Higher-revving engines are a challenge for the valvetrain, and the overhead cam engines used in these smaller engines are much less problematic than their pushrod counterparts. The connection between the cam lobe and the head of the valve is far stiffer, and has fewer components, so controlling the valve at high speed is easier.

Race engines are even more of a challenge. As engineers, we want to open the valves further and for them to open and close very rapidly, and we want our engines to operate at higher speeds than an engine from a typical passenger vehicle. So, where the camshafts in a passenger car are driven by a belt – or by chain on a motorcycle – most bespoke race engines use a geartrain to drive the camshafts. The advantages of gears over belts and chains are increased stiffness and greater accuracy, particularly over extended periods of time. Belts and chains stretch over time; gears do not.

Cam drive gears are made from steel, but their loading is complex as the torque transmitted is far from uniform through the engine’s operating cycle. Despite this, some of the gears in a race engine geartrain can have impressively small face widths (the term given to the width of the gear tooth between the faces of the gear at the pitch circle).

Moving inwards from the ‘working’ portion of the tooth, the root radius needs to be considered. While these radii are usually generated by the action of the gear cutter, some people advocate producing a more optimal form in this area to reduce stress concentrations.

It is common to reduce the width of the gear between the hub and the teeth, but the designer needs to be careful to provide sufficient ‘depth’ of material to support the teeth. The narrowed area, which reduces the mass and inertia of the gear, is often provided with holes and slots to further reduce mass, but if too many holes are used then cracks can develop. There may also be a penalty in terms of oil drag and shear from using holes rather than simply making the web between the hub and rim of the gear thinner.

The number of teeth chosen for the gears is dictated to some extent by the 2:1 speed ratio required between crankshaft and camshafts, but the choice is made so that the same highly loaded teeth are not always in contact with each other. So, we find that prime numbers are often used here.

It is not always possible to find gears that will fit in the allowable space and operate at the exact centre distances required. In such cases, it is necessary to adjust the profile of the gears in a design process called ‘profile shifting’, which allows a gear to operate properly but with a slightly larger or smaller diameter than its tooth size and number of teeth suggest. Profile shifting needs to be done carefully, however, so as not to compromise the strength of one or both gears.

Written by Wayne Ward

What we need is a material with temperature resistance, high thermal conductivity, thermal expansion which is close to that of the cylinder head and good mechanical properties, especially fatigue resistance. Often, as engineers, we find ourselves ‘painted into a corner’ in one of these areas. We may need higher thermal conductivity but can only find it in a material with lower strength, or we may find another characteristic that we like but find that the thermal expansion coefficient puts the cylinder head at risk due to high stresses. In order to solve another problem, we may need to lower stresses by widening the valve-to-seat contact, which can harm performance.

The traditional materials have been copper alloys, namely different types of bronzes. In recent years, beryllium-copper alloys have become a firm favourite; often we find a mix of two alloys in the same engine with one alloy – usually a high-strength type – specified for the inlet seats, and a different, lower strength but high thermal conductivity alloy used for the exhaust seats.

Sometimes, however, we find that we can’t get the seat material to live for long enough. The long-life engines, which once comprised only endurance racers but nowadays include series such as Formula One and MotoGP, may need a different class of material altogether. Powder-metal valve seats are based on a ‘pre-form’, which consists of a sintered metal (often steel) that is deliberately made porous. This is then infiltrated with an alloy that lends the structure stiffness and thermal conductivity. The infiltration alloy is generally a copper or bronze.

The thermal conductivity of the powder-metal seat material, compared to one of the normal bronze candidates, is compromised owing to the amount of material now displaced by the pre-form (which is of lower conductivity). However, because a significant proportion of its volume consists of a high-strength material, with increased heat resistance, the combined properties of the material mean its strength is markedly increased, as is its resistance to deform gradually at the higher end of the operating temperature range.

Such materials are essentially a type of metal-matrix composite (MMC) but in powder-metal valve seat materials, the reinforcement is in much greater proportion and of much greater size than in a traditional MMC. It is an unusual material in that the strengthening reinforcement and the matrix (if we wish to consider the softer, weaker material as the matrix) are both continuous structures, whereas traditional powder reinforcements in MMCs are designed to be discontinuous and evenly distributed within the matrix.

Written by Wayne Ward

Like me, you may be surprised to learn that some companies have combined the function of these components, the merits of this being lower cost, improved reliability, lower mass and reduced complexity. The component resulting from this unusual marriage, known as a lash lock, is actually, as is the case with any conventional valve lock, a pair of components that allow the locks to engage with a groove cut into the valve stem.

Where the lash cap and valve locks are separate components, the lash cap contacts the top of the valve stem, but in the case of lash locks, this is not always the case. In order to locate properly in the stem groove, the lash cap part of the component must remain clear of the valve in all cases. Allowance therefore needs to be made for the case where the groove is in its most distant position relative to the end of the stem. The load transfer is then not via the top of the valve stem but through the wall of the lash lock.

The result of having a split component is that the top surface of the lash lock is not continuous and flat – there is a cut that crosses it. Thankfully, on the components I have seen pictures of, the cut is not a simple straight line but is in an ‘S’ shape, meaning that the end of any rocker will not come to an abrupt ‘valley’ as it traverses the top of the lash lock. Such a straight cut might have an upsetting influence on valvetrain dynamics if the rocker tip were to ‘drop’, however slightly, into the groove and have to climb out again. If we rely on having a nominally flat top to a conventional lash cap for reasons of lubrication – to sustain an oil film in sliding or rolling motion – the gap in the lash lock’s top surface will adversely affect this.

If these lash locks were used in overhead camshaft (OHC) engines with inverted bucket followers, this would not be a concern. The lash cap, where it is used to change the valve clearances, is a very simple component to change for one of the correct thickness, and certainly for OHC engines this is the way valve clearances are adjusted. However, for a lash lock, the act of changing the component is far more complex, as we have to compress the valve spring as we would when installing or removing the valve locks.

It remains to be seen whether such components take an increasing share of the market from conventional valve locks and lash caps, but at the moment they seem to be a niche offering.

Fig. 1 - A CAD model of a combined lash cap and valve lock

Written by Wayne Ward

When excited at certain frequencies, the spring will vibrate at one of its natural frequencies, and in such situations the stresses in the spring may be far higher than those in which we intend the spring to operate. It is therefore imperative that engine designers consider techniques and methods to reduce the tendency of the spring to operate in this resonant condition, which is often known as spring surge.

The first thing that should be done is to ensure that none of the major harmonics of the cam profile coincide with the natural frequencies of the spring. A cam profile is not a simple sinusoidal wave, and can be analysed to show the strength of a number of harmonics. Basically, the profile is made up of sine waves of frequencies that are integer (whole-number) multiples of the actual frequency of valve opening. The strongest of these are the low-number harmonics, so we need to avoid these if possible.

The exact number that needs to be avoided is not universally accepted, and depends on the experience of the engine designer. I’ve heard of people avoiding as few as the first five harmonics, and others who choose eight or more. This essentially means that the natural frequency of the spring and mass system (the mass being that of the components, the action of which the spring controls) needs to be N times the basic excitation frequency. So, if the engine revs to 18,000 rpm and the cam turns at 9000 rpm, which equates to 150 opening and closing events per second, we need the natural frequency to be at least 150 N. This is the strongest fundamental defence against surge.

Where two or more springs are used to control each valve then we can have them specified and supplied with a light interference fit. It is usual to specify the springs with different natural frequencies so that, if surge should occur, only one spring will resonate at a time. The damping action due to friction between the resonating spring and its partner reduces the damage done by converting some energy to heat and, where space allows, some spring suppliers also supply special dampers that fit between pairs of nested springs.

Progressively wound springs, which have tighter coils at one end, are also effective. As the spring compresses, the most tightly wound spring coils come into contact, changing the  effective number of coils and increasing the spring rate, which is a measure of the spring’s stiffness. The natural frequency of the spring-mass system is proportional to the square root of the spring stiffness, and where the change in stiffness throughout the opening event is large enough, the fact that the excitation no longer coincides with the natural frequency of the spring effectively controls vibration.

Where a single spring is used, beehive or tapered springs are also another way of providing a spring with a changing natural frequency throughout the lift curve, achieving the same effect as a conventional progressively wound spring.

Written by Wayne Ward

The fact is that there are countries – the US and Australia for example – where cheap gasoline makes daily motoring with a big engine perfectly affordable. While there is an increasing permeation of European and Asian cars into these markets, pushrod engines remain stubbornly popular. Engineers from Europe and Asia ought not to labour under the impression that the Americans can’t understand the benefits of overhead cam (OHC) valvetrains, because they do. The likes of GM and Ford know very well how to make OHC engines because they make them in their millions.

The fact is that overhead valve (pushrod) engines are very compact. From the crankshaft to the top of the cam cover, an OHC engine is much taller. When GM looked at new engines for its latest incarnation of the Chevrolet Corvette, it assessed OHC designs. GM is a huge company, with resources to match, and a lot of clever engineers there looked at the options and concluded that a big pushrod V8 was what they needed. The OHC V8 that provided the requisite performance, however, was simply too tall to fit beneath the desired bonnet/hood contour.

There are other alternatives though to allow an OHC engine to fit under that hood – a smaller, higher-revving engine might have done the trick, as would a turbocharged or supercharged engine of smaller displacement and dimensions. After all things were considered, GM came to the conclusion that the simplest and most effective engine that gave the required performance with the required dimensions was the pushrod engine.

Until legislation or economic factors force manufacturers into making big changes to their car designs and the engines that power them, people will still want big V8s while they can afford to run them. While there is sufficient demand for big-capacity engines, we can assume that pushrod valvetrains will remain a popular choice when people come to assess their power unit requirements, just as GM has done, mainly owing to their advantage in terms of engine height. Someone once said to me, “The biggest problem with a pushrod engine are the pushrods”; that may be true, but the engines still have a valid place in the market, based on sound engineering judgement.

Pushrod engines still do very well in open competition against OHC engines. In endurance racing for example, there has been a protracted duel between Corvette and Aston Martin for many years. The rules aren’t exactly the same though for both engines: there are breaks given to the Corvette owing to the fact that the engine has two valves per cylinder, but the breaks are relatively small. Of course this leads to some complaining, but of course the option is there for the complainant to use a pushrod engine of their own!      

Written by Wayne Ward

In racing, the overhead valve (pushrod) engine has a mandated stranglehold in NASCAR, and where large-capacity engines really count – in drag racing – they reign supreme. Where pushrod engines compete on even terms, such as Le Mans, they are more than capable of holding their own.

Each component in the racing pushrod valvetrain is very specialised, and the lifter (cam follower) is no exception. Apart from the Mercedes-Ilmor 500I Indy engine, people have generally stuck to using a translating cylindrical follower. Where the design engineer has the freedom to use a lifter with a bearing, he will generally do so. It allows him to open the valves with higher velocity and get more gas flow through the engine. Airflow means power, and if a good job is made of the tuning, an engine using roller followers will always beat one equipped with flat-faced followers. Almost without exception, roadcar manufacturers have also adopted the roller follower.

There are variations on this basis design, with one of the main distinguishing differences being the type of bearing used for the roller. The simplest bearing is of course the plain bearing. In the plain bearing lifter, the roller can be bushed or bushless, but the principle remains the same – the roller and shaft are separated by a thin film of oil, as is the case with the bearings for a typical crankshaft. This arrangement has the advantage of being very stiff – the roller has a thick section and the shaft can also be stiff.

The roller bearing variant uses a needle roller bearing between the shaft and roller. The roller and shaft surfaces must both be very hard in order to resist rolling contact fatigue, and the stresses involved are very high compared with the plain bearing variant. The load is carried by a small number of rollers which are essentially line contacts. Although elastic deformation provides some increase in actual contact area, it requires that the components are very specialised, and the material limitations of the follower shaft and roller bore need to be borne in mind when designing new cam profiles; for plain bearings contact stresses are hardly a concern. The stiffness of the roller bearing arrangement is also lower than the plain bearing type. For a given follower roller outside diameter, the section is diminished by the space taken by the bearing. An increased shaft diameter can help improve stiffness and reduce contact stresses for a given follower load.

Additional to the above considerations are friction and wear. The roller bearing is felt to offer lower friction, and at start-up we have a rolling contact scenario, where the plain bearing variant is in sliding contact until the oil film is fully established. The roller bearing type is probably less prone to failure due to lack of lubrication at start-up.

Think carefully when specifying the type of follower, as friction, wear and stress all need to be considered.

Written by Wayne Ward

Collets are generally too small and light to make a significant difference to reciprocating valvetrain mass; there is much greater scope to reduce mass by looking to components such as the valve, springs and spring retainers. However, the engineer should not be drawn into thinking that these simple components have suffered from a lack of development. Their interface with the valve and the spring retainer are critical to the reliable function of the valvetrain. Stress concentration and fretting where poor design choices have been made or incorrect materials have been chosen are two reasons why collets can cause valves to fail.

Surface treatments and coatings also play an important role in providing a suitable collet. Coatings can be an important way to reduce problems owing to material incompatibility or simply using a material whose surface is not suited to the application. For example, if the lock and retainer materials are too similar, the materials are soluble in one another, and there can be problems with the materials seizing.

There are situations where both aluminium and titanium are used for collets, and in both cases it is probably wise to consider coatings to prevent damage during installation and service. In particular, in the case of titanium collets being used with titanium valves and retainers, there is a danger that cold-welding of the surfaces could take place with very little load. Protecting the interface by coating one of the components must be considered. It is quite often the case that the valve will be coated with a hard, thin engineering coating such as chromium nitride, but sometimes titanium collets are also coated, especially where there is no coating in the bore of the retainer.

Even when there are no particular concerns over the collet material, it is common to find that collets have been subjected to ‘oxide’ surface treatments to prevent corrosion. We should not confuse the term ‘oxide treatment’ with rust though – the process here is what we might usually call ‘black oxide’ or ‘blackodising’, where a very thin layer of the surface is converted to an oxide known as magnetite, which is then subsequently treated with oil. The oil is absorbed into the component surface, providing a degree of corrosion prevention.

A number of companies who offer bespoke collets to race engine suppliers and builders say they will coat or surface treat collets to suit customer requirements, and such coatings will include soft metal coatings such as silver. The aim of coating or surface treating the collet is not to reduce friction – it is desirable to have a high friction coefficient at the interface between the collet and the retainer in order to reduce movement.

Written by Wayne Ward

Controlling the valvetrain at high speed is much easier when the valvetrain components are stiff and light. One advantage of pneumatic valve return systems (PVRS), as found in MotoGP and Formula One, is that there is no valve spring mass, and the ‘piston’ that in effect replaces the spring retainer is very light indeed. Many of the diagrams showing PVRS though have the ‘piston’ as a far larger and more complex component than is actually the case in a modern system.

The valve spring retainer has to be as light as possible. A previous article on this subject looked at materials options, but in this article I want to look at some of the design features that make for a low-mass and reliable retainer.

The retainer’s main function is to retain the valve in its compressed state, providing the correct level of pre-load and reliably coping with the stress cycles imposed on it. The valve retention involves initially compressing the valve and then locking a pair of collets in place so that the valve remains loaded against the seat when the valve is closed or, when the valve is open, the follower should (in most circumstances) remain in contact with the cam lobe.

Collets are tapered on the outside, and in order to provide a reliable seat the cone angle on the retainer must closely match the collet angle. It is also important that the surface finish in the collet bore is good, otherwise there may be some wear and slight loss of pre-load. A certain amount of material around the collet bore is needed to prevent the collet being pulled through the retainer – any movement of the collet subtracts from the spring load, so we need to guard against this by providing a sufficiently stiff ring around the collet. It can be tapered though to reduce mass.

The springs are retained in a concentric relationship with the valve by having the spring location (or multiple locations if nested springs are used) machined concentrically to the collet bore. If the springs are not concentric then the loads with the valve won’t be either, and bending stresses will be increased. The transition between the compression face of the spring location and the radial location surface also needs to have a generous radius. This is the point of highest stress in most circumstances, and you should be sure your valve spring has a large enough chamfer on its upper end turn to fit this radius on the retainer without interference.

The top of the retainer represents a large mass, and sometimes there is a tendency to make this too flexible – remember that there will be a loss of spring load compared to that expected for a given valve lift as the retainer deflects. This can’t be prevented, but it can be mitigated.

Written by Wayne Ward

With timing gears, what we want often want to do is improve durability while decreasing inertia and component mass. A number of surface treatments can help us achieve these goals.

Carburising, also commonly referred to as case-hardening, is a technique that not only improves wear resistance but imparts high levels of compressive residual stresses into the surface of the component. This gives the gear a much greater fatigue strength, as it effectively lowers the maximum tensile service stresses at the component surface. This helps not only in terms of preventing tooth fatigue breakages, it also provides a degree of protection against rolling and sliding contact fatigue that can lead to pitting damage on the tooth flanks. This process cannot be applied to existing gears; the steel used must be of a certain composition suitable for carburising, and it is likely that the teeth will need to be ground following carburising in order to remove the effects of distortion that often occurs when this process is used.

Nitriding is another process that imparts compressive residual stresses to the surface of a component, and it also needs to be done on a special composition of steel. It is used for gears, but less commonly so than carburising.

Shot peening is a mechanical process that aims to put the surface of the gear into a state of residual compressive stress. The usual method of shot peening is to propel a high-velocity stream of hard media toward the component surface; there are other methods of shot peening, such as ultrasonic, but these are very much in a minority.

Shot peening is often combined with carburising in order to give a high degree of compressive stress – the resulting fatigue resistance offers an improvement over either of the processes in isolation. It can of course also be used in isolation, and on existing components with success.

Multi-stage peening, using different peening media and intensities, is often specified for the timing gears in race engines, and this allows an optimum compressive stress profile, with the maximum level of compressive stress at the surface. After initial peening, the maximum level of stress can sometimes be found below the surface: there is often an advantage to have the maximum compressive stress at the surface, especially where a component is loaded in bending.

There is a well documented increase in fatigue strength owing to improvements in surface finish, so we can expect to find gains in durability by using treatments that decrease the surface roughness of the gear teeth. Polishing or superfinishing processes, combined with carburising and shot peening, have been shown to be very effective.

