimised to a greater degree. This was thanks to the complex patterns that could be created through rapid prototyping, allowing for internal structures that were not possible with traditional casting methods.

The end result in the case of the Minardi gearbox was a weight saving of 25% and a reduction in package size of 20% over the previous magnesium casing. Titanium is still used extensively in Formula One transmissions, although the current trend is to combine it with composite components.

Composites:
The first carbon fibre Formula One gearboxes appeared in the late 1990s, but considerable problems were experienced in relation to the heat rejection properties of the material. Today, the most common use of carbon in transmissions is in conjunction with metallic - usually titanium - structural components.

One of the first teams to take this approach was Renault with its R23 car in 2003, which used a casing with a titanium lower section and carbon fibre upper section. However, the design was dropped the following year in favour of a 100% titanium unit. The current trend is now for the structural 'bulkheads' of the gearbox to be fabricated from titanium, with the rest of the structure moulded from composite material, providing a very light yet exceptionally stiff structure.

Fig. 1 - The fully cast titanium gearbox produced for the Minardi team provided significant weight savings

Written by Lawrence Butcher

In terms of the gearbox itself, the main advances made by manufacturers have centred on reducing frictional losses. With the various engine offerings all falling within a few horsepower, even a small percentage reduction in power loss can prove decisive, especially on tracks where restrictor plates are mandatory. Surface finishing techniques have played a major role in these frictional reductions, with all the transmission manufacturers using either a media or chemical-based finish on components such as gears and shafts.

Chemical processes are gaining greater favour, thanks to their ability to improve the material characteristics of a component as opposed to simply smoothing its surface. Not only does this attention to finishing reduce friction between the gears, it also increases their durability. Combined with an increasing trend towards ground as opposed to milled gear sets, this has meant that the meshing between gears at a microscopic level has improved. This in turn has allowed engineers to reduce the size of the gears, further reducing parasitic losses through a decrease in rotating mass. One manufacturer has even gone so far as to add scallops to the gear face to remove yet more mass.

The gearbox is not the only area that has seen improvements, with the ring and pinion being the subject of considerable development by teams and parts suppliers to improve both performance and reliability. For many years, teams would use an off-the-shelf, machine-cut ring and pinion gear which was less than optimal in terms of performance and subsequently treated as a disposable item after only a few races.

One manufacturer though has bucked this trend by producing a bespoke item, featuring the latest advances in gear manufacturing technology. A key factor that enabled the company to make considerable gains in performance was access to advanced gear design software, which allowed engineers to optimise the profile of the gear teeth for maximum efficiency. This design capability, combined with the use of a proprietary steel formula and features such as gun drilling of the pinion shaft, has resulted in a ring and pinion that is lighter and more durable than anything currently available.

In practice, the new components can comfortably be used for ten-plus races without any reduction in performance, whereas the old gears would be replaced after only three. However, crew chiefs are still unwilling to run them for their proven usable life, so have been replacing them after five or six races. This led the manufacturer to reassess the gear design: if the gears were going to be used for only five races, there was no need to overbuild them. The result was a new, revised version of the ring and pinion, with a shorter service life but even greater efficiency gains.

This is only a brief glimpse of the work being undertaken to improve what is essentially a 1950s-vintage driveline. But rest assured, although its roots may be in the past, the modern NASCAR transmission is not short on cutting-edge technology.

Fig. 1 - A current-generation four-speed NASCAR Cup gearbox

Written by Lawrence Butcher

Honda is currently the only team on the grid to admit to using a system, and it certainly made an impact on its debut at the Sepang pre-season tests in 2011. Honda Racing's Shuhei Nakamoto said at the time, "Obviously HRC is not using anything illegal, but it's true that we have something new on the transmission that currently allows for faster gear changes."

Seamless shifting as such is not unique in motorcycle transmissions, with Honda having developed a dual-clutch transmission (DCT) system for its larger-capacity consumer bikes. However, by its very nature, a DCT system is large and bulky, exactly the opposite of what is needed in a GP bike, as any performance gains would be offset by added weight.

A system similar to those found in Formula One would also not be as appealing. While a subject in its own right, essentially Formula One seamless shifts engage two gears at once, giving about a 4 ms window to disengage one or other gear, before the box destroys itself! High-pressure hydraulic systems are used to ensure that the first gear can be disengaged, preventing this happening. This is not an attractive proposition on a GP bike, because of the packaging requirements of the hydraulics and the possibility of locking the wheel if the system fails.

There are several motorcycle seamless shift systems available, but the GP teams are keeping tight-lipped about whose technology they are using. Two British companies, Xtrac and Zeroshift, produce the technology, but neither is willing to confirm their involvement with a particular team. Zeroshift though was willing to explain the workings of its system.

In a regular bike gearbox, the selection of the gears is still achieved by a sliding dog ring, which spins freely between shifts. The Zeroshift transmission eliminates the dog rings, and instead uses 'bullet rings' to engage the gears. A gear is selected when one ring is moved until its bullets hook onto drive teeth on the side of that gear. A second bullet ring moves in the same direction, with its bullets filling the gap between the teeth, eliminating any slack between the gears. Eliminating this slack is what creates seamless upshifts and downshifts. The shift sequence is illustrated below.

Fig. 1 - The Zeroshift system eliminates the gear dogs found in a standard racing transmission, replacing them with shifter 'bullets'

Fig. 2 - Neutral selected, with first gear about to be engaged

Fig. 3 - All three bullet teeth then mesh with first gear

Fig. 4 - Drive is maintained to first gear as the bullet starts to move in readiness for the second gear

Fig. 5 - The second gear ring is now engaged, but drive continues to be transferred to first gear

Fig. 6 - Second gear fully engaged with first gear bullet now merging into second gear ring

Fig. 7 - First gear has been fully disengaged, with second gear now transmitting all the power

This not only gives faster shift times, it also presents the rider with a far more stable platform under shifting. With a seamless transmission of power, it is even feasible to upshift in corners in adverse conditions, something that would be guaranteed to throw you off if attempted with a regular quick-shift system.

The system could also have a place in the long-term future of bike racing, notably if GP bikes were ever to become fully electric. Zeroshift has put considerable work into developing the system for use with electric vehicles and hybrids.

To date, nearly all electric vehicles have used a direct drive system, but the space limitations within a motorcycle frame limit factors such as motor and battery size. It is here that a transmission system can give an electric motorcycle greater flexibility by better using the motor's torque. While the days of all-electric racing are probably still a long way off, it is solutions like this that will help make it truly feasible.

Written by Lawrence Butcher

Of these the key development has been the appearance of semi-automated shift systems that do not rely on the driver for accuracy. A driver can never be 100% consistent in the way that they select a gear, and this places incredible strain on the teeth of dog engagement gears. In the 1980s, Porsche tried to address this problem - with, it has to be said, considerable success - through the use of synchromesh engagement on its sportscar gearboxes. But while this prolonged the life of the gear sets, it also increased the shift duration which, if added up over a 24-hour race distance, accounted for a considerable quantity of lost time.

It was not until the advent of pneumatic shift systems in the late 1990s that truly reliable operation of dog engagement gearboxes was possible, with the advances in sensor and computing technology making their operation practical. As with many things in motorsport, pneumatic shifters originated in Formula One, with McLaren experimenting with systems in the early '90s. However, these systems relied on high levels of air pressure, in the region of 100 bar, which not only made sealing and reliable operation an issue but meant consuming a sizeable quantity of power to generating these pressures. That was until Tyrell made a breakthrough by developing a system that operated at far more manageable pressures of 8-10 bar.

Most pneumatic shifters consist of four main components - an electronic control unit, a valve block, a shift actuator and hand controls for the driver. The most important component is the control unit, often referred to as the GCU. This interfaces with a number of gear position sensors within the transmission and with the car's ECU, and controls operations such as throttle blipping and actuation of the shift mechanism. The relatively low operating pressure of the system means that sufficient air pressure can be maintained with only a small compressor, which is often housed either in the GCU or, in some cases, incorporated into the gearbox.

The final part of the jigsaw is the driver controls, which invariably consist of two paddles mounted on the steering wheel, one controlling upshifts and the other downshifts.

Electronic systems are also available that replace pneumatics with an electrically operated shift actuator. The control side of the system works in an identical fashion to a pneumatic unit, relying on a combination of input data from gearbox, chassis and engine sensors. However, the actual movement of the gear shifter is completed by a push/pull electrical solenoid, powered by a high-voltage supply from the system's controller. The system can also accommodate non fly-by-wire throttle systems, with a separate solenoid controlling the throttle shaft to initiate a 'blip' on downshifts.

