The forces exerted by these rotating masses are proportional to the stroke of the crankshaft and the square of engine speed. As engine speeds have risen, so has the need for crankshaft counterbalancing. Although we see many crankshafts with a pair of counterweights opposite each crankpin, this is not actually necessary, and there are crankshafts with four pins that have four counterweights, and in some cases only two.

There is always some discussion about the proper level of balance to be provided by any crankshaft counterweighting. When engineers refer to a percentage balance factor, this is to do with to the percentage of the reciprocating mass that we aim to balance. A more accurate way to describe it would be 100% of rotating mass and 50% of reciprocating mass – 50% is often the ‘standard’ that people aim for, and this can give a reasonable compromise between crankshaft mass/inertia, peak bearing loads and vibration. However, balance factors of between 20% and 80% are commonly used for various engine architectures, various running speeds and types of use, from light use on the roads to much higher duty cycles in motorsport. Sticking to a favourite value or something that is recommended for a particular brand of engine is often not the best case for engine performance.

Knowing the masses involved, speeds of rotation, cylinder pressures and so on, a good estimate can be made of the bearing loads throughout the engine cycle and at various relevant operating speeds, and if you have been involved in the design of a bespoke race engine then this may be a process you have worked through (unless you have a kind bearing supplier who will do the work on your behalf).

Depending on operating conditions, it might be found that a different balance factor would be a performance advantage, and it is possibly at this stage that the bearing supplier will want to be recompensed for its extra input if you want them to work through all the possible permutations. While the peak bearing load might increase, for example by changing to a lower balance factor, the mean bearing load through the operating cycle in the engine’s operating speed range may fall significantly. By comparing the expected frictional losses for various levels of balance factor, it might be decided to look at a crankshaft with a balance factor of less than 50%.

Like so many other aspects of race engine design, the choice of balance factor is a matter of compromise. In order to run the balance factor that is most advantageous for engine performance, the engine supplier may have to accept lower mileages between engine rebuilds (if the bearings are the limiting factor) or having to use more expensive bearings with a high load rating. 

Written by Wayne Ward

However, the Norton race bikes were anything but. They were incredibly fast, once the engines ‘picked their feet up’, and I remember watching them reel in the faster-starting opposition at Mallory Park in Leicestershire, England, in the late 1980s. These engines won British Superbike and Formula One championships and an Isle of Man TT race. There was even special dispensation given for the machines to compete in the 500 cc Grand Prix series in 1991 and beyond.

For every supporter, however, there was someone vehemently opposed to the Nortons, saying that their 588 cc capacity was nonsense and that they were actually twice as large. That depends though on how the engine is judged to operate. If it is classed as a two-stroke, then 588 cc is probably correct, as the engine draws in the same amount of air as a 588 cc two-stroke. You see what is coming next – the engine also draws in the same amount of air as an 1176 cc four-stroke, but it is neither, so the debate continues to rage.

In addition to the discussion over capacity, there is some debate among engineers as to the correct name for the Wankel’s crankshaft. Thankfully, as with the conventional crankshaft, engine swept volume – that is, displacement – is directly proportional to the throw of the crankshaft. The shafts, compared to those in a reciprocating engine, use much lower throws; in our usual parlance we might say they have a high degree of overlap.  There is no requirement to provide thrust surfaces on the eccentrics (the equivalent of crankpins) as the thrust of the rotors is taken by the housings. There are not the large inertia forces to cope with that we have in the reciprocating engine, and with a large overlap the stresses on the bearings are modest.

Where more than two rotors are used, the crankshaft/eccentric shaft needs to be built up in stages. Each additional rotor requires an eccentric sleeve extension to be added. There have been a number of three and four-rotor engines used in racing – the Le Mans-winning Mazda 787 for example had a four-rotor engine called the R26B, and the extra rotor sections are keyed to the central section, with one fitted at each end in the case of the four-rotor design. Lubrication of the bearings was a simple matter, with eccentrics and main bearings fed from a large drilling running along the axis of the R26B’s crankshaft.

Rotary engines were always rare in racing. They were fast, but their notorious thirst for fuel meant they needed a fuel tank for many race series. Where air restrictors are used to control output, rotaries require a significantly larger restrictor area in order to be competitive. Mazda found racing success some years ago with a rotary engine, but nobody is now seriously considering rotary racers, despite a short-lived attempt by Norton a few years ago to repeat its one-off rotary TT success.

Fig. 1 - The late Steve Hislop won an Isle of Man TT on a Norton with a twin-rotor engine

Fig. 2 - The displacement of the rotary engine is directly proportional to the ‘throw’ of the crankshaft, but beyond this, engineers fail to agree unanimously on engine capacity. This example is from a two-rotor engine

Written by Wayne Ward

The crankshaft itself rotates in bearings in the crankcases, and at very high speeds in some cases. MotoGP engines rev close to 20,000 rpm, while in the mid-2000s Formula One engines were in the same speed range. I have an old 750 cc motorcycle from 25 years ago that will run happily to 14,000 rpm.

Where we have rotating motion, we have centrifugal forces. We often see these as unhelpful in the context of an oil system as we need to run high oil pressures in order to feed oil into the crankshaft oil drillings, fighting the centrifugal forces that are trying to return the oil to whence it arrived. However, it is possible to take advantage of these forces for the good of the oil system and hence the health of the race engine.

Two problems we might suffer from in a race engine as far as the oil system is concerned are aeration and contamination. Excessive amounts of air in the oil can lead to problems with lubrication, and bearings are especially susceptible to damage through oil that contains too much air. Contamination in the form of solid debris is also particularly damaging to bearings, as well as other components. Anywhere that hard, solid particles can enter a narrow sliding contact presents opportunity for damage. For both these situations, we can use centrifugal forces to help us.

The key is to get the main feed of oil onto the crankshaft centreline. This can be achieved when feeding from a gallery into a main bearing, but it is conventionally done with a nose-feed crankshaft, where the main feed of oil is fed into the non-output end of the crankshaft.

In terms of air separation, we have two options. One is very passive and is by the simple expedient of a small oil drilling on the centreline of the crankshaft, passing into the crankcases; the air, which is forced towards the crankshaft axis owing to its very low density, can pass through without dragging very much oil along. The other option is to use a device to take advantage of the difference in angular momentum between the crankshaft and the oil as it is introduced. This is an active air-oil separator, whose function is the same as that of the simple on-centre drilling, but it should be much more effective.

Regarding solid debris, this is again normally much more dense than the oil. If the oil is moving sufficiently slowly axially, it will have time to pick up enough momentum to force the solid debris to the outside of the drilling. By introducing a ‘dam’ at some point along the cavity into which the oil is introduced, the solid debris should be retained there. A scheme working on the same principle was used in some of the Rolls-Royce Merlin engine variants. It must be said though that trying to include the active separation scheme and the debris centrifuge may not be possible in most engines owing to lack of space.

Written by Wayne Ward

As with many aspects of powertrain design, it is possible to combine the thinking behind two concepts, and so it is with the visco-elastic damper. We have described the basic concept of an inertia damper where an inertia ring reacts against vibrations via elastic elements, normally either springs or elastomeric elements such as O-rings. Elastomers can act very effectively as dampers, but we have to be aware of the rate at which we are putting energy into any system containing them. Beyond a certain wattage (or BTU per hour for those who prefer imperial units) per unit volume, the elastomer will quickly degrade into something unsuited for further use as a damping element, so the critical rate of heat addition to a volume of elastomer depends on the exact type and grade used.

So, if we want some controllable damping but don’t want to rely on an elastomer to provide it, is there an alternative? Well, if we are willing to accept a degree of extra complexity, the answer is yes. A visco-elastic damper incorporates springs that have very little damping, and a liquid that takes care of the damping. It allows us to have independent control over the behaviour of the elastic elements (springs) and damping which we don’t have in an inertia damper with elastomer elements. A viscous damper has damping in abundance but no springing.

In a visco-elastic damper, when there is vibration present, fluid is forced to travel through small orifices between pairs of chambers, and the motion of the inertia ring relative to the crankshaft is also controlled by springs. It can be likened to a rotary version of a very simple suspension unit (shock absorber) that we might find on our car or motorcycle. Owing to its elastic elements, the visco-elastic damper can be tuned to react to vibrations that occur at certain frequencies, corresponding to those at which we know serious vibrations occur.

The extra complexity in manufacture comes from having to machine precise orifices for metering the damping fluid and to incorporate seals that will prevent the damping fluid from escaping. There is also the matter of having to put the assembly together with the correct amount of fluid in there. It is fortunate therefore that there are companies which supply such complex units.

I don’t know if any race engine suppliers are using such units, owing to their comparative bulk, but they are certainly used for controlling torsional vibrations on crankshafts and camshafts in industrial engines. Don’t read anything significant into the fact that these are fitted to crankshafts to infer the size of the industrial engines which are equipped with viscoelastic dampers; I have seen more than one highly optimised race engine fitted with camshaft dampers.

Written by Wayne Ward

There are some notable designs of damping device though that could be adapted for use in modern race engines, and the ease with which this is achieved will depend on the space in which the damper can be accommodated.

There are a number of variations of pendulum damper, and although each damping element acts as a pendulum, very few will bring to mind that in a traditional clock.

Several designs of pendulum damper are used on individual crank throws. These can be attached to the side of the crank web on either side, close to the main bearing, or can be designed to form part of the counterweight. It is strange to imagine the counterweight not being firmly fixed to the rest of the crankshaft, but it is this freedom of movement that allows the counterweight mass to move and to oppose the action of the torsional vibration. Some designs have been used successfully in marine and aero engines, and there is no reason why they couldn’t be applied to a modern race engine. Clearly, the stresses in the component supporting the counterweights would be considerable at high engine speeds, so such designs might lend themselves best to lower speed engines.

A variation on the pendulum damper idea is one that might find use more easily in a race engine as a crankshaft damper, as it can be mounted in a carrier and fitted to the end of a crankshaft, as per the more usual elastomer, sprung or viscous dampers. In this design, a number of metallic ‘rollers’ sit within a housing that has compartments into which the rollers fit. When these were first introduced, they were known as ‘roll-form absorbers’, referring to the form of the pendulum as a roller. The difference between the diameter of the roller and that the compartment affects the frequency at which the damper works best.

There may be a damping fluid in each compartment holding a roller, but this is not always the case. The damping fluid helps the device dissipate energy, but it is not required in order for the damper to reduce vibration amplitudes. Problems have been encountered with this type of damper where each pendulum has been made too light and therefore oscillates over a large angle about its mean position, or where there has been insufficient or excessive lubrication*.

The principal problem with large roller pendulum amplitudes is that of slipping, as the change from a pure rolling operation to a combination of rolling and slipping makes predicting the damper’s behaviour difficult. 

* Ker Wilson, W., “Practical Solution of Torsional Vibration Problems”, vol. 4, 3rd edition, Chapman and Hall, 1968

Written by Wayne Ward

Race engines, with their relatively narrow power bands and narrow operating ranges (especially for series running wide-open throttle on ovals), tend to run into the lower speed resonances only as a short-term transient, as the engine increases or decreases speed. Where engines run for extended periods outside their designed operating range, for example during safety car periods, it is common for there to be certain engine speeds which are avoided completely through gear selection.

In the ‘good old days’, when ships were powered by reciprocating internal combustion engines, it was a legal requirement to have calculated all torsional resonances for each engine installation throughout the engine running range, precisely in order that damaging torsional resonances could be avoided.

It was during that period, when crankshaft vibration really could be a matter of life and death (imagine losing drive due to a broken propshaft halfway across the Pacific), that many of the damping devices that have been used with varying degrees of success were first developed. Some of these are still used in racing, while others have been adapted to other uses in engines, for controlling torsional vibrations in the cam gear drive for example.

One such device (and there are many) is the viscous damper, of which there are variations. One of the more popular is similar in principle to the inertia damper described previously. An inertia ring is supported on either its inner or outer diameter in an enclosed volume containing a special damping fluid. During normal operation, the ring rotates along with the damper housing with minimal losses. Only when vibrations begin is there a significant speed difference between the inertia ring and the housing. The oil in the carefully controlled gap between the housing and the inertia ring is therefore subjected to shear, which in turn exerts a torque on the housing and hence on the component to which it is fastened, in this case the crankshaft.

As with the rubber-damped inertia damper, the viscous damper has operating limits – the amount of energy being put into the damping fluid needs to be controlled. Just as a rubber damping element will rapidly degrade if too much energy is dissipated through it, so the damping fluid can also be damaged. The effect is one of a large change in viscosity, and hence the damping characteristics are changed.

The damper can be tuned by dimensional changes to the inertia ring and housing, but the parameters that show the greatest sensitivity to changes are the fluid inertia and the gap between the housing and inertia ring which, along with the speed difference between the housing and ring, dictates the shear rate of the fluid.

Written by Wayne Ward

The problems are most serious where engines are driven at a near-constant speed on full throttle. If this coincides with a critical frequency at which the cranktrain resonates, the amount of energy involved means serious damage can very quickly result. To change the torsional characteristics of the cranktrain so that the operating speeds and resonant frequencies are a safe distance apart can mean that components such as crankshafts, con rods and piston assemblies would need to be redesigned to change their stiffness and inertia. Such changes may not be practical, so we need to see what else can be done without compromising the performance of the engine.

This is where dampers come in. There are many different kinds of damping solutions for cranktrains and valvetrains, but they all aim to reduce the amplitude of vibrations in the operating range of the engine and the high stresses associated with these amplitudes. A good reference for anyone interested in this is the fourth volume of the epic treatise on torsional vibration by Ker Wilson*.