A number of companies find an advantage in making timing gears from through-hardening steels, and dispensing with carburising or nitriding, often using both peening and polishing to acquire a useful degree of compressive residual stress. There is a focus article in Race Engine Technology, issue 74 (November 2013), which deals with carburising, nitriding, peening and polishing in more detail.

Written by Wayne Ward

In the case of the inlet valve, heat flux is only very modest compared to that in an exhaust valve. The inlet valve is kept cool owing to the flow of cool inlet charge over it, and very little heat is rejected to anywhere other than the valve seat. The exhaust valve is a different matter, with extremely high heat flux not only into the head but also the area behind the head. Here, there can be considerable heat transfer through the valve guide.

However, the proportion of the heat rejected through the valve guide, and so the total amount of heat transfer, is increased greatly when sodium cooling of the exhaust valve is used. The result of the higher overall heat transfer is that the valve operates at lower temperatures, which generally means that the allowable stresses for the valve increase. Whether the design and development engineers take advantage of this by reducing valve cross-section, or simply have a greater factor of safety, is a choice for them to make.

There are other benefits to having this greater heat transfer through the valve guide. As the valve head runs cooler, the instantaneous contraction as it comes into contact with the relatively cold valve seat is reduced. This in turn reduces the tensile stresses in the periphery of the valve due to the seat area trying to contract around a hot ‘core’. So, in addition to the increased allowable stresses, we can lower the tensile stress in the periphery of the exhaust valve head during part of the valve’s operating cycle by rejecting more heat through the valve guide by using sodium cooling.

The inlet valve does not however benefit to the same extent from increasing the heat transfer through the valve guide, although the effects noted will be the same. The real benefit for an inlet valve is likely to be due to the reduced valve temperatures. There is some heat exchange from the valve to the inlet charge, with the valve being cooled by the inlet charge. It is clear that the charge must be heated by the valve, and this naturally has a small but deleterious effect on charge temperature. The slightly increased charge temperature will give rise to lower charge density and lower volumetric efficiency.

So, there is an argument for increasing heat transfer through the inlet valve guide, but probably more for reasons of slightly increasing performance than reliability. There isn’t strong evidence of this having been used widely in racing. Some of the 1950s works Norton motorcycles used sodium-cooled inlets, but whether Norton’s reasoning was to improve performance or to try to solve a reliability problem isn’t known.

In terms of reliability and performance, the use of sodium-cooled inlet valves is likely to look more attractive if the valve is hot, and the increasing use of turbocharged engines might see such valves resurrected in modern race engines.

Written by Wayne Ward

Other types of cam followers, such as finger followers, may offer a number of advantages – including lower reciprocating mass, more latitude with valve lift curves and possibly also decreased friction – but there are significant practical advantages with inverted bucket followers. They are essentially cylindrical components that can be produced by conventional turning on a lathe, and can be finished by grinding on conventional grinding machinery. Finger followers are much more complicated to manufacture, with a great deal of complex milling, and it is difficult to achieve the required accuracy between the pivot and the faces that contact the follower and the lash cap.

The correct valve clearances are generally achieved by selecting a lash cap (often called a tappet shim) that sits between the top of the valve and the inside of the bucket. However, there are engines which have used inverted bucket followers but which have taken a different approach to the adjustment of clearances, choosing instead to fit a larger shim on the top of the follower, between the follower and the cam lobe.

In order to be able to use the maximum possible valve opening and closing velocities, the shim needs to be as large as possible. The need to retain the shim means the follower top needs to have a wall, which is responsible for keeping the shim in its correct position. This means that, for a given follower outside diameter, the shim diameter must be smaller; as this represents the maximum contact diameter, the valve opening velocity is limited (valve opening velocity for a translating flat-faced follower is proportional to the maximum contact eccentricity from zero opening). This can represent a significant disadvantage.

This approach does however allow the use of a low-density bucket, providing that the contact stresses are maintained within material limits. If we take the simple example of an 80 g reciprocating mass being accelerated at 1000 g and 1000 N spring force, the force required  is 1000 + (0.08 x 1000 x 9.81) = 63,784 N.

If we were to use a titanium bucket with a yield stress of 900 N/mm², we would need an area of 69.76 mm to resist yielding. This represents a circle of just less than 9.5 mm diameter. So, if we were to use something similar to a lash cap on the underside of the follower and a large lash disc on the top, we can use a much lower density material for the cylindrical body of the follower assembly. This would reduce the reciprocating mass of the valvetrain and thus lower the requirement for contact area.

As passenger cars use much more modest valve acceleration rates and lower spring forces than are usual in race engines, they can make even larger steps in the use of low-density materials. I have seen cam followers made from anodised aluminium for the follower body with a hardened and ground steel lash disc between the follower and camshaft. A steel ‘lash cap equivalent’ would provide sufficient contact area beneath the follower to prevent yield.

Written by Wayne Ward

With flat tappets, the lift velocity is limited by the size of the tappet, which again is limited by NASCAR. A roller lifter allows higher lift velocities, and within a given valve event duration this extra latitude can be used to maximise lift and the area under the lift-angle curve. It is little wonder that roller lifters are almost always used in passenger cars, and the same applies in racing where allowed. Even though roller lifters are more complex and expensive, very few passenger cars have continued to use flat lifters in recent years

One disadvantage with the use of roller lifters though is that, for the roller to be properly aligned with the cam lobe, the lifter must be prevented from rotating within its lifter bore. There are several ways of achieving this.

Chevy’s ‘lifter tray’ solution is a one approach, where a flat on the lifter body locates on a corresponding feature on a plastic lifter tray. One problem sometimes reported with lifter trays though involves oil collecting in the trays, so additional holes are often added to ensure oil cannot pool in the tray.

Another solution, often known as a ‘dog bone’, is where a metallic plate is bolted to the engine block, and is equipped with flats which locate on corresponding flats on the lifter body. It is similar to the lifter tray approach, but the dog bone is a much smaller component than the lifter tray, and suffers none of the oil drain problems reported with trays. In common with the lifter tray solution, the dog bone gives the lowest possible lifter mass, as no reciprocating mass is added.

The lifter bore can be machined to allow an anti-rotation ‘key’ to slide within it. This key is formed as part of the lifter, and this is possibly the most compact solution to the problem. It does add a little mass to the reciprocating follower though, and requires a more complex lifter bore.

Another idea, although much heavier in terms of reciprocating mass, is to link a pair of lifters on one cylinder with a relatively loose articulating link known as a tie bar. The link itself is quite a simple component, although it has the disadvantage of adding to the reciprocating mass of the components on the pushrod side of the rocker. However, only part of the tie bar’s mass is ‘seen’ by each follower, as the mass centroid will move only half of the average lift of the two valves on any cylinder.

Written by Wayne Ward

The pushrod valvetrain is the most striking difference between OHV engines and overhead cam (OHC) engines used elsewhere in the world. The pushrod is one of the most critical components in the engine – not only does its design have a huge effect on the stiffness and resonant frequencies of the valvetrain, it also has to run very reliably in spherical sockets at each end of the pushrod, where conditions for lubrication aren’t great.

The main reason that lubrication conditions aren’t ideal is that the pushrod is constantly moving relative to its pivot, but only at a very low velocity. It is difficult therefore to generate a satisfactory oil film that can reliably separate the components under all operating conditions. In general, race engines place higher loads and stresses on components and contacts between them. This is especially true of valvetrains, where high rates of valve acceleration are one of the development engineer’s tools when producing improved valve lift profiles.

As such, the materials from which pushrods are made is critical. Some engine suppliers choose to use three-piece pushrods, where the main ‘tube’ and the ends can be made from different materials with very different properties. However, a number of engine suppliers, right through to some of those supplying winning engines for Sprint Cup, choose to use single-piece pushrods.

The material choice for such components is critical. If we look at NASCAR engines, the use of steels is mandatory, but it may not be wise simply to choose the first very hard steel that springs to mind for a single-piece pushrod for various reasons. First, some very hard steels are also very abrasive, especially where the surface is not machined to a very fine finish; this may promote wear of the sockets at each end of the pushrod. Second, for the main part of the pushrod, we want something with a degree of toughness in order to cope with high cyclic loads and shocks.

However, choosing a steel with high toughness is also not without risk. Many high-toughness steels may not be capable of being made sufficiently hard to withstand the contact stresses. In specifying the steel for a pushrod, the engineer needs to be sure that he or she has the correct combination of toughness and strength. They also need to ensure that the material is not too abrasive in its finished condition.

A popular choice for single-piece pushrods are tool steels. There are a number of such steels which, when suitably heat treated, are suitable for high-spec pushrods. For example, some companies offer single-piece pushrods in H13.

Written by Wayne Ward

In their unending quest to make the engine perform better through improved breathing, development engineers will try to open and close the valves in an ever more aggressive fashion, and will want to achieve this with the smallest and lightest spring. The stresses in the springs are high as a result, and without careful surface engineering they would have a very short life.

The spring wire itself is very special. It has extremely high strength compared to any ‘normal’ specification of steel, and special attention is paid to the cleanliness of the material. In order to improve the fatigue life of the material, there is an advantage in placing its surface in a state of compressive stress. As the stresses are highest at the surface, any residual compressive stress before service loads are applied effectively reduces the stresses in the component when in use. For example, if there is 500 MPa of compressive stress before loading, and the service load results in a stress of 1800 MPa, the algebraic sum of these is 1300 MPa. While compressive residual stresses don’t reduce the stress amplitude, however, they can beneficially offset the mean stress significantly.

The two main surface treatment processes used in the manufacture of racing valve springs are nitriding and shot-peening. Nitriding is a surface hardening process during which nitrogen is diffused into the surface of the steel. The depth of nitriding on a valve spring will be small compared to that on a larger component such as a crankshaft; while it is possible to nitride small sections completely, we want to maintain a tough, ductile core. Nitriding is known to impart significant residual stress, and the process can be tailored to increase the level of that stress.

The process takes place between 400 and 600 C; the higher the temperature, the greater the rate of diffusion of nitrogen into the surface. The temptation would be to use a higher temperature to shorten the process time, but this is not without penalty. Nitriding can lead to softening of the core material, so some valve spring steels have been developed especially to resist the softening effect of nitriding temperatures.

Shot-peening is a well known technique for increasing the fatigue strength of many components. For the highest specification of valve springs, multi-stage peening is employed using different sized media and peening intensities. The aim here is not only to ensure a high level of compressive residual stress but to make sure that the maximum occurs at or very close to the surface while also having a useful depth of compressive residual stress.

Where both nitriding and shot-peening are used, the nitriding process is done first, as the nitriding temperatures would relieve the stresses introduced through peening. Another advantage of peening after nitriding is that the amount of residual compressive stress due to peening is increased, as the substrate is harder and stronger. According to Suda and Ibaraki*, a combined nitrided and shot-peened material can have a 40% higher fatigue strength compared to the non-nitrided and non-peened state.

* Suda, S. and Ibaraki, N., “The Past and Future of High-strength Steel for Valve Springs”, Kobelco Technology Review, December 2005

Written by Wayne Ward

The development of the valvetrain often includes new camshafts, valves, valve springs and spring retainers. The ‘forgotten’ components in valvetrain development though are the collets – also widely known as ‘valve locks’ or ‘valve keepers’ – although they too have a very important role to play. The reason they are forgotten is possibly because they are so small and there is little scope for development.

The role of these components is to allow the spring retainer (or piston, for pneumatic valve return systems) to be positioned on the valve. They act as a pair and have a tapered outside surface to allow them to lock into the retainer or piston securely due to friction. On the inside, there is a ‘ridge’ which is used to position the collet axially on the valve; the valve has a corresponding groove into which the collet ridge can locate.

The design of the internal ridge dictates the groove geometry on the valve; as this groove is a point of stress concentration, we would like this concentration to be minimal. Compared to passenger car and industrial engine valves, racing valves are often much more highly stressed, so racing collets have a semicircular ridge, whereas many road valves have a simple ‘square-cut’ groove to accommodate a much simpler collet ridge.

The whole interaction of the collets with the valve needs to be considered carefully, especially with highly stressed valves. Not only does careful attention need to be paid to the groove into which the collet ridge locates, but there are sometimes further relief grooves to prevent hard contact with sharp edges at the lower end of the collet.

Ducati has been unusual in that it has not used a split collet, but instead has used a split steel ring to provide a location on the valve and a single-piece ‘collet’ which is constrained from axial movement. In rejecting the use of valve springs in favour of a desmodromic system, Ducati requires a method by which to shut the valve. This is a done by a valve-closing camshaft, which operates the valve via a rocker acting on the single-piece ‘collet’.

There are a wide range of materials used for collets. Steel, as is commonly used for passenger vehicle engines, remains the most popular and cheapest option. Given the tiny mass of these components in comparison to the other reciprocating masses in the valvetrain, people only generally tend to consider lower density materials for collets once the larger reciprocating valvetrain components have been optimised, as the mass reductions available are so small as a percentage of overall reciprocating mass.

Titanium is another option, and this is offered by a number of motorsport valvetrain component suppliers. Care has to be taken though when using titanium collets. People who have taken the step of using titanium collets are very likely not to be using steel retainers. Titanium also has some wear problems, especially in contact with other titanium parts, so caution needs to be exercised to prevent surface damage. A coating on one or other part may be required. Other materials could also be considered, such as high-strength aluminium.

Fig. 1 - A typical valve collet

Written by Wayne Ward

In a number of articles in Race Engine Technology magazine, the importance of maintaining control of the valvetrain over the working range of the engine has been discussed. ‘Maintaining control’ means having an understanding of how the valvetrain is working dynamically. As simulation and testing has become more widespread, what has become clear to people is that valvetrain stiffness is important, as is component mass.

The valvetrain can be considered as a ’spring/mass’ system, albeit a complex one with many different degrees of freedom, numerous springs and numerous dampers. You may think you have only one set of springs, but every component in the system has a spring rate and a damping characteristic. The spring retainer is an important part of the valvetrain in terms of mass, because it is reciprocating, but also in terms of stiffness. If we try to pare the retainer down too much in pursuit of low mass, we might compromise stiffness (effectively making the spring less stiff) and increase stress.

Therefore, consideration of spring retainer material is an important design consideration. Passenger car retainers are generally made of steel, and this remains a popular option for race engines. High-strength steel has a high fatigue limit, and a number of surface treatments such as carburising, nitriding and nitrocarburising can improve this significantly. The downside to using steel is mass, and effort needs to be put into the design to minimise unstressed material to keep mass low.

Titanium is another popular material for spring retainers, as it has a low density compared to steel and good mechanical properties. It does suffer though from poor wear characteristics, and these parts are often coated on contact faces (collet support face and spring contact faces) with a coating such as titanium nitride or chromium nitride. Titanium metal matrix composite (MMC) materials would be a good spring retainer material, and such alloys are proven for valvetrain applications (1).

High-strength aluminium alloys are also used for spring retainers, as are aluminium-based MMC materials. Aluminium has lost some of its initial popularity as a retainer material, possibly because people failed to understand that aluminium components need to be looked after more carefully and replaced more often than steel or titanium components. There is no actual problem with using the material, so long as you understand its inherent limitations.

Typical high-strength aluminium alloys used for spring retainers are 2024 and 7075, and in addition to those bespoke designs produced for race engines, many companies still sell aluminium retainers for tuned roadcars. Aluminium MMCs have been used in race engines for at least 20 years (2) and, compared to a conventional aluminium alloy, benefit from dramatically increased stiffness with comparatively little increase in density.

References
1. Hunt, W., Miracle, D.B., Automotive applications of metal-matrix composites, in ASM handbook, ASM International, vol 21: Composites, 1029-1032, 2001
2. Dwivedi, R., Altland, G., Barron-Antolin, P., Leighton, J. et al, “Applications of Metal Matrix Composites in High Performance Racing Engines”, SAE Technical Paper 911770, 1991

Fig. 1 - A sectioned CAD image of a titanium spring retainer for a racing application

Written by Wayne Ward

In a bespoke race engine, you will quite likely find that the camshaft drive is via a series of gears, which are generally straight-cut spur gears. While helical gears are generally quieter in operation than spur gears, they generate axial thrust and require thrust bearings to cope with more load. Friction and wear are considerations in choosing straight-cut spur gears over the helical type.

Compared to a belt or a chain, these can represent a significant weight disadvantage, so why use them? There are several reasons, chief among which is precision. The improvement in precision is aided by the fact that gears are very much stiffer than either belts or chains. All cam drives are elastic systems, and they have to cope with rapidly varying loads. With lower stiffness, the elastic deflections are greater and precision is reduced accordingly; with a system lacking in stiffness we cannot be sure of the exact relationship between crankshaft angle and camshaft angle. This means that, for an engine equipped with chains or belts, there has to be an extra allowance for elastic deflections in the system compared to an engine with gear-driven cams.

This allowance might be that the timing using standard cams is more conservative than we would like to use or, where bespoke cam profiles are used, these may not be as aggressive as the engineers would like. Aggressive opening profiles, with high rates of acceleration, cause higher loads which then exacerbate the deflection in any cam drive. However, gears are less affected than either chains or belts in this regard.

A definite advantage of gears over belts or chains is the lack of tensioner and the fact that there is no periodic adjustment or maintenance. There are many race engine failures due to poor adjustment of chain tensioners or incorrectly functioning tensioners.

In high-speed engines, gear drives seem to be favoured. The last chain drive designed into a Formula One engine was during the mid-to-late 1990s by Cosworth. More recently, Aprilia caused some controversy by changing from chain-driven to gear-driven cams in the middle of the 2010 season, a move which was legal within the rules but which wasn't welcomed by other teams, who could see a very fast bike becoming even faster. The race kit parts and engine modifications to convert the Aprilia engine were reputed to be very expensive. However, Ducati is known to have a 'hybrid' chain and gear drive system for its desmodromic valvetrain on its twin-cylinder superbike engine.

Written by Wayne Ward

Although they have a very similar cycle of operation, the inlet and exhaust valves are subjected to very different thermal cycles, with the exhaust valve running much hotter than the inlet. This is because the inlet valve will have relatively cool air or air-fuel mixture intermittently flowing past it, whereas the exhaust valve is heated by very hot burnt exhaust gases flowing past. Consequently, the exhaust valve has much higher demands in terms of material temperature resistance and has greater cooling requirements.