The improvements in shift systems probably represent the biggest single factor in increasing the life of a transmission. The biggest risk to a transmission is incorrect meshing of the gear dogs, a problem that semi-automatic shifters almost eliminate. No system is infallible though, and this was highlighted during the 2011 race when the shift system on the Robertson Racing Ford GT, competing in the GTE Am category, failed. The incident proved an invaluable lesson in the benefits of having a back-up system, as fortunately the original, lever-operated sequential shift system had been left in place and some quick thinking by the pit crew saw the paddle shifter disconnected and the traditional set-up reinstated.

Fig. 1 - A pneumatic shift actuator installed in a Lola chassis (Courtesy of Geartronics)

Fig. 2 - The components that make up a pneumatic shift system - the GCU with internal compressor, shift actuator and valve block. (Credit: Megaline)

Written by Lawrence Butcher

In relation to transmission design, this has meant the removal of a number of performance-enhancing components, notably the use of centre differentials on the prop shaft and, in 2006, a ban on active diffs. These regulations give designers far less scope for development, and have resulted in some fundamental changes in gearbox design.

In the previous generation of WRC cars, gearbox layout and location was an area where considerable performance could be gleaned. Thanks to the nature of most of the base cars used in WRC competition, a transverse engine and gearbox is the almost universal starting point. In the past, designers found benefit in rotating what was originally a transverse gearbox through 90º, when converting them to four-wheel-drive specification.

This put the gearbox and centre differential of the rally version's four-wheel-drive system in line with the car's longitudinal axis, which in turn simplified the drive to the rear differential. This approach also moved the mass of the gearbox further aft, a desirable trait for improve handling balance, and was a tactic adopted by most WRC contenders in the early 2000s.

However, this relocation brought the added complication of creating a transverse transfer case running across the back of the engine, adding both weight and packaging issues, for what was only a small improvement in handling balance. After some experimentation it was discovered that similar results could be achieved by using a regular transverse layout, a format adopted by all the current manufactures.

The introduction of the smaller, 1.6 litre forced-induction engines has also had a knock-on effect on gearbox design. The wide power band of the previous generation of 2.0 litre turbocharged engines meant only five gears were deemed necessary, with manufacturers even experimenting with four speeds. With the reduction in torque brought about by the smaller engine, the latest generation of WRC cars are exclusively six-speed.

As of 2006, active front and rear differentials, which used electronic control units to vary locking characteristics, were no longer allowed in competition, and this was followed by a ban on active centre diffs in 2010. The system now allowed consists of front and rear mechanical differentials, with a manually operated centre clutch to disengage the rear prop shaft (for use when the driver operates the handbrake to initiate turn-in).

Rule 5.2b of the regulations also bans the use of viscous or hybrid viscous-plate differentials, stating: "Mechanical limited-slip differential means any system which works purely mechanically, ie without the help of a hydraulic or electric system. A viscous clutch is not considered to be a mechanical system." The result of this ruling is that a car must be set up at a compromise for a rally distance, with differential settings capable of dealing with a wide range of surfaces.

Mechanical differentials are still adjustable by the drivers, but only to a limited extent. On most units the pre-load can be loosened or tightened manually between stages. But any more meaningful adjustments can only be made in the Service Park. There are rumours, however, that some teams have found ways of exploiting the central clutch to vary power distribution from front to rear, but no official comment has been forthcoming.

Overall, the regulatory changes introduced over the past five years have greatly simplified WRC transmission (and overall car) design, bringing it closer to that of other rallying classes. While there is some debate as to whether this has brought about real cost savings, it will undoubtedly pave the way to international competition for teams graduating from national level Super 2000 competition.

Fig. 1 - A current-generation six-speed sequential WRC gearbox, featuring an externally adjustable mechanical diff (Courtesy of Xtrac)

Written by Lawrence Butcher

So how do we engineer a composite to have the necessary mechanical thermal and frictional properties required for a clutch plate material?

Carbon-carbon composites are the answer. Carbon exists in a number of different forms and the term 'carbon/carbon' relates to both the reinforcement and the matrix material - both being carbon, although different forms of it. This composite consists of a fibrous carbon material within a graphite matrix. As with fibre-reinforced polymer composites, directional strength can be determined by selecting the direction of the fibres. Because of the multi-directional loading to which the clutch plate is subjected, a randomly orientated fibre is used.

The material combines the stiffness and strength of polymer-reinforced composites with the excellent thermal properties of the graphite material. Because of this, the carbon fibre-reinforced graphite composite exhibits a low coefficient of thermal expansion (and therefore low deformation) and the excellent frictional properties required of a clutch plate material. And the relatively low density of the material means it also exhibits an extremely low inertial weight. In addition, depending on the frictional and wear rate properties required, the porosity can be controlled by the composite constituent properties and the manufacturing process.

Unlike glass- and carbon-reinforced polymers, the manufacturing process is slightly more complex. Carbon-carbon composites are made using various methods, to extract the graphite from a natural carbon product through carbonisation (charring). In the early development of carbon-carbon technology, Chemical Vapour Infiltration (CVI) and Liquid Phase Infiltration (LPI) techniques were used to fabricate the composites; in some cases they still are. The CVI process involves the infiltration and pyrolysis (thermochemical decomposition in the absence of oxygen) of a vapour that deposits carbides on the fibrous carbon surface. Further heat treatment forms the crystalline graphite from the deposited carbon.

LPI is a similar process that involves the infiltration of the composite with a carbon-rich resin or similar; the pyrolysis process then extracts the carbon from the resin, and further heat treatment or carburisation forms the graphite from the carbon. This process is then repeated to achieve the required density and porosity of the composite.

Because of this lengthy and complex process, the Across Corporation has patented a process that uses a pre-formed carbon fibre yarn that aims to cut the number of processes required in the manufacture of carbon-carbon composites - the CVI process consumes large amounts of energy - in order to reduce cost and throughput times, and increase quality.

Typically exhibiting much greater longevity than that of sintered clutches, combined with the increased durability at elevated temperatures and an extremely low inertial weight, it is easy to see the importance of this aerospace-derived material in performance automotive applications.

Fig. 1 - Carbonetic Paddle clutch exhibiting the carbon-carbon friction material

Written by Chris Thwaites

Further to applications of surface coatings in other components, geartrain components are ones that can really use the benefits of DLC coatings, to give improved surface hardness and lower frictional losses. Ball-on-disc tests show that some DLC coatings give coefficient of friction values of as little as 0.05. Reports suggest that, in some cases, geartrain efficiency has been improved by up to 20%, owing to the lower frictional losses thanks to the coating's improved lubrication. The coating's properties enable this improvement by reducing the inclination of the lubricant adhering to the gear material.

These improvements in lubrication have helped to lower gearbox temperatures. There are improvements in other areas as well, including the resistance to scuffing, particularly during running-in periods. Further gains could be made within geartrain products due to additional use of such coatings to facilitate improved strength, allowing the use of lighter materials that, for example, aim to decrease rotational inertia.

Although there are a number of alternative coating processes, including titanium and chromium-based finishes, few offer the high hardness values or covering properties of a DLC coating. While exhibiting increased hardness properties over the substrate material, the coating also demonstrates a higher Young's modulus resulting in prohibitive compressive surface stresses, which in some cases can limit the overall film thickness applied to the material. Because of this it is common for surface treatments to be applied in a multi-stage process.

Even though material properties are improved, the coated components' surface finish is only as good as the surface to which it is applied. The component still needs the necessary finishing to the required specification for the coating to achieve the properties of which it is capable. Even though there is the benefit of improved surface hardness through the DLC process, the substrate material will still need heat treatment to ensure the correct depth of hardness values and to guarantee operational performance.

As can be seen from the applications of high-hardness surface coating treatments in transmission components, there are clearly performance benefits to be had at a cost - but there are also a number of design and manufacturing considerations to take into account as well.

Fig. 1 - Today DLC is used on all manner of geared components (Photo courtesy of Sulzer)

Written by Chris Thwaites

In motorcycle clutch units there are two main methodologies of assuring the pressure plate compresses the clutch plates when the clutch lever is de-pressed, namely using either a series of coil springs or sets of diaphragm-type springs. But which is more suitable? Depending on the manufacturer, OEM and aftermarket clutches can take the form of either type in order to satisfy a variety of criteria including the feel of the clutch, compatibility, adjustability, manufacturability and durability. While coil spring units find their home in most British Superbike teams, many World Superbike teams have favoured the diaphragm spring alternative. From both a manufacturer and customer view point, there are a number of areas to consider.