One type of damper that is commonly used is the inertia damper. Here, an inertia ring is contained within a housing, and its rotation is constrained by the friction applied to it by pre-loaded elastomer elements. The ‘tuning’ of the damper is effected by changes to the inertia of the ring, the number, dimensions and damping characteristics of the elastomers and the pre-load on them.

During normal operation, the tendency is for the inertia ring in the damper to oscillate very little, perhaps a few tenths of a degree. When torsional vibration begins, the amplitude and energy in the torsional oscillations builds quickly, and where a damper is fitted the inertia of the ring means that its angular displacement from its normal position is increased, straining the elastomeric elements in the system. This exerts a ‘restoring torque’ on the cranktrain, thus controlling the amplitude of vibration. If properly developed and tuned, such devices can practically eliminate damaging resonances.

There are energy losses associated with dampers, but under normal conditions these are absolutely minimal. The only time when significant losses are found is during periods of torsional vibration, when energy is being dissipated in the form of heat. 

* Ker Wilson, W., “Practical Solution of Torsional Vibration Problems: Devices for Controlling Vibration”, Chapman and Hall, 1968, ISBN 0-4120-8580-1

Written by Wayne Ward

There are a number of ways to optimise any engine component. The traditional way is to run ever more highly stressed components in an engine until one breaks. Running engines on dynamometers, no matter how sophisticated, is unlikely to find all the problems – failures will inevitably occur at the track, which is the only real test of a race engine. However, developing components simply by building lots of engines and running them on the track or dyno is extremely costly in terms of time, effort, money and reputation.

Simulation and data analysis is one way in which we can do much of the development of a crankshaft without committing to cutting any metal or building any engines. By examining previous successful and unsuccessful designs we can arrive at some design allowables, and this is a good basis to begin using finite element stress analysis. Ideally we need to compare fatigue safety factors to be sure of a good design, but many people find reasonable success in improving designs by staying within known maximum stress limits.

Simulation of torsional behaviour is very important – we don’t want to build an expensive engine that fails because of a major torsional resonance in the running range, especially if it coincides with an almost constant engine speed as we might find on high-speed circuits. To simulate torsional behaviour we need to model the behaviour of all those parts of the engine that affect on cranktrain torsional vibration, which means not only the crankshaft and adjacent rotating/reciprocating components but possibly also the valvetrain.

The use of CFD in crankshaft analysis is important when looking to optimise the oil system. There are definitely gains to be had from looking at the interaction of the crankshaft with the bearings – for example, a number of clever developments have resulted from the use of CFD in relation to oil feeds.

Equally important is using CFD to model the two-phase behaviour of the oil-air mixture entering the crankshaft. As much as we like to think that we are pumping pure oil through the engine, it is a fact of life that there is likely to be some air in it (you may be surprised how much), no matter how well designed the oil tank may be. For nose-feed crankshafts especially there are a number of ‘tricks’ that can be developed using CFD to decrease the proportion of air in the oil.

We can also simulate the effect of oil ‘depletion’ along the crankshaft. The crankshaft acts like a centrifuge, as the dense oil naturally tends to be flung to the largest radius. As the oil flows along the crankshaft there can be a tendency for the proportion of air to increase as the oil in the galleries is consumed. CFD can give the cranktrain engineer an early warning about how close they are to having a real problem.

Written by Wayne Ward

The engine is a four-cylinder Nissan unit, not production-based but developed for racing in the DeltaWing in England by RML. I had a good look at a crankshaft, and was really impressed by the detail. I’ve seen the crankshafts in many of the modern-era Formula One engines, so for a crankshaft to stand out, it had to be novel.

First, there were bolt-on counterweights. This is not especially novel in itself, although very few outside Formula One are prepared to take this risk. I’ve designed both counterweights and their bolts, and you need to be very sure about the calculations. When they part company with the crankshaft, they often simply exit the engine and car by the shortest route. The reason for the use of this type of counterweight might not be the same as in Formula One, as in to achieve the very lowest inertia, but to allow machining access for some pretty unusual crank machining without having to lose counterweight mass. Some people choose to accept big holes and slots in counterweights in order to get the crankpin drillings that they want, but this would not have been possible in the DeltaWing, so bolt-on weights were used.

The designer had used some very large crankpin drillings on the conventional inline four-cylinder crankshaft, and these had very generous radii on their exits. They had also taken great pains to be able to machine into the ends of the main bearings where these could be reached by machine tools, again aided by the decision to use bolt-on counterweights, which would not have been assembled onto the crankshaft until well after all of the machining was complete.

There are some real advantages to machining the main bearings – it not only lowers crankshaft mass, and to a lesser extent inertia, but it improves the distribution of stress. The load path is forced away from its natural route and off the plane formed through the axes of the crankpin and main bearing by removing the material through which it would normally take. Instead of the stress being concentrated, it is split and therefore gives a lower stress concentration factor.

This has been proved in the past, especially by the work carried out several decades ago by Matinaglia and Lurenbaum, which is summarised in “The Internal Combustion Engine in Theory and Practice”. Crankshaft stresses were lower and durability improved where main bearings could be bored. This technique lent itself to long-stroke engines whose main bearings could simply be bored from one end, but with some clever thinking the DeltaWing engine showed that it can be achieved in a more modern race engine.

Fig. 1 - Clever machining facilitates clever design for an engine to go in an innovative car 

Written by Wayne Ward

The 2010 article mentioned that welding is sometimes used to prevent the various parts of the crankshaft from losing their angular alignment if the interference between crankpin and crankwebs is insufficient. In addition to angular movement being possible, it is sometimes found that the crankwebs tend to move axially in relation to each other.

There are two main problems with this particular axial movement. The first is that the rod end float can be diminished to zero, with the rod thrust faces in constant contact with the crankwebs. This can lead to rapid thrust washer wear and an increase in friction. If the movement is significant enough, it can cause the piston to be thrust to the side, effectively rotating the area of maximum contact pressure from its usual location. That can mean the piston wears more quickly as the ‘bridge’ between inlet or exhaust ports, against which the piston usually bears, is no longer loaded as intended. The second problem with axial movement of the crankwebs is that the rod thrust clearance can increase; again this can allow unintended off-centre loading of the piston-to-cylinder contact.

It is common to find axial movement of the crankwebs combined with some angular misalignment too, which should not be surprising given that there is insufficient interference to prevent movement.

Welding is a common way to prevent movement, but there are some pitfalls in welding the crankpin to the web(s). If there is too much heat added to the crankshaft during the operation, there is a danger of distorting the crankshaft, thereby making it unfit for use. The crankshaft must be assembled accurately before welding, otherwise we may simply be condemning an engine to run with an inaccurate assembly, even if the welding is sound. If too much heat is introduced then the weld itself might weaken the crankshaft components to the extent that the fatigue life of the crankshaft as a whole is reduced.

The welding is done on the outside face of the crankweb, where the crankpin is pushed through. If proper weld preparation is done on both components, and a skilled welder does the job, it may not be necessary to dress the weld, but it is common to see the weld dressed either for neatness or to prevent contact with adjacent components. Often the weld does not go right around the crankpin, especially where the engine is relatively short stroke. It is common to have just a short run of weld closest to the outside diameter of the crankweb, or to create an intermittent weld around the crankpin.

Even if the process is carried out perfectly, there is the disadvantage that the weld has to be ground away before the crankshaft assembly can be serviced.

Written by Wayne Ward

When cylindrical grinding takes place, the grinding wheel, which is usually much larger than the journal or crankpin being ground, rotates at a much higher speed than the workpiece. The ratio of surface (peripheral) speeds between the grinding wheel and workpiece is very large, so one might expect that the direction of rotation of the crankshaft during grinding would be unimportant, but there is widespread agreement that the direction of rotation of crankshaft travel should be in the same direction as that of the grinding wheel. This means that, at the point of material removal, the surfaces are travelling in opposite directions (if this is confusing, think of gears in mesh – they rotate in opposite senses but in the mesh area the teeth are both travelling in the same linear direction and at the same pitch line velocity).

The act of grinding and material removal tends to shear and lift material from the surface. When the grinding takes place as described above, the action of the grinding wheel on the ground surface acts to ‘flatten’ any material that has been pulled up from the surface (imagine a splinter of material still attached to a piece of wood).

In choosing the direction of crankshaft travel, we need to refer to the direction of rotation of the crankshaft in the engines, as we want the ‘splinters’ still attached to the crankshaft to be flattened by the rotation of the crankshaft rather than be torn out. During grinding, the crankshaft should rotate in the same direction as it does in the engine.

Any polishing operation should also aim to achieve the same effect. The end result should be that any ‘splinters’ should be as flat as possible to the crankshaft surface. Polishing is often carried out to achieve a level of surface finish that is not readily achieved by grinding.

There are several methods of crankshaft polishing. The simplest is manual ‘tape polishing’, where a special-purpose polishing tape is held with manually applied pressure against the crankshaft while the crankshaft rotates. The tape is very strong and has very fine particles of abrasive (commonly industrial diamond) embedded or bonded onto its surface. The polishing action takes place over the ‘wrap’ angle formed by the tape, and the maximum polishing pressure is nominally at the halfway point of the wrap angle.

A better method than manual polishing is continuous tape polishing, where a slowly moving continuous tape ‘belt’ polishes the crankshaft. It has many advantages over the manual – for example, the fact that it is not a manual process removes the ‘artisan’ aspect of polishing; and the pressure and duration can be easily controlled. The process is also much more repeatable than manual tape polishing, so the machine operator can be put to good use on other skilled tasks.

An alternative manual method is to produce a specially sized grinding lap into which polishing tape is fixed or a polishing compound is used; this arrangement is sometimes referred to as a ‘nutcracker’. This results in a more even pressure during polishing and the pressure is more easily controlled, either manually or by pre-loading the nutcracker using a spring of known load.

Written by Wayne Ward

Grinding and polishing are the processes by which all these bearing surface specifications are achieved.

The grinding process needs to be carefully undertaken and has to be considered when machining the crankshaft prior to grinding. Where crankshafts are surface hardened, the depth of the hard ‘case’ on the component is relatively shallow. In order to maintain a constant depth of hardened case, the pre-grinding machining stage needs to be a constant offset from the ground surface. If too much material is removed, the crankshaft will lose the benefit of both the hard-wearing surface and the residual compressive stresses that offer improved fatigue.

Unfortunately, the most common mistakes with grinding allowance happen in the fillets of crankpins and main bearings, where the stress concentration factors are high and from which fatigue cracks are likely to emanate. Equally, there is the danger of not removing enough material, as that can leave a hard, friable layer at the surface which then easily flakes in service.

The resulting debris is very abrasive and can cause further damage in the engine. I have seen an engine that this has happened to, and although the damage was limited to the crankshaft and some of the con rods, it caused a much more rigorous inspection regime to be enacted at the crankshaft manufacturer at the customer’s behest. This was caused by insufficient grinding allowance being left on the crankshaft before nitriding. It should be noted that both under- and over-removal of the hardened case can result from poor machining prior to grinding, or lack of accuracy in grinding.

Even when the dimensions from the drawing are closely adhered to, damage can be caused by grinding if the operation causes excess heat at the surface. Grinding cracks may be parallel in appearance, or a network similar to dried mud. The heat caused by ‘abusive’ grinding can be enough to cause a transformation in the structure of the steel, although cracks can often form without the surface reaching temperatures sufficient to cause a change in structure. Factors affecting the heat generated in grinding include material removal rate, coolant flow rate, the condition of the grinding wheel and the grinding wheel material. Even if the crankshaft is not cracked, abusive grinding can leave the surface of the component in a state of tension, which can be disastrous for fatigue life.

Written by Wayne Ward

Many of these engines used roller bearing crankshafts, where either the main bearings, con rod bearings or both used rolling element rather than plain bearings; these days it is far more common to find four-stroke engines using plain bearing crankshafts. Two-stroke engines use roller bearing crankshafts owing to their need to lubricate the bottom end of the engine with a mixture of fuel and oil. Quite often with two-strokes, the big-end bearing runs directly on the crankpin, and the main bearings have a bearing inner race pressed onto the crankshaft.

That is not to say that, for four-stroke engines, roller bearing crankshafts slipped into the mists of time with the likes of the 50 cc twin-cylinder Honda, its bigger six-cylinder 250 cc cousin, or their contemporaries. Ducati has used roller bearing crankshafts on its Superbike engines for many years, right up until the recent release of the 1199 cc Panigale, which uses plain bearings.

There was also a brief flirtation with roller-bearing crankshafts in Formula One in the V10 era. The reason for this was the promise of lower friction, but while some people found this to be true, with significant reductions in friction, others found the gain to be negligible. According to one leading engineer, this variance was thought to be due to the differences in the amount of oil flow through the plain bearings that the roller bearing crankshaft was being compared to. As the roller bearing requires little oil to provide lubrication then some engines which, as a plain bearing engine, required lower flow rates would see less benefit compared to one which required higher flow rates.

The complexity of a roller bearing crankshaft, and the extra space occupied radially by the bearing arrangement, means it has to offer a very useful increase in performance in order to offset the disadvantages. The FIA effectively closed this avenue of development by outlawing ceramic rolling elements on which the Formula One crankshaft bearings relied. Since then though there have been advances in bearing element steels that might have made such a crankshaft an attractive proposition, had performance development been allowed.

In terms of the crankshaft itself, for the most compact installation of a roller bearing the rolling elements run directly on the crankshaft journals. We have to take account of contact stresses between the bearing rollers and the crankshaft, and particularly subsurface stresses. In this case, the depth of hardening and the variation of strength with depth needs to be carefully considered in order to avoid subsurface fatigue failures.