The valve seats are the primary means of removing heat from a valve, and therefore one requirement for the seat material is high thermal conductivity. The combination of valve-to-seat contact area and thermal conductivity dictates the potential rate of heat removal for a seat.

Valve seats are often produced from copper alloys, with beryllium copper being popular owing to its combination of strength and thermal conductivity. It is common practice to produce the inlet seat from a stronger material that is less thermally conductive. Copper alloys are also excellent bearing materials: they are very resistant to seizure and do not react chemically with any of the fluids passing through the ports and combustion chamber.

The thermal limit for copper exhaust seats is sometimes approached for exhaust valve seats in boosted engines with high exhaust gas temperatures, where the seat can sometimes begin to melt. In production engines, valve seat inserts produced by sintering powdered ferrous materials are often used, and some companies have developed sintered materials that are infiltrated with copper to improve thermal conductivity.

The valve guides are also commonly made from copper alloys. As less heat is transferred through the guide than the seat, their requirements in terms of thermal conductivity are lower, and the materials used are generally not the very expensive alloys containing beryllium. Having said this, engines using sodium-cooled valves transfer a much greater proportion of heat through the guide; the liquid sodium is very effective at taking heat from the valve head and transferring it to the stem. The strength requirement is also lower; in an ideal situation there should be no lateral load on the valve, although in the real world there is often a degree of side load from valve actuation and thermal distortion of the head.

While the guide and seat are machined concentrically with great accuracy, we would be naïve to think that they remain perfectly concentric in use. The lateral loads are low enough to allow the use of other materials. Aluminium is one possible candidate, as it has low density and high thermal conductivity. Cast-iron guides are widely used in heavy-duty production engines, and powder-metal ferrous metal guide materials have been developed in which lubricating fluids are stored in the pores of the sintered structure. There are also sintered guide materials that are infiltrated with solid lubricants.

Written by Wayne Ward

However, not all lash caps are the same, so here we will look at some of the variations in their design.

The most commonly used type of lash cap is the 'inverted cup'. These are relatively simple to make, and can be bought to suit many valve sizes at only modest cost. The 'base' of the cup is the important feature here; its thickness is carefully controlled and both upper and lower surfaces are carefully machined to a fine finish. The wall of the cup is simply there to keep the lash cap centred on the valve stem.

There are several types of lash cap that are even more simple than the inverted cup. Very small plain discs can be used, but these must be prevented from moving. This can be done easily where inverted 'bucket' followers are used in an overhead cam engine; a small recess is provided in the underside of the follower.

However, any marginal saving in complexity and mass with this design when used with the 'bucket follower' is cancelled out by the extra complexity involved in the design and manufacture of the follower. They are also more difficult to fit than an inverted bucket lash cap, and great care has to be taken in their installation to ensure they remain in place. Again, where bucket followers are used, a recess can be machined into the top of the follower and a large disc is used as a lash cap.

These have proved popular in production engines, and a number of manufacturers use aluminium followers with steel shims. The hardened shim provides the contact face between the cam lobe and the follower, so the aluminium follower skirt acts only as a guide for the follower in the bore. A further variation on the flat disc shim is one where a relatively small shim is prevented from moving because it is restrained by a corresponding bore on the spring retainer.

Turbocharged and supercharged engines sometimes require a special type of lash cap. Occasionally a plenum explosion will occur that reaches such a pressure that the valves open when the cams aren't controlling them. As there is nothing accelerating the follower, the lash cap can find itself with sufficient clearance to escape from its intended location. Such lash caps require special design features to ensure they cannot be separated from the valve, except when the engine is rebuilt.

Some lash caps are provided with a small hole in the centre; this is not to save weight but to ensure there is no air or oil trapped beneath the shim on fitting. If there is trapped air or oil 'holding off' the shim from metal-to-metal contact, incorrect valve clearances can be set. When the air or oil is displaced during service, the valve clearance increases, and performance and reliability can be compromised.

Written by Wayne Ward

However, one fact the flow bench and CFD will confirm for us is that bigger holes will flow more air for a given pressure difference. Therefore, they will always tell us that bigger valves, used with correspondingly large valve seats, will flow more air. The ability of an engine to flow and, most important, trap air is a measure of its potential to produce power. To a given mass flow of air we meter a set amount of fuel in a set ratio, and the amount of fuel and air trapped in the engine dictates the amount of energy that will be liberated by combustion.

What people have found though, at great expense, is that the flow bench and steady-state CFD are not the best tools for sizing valves. Unfortunately for design and development engineers alike, the best method for those without huge computing power at their disposal is to build engines and run them on the dyno. If you were to undertake this painstaking and expensive development work, you would find a peak in performance, either side of which there would be a drop-off in output. One hugely experienced engineer once remarked to me, on the subject of designing and developing fixed-displacement engines, that it would be less of a mistake to select the wrong cylinder bore size than to select the wrong valve size. What we can say is that it isn't necessarily the best course of action to squeeze in absolutely the largest valves possible into a given combustion chamber.

One measure that we do have that might help us select inlet valve sizes that are somewhere close to optimum is that of mean flow velocity. This is often referred to as the Lovell factor or 'mean mach index', and there are plenty of references to this in Race Engine Technology magazine where sufficient information has been possible to undertake the analysis.

As with other measures of engine design, such as mean piston speed, we find that most race engines, large or small, do not differ hugely in terms of intake mean flow velocity. Those that are a long way outside of the normal mean flow velocity range of 65-75 m/s either have something wrong or the people developing them have taken a very unusual development route. For instance, the Corvette C5R Le Mans 5.97l entry from 1999 had mean inlet valve flow velocity of almost exactly 70m/s. For 2000, displacement was increased to almost 7 litres, and the mean inlet valve flow velocity increased to almost 75 m/s. The last Toyota V8 Formula One engine had a mean inlet flow velocity of 68m/s at peak engine speed.

Bear in mind that the V8 engines had been allowed to run much faster than their current mandated maximum engine speed of 18,000 rpm. During 2006, when engine speeds were not capped, the mean inlet flow velocity at maximum engine speed would have been around 75 m/s. The 7 litre, two-valve-per-cylinder, 24-hour Corvette behemoth and Toyota's beautiful, high-revving, lightweight Formula One engine were similar in at least one respect... .

Fig. 1 - This beautifully machined Formula One combustion chamber houses inlet valves whose sizing appears to share much with some very different race engines

Written by Wayne Ward

A fundamental limitation of the inverted bucket tappet is the opening and closing velocity of the valve. The valve spacing dictates the spacing of the tappet axes and the maximum diameter of the tappet. This maximum diameter in turn limits the valve opening velocity. If we assume that the follower's axis intersects the camshaft axis (most do) then it can be proved that the instantaneous distance from the follower axis to the cam-to-follower contact is equal to the lift velocity, when the velocity is measured in mm (or any other unit of length) per radian of cam rotation.

If we had a 30 mm diameter tappet, the maximum velocity - assuming we were happy to sweep the cam profile contact point right to the edge of the tappet - would be 15 mm per radian. In practice, the limit is lower than this, owing to edge detailing on the top of the tappet and a desire to keep a certain finite width of lobe sitting flat on the face of the tappet. If we wish to use a symmetrical lift profile, it is unlikely that we will want there to be an offset between the camshaft axis and the follower axis, lest we limit velocities further.

If the valvetrain designer cannot increase the diameter of the follower then he can avail himself of extra velocity by using a tappet with a curved top. This has the same geometric effect as changing from a flat-faced lifter to a roller-lifter on a pushrod engine. While maintaining the same valve lift profile, the point of contact remains closer to the follower axis for any value of pressure angle and follower curvature. For a given pressure angle, a smaller cam follower curvature decreases the distance from the follower axis to the point of contact. This can be very handy if your new valve lift profile developed with your engine simulation software overhangs the edge of a flat tappet and you are prevented by engine geometry from increasing the tappet size.

Followers with curved contact surfaces have disadvantages though. They are taller and heavier than their flat counterparts, and the increase in mass can require a different valve spring. Such followers also require there to be an anti-rotation feature on the follower which locates with a corresponding feature in the follower bore. In order to maintain the correct alignment of the follower with the cam lobe, we therefore need to increase the complexity and cost of the follower and the machining in the cylinder head or cam carrier.

Written by Wayne Ward

For the top level of NASCAR racing - Sprint Cup - the governing body mandates the use of cam followers which are free to rotate within the lifter bores, and which therefore are cylindrical. NASCAR mandates flat-faced lifters having a maximum diameter of 0.875 in (22.22 mm). In the modern era, flat lifters are almost an anachronism; no new production pushrod engines have been designed around the flat follower for years, and nobody in racing would use them out of choice, as they limit valve opening velocities. In NASCAR, it is likely that the flat follower is used in order to limit development of better breathing engines, and hence limit engine output.

Roller lifters, which have at their nose a rolling element bearing, are less limited in terms of valve opening/closing velocities and so offer much more scope for development of valve lift profiles. A flat-faced lifter is limited in terms of lift velocity by its diameter; higher lift velocities require larger-diameter lifters. For every inch of lift per radian of cam rotation (one radian being about 57.3 degrees) the cam lobe-to-lifter contact will sweep out from the lifter axis by 1 in. With an infinitely thin cam, we can see that a Sprint Cup cam could only attempt to lift (or close) at a rate of 0.4375 in per radian. In reality, the maximum velocity used would be less than this, as it is common practice to maintain a certain percentage of the lobe width on the flat face of the lifter at maximum velocity.

With a flat-faced lifter, provided sensible contact stresses are maintained, acceleration causes no problem for the lifter. The roller lifter limiting velocity is a function of the pressure angle - that is, the angle between the normal to the contact and the axis of the lifter. The pressure angle is a function of both velocity and acceleration, and if this angle is too large, the lifter will tend to bind in the bore. There isn't room here to expand on the mathematics behind this, but it is sufficient to say that the valve lifting and closing velocities may be comfortably higher than with a flat-faced lifter.

A roller lifter may also be used at higher levels of stress without suffering surface fatigue of the cam or lifter. Lubricated rolling contacts can withstand much higher levels of stress; rolling element bearings are a prime example of this, with the surface stresses used routinely being in excess of the material's tensile strength. One cam manufacturer I spoke to in connection with an article on camshafts published in Race Engine Technology issue 48 said his recommended limiting stress for a roller lifter was more than twice his recommended limit for flat-faced or other lifters with simple sliding contacts.

There are though some disadvantages to a roller lifter. They are more complex, and therefore more expensive compared to a flat-faced type. They are also less stiff, owing to less direct load paths.

Fig. 1 - Roller lifters have a number of technical advantages over flat-faced lifters

Written by Wayne Ward

While Harley-Davidson puts a good deal of support behind its factory team, and S&S has managed to get NHRA to open up the specs a bit on the Buell engine, the Suzuki teams are really behind in terms of technology. Only one team builds its own engines: White Alligator Racing, with second-year rider Jerry Savoie.

In 2011, Savoie began the season with a standard Suzuki motorcycle, but at the prestigious Mac Tools US Nationals, he changed to the Hayabusa Suzuki that was developed by Don Schumacher Racing, before that team decided to pull out of the class. Savoie, an alligator farmer by trade, bought the entire stock of motorcycles and spares from Schumacher and had Mark Peiser build his engines.

The partnership worked well, and Peiser had the new Hayabusa motorcycle ready for action at the US Nationals in the first week of September. There were only two drawbacks Peiser encountered as he began this new adventure with the engine in the new Hayabusa chassis - a dearth of crankshafts and valve spring failures.

"When we bought all the equipment from Schumacher, we had probably 15-20 valve springs that came with the deal. We were running out of springs and I discovered they were a special order spring for this application," Peiser explained.

His supplier had problems duplicating the original springs that came with the Schumacher-built engine, and Peiser stressed it was not the vendor's fault. "It was all about the wire to wind the spring; they got an inferior wire and it made the rest of our season very, very difficult," as Savoie had qualified for the play-off Countdown to the Championship that decides the class champion by the end of the year. He was also in the running for the Auto Club Road to the Future rookie of the year award, so any misstep was costly.

"We just struggled throughout the Countdown with worn-out springs," Peiser said. "The ones I got for the Nationals just wouldn't live: they would lose about 30 lb off the seat pressure on the dynamometer - not even on the track. We tried a different one with a different heat treatment, but that didn't work because they were too brittle and they broke. I broke two in one run at Dallas last fall," he said. It was a dilemma that pushed Savoie back to eighth in points at the end of the year.

The supplier has new wire "and it looks like those springs are going to be good," Peiser said.

He uses a double-valve spring with four coils each on the outer and the inner spring. The material is tool steel, and he said the harmonics are measured at 640 Hz, "Which I believe to be good," he said. Installed pressure is 220 lb and the new springs "only dropped about 5 lb on the dyno, which is good. On the track they'll probably lose 7-8 lb and then settle and stay that way so I can get 20 runs out of them, and that will be huge," for a man who was changing valve springs in 2011 like they were socks.

Peiser checks every single valve spring when he receives it. "I have a certain number that I check them at," he said. "I'll go through the whole batch and check every one, and actually write on the ends what that number is. That way I can run a matched set. If I've got 100 springs, I'll put springs in the engine that have the same installed pressure," Peiser said.

"I'm a valve spring fanatic," he admitted. "I'll change springs when they probably don't need to be changed just so I know we've got good, fresh springs in an engine. I know how important that is and I'll change them every ten to 15 runs instead of 20, just to be on the safe side." The manufacturer shot-peens the springs, he said, and Peiser assured me he doesn't make any changes to his springs other than measuring them for seat pressure.

New valve springs didn't do all the tricks to keep the Suzuki running though. Over the winter, Savoie decided to park the Hayabusa and move to a Buell. "With a Suzuki, you're kind of out there by yourself [without any factory support], but we're not going to abandon the Suzuki; we may run them both."

Fig. 1 - Mark Peiser notes the seat pressure on every valve spring he receives in order to get matched sets (Photo: Anne Proffit)

Written by Anne Proffit

Many naturally aspirated race engines (and some production engines) use titanium as an exhaust valve material. Titanium alloys benefit from low density and are hence attractive to a valvetrain designer. The material may be the same as the inlet valve material in such instances, although there are some titanium alloys available which are chosen specifically for exhaust valve use. These come at a premium, but will run reliably to higher temperatures compared to 'old favourites' such as Ti-6242. Polmear (1) suggests that exhaust valve materials such as Ti-834 have found use in production engines. As ever, motorsport benefits from aerospace development of materials: most high-strength titanium alloys used for valves were originally developed for gas turbine engine compressor blades.

While titanium has found use in turbocharged engines as a racing exhaust valve material, this was in the days of the comparatively low-boost and alcohol-fuelled CART engines.

Steel materials have proved a very popular choice of exhaust valve material for naturally aspirated engines. Austenitic steels such as 21-4N remain popular today for racing applications, and more than 40 years ago it was cited in technical papers and books on the subject of valves (2, 3). These days there are materials with slightly better mechanical properties in this category, such as DIN 1.4882

For the more extreme applications, we have to look beyond steel materials and towards superalloys. Typically specified for turbocharged race engines, these alloys are nickel-based and offer much improved strength and stiffness at temperature. Nimonic 80A is a typical offering from valve manufacturers. Owing to the consistently high market price of the elements from which such materials are made - mainly nickel and chromium in the case of Nimonic 80A - and the difficulties in machining them, such valves don't come cheap, but they do make the difference between an engine working and not working.

(Sources: G&S Valves Technical Information, Issue 2, 2003, for 21-4N, DIN 1.4882 and Nimonic 80A, and Timet datasheet for Ti-834)

Not only must we consider the strength of the exhaust valve material at working temperature, we must also consider other properties such as creep and corrosion resistance. Creep is a measure of the 'relaxation' of material, and is measured by observing time-dependent strain under a fixed load, or by observing time-dependent stress under fixed strain. Books containing relevant creep data include that by Conway (4).

References
1. Polmear, I.J., "Light Alloys", 4th Edition, Butterworth-Heinemann, 2005
2. Cowley, W.E., Robinson, P.J., and Flack, J., "Internal Combustion Engine Poppet Valves: A Study of Mechanical and Metallurgical Requirements", Proc. IMechE Auto. Div. 1964
3. Smith, P.H., "Valve Mechanisms for High-Speed Engines", Foulis, 1967
4. Conway, J.B., "Stress Rupture Parameters: Origin, Calculation and Use", Gordon and Breach, 1969

Fig. 1 - Material selection is key to successful valve operation; strength is only part of the equation

Written by Wayne Ward

Anything that doesn't behave as an infinitely stiff member (which means everything) will affect the motion of the head of the valve. In the time since Spintron rigs have been widely used, the importance of pushrod stiffness has been understood, with the result that pushrods have grown in cross-section (and mass) in order to increase stiffness. The limiting factor in some cases is now the available space in which the pushrod can articulate.

One important component that isn't part of the fundamental opening mechanism, but which does a very important job in closing the valve, is the valve spring. The problems with resonance of springs are well known and have been widely documented. The most recent article on valve springs in Race Engine Technology magazine (issue 54) has a number of references, one of which by RR Tatnall* makes for worthwhile reading.

In order to avoid the phenomenon of spring resonance - often referred to as spring surge - we need to avoid the natural frequency of the spring. When this is calculated, we will find that it is very much higher than the operating speed of the engine. The reason here is that we need to avoid the lower cam profile harmonics that could excite the spring, causing it to surge. With sufficient knowledge of our cam profile we can certainly do this.

So, how are pushrods involved in this? Well, choosing an incorrect pushrod stiffness can also excite valve springs. When the pushrod vibrates, as it is wont to do when struck at its lower end, this can excite the valve spring and send it into surge. We need to take care to calculate all the significant natural frequencies for the pushrod to ensure we don't get ourselves into a pickle with spring resonance. Maintaining the engine at the speed at which excitation occurs would lead to a very rapid mechanical failure.