Within the coil sprung unit, depending on the number of springs used in the clutch unit, changing and varying the spring rates among the set allows an almost infinite degree of adjustability, and enables the user to determine the amount of engine braking or slip required from the clutch, as well as allowing the capability to maintain consistent performance as the clutch pack wears. In addition, some have reported that the coil spring can give a more linear clutch lever feel due to the nature of the spring rate as it is compressed.

On the dirt bike scene, particularly in Enduro events, many riders of the 450 cc four-strokes - chiefly the Honda - have found extra longevity and resistance to clutch fade by replacing the standard four-coil spring unit with an aftermarket six-spring unit, due to a more uniform loading of the clutch plates. Further reduction in fade and greater resistance to stalling have been found by replacing the OEM unit with a Back Torque Limiter, not only through the benefits afforded by the device but by providing additional oil flow through the clutch unit and consequently more effective cooling.

The diaphragm spring units, specifically the STM unit, typically contains two diaphragm-type springs, by replacing the secondary spring to vary the level of engine braking, and the primary spring allows the level of slippage under acceleration to be maintained - for instance, if power output was increased, the spring tension could also be increased to help maintain clutch engagement under hard acceleration.

From a manufacturing point of view, providing a multitude of spring rates could cause headaches; forming ribs within the springs' 'fingers' is one way to increase clamping force, though this method alone would require significant and costly tooling requirements. Diaphragm springs have enabled clutch designs to become more compact and maintain a light clutch actuation, not only within the motorcycle industry but across the automotive industry. Also, because of the nature of the spring design, fatigue failure could be problematic but spring design, along with metallurgy advances, have improved performance. Although we have yet to delve further into cost or supplier loyalty, there is clearly a lot of food for thought.

Fig. 1 - A Sigma performance 'Slipper clutch' coil spring unit for a Honda CRF 450, showing the clutch centre and the amount of rotational movement available

Written by Chris Thwaites

The Oulton Broads and surrounding areas, in Suffolk,

England, have hosted hydroplane racing for many years. These high-speed vessels can reach speeds in excess of 50 knots (about 57 mph) using surface-piercing technology, most of them running two-stroke engines from manufacturers such as VRP, Rossi and Konig, the latter found in GP motorcycles ridden in the 1960s and '70s. Czech manufacturer Konny reproduces components for these classic Konig engines and transmissions. The technology has been used in high-speed applications for many years and, more recently, in Mercury Racing's M8 drive package where it aims to exploit the full potential of the 1350 hp 9 litre V8.

Conventional screw propellers are designed to operate fully immersed below the water's surface, but the nature of propeller design - and in fact any pump operating in water - can cause it to become exposed to cavitation conditions. The surface-piercing propeller drive offers reduced appendage drag over traditional stern drives and outboards, with the (possibly most critical) advantage that they do not suffer from the problems associated with cavitation. This alternative solution functions with the propeller semi-submerged, operating aerated, meaning it is not exposed to cavitation. You will instantly recognise a surface-piercing drive by its rooster tail and the separated sheets of wake caused by the propeller repeatedly entering and exiting the water's surface.

Unlike conventional propellers, surface-piercing drives operate with their low-pressure face completely aerated - there are a number of different phases of operation before the fully ventilated regime - that is, the face of the blade is completely covered with an air cavity. These propellers therefore rely on creating large amounts of lift from the positive pressure face due to loss of thrust on the suction side.

Numerous profiles have been explored to create maximum thrust in this ventilated condition. According to one theory, the most efficient foil shape offering the greatest lift:drag ratio has all its lift concentrated at its trailing edge. If you look closely at the trailing edge of some high-speed propellers you will notice heavy cupping, much like a Gurney Flap, which improves thrust by increasing the effective pitch of the propeller, and therefore supports this theory. Also, the distinctive trailing edge offers an attachment area for the air cavity, whilst helping to maintain blade strength and maximise blade area.

The propellers found on these high-speed vessels are easily recognised by their square trailing edge and thin leading edge, usually constructed from high-strength, corrosion- and fatigue-resistant stainless steels or nickel aluminium bronzes. Surface-piercing propellers, in particular, use these exotic materials due to the complex nature of the forces acting on a semi-submerged propeller. In addition to this, high-strength materials facilitate a smaller hub diameter, therefore reducing drag; they also permit the leading edge to be as thin as possible.

Having spoken to David Jones, who runs a 500 cc two-stroke Konig engine and surface-piercing drive in his classic hydroplane for the Area 31 Powerboat racing team based on the Oulton Broads, it seems that drivers have their own opinion and preference - drivers favour different transom heights, propeller diameters, pitches and the number of blades run on their raceboat. Area 31 runs a two-bladed propeller, while competitors run three-, four- and five-bladed items that are often hand-modified, offering different characteristics. It has also been found that excessive propeller diameter has caused chine walking on his raceboat.

Konny offers a number of different gear ratios for these engines, from 11:15 to 14:14 (1:1) depending on the course, propeller choice and running engine speeds of about 11,000-15'000 rpm, so propeller speeds are high. At these high speeds, therefore, cavitation would be unavoidable when operating fully submerged, yet propulsion with merely a quarter of the blade remains the most efficient method.

Fig. 1 - David Jones' A31 runs a three-bladed propeller

Fig. 2 - Mercury Racing's M8 drive exploits surface-piercing technology

Written by Chris Thwaites

One particularly enterprising car constructor has developed, and very successfully used, its own gear-drive conversion in place of the existing chain drive. I recently spoke to Del Quigley of DJ Racecars about this transmission and what led him to produce it initially. Quigley explained that he originally used a chain drive on his early Hayabusa-engined cars but that the output of the engine was "destroying chains rather quickly", referring to the 530-section chain.

There is a buoyant market for tuned bike engines in car racing, and the engine specification that was breaking so many chains was producing 220 bhp. The chain run on this chassis, at 220 mm between sprocket centres, was much shorter than in the bike, and this could have been a contributing factor in the shortened life of the chain. Short chains are suspected of having shorter life for a number of reasons, two of which are the shortened time between loads on any given roller, and the smaller angle of chain 'wrap' around the driving sprocket, with fewer teeth consequently carrying the load.

So the company set about designing and making a bespoke gear drive for its cars, and the first one was run in 2003. DJ Racecars, which is based in Derbyshire, England, manufactures all the machined parts except the gears.

The gear drive is more than capable of handling the torque of the Hayabusa engine, as Quigley explained, "The gear drive came into its own as we have very successfully used it on two supercharged cars, one of which produced 353 bhp at the rear wheels." This output is more than double that of the original Hayabusa engine, for which Suzuki claimed 175 hp before the 2008 model upgrade.

As can be seen from the accompanying photos, the casings are machined from billet aluminium. Although better known for composite chassis and wing manufacture, the company has used its existing CNC machinery to produce some well-machined lightweight casings and other components.

The method by which the gear ratio is changed is by selecting pairs of gears that fit behind an easily accessed cover on the left-hand side of the car. One of these gears fits in place of the original sprocket, while another fits by using interchangeable bearing housings. The hardware for each ratio therefore comprises two gears and two housings, complete with bearings.

In those series that the company is most heavily involved in, hillclimbing and sprinting, there is no requirement for a reversing mechanism. The gear train drives the rear driveshafts via a Quaife differential, often referred to as the 'Radical diff' after the racecar manufacturer of the same name.

Thus far the gear drive has been fitted to five bike-engined cars and has, according to Quigley, been "very reliable".

Fig. 1 - The gear-drive transmission shown installed in a hillclimb chassis (Courtesy of DJ Racecars)

Fig. 2 - This photo shows the gears that are changed in order to vary the drive ratio (Courtesy of DJ Racecars)

Written by Wayne Ward

In Britain, besides sportscars such as the Radical, there are a number of manufacturers providing cars powered by bike engines to compete in circuit racing, sprinting and hillclimbing. Often the bike-engined cars can punch 'above their weight' and achieve good results in a straight fight with much more powerful opposition.

Bike engines are relatively cheap and very light. The low mass is important in these lightweight, nimble cars. Moreover, there's a good number of very powerful stock engines from 600 cc to 1400 cc, and conversion kits to take them to even greater capacity. The Suzuki Hayabusa engine is a popular choice for increased capacity, and a number of them in supercharged and turbocharged specs race in the UK.