Written by Wayne Ward

The crankshaft is one component where we can make substantial gains in performance by reducing friction. The losses in bearings due to oil shear are substantial, and any reduction in shear losses are worth chasing. Within the limits of the bearing materials available to us, we can reduce the bearing area in order to reduce friction. Although a loaded bearing is a more complex case, we can learn something from studying a plain shaft running concentrically within a plain collar with oil between the two; this example was used in the Race Engine Technology article on crankshafts in issue 65. The torque required to turn such a shaft is proportional to the third power of the shaft diameter, but directly proportional to the bearing width. There is therefore a much greater benefit to be gained from reducing bearing diameters than width where friction is concerned.

The strategy of friction reduction through decreased bearing diameters is not without risk though, as any useful decrease in the diameters will cause a significant reduction in the torsional and bending stiffness. A large change in torsional stiffness may also serve to bring a torsional resonance within the operating range of the engine, leading to reliability problems. A usually ‘bullet-proof’ engine may therefore become prone to crankshaft failure, or other problems may be introduced if the natural frequency of the cranktrain is changed markedly.

Even if torsional problems are avoided, there is a limit to the useful reduction in crankshaft bearing diameters, owing to a lack of stiffness in bending. Just as torsional stiffness is proportional to the fourth power of diameter, so is bending stiffness, as both are directly proportional to the moment of inertia (second moment of area). Excessive bending in the crankshaft will lead to loading of the edge of the bearing rather than presenting a relatively even load across the face. When a bearing is loaded in the ideal sense – that is, with the axis of the shaft parallel to the axis of the bearing – the maximum pressure is at the centre of the bearing. When edge loading occurs, the bearings are quickly damaged and friction tends to rise again.

One possible solution to limited amounts of bending in the crankshaft is to have ‘barrelled’ crank journals running in cylindrical bearings, or cylindrical journals running in barrelled bearings. The crankshaft may also be damaged by the high contact stresses due to the edge-loading condition, so as we reduce bearing diameters in an attempt to minimise friction, there will be a situation of diminishing returns and eventually no further friction decrease will be seen. In fact going further is likely to increase friction.

Written by Wayne Ward

To produce a fully ground elliptical or conic fillet requires either the use of a specially dressed wheel (‘dressing’ is the process of producing the desired form on a grinding wheel) or to have a grinding wheel with a suitably small corner fillet and to control the position of the wheel accurately during grinding. There is a separate cause for concern though if we are trying to produce an elliptical or conic fillet on a surface-hardened crankshaft, in that we need to accurately ‘offset’ the final profile during the manufacturing stage, meaning that there is an extra emphasis on precision. The use of specially dressed grinding wheels is nothing new in crankshaft manufacture, as ‘barrelled’ main bearing journals have this requirement.

If we are to produce any form of ellipse using conventional turning, the operation is pretty straightforward if using CNC machinery, which is not restricted to following simple radii. However, in recent years, as explained in a recent article on crankshaft manufacture, there has been a marked trend toward the use of CNC milling to rough and finish crankpins. Here, the options for producing the required conic section fillet are to use a cutter with a smaller corner radius and to generate the corner radius by careful control of the cutter, or to use a cutter with the desired corner profile ground onto it. Either method would work, but the latter is perhaps preferable in terms of reducing machining time.

So, if we can produce an elliptical or conic section fillet, would we want to do so deliberately? Well, there are some pointers in standard texts on the subject. “Peterson’s Stress Concentration Factors” refers to work from the 1940s which examined the stress concentration factors for a flat stepped bar in bending for elliptical fillets of varying ‘aspect ratios’ between 1 (that is, a simple radius) and higher ratios (ellipses). For this simple case there appears to be a clear advantage with the elliptical fillet, with the theoretical stress concentration factor, Kt, being lower for elliptical fillets, although the magnitude of this advantage is generally blunted somewhat as the various factors are applied to give a fatigue stress concentration factor. However, it should still present a quantifiable advantage.

Fortunately, living in this age of powerful computing, we are able to simulate more complex load cases and geometries quickly using FEA. A ‘quick' and dirty’ comparison using geometry approximating a simple, low-overlap crankshaft with no boring of the pins or main bearings showed that there was no advantage to be gained by using elliptical fillets, although for this particular geometry there was surprisingly little disadvantage owing to the geometries simulated. The case was one of bending, and the fillet ellipses simulated were a) semi-major axis = semi-minor axis = 3 mm (that is, a plain 3 mm radius); b) semi-major axis = 2 x semi-minor axis, semi-major axis = 3 mm, parallel to crankpin axis; and c) semi-major axis = 2 x semi-minor axis, semi-major axis = 6 mm perpendicular to the crankpin axis.

Although there exists the manufacturing techniques to make crankshafts with elliptical or conic fillets, from a very limited simulation there appears to be little incentive to make such a crankshaft.

Written by Wayne Ward

It is very likely that you will encounter a crankshaft with counterweights, and you might notice that the edges are chamfered or bevelled. Some people design and manufacture counterweights with more extreme bevelling, sometimes referred to as ‘knife-edging’.

The aim is to reduce windage losses in the crankcase, allowing the crankshaft to cut through the mixture of oil, air and blow-by gases with minimum disturbance. However, there are two schools of thought on the significance of this effect.

If we imagine the air-oil mist to be a stationary dense ‘fog’ through which the crankshaft counterweights must pass, it seems logical that we should make some real effort to shape the crankshaft (and possibly other moving components on the crankcase) such that the amount of energy lost through ‘aerodynamic’ effects is minimised. The knife-edging will have its greatest effect at the largest radius on the crankshaft. Removing material here while maintaining the same balance factor means increasing the breadth/width of the counterweight, assuming that we don’t make its radius larger. This knife-edging generally means that the counterweight centre of mass moves slightly back towards the crankshaft axis, with the larger counterweight mass compensating for this. So, we have larger crankshaft mass and inertia for an expected aerodynamic gain.

The second school of thought says that the oil-air mist in the crankcase is anything but still, and circulates rapidly in the direction in which the crankshaft is travelling. Thus, the difference in speed between the crankshaft and the oil-air mist is not as high as we might first assume, and any aerodynamic losses are consequently not as high. This school of though leads us to spend less effort on knife-edging of crank counterweights. It can’t be argued that anything other than crankshaft motion has imparted significant angular motion on the crankcase oil-air mixture, but we have to note that its effects are likely to be small and most noticeable when the engine is being asked to accelerate.

Lots of people offer knife-edge crankshafts, and one has to take a view on how significant this effect is. It is likely to have a very small effect on steady-state output, and if the output can be demonstrated to have improved then it might well be for reasons other than the intended aerodynamic drag decrease. If we look hard at the other effects in play here, we might find that there are other clues as to where we might minimise cranktrain friction losses.

Written by Wayne Ward

In the vast majority of cases, racing crankshafts are made from steels that are subsequently nitride hardened. Although processes like carburising (also known as case hardening) have been used in the past, nitriding is now used to almost the complete exclusion of other processes for highly stressed crankshafts.

There are several advantages to nitride hardening. The obvious one that we might draw from the name of the process is that nitrides are formed at the surface to create a hard, wear-resistant layer. However, just as important is the level and depth of residual compressive stress created. This has a significant and positive effect on the endurance of the part, and for this reason there are lots of applications for nitride hardening where there is actually no requirement for a hard surface.

The same can be said of carburising. Compared to carburising, however, nitriding takes place at a relatively low temperature, and there are no quenches in the treatment. Distortion is therefore minimised.

There are some drawbacks to nitriding though. The process generally takes a very long time and so can be quite costly. The longest cycles I have heard of for racing parts take more than seven days in the furnace; although a week or more in a furnace is not typical for a nitriding process, it can form a significant part of the cost of a component.

Perhaps of greater importance is the fact that the nitriding process, when one takes into account shipping times and so on, can constitute a large proportion of the overall manufacturing time, and this needs to be planned for. Unless your nitriding supplier has multiple furnaces running long cycles at staggered intervals, your wait for parts to be turned around can be much longer than the nitriding process itself. For example, a 90-hour cycle is not unusual, and missing the start of this cycle by a day means you will need to wait an extra three days for the next process to start - assuming of course that another 90-hour cycle is scheduled to run.

There is significant interest therefore in steels that can be nitrided more quickly. The rate of diffusion of nitrogen into the surface of the component and through the material immediately beneath the surface is controlled in part by the process temperature. This is often limited by the tempering temperature of the steel. If the tempering temperature is exceeded, the steel will begin to soften. During a long cycle, the effect is significant.

Nitriding steels have been developed that have higher tempering temperatures. This means they can be safely nitrided at higher process temperatures without risking softening of the substrate material. Such steels offer the possibility of much reduced nitriding process times, and therefore faster overall manufacturing times for key components. It is quite possible that the component costs can be reduced.

Written by Wayne Ward

For a recent article on crankshafts for Race Engine Technology magazine, crankshaft suppliers revealed that the proportion of their racing crankshaft business which has nose-fed oil galleries has increased in the past two or three years. There are some basic advantages of nose-fed crankshafts in terms of requiring less oil system pressure and, possibly, some simplification of the maze of interconnected drillings arranged between the oil pressure pump and the numerous entries to the main bearings.

Some nose-feed crankshafts feed main bearings and crankpins, while others feed crankpins alone. In the latter case, the oil drillings in the cylinder block assembly are probably not appreciably simplified by using a nose-fed crankshaft. Where the crankshaft carries the oil flow for the main bearings and crankpins, the simplification of the oil drillings might improve the durability of the cylinder block. Another issue is that drilled and grooved main bearing shells would not be required, and this might improve bearing life.

The work by Lurenbaum and Martinaglia referenced by Taylor* in his book looks at the various methods for increasing crankshaft fatigue life through the disposition of material in a crankshaft. There are several ideas used widely in production and racing crankshafts these days, but there are others that do not lend themselves to modern short-stroke race engines. In particular, the boring of main journals is problematic where there is a requirement to accommodate other oil drillings. Also, depending on the size of the main journals and crankpins - and considering overlap - the size of through-bore may not be worthwhile.

The boring of crankpins and main journals is an effective way to manipulate the load paths in the crankshaft, thereby improving fatigue life. Of course, hollow main bearings and pins have the advantage of reduced mass and inertia. Boring pins through also means that smaller counterweights can be used, which again reduces mass and inertia further.

However, where a nose-feed crankshaft is used, and where the oil galleries are of an appreciable size, the effect of the interconnected drillings forming the oil gallery in the crankshaft will tend to spread the load path and direct it away from the area that is normally the worst case for stress concentration. Although the layout of the drillings is not the same as the crankpin and main bearing bores shown in Taylor's book (they run parallel to the crankshaft axis) the overall effect will be similar.

In fact, owing to the proximity of the drillings to the main bearing and crankpin fillets, the effect may even be enhanced compared to the drilling patterns shown in Taylor. While the spreading of the load path will lead to increased stress in other locations, providing the overall maximum stress is reduced, the effect is likely to be positive in terms of fatigue life.

* Reference
Taylor, C.F., "The Internal Combustion Engine in Theory and Practice", vol 2, MIT Press, 1982, ISBN 0-2627-0016-6

Written by Wayne Ward

Traditionally, the work involved in producing the basic functional shape of such parts has been done on the lathe. We can imagine taking a billet, casting or forging, and machining the main bearing journals, as these lie 'on centre', the axis of the rough part as it is mounted in the machine. However, when we come to machine the crankpins, each of which lie off-centre, life becomes a little more complicated. The billet has to be accurately offset within the lathe, and the cuts taken at this stage are intermittent, which means that the lathe, its foundations and tooling all need to be very stiff to withstand the 'impact' of the cut starting on each revolution.

A single-cylinder crankshaft or a 360° twin inline twin-cylinder crankshaft is relatively simple, as it has a single off-centre machining operation in the lathe. For a typical inline four-cylinder or flat-plane V8 crankshaft, there are only two pin positions to consider, but these need to be in the correct position relative both to the crankshaft axis and each other. Other configurations are yet more complicated, with a cruciform or crossplane V8 crankshaft requiring four offset positions, and an inline five or a V10 requiring five crankpin positions, all of which have to be accurately positioned relative to the crankpin and in the correct angular relationship to each other.

The continuing trend towards smaller main bearings and smaller journals makes the manufacture of such crankshafts more difficult. While it is possible to use a 'steady' when turning on-centre, as is the case when machining the main bearings, the same can't be said when turning crankpins. The crankpins need to be accurately machined before grinding so that the amount of material removed during the grinding process does not leave one area with a very thin hardened case; this is especially important for crankshafts that have only a thin nitrided case.

A more modern method of accurately, consistently and quickly producing a crankshaft is to take advantage of CNC machining. The crankpins can be produced with the billet always rotating between centres. The method here is to move the tool continuously when machining the crankpin, and this kind of work is well suited to a 'mill-turn' type of machine, where a powered milling cutter replaces the manually controlled tool post of the lathe mentioned before. The spinning tool 'chases' the crankpin around as the crankshaft is slowly turned on-centre.

The advantages of this method of manufacture are numerous. An important one is that the work can be done while using lathe steadies, minimising workpiece deflection. The cuts are also light and continuous, again minimising workpiece movement, and only one set-up is required; with the number of set-ups reduced to a minimum, the chances of errors creeping in are therefore mitigated.

Written by Wayne Ward

Where plain annular contacts are used, we are only able to generate pressure in an oil film due to 'squeeze film' operation - that is, where there is a decreasing clearance between the two annular surfaces. Otherwise, we rely on enough lubricant being forced through the contact to prevent metal-to-metal contact. There are a number of design features that can be employed on the thrust face of the con rod to allow hydrodynamic lubrication to become established. Very little thrust area is required as the forces, and thus stresses, involved are only very small. A full annulus is generally used because people are comfortable with this solution, especially given the difficulties in generating pressure in the oil film.