The pushrod, in common with everything else, has more than one natural frequency, and more than one fundamental mode of vibratory movement - it can vibrate torsionally, longitudinally and laterally (bending). Each of these will have more than one natural frequency, with different 'mode shapes' depending on the order of vibration. Anyone with access to a finite element analysis package, and who hasn't tried this before, would find it interesting to look at the natural frequencies and vibration modes of this simple component.

In designing the pushrod for our race engine, we need to have an eye not only on providing sufficient stiffness to help with valve control, we also need to be mindful of the natural frequencies of vibration and how these will interact with the valve spring.

Reference
* Tatnall, R.R., "Fatigue of Helical Springs", ASME Proceedings of Spring Meeting, 1940

Fig. 1 - The interaction between pushrod and valve spring is critical. Incorrect choice of pushrod stiffness can lead to destructive spring surge

Written by Wayne Ward

Ron Shaver of Shaver Engine Specialties, in Torrance, California, builds Tony Stewart's World of Outlaws 410 cu in sprint engines. And, like many other parts of his engines, valve springs are a vital area where risk and reward have to be balanced. We spoke to Ron Shaver to find out the state of his particular art when it comes to valve springs in the harsh environment of World of Outlaws.

Talking about the design he runs, Shaver says, "We have to run fairly high loads in the spring to cope with the valve gear at the rpm we run. We have experimented at length and have developed packages that work well. Currently we run a double spring with an inserted damper on the 410 cu in engines. This current spec of spring is made for the sprint car crowd and for sportsman drag racing. It has an outside diameter of 1.550 in and inside diameter of 1.42 in on the outer spring; the inner spring's diameter is 0.706 in.

He uses the same spring for both intake and exhaust. These springs are made of chrome silicon steel alloy. His supplier uses a proprietary heat treatment and micro-peen combination to give the best possible properties for the material and the best possible surface condition to try to reduce failure-inducing surface defects to an absolute minimum.

The install load is 275 lb, and install height is 2.100 in. Open load is 805 lb with a corresponding open height of 1.200 in. The maximum coil bind is 1.150 in and maximum lift is 0.800 in, Shaver says. The spring has a frequency of 29,368 Hz and, with its damper installed, inner frequency is 28,434 Hz.

If every spring was as good as the best one, life would be much sweeter for everyone using wire valve springs. In fact, Shaver considers valve springs a "voodoo area of the valvetrain. You find one [spring] that works great - and I'll never understand this - and you run them for a year. Then you go to another batch of the same design and the springs don't work the same."

Having worked with every valve spring maker, Shaver has found discrepancies from batch to batch with all of them. There are several theories about the variations in spring reliability. Shaver believes it's the heat treatments used, and cites an example where a batch of springs typically heat treated to 54.1 on the Rockwell C scale were treated to 54.4 and failed.

To try to reduce issues, Shaver carries out a series of his own checks. He makes sure the material is consistent through Rockwell checks, and every single spring goes through a spin-checker and a cycle crush to check it and set it in properly. "We can't check the heat treat and we can't always catch the tiniest of defects, but we work hard at visual inspection and the other physical checks."

While he used to have to finish the ends on his valve springs, Shaver finds the ends on new springs are "pretty well finished; they're well done. We do a visual on all the springs, check them all for pressure on the Spintron and try to run each for a million cycles, checking every 500,000 cycles, the put them in and run them."

Naturally, the spring doesn't work alone, and much work has also gone into developing the components that work alongside the valve spring to make sure they all work well together. While he initially began running steel retainers made of 300M material, he found some breakage. "We made adjustments for that, and in a week we had new tool steel retainers with a different type of shot-peen," he explains.

"We've taken half the spring step on the retainer and put it on the spring seat to cut down on the weight of the retainer, because that's a moving weight and we want to get rid of all the valve gear weight that we can. That helps reduce loads all through the valvetrain - including the spring. We did try titanium retainers, but the spring damper used to eat up the surface of the retainer, so we now stick with steel."

He doesn't flood the springs, as some engine builders do, but he does spray them for cooling from the rocker bar.

"Aside from material or heat treat issues, we have to juggle the amount of direct oil cooling. If we have too little oil you can see it right away because the spring will become straw-coloured. When we see that, we send it right in for testing to confirm the cause of the problem, then we can adjust the cooling accordingly."

And the result of the extra care taken - once confidence is built up in a spring valve gear package - he may even send springs out for a second time after he rebuilds an engine after 550 race miles, meaning a spring could easily stay in an engine for up to 1100 race miles.

With all the development and testing that goes on, it's not a trivial thing to change the spec of the spring. Shaver thinks it will happen only about once a year, sometimes more often, but these changes are usually driven by improvements in material or processes at his spring supplier.

In the harsh environment of the World of Outlaws, upsetting the "voodoo area" is never done lightly at Shaver Engine Specialties.

Fig. 1 - Shaver Engine Specialties uses a double valve spring with inserted damper

Written by Anne Proffit

The question of whether, faced with a clean sheet of paper, engineers would invent something as complex and hard to control as a poppet valve for a high-speed engine is interesting, but not one that we will discuss at this juncture.

As engines have developed over the decades, so have their components, especially in recent times as speeds and output have risen. Steel valves were replaced by more corrosion-resistant austenitic grades. Higher strength superalloys have been adopted for the extremes of turbocharged and supercharged engines, and titanium alloys have remained a favourite of those chasing maximum performance from naturally aspirated engines, and even some companies using lightly boosted engines.

One material that is rarely mentioned in terms of poppet valves though is aluminium. The material offers some real advantages, especially its low mass, which means less closing force is needed, and lower closing forces require a lighter and shorter spring, and therefore a shorter valve. This means the height of an engine designed from the outset around such a valve could be shorter, with the attendant savings in casting mass.

There are a few such critical components in an engine where any mass saved can be multiplied by saving mass from other components as a direct consequence. However, any advantage in terms of valve control due to the lower density of an aluminium valve is offset by having a low modulus material.

You may ask, "Has anyone actually tried this?" The answer is yes, at least once, on a very advanced race engine. The material used was a metal matrix composite (MMC) aluminium material, based on a 2000 series alloy. The MMC in question was not specially formulated for high-temperature use and so was something of a compromise. Since then, however, there have been a number of material advances, both in aluminium MMC materials and in the availability of wrought alloys with improved high-temperature properties.

While these materials may not be suitable for turbocharged engines, there is some interest in their use in naturally aspirated applications, and at least one large valve company involved both in racing and mainstream automotive supply is producing aluminium poppet valves. It would be very interesting to see the results of engine testing with one of these new materials, or something formulated especially for the purpose.

Fig. 1 - Might we see aluminium poppet valves in race engines?

Written by Wayne Ward

This means that the relationship between the tip of the valve, the roller tip and rocker pivot are as designed. Not only does a badly set-up mechanism deviate away from the designed valve lift profile, but the relationship of the contact between the valve tip and roller compared to the valve axis is wrong. This can lead to unintended bending stresses in the valve, and this in turn can lead to problems such as valve guide wear.

There are three major variables in the equation here, each of which can be modified independently by component selection. First, we have a variable-length valve, by virtue of having a range of lash caps at our disposal of varying thickness. Second, we can order pushrods of almost any length we might desire. The third variable is the height of the rocker stand axis.

Essentially, for a given combination of rocker geometry, valve geometry and valve lift, there will be an ideal height from the top of the lash cap to the axis of the rocker. While people can have different ideas about which point in the valve lift cycle the roller axis and valve axis should intersect, most would deliberately avoid putting excessive bending into the valve stem. For teams running the rollerless rocker designs that have been used by some of the Sprint Cup teams in recent years, the problem of bending is likely to be more of a concern, with a greater coefficient of friction being present between lash cap and rocker.

The suppliers of pushrods and associated components are pretty clear that you need to have this relationship between valve tip and rocker pivot axis correct, and only then specify the pushrod length. People who simply try to accommodate a collection of good parts they have to hand are inviting trouble.

A number of suppliers manufacture checking tools that are simple to use and ensure that an engine builder always gets the correct geometry. Of course, if you are running something unusual and special to your engine, you will need to design your own jig or modify an existing one.

Other suppliers specify a procedure during which you check by eye that the roller contact moves across the top of the valve, with the extremes of travel being equidistant from the valve axis. If the centre height of the rocker pivot is too high compared to the valve tip, the options are to fit a longer valve or a thicker lash cap, or machine the base of the rocker stand or the seat area on the cylinder head. If the pivot height is too low, the options include fitting shorter valves, thinner lash caps or shims under the rocker stands. The procedure, whether using a bespoke tool or checking the 'sweep' of the rocker over the valve tip by eye, may involve building the components onto the head a few times.

In the case that the rocker stands incorporate mountings for a number of rockers, they cannot be shimmed to each individual valve, and adjustment of lash caps is likely to be required.

Fig. 1 - The height of rocker stands carrying individual rockers can be adjusted to suit the individual valve they operate

Written by Wayne Ward

Stanton Racing fields both Mopar and Toyota Midgets in USAC competition. The company took over the build and rebuild procedures for California-based Toyota Racing Development about three years ago, and since then it has reworked the engine and focused on longevity as well as power.

"A good example of our valvetrain in one of our Toyota engines are the results for Tracy Hines, who went 18 nights on the same set of valve springs," Milholland says. "They came back to us and were down only about 10 lb." Stanton Racing used to have to change the springs after four to six nights due to breakage, and were losing upwards of 30-40 lb, but changed its valvetrain package after the 2009 season. "We fell in with a company that offers a good overall spring - and it all works. They found the magic bullet," Milholland says.

Of course, the mandated 8700 rpm limit for all USAC pushrod Midget engines has also helped keep the valvetrain happy. "We were running 9300 and 9400 rpm, and that was really twisting the springs, because there's a huge difference in harmonics from 8700," he says. "At the same time, we've got customers who get into the rev limiter so hard and for so long that they were getting deflection and other issues. It seems the valve springs are staying a lot more consistent with the lower rpm, more than anything else."

Milholland points out that his company - as with every other National Midget engine builder - is somewhat limited by the outside and inside diameters on the springs. "It all depends on the size and shape of the cylinder head," he explains. At the moment, Stanton Racing uses a double beehive spring with inner damper; it has an outside diameter of 1.500 and inside diameter of 0.790, with coil outside diameter of 0.210 and insider diameter of 0.150. Weight on the valve spring is 145 g, and Stanton is using the same spring for intake and exhaust.

This steel product - steel retainers and seats are also specified - has been through a nitriding process that allows for the kind of wear and fatigue in typical USAC National Midget races, which are held on both asphalt and dirt-track circuits of varying lengths and preparation. "We rely on the manufacturer for consistency of the material," Milholland says. "We don't have a good way of analysing material consistency here; our focus is on spring tension."

When a typical USAC National Midget engine is built, spring tension starts at about 270 lb of pressure on the seat and may soften by about 10 lb before rebuild, after ten to 14 races. "It's really inconsequential," Milholland says. "We use an install height of around 2 in installed at 270 lb open; our camshaft package will be open at around 800 lb, actually in the area of 795-805 lb."

Stanton Racing finishes the ends of the valve springs itself. "We've got a little procedure we go through as far as trimming the ends off them. We use a high finish sanding cone, maybe 400 grit or so, to finish them before insertion," Milholland says.

Cooling is accomplished through spray procedures. "It's a bit different for the Mopar and the Toyota designs," he says. "Some of the Toyota designs have oilers through the valve covers, and others come through the rocker arms. We're spraying them, one way or the other, but it's a different design philosophy for the Toyota and Mopar engines."

Before Stanton Racing came up with its current package, it was breaking springs and needed to find products that would last. For the past couple of years, none of those problems have been evident.

"We used to worry about softening of our springs - that's the only type of wear we experience, aside from breakage - which we haven't had since we put our new package together," Milholland says.

At the final race of the USAC National Midget season, the 71st annual Turkey Night Midget Grand Prix at Toyota Speedway in Irwindale, California, Midget engines prepared by Stanton Engines held the night. Its Toyota engines set fast times with a new track record and won the race. And no, there were no valve spring issues.

Fig. 1 - Stanton Racing uses the same valve springs for intake and exhaust on its National Midget engines (Photo: Erik Milholland)

Written by Anne Proffit

material with a view to having the valves made in a certain homogenous metallic alloy.

If the parts of the valve are studied in isolation, we might choose to make some parts in different materials to achieve optimum performance. To an extent, through the use of coatings, we are already able to place a different material at the surface of the valve for tribological reasons, and it has been possible for many years to provide hard-wearing valve tips using materials such as 'stellite'.

There are manufacturing techniques applied to valves that can offer different properties in the head and stem. Production methods such as stem extrusion and head forging (upsetting) provide different levels of work and grain orientation in the stem and head. However, through the use of techniques such as friction welding, we can produce the head and stem in different alloys of a given type of material. Indeed, this technique is often used to provide hard-wearing valve tips.

The friction welding process is widely used in industry, but it is generally for ease of manufacture rather than a method for joining dissimilar material grades. There are instances where the technique is used for aerospace and nuclear applications to join materials that differ widely in nature, such as aluminium and steel. The technique is clearly very versatile and can also be used to produce excellent bonds between similar materials, such as different alloys of a given basic material or some more widely differing material pairs. Titanium is an obvious application for such a technique, where a 'stem alloy' could be joined to a 'head alloy' with ease.

Most of the deflection in a valve is in the stem, so naturally we might choose to use a material with the highest possible modulus here, and pay less attention to the density. The head of the valve could be made from an alloy selected for a combination of low density and other properties such as temperature resistance.

While there isn't a great deal of difference between many of the suitable titanium alloys in terms of the properties that we might choose to optimise for our valve, there is the opportunity to join titanium to other materials. We might for example, if the rules pertaining to our race series allow, join a titanium valve head to a titanium aluminide stem (or vice versa). At the moment, titanium aluminide is still a very expensive material, and the manufacture of whole valves from this material would be comparatively costly. However, most of the gain might be realised at significantly lower cost if only part of the valve were to be made from it.

The technique is cost-effective enough for it to be used for the manufacture of some steel racing valves, with a higher-grade material being used for the valve head.

Fig. 1 - Friction welding is used to provide economical valves by using expensive materials only where they are really needed

Written by Wayne Ward

The same reasons that pushed people along the development path of increased pushrod stiffness also led them to consider again the use of steels and other high-stiffness materials for rocker manufacture. Beyond material changes and the inevitable geometric changes that come from having used a new material, there are a number of other changes that can be made to improve the stiffness of the assembly, and which lend themselves in particular to the use of steel.

An obvious target for improvements in stiffness coupled with mass reduction is to remove the adjuster from the pushrod side of the rocker, which many people have done. While this brings with it the need to carry a greater inventory of parts to cope with the range of tolerances on other pieces, it is a worthwhile venture. At the opposite end of the pushrod, it is possible to dispense with the roller bearing which runs on the lash cap. The bearing itself has some flexibility, as does the shaft that carries it and the portions of the rocker supporting the shaft. By running the rocker tip directly on the lash cap, the stiffness of the rocker can be increased.

There are some people in NASCAR who have run this system with success. The form of the contacting tip needs to be carefully considered if it is not to have an excessive rate of sliding as it operates the valve. Excessive sliding can cause lateral forces on the valve, which introduce friction and extra complication when trying to control the valve.

The form of the rocker tip used for this system takes cues from the finger follower designs used in recent Formula One engines, being based on an involute profile; this reduces sliding by making the contact one that primarily experiences a rolling action, similar to a gear acting on a rack, which is the other main application for involute profiles acting against flat surfaces. With high-strength steels and the use of modern, hard, low-friction coatings such as DLC, high levels of contact stress can be withstood. Rockers that operate without a roller can be made lighter than a roller-rocker with equivalent stiffness, or can be made stiffer for the same mass as a roller rocker.

It is also possible with careful consideration to remove the roller bearing at the rocker pivot. Again, this can stiffen and lower the mass of the rocker. Careful thought needs to be given to lubrication of the contact and to the materials used, as well as the correct clearance. The main technical problem here is that a supply of pressurised oil needs to be directed to the contact between the rocker and the shaft. Depending on the position of the existing oil galleries within the head, and the amount of material in the head casting in which to machine the required drillings, this could be a difficult job.

Fig. 1 - The action of a rollerless rocker contacting the lash cap should be similar to the action of a rack and pinion

Written by Wayne Ward

He's developed the Ford's engine throughout last year and into this, and the major problem he's had with his engine is valve springs. Like most Pro Stock entrants, he's increased his rev range into the 11,000 rpm range and, as he says, "It's killing my valve springs."

The triple-spring item Morgan uses hasn't been able to last more than three runs without having to be replaced. "Think about that: since Brainerd [in August], I've got [through] $10,000 in valve springs," he said at Dallas in early October. "That's three events and four tests - so I hate to even fire the engine up on the dyno now."

The valve springs he's using first run at 470-500 psi. "We don't ever want them to go below 425 psi," Morgan says. "They'll start at 525 and lose 50 lb, and they stay in that area. My car will run better the second round because that's the second time those springs have taken a set."

On the other hand, he says, "They'll have less pressure on the engine the second round, and they'll run engine speed it takes to run, but I still have to change them. I have no choice."

The culprit for his failures is high engine speed, but if he backs off he won't make the show. "I feel bad for the valve spring companies because they are trying to supply us with good parts, and they're doing a good job - but we need better parts. They just can't stand the extra revs going from 10,000 rpm to 11,000 rpm. They just don't last as long as they should, and that's a shame."

He's shattered some valve springs and realises there's a chance of wrecking the rest of the valvetrain when that happens. "You tear the tip up into the pushrod, the adjuster up into the rocker arm and possibly wreck the end of the valve," he says. "I worry about what else it's doing when the spring breaks, like the valve tagging the piston. Valve release is perfect ,and if it does tag the piston it comes right back. Then you're having to change the valve, the gasket and the springs because you're not going to take a chance on running them out, unless you're an idiot."

The spring manufacturers have been looking at heat treatments to stop springs being so brittle. "They're trying different heat treatments and they're trying really, really hard," Morgan says. "It's possible they got some bad wire, and that's a problem in their business.

"We really think pneumatic springs are the way to go in this class. We've been pushing for this for years, but the NHRA won't listen to us. We'll spend close to $60,000 this year on valve springs for one car - that's a lot of money. We haven't changed our cam profile on any of these motors but the valve spring process has changed, and right now we're all having valve spring problems.