There are two main problems, however, with using a bike engine for car use. First, it has no reverse gear, and in many cases the regulations require that the car is able to move in reverse under its own power. Second, there is no differential, so there is a ready market for limited-slip differentials with a reversing capability, and a number of manufacturers currently service this requirement.

I recently spoke to Elite Racing Transmissions from Stoke-on-Trent in England about its MX200 reversing differential, which is made specifically for bike-engined cars. The MX200 differential is designed for bike engines mounted longitudinally, such that the input shaft runs parallel to the direction of car travel, and this means it is suitable for front-engined and rear-engined cars alike.

The unit features a Salisbury-type limited-slip differential - a system based effectively on a friction clutch that limits the slip between the two output shafts as their speeds differ - and a particular design feature is the system of easy-access drop gears that allow "a rapid change of the overall final drive ratio" according to Elite's sales manager Richard Homer. For this particular system the range of final drive ratios can be varied between 2.67:1 and 8.75:1.

On that point, there is a useful resource on the Spire Owner's club website (www.spireoc.co.uk), which has a handy calculator for drive ratios for bike-engined cars. There are a number of companies manufacturing small, agile cars for hillclimbing and sprinting, such as ADR Engineering, which recommends and supplies this differential with a each complete car or kit.

This particular differential from Elite weighs in at 32kg, and is designed not only to cope with the extra power afforded by engines with increased capacity, but also with the often substantial increases in power and torque associated with turbo and supercharged applications, with Homer saying it is "designed for turbo Hayabusa power".

Fig. 1 - This limited-slip differential is especially designed for bike-engined applications, incorporating a reverse gear (Courtesy of Elite Racing Transmissions)

Written by Wayne Ward

Anyone who follows Formula One closely will remember the dominant position that Williams had at that time. The CVT's development happened during a golden era for Williams; having the best car, they attracted the best drivers and for years were considered 'kingmakers' owing to the fact that, more often than not, the world champions drove a Williams.

The nature of the relationship between Williams and Van Doorn was one of exclusivity. Had the CVT not been banned, the route to development of a successful system for other teams would have been made harder by this deal.

The FIA introduced a rule mandating that Formula One transmissions should have "a minimum of 2 and a maximum of 7 discrete gear ratios", thus outlawing the CVT system. The fact that the car test had been a success escaped to the outside world; the rule was introduced suddenly and within two weeks of that first car test. But there were few protests from within the team when the rule was introduced.

So what would the implications have been for the hard-working engineers at Renault Sport's Viry-Chatillon base and the engineers at Williams, specific to the use of this transmission for racing? As mentioned in the previous article, they would have needed to re-optimise the engine for a single speed and undertake much durability work in order to make the engine survive such a high duty cycle. At the time, the maximum output was at a lower speed than that at which the Formula One engines run today, but it would have been a very big task.

Boulanger says, "In a race mode we would have had to design a new specific engine further optimised around peak power that could last 90 minutes at 15,000 rpm continuous." There was also the matter of the size and mass of the new transmission, of which Boulanger remarks, "Yes, it was quite heavy and big (in volume), which is something Patrick Head and Adrian Newey did not like!"

We can easily understand that heat rejection would be much higher, given the increased average engine output, so there would have been a marked increase in the size of the radiators, and there would also therefore have been an aerodynamic penalty to this extra cooling requirement. This would have had its own weight penalty with larger sidepods, larger radiators and more fluids being carried on the car.

We can see that, beyond the initial successful test where the concept was proven to work and to be very effective, there would have been a tremendous amount of work to do in order to make any car using such a system race-worthy. In fact, had the FIA not banned the system - and had Williams and Renault carried out the required work - their position of dominance would have been absolutely unassailable, and it could have changed Formula One for a long time.

Fig. 1 - The Williams CVT Formula One car. Will we ever see its like again?

Written by Wayne Ward

Mated to the current Renault V10, the CVT allowed the engine to run at a constant speed where power output was greatest, and for the transmission ratio to be controlled to provide the optimum tractive effort. Arnaud Boulanger, who now holds a senior role at the Renault Formula One team, was involved in the project at the time and gave us his insight.

If you watch the video footage of the test, you will notice that the noise of the car is almost constant. "Most people at the time thought it would spoil or destroy the show to the public because one could not relate noise to speed," says Boulanger. "I honestly believe the fans would have hated it".

The only variation in sound comes when the car drives from its 12,500 rpm 'idle' speed corresponding to peak torque, used for launch, to its constant 15,000 rpm running speed (peak power engine speed in 1993). The potential advantage of the CVT was described as "massive" - the average power around a given lap would have been a vast increase over a car equipped with discrete ratios.

Little engine work was done for the brief testing, but it is certain that the engine would have had to have been changed to cope with the requirements of CVT.

The first of two obvious points is that, for CVT use, the engine would be optimised around a single speed rather than the convention of being tuned such that it is driveable over a wide range. The second important impact of CVT on the engine is that the duty cycle is very much greater, and therefore many of the components would need to be redesigned to withstand the rigours of being operated at constant high speed and load. There is little doubt that this could be achieved, but it would have taken some work at the time.

I asked Boulanger his opinion about the 'arms race' that CVT would have brought to Formula One. Would everyone have had to develop CVT, had it not been banned by the FIA? "Yes - not because the public would have liked it, but they would have been forced into it," he says.

Despite the fact that Boulanger believes that the public would have hated CVT in Formula One, such was its potential advantage that everyone would have been forced to follow because the penalty of not doing so would be so damaging. In the next issue of RET-Monitor we will look at some other aspects of this fascinating project.

Fig. 1 - A cutaway illustration of the CVT system

Written by Wayne Ward

When using a swing axle, this was usually achieved by using a ball-and-socket arrangement integrated into the differential output shafts, while fully independent and de Dion systems required some form of universal joint at each end of the driveshaft.

For many years the Hookes joint, usually made from Hardy Spicer components, was the preferred option, but this could not solve the whole problem in isolation. As the wheels move from full rebound to full bump, they move through an arc of varying radius that seldom coincides with the centre of articulation of the inboard joint. That means the instantaneous distance between this and the corresponding centre on the outboard joint will vary through the range of suspension movement.

On modern cars, which tend to have restricted movement and longer wishbones, this effect is much reduced (although still present) but on older cars with narrower track dimensions and several inches of movement the effect is significant - several tenths of an inch.

In effect, to complete the installation a driveshaft of varying length was required. There were a number of different solutions used, none of them completely satisfactory. The simplest, used on many early Lotus and also on a whole generation of lower formulae cars, was to fit a rubber 'donut' between the inner Hookes joint and the transmission output flange. This could 'soak up' a limited amount of variation and misalignment, but put severe shearing stresses though the rubber, and failure was not uncommon.

A more refined solution was to split the driveshaft into inboard and outboard section, with the inboard sliding axially within the outer on a coarse spline. This could handle the length variation but not always successfully under extreme torque loadings. In this scenario the spline would 'stick', and would in effect become an extra, and rigid, suspension member.

This would momentarily lock up the rear suspension, in much the same way as the spring damper unit would. In turn this would lead to a sudden loss of rear tyre grip and an inevitable 'moment' for the driver. Removing the driving force would then 'unlock' the spline, and movement and grip would be restored.

The elegant answer to this problem was to fit a series of ball bearings between the inner and outer splines so that relative movement between the two no longer depended on sliding friction, but on a rolling bearing. As an engineering solution this was generally successful but it was also heavier, bulkier and - perhaps most significantly - far more expensive, as a number of precision-machined components had been introduced into the assembly. BRM made its own, which were much admired; others modified roadcar units, notably those produced by Mercedes.

A clever solution that avoided the complication and cost of the sliding splines was sometimes used where the amount of axial float would allow it. This was to remove one of the hub bearing shoulders within the rear bearing assembly, and let the hub unit float axially to take up the length variation. A chartered engineer would probably have frowned on this but the racers of the day used it to good effect.

Alternatives to the Hookes joint really only became viable with the move to front-wheel drive roadcars in the 1960s and 1970s, and the investment in constant-velocity joint technology that this brought about, notably from the Lobro company. Before this, it had been necessary to source this type of joint from the Volkswagen parts bin, with variable quality and results, or from Porsche, at a much higher cost. Later still, the tripod joint, also roadcar-derived in many instances and which can generally operate more efficiently at higher angles, tended to be favoured.