However, we are not limited to producing a full annulus on the crankshaft to bear against the con rod thrust face, and there are a number of reasons to consider something other than a complete annulus. Most obviously, there is a weight saving to be made, especially if a portion of the thrust face is removed at the furthest distance from the crankshaft axis. Not only can this weight saving be realised, but weight on the opposite side of the crankshaft axis can also be saved as there is less mass to balance. Fig. 1 shows the sort of material removal that might be considered if weight is to be removed from the outer part of the rod thrust face on the crankshaft.

Less obvious is the fact that equally significant mass savings, accompanied by very useful reductions in stress concentration factor, can sometimes be made by removing part of the rod thrust face on the crankshaft at the point closest to the crankshaft axis. By removing the innermost part of the thrust face, we can create a more generous fillet radius, especially where we undercut the crank web. Undercutting and removing material from the crank web, when done judiciously, is a proven method to reduce stress concentration. I would refer the reader to Taylor*, which neatly summarises earlier work by other engineers on the subject of crankshaft stress concentration.

Care must be taken when interrupting the rod thrust face on the crankshaft. Careful detail of the leading edge of the thrust face is required to prevent wear. The edge should not be left sharp.

Another advantage of not having a complete annular thrust area is that such a design is actually likely to provide the hydrodynamic oil film which is impossible to create between parallel flat surfaces.

* Taylor, C.F., "The Internal Combustion Engine in Theory and Practice", vol 2 (Combustion Fuels, Materials, Design), MIT Press, 1985, ISBN 0-2627-0027-1

Fig. 1 - Reducing the thrust area between the crank and rod can bring a number of benefits, including lower crankshaft mass and inertia. Compare the left side, having reduced thrust area, with the right

Written by Wayne Ward

The current Yamaha R1 is a four-cylinder engine that has made use of a cruciform crankshaft to give an uneven firing order. This was based on the successful YZF-M1 MotoGP race bike that enjoyed world championship success with Valentino Rossi and Jorge Lorenzo. Whether this is a real benefit for the road motorcyclist or the racer is unclear. Would Rossi and Lorenzo have enjoyed so much success with a flat-plane crankshaft in their bike? I would suggest that they would have been equally successful.

It seems similarly wrong to suggest that Yamaha would encumber their MotoGP bikes with unnecessary complication. Perhaps the riders prefer the feel of such machinery. You will have noticed that uneven firing in production vehicles is mainly restricted to motorcycles. Such machines are rarely designed with smoothness in mind, and one might argue that an uneven firing order adds to the aural appeal of a motorcycle.

Roadcar manufacturers, on the other hand, have no truck with such uneven firing orders - smoothness and lack of vibration and noise are highly valued. There are some vee engines where even firing would not be possible if the rods were run on shared crankpins. A good example of this are the 90º V6 engines used by a number of manufacturers. These are used either because the manufacturers have a 'modular' engine family with a 90º vee angle, or by design because a 60º vee is felt to give too tall an engine, and a 120º angle is either too wide or does not package nicely in a vehicle. For a V6, both 60º and 120º three-pin crankshafts give even firing intervals.

For those who choose to run such production-based V6 engines in their racecars, when a racing crankshaft is manufactured a choice has to be made whether to carry on with a split-pin crankshaft or commission a three-pin crankshaft. The split-pin option is more expensive as it has three extra crankpin grinding operations, but allows other pieces of hardware and electronic circuitry to be retained. The engine control unit would need to be reprogrammed if changing to a three-pin crankshaft - we don't want half of the engine firing 30° early or 30° late.

Where a pushrod V6 is concerned, switching away from the even-firing split-pin crankshaft would require new camshafts to be commissioned as well, owing to the fact that, between the adjacent rods, we need to provide space for a 'flying web' to bridge the two pins. We also need room between the two adjacent big-end bearings for fillets.

In providing for the flying web and the pin fillets, the 'bank stagger' is the limiting factor. A very compact engine will have less bank stagger and thus less opportunity to provide wider bearings. It is easy to see that with a split-pin crankshaft we might become limited on rod bearing width, and this might dictate the maximum state of tune of the engine.

Fig. 1 - The space left by the bank stagger has to accommodate 'flying webs' and fillets

Written by Wayne Ward

At the end of the previous article, I said we would further develop the idea of air and debris separation.

Air in oil is a real nuisance, especially where we want to use the oil to provide the lubricant in a hydrodynamic bearing, as is the case with the main and big-end bearings. Discussions with crank manufacturers in the past has revealed that it is common not to feed the main bearings via a nose-fed crankshaft, but only to feed the big ends. Nose-feed crankshafts are very much the domain of the specialist engine design company, and many crank manufacturers will only make such parts to designs supplied to them.

The separation of air and oil relies on the air-oil mixture having a rotational velocity applied to it. If the aim is to provide proper separation of the air and oil in a short length, then we need to ensure that the air-oil mix undergoes rapid angular acceleration. It may not be enough though to rely on the transfer of momentum from the crankshaft to the oil by shear stresses alone, especially when dealing with high flow rates. In this case, we may need a rudimentary paddle wheel or impeller to impart the required energy. This has the added effect of causing small air bubbles to coalesce. Larger air bubbles can be separated more easily as the ratio of centrifugal forces to viscous forces is increased (look up 'Stokes Flow') although this effect is probably minimal.

Once we have used the nose-feed cavity to separate the air and the oil, what do we do with the air that is now contained in a theoretical cylinder on the crankshaft axis? Common practice is to drill through into the first crankcase cavity, into which the air is then expelled; we have to accept though that some oil may also find its way through the hole. At the far end of the crankshaft we also have to consider trapped air, so we should provide a 'bleed' through which this air can escape, and this is commonly also connected to a crankcase volume.

The separation of solid debris in the oil again relies on there being a difference in density. It is certainly true that our filtration should be sufficient to contain debris that could prove damaging, but I'm sure I'm not the only engineer to be caught out by faulty assembly, damaged filters or debris that has become loose after engine start-up, following lots of flushing, ultrasound and so on, especially if cast galleries are used.

In the case of solid debris, we have to provide a 'dam' to retain the debris on the outer wall of the nose-feed cavity. Such a dam can be provided by a machined step or an undercut. I expect that fewer people consider the separation of debris than consider air when designing a nose-fed crankshaft, especially compared to decades ago, where good filtration could not be relied on.

Fig. 1 - The Rolls-Royce engine in the Supermarine Spitfire used both air and debris separation in its nose-fed crankshaft

Written by Wayne Ward

The idea behind them is simple: the oil is introduced to the crankshaft along its centreline rather than radially inwards via the main bearings. In feeding the oil into the crankshaft at a point on its axis, we lower the oil pressure requirement.

There are other advantages to nose-feed crankshafts. We can use the centrifugal action of the rapidly rotating shaft to improve the quality of the oil feed to the bearings, in terms of minimising both aeration and hard particles that may have got into the oil system. We may have a three-phase system, of which we want only the oil (the other two being solid debris and air) to be sent to the bearings. Centrifuges are a common industrial method of separating all kinds of materials, and the cavity of a nose-feed crankshaft that receives the oil is, in effect, a small centrifuge that we can use to help separate the oil from the other phases.

If we can get the fluid to spin at a high enough rate then we can separate the oil in a short length. If we ignore the issue of having solid debris for a moment, it is easy to imagine that the oil, being several hundred times more dense than any air entrained within, will tend to travel radially outwards, forcing the air radially inwards towards the centreline of the crankshaft. This is the same simple principle that governs rotary mechanical air-oil separators that can sometimes be found on race engines, except that the proportion of oil and air are much different in those devices.

If we extend the same principle a little further, this time including the solid debris that might have found its way through a damaged filter or might have been present in the oil system at first build, then this should make its way to the outside wall of the nose-feed cavity. Owing to the smaller difference in density of the oil and debris compared to the oil and air, and taking into account the viscous drag that the oil imparts on the debris, the process of centrifugal separation is not as rapid when trying to separate debris as it is with air.

There is another issue to consider once we have grasped the principles of centrifugal separation and convinced ourselves that we can make it work - how do we ensure that neither air nor debris are likely to enter the galleries that connect the nose-feed cavity to the main and big-end bearings? This will be the subject of further articles.

* Rubbra, A.A., "Rolls-Royce Piston Aero Engines - a designer remembers", Historical Series no 16 :Rolls Royce Heritage Trust, 1990. ISBN 1-8729-2200-7

Fig. 1 - Operating on the same principles as a swing carousel, the nose-feed oil system of a crank can separate oil, air and solid debris

Written by Wayne Ward

With increased engine speeds comes increased component surface speeds, so crankshaft counterweight speeds are significant. Little wonder then that a common practice on bespoke racecar crankshafts is to 'knife-edge' the leading and trailing edges of the counterweights. However, this strategy is often not as effective as many people might imagine. Because of viscous effects, the air-oil mixture in the crankcase also travels around at a significant speed, and the frictional drag on a body is related to the surface speed relative to the atmosphere through which it is travelling.

Often the more significant effect is that of oil shear between the outer walls of the crank counterweights and the inside walls of the crankcase cavities. Even a local small gap can have a significant effect.

The strategy of keeping the smallest possible amount of oil in circulation in the crankcase also has great value, as it means that viscous drag losses caused by the crankshaft sweeping through the air-oil mist are kept to a minimum, and there is less oil on the crankcase walls that might be subjected to shearing.

It is common to see crankshafts that have the same detail on the leading and trailing edges of the counterweight. In these days of fully 3D-machined counterweights, should we not expect to see the leading and trailing edges dealt with somewhat differently? We know in which direction the crankshaft will rotate in an engine, so we can identify leading and trailing edges with ease. Look at the wings on a racecar or aircraft, blades for gas-turbine engines or helicopter rotor blades. Despite widely differing designs, what they have in common is a relatively blunt leading edge and a relatively narrow 'sharp' trailing edge. The trailing edge ensures that the flow from both sides of these 'wings' is merged without significant separation of flow and recirculation in the wake of the wing. Separation is not such a problem at the leading edge of the wing, with pressure gradients helping to keep flow attached.

If we are to consider aerodynamic counterweights, ought we not also consider something more highly engineered than the 'zero-lift' devices we often see? The possibility to design a counterweight to pull oil off the crankcase walls and into the high-speed flow outside the boundary layer may have benefits in lowering shear losses and increasing scavenging.

Engine designer the late Hiro Kaneda remarked that while people look to the cylinder head for performance, the bottom end of the engine is a 'treasure chest' in which improvements in performance can be found. Certainly the mitigation of friction in the crankcase can yield significant gains (depending on how bad the situation is before development work is done). As this continues, the 'treasure' may become more sparse and harder to find. Clever thinking with counterweight design might be one way to reduce friction further.

Fig. 1 - A crankshaft with carefully machined leading and trailing edges on the counterweights

Written by Wayne Ward

The reasoning can be found in the width of the human knee. Complete with protective clothing, a biker is at least 350 mm across the knees, even with them clamped tightly together. With a chassis in between them, the rider's knees are a lot wider than most engines. Using similar reasoning, ankles are about 125 mm wide each, and they have to go each side of the swinging arm (which, in turn, must be wider than the rear tyre) so there's no point in trying to make an engine much narrower than this. (In the previous Grand Prix motorcycle racing formula, for 500 cc two-stroke engines, Honda made both V2- and V4-engined bikes, and the difference between them in overall width amounted to only a few millimeters.)

Perhaps surprisingly, having a light crankshaft of low inertia seems to give no particular advantage in bike racing. Riders are usually happier with the bigger flywheel effect of a heavy crankshaft. In other forms of bike racing, like motocross or flat-tracking, it is common for riders to optimise the crankshaft inertia with bolt-on flywheel weights. Bizarrely, the motorcycle crankshaft is thus now a piece of sports equipment which, like a shotgun, cricket bat, or golf club, has an optimum inertia that must be tuned to individual competitors.

A current MotoGP bike spends a surprisingly small proportion of a race developing full power. Wide-open throttle (WOT) is commonly used for less than 20% of a lap, even on high-speed circuits. Critical to improved lap times is best possible use of available grip when accelerating away from the apex of a corner. To this end, MotoGP bikes rely on a combination of advanced traction control systems and measures to improve the riders' 'feel' of the tyre grip.

A racing motorcycle can bank to more than 50º from the vertical, and its movements in yaw and pitch are much more significant than those which a racecar experiences. The inertial and gyroscopic effects of the crankshaft are certainly large enough for riders to notice, and they prefer the feel of a heavy crankshaft because the flywheel effect of a high-inertia crank slows down its rate of response and makes it easier to control.

In trying to improve the performance of their I4, Yamaha engineers speculated that an inherent (and hitherto unnoticed) advantage of the vee-configuration engine was the constant crankshaft speed, due to one piston being somewhere near to maximum velocity as its partner (which shares the same crankpin) is at standstill. By contrast, the cyclic speed variation of the crankshaft of a conventional I4 engine is relatively large, because all the pistons change direction - that is, come to a halt - at the same time, and all of them reach their maximum speeds at more or less the same time.

Yamaha's solution, the 'crossplane' layout of the latest I4 MotoGP crank designed to mimic the V4's low 'inertia torque', gave rise to a more constant crankshaft speed, and also to a lop-sided exhaust note due to the irregular firing intervals.