"We start making everything stronger with the camshaft, then the lifters, the pushrods and the rocker arms," Morgan explains. "Now we don't have a valve spring good enough to work with all that. It works, but you get no more than three runs on them, and it's got me devastated. I hate even to warm up the engine, because you're talking cycle time."

Fig. 1 - Larry Morgan uses intake valve springs that install at 525 lb at 2.300 in (Photo: Anne Proffit)

Written by Anne Proffit

Production car makers have long realised the benefits of low valvetrain mass. Low-mass valves and associated components require lighter springs to control the valve; these springs are shorter and therefore the camshaft height can be reduced. A smaller, lighter engine needs a smaller and lighter support structure and so the valvetrain mass reduction 'permeates' throughout the design of a new car to allow a significantly lower vehicle mass.

In a race engine, our primary target is more often performance than overall engine and vehicle mass reduction. Within an engine of essentially fixed mass, an effort to reduce valvetrain mass is still well rewarded by the ability to reduce friction a little, or to maintain a satisfactory level of valve control while using valve lift profiles with higher rates of acceleration. Certainly any friction reduction resulting from running lighter springs is welcome, but the real gains are likely to be the result of new valve lift profiles.

When we have exhausted the possibilities for lightening the stem by drilling, and have taken advantage of the optimum materials for our design, what options are left? Well, we can look to the remaining significant solid mass and reduce its weight. Hollow-headed valves are not new - the concept is older than most of us reading this (unless you are extremely ancient). Their original use was for sodium-cooled valves. The internal cavity, which was joined to the hollow stem, allowed the sodium to remove much more heat from the head of the valve than was possible using a drilled stem alone. However, in more recent times, their application to modern bespoke race engines has been limited by regulation on one hand, and by cost on the other.

NASCAR prohibits the use of hollow-headed valves, and it is possibly in NASCAR that the best use could be made of such a regulation, where the valves are physically large. In Formula One, where many avenues of technical development have been barred, hollow-head valves have remained a possibility; the only obstacle to their use has been the 'cut-off' provided by the homologated engine regulations. Engine makers that didn't have them when the engines were 'frozen' would find it hard to convince the FIA to allow them on grounds of cost-saving or reliability.

As the valve size becomes smaller, the technical difficulties of making an equivalent percentage mass saving are increased, as geometric similarities mean that wall thicknesses are decreased. The basic strategy of hollow-valve production is to machine the combustion side of the valve, and then weld a closing plate (or plates) in place before finish machining the valve. The methods of achieving a good weld need to be carefully considered. Laser welding, electron-beam welding and diffusion bonding would all seem to be good candidates for producing high-quality joints.

As we can imagine, the costs involved in producing such labour-intensive valves are considerable. There are a number of extra machining operations compared to a conventional valve, and there is a requirement for considerable investment in welding equipment and trained employees to use it. As such, even in race series where these items remain with the regulations, their use is very likely to be limited, if they are used at all. The number of valve manufacturers who are able to produce such components is naturally also limited by the size of the market for these items.

It should be noted though that the advantage of hollow-headed valves has not escaped the attention of roadcar manufacturers. At least one large manufacturer of poppet valves is working on hollow-head valves for road vehicles. The development doesn't take head-hollowing to the extremes that bespoke race valves would, and aims to increase valve cooling efficiency rather than presenting an avenue for performance development.

Fig. 1 - Hollow-headed valves are not a new concept; this shows a typical aerospace design used decades ago

Written by Wayne Ward

Roller lifters are more complex and have moving parts, but offer better engine performance due to the extra latitude given to development engineers in designing new valve lift profiles. The diameter of a flat lifter limits the opening and closing velocities of the lift profile according to the formula:

where r is the radius of the lifter, and w is the width of the cam lobe. Note that the formula assumes that the lifter bore intersects the camshaft axis, and that the middle of the lobe width is coincident with the lifter bore axis.

This is the same limitation which designers of overhead cam engines find when using flat tappets. This limitation can lead to unconventional valve acceleration profiles in order to maximise the air-flow into the engine.

Engines, while appearing to be very stiff, naturally move about under load and at temperature. Race engines are highly loaded, and high loads lead to greater deflections. In the case of the flat-faced lifter, camshaft and block deflections mean that the cam can be slightly misaligned relative to the lifter bore. This can lead to 'edge' loading of the cam - that is, where the theoretical line contact between cam and lifter is replaced by a point contact between the lobe edge and the lifter. This situation can cause premature damage to the face of the lifter.

There are a couple of design features that can improve the chances of the lifter not failing due to locally high contact stresses. The first is to carefully deburr the edges of the cams, replacing the sharp edge with a small radius. This does much to limit the contact stress. The second is to add a very slight dome to the flat face of the lifter. This gives no advantage in terms of increasing the available lift velocity, but means that the camshaft and lifter can accommodate a reasonable degree of dynamic distortion without edge loading being a factor.

Of course, this means that the contact is designed to be a point contact, but with the cam being flat and the follower having a very large dome radius, the contact stresses are not too high. The 'height' of the dome (measured from the edge of the flat surface to the top of the dome) is very small, and likely to be less than one thousandth of an inch.

Mechanically, the flat-faced lifter is simple, although it is a little more complex than the analogous flat tappet of an overhead cam engine. In many cases it has to transfer pressurised oil to the bore of the hollow pushrod in order to provide lubrication to the contact between the pushrod and rocker.

As we know, valvetrain stiffness is critical, and this is one area where the flat-faced lifter is superior to the roller lifter. With no bearings, shafts or shaft supports - all of which deflect to a greater or lesser extent - the flat-faced follower is much stiffer in compression.

Fig. 1 - Flat lifters, as might be found in a Sprint Cup engine (Courtesy of Comp Cams)

Written by Wayne Ward

The K&N Series travels to circuits around the coastal area and, like many in the NASCAR regional fold, accepts both 'spec' (crate) and 'open' engines for competition. In trying to lead competitors to the crate small-block, NASCAR has made it difficult for engine builders to be creative when using 'trickle-down' engines such as the Ford C-3, used in Nationwide Series competition nearly a decade ago.

According to Garrett Jacobson's chief operating officer Gregg Jacobson, his valve spring of choice in a very complex valvetrain system has been a double spring provided by PSI Springs; he's been working with the company for about five years, he says.

Jacobson specifies double springs of tool steel in light of the duration of valve opening, valve lift and acceleration. "We also take into consideration the size of the base circle and the stability of the lobe," he says.

In the build-up of the Ford C-3, Garrett Jacobson Motorsports carries out extensive testing of the valvetrain "to ensure that the system does not fail". Reliability is key, he says, particularly with teams that are on the road for long durations and not always able to return an engine for examination at the specified time of 800 track miles. K&N Series races can last anywhere from 50 laps to as many as 200 laps, and are held on both short and longer oval circuits.

While Jacobson does specify the particular part he wishes to use in the Ford C-3, he notes that the "spring manufacturer does an exceptional amount of research and development on their own on our behalf. That being said, it's still up to us to make the final determination that the spring we specify has to be stable on our cam profile." He makes this determination with in-house dynamometer and Spintron testing to confirm compatibility of the valve spring provided to him.

The double valve spring used in this particular application is not a bespoke unit. "PSI has been working with NASCAR race teams for a long time," Jacobson says. "This is to our advantage, because they have done the research and development on similar applications [before Jacobson's build-up of the Ford C-3]. There is no reason for us to reinvent the wheel in this particular case."

Jacobson says he takes the data from his own in-house testing and crosses it with PSI's "so we can come up with a viable part that will do a reasonable job for our engine project. If we come up with an issue during our own r&d process, we can always get in touch with the tech team at PSI to help us come up with a viable solution."

In the interest of building his Ford C-3 engine to compete against the crate engines, Jacobson says, "We do as much development work as we want to within this class. We are always looking for that competitive edge." While he may be a bit constrained by the rules and limits set by NASCAR, he says, "We can do what we want - as long as we stay within the limits they set for us."

Fig. 1 - The PSI double valve spring is a joint venture between the manufacturer and engine builder as they work to come up with the best compromise in a highly stressed engine (Photo: Anne Proffit)

Written by Anne Proffit

The subject of hollow poppet valves that are cooled internally through the use of sodium or a similar material has been discussed previously, both in Race Engine Technology magazine and in my RET-Monitor articles. The aim here is to transfer heat from the valve head to the cooling system via the valve guide.

However, where internal cooling isn't feasible or allowed under the rules, we have to look to other ways to keep the heads of the valves cool. Exhaust valve cooling is important for reliability reasons. Exhaust valve temperature is an important limiting factor for exhaust valve materials, and is the driving force behind the use of superalloy materials and new titanium materials for exhaust valve applications. It is also important to keep the head of the inlet valve cool, because the transfer of heat from a hot valve head to the incoming charge affects volumetric efficiency. What we want to do is transfer heat from the valve head to the cooling system as efficiently as possible. So, what can we do to achieve this?

Increasing the width of the valve seat is an effective method of increasing heat transfer, but this can have a significant negative effect on flow coefficients, and seat contact widths are generally kept to a minimum for this reason.

The use of high thermal conductivity valve seat materials is another effective way of increasing heat transfer. Beryllium-copper alloys are commonly used for this purpose, but beryllium-free alternatives are available that offer similar thermal properties. A range of more exotic materials offer even higher thermal conductivity, and computing and power electronics applications are, in many cases, driving the development of high thermal conductivity materials.

The use of thin-section valve seats can be used to reduce the distance and the thermal resistance between the hot valve and the cooling water, although this has some practical limits when using machined valve seat inserts. The practice of plasma-spraying valve seat material directly onto the cylinder head before machining is widely used in production car engines, but has only found limited application in bespoke race engines. Although the technique is very effective, it is also expensive when setting up to process a very small number of cylinder heads.

Water cooling of valve seats is another possible method, and this is commonly used in very large engines and piston aero engines, where the technique has been used for many years. Simply connecting a machined recess to the water jacket by using one or more drillings will have only limited effect unless there is a pressure differential between the inlet and exit drillings to the valve seat to encourage water to flow around the seat.

We also need to be aware of the possible negative implications of seating a hot valve onto a well-cooled seat.

The contracting of the rapidly cooled seat surface on the valve which results when the valve contacts the cool seat can lead to some valve distortion and tensile stresses in the rim of the valve. These cyclic tensile stresses can lead to fatigue failure in the form of valve 'chipping' and cracking.

Fig.1 - This is a typical water-cooled valve seat insert from a piston aero engine. While this is not suited to compact race engines, the task of removing heat from the valve is an important consideration

Written by Wayne Ward

For many years, engine designers were, in large proportion, blissfully ignorant of the behaviour of the valvetrain, partly because there was no method by which dynamic valve displacement could be measured, and no easy way to calculate the behaviour of the system. With the advent of both specialised software and test machinery, however, the very thin and flexible pushrods of years past have been supplanted largely by much more substantial items with greater stiffness. In terms of being in proper control of valve motion, these are undoubtedly the correct way to go, but moving in this direction can appear to lose performance initially.

The problem (if it is indeed a problem) is that, in changing to a much stiffer valvetrain, the motion of the valve can be changed substantially. In developing cam profiles to work with the more flexible pushrods, engineers are likely to have developed longer-duration cams, simply to cope with the valve closing earlier than the designed cam lift profile would have suggested. The high acceleration rates at valve closing cause the valvetrain to 'compress', with the pushrod becoming shorter, the arms of the rocker bending slightly and the valve stem also being compressed. The result of this is that the valve seats before we would expect it to, given an infinitely stiff system. The same effect pertains at valve opening, where the opening response of the valve will lag behind the assumed valve lift curve.

Those teams with access to a valvetrain dyno and the ability to measure valve displacement in real time are able to measure this effect. Providing the correct data is entered into suitable calculation software, the same effect is predicted, although simple hand calculations or spreadsheets should be enough to convince the mathematically inclined engineer that the effect on valve timing is both real and substantial.

If you have an engine for which you have developed cam profiles, and you don't have the resources for calculation software or a valvetrain dyno, then you have to take a decision about the benefits of moving to a stiffer pushrod. If you intend to gain more performance through better 'breathing', this will probably mean changes to more 'aggressive' valve lift profiles with higher rates of acceleration, exacerbating the symptoms of insufficient valvetrain stiffness. Changing stiffness at the same time as changing valve lift profile is likely to lead to one effect masking another, so if you have to persevere without analysis or a valvetrain dyno, in changing to a stiffer pushrod you should be prepared to go through the pain of developing your original cam profile to get you back to your starting point in terms of engine performance.

For those with a larger budget and access to analysis and testing, moving to a stiffer pushrod is largely a matter of understanding the implications on the valve lift curve, and the consequent effects on 'breathing' and working around this, while having the benefit of improved valve control.

Fig. 1 - Pushrods such as these, with wide shafts, are becoming popular as valve motion is better understood.

Written by Wayne Ward

Neff, the crew chief who helped Gary Scelzi to his 2005 championship, came to JFR in 2008 as a driver, with tuning duties performed by John Medlen. He earned the Auto Club Road to the Future (rookie of the year) award that season and took his first Funny Car victory in the season finale in 2009.

Neff then stepped away from the driver's seat and tuned Force to his 15th career Funny Car title in 2010, and when Force Hood stood down, he got the opportunity to race again - and tune his own car in the process.

Like most crew chief tuners in NHRA, Neff pays special attention to valvetrain issues and valve springs. And like the entire JFR team, he specifies a PAC double-spring steel valve spring.

"It's what most people run in this class, with an inner and outer coil," he says. "We run about 425 lb of spring pressure, with a little bit more on the intake than exhaust. They fall off a little bit on each run (between 10-25 lb of pressure) and the intakes last eight runs, sometimes more than that. We use the exhaust valve springs for as many as ten to 12 runs" down the 1000 ft dragstrip.

JFR has been running the same-spec valve spring for years, Neff says, even though the team has gone to a Ford/JFR-designed cylinder head. "We keep track of how many runs are on them, and know from past experience how many runs we can do before we encounter failures," he says.

Failures come in the Funny Car class as a result of the customary wear and tear. "The springs just wear out traditionally over time, but we've had some break in the past. They break in the middle and then the piston runs into the spring and you've got some big, big problems," he says.

"You've got to have some kind of cut-off point for all these parts, otherwise you have catastrophic failures, and you can't have that in this business." The same-spec 6.90 oz (195.6 g) valve spring has been on all JFR Ford Mustang cars for the past two years, according to Neff.

Everything seems to be going to plan for Neff and his Ford Mustang Funny Car. He is the first in his class to qualify for the NHRA's Countdown to the Championship six-race play-off series that begins following Labour Day's Mac Tools US Nationals at Lucas Oil Raceway outside Indianapolis. Leading his category into the Countdown is the current goal, so Neff is rather guarded about his tuning work.

JFR has used several manufacturers' valve springs in the past "but these PAC springs hold up the best. They're the best I've ever run on dragsters, Funny Cars or any other cars I've worked on," Neff says. It's a flat-top spring and "the thing is wound so it's got a lot of room in the middle for the high-lift camshaft we use. They're wound together right on the end, with the retainer so that they go up and down in a straight manner. They try to get crooked so we're very careful with how we manage them."

Fig. 1 - John Force Racing PAC steel double-spring valve spring (Photo: Anne Proffit)

Written by Anne Proffit

There are a number of surface treatments that aim to reduce friction and make the stem less susceptible to wear, and these have typically included heat-treatment processes such as nitrocarburising of steel valves and processes such as coating the stem with titanium nitride, chromium nitride or the ubiquitous DLC. All of these increase the hardness of the stem surface and lower friction compared to an uncoated valve.

However, none of these treatments fundamentally change anything about the lubrication conditions or help to produce or sustain an oil film. Study of a textbook on fluid film lubrication will reveal that flat surfaces in linear sliding will not generate pressure in an oil film (thrust bearings are an example of this). To generate an oil film, there is some change in the surface required, such as dimples, steps, grooves and so on, or some movement of the surfaces towards each other (squeeze film lubrication).

An alternative to coating valve stems would be to texture them so as to provide conditions more conducive to forming an oil film. Race valves are available which have textured stems, and the benefits in terms of lubrication are twofold. The unusual texture of the stem, in conjunction with relative movement, naturally creates conditions that produce pressure in a fluid film, but the texture can also help to retain oil within the small clearance between the stem and guide. The texture is made up of 'waves' in a regular and repeated pattern along the guided portion of the valve stem.

The problem with lubrication of a contact such as a valve and guide is that the motion is reciprocating rather than continuous, and this is not helpful in sustaining a lubricating film. Hydrodynamic lubrication, as the name suggests, relies on there being some relative movement between components. Even with texture, or other features, lack of relative movement means that a hydrodynamically generated oil film cannot be sustained.

You may be asking why there is any requirement for lubricating oil to be present when the valve is not moving. The answer is that the presence of oil is of benefit during the period when the hydrodynamic oil film is being re-established. Until fully established, the contact is one of mixed or boundary lubrication, with a mixture of dry and lubricated contact. The 'mixed' coefficient of friction depends on the coefficients of friction for both lubricated and dry conditions, and falls as fully hydrodynamic conditions are established.

To minimise friction during the mixed lubrication part of the valve motion, it is an advantage to decrease the dry friction coefficient of the stem/guide contact. The textured stems are supplied with a polymer-coated stem, which is often used in conjunction with an underlying hard PVD coating.

Fig. 1 - 'Snake skin' textured valve stems are reputed to offer significant improvements in terms of friction and wear of the stem/guide contact (Courtesy of Sinus Valves)

Written by Wayne Ward

As the use of valvetrain dynos such as the Spintron rig has become common, so people have realised the importance of component stiffness. Pushrod dimensions have increased markedly as a result in the search for stiffness, and a great deal of effort has also been put into augmenting the stiffness of rockers, with these parts commonly made in steel.

The more anonymous components can have an important effect on system stiffness, and here we must include rocker shafts and stands.