Both types, constant velocity and tripod, deal with the length variation issue by allowing the balls or rollers to move along grooves within the housing, their cylindrical form allowing for angular misalignment of the shaft to the housing axis. Ironically the basic principle is almost identical to that used in the simple ball-and-socket type of joint used back in the 1930s.

Fig. 1 - A Hookes type universal joint with a sliding spline driveshaft. This F5000 lola also uses inboard rear disc brakes

Written by Peter Elleray

Most 'racers' are familiar with using the traditional straight line plot of road speed against engine rpm as a method of choosing ratios. Whilst this gives an indication of maximum speed in each gear and helps to assess the drop in engine rpm associated with changing up into the next ratio, its use is otherwise limited. In particular it does not

introduce any engine parameters into the equation, and so the estimate that we have of rpm drop is only as good as the driver's intuitive feel for how well the engine performs in this rpm range.

The tractive effort curves, which quantify applied motive force at the ground, include all of this information and more, and so they can tell us how well our ratios are suited to the engine's characteristics as well as the fundamental track parameters that have driven their choice.

We noted in Part 1 that engine power will rise to a peak before starting to fall again. This invariably occurs before we reach the engine's 'safe' rpm limit and the temptation is to make full use of this. But depending upon the rate of fall off in power after the peak is reached, this may not always be the right thing to do.

We can see this in the tractive effort curves. The torque curve from which they are derived is in itself related to the power curve by the equation P=n.T (P=power; n=rpm; T=Torque). If the power were to remain constant, an increase in rpm would see a fall in torque, so when power falls torque drops off more steeply. The tractive effort curve develops a negative slope which increases. We can see that if the engine is taken to maximum rpm we can reach a point at which we are making less tractive effort in that gear than if we had selected the next one. In other words we would be accelerating faster in the higher gear once past the point where the curves cross over.

If we also include the aerodynamic and mechanical drag force curves on the same diagram then the point at which the tractive effort curve(s) cross will represent our theoretical top speed. This does not have to be in our highest ratio gear, and in an 'overdrive' transmission it will not be.

Although we tend to think of engine performance in terms of 'power output' it is much easier to analyse it 'on track' using these torque derived curves. However, we can also plot curves of power in each gear against road speed. A set of these are shown in the lower corner of the ratio chart. We can clearly see the effect of the fall in power after peak power rpm is reached and how the power developed at the same road speed in consecutive gears can overlap.

The area which lies below the tractive effort curves in our graph is the product of Forces and Velocities and is therefore measured in units of power. The key to optimising the power of the engine throughout the speed range of the track in question is to maximise this area. If we reach maximum rpm in a gear before crossing the next T.E. curve then we must add a vertical boundary to the set of curves which will intersect this curve. If we cross over the tractive effort curve for the next gear by remaining in gear until maximum rpm, then we are cutting out a small, roughly triangular shaped section of the available area every time that we do so, which represents wasted power.

We can see that we can draw a curve that will run from top left top to bottom right on the graph and which is tangential to each tractive effort curve. This is the curve we would obtain if we had an infinite number of gear ratios, instead of the discreet number shown. In other words it represents the power exploited if we had a continuously variable transmission system (CVT). The equation of this curve would be P=k.V.

The use of these curves is not limited to the choice of gearing. They can be useful in mapping an engine, and in many ways are a key element in bringing together the sometimes conflicting requirements of the engine designer and the chassis designer.

Fig. 1 - Comparison between conventional ratio chart and power/speed based chart

Written by Peter Elleray

Why should this be so, and how do we then determine what we need?

Essentially we seek to transform rotational motion and energy at the engine flywheel into linear motion and kinetic energy at the vehicle tyre contact patch. To do so we require a mechanical connection between the two. Consider something as simple as a slot car. This does not employ a gearbox, but a single pair of gears in mesh, one on the motor shaft and the other on the solid axle to which the driven wheels are rigidly attached. Depending upon the characteristics of the electric motor there will be a variation on the number of teeth on each gear - the gear ratio - but this will be fixed. Speed control is achieved completely by the hand held rheostat operating on the electric motor. We often call this hand controller the 'throttle' although the electric motor does not breath air to produce power as an internal combustion engine does.

In theory and with a sufficiently powerful motor there is no reason why this type of control cannot be used on a full size electric vehicle. Multiple ratio gearboxes are only required where the torque and power characteristics of the motor unit are such that, if geared directly through a fixed ratio to the road wheels, it cannot provide sufficient torque throughout its speed range to impart the required in line tractive effort throughout the vehicles operational speed range.

With an internal combustion engine this is almost universally the case. It's characteristic power curve rises to a peak from zero before falling sharply away again, which imparts a 'torque curve' that will also eventually fall with increasing speed . The phrase 'running out of steam' comes from an earlier era and a different type of powerplant with different characteristics again, but is still used to describe an internal combustion engine that can no longer sustain acceleration, or cope with an incline.

We use gearing to obtain twin objectives. On the one hand, it is intuitive that if we 'gear down' from the motor at a ratio of 2:1, then the geared axle will revolve at half engine speed. Our road speed will then depend upon the diameter of the driven tyres. At the same time, we will have increased the torque acting through the driven axle by a ratio of 2:1 when compared with the torque at the crankshaft of the engine. Conversely, when we 'gear up' at 2:1 we double speed and halve torque in comparison to the motor. On a typical car, we seldom find ourselves 'gearing up' except in the highest gear, which in a road car is referred to as an 'overdrive', and, in any case we put all of the torque through the gear train itself through one 'final drive' which reduces speed at something in the order of 3:1 and increases the torque to the driveshafts by 3:1.

In effect then, we are gearing down in all ratios between crankshaft and road wheel. Our lower gears, with the highest numerical ratio, give us the highest torque multiplication so that they can provide high tractive effort from a standing start, or to get out of a slow corner, but the vehicle will be speed limited in them. The mid range gears provide a balance between tractive effort and speed, whilst at the top of the range, we must sacrifice force for velocity.
To visualise this process and the estimate what ratios we might require, we plot the tractive effort in each gear against the road speed. Tractive effort is obtained by multiplying engine torque and the overall gear ratio (gear and final drive), and dividing by tyre radius. Road speed in each gear is obtained from engine rotational speed using the same factors, and so in effect we obtain scaled versions of the engine torque curve in each gear, plotted against the road speed range that we obtain through the engine rev range in that gear. In the lower gears these will appear to be scaled up along the vertical axis (which represents Tractive Effort) and down along the horizontal, or speed axis. In effect they looked 'squashed' horizontally and stretched vertically. In the higher gears the opposite occurs. A typical series of tractive effort curves is shown in the accompanying illustration. This type of graph, when used with a conventional gear ratio chart, can tell us a great deal about performance, and this is something we will develop next time.

Fig. 1 - Traction effort graph.

Written by Peter Elleray

Putting some numbers to this, consider a Formula Three car whose engine is dry sumped and running a 140 mm diameter race clutch on a lightweight flywheel. It will have a crankshaft axis that is less than 100 mm above the bottom of the engine. Running with Avon slicks at the rear, the drive flange on the rear hub will be approximately 285 mm above the contact patch. If the car is running at 30 mm rear ride height, then that gives a 155 mm step up height between the crankshaft and the final drive output flange. A conventional, in line, longitudinal transaxle, with drive entering through the lower shaft and exiting through the upper will typically have shaft centres of just over 3 inches - around 78 mm. The rest is taken by driveshaft angularity. This can be reduced by making the offset between the differential and output flanges as small as possible, and the rear track as wide as possible.

On a relatively low powered race car, and within modern machining techniques employed in finishing the tripod joints, this is something that you can 'get away with', but on higher output race cars this is usually not the case.

Designers have therefore consistently looked beyond the simple twin shaft transaxle in an effort to reduce the size of the offset - and the problem. One solution is to introduce a step up gear into the assembly, either in initial design or as an add on. Depending on the requirements of the car, this gear can be positioned at several different points along the drive line.
Positioning it immediately after the clutch will lift the complete gearbox assembly aft of this by a proportion of the step up gear centres, the exact figure depending on whether the gears are arranged to sit above one another. This might benefit the ground effect tunnel area on a full ground effect car by reducing the size of the 'bump' needed to clear the diffuser roof, but it will also raise the transmission CoG, therefore negating to some degree the benefit of a shallow sump and small diameter flywheel. The input shaft would also be very short and therefore liable to weaken under shock loading.