As is often the case in race engine design, an elegant theory has failed to make much impact in the real world. The Yamaha MotoGP engine is down on power compared to its V4 rivals. Yamaha released a production sports bike with a crossplane crank, which was hailed with great fanfare by journalists for its uncanny grip out of corners, and yet in racing for modified production bikes - where the crossplane crank competes with conventional 'two up, two down' I4s - the Yamaha is only tolerably competitive in the World and British SuperBike series. It also made no impression on the leader board at the 2011 Isle of Man TT races, a 'real roads' event where one might expect the Yamaha's supposed improved traction to be most significant. All the TT races this year were won by conventional flat-crank I4s.

Fig. 1 - Yamaha's crossplane I4 crankshaft

Written by Ian Cramp

There are some important changes and restrictions affecting the design of crankshafts for 2014, which go beyond the obvious requirement to deal with two fewer con rods.

There is a requirement to have crankshafts with only three crankpins; there is no option to have either a six-pin crankshaft or one with 'split' or offset pins. Both offset pin cranks and six-pin cranks are commonly found on production six-cylinder engines, particularly ones whose bank angles are 90º (as is the mandated bank angle for the 2014 Formula One engines). The main benefit of this rule would seem to be the prevention of effort and money being expended on investigating alternative crankpin arrangements.

There are two other new crankshaft regulations that would appear to be linked. These are that the crankshaft height is mandated at 90 mm from the bottom of the engine, and that no material with a density greater than 9000 kg per cubic metre can be assembled to the crankshaft.

In the past, much effort was put into the design of Formula One engines that had the lowest possible crankshaft axis height above the bottom of the engine. Not only could the engine thus be positioned lower in the car, but the gearbox input axis was also positioned low down. With an increased stroke and much higher cylinder pressures in the new turbocharged engines, Formula One engine designers would need to increase crankshaft axis height, but with the new rules, they no longer need to design to a minimum.

Combined with mandated fixing locations between the engine and gearbox, this allows engines to be swapped season-to-season, giving privateer teams more opportunity to use the most competitive engines. Quite what this does for independent engine suppliers isn't clear, but it probably doesn't encourage them into the sport/business.

Given an easily achieved crankshaft axis height, there is less incentive to use tungsten counterweighting. However, in the search for minimum crankshaft inertia, people would undoubtedly turn to tungsten crankshaft counterweights, as is the case in other, lower-budget series and even some high-end roadcars. As a method of producing a low-inertia crankshaft, tungsten is very effective. Its removal from the rules doesn't stop people working toward low inertia; it simply means the crankshafts inside these engines will be increasingly intricately machined.

Any gears in the timing drive from the crankshaft to the camshafts must be a minimum of 8 mm wide, having previously not been subject to any restriction, and this logically means that the gear on the crankshaft is also a minimum of 8 mm wide.

The rule prohibiting welding between the front and rear main bearings remains, making the use of the 'hollow crankshaft' unlikely in the near future, although novel manufacturing technologies may allow hollow crankshafts without breaking this particular rule.

Fig. 1 - The option of running a 'split-pin' crank, as per this example from a Honda NSX, isn't an option for the 2014 Formula One engines

Written by Wayne Ward

Sharp's article, and the preceding one (2), discuss the reasons for choosing induction hardening in preference to nitriding. In this case, it was felt that induction hardening would be better suited to a roller bearing crankshaft, although nitrided crankshafts have run successfully with rolling element main bearings. Rolling element bearings impose considerable contact stresses on a crankshaft, and the maximum stress actually occurs below the surface, at a depth which can be beneath the diffusion zone of the nitriding. Subsurface fatigue can occur, leading to pitting and spalling, and soon to overall failure. A deeper hardened layer means that the material at the point of maximum subsurface stress is stronger and harder, and thus is more able to sustain a given amplitude of cyclic stress.

Carburised crankshafts fell out of general favour many years ago as nitriding became more popular. Compared to nitriding, carburising (also known as case hardening) requires much higher processing temperatures, and distortion can be a serious problem. Given the increasingly slender proportions of crankshafts in modern race engines, distortion can be a serious concern. In this process, carbon is diffused into the surface of a low-carbon steel to produce an outer layer which is very hard and wear-resistant. In common with nitriding, it offers a considerable amount of residual compressive stress at the surface, which is proven to improve endurance limit - 800 MPa of compressive stress at the surface is typical of a 0.25 mm deep carburising treatment (3).

Modern carburising methods that cause less distortion have been developed, and perhaps here is an alternative to nitrided crankshafts where roller bearings are used. Low-pressure carburising, also known as vacuum carburising, is becoming more popular, allowing some gears to dispense with post-carburising grinding.

In comparison to nitrided surfaces, carburised parts can be treated to have a much greater depth of hardening. The process is also shorter than typical nitriding processes and generally less costly.

Nitrocarburising is also used for race crankshafts, but not as widely as nitriding. The process is done at a slightly lower temperature than nitriding, and both carbon and nitrogen are diffused into the surface. It is capable of producing high levels of compressive residual stress but, in comparison to nitriding, the depth of the hardened layer and the depth of compressive stress are very much reduced. The hardened layer is so thin that it allows for no post-hardening material removal without seriously compromising the results of the hardening. Grinding is certainly out of the question if surface hardness and compressive residual stresses are to be preserved in the fillet radii of the crankshaft.

References
1. Sharp, T., "Which Surface Treatment Is Best?"
2. Sharp, T., "Which Hardening Method is Best?"
3. Tobie. T., et al, "Systematic Investigations on the Influence of Case Depth on the Pitting and Bending Strength of Case Carburized Gears", Gear Technology Magazine, July/August 2005

Fig. 1 - While nitriding dominates surface heat treatment methods for race crankshafts, there are alternatives (Courtesy of Arrow Precision)

Written by Wayne Ward

The general trend for trying to reduce engine friction means that modern design practice is often in the direction of reducing bearing diameters. This means greater care needs to be taken in manufacture, especially heat treatment, if serious distortion is to be avoided. Manufacturing stresses, with surfaces commonly stressed in tension by the shearing action of machining, mean that distortion is likely in heat treatment where these are excessive.

The traditional method of rough-machining the crankshaft was to turn each crankpin on a large sturdy lathe with the crankpin on centre and the central axis of the crankshaft offset by half of the crank stroke. The forces involved were considerable, and the machining was time-consuming. The advent of modern machining inserts has improved things, but these types of tools are often not forgiving of the type of intermittent cutting resulting from eccentric turning.

What has markedly improved the machining of crankpins is the modern machining centre, where cranks can be produced on lathes with milling attachments or on milling machines equipped with a fourth axis. In these circumstances the pins are produced by a milling cutter that follows the motion of the pin while the crankshaft blank is slowly rotated. Not only are the machining forces lower, but there are other advantages too. The rough-machining operation is quicker, and management of machining swarf is much improved. Crankshafts thus produced are likely to suffer less from distortion in heat treatment.

Modern machining methods have also improved the accuracy of manufacture and reduced the amount of manual dressing of features. Features that previously required careful manual processing were the edges of bevels on the webs between crankpins and main bearings, as well as the edges of oil holes. Both types of feature are commonly finished now by CNC machining methods, and the surfaces of crank-bevel features are no longer restricted to swept surfaces defined by eccentric machining on a lathe.

The consistency of oil-hole dressing is very important, as these features are often the weak point of a crankshaft as far as fatigue is concerned, and it is likely that incorrect oil-hole preparation - or even forgetting to dress one hole - would render an otherwise optimal crankshaft unfit for use. Where the edges of oil holes are machined, there should be no excuses for not having consistency of form and finish before nitriding.

The use of 3D machined surfaces, courtesy of modern CNC machining processes, opens the door to more complex shapes being used, and for the crankshaft designer to design lighter crankshafts which (with some thought put into the design and analysis process) can be more resistant to fatigue failure.

Race crankshafts are always likely to remain difficult and time-consuming to produce, with specialist producers being able to invest in new machinery. Perhaps the most impressive crankshaft manufacturing machinery is that used for production crankshafts, where the time taken for the entire machining operation is measured in seconds rather than tens of hours.

Fig. 1 - Modern crankshaft manufacture has given us shorter manufacturing times, lower levels of machining stress and the opportunity to design a light, stiff, fatigue-resistant part (Courtesy of Pankl)

Written by Wayne Ward

Then there are other regulations that outlaw materials which nobody would choose to use anyway. It is into this category that titanium falls when being considered for crankshafts. There are stories that, a couple of decades ago, someone tried (unsuccessfully) to make a titanium crankshaft for Formula One racing. I have no doubt that it is entirely possible to make such a component, but what I seriously doubt is whether it would represent any improvement over a steel crankshaft. There are a number of reasons for this.

Titanium is notoriously poor in sliding contact. Indeed, steel is nitrided for wear resistance in highly loaded crankshafts. Titanium would need some very special surface treatments to allow it to run successfully without suffering serious wear in a very short time. For the sake of argument, let us assume that we are able to overcome this hurdle using a combination of treatments.

While there are a number of high-stress points in a crankshaft - fillet radii and oil drilling exits being prime examples - an important aspect of the design process is to consider the stiffness of the component. This, in addition to other design variables and component masses, has an effect on the torsional behaviour of the crankshaft system and dictates the critical torsional frequencies.

If a steel crankshaft had been optimised in this respect and the requirement for a titanium crankshaft would therefore be to match the stiffness, it would need to be much larger in diameter. In terms of torsional stiffness, the product of the shear modulus and the second moment of area needs to be matched. In choosing titanium, this can be done easily, as the second moment of area is proportional to the fourth power of diameter. Shear modulus is proportional to elastic modulus, and to match the shear rigidity of a steel shaft - the product of shear modulus and second moment of area - the diameter of the titanium shaft needs to be greater, according to the following equation:

We will find that the titanium shaft needs to be about 16% larger than the equivalent steel shaft to maintain torsional rigidity, and therefore the titanium shaft has 35% greater area. The product of area and density gives the titanium shaft a 24% lower mass, so it appears to be an attractive material, if the surface can be treated such that it will last in a race engine.

However, the real disadvantages come when considering the effect on other components. As a consequence, the bearing surface speeds are 16% higher. In terms of friction this is a big disadvantage, and affects both the big-end and main bearings.

The big end of the con rod needs house bearings sized to accommodate a 16% larger crankpin. The means a much heavier con rod, and by moving the rod bolts further away from the applied load means they have a greater component of bending. This, along with other design changes to the con rod, may necessitate a redesign of the bolt.

It is common, certainly in the case of 'vee' engines, that the lowest point of the con rod locus is defined by the head of the con rod fastener. By moving this away from the crankpin centreline, the locus sweeps lower in the crankcase, reducing clearance to the floor, and in the case of some highly optimised engines, this would require a deeper sump and moving the whole engine up in the car.

Titanium is an attractive material, but as the saying goes, we need to 'pick our battles carefully' when we look to substitute steel for titanium.

Fig. 1 - Formula One crankshafts have small bearings to reduce friction and a low crankshaft centreline height. It is unlikely that using titanium for crankshafts would be considered

Written by Wayne Ward

One method of transferring oil between main bearings and crankpins is to have oil drillings running along the length of the crankshaft, with drillings from the surfaces of the crankpins and main bearings intersecting these radially. Where the oil drillings break out of the crankshaft, we need to plug these holes. A simple solution is to tap the ends of the holes and plug them with a threaded bung. This can be as simple as a grubscrew used with a thread-sealing compound or thread-locking compound. A variation on this is to machine the screw to have a plain portion on its 'nose' and to fit an O-ring to provide a positive seal. To provide some extra security, where access allows, these can be 'staked' in place.

The use of a machined press-fit plug is common, and there are a number of proprietary types of two-piece high-pressure sealing plugs that are also popular. These usually have an outer piece which is a transition fit in the hole provided, and an inner piece which is a force fit, expanding the outer when fitted, providing a heavy press-fit without any chance of damaging the hole in the crankshaft.

While 'nose-fed' cranks do not represent a large proportion of race crankshafts, they are popularly used for short-stroke applications. In a nose-fed crankshaft, oil is introduced at the nose of the crankshaft on the crank axis, and is transferred along the crankshaft via drillings that are machined from the bevels next to each crankpin. Again, there is a need for these drillings to be reliably sealed. Such drillings can be responsible for locally high stresses and so real care is taken to design something that doesn't cause very high stress. Again though, threaded plugs or interference-fit plugs can be used, and where this is the case, it is common to stake them in place and use a sealing/retaining compound.

However, threads can be responsible for high levels of stress concentration; special taps that form a controlled radius at the major diameter of the internal thread are a good way to minimise the stress concentration factor. A better way though is to dispense with threaded plugs and look to another method of plug retention. If a plug can be located against a stop, a retaining clip can be used, and a groove to accept a round wire clip, similar to a piston-pin circlip, is a good way of keeping the stress concentration to a minimum.

Fig. 1 - The design of a reliable sealing method for the oil drillings in this nose-fed crankshaft needs careful thought

Written by Wayne Ward

If we are to take fullest advantage of the technique of replacing steel with tungsten, we need to replace the maximum mass - or, more accurately, reproduce the optimised counterweight moment by using tungsten. Tungsten achieves the same moment with lower mass and inertia. To summarise this 'simple' method, we bolt a big chunk of tungsten onto the crankshaft, with the bolts pointing radially inward, or more commonly parallel to each other and in the same direction as the mass centre of the counterweight acts.

The point about this method that makes it less widely used than we might expect is that it is one of the more risky methods by which the material can be attached. However, in light of the more positive reasons above, we should not be surprised to find that, in those rare production engines that use tungsten counterweighting, this is the chosen method - precisely for reasons of effectiveness and economy.

The reason that this method is risky is due to the high engine speeds in many motor racing series. The force acting on the dense metal mass at any given engine speed is proportional to its mass, the distance between the crankshaft axis and the centre of mass of the dense metal part, and the square of the rotational speed of the crankshaft.