In order to be kept stiff, rocker shafts have to be short and well supported. In cases where they are bolted directly to their rocker stands, this is the reason why the bolts are kept as close as possible to the rocker. With the deflection of a simply supported beam being proportional to the cube - that is, the third power - of the distance between the supports, we can see the reason why it is important to keep the supports close to the rocker.

The diameter of the shaft is also important, the deflection of the shaft being inversely proportional to the cube of the shaft diameter. Therefore, even a small increase in shaft diameter is a very potent way of increasing shaft stiffness and minimising deflection in service.

The rocker stand has an important part to play in the pushrod valvetrain, and the structure supporting the rocker shaft needs to be designed with stiffness in mind. The stands are commonly made from aluminium alloys, but also steel can be used where stiffness is of particular importance. Titanium would be another acceptable material for this application, being intermediate in terms of both stiffness and mass when compared to aluminium and steel.

There is, however, little point in fastening a very stiff structure to one that is very flimsy in comparison, because we will find that the dynamic behaviour of the system is dictated by the stiffness of the least stiff component. We therefore need to ensure that both the rocker stand and the structure to which it is mounted are sufficiently stiff. Where a stiff rocker shaft is mounted to a well-designed, stiff rocker stand, but the rocker stand is mounted to a thin, poorly supported section of the cylinder head, the full benefit of the stiff shaft and stand won't be fully apparent. That is why we sometimes see some really bulky rocker stands which extend well beyond the area apparently required to mount it to the cylinder head. This is an attempt to transfer the valvetrain reaction loads to a stiff part of the cylinder head structure.

Fig. 1 - This rocker and stand assembly is very stiff. The steel shafts are very short, using reduced flange 12-point nuts to secure them, and the rocker stands are made of steel (Courtesy of T&D Machine Products)

Written by Wayne Ward

Wayne Gaskamp has been able to accomplish just that with his 1800-plus hp engine that propels Mike DeClark's F Bomb drag boat in two categories of National Jet Boat Association (NJBA) competition, the seven-second and unblown fuel jet classes.

The NexTek-polished Manley triple springs of lightweight tool steel have been through at least 12 races over the past two seasons and make 12-15 passes down the quarter-mile during each race meeting. There are many, many test sessions between events to hone the tune-up for both engine and boat.

"Every time I take the motor apart we look at the valve springs," says Gaskamp. "We check the springs for pressure and they've always been good, so what I do is change the keepers - they take the real beating - and the seals, lash the valves and put it all back together."

Installed pressure is 350 lb at 2.100 in and open pressure is 1010 lb at 1.200 in. Maximum lift is 0.900 in, with outside dimension of 1.677 in and inside dimension of 0.635 in. The polished spring is intended to reduce friction, improve fatigue life and minimise lobe loss. Apparently it's living up to its billing.

While Gaskamp runs his V8 engine at 7500 rpm (8000 at the hit, and a maximum of 7500 through the balance of the quarter-mile pass on the water) the spring is capable of handling 10,000 rpm, which would lessen its longevity proportionally. Installation in the drag boat gives the spring a softer usage, thereby prolonging life greatly for the spring. "They take the beating extremely well; they're really good."

Gaskamp says he hasn't lost any pressure in the past two years and believes polishing of the spring, rather than coating it, has a lot to do with the longevity, "Keeping the springs alive and not breaking the ends is the objective. On this engine I use this particular spring because we have a lot of lift. It has a lot to do with the camshaft, your set-up and your geometry. The geometry of these springs is right on for the rocker arms; it's just happy where it's at," says Gaskamp.

Even with all his success in the current configuration - boat driver DeClark set the performance record in the NJBA's unblown fuel jet class of 7.15 s Elapsed Time at 144.46 mph in Bakersfield, California, this past April - Gaskamp is looking ahead to possibly changing the spec on his valve springs. This would happen following the current six-race season, on hold now until September, where DeClark handily leads the unblown fuel jet category and lies second in the seven-second class for drag boat competition.

"We're looking at the newest double spring for possible use," says Gaskamp "The designated lightweight dual drag race spring's advantage is its lighter weight. It'll rpm (or rev up) easier because of the lighter weight and, of course, they're a bit smaller because they're a dual spring, but they have the same pressure as our current, triple spring. There are less harmonics involved, as well."

Fig. 1 - This triple-spring Manley valve spring has been in the 565 cu in V8 for two years (Photo: Wayne Gaskamp)

Written by Anne Proffit

In employing high velocity and acceleration for our valve-lift profiles, we subject the whole valvetrain to higher loads and increased levels of stress. Contact stresses are higher, and the loads in the mechanism driving the cams is also increased.

We often hear of valve control, or rather people working hard to avoid loss of valve control. Loss of valve control can encompass valve jump at the nose of the cam, where spring loads can become insufficient to maintain contact between cam and follower where the negative acceleration is too high, and there will come a limit for any cam profile where the mass of the valve and associated components and the speed of the engine conspire to overcome the available spring force. Contact is lost at this point, and this can be at or around the nose, but depending on the forces at work can be on the opening or closing flank of the camshaft.

For a given valve-lift profile that loses contact at a given speed, typical 'cures' are to increase the amount of spring load available or to decrease the mass of the reciprocating valvetrain components. Both are laudable aims, but we can go too far with both schemes. In the case of trying to decrease valve mass, people will often substitute steel materials for titanium, or move to a hollow-stem valve (or both).

It can be useful to consider the valve as a spring-mass system. The piece of the valve we are interested in controlling is the seating surface and, being part of the valve head, this is part of the mass on the end of the axially flexible valve stem. Of course, the valve head is also flexible, but the deflections in the valve head are generally low compared to the deflection of the valve head relative to the opposite end of the end of the valve, where the actuation takes place.

The valve head oscillates at the natural axial frequency of the valve. If we ignore for a moment any residual vibrations in the system, there is a strong deflection provided by the acceleration peak, which occurs just as the valve is closing. Where the stem stiffness has been markedly reduced through changing material or drilling - especially where the head mass has not been tackled to reduce mass - this deflection and the amplitude of any vibrations can mean that the valve seats prematurely. Valve lift profiles are generally designed to have a known valve seating velocity provided by a constant-velocity closing ramp. Early closure due to stem compression can mean that the valve closing ramp is not used, and high closing velocities therefore result.

Higher than expected velocity when the valve is seating can lead to valve bounce, which is literally the valve bouncing of its seat. This can have a negative effect on performance, but also poses something of a reliability risk to the valve. Other vibrations in the valvetrain can also lead to valve bounce - spring surge, for example.

It is probable that there are residual axial vibrations in the valve remaining from the initial acceleration peak when the valve opens. To remove one possible cause of valve bounce, we need to be sure we have not made the valve stem too flexible.

Fig. 1 - In order to avoid high-velocity valve seating and valve bounce, stem stiffness should not be reduced excessively in order to save mass (Courtesy of Sinus Valves)

Written by Wayne Ward

In recent years, NASCAR has allowed the various competitors in its Sprint Cup race series to design a bespoke engine, albeit based on the traditional pushrod valvetrain and with certain dimensional constraints such as in bore spacing, camshaft height and lifter diameter. This was done after deciding to allow Toyota to enter the fray with a bespoke engine, as it had no pushrod V8 in its production vehicle range on which to base a NASCAR engine.

However, if we look a little further back, every competitor's engine was based on a production pushrod V8, all of which would have been equipped originally with steel rockers. Steel is a material for which manufacturing methods are well understood and which is available in good quality at the right price. While the road engines carried on with steel rockers, race engineers looked to lighten the rockers and eventually made the leap to aluminium materials. Aluminium is much less dense than steel, and is therefore a very attractive metal from which to make parts that need to be moved back and forth rapidly. It remains a popular material for many race rockers.

NASCAR, however, is a 'big-bucks' enterprise, with huge crowds at the races and many millions more watching on TV. Naturally there is superb and evenly matched competition between engine manufacturers competing for an advantage and striving to put their cars in the winners circle.

In terms of rockers, what constitutes a good direction for development? Well, according to senior engineers at the top engine manufacturers, stiffness and low inertia are two of the primary goals.

Stiffness is critical to the success of the high-speed operation of the valvetrain. As the regulations effectively limit the engine speed by use of the 'gear rule', the development focus of the engine suppliers is to improve the breathing of the engine to liberate more power. As such, cam-profile development naturally works towards more aggressive valve-lift profiles, lifting the valve more quickly, holding it open for longer and slamming it shut more rapidly, thereby maximising the area under the valve-lift curve. Clearly this must be done within sensible constraints, but it remains the case that, owing to the high acceleration rates expected of the valve without losing control, stiffness is at a premium. Compared to steel, aluminium does not score well where space is limited.

In recent years, the balance has tipped again toward the use of steel as a material for NASCAR rockers. It offers a stiff structure, along with the possibility to remove the traditional pushrod-side adjuster screw. Even without any form of adjustment, aluminium would not survive the stress placed on it by the pushrod contact and so would require a steel insert. By designing the steel rocker with both stiffness and low mass in mind, it has now become the material of choice once more for the most highly developed race rockers.

Fig. 1 - This steel rocker is typical of those enjoying success in NASCAR's top Sprint Cup division (Courtesy of T&D Machine Products)

Written by Wayne Ward

Time Attack, which is a form of American road racing where the car-driver tandem compete for a time slip is "a bit of road racing with a bit of a drag racing element to it," according to Rob Cardona, WORLD Racing's manager.

WORLD Racing competes in the unlimited category using a 2AXFE 2.4 litre all-aluminium engine with a single turbocharger in its Scion tC. "Right now we're running about 1000 horsepower and maybe a little bit more; we can run up to 50 pounds of boost, depending on the atmosphere and the racetrack," said engine builder Gary Kubo.

Because the team designs its engine to run at a moderate rpm range, typically not exceeding 7800 rpm, Kubo ends up with a broad power band where, he said, "The sweet spot is between 4700 and 5700 rpm, and on some tracks with long straightaways we actually get to 7800. This way there's less cycles on the valve springs and obviously less rpm and less torture to the valvetrain."

Incredibly, the team is running a single beehive spring on its engine at 110 lb of pressure installed. "We're probably going to have to up that as we go up on the boost," Kubo said.

It might seem impossible to believe, but the team has used simple inspection procedures on its valve springs over the past few years. "We just check them and they've maintained the same pressure for the last three years," Kubo said.

The team has been using the same valve spring spec - equal for both intake and exhaust - for those three years. "Before that, we used OEM springs, which were good for about 700 horsepower. But because we didn't want to bust a retainer in half or something like that, we changed. We didn't want to take a chance because the boost pressure kept going up," he said.

At that time the team was running gasoline but it has since changed to methanol as its fuel of choice. "We decided to be safe and go along with the titanium retainer and a billet keeper," said Kubo. "This spring is made of a proprietary material from the manufacturer but they still manage to keep their pressure - even with the high temperatures coming with all our boost [of up to 50 psi]."

The rationale Kubo gives is that the rpm level is so low. "If the revs were higher we'd need a double spring to resonate the harmonics, but we've had no harmonics problems. We've never broken a valve spring, never dropped a valve, never lost a keeper, nothing."

Every time the engine is serviced, which is in the neighborhood of 30 hours, the team looks at the springs. "If we push an engine hard, that's about six races, and whether it's good or not, we're going to take it apart and check all components."

Kubo expects that the team will eventually be forced to go to a double spring. "But we haven't found the necessity yet, so why spend the extra money? That would be based on rpm and boost because, when you start running a lot of boost, the pressure wants to pop the intake valve open before the cam opens it, so you need seat pressure to hold the valve shut.

"That," he confirms, "is our biggest problem. On the exhaust side, mainly it's the closing of the exhaust valve that's critical because you want to make sure it shuts and cuts all the heat from the exhaust."

Still, most of the development work at WORLD Racing is based on the single spring; the team hasn't delved into 500-width cams. "That's because we'd have to build a 10,000 rpm engine, and that's more money," said Kubo. "We've found our sweet spot under 8000 rpm and that's more impressive than having an expensive engine where we'd have to replace the valve springs every session. That's nothing to brag about - it works so let's leave it!"

Fig. 1 - Beehive valve spring for the 2AZFE engine (Courtesy of Ferrea)

Written by Anne Proffit

We should briefly restate here the reasons for hollow valves. The first and more obvious reason is mass reduction in cases where a reduction in stem axial stiffness does not compromise the proper action of the valve or present problems through lack of bending stiffness. The second reason is to allow the use of an internal coolant, which is commonly metallic sodium. This has been a common measure for decades for some non-racing applications and its use is widespread in racing. Indeed, valves where the hollow stem is joined to a hollow head can be used where the rules allow.

One factor that every kind of hollow valve has in common is the requirement for the hollow stems to be closed. Naturally this is more critical in the case of internally cooled valves, where lack of sealing could allow the escape or oxidation of the liquid coolant, rendering the cooling effect less effective.

There are various methods for closing hollow valve stems. One of the simplest is to insert a plug, and is suited to hollow stems drilled from either end - clearly, it is possible to drill the hollow stem from the tip end or the head end. An interference-fit plug is a simple way to produce the desired sealing effect and, where the same material is used for the plug as the valve, the required interference is very small; it is also possible to weld a plug or plate to close the hole. Laser welding or electron-beam welding would be suitable methods for this operation, as both produce clean, consistent and very accurate welds.

Swaging is one method which is totally mechanical and requires no extra parts to be attached to the main body of the valve, either through welding or the use of an interference fit.

Swaging is a process of radial deformation, and basically the hole is squeezed closed. The method is commonly used in a number of applications, both in motorsport and general automotive use. Hollow driveshafts are often swaged to produce the 'necking' down from the large diameter of the hollow shaft to the splined ends. Clearly, before the swaging operation, there needs to be sufficient cross-sectional area to produce a solid stem. If we take the example of a stem with a 2 mm internal drilling and a 6 mm finished stem diameter, the pre-swaging diameter will need to be 6.235 mm, as shown in this calculation:

After the swaging operation, there will obviously need to be some machining, so the pre-swaging outside diameter will need to be somewhat greater than this calculated diameter. A side-effect of swaging is that, with it being a process of mechanical deformation, the material thus deformed will be work hardened.

Fig. 1 - Hollow valves are commonly used in racing but require a reliable way to close the drilling, especially where sodium cooling is used

Written by Wayne Ward

What we can say is that the faster the cam tries to lift the valve in an angular sense (measured in inches per degree, millimetre per radian and so on), the larger the distance 'swept out' by the cam profile across a given flat follower. Where the tappet size is the limiting factor for lift velocity, a common development is to use larger diameter tappets. However, in race series such as NASCAR Sprint Cup, the diameter of the tappet/lifter is limited to a maximum value in the regulations. In the case of Sprint Cup engines, this diameter is 0.875 in (22.22 mm). Lifting at too fast a rate will see the cam contact reach and go past the edge of the camshaft. By going too far over the edge, we can reach a point-contact condition that is very bad for friction and wear, owing to poor lubrication conditions and high contact stresses, which can lead to subsurface fatigue.

Limiting maximum lifter diameter is one method by which engine performance may be limited, or certainly aggressive development in the direction of increased lift velocities is curtailed. People will always try to push these limits, by taking the cam/lifter contact as close to the limit as possible, but also looking to maximise the cam angle over which high lift velocities are maintained.

Where rules don't limit lifters to be of the flat, solid type, people will generally turn to roller lifters, as the relationship between lifter velocity and contact point is very different. As the name suggests, a roller lifter has a rolling element incorporated in the lifter foot. These need to be carefully aligned so that the axis of the roller is perfectly parallel to the axis of the camshaft. They don't need to be constrained tightly to achieve this as the action of the cam on the roller tends to self-align the lifter perfectly to the cam. Generally though, they are constrained within certain limits, and typical methods of doing this are keyways, tie-bars and flats on the lifter which align with corresponding features on adjacent components. The devices guard against the tappet axis finding itself perpendicular to the camshaft axis at engine start-up where damage might occur.

Keys fitted in the lifter require a corresponding keyway in the lifter bore, slightly complicating the machining of the cylinder block, but these are very popular, given the very small amount of extra reciprocating mass required to prevent lifter rotation.

'Dogbones' or tie-bars, which are a connecting 'link' between adjacent lifters, are an easy fix to the rotation problem and can run in plain cylindrical bores, requiring no special machining on the block. The do however carry a larger weight penalty and so aren't often used for highly optimised race engines.

Flats machined onto the lifter can be a simple method of restraining rotation, and there are a number of ways in which these can be used to prevent lifter rotation. Corvette for example uses a 'lifter tray' with corresponding flats in its production vehicles.

In the NASCAR Camping World Truck series, roller lifters are mandated but, given a choice, everyone would use these anyway as they allow much more aggressive lift profiles without the risk of component damage. In the real world of production automotive engines, where there is no restriction concerning maximum lifter diameters and no obstacle barring the use of roller lifters, the flat lifter is an anachronism and they are now only very rarely used.

Fig. 1 - A tie-bar or 'dog-bone' is an effective way to align roller lifters with the camshaft, but they carry a weight penalty (Courtesy of Comp Cams)

Written by Wayne Ward

For Bob Vandergriff Jr's C&J Energy Services National Hot Rod Association Top Fuel dragster, this year competing full-time in the Full Throttle Drag Racing Series of 22 events, the choices are predicated on a change to the set-up configuration for the 2011 season.

Working to gain reliability on the dragstrip and to be competitive with teams that have had full-season set-ups far longer than they have, Vandergriff's team is spending the first few races of the year trying to find a happy combination that will work best for their engine.

The objective is to be aggressive enough to qualify and to make the field of 16 for Sunday eliminations, but also to keep parts in the engine from ending up as a molten mess after a run. For these reasons, according to crewmember Rob Hauser, the team has chosen to use double-spring PAC valve springs in the current combination as they find the sweet spot.

"The objective is not to beat up the valve springs but to take them out before they've lost their longevity," he told me. "We test them every run while the head's being serviced, check each one on our spring checker and, as long as the spring maintains the correct level of spring pressure, we'll continue to use them."

The steel double spring the team has chosen - for both intake and exhaust valves - has 895 lb/in pressure at production. Installed pressure is 550 lb and, when open, the spring generates about 1150 lb, he said. "Once it falls below that level, we'll take the spring out and set it aside. We may then sell it to an Alcohol team or a smaller-budget team to offset some of the costs.