The alternative, of placing the step up either just in front of, or maybe behind the final drive, will produce the exact opposite set of gains and losses. As ever, there is no one correct answer in race car design, everything is a compromise.

Another solution, which saves the additional weight of the step up gears, their bearings and so on, is to employ a hypoid bevel final drive. On paper this always looked quite attractive, and was used in Indycars quite extensively in the 1980s, where the rear tyres were 28 inches diameter and the conventional Hewland DG gearbox featured only a 3 ½ inch shaft centre. The downside was the requirement for a more sophisticated lubrication system and the special oils needed for the hypoid. These gearboxes fell out of favour after a certain amount of unreliability - they were also quite expensive to manufacture. As an aside, it was also possible for the drawing office to produce a single speed with 5 reverse gearbox by getting the hypoid on the wrong side of the crown wheel, but this could not be corrected by flipping it over as on a conventional pinion gear. Not that this ever happened of course….

A more pragmatic approach, which again has a certain elegance on paper, is to tilt the complete engine and transmission assembly up about the nose of the crankshaft - typically between 1 and 3 degrees of tilt would be used. The Arrows A2 Formula One car employed this, primarily to clear the ground effect tunnels, and it was common in Indycars as well. Again, this lifts the whole CoG of the transmission and engine again, and in that respect is slightly illogical.

Perhaps the best answer comes with the use of a transverse box, for then the gear trains can be so arranged that the final drive comes out more or less where you require it - the step up gear is necessary in any case to turn the drive through 90 degrees.

Fig. 1 - An early hypoid final drive in the Lotus 12.

Written by Peter Elleray

Starting with the basic operation of a conventional ‘push’ clutch, the ‘push’ that is delivered by a slave cylinder refers to the force that is required to disengage the clutch plates. Upon depressing the clutch pedal, hydraulic fluid is forced into the cylinder which is then thrust out towards the engine to accommodate the extra fluid, taking the release bearing, or bobbin, which is mounted directly onto its nose, with it. This is already in contact with the inner tips of the clutch fingers, but exerts no force on them when the clutch is engaged. We saw last month that the fingers act as radial leaf springs, and that they are preloaded on assembly to obtain the clamp force. They pivot radially on pillars mounted to the cover roughly half way between the input shaft axis and clutch cover outer edge, their outer tips in contact with the plates and applying the clamp load axially. The thrust from the release bearing now applies a bending force to the finger, which is sufficient to overcome the preload within them and therefore the clamping load and so the clutch is disengaged. Notice that the internal forces react in such a way as to push the bellhousing, to which the slave cylinder is rigidly mounted, off the rear face of the engine block when disengaging the clutch.

The pull clutch operates in the reverse manner. The slave cylinder and bobbin are in contact with the inner tips of the clutch fingers, as before, but the assembly is now mounted to a spider that is itself mounted onto the back of the engine block via pillars which just clear the perimeter of the ring gear. When the pedal is pushed in, fluid enters through the spider into an internal chamber arranged so that the slave cylinder now moves away from the clutch, i.e. in the opposite direction, pulling the fingers with it by means of a ring turned into the end of the cylinder body. The fingers are now pivoted on the cover just inboard of its periphery, i.e. outboard of the set of bolts or studs that bolt the clutch assembly to the flywheel. This achieves two things. Firstly, with the pivot outboard of the point of contact between the finger and plate, it compensates for the cylinder moving in the ‘wrong direction’, and ensures that when it does, the force applied still acts in the correct direction to release the preload in the fingers and hence the clamp load in the plates. Secondly, and fundamental to the design, the mechanical advantage obtained by the finger is significantly increased as the effective length of the lever arm now spans the radial cross section of the cover plate. Note also that whereas the conventional ‘push’ design tried to separate the transaxle from the engine when declutching, the pull type accounts for the reaction forces within the assembly of spider to engine block.

Together these factors can then be used to improve feel, or reduce pedal load, or increase clutch capacity without increasing pedal load, or travel. The fulcrum design is simplified, and wear reduced so that the clutch will remain more consistent for longer. The design is also lighter.

As far as installing a pull clutch into an engine installation designed for a push clutch, it may not be so simple. Some modern race engines have the mounting for the release bearing spider integrated onto the rear face of the block, others do not. An adapter plate is then required to provide them. The original slave cylinder mounting, which will have been machined into the front face of the bellhousing or transaxle will then be redundant, but room will be required to accommodate the spider itself, which is unlikely to have been accounted for in the original design as it would have been wasted space. Finally, in addition to the clutch unit itself, the slave cylinder and bearing will be bespoke to the clutch assembly, and new routes will have to be found for the clutch feed and bleed lines.

Written by Peter Elleray.

High School physics tells us that the frictional force (F) created between the clutch plates will be equal to the Normal, or Clamp Force within the assembly (N) multiplied by the Coefficient of Friction (µ) of the plate materials; F = µ x N. We can visualise the friction created as a series of forces that act tangentially from the inner diameter of the plates to the outer. The Torque then carried will be equal to the sum of these Radial Forces multiplied by the distance of each one from the input shaft axis. This will in turn be equal to the average Radial Force obtained (R) x the mean radius of the plates (r); T = R x r.

Since the average Radial Force is equal to the product of the average coefficient of Friction (µ) and the Clamp Load (R= µ x N) , finally we obtain

Torque Carried (T) = µ x N x r.

In the past 20 years or so, the trend has been to reduce the diameter of the clutch assembly in order to lower the complete engine and transmission system within the race car. At the same time the inertia of the clutch assembly was itself reduced which in turn could be used to speed up gear changes and enhance engine response. Outside of Formula One, where it is even lower, a typical race clutch diameter might now be as low as 140mm, whereas once184mm (7 1/4”) was common. In turn this has led to a reduction in the diameter and working area of the clutch plates and hence ‘r’ in the equation above. At the same time the Torque carried has gradually risen, which means that either µ or N must rise in turn in order to maintain drive.

There are practical limitations on both of these parameters which are a function of the friction material technology and the capacity of the hydraulic system that is used to release the Clamp Force when required. A typical master cylinder and slave cylinder might have an area ratio of 4:1, which means that the force applied at the release bearing will be 4 times that at the master cylinder pushrod. In turn, a pedal ratio of 3:1 will give an overall mechanical advantage of 12:1. A typical release load of 450daN therefore means that 38kg will be required at the pedal, with 650 psi pressure in the system. This gives a reserve of about 50% on a typical system, which will be plumbed using a mix of dash 3 hard and flexible lines and fittings, and in practice it is wise not to reduce the reserve much beyond this.

As far as friction coefficient is concerned, one of the limitations of traditional materials was that this fell as temperature increased. With the introduction of Carbon as the friction material in the 1980’s this was eliminated and so the coefficient in service effectively raised.

Alternatively more plates can be added. A typical 140mm dia twin plate race clutch might be capable of transmitting 750Nm, whereas a triple plate increase this to 1100 Nm. In practice 4 plates are the limit. The efficiency of the clamp system is another area that can help increase effective capacity and this has led to the development of the ‘pull clutch’, which we will look at next month.

Written by Peter Elleray.

Central to its function has been the move from a H-gate to sequential gear selection, for with this it is much easier to provide powered control of the actual selection mechanism. In the modern box this normally takes the form of a drum, with a pathway machined into it, whose axis is parallel to that of the actual gear shafts. Depending on the type of system fitted to the car, as the driver pulls at his steering wheel paddles, an actuator will be operated either by pneumatic, hydraulic, or fully electric means, which will in turn move a peg or lever that engages in the track of the drum, and so rotates it. As it does so it will sequentially pick up conventional selector forks that will move face dogs into engagement with the next gear.

The historical ancestor of the concept – in the racing world – is the gearbox designed for the Cisitalia GP car for which Porsche was responsible, immediately post war. This was never raced, but in 1956 a similar system was adopted for the first single seater Lotus, the front engine type 12 Formula Two car. For his first tentative steps into single seater racing, Chapman boldly commissioned a bespoke unit that was typically somewhat in advance of anything else available ‘off the shelf’ at the time – of which there was precious little in any case. It was, in effect, a scaled up motorcycle unit, with five forward speeds occupying no more than 80 mm in length, and an all up weight of just 22,5 kg. Shaft centres were 99 mm with a further 30 mm step up available from a hypoid final drive. In its original form the manually operated unit was fully sequential. One set of gears was permanently splined to the pinion shaft and the other free to rotate about the first motion shaft. Rather than using a rotating drum, selection was by a sliding sleeve that sat over the first motion shaft. On this was formed a six sided spline which engaged with similar splines formed on the inner annulus of each gear. Bronze spacers between each gear formed a neutral position, with the cockpit lever always returning to the same, central, neutral position after down- or up-changes.