Such are the forces involved that it can be necessary to use some pretty special fasteners. It is not possible to measure the stretch on these critical fasteners as we often prefer to with con rod bolts, whether or not the bolts are 'off-the-shelf' or a bespoke creation. Therefore we need to be sure that, taking into account the significant uncertainties involved in bolt tightening using conventional tools, we can preload the bolts to the required amount, giving a sufficient factor of safety against joint separation. For example, it may be necessary to commit to testing a number of instrumented representative assemblies in order to understand the torque-preload relationship, if this is the method to be used.

The risks of a mass becoming loose are significant - it will almost certainly puncture the engine, and possibly the car too, and land somewhere it was not intended to. If you are unlucky enough for this to happen, the damage to the engine can be extreme, ranging from comparatively 'mild' punctured casings to serious damage to many engine internals and castings.

This method should be used only where the forces are well understood, safety factors are proven and confidence exists in the design and manufacture of the fasteners and in their tightening.

Fig. 1 - This production engine from VW uses heavy metal counterweights on its crankshaft

Written by Wayne Ward

There is another popular option though, which is to install threaded inserts either radially inwards or in that general direction, i.e. parallel to a line joining the counterweight mass centre and the closest point on the crank axis. In a thick crankshaft counterweight, it is easy to imagine a series of holes, drilled and tapped radially inwards from the outside radius of the crankshaft. Into each of these can be screwed a tungsten insert provided with the appropriate thread.

The idea is very simple, but there are considerable forces acting radially outwards on the tungsten due to its mass, so there are a number of important considerations to account for in terms of component stresses before deciding that this method is the one for your application.

First, the engineer needs to be satisfied that the insert isn't going to come undone, as some troublesome fasteners tend to do if they lose preload. These will definitely be more trouble than the average loose bolt should they decide to exit the crank!

There are several chemical methods - thread-locking compounds, bearing-retaining compounds, high-temperature adhesives - that might be used to lock the insert in place. However, a popular method is to 'stake' the inserts in place by mechanically deforming the insert, crankshaft or both, often by making a series of marks using a hammer and centre punch around the periphery of the insert at the outside of the counterweight.

The photograph here shows a V10 Formula One crankshaft from several years ago that was designed to be used with radial-type inserts. The inserts to be used with this crankshaft (not installed) aren't true radial inserts, but instead run parallel to the plane through the crank axis and crankpin they are balancing. As can be seen from the photo, the inserts are short pieces, confined to the thick 'rim' of the counterweight. Another important consideration is to calculate whether the length of engaged thread is enough to prevent thread stripping when the insert is under maximum load.

There is the more conservative (and costly) option of welding a cap over the end of the hole, but this also serves to displace tungsten from the outer radius of the counterweight, where it is most effective.

An alternative to the radial-threaded tungsten insert is one that is installed in a hole through the counterweight drilled perpendicular to a plane containing the crank axis and the counterweight into which the insert is to be installed. This method allows for one or two tungsten bars to be inserted into threaded holes, and then to be staked or otherwise kept in place by welded caps again. This method is not as popular as the radial type insert.

Fig. 1 - With 50 tungsten inserts, the careful assembly of this V10 crankshaft is critical to reliability (Courtesy of Pankl)

Written by Wayne Ward

The advantages of adding a high-density material to a crankshaft - or rather, replacing a given volume of the base crankshaft material with something more dense - are not lost on production engine manufacturers, with at least one company having offered production vehicles with tungsten alloy fitted to the crankshafts of some premium models for a number of years.

The material most commonly used for high-density counterweights is tungsten - or, to be more precise, one of a number of tungsten-based alloys. These generally contain nickel and other elements such as copper or iron. The density of such alloys lies in the 17-18g/cm3 range, more than twice that of steels.

A number of companies we questioned for the RET article said they have produced crankshafts incorporating high-density alloys, fitting them using a number of methods. The most commonly preferred method is based on cylindrical tungsten inserts press-fitted into holes machined into the steel counterweights of the crankshaft; these holes run parallel with the crankshaft axis.

This method has several points which make it a favourite:

1) The machining required to accept the inserts is made up of a number of simple cylindrical bores, and so can be done economically.
2) The inserts themselves are simple cylindrical pieces, easily made on conventional lathes
3) The method of attachment creates a high frictional force between the insert and the crankshaft, and in the direction that the insert is installed the forces which would act to overcome this friction are small. A large factor of safety against loss of the material is therefore easily achieved.

Where crank suppliers are making crankshafts that incorporate high-density materials, this method is least likely to cause problems. One company questioned said that, when fitting tungsten to a crankshaft by any other method, this is always done to the customer's design and specifications, and clearly at the customer's own risk. The photo here shows a crankshaft with tungsten added by this method.

The question of the material being able to break when put under the combined action of centripetal and press-fitting forces is reasonably easily dealt with by carrying out some basic calculations. The method of fitting places no great demands on the tungsten insert itself, so the alloy chosen for the inserts in this method does not need especially high tensile strength, thus giving another cost saving. High-strength material, which is generally used for military purposes as armour penetrators, commands a premium over normal ballast materials.

There are a number of other methods which are used for adding tungsten, and these will be examined at a later date.

Fig. 1 - This Peugeot Formula One crankshaft has tungsten inserted parallel with the crankshaft axis (Courtesy of Vitesse Engineering Services)

Written by Wayne Ward

For a substantial proportion of certain types of engine, however, it is typical to find a crankshaft assembled from a number of pieces. This is generally the case where needle-roller bearings are used for the big-end bearing.

The types of engine that commonly use this method of construction are either single-cylinder four-stroke engines such as we find in motocross machinery, or two-stroke engines. Two-stroke engines find common use in motorcycle racing - although sadly even the 125 cc Grand Prix class is soon to disappear - and karting.

The crankshaft is assembled with the bearing and con rod in place on the crankpin, with the con rod being of the solid type with no split across the big end, allowing it to be clamped around the crankpin of a 'conventional' crankshaft.

In conventional two-stroke race engines, crankcase lubrication is looked after by oil that is either pre-mixed with the fuel or metered in some proportion to the fuel by a pump from a tank. After it has fulfilled its lubrication task, the oil passes into the combustion chamber where it is combusted with the fuel. Given the marginal lubrication, the needle roller is a natural choice for such engines.

Single-cylinder four-stroke engines can also dispense with a complicated crank oiling system by using the needle-roller big-end bearing. There is some evidence that needle-roller bearings can offer advantages in terms of friction, and in the V10 era some teams are known to have tested if not raced engines with needle-roller main bearings using ceramic rolling elements. This is borne out by the fact that the regulations were specifically altered a few years ago to outlaw ceramic bearing elements by mandating an iron-based alloy for rolling bearing elements.

The crankshaft for a single-cylinder engine can typically be made in two or three pieces. The two-piece crankshaft has one main bearing, a crank 'cheek' and the crankpin as a single piece, and a second piece which is a main bearing journal and a crank cheek. The crankpin is a heavy interference fit into the cheek of the second piece. A three-piece single-cylinder crank has a separate crankpin.

Multi-cylinder crankshafts can be assembled in this way and a great many two- and four-cylinder two-stroke engines were built like this.

Problems can arise in the incorrect alignment of the different pieces, causing strain on the crank when being fitted and excessive loads on the bearings. The depth of the press fit needs to be accurately controlled in order to prevent the con rod becoming trapped between the two crank 'halves'.

To be assembled correctly, proper tooling should be used. Furthermore, the amount of interference needs to be tightly controlled. Insufficient interference can see the assembly twist in use, and it is common to see racing cranks that have been welded to prevent this.

Fig. 1 - This crankshaft assembly from a four-stroke motocross machine has a split crankshaft and a single-piece con rod (Courtesy of TSRMX.co.uk)

Written by Wayne Ward

With very few exceptions, four-stroke engines use camshafts to open poppet valves, and the cams need to be driven at a fixed speed ratio to the crankshaft, and timed to the motion of the piston very precisely. While it would perhaps be convenient to do so, nobody drives the cams electrically, and so a mechanical connection between the crankshaft and the camshaft(s) is necessary.

In the case of belt-driven cams, there is the simple solution of bringing a plain shaft out of the engine through a seal and fitting a pulley. For high-speed overhead cam race engines, however, the general preference is to drive the cams via gears or chains. While chains are still in wide use, gears have largely supplanted them for bespoke applications but, in both cases, we need to have a driving gear or sprocket on the crankshaft.

It is of course possible to cut a gear or sprocket integral with the crankshaft, and in many cases this is exactly what is done. In the case of motorcycles, gears are often cut integral with the crank in production engines, and when these are used as the basis for a racing engine there is no choice but to follow the production manufacturing methods.

In some other bespoke engines, the cam and auxiliary drive gear is cut integrally on one end of the crankshaft. There are advantages to this - fewer components are involved, no risk of the gear becoming loose and it is ultimately a lighter and neater solution. But the manufacture of the crankshaft is necessarily more complex, and the parts are therefore more expensive and take longer to manufacture.

In many applications, a gear or sprocket is separately made and then fixed to the finished crankshaft. This is quite common, but the number of components is increased, with often multiple fasteners and dowels locating and holding the gear in position.

This method, however, has its advantages. The gear, if worn, can be easily replaced and there is the option of making the gear from a different material from the crankshaft. Nitride-hardened steels are common for racing crankshafts, but some people have reservations about nitrided gears, preferring to use a carburised gear instead.

I spoke to UK crankshaft manufacturer Arrow Precision regarding integral gears, and asked managing director Ian Arnold about the popularity of integral gears. "About 20% of our race cranks have integral gear or sprockets," he says.

It is clear that, from a design point of view, this is an attractive solution, but in terms of manufacture, the extra complexity costs time and, of course, money. There is about an extra week's worth of work in having an integral gear or sprocket cut, and the extra cost depends to a large extent on the number of parts being made. For a one-off special, the increase can be as much as 10%.

Fig. 1 - This nitrided V6 crankshaft has an integral gear (Courtesy of Arrow Precision)

Written by Wayne Ward

In terms of geometry, there are two types of crankshaft design features that most often prove to be the initiation site of fatigue damage. The first are the fillet radii which we find on the crankpins and main bearing journals. The oil hole exits on the crankpins, and main bearing journals are also common sites for fatigue damage initiation, so it is these that we shall look at a little more closely in this article.

The geometry of the oil-hole exit and the level of surface stress here is critical to providing an adequate safety factor against failure. The crankshaft is stressed in both torsion and bending. In both of these cases the stresses occur at the extreme fibre of the material.

So, if we imagine a beam in bending with a vertical load applied somewhere along its length, we can expect to find that the maximum tensile or compressive stresses will be found at the top and bottom of the beam. In terms of a cylinder in pure torsion, the maximum shear stress occurs at the surface.

Other than simplicity, there is little that could be recommended in simply drilling the oil hole in the correct position. By doing this, a sharp edge would be provided at the point of maximum shear stress and, owing to the resulting stress concentration due to the hole, this point will have a very high level of stress. It is common therefore to provide a nice radius on the oil-hole exit, and this can be done carefully by hand or as part of the machining of the crankshaft.

It is important to select the appropriate stage during manufacture at which to carry out this procedure. If the material is removed after nitride hardening, the surface of the crankshaft will be robbed of its protective compressive residual stresses. It is better then to provide the radius before surface hardening and then to polish the area at a later stage, removing very little of the layer imbued with the compressive stresses.

There are other solutions to the radius provided on oil-hole exits, and on production crankshafts we commonly find a simple chamfer, but this is not as effective as a properly formed radius. Other simple milling operations are used for racing crankshafts which are thought by those using them to be as effective as the radius on an oil-hole exit, but which are not as costly in terms of manufacture.

Textbooks inform us that the maximum shear stresses occur at an angle of 45º to the axis of a cylinder loaded in torsion. We can therefore resort to more elaborate schemes for removing material from the oil-hole exit, and these include material being removed at an angle of 45º to the crankpin or main journal axes.

Fig. 1 - Poorly finished oil-hole exits can lead to early crankshaft failure

Written by Wayne Ward

When feeding the connecting rod's big-end bearings with oil in the 'conventional' manner of pumping in oil via the main bearing journals, we need to provide a groove in the bearing shell so that the big end is fed continuously and doesn't suffer from having an intermittent supply of lubrication. Therefore, in making space for the groove in the bearing shells, the bearing becomes wider.

The implication of this may be that, where loads are high and space is limited, the main bearing fillet radii might not be as generous as we would like. Where no groove is needed a narrower bearing can be made to support the service loads, thus leaving room for generous fillet radii.

These more generous fillet radii lead to a lower stress concentration factor and therefore a greater factor of safety against fatigue failure in this area of the crankshaft. Arai's work which is referenced in Peterson's Stress Concentration Factors* shows the effect of increasing fillet radius for a simple case over a range of crankshaft overlap ratios.

Feeding the crankshaft via the nose requires a sealing arrangement on the crankshaft axis at the opposite end to the output. The oil needs to be supplied on the crankshaft axis, and this can be via a static tube poking into the crankshaft nose or by an extension of the crankshaft poking into a static part of the oil supply. In either case, a seal must be provided.

In the first case, where a tube protrudes into the crankshaft nose, the seal will be in the crankshaft and rotating with it; in the second case the seal is static. In the first case we need to ensure that, at maximum operating speed, the load on the seal lip is not relieved by centripetal forces to the extent that the seal no longer works as expected.