"We use a double spring instead of a triple - it's two separate springs, instead of having a damper in between like the titanium springs," Hauser explained. "We like this steel spring - it seems to be 'living' quite nicely - and they last longer than the titanium springs, maintain the pressure and cost a heck of a lot less." Weight is a bit higher, at 5 oz (141.75 g), but the smaller diameter offsets this additional weight of steel versus titanium.

The valve springs in use at Vandergriff Motorsports are a standard spring and are straight-cut parts, not a conical or beehive valve spring. "In the older spring we used, in order to get the spring pressure we needed, it had to be a triple spring, but now the spring is actually physically taller than we used to run. The wire diameter is smaller now so we can still get a decent coil bind to allow us to have the camshaft the size we wish to run without having problems with coil bind."

When coil bind occurs in a Top Fuel engine, the valve fails to open and can break pushrods and/or rocker arms. "When one coil touches the other and they're all squished together, we get breakage on the opposite side," Hauser said.

"Since this valve spring is taller, with about half again more rotation than what we used before with a triple spring, we get higher spring rates so that when we do squeeze it down, it goes right back to its 550 lb operating pressure. In the past, the larger spring would only give up about 425 lb of spring pressure, but the two of these together work out to 550 - this is enough to get the valvetrain right where we need it.

"We're using this valve spring as part of our new tune-up for the 2011 season, and we're very happy with the service life we're getting. We had instances in the past with premature failures and that would cause catastrophic failures, which are mega-expensive," Hauser said. "It costs enough just to maintain our engines without catastrophic failure - that adds a whole new dimension to the situation."

Fig. 1 - Vandergriff Racing is using a double steel valve spring this year to complement its new tune-up configuration (Photo: Anne Proffit)

Written by Anne Proffit

requires a smaller supporting structure, and the resulting car is much lighter and more fuel efficient, and requires less outright engine output to perform well.

By turbocharging any engine, we increase cylinder pressures and temperatures, and in doing so we make increased demands on a number of components, from spark plugs to pistons to bearings and valves. The increased operating temperatures for valves in turbocharged engines means that material selection and management of heat become more critical.

Many turbocharged engines will use valves made of high-temperature 'superalloy' materials. Inconel and Nimonic are commonly used, and these superalloy materials retain a large proportion of their room-temperature mechanical properties at high temperature. Beyond certain temperature limits, the more usual valve materials, such as titanium and the martensitic and austenitic steels lose a large proportion of their ambient temperature mechanical properties, limiting engine performance. There are a lot of high-temperature materials to choose from, and motorsport benefits from the development of materials for aero gas turbine engines by seeing new materials coming through which are suitable for our use.

The management of heat in the valve, in any type of race engine, is critical to the durability of the component. Hollow valves with an internal coolant are very successfully used in many race series to remove heat from the valve head, transferring it to the valve guide and thence to the engine coolant via the cylinder head. The internal coolant is often pure metallic sodium, although others have been used with success.

Some valve guides are directly liquid-cooled, intersecting liquid-cooling channels in an attempt to make the transfer of heat more efficient.

In valves with no internal coolant, a large proportion of heat energy is transferred via the valve seat interface during the period when the valve is closed. A wider seat improves heat transfer, but this can be at the expense of flow coefficients. The choice of seat insert material is important, and materials combining good strength with high thermal conductivity are ideal candidates.

As with much in engine design though, improvements made to the cooling of the valve do not come without penalty. Improved cooling via the valve seat causes the tensile stresses in the periphery of the valve head to increase, and this can have an effect on the durability of the valve, and in some cases can lead to the appearance of fatigue cracks.

A number of people are reported to have used, or to be using, thermal barrier coatings on the face of the valve to minimise the heat transfer through the valve and to reduce valve operating temperatures.

Fig. 1 - This valve has a hollow stem; partly filling this with a coolant (typically sodium) improves the transfer of heat from the valve head (Courtesy of Zanzi SpA)

Written by Wayne Ward

So, what are the advantages of producing a rod in three pieces? Well, there are many. A primary reason is one of economics - particularly relevant in recent times.

The pushrod ends need to be hard and wear-resistant, being durable without causing damage to their adjacent parts during sliding contact. The ends are quite often made of very hard steels, commonly tool steel, and the cost of producing complete pushrods in tool steel can be prohibitively expensive, especially given that there is no requirement for the material to resist high contact loads except at the ends of the pushrod.

The central part of the pushrod can be made from a cheaper material that simply fulfils the requirements of being sufficiently strong and stiff. Furthermore, an assembled pushrod has the advantage that the ends can be subjected to special surface treatments aimed at improving wear characteristics which may not suit the centre section, or which would prove an expensive option if carried out over the whole length of the pushrod.

A single design of centre section can be fitted with any one of a large selection of ends, and these can be different at each end, compatible with the mating parts (lifter and rocker). A selection of different pushrod end designs is shown in the photo here.

This means that a special-purpose pushrod, compatible with components from different suppliers, can be quickly assembled without needing to resort to a bespoke design. Depending on the design of the ends and their method of fitting, it may be possible to exchange end pieces that show signs of wear. Some people choose to 'refurbish' pushrods in this manner; others choose to retire rods whose ends have become worn.

Pushrods can be assembled to the correct length quickly by holding an inventory of different lengths of centre sections to which the ends are fitted. Again, this is economical compared to holding a wide stock of solid single-piece pushrods.

Having made the economic case, there are a number of high-profile and high-budget engineering-led race engine suppliers who use three-piece pushrods - NASCAR Sprint Cup included - so the engineering case for this type of construction remains strong.

There are others for whom the economic considerations are not so strong, and who see benefit in single-piece rods, machined entirely from a material with the correct specifications to resist wear in the highly loaded contacts at each end, although they often come at a significant premium.

Single-piece rods cannot be assembled wrongly, however, nor do they suffer from the assembly becoming loose over time, and compared to a three-piece rod of the same geometry they should be slightly stiffer. Since the advent of widespread 'Spintron' testing, pushrod stiffness has been found to be very important to proper valve control, and NASCAR Cup engine builders and others have moved to much stiffer designs in recent years.

Fig. 1 - Part of a wide range of pushrod ends fitted to 'standard' centre sections (Courtesy of Manton Pushrods)

Written by Wayne Ward

By the time the engine was completed it developed 698 hp and 670 lb-ft of torque. "When we were running the stock engine with a higher boost, we noticed the valves were getting tweaked a little here and there, but what do you expect with a stock unit putting out that kind of horsepower?" says team manager Eric Cantore. The Lambda V6 normally produces about 300 hp but, says Cantore, "We've doubled that and then some."

Even with the original equipment engine, RMR found very little wear on the valve springs. "We were pretty much using all the elasticity available. We basically went back to that very soft material and the valves were starting to wander a bit [when the engine was upgraded]," Cantore says.

Still using a stock valvetrain, RMR needed to get the right valve spring for its larger capacity engine. Cantore called Ferrea Racing Components of Fort Lauderdale, Florida, for advice. "They said they could make things a bit lighter and stronger, and gave us a different set of springs. We tried it out, ran it for three events, took the head down and checked the springs, and everything was right back to new," Cantore says.

"Spring pressure was at the same 180 pounds spring rate - the stock springs are 80 pounds. The springs retain pressure throughout and the retainers are beautiful. The valves are straight with no pinging on them; we use a solid valve, not hollow material. Ferrea have been able to bring in materials that are almost as light as titanium but without the cost and any further worries," Cantore says.

Because this is a turbocharged application, the exhaust side retains heat, so RMR uses a spring developed from high-strength alloys, focusing on premium-grade chrome silicon steel. The springs are heat-treated and stress relieved to increase spring life. "We're still using a high-strength alloy - a silicon steel - but we expect to work with a tool steel mixed with different alloys once it's tested," Cantore says.

The 30 gramme valve spring is a single spring with titanium retainer that is CNC machined from aerospace-quality titanium. The retainers are heat-treated as well.

"We only rev the engine to about 7600 rpm, which is not very high. It doesn't need to be because we see our power start to taper off at about 7200, so a larger spring isn't necessary," Cantore says, adding that they've had no problems with the small, single spring and that the valves look good after each use.

The team does leakdown and pressure checks after every event and used these springs for five contests, including a Brazilian hillclimb last December. They intend to refresh all valvetrain equipment before the first Formula Drift event at Long Beach, California, in April.

The springs have a flat top with a bucket tappet and shim underneath. "We gave Ferrea a cylinder head and they went and cut the valves, floated it and told us what would be best for our application," Cantore says. "The valve spring worked great at Pikes Peak and in all of our Formula Drift events."

The valve springs are of equal size for intake and exhaust. "With the lower revs they don't see any spring pressure loss," Cantore says.

Fig. 1 - Ferrea valve spring and retainer (Photo: Anne Proffit)

Written by Anne Proffit

Among these exhibits, there is sometimes a single, horribly distorted part, proudly displayed. What possible use can this be? One of the primary requirements for a valve to operate in a satisfactory manner is for it to be very straight. Straightness tolerances are measured in microns over the length of the valve, yet these specimens are severely twisted.

The reason a valve manufacturer or supplier wants you to see the twisted valve is to prove to you that their valves are possessed of a large measure of ductility - in other words, the valve will behave plastically beyond yield. This is of little consequence in normal operation, but can be a real saviour in the event that the unexpected happens.

There are many reasons why a valve might see loads that are far different in terms of magnitude and direction from those we can expect in normal service. Normal operating loads come from the valve being opened (stem compression) and closing - tension in the stem and seating loads, which may be compressive, tensile or bending depending on the location in the component. There are also some loads due to friction in the valve stem. However, none of these require a large degree of ductility.

When something goes really wrong - for example, the inlet swallowing something it shouldn't have, such as a screw in the inlet becoming loose and finding its way down the inlet port - valve loading cannot be predicted. Such debris can come between the valve and its seat as the valve is closing, or it can clatter past the valve and into the combustion chamber. It will inevitably lead to huge bending loads in the valve, causing plastic deformation.

If the valve head then breaks off, the problems are compounded, and there is then a greater chance that the other valves will fail and that one of the pieces will subsequently puncture and pass through the piston. Where materials have little plastic deformation beyond yield before they fail, we commonly term these to be brittle, and overloading will cause them to fail by rupture.

If the valves are ductile and remain in one piece - albeit twisted - the damage to the engine can be limited, and perhaps restricted to replacing the cylinder head in question and a cylinder liner. Where the valve head becomes detached, our next line of defence is the piston, but if this is breached and the valve head reaches the bottom end of the engine, the engine can effectively be scrapped, requiring the replacement of major structural components.

Fig. 1 - The damage to the engine would probably have been much worse if these valves weren't so ductile.

Written by Wayne Ward

What came to mind is what this statement means in regard to other engines. The reason I started thinking about this was a statement I came across in an article in Race Engine Technology concerning pushrod systems. The comment was that valvetrain systems using pushrods are at no disadvantage compared to overhead camshaft systems with direct valve actuation in those race series which, by their regulations, have intake restrictions, generally restricting rpm as consequence.

So, when one takes the high-revving part out of the equation, what would be the definition of high performance? In the case of a valve actuation system, high performance most definitively means stability of the system in relation to its loading. And this is where it becomes interesting to compare engines in entirely different corners of the automotive industry.

In issue 29 of Race Engine Technology (March/April 2008) an interesting article on pushrod system technology was written by Ian Bamsey and Wayne Scraba. The goal of the article was to show which parameters of a pushrod valvetrain system directly influence the system's functionality and stability. In the article, exhaust pushrod loading in an NHRA Top Fuel engine was mentioned, being in the area of 7000 lb force. From my experience with other performance engines - it will become clear below what that performance engine experience is - I very well know that having these kinds of loads means the valvetrain components need to be robust, and the system's behaviour very stable, for there not to be any collapse.

Which other engines in the industry also use pushrod valvetrain systems and show comparable loading of the valvetrain? For one thing, there are not many engine designs that can be compared here. A number of heavy duty (HD) truck engines still depend heavily on a pushrod systems due to its compactness in regard to engine size. Because of their required normal engine life of about a million miles, one might think these Kings of the Road are not as highly loaded as Top Fuel Engines, which by comparison are required to have only a marginal engine life. However, and this is where I try to translate my experience from HD truck engine development to the world of race engines, pushrod loading of the Top Fueller is at about the same level as the HD truck engine.

The major differences in the introduction of the loads between the two engine types and applications are combustion and rotational speed. Looking at rotational speeds, it can quickly be seen that the NHRA Top Fuel engine, running near to five-digit speeds, has clearly the more difficult load case. When considering the combustion loads, however, it's a different story. Based on the charging pressures, one can compare 5 bar for the Top Fuel engine with compressor with 3.5 bar for a HD Truck engine (single-stage charging, as is still the case in most of these engines).

Although from a relative factor the percentage of valve diameter to piston diameter of the NHRA engine is far larger, with 40% over 30% of a HD truck engine - not that strange, given the swept volume difference of 8.2 litre V8 versus a 13 litre-class inline six-cylinder engine - the absolute diameter is about the same, at 40mm. Assuming combustion pressure over the four-stroke cycle and exhaust valve timings, the pressure on the exhaust valve could be determined. I do not have pressure traces for a Top Fuel engine, so it is difficult to provide comparisons at the moment, but it would be interesting to follow up on this. If there is any data out there, let me know.

But, and this is where the comparison goes off a little, a HD truck engine has a engine compression brake functionality that relies on opening the exhaust valve at the moment of highest compression, in order not to get energy back when the piston goes on its way down again. The loads introduced exactly at the moment of start actuation are far higher than in normal engine operation, but equal to Top Fuel pushrod loads.

So, having explained the differences in application and load cases, the detailed environment and load cycle might not be fully comparable, but looking at the possible consequences of these loads, the required engineering solutions are exactly the same, mainly discussing cam follower roller robustness and pushrod dimensioning.

I would like to look a little further into the pushrod dimensions, just to realise where the differences are and what the reasons behind the design choices are.

The above comparison would mean that the pushrod valve actuation system of the Top Fuel engine would, dimensionally speaking, need to be more or less the same. However, if we look at the dimensions of the pushrod itself - Top Fuel , 14.3 mm x 4.8 mm (0,562 in x 0.188 in) and HD truck, 18 mm x 3 mm (about 0,687 in x 0.118 in) - then we can conclude that there are clear differences.

The Top Fuel engine has a smaller outer diameter and, as a consequence, a higher wall thickness. A principal look at the buckling resistance (Inertia goes to the fourth power), only on diameters, will show that the HD truck pushrods are significantly more robust than those on the Top Fuel. Assuming these components for both the Top Fuel and the HD truck need to be designed for infinite life, this would mean that the pushrod material specification must be significantly different.

Looking at the material specs confirms this. For example, heat-treated 4135 (1.7220 EU) is 1900 N/mm2 (275 kpsi); normal cold-drawn St37 (1.0037 EU) is about 300 N/mm2 (44kpsi) - a significant difference.
Looking at the dimensional difference of the two engines, 8.2 litres against 13 litres, and the bore size of roughly 100 mm to 130 mm, that would mean the available space for the pushrod would be, as rule of thumb, 100 / 130 x 18 = 14 mm.

This brings us back to the start of this article, where I started from comparable pushrod loads to be withstood within the maximum available clearance in the engine structure. Due to the smaller size of the Top Fuel engine, materials with higher material properties are an absolute must.

So, putting the statement 'high performance' in perspective, it is nice to see that, as with many other examples, current HD truck engines, running for enormous amount of hours, are almost race engines, being produced as cost-effectively as possible. Think of it as a Top Fuel car pulling a trailer over the highway.

Fig. 1 - Typical pushrod system (Image from Wikipedia UK, originally uploaded by IJB TA , GNU Free Documentation License)

Written by Dieter van der Put

Dunn is pretty much of a throwback to the earlier days of straight-line competition, a man who prefers not to use computers in his diagnoses but rather to let his innate knowledge dictate the manner in which he attacks tune-ups for a 1000 ft catapult down the dragstrip in an 8000 hp nitromethane-burning Funny Car.

Jim Dunn Racing's Flopper, sponsored by dog-food producer Canidae, uses valve springs produced by Noel Manton of Manton Engineering in Lake Elsinore, California. Collaborating with Manton over a period of about 15 years, his choice is a tool steel triple spring. "We have to have triple springs to get the spring tension we want on the seat," Dunn says.

When initially installed, the valve spring - produced by pushrod manufacturer Terry Manton's father Noel - has about 550-600 lb of spring pressure. "They'll lose about 25-30 lb on the first run and then go down each run by 'X' amount, from the heat and vibration." The valve spring is not terribly dependent on Dunn's tune-up. "It doesn't kill them that much," he says.

Because he's running on a budget that would leave many amateur drag racing competitors in shock - Dunn operated his professional team on less than $1 million over the 23-race 2010 season - he has to be very careful where the money is spent. To be on the side of safety, he and his volunteer crew examine valve springs after every run to make sure the seat pressure isn't below his specified number.

Dunn says he gets somewhat less than the 20-run figure quoted by much of his competition, which has more money to spend on the spring and its ancillaries. And he has developed his own way to get his valve springs to last longer in his high-compression engine, which he runs at a standard 6.9:1 level, the upper reaches of the class.

"When they come down to 425 [lb], we'll put a 60 thousandth shim under it and that brings it up past 500 lb. When they come back down to 425 again we'll take them out because they'll bind up," he says.

Dunn's four crew members will check both intake and exhaust valve springs with every pass down the racetrack. "We'll probably go through one or two each pass and it's not bad at all," he says. Using the same size and spring pressure for both sides of the engine, Dunn says that about 90% of his competitors run the same type of spring he does.

"The rich guys run titanium on the intakes, and they cost about four times as much as the steel," he says. "What's bad about the titanium springs - for a guy on a budget like me - is that they don't go away. Titanium has a memory so they'll keep coming [back] up. After about 20-25 runs, though, they'll disintegrate and [the resulting failure] goes all through your motor and you get a 'big one'. Now, with our springs, you measure them and they're shorter because they start collapsing." It's his best opportunity to keep an eye on wear.