The ‘box was notoriously variable in its reliability, gaining the nome de plume of ‘Lotus Queerbox’ something that has been put down variously over the years to poor lubrication, undersized gears, and reuse of out of life parts in what was still an organisation that existed on the margins of viability. By 1961, when it was replaced by a ZF unit, it had been adapted for mid-engine use, the final drive had been changed to a spiral form after lubrication issues with the hypoid, and the change had been modified so that the lever now migrated as gears were selected, two racks being required in the system to achieve this.

But Chapman never completely gave up on his original transmission concept. In 1974, by which time Formula One Lotus’s were using the same dog clutch H-shift Hewland FG400 gearboxes as everyone else, he introduced a change system operated by a button on the gearlever, thus deleting the use of a conventional clutch pedal for all but starting. The idea was to speed up changes and reduce the time spent in freewheel, at the same time allowing the drivers to left foot brake. Even then this was recognised as a potentially faster way around the lap, but the system was driven by the starter motor with the conventional clutch release operated electrically from this, and instead of clutches gear shifting the end result was more often a series of burnt out motors and electrical short circuits that killed the engine.

Still he pursued his two pedal philosophy and in 1978 reintroduced a version of the ‘queerbox’. Instead of using splines, engagement was now by ball bearings thrust into position with each gear by a bobbin running inside the input shaft. Ground effect aerodynamics were just beginning to be applied to the type 78 and 79 and clutchless, two pedal control, by which pitch change under acceleration and deceleration could be more closely controlled, suddenly opened up a whole new relevance for the technology...

Written by Peter Elleray.

When the mid-engined car began to take over in the late 1950s, the transmission naturally remained at the rear, in unit with the final drive, but now a conventional front-engined road car gearbox no longer provided a decent starting point. The early Coopers instead used a road car based Citroen transaxle adapted from the front-wheel drive Light 15, whilst other transaxles of the era were based around the rear-wheel drive Renault and VW gearboxes – the Hewland. In rotating them about the final drive axis to sit behind the engine, the gear cluster now stuck out behind the rear wheel centreline.

It so happened that the other early examples of mid-engined Formula cars produced by BRM and shortly afterwards by Lotus used transaxles that were derived directly from their existing front-engined Grand Prix cars, and in both of these cases these units were in-line. The BRM already had its gears behind the final drive, so that in the front-engined car the driver did not have to sit above them, whilst the Lotus had been designed initially for an offset driveline and laydown engine installation. As a result, the BRM unit was almost a straight fit, whilst the Lotus, with its gears in front of the differential needed extensive redesign work. In this perhaps conservative rework, the gear clusters ended up in the same position as on the BRM.

Thus in 1959, when Cooper customer Rob Walker commissioned his own transaxle for a BRM engined Cooper special, ex-Maserati engineer Valerio Colotti designed a unit that was also in-line, with gears behind the diff, and so, fortuitously when Walker switched to a Lotus 18 the following year, the Colotti ‘box became a retro fit for it, and many Lotii to follow. Along with its derivatives this gearbox would go on to become available to all .

A bespoke 5-speed in-line gearbox followed from Cooper – the C5S – in 1960, and so by the early 1960s there was a small cottage industry in the UK and in Italy where you could go to buy a race car transaxle. By force of circumstance all of these units were in-line, all with gears behind the differential, and this established a pattern that would hold up for the next 25 years.

The few exceptions tended to originate from a manufacturer racing programme, for no specialist Formula One constructor of the time had the resources, either financial or otherwise, to develop their own transmissions. Gradually the Hewland began to predominate, although for many years Lotus stayed faithful to the ZF company’s products, chosen to replace its own design and so again in-line, with gears behind the diff. BRM redesigned their own gearbox several times, always maintaining the basic layout, but as new constructors came along, they naturally turned to the proven and available product, which was the Hewland FG and DG series.

Along the way, the Honda companies first Grand Prix car, which employed a transverse engine, also employed a transverse gearbox – it was to all intents and purposes a scaled up motorcycle powertrain, whilst both Alfa Romeo and Porsche successfully employed in-line units with the gear cluster ahead of the differential on the sports racing cars. The Alfa unit appeared briefly in Formula One in 1972 in the back of a March, before disappearing again, and it was not until 1975, and the appearance of the transverse 5-speed unit in the Ferrari 312T, that the pattern was broken. For many more years the Ferrari would be alone in using this layout, and when Lotus decided, a few years later to redesign their 1950s gearbox for current Formula One use, they maintained the in-line layout, with overhung gear shafts.

This gearbox was sequential, and so we can see that about 30 years ago, the set of ingredients that made up the modern race car ‘box – twin clutch units aside – had arrived but had not yet been collated together into one, homogeneous design.

Written by Peter Elleray.

Consider the conventional passenger or race car control system:

First, we have the right hand ‘throttle’ pedal – if you press with your foot you go faster, release your foot and you slow down – logical enough. The second pedal, well, if you want to slow down a lot faster, or in an emergency, press this, we’ll call it the Brake… Sooner or later you will need to deviate from a straight line, and so we provide a ‘wheel’ in front of your seating position... rotate it to the right and you go… right. And vice versa.

But now consider the remote lever, which goes up and then down through a gate, then side to side, up and down again, and furthermore needs to be in a different – although not necessarily the same – part of the gate depending on your road speed, engine revs, and type of manoeuvre... And furthermore, it needs a third pedal to control it before operation… Who dreamt that up?

Given that some form of multiple, intermediate gearing between engine and road wheels is necessary with an internal combustion engine, it’s hardly surprising that since the dawn of the motor car alternative forms of automating the change process have been pursued.

In the race car, the issue has been more one of the impact on this process on the cars performance rather than of difficulties in executing it – although for many years ‘transmission’ failure originating from the pilot’s desperate attempts to shift from one gear to another in the minimum possible time were a common cause of race car unreliability.

In the modern Formula One car we have arrived at a situation where the ‘twin clutch’ transmission is used. We have arrived at this via automated clutching and declutching of a traditional manual shift gearbox using driver actuated microswitches, a solution that future historians will regard as ‘interim’. Whilst these systems can provide a useful saving in the time the lag between applying tractive effort in one gear and then the next, they are as significant in the way that they have removed the ‘driver variable’ in the shift process, and hence increased reliability, particularly when their use spread to endurance racing.

The twin clutch system still operates a traditional manual box – but by providing two clutches to operate non sequential gears, it is possible to have the next gear ‘pre-selected’ with the second clutch almost instantaneously with the current gear being disengaged by the first. Naturally this process is operated by the driver via microswitch.

This is somewhat more intuitive than being presented with an H-pattern gear lever and foot operated clutch pedal, but still falls some way short of full automation.

The use of the torque converter on race cars has been limited over the years due to their weight, complexity, power consumption, torque consuming ‘slip’ and limiting effect on operator override. A notable exception was the unit developed by General Motors in the 1960’s and raced for several years on the Chaparral Can-Am and endurance race cars. It was no coincidence that they were used in conjunction with large capacity engines with a broad range of torque.

Ironically, to find a more appropriate predecessor for today’s twin clutch boxes we need to look back into the pre-War British racing scene, where it was common to see the ‘Wilson’ epicyclic pre-selector box used. The early ERA’s, various Alta’s a number of ‘Brooklands’ specials and others were so equipped. These boxes, although heavy and bulky in comparison even to the crude dog boxes of the day, were popular because they achieved the same end result as the modern Formula One ‘box. No clutch was fitted, but two brake bands used to provide ‘slip’, and allow for pre-selection via a conventional lever which was then actuated when required by stamping hard on a foot pedal in place of the traditional clutch. This was the ‘30s equivalent of a 21st century microswitch and processor unit, and allowed the driver to keep both hands on the wheel at the critical moment, which even then was recognised as a more stable and controllable driving technique.

So, in some respects, after decades of trying very successfully to perfect a crude and counter intuitive system we have come full circle. Another example of the past rewarding study?

Written by Peter Elleray.

Generally the loads that we will ask it to absorb are a significant proportion of, or even a multiple of all up vehicle weight, and they will act in a variety of directions and combinations.