The effect of this might only be to spill excessive oil into the crankcase, but equally it might temporarily starve the crankshaft of oil. Given that this is most likely to happen at high engines speeds, it is critical to ensure sufficient sealing lip loads in all circumstances. This will clearly be more of a concern when the engine speed is high and the seal is large. We might therefore conclude that it would be better to have a static seal, although both types have been used successfully in recent racing engines.

Although centripetal forces might concern us in this instance, they can be used to good effect in improving the quality of the oil flow in the crankshaft, and in the next article I will look at how we might use these forces to our advantage.

* W.D. Pilkey and D.F. Pilkey, Peterson's Stress Concentration Factors, 3rd edition, Wiley 2008, ISBN: 978-0-470-04824-5, see page 451 and relevant text

Fig. 1 - This F1 crankshaft has no evidence of any oil holes on the main bearings, and is of the nose-feed type

Written by Wayne Ward

The concept of the 'nose-feed crankshaft' is not a new one, being used in some production engines, and having been successfully employed on the Rolls-Royce Merlin engine. The idea is a simple one; oil is pumped into one end of the crankshaft (normally called the 'nose' end, hence the name) and the pressure in the oil system forces it along the crankshaft through a continuous gallery. There are many forms that this gallery can take, but a great many such crankshafts employ angled drillings which enter the web cheeks at an angle in the side of the crankpin and continue through to the crankshaft centreline where they meet a drilling from an adjacent crankpin.

There is a feeling that the oil pressure in the crank oiling lubrication circuit can be maintained at a lower level than with a crankshaft being fed via the main bearing, owing to the fact that there are no centrifugal forces to overcome in order to move the oil initially toward the crankshaft centreline, and no viscous forces in the main bearing opposing the flow of oil into the oil hole. In running a lower oil pressure in the crank oiling circuit, there can be a number of gains to be had in terms of reduced friction from various sources.

However, we have to be mindful of other factors when deciding to provide oiling by this method. We need to be sure that, given our new lower oil circuit pressure and the fact that we are oiling each and every crankpin from a single source, we can provide sufficient flow to the crank such that the furthest pin from the oil entry to the crankshaft receives sufficient flow to satisfy its lubrication needs. Pressure losses in the galleries can be significant and need to be taken account of in this regard. Given this concern, we also need to decide if we continue to lubricate the main bearings from the main oil gallery, or via the flow that we have provided through the crankshaft. Each solution has its own merits and pitfalls. The accompanying picture of a V10 Formula One crankshaft has no sign of oil holes on the main bearing diameters. We can therefore presume that it is a nose-feed type, as we might reasonably expect.

There are certain advantages, besides lower oil pressure to the nose-feed crankshaft; one of these is that we can take advantage of the centrifugal forces present to help increase the quality of the oil flow by providing both scope for de-aeration of the oil flow and filtration.

Fig. 1 - Echoing contemporary design trends, this V10 F1 crankshaft is of the nose-feed type.

Written by Wayne Ward

The crankshaft which I mentioned, from the Mercedes SSK, was physically very large and of sufficiently large throw to allow much of the material removal for lightweight and mitigation of stress concentration to be removed by means of machining, although it is almost certain that the 'blank' or 'rough' crankshaft was originally cast.

Certainly, we would struggle to provide such elegant design features in many of our modern racing engines, especially where short-stroke designs have been favoured in the search for ever-increasing engine speeds as we chase higher output. The room for machine tool access is simply not available, and where it is available, we may struggle owing to lack of rigidity in the tooling. Some other modern methods of material removal may prove to be a little more fruitful in this respect, although practices such as spark erosion still require a certain amount of access and, whilst they don't suffer the same problems as machining due to lack of rigidity, they can have their own problems.

So, do we have to discount Matinaglia's good advice, and the results of his excellent research and experiment? Well, that would depend on our available budget, our knowledge of manufacturing processes, and our willingness to think of the possibilities. In the press rumours abound that some Formula One engine manufacturers had looked at hollow crankshafts several years ago, and the anecdotal evidence is strong enough to believe that it has more than a grain of truth to it. The gains to be had are sufficiently large in terms of reduced inertia (if reduced inertia is indeed your goal) and mass that it is almost inconceivable to think that this hadn't been considered. These gains alone are clearly worthwhile, but if we consider that these low mass components may come with an attendant benefit of lower critical stress, it makes the case for such research compelling.

In my previous employment with one Formula One engine manufacturer I made proposals about the manufacture of hollow crankshafts, and I had arrived at a workable solution in terms of manufacturing techniques before the project was shelved, owing to forthcoming regulation changes and requirements for homologation of engine designs. There are a number of feasible solutions to providing a hollow crankshaft using a number of commonly-used industrial techniques, although many of these techniques aren't widely used in motorsport.

At the moment, we commonly hollow crankpins, both to save mass and to mitigate stress concentration (as seen in the accompanying picture of the Peugeot V10 Formula One crankshaft), however, we rarely see anything which goes further toward the goal of a hollow crankshaft.

However, for those working in an environment where rules don't shackle engineering excessively, with a development budget, an enquiring and creative mind, and a good knowledge of manufacturing techniques, the possibility remains to create a crankshaft with hollow mains and crankpins. Perhaps Matinaglia's work for the mighty Sulzer won't remain only as an interesting technical dead-end as far as racing is concerned.

Fig. 1 - A hollow crankshaft - a technical dead-end?

Written by Wayne Ward

Once this strategy of introducing compressive residual stress has been employed, if you still experience crankshaft failure, the path that you might reasonably follow is to provide more material, or more generous fillet radii. However, design changes in this direction are generally quite limited owing to space restrictions and the geometry of other components and this is especially true of highly optimised racing engines, which we design to be small, light and compact where possible.

Something that we can do to improve the situation is to remove material from areas of the crankshaft by means of machining operations. Probably the best-known works on the subject are the 1943 Sulzer Technical Review paper by Martinaglia written in English and the paper in German by Lurenbaum, both of which are referenced in volume 2 of "The Internal Combustion Engine in Theory and Practice" by C.F. Taylor. These papers detail changes made to the design of a crankshaft by removal or redistribution of material from the main bearing shafts, the crankpins and the webs and the results of torsional fatigue strength tests on the components. Today we would be able to use finite element methods to good advantage to get the results using a fraction of the time and money. Some of the material removal details that they describe lend themselves to large cast crankshafts, particularly where crankpins and main bearings are made hollow with undercut bores. These design features would be really tricky and probably impossible to implement in many of today's racing engines which rely on the structural durability afforded by machining from a high quality billet of steel. The results of the tests are certainly very impressive, with an improvement in torsional fatigue strength of over 200% compared to the basic 'reference' design. For those familiar with adding more material and therefore more weight to increase fatigue strength, the concept of material removal to achieve the same end might not seem intuitive. However, it is a proven technique as demonstrated by the results of the experiments referred to above, and the general method of material removal to reduce stress concentration factors is well known and well documented.

I have been lucky enough to see a machined crankshaft destined for a Grand Prix motor racing engine which has many of the design features described in the references. This clever piece of design work, destined for the famous pre-war Mercedes Benz SSK (car shown in picture) originated on a drawing board in Europe several decades ago, before the idea of every design engineer having a 3D CAD system and his own personal computer had been thought of.

Fig. 1 - 1930 Mercedes-Benz SSK Roadster.

Written by Wayne Ward

If we imagine the case of a simple cylinder loaded in torsion, the shear stress increases linearly from zero at the centre of the bar, to a maximum at the surface. The magnitude of this stress is very easily calculated using the equations found in a university level textbook on engineering. The stress can be calculated for a wealth of more complex cases by referring to a book such as Roark’si.

This stress needs to be balanced against the significant strength of the material depending upon its use. In terms of a crankshaft, the significant strength is fatigue strength, which we must calculate for the case in hand based on the material, size of component, surface finish, service temperature and the correct stress concentration factor.

Returning to the simple cylindrical case, how do we factor the compressive stress into our information? We ‘add’ the stress due to our to our service load to give a level of stress which takes into account the stress field in the part prior to service loads. The stress field due to any process which puts compression into the surface layers necessarily puts the core into tension, and the ratio of the areas of the compressively stressed surface layers to the core area dictates the level of tension in the core. For a cylinder of significant dimensions relative to the stressed surface depth loaded in torsion, we don’t need to be worried too much about this effect, but should be wary when high tensile loads are applied. The high level of compressive surface stress means that the high stress due to torsion is mitigated to a large extent, and the fatigue life is markedly increased as a result.

In the case of crankshafts, we often find failures occur at the junction of the crankpins or main bearings with the crank webs. Generously proportioned fillets are essential, but not enough on their own without nitriding or additional processes. There may be circumstances where nitriding alone does not provide sufficient levels of compressive residual stress to allow a given design to work, and in these cases we need to look at extra processes to give us the advantage we need.

Fillet rolling is a process which has been widely used on production engine crankshafts for many years. In this process a tool, which generally takes the form of a ball or wheel is loaded against the surface of the fillet and rolls against the crankshaft as the component is rotated in a machine. A great advantage of this process is that it can be done with conventional machine tools provided that the machine bearings are sufficient to stand the tool load which can have a significant thrust component. However, purpose designed fillet rolling machines are available which are much better suited to the purpose.

Written by Wayne Ward.

Many years ago, crankshafts were designed with external oil gallery circuits made of thin metal tubes fastened to the crankshaft and rotating with it. These are only practical on large slow-speed engines, and would not be a realistic proposition on a modern racing engine. However, the large racing engines and piston aircraft engine of yesteryear have provided some ingenious methods for transporting oil to the connecting rods, and some of these methods are still in favour today. More of this a little later. First we will look at what has been general practice for many and, in a large number of cases, continues to be.

Common practice among road and racing crankshaft manufacturers has been to feed the main bearings with oil from a pressurised gallery. In order to feed the connecting rod, the pressurised oil on entering the main bearing, feeds into a central circumferential groove that commonly covers all 360 degrees of the bearing. From this groove containing pressurised oil, a drilling, or series of drillings takes the oil to the crankpins. This can often be done with a single compound angle drilling between the crankpin and the main bearing for each oil feed. However, some engine designers and some crankshaft manufacturers prefer to have axial drillings along the crankshaft, with drillings perpendicular to the crankshaft axis taking oil to these axial drillings from the main bearings and from the axial drillings to the crankpins.

Anyone who has studied crankshaft design, even briefly will know the importance of the oil holes to the fatigue life of a crankshaft. Along with main bearing and crankpin fillet radii, the oil holes are the prime causes of fatigue failure of crankshafts. Torsional fatigue failures of crankshafts often show cracks that have initiated at the oil hole exits, the condition of which is critical. It is absolutely essential to provide proper instructions about how these oil holes are to be treated, as left in the as-drilled condition, failure is likely to follow. At the very least, these should always be provided with a small radius to break any sharp corners, and this should be done before any surface treatment such as nitriding. After surface treatment, these radii should be polished to remove any imperfections or, in the case of nitrided parts, any ‘white layer’. It is important to radius the edges before nitriding so that the all-important residual compressive stresses are maximised in this critical area. If we were to radius the edges after nitriding, we would remove not only the material containing these stresses, but we would also have a more difficult task of producing these radii.

Written by Wayne Ward.

Often this is achieved through careful design of the crankshaft and the counterweights are incorporated into the crankshaft. What we aim to do in providing a counterweight is to achieve a certain moment relative to the crankshaft axis. Depending on the configuration of the crankshaft these may or may not be directly opposite the crankpin to which it is closest. The product of the mass of the counterweight and the distance of its centre of gravity from the crankshaft axis mean that the required moment can be provided with minimum mass by having the centre of gravity as far from the crankshaft axis as possible. However this strategy leads to increased inertia which causes loss of acceleration of the engine.

There is generally some restriction on the maximum radius of the counterweight imposed by adjacent components such as the base or walls of the lower crankcase or quite commonly, the combination of con rod length and the radius of the piston pin boss cause a restriction at bottom dead centre. Therefore we have to contain our counterweighting within the envelope defined by these restrictions. In many cases it remains feasible to integrate the necessary counterweighting with the crankshaft.

However, we may not be able to do this in all cases, or we may have a requirement to reduce inertia to an absolute minimum. There may be a further advantage in reducing the maximum counterweight radius to reduce the crankshaft centre height relative to the bottom of the engine. In some racing formulae this height dictates the gearbox mainshaft height and therefore has an impact on the rest of the car including aerodynamics. It is for this reason that Formula One engine suppliers invested lots of time and effort to reduce crankshaft centreline height a few years ago. Even before the development of the engine was frozen, there was a regulation introduced which imposed a minimum distance between the bottom of the engine and crankshaft axis.

In Formula One, in order to meet the required amount of counterweighting within the restrictions, it has not been possible for some time to use integrated counterweights within the same single component. It has been accepted practice in Formula One, and in a number of other applications to use heavy metal to provide the some of the counterweighting. There are now series production road vehicles using heavy metal counterweights in addition to some heavy goods vehicles.

By heavy metal we are most commonly, but not necessarily, referring to one of the commercially available tungsten alloys. These have densities which are far greater than that of steel, generally being in the range of 17,000 – 19,000 kg per cubic metre compared to approximately 7850 for steel. The FIA has again imposed a restriction in Formula One to prohibit the use of any engine materials with a density greater than 19,000 kg per cubic metre.

The options for fitting these separate pieces to the crankshaft are varied and can, by themselves, impose come other demands on the mechanical properties of the heavy metal weights and their associated components. If you are considering the use of heavy metal weights you must be sure that the method you use to secure these is reliable. The loss of a counterweight will not only damage the engine, it is a serious safety risk. Tungsten alloys are often used for armour penetration and it should be no surprise then to find that these parts can, if they become detached from the crankshaft exit the engine and the car in a pretty straight line, i.e. through the side of the engine and the bodywork or whatever else is in the way.