First, if it is to do its job of providing the rearmost section of chassis, we ask it to take the same torsional loading as the chassis itself. We also ask it to accept lateral bending loads from cornering forces, and vertical bending loads from car weight, load transfer and aero forces and the magnitude of these moments will increase with the distance from the rear axle. This in part explains why most casings are nearly always much larger in cross sectional area towards the front than they need to be to house the gearbox itself.

Second, we almost always hang the majority of our rear suspension either directly from the casing, or from subframes mounted to it. We have to provide reinforcement locally where we choose to feed the loads in, and we will generally try to tie opposing sides together internally with webs and diaphragms to stiffen and distribute these high loadings through the case. These, in turn, often have to double as bearing supports for the gear shafts themselves.
Thirdly, we invariably cantilever the rear wing structure off the back, and sometimes bolt up the underbody tunnels or diffuser, once again feeding in high point loadings.

Finally, we often mount a pneumatic airjack to the casing, and so, when stationary in the pits, we ask it to support 60% or more of the car’s weight.

All of this in addition to the internal loads set up by the gears themselves, in the crown wheel and pinion, and by the clutch…

Amazingly, when one considers that the average wall thickness of a case can often be only four or five millimetres of magnesium, we almost always get away with all of this and still retain a transmission system inside that works for us – most of the time. We usually achieve this despite being obliged to make provision for sensibly sized access holes in order that we can open the box up and change the ratios in a hurry. And although we would like to make the complete unit from diff housing to engine in one unit, practicalities invariably force us to split the whole into front and rear sections.

As a final challenge, we almost always shroud the box in tight fitting bodywork, and pack the exhaust system in as close as we possibly can for good measure, so that the case itself runs at elevated temperature, to the detriment of the life of the gears within, and the stiffness of the case itself.

This stiffness, both in torsion and in bending, is every bit as vital in the transmission as in the chassis, or in a stressed engine. A structure is only ever as stiff as its most flexible component, and if the gearbox has a lower unit stiffness than the chassis – which it often has – then this will be the limiting factor in the key axle to axle stiffness figure. The relatively large cross sections that we can usually afford to design into our gearcasing at its forward end is a bonus when it comes to increasing stiffness, but by the time we reach the rear axle we invariably seek to reduce it to the practical minimum to house the final drive and differential assembly, because we are almost always looking to maximise the diffuser or tunnel area that activates the aerodynamic underfloor. This works against us and it is not uncommon to see a graph of torsional deflection, plotted against the wheelbase, that suddenly starts to climb in the vicinity of the diff housing. Sometimes this comes down to detail design, and it is always standard practice to dowel individual sections of the casing to one another, and to provide generous flanges and internal ribs.

Lack of Stiffness is not just the enemy of wheel loadings though, a flexible case can cause havoc with the gears themselves, which in extreme cases can slip out of mesh or engage more than one or a time.

All in all, when it works, it is a major engineering achievement!

Written by Peter Elleray.

Also, mechanical gear selection is very old hat and most single-seater drivers fully expect to see paddles behind the steering wheel to change ratios. But the actuators and their systems, whether they be hydraulic, pneumatic or electromechanical all take up extra space.

These were issues gearbox manufacturer Hewland was able to address when it started work on a revamped version of its mid-sized NMT [nineties medium transaxle] unit. Obviously the sensible solution is to incorporate these systems within the gearbox saving on size and weight. But they also contribute complexity and cost to an already price-sensitive component of a racecar. The big problem stems from the speed at which conventional oil pumps can run before they start to suffer cavitation and wear. Usually this is addressed using step down gears and an ancillary shaft to avoid overdriving the pump. But any such complication is unwelcome due to the pressure it puts on space and price.

Hewland's solution was to completely rethink the pump and dump the old three-gear design for a new concentric lobe unit that can run at higher speeds without cavitating. This could now be positioned on the end of the mainshaft running at gear speed. The pump is still a single-stage and the gearbox is still wet-sump although a double floor in the main casing baffles the oil, reducing surge.

Not content with this, Hewland then looked at how to drive an air compressor for the gearchange. Often this is dealt with by an electrically-driven compressor that draws energy from the battery and builds pressure in a reservoir. Hewland knew that removing yet one more ancilliary system from a racecar would be a welcome benefit for its customers. The solution was to extend the oil pump drive out of the back of the gearbox and straight into the centre of a compressor. The adoption of the lobe style pump made this possible and the in-house units were made larger specifically to allow a big enough drive to be extended through the centre.

Again this had gear speed considerations and careful design and specification of the unit was necessary to ensure it could work at all the revs required. In its first application, the new Formula Two spec car, the 'box would be operating at a mere 8250 rpm. However, on other applications, it can expect to run as fast a 12,000 or even 15,000 rpm. These demands led to Hewland again designing its own unit to fulfil that duty cycle reliably. The result is once more very compact and fully integrated into the overall unit. Pressure from this compressor actuates another refinement in the new design, single-acting shift actuators sitting inside two turrets either side of the top of the gearbox. This is not the first time Hewland has built a compressor into one of its gearboxes but it is the first time it has been achieved so simply and economically.

The position of both the oil and the compressor pumps has made them much more easily serviceable and self-contained. Even the oil filter has now been incorporated into the box with a paper element inside a transparent bowl that is both easily accessible and provides a fast visual check of the transmission oil level. A gauze filter for the feed to the pump is still incorporated to capture any larger pieces of debris before they enter the pump. The finer paper element has to be upstream of the pump as it creates a greater resistance to flow and pumps are better pushing than pulling.

Written by Charles Armstrong-Wilson.

A back torque limiter is a clutch that deliberately slips in a smooth and controlled manner as the rear wheel tries to turn over the engine. Most clutches are simple ramp style mechanisms, a one way ramp and ball mechanism under the centre of the clutch pushes the centre of the multiplate clutch up as the rear wheel tries to drive the engine through the transmission. As a result the clutch disengages as the ‘rising centre’ pushes the pressure plate of the top of the plate stack, disengaging the motor.As soon as the clutch disengages of course there is no source of power to drive the clutch centre up the ramp, and the centre starts to fall, allowing the clutch to re-engage and the process to start again. In practice, depending on the angle of the ramps and the spring pressures chosen the clutch finds an equilibrium that allows only partial engine braking to be transferred through.So why do motorcycles need this mechanism when the engine braking could be a valuable additional brake? It’s all about the way the chassis of a bike is allowed to pitch around to maximise grip in certain circumstances. Lets just consider a bike coming out of one corner and diving into another.. what happens when and why?The bike is leaned right over, the rider applies throttle and the resulting centrifugal force brings the bike upright. As power continues to increase it is better to have the largest possible contact patch on the rear tyre. The bike rolls back on its suspension and loads up the rear tyre, compressing the sidewalls and flattening out a larger than normal contact patch. In addition the rear drive chain is acting to try and extend the rear suspension by the nature of the relation ship between the front drive sprocket, the swingarm pivot point and the rear sprocket. This adds additional pressure on the contact patch and helps move some of the bike weight forward.As the bike enters the next corner the brakes are applied and the front suspension dives, transferring load to the front tyre, flattening that contact patch and allowing more aggressive braking. As the rear tyre sees load transferred the front however the same settings that pushed the rear tyre in to the ground under acceleration are now trying to lift the entire wheel off the ground. The transfer of load to the front, the pitching forward of the entire motorcycle and the desire of the rear wheel to hop as it tries to use whatever grip is left to turn the engine over compression means that any mechanism that can disengage the engine will add to the available grip and will greatly stabilize corner entry feel. One additional benefit is that the ‘slipper clutch’ acts to help reduce engine over-rev as well.One alternative solution is to use the engines throttles to increase the effective tickover level corner by corner to eliminate the engine braking, this has been very successful in recent years amongst race classes using sufficiently sophisticated throttle systems.

The current trend of the increasing use of motorcycle engine/gearbox units in small displacement racing cars, begs the question of whether the slipper clutch is relevant here. In a racing car the potential benefits are more in the way of stopping over-revving and reducing stress on the gearbox, primary gears and crank bearings near the primary gears. Motorcycle constructors typically design gearboxes that are smaller and lighter than their automotive counterparts; this is because the levels of grip are so low and there is no point carrying around a gearbox that is heavier than you can use. Of course, if the bike engine is put into a car chassis, with loads of grip, it can cause problems. The same for over-revving on down shifts. The grip levels afforded by a race car will easily get a motorcycle engine over-revving, without any of the hopping and skipping of the rear wheel that occurs on a bike but which also helps prevent severe engine damage.