Written by Wayne Ward.

Before we start, we should note that nitriding, whilst overwhelmingly popular for racing crankshafts is not universal. There are various other treatments providing local improvements to the fabric of the crankshaft close to the surface which have been used and remain popular today. ‘Tufftriding’ or variants of this nitro-carburising treatment still have some proponents, and it does provide a wear-resistant surface. However, ‘Tufftriding’ is only a very shallow treatment. Induction hardening (or more rarely flame hardening) is applied to bearing surfaces and fillets and aims to increase the bulk hardness of a layer near the surface, locally increasing strength and hence wear-resistance and fatigue properties. Carburising, or case hardening, is now largely out of favour, although it has been favoured in the past, before nitriding had been invented, or whilst nitriding was at a less advanced stage of development. This process diffuses carbon into the surface of a low-carbon steel, increasing its hardenability. This can impart large increases in hardness, with considerable case-depth and an attendant increase in wear-resistance and fatigue strength. However, owing to the harden, quench and temper nature of this treatment, distortion can be a very serious problem.

Hence we come to nitriding, which is favoured by most of the manufacturers of racing crankshafts, especially for single-piece machined items. Nitriding aims to diffuse nitrogen into the surface of specially alloyed steels, rich in elements which are strong nitride-formers, most especially Chromium, Molybdenum and occasionally Aluminium, Vanadium and Tungsten. There are two main types of nitride hardening processes, namely gas-nitriding and plasma nitriding. There are similarities in the two processes, although the methods are very different. Both take place at around 500°C (900°F) and both are relatively long processes, often taking tens of hours and, in some cases, well over 100 hours. The relatively low processing temperature compared to carburising means that distortion should be much less pronounced.

Gas nitriding is the older process of the two and still the most popular in terms of numbers of parts treated. The parts are suspended in a furnace with an atmosphere which is rich in nitrogen (most often dissociated ammonia). Gas nitriding produces a hard friable layer, often called the white layer. This can be very wear resistant, but owing to its brittle nature can become detached and cause problems as it forms very hard debris. Moreover, fatigue cracks can form in this layer, hence the reason why crankshafts are generally ground after gas-nitriding. Removal of the white layer in fillets is very important, and some manufacturers test parts using a simple chemical test to specifically check that this has been achieved.

Plasma nitriding is generally carried out under low pressure (a few millibars being common) with the parts being electrically charged and forming a cathode. The glow seem emanating from the parts has led to this process often being called glow-discharge nitriding. The heat produced by the discharge is what allows the nitrogen to diffuse into the material. There are some treatments which have relatively poor ‘coverage’ in small grooves or holes, which has caused some concern for manufacturers of crankshafts. However, there are newer treatments which can overcome this lack of penetration and can therefore treat oil holes etc.

Both methods are used at the highest levels of motorsport, although for more mainstream applications, gas nitriding remains more widespread.

Written by Wayne Ward.

When we specify the material for the crankshaft, we need to be careful not only to specify the composition of the material, but also the level of mechanical properties that we can expect. In general, we do this by ordering the material with a given property or combination of properties within a range that we are happy with. We may deem that we want a certain level of ultimate tensile strength (UTS), or yield strength, or elongation etc. Particularly with reference to UTS, there is a very well understood relationship between this mechanical property and hardness. This allows us to specify a level of hardness for the material which has the advantage that the test can be carried out with (relatively) inexpensive equipment on any reasonably clean flat or round surface.

The method by which we achieve this desired level of hardness or strength is by the processes of hardening and tempering. Again, many of us here will understand this process in detail and specify this kind of treatment as a part of our everyday job. However, we will cover the basics of the process here without going into the ‘nitty-gritty’. There are lots of excellent textbooks on the subject, and we can cover it in later articles if there is any feedback or comments asking for this.

The hardening process involves getting the steel very hot indeed – each grade of steel will have a given temperature above which the entire structure is known as ‘austenite’. This is generally above 900 degrees C (>1650 degrees F) and depends on the grade of steel and the properties desired. In general, for most steels, austenite is not an equilibrium phase, and over time (in some cases a very long time) will change into something else. Austenite is a relatively soft phase and not many steels used in engine components use these types of steel (although poppet valves are a notable exception). However, in the controlled process of hardening, we ensure that this change happens by quenching the steel. Where the grade of steel requires that the temperature change is very rapid, the steel may be submerged in water, oil or a polymer quenchant. However, for some steels which are quite highly alloyed, the quenching can be done in air, which leads to less distortion. Generally, what is produced (or desired at least) is a phase called martensite.

Martensite is an equilibrium phase, but is extremely hard and brittle and requires a further process to make the steel usable as a crankshaft. This process is called tempering, and involves getting the steel moderately hot in the case of a crankshaft (somewhere between 500 and 700 degrees C (approximately 930 – 1300 degrees F) for a nitriding steel depending on the exact grade and the level of strength desired). Steels for other applications will temper at higher or much lower temperatures (for example, common gear steels are tempered at around 150 degrees C). As for stabilisation (stress-relief), tempering depends on a combination of time and temperature. If the temperature in service (or during the rest of the manufacturing process) exceeds the tempering temperature, there is a danger of loss of mechanical properties.

Nitride hardening can affect the manufacture of crankshafts in this way. Lower tempering temperatures, in general, leads to higher hardness and strength and lower ductility. If one was to specify a high hardness, requiring a low tempering temperature, it may be that the core material would be softened by the nitriding treatment. This effect can be mitigated by careful choice of alloy if high core strength is felt to be a requirement.

However, it is quite common for crankshaft blanks to be ordered pre-hardened and tempered to the desired core hardness – careful manufacturing minimises the amount of stress-relief required, and proper specification of the nitriding process poses no danger of diminishing the required properties.

Next month we will look at what options we have for nitriding, how these processes are done and what the results are.

Written by Wayne Ward.

As we saw last month, the materials in current Formula One use are mainly nitriding steels. By nitriding steels, we mean those with a composition containing elements which are strong nitride-formers, chief among these being chromium, aluminium, molybdenum and vanadium.

As you can see from looking at the tables in last month’s article, both of the steels shown contained chromium and molybdenum, with 3S132 also containing some vanadium. EN41 is a steel which contains significant amounts of aluminium (>1%), although this doesn’t seem a popular choice for crankshafts.

However, before we come to the important subject of nitriding, we will look at the other heat-treatment processes which are used prior to nitriding. In general, the racing crankshaft, certainly in Formula One, is machined from a solid cylindrical billet of material. This material, unless specifically requested to be delivered in a hardened condition, will be delivered with a standard heat-treatment from the mill, and will typically be in a softer condition than you will want to specify for the core hardness of a racing crankshaft. This is not seen to be a problem as there will be several stages of heat-treatment before the nitriding treatment is undertaken. After initial rough machining, the crankshaft has had a lot of material quickly removed and there are machining stresses in the steel, which can lead to unexpected distortion further along the production route. The advantages of having a softer material at this stage are (a) the material removal can often be quicker and (b) owing to the lower shear strength of the material, the machining stresses are lower. The traditional method of crankshaft manufacture, by offset turning using big-stiff lathes, was quite a violent method of rough machining thereby putting a lot of stress into the material. Whilst it is impressive to see the old methods used, the shallower cuts and higher speed machining methods used on CNC machining centres mean less machining stress.

At this stage, with a reasonable amount of stock left on the rough crankshaft, it undergoes a stabilisation treatment, often called stress-relieving or stress-relief annealing. This is done at a reasonably high temperature, and is made quicker by using increased temperatures. The relationship between the degree of stress-relief and time / temperature is a complex one; there is a parameter which contains the time and temperature variables, and increasing the parameter leads to a greater degree of stress relief. Let us, for example, say that a given steel will have 90% of its stress relieved by five hours at 600°C. The same steel would need almost 100 days at 500°C to achieve the same effect. 600°C is a typical temperature for this type of treatment, and is high enough to cause significant temper-softening in materials which are hard enough to use for racing crankshafts. It would therefore be of little use to order pre-hardened material from the mill. Sometimes there is more than one stabilisation treatment used before the hardening and tempering treatment which will define the final core hardness of the crankshaft.

Next month we shall continue to examine the heat-treatment processes to which a crankshaft is subjected before it is finished.

Written by Wayne Ward.

The bar-stock used to make an Formula One crankshaft is not necessarily much different from that used in other formulae; indeed some manufacturers in ALMS / LMES use the same basic composition of steel. As was also mentioned last month, there are several different grades to choose from for a given composition, depending on the 'cleanliness' of the material. By cleanliness, people often mean the percentage of sulphur and phosphorous in the material. Sulphur is commonly added to basic engineering steels in order to improve machinability. However, it can lead to problems by combining with metallic elements to produce large inclusions of sulphides in the ingot. These can then, by the process of rolling form long axial defects (also called ‘stringers’) in the material and this is one of the reasons why engine manufacturers steer clear of the basic 'EN' grades of steel in the UK. If you look at the composition of steels, it is common to see the percentage of sulphur and phosphorous quoted for engineering grades, and often a combined figure for the two.

If we look at the percentage of these two important elements in a steel composition commonly used for racing crankshafts in engineering and aerospace grades, we can see the difference in the quantities of these important elements.

Table 1: Composition of engineering and aerospace grades of a crankshaft steel. Maximum Sulphur and Phosphorous contents shown in bold.

As can be seen, the basic compositions of the steels are very similar, and 3S132 also conforms to the composition for EN40C.

Steel manufacturers quote these figures because of their well-known effect on the dynamic mechanical properties of the material. They have very little effect on other qualities of the steel such as hardenability or corrosion resistance for example.

This drive for cleanliness in the steel not only means that we are likely to see less gross axial inclusions (also called 'stringers'), but that we also get a very useful increase in fatigue strength. One European steel manufacturer presents data for nitrided steel which shows over 30 percent improvement between the basic and double-remelted grades of steel at 107 cycles. This remarkable figure means that a design engineer, armed with such data, can design to very much higher allowable stresses, or that he has the comfort of a higher safety factor against fatigue failure.

The main effect on fatigue life is the effect of the nitriding treatment itself. This surface hardening treatment not only produces a very hard, wear-resistant surface, it imparts a high level of compressive stress on the crankshaft. This compressive stress greatly increases the fatigue strength of the component, and is especially effective where stress concentrations are concerned. For the crankshaft designer the main stress concentrations are oil holes and fillet radii. References to the benefits of nitriding with regard to fatigue can be found in metal fatigue textbooks published over a number of decades. We shall look further into the available nitriding processes in a later article and the various solutions which people use for their racing crankshafts.

Written by Wayne Ward.

The banning of ceramic bearing elements threatened to deal the roller bearing a fatal blow in Formula One, but there are new steels available today which mean that this remains a viable proposition.Torsional vibration is another problem to be dealt with by the racing engine designer. This can be calculated in the first instance by hand calculations as can be found in the early books on the subject published decades ago. Ker Wilson’s five-volume epic on the subject has some good material for the modern engineer to digest. Many years ago detailed calculations for torsional fundamental frequencies were a legal requirement for the many people involved in engine design for marine applications, and it is thanks to this that literature is available to help the engineer who does not have expensive finite-element methods to provide his answers.The question of lubrication can be dealt with by either feeding the crankshaft journals from the main oil gallery, and subsequently to the crankpins via angled drillings. The alternative, which is a favourite in Formula One, is to use a nose-fed lubrication system. This method was used on the Second World War Merlin aircraft engine, and has the advantage of requiring lower oil pressure than the gallery fed system as the oil does not have to overcome centrifugal forces in order to travel radially inwards to longitudinal galleries on the crankshaft.

A V10 Formula One engine manufacturer achieved good results from a gallery feed crankshaft by employing some clever design features which allowed them to run quite a low main gallery pressure, however, this engine was never raced. The choice of location, size and shape of lubrication drillings has a profound effect on crankshaft stress, and failing to take this into account leads to many failures in service.The vast majority of crankshafts are made of nitriding steel grades of varying quality, and then nitride hardened to give a hard, fatigue resistant surface. Opinions vary as to the optimum depth of the nitrided case, and the range of depths employed in Formula One varies from around 0.4mm to 0.9mm. The specified hardness of the core also varies widely. Nitride hardening is felt to be most effective on steels of higher core hardness, but many engine manufacturers feel that core hardness does not need to be very high. One final note about material choice concerns the quality of the steel, or its ‘cleanliness’. Fatigue studies by a well-known steel manufacturer show that the fatigue strength of a nitrided steel is strongly influenced by how ‘clean’ it is, with the more expensive double-remelted grades showing a very clear advantage.

Owing to the restrictive nature of the Formula One regulations, engines had begun to converge toward today’s maximum allowed bore of 98mm, especially before the decreasing rev limits recently introduced. The 98mm bore gives a very short stroke of less than 40mm, and it is thought that the majority of manufacturers use this bore size or something very close to it.The matter of counter weighting in Formula One is dealt with using tungsten alloy pieces which are fastened to the crankshaft using threaded fasteners. Some previous designs had inserts which were inserted into the steel crankshaft, but this does not give the same benefit in terms of reducing dimensions and inertia. Depleted uranium has been considered but is now outlawed under the maximum density regulation: “No material with a density exceeding 19,000kg/m3 may be assembled to the crankshaft.”Coatings are sometimes used on Formula One crankshafts with DLC being popular. This is not universal however, as some manufacturers had not developed this technology before the engine homologation regulations came into force.The challenges for the crankshaft designer are many and varied, but also interesting. He must provide a part which is sufficiently strong and stiff whilst keeping mass, inertia and friction in mind.