Valve sprigs don't break very often, Dunn says. "I think I've broken one in the past five years - and that was this year! It was the first one I ever broke in the centre, but of course, those [flat] ends will break off." But that means nothing in the scheme of things to Dunn.

When he has used a valve spring to the fullest extent, Dunn says, "I'll take them out and give them to the Bonneville guys. Heck, they run 350 lb of seat pressure, so we take out the shims and those guys have a new set of valve springs."

Fig. 1 - Manton Engineering triple valve spring, as used in Jim Dunn Racing's Funny Car (Photo: Anne Proffit)

Written by Anne Proffit

There have been a number of uses of reed valves on four-stroke engines, some of which have been in the induction system, meeting with various degrees of success or lack of it.

UK-based Performance Bike magazine built a single-cylinder, four-stroke engine with a reed block in the intake some 20 years ago. The ethos of the machine was to provide something light and responsive. A new water-cooled head was commissioned from Silk Engineering in Derby, England, to fit on an existing engine, and the idea of the reed was to improve the engine's volumetric efficiency at speeds away from the 'tuned' speed of the engine by not allowing charge to escape back down the inlet.

However, the combined effect of having a large restriction in the inlet and something that so fundamentally affected the way the pressure waves behaved meant that the engine didn't behave as expected when it first ran. A lack of enthusiasm and money from the magazine consigned an interesting (but possibly technically doomed) project to the dustbin, and the whereabouts of the engine isn't known.

An interesting engine operating on similar principles is being developed by an enthusiastic engineer in Sweden, on very limited funds. The intake of the 'Ellwood Hybrid' 500 cc engine breathes via the crankcase, induction to which is controlled by reeds. As the piston descends, it compresses the charge in the crankcase and this is admitted to the long intake port using yet another reed.

There are, of course, two piston strokes that serve to compress the charge and so, while the intake valve is closed, there is a store of lightly compressed air trapped between the final reed valve and the conventional poppet intake valve. The engine, based on an old Godden speedway engine, has been run in a motorcycle and campaigned in circuit racing, drag racing and ice-speed record racing.

The original 500 cc version of the engine has been developed over 15 years, but a 1300 cc version is now also being built. The problem with this latter machine though is that the engine charge has to pass through an otherwise conventional four-stroke bottom end, replete with oil. The charge therefore 'picks up' a great deal of oil that is then carried through the rest of the inlet tract and into the combustion chamber, where it will be problem both for combustion and emissions.

The technology would perhaps benefit from using two-stroke oil in the bottom end, as the oil is designed with combustion in mind. However, in most two-stroke applications, the bottom end is essentially dry at start-up and oil is constantly metered to the crankcase, and is proportional to throttle opening. Simply using two-stroke oil in the crankcase in place of conventional lubricating oil would not solve the problem of excess oil in the mixture.

Readers wanting to see this novel engine in action can do so at the Swedish Landracing Speedweekend on Ice, 11-12 March 2011, where the 500 cc Hybrid will be running on nitromethanol.

Fig. 1 - The 'Ellwood Hybrid' 1300 cc engine, currently under development, uses reed valves in multiple locations in the inlet

Written by Wayne Ward

Often the reason is product cost. These race series rely on performance parts that are not as sophisticated as the high-level series, or not even production components. That's not to say these components are any worse, but what could be the downside of using them in your valued race engine?

In this article I want to focus on friction-welded pushrods, specifically on the welding process. As stated in my earlier RET-Monitor article, pushrod ends are typically pressed into the pushrod tube to achieve the function of the pushrod, actuation of the valve train.

When using a pushrod with pressed-in rod ends, the press fit can become loose, allowing the rod ends to hammer the tube when the engine is running and risking engine damage, potentially ruining the engine. To make sure this can't happen, it is common practice to weld the rod ends to the tube after press-fit of the rod ends. Both laser and friction welding can be used. The welding itself is a critical process, as the cup is a hardened, high carbon-content component that is not ideal for welding.

An alternative manufacturing process could be to reverse the welding and heat treatment processes, which means hardening the rod ends after welding. The drawbacks with this though are that the rod-end hardness will be lower than with loose rod ends and that it requires a significant investment and increase in product cost. Based on the assumption that we are looking for a cost-effective alternative, the overall volume of high-performance pushrods does not allow for these kinds of investment levels.

Friction welding is mostly chosen when the priority is product cost. Compared to laser welding, however, it is not as easy to control, and in order to ensure sufficient quality of the weld, proper process control is essential at the end of the manufacturing process.

A typical test procedure consists of applying a rotating bending load onto the rod ends. This allows the total welded areas to be loaded and maximises the chances of the more severe internal defects being found.

A comparison of this test procedure can be made with the functional loads on the pushrod, where a radial force acts on the rod ends due to the layout of the valvetrain systems and the arch-shaped movement of the rocker. What needs to be taken into account, however, is the fact that the test loading required can be physically applied to the rod ends.

The rod itself needs to be held firmly in position at the moment the test load is applied. Tube material has lower mechanical properties and therefore is not as strong, so there's the risk of the tube being damaged or even deformed during the rotational bending test. On the other hand, when the loads are reduced so as not to damage the tube, one might doubt the representativeness of the test procedure.

This test procedure is called non-destructive. That means the part will not - or, as mentioned, should not - be damaged during the test and can therefore be sold to the customer.

Also, in order to be able to determine the correct test boundary conditions such as test load, load direction, rotational speed, amount of revolutions in the test, test duration and so on, the load at which the part will actually fail needs to be known. To gain enough information and experience, a destructive test may be needed. The goal of a destructive test is always to find out the circumstances in which a part fails, after which the results can be compared to the initial design guidelines and assumptions.

So, given that the focus here is on the friction weld rather than the rest of the pushrod, we should ask ourselves which test will provide sufficient information about the weld itself. To investigate the welded zone, a tensile test can be performed. This will provide a good insight into the strength of the welded zone and which area is most critical.

For example the detailed shape from tube to rod end, including wall thickness and the original press-fit area, can be analysed. If the rod is designed and welded properly, the heat-influenced zone will suffer from an initial crack under tensile stress in the pushrod, just before another area starts to fail. If the welding process is not yet as stable as required, welding defects will result in preliminary failure of the part. As long as this component does not find its way into a high-performance race engine, this can be seen as an advantage and possibly improve the product and process.

My goal here has been to focus on the processes following the pushrod's design phase, and not limited to the manufacturing process. Quality control is also just as important in order to ensure that only quality parts leave the supplier. The example of the friction welding process should make engineers aware of the fact that knowing the loads and testing them are two different things. And since the race engineering industry consists not only of high level professional race teams but many race enthusiasts as well - who often use standard automotive products - I hope quality will not be taken for granted.

Fig. 1 - Pushrod breakage after tensile quality testing

Written by Dieter van der Put

"For the past year-and-a-half, the majority of our racing valve springs, retainers and spring seats are coming from PAC Racing Springs. The valve spring is a double spring with a damper insert," he says. "We use the same spring for both intake and exhaust applications because we have a bit more intake lift than exhaust lift over the nose (spring pressure)."

Buying an off-the-shelf part from PAC - or any other vendor - means building trust with the manufacturer. "Sometimes the manufacturer's suppliers don't always have the same metals, and there are some aspects in the making of a valve spring that are not totally under control," Hardesty says. "In manufacturing there can be some digressions (that can ultimately kill an engine)."

The current spring that Shaver uses for its World of Outlaws engines for four-time champion Donnie Schatz and his True Speed team, which is owned by NASCAR Sprint Cup star Tony Stewart, is distinguished by its raised seat. "Instead of the spring seat being flat, it's got a little bump. The normal 100-thousandths or so on the retainer is now shared by the retainer and the spring seat, so that takes a little weight out of it," Hardesty explains.

"The reason for using the lightweight steel retainers instead of titanium is our damper," he says. "The damper tends to tear up the titanium, so we were always having to replace the titanium. The steel ones, while more expensive, last three times as long."

On its state-of-the-art engines, SSRE requires rebuilds every ten to 12 races, or about 400 miles. "Valve springs are now seeing 20-24 races, and that's unheard of in this business," Hardesty says.

This 164 g PAC spring meets increased pressure over the nose from what Shaver's engine group used several years ago. "We used to run about 700 lb over the nose and now we're seeing about 800 lb of pressure," Hardesty explains.

"In sprint cars we see a lot of uncontrolled valve motions because they start slipping the tyres. A driver might be running 7500 rpm and something happens on the track - suddenly he goes to 8500! We've got to have some safety factor in these big parts."

Part of the reason for this incredible longevity is due to lighter weight of the intake valves, and the rest comes from the use of a 300M special retainer. Here, Hardesty says, "The half of the step for the inner is taken away and we've put that half on the spring seat itself."

While initial cost for the steel retainer is a higher, the overall cost is lower because it's not necessary to buy a new set of springs each time the engine comes in for its rebuild.

The sole drawback to the spring currently used by Shaver Specialties is the need for better lifters, and for that Shaver specifies 905 mm Jesel lifters.

"The stuff we were using five or ten years ago won't stay in - the bottom of the lifter is where the roller bearing swells - and then the clearance would stop," Hardesty says.

Shaver hasn't changed its spec on springs for more than 18 months.

Because engine builders compare notes on what they've tried or seen in their travels, Shaver Specialties learnt about this particular spring from one of its competitors. "While we were all whining about spring life, they said they'd tried this spring on one motor and that it worked pretty good," Hardesty says. "It's actually a drag racing spring made for severe duty over a short period of time, and while we were hesitant to try it, we did so out of desperation and it's a really, really nice valve spring."

Fig. 1 - Valve spring with intake and exhaust valves, seats and steel retainers.
Fig. 2 - Valve spring with steel retainers
(Photos: Anne Proffit)

Written by Anne Proffit

Two-stroke engines have been used widely for grand prix motorcycle racing (although this era is soon to come to an end), motocross, snowmobiles, jet skis and so on, and are currently enjoying something of a resurgence for applications where low fuel consumption is of importance. Just as the motorcycle race classes and the bike manufacturers are turning away from two-strokes due to emissions, other companies are embracing the technology.

Two-stroke engines have no need for the valvetrain that we find in a four-stroke engine; the movement of the piston past carefully shaped ports controls the admission of charge to the cylinders. Fresh charge is drawn into the crankcases where it is compressed by the descending piston. The piston then uncovers ports, through which the charge is passed to the cylinder.

The principle requires a device to prevent the charge that's compressed in the crankcase from being expelled back to whence it came. This function is usually dealt with by a reed valve but there are other systems, the most common of which is the disc valve system. These systems - the reed or the disc - act as a one-way valve, opening when the crankcase pressure is below atmosphere and closing as it rises again.

The reed valve - sometimes called a leaf valve - is a profiled flat component that sits across an aperture in its closed position. These are commonly used in pairs on a V-shaped block, as shown in the photo here. The opening of the reed valve is restrained by a 'reed stop' that prevents the reed opening so much that it becomes permanently deformed.

The materials used to make reed valves vary widely, although steel was a popular choice for many years. Available in a wide range of thicknesses as shim stock, it is in many ways an ideal material.

In recent times though, steel has been replaced to a large extent by composite reeds. Glass-fibre reinforced plastics were very popular for this application, offering good fuel resistance and low mass. These have, however, as we might expect, been supplanted largely by carbon-fibre composites.

A further development has been the use of two-stage reeds. These comprise an inner reed which itself has an aperture, and a smaller outer reed that opens initially. When the air flow rate reaches a certain level, the pressure differential across the inner reed causes it to open, allowing the full potential airflow into the engine. This is said to improve part-throttle and transient behaviour.

In these two-stage reed valve arrangements there is scope to use different materials for inner and outer reeds, and to change the shape of the aperture in the inner reed to tailor the pressure differential at which each opens.

Fig. 1 - The operation of an induction reed valve as applied to a two-stroke engine

Written by Wayne Ward

Looking for a typical example of such an evolution, my interest was awakened in the BMW approach with its boxer motorcycle valvetrain systems, where clear progress can be seen from very basic valvetrain systems to the current finger-follower systems. This development connects motorcycle and car technology, originating from motorsports initiatives.

In the past, BMW boxer motorcycles have had their engines transverse to the driving direction, to improve air cooling of the cylinders. The consequence is the wider overall width of the engine, leading to lean angle constraints in motorcycle racing. To reduce width as much as possible, BMW has introduced two major updates of the valvetrain mechanism over the past 20 years or so - the R259 series in 1993, with the Hi Camshaft valvetrain mechanism, still using pushrods and, in 2008, the HP2 models, with overhead camshafts and finger followers.

Before 1993, BMW had developed its boxer models with a low, in-block camshaft centrally located below the crankshaft, flat cam followers and 'long' pushrods, actuating rockers and valves. Issues such as floating valves were common when increasing rpm for higher output for specific race applications. All kinds of solutions were explored to improve the system at the time, by optimising the mass of the system and using spring-backed cam followers.

From 1993 a totally new boxer engine was introduced, which was equipped with a so-called high camshaft position. In the past, a number of engines were equipped with high-mounted camshafts, most of them still located in the engine block, for ease of manufacturing. In this new design though, a separate bracket on the cylinder head incorporates the camshaft, a set of cam followers (bucket type) and the valve rockers, the latter two components connected through two short pushrods made from aluminium with hardened rod ends.

During the R259's development, BMW also undertook some studies in parallel, which included a study with desmodromic valve actuation on the so-called BMW R1, all to investigate ways to decrease the overall width of the engine.

After the introduction of this new boxer engine, the top end was gradually optimised for higher rpm durability, reducing valve-sided mass by decreasing valve stem and head, required for the higher revving R1100S engine. Mainly due to the fact that the BMW boxer engine would never be a real race engine, apart from appearances in the Dakar rally, the 24h of Le Mans and the Boxer Cup, no further specific race engineering on a professional level was done on the engine to improve performance significantly.

Since then, several engines have been introduced by BMW based on a new top end, this time with overhead camshafts and finger followers, and based on its Formula One experience. In 2004, the first motorcycle with this new top end was presented - the new K series. For the BMW boxer engine, the first public use of this finger-follower valvetrain was during the Le Mans Endurance race in 2007, the official debut of the so-called HP2 Sport engine.

This is the last engine that is 'updated' to this finger-follower valvetrain system, although perhaps not based on performance only but in any case on the modular approach in the BMW engines.

And where the boxer engine (or flat twin) seemed to be the ideal engine concept for a pushrod valve actuation system, the requirements of today more or less impose direct valve actuation, which means the high-performance BMW 1000RR.

Pushrod systems are prescribed for a number of race classes, which enforces pushrod system developments. But when this is not the case, different choices can be made, making pushrod systems the 'unusual' system - more or less what desmodromic systems have been called for years.

As I said at the beginning here, although this article is not about the pushrod as a component, it is about a structural approach and the evolution of the top end of an internal combustion engine in a broad range of applications, including race engineering. In that respect BMW has defined its own clear route on valvetrain systems, moving from pushrod-actuated systems to finger followers for its boxer engines - the same as in its other ranges of engines.

Fig.1 - BMW R1 prototype with desmodromic valve actuation (Courtesy of Hans-Jörg Milse)

Written by Dieter van der Put

"The revs on these engines has gone up, and trying to find the ultimate performance level of these things is becoming even more difficult," he says. "You have to realize that these Pro Stock engines, volumetric efficiency-wise, are actually above a Formula One car. We have extreme valve lift, and our valve sizes are bigger. These are not endurance motors - they run full throttle for seven seconds and that's about it."

Because of the short duration involved in a Pro Stock race, Johnson can make a lot of compromises. "These engines run way over an inch of valve lift at the valve, so it's net lift - not gross lift - and we're looking at engine speeds approaching 11,000 rpm. We need to factor those things in. We don't run a lightweight valvetrain," he says.

The life expectancy on Johnson's PSI valve springs is not very good, he tells me. "Valve springs are good for about six to eight runs on the intake side and about ten runs for the exhaust springs. We run triple springs on the intake and dual springs on the exhaust side without any flat-wire damper insert as used in some other racing series. The dual spring uses its self-dampening characteristics and friction dampening, and the triple spring on the intake side uses friction dampening on the spring.

"The weight of the springs contributes to valvetrain weight, and the duration is a lot longer, so the characteristic is valve velocity on the exhaust side," he says. "It is typically less than the intake so we can get by with a lesser spring, so to speak." Some of his compatriots are known for using triple springs on both intake and exhaust, where Johnson feels the double spring on the exhaust side is more than sufficient.

Incredibly, Johnson changes his spring specification on a race-by-race basis. "We change our specifications almost every week. We normally feel the valvetrain is a weak link, so we change the cams," he says.

"If the valve springs aren't breaking, then we're not aggressive enough with the camshaft. There are a lot of things that go hand in hand. It's a continuous process, always changing."

Under what he considers normal conditions, Johnson is going to keep the same valvetrain combination going for any given weekend. "If you change the cylinder head design, con rod or anything that has anything to do with air flow on these engines, you're going to be changing the camshafts, pushrods, roller tappets and so on," he says. Rocker arm ratios are always in change, and rocker materials are always being changed. Our valvetrain specs are in constant development."

Because NHRA stipulates steel for Pro Stock valve springs, it scotched Johnson's revision of using a pneumatic valvetrain in the late 1980s and early 1990s. "I did perfect a pneumatic valvetrain but it's not allowed right now, even though it would probably save each competitor about $50,000 a year," he says.

Johnson is always on the look-out for new valve spring iterations and works closely with his supplier to discover new ways to get the springs to match well with his ever-developing camshaft specification needs.

Fig. 1 - Intake (left) and exhaust valve springs used by Warren Johnson in NHRA Pro Stock competition. Johnson has one victory in the 2010 season to his credit (Photo: Anne Proffit)

Written by Anne Proffit

Titanium is much more favourable in terms of component mass, and is widely used in motor racing. The material comes with problems though; titanium requires surface treatments to prevent oxidation. Also, titanium has low stiffness, and its specific stiffness is equal to that of most other alloys, meaning that to arrive at the same stiffness in compression, we would need about the same mass of titanium as we would steel, for example.

Titanium aluminide, as we have seen, has better high-temperature capabilities than titanium and is both less dense than titanium and has a higher modulus. It is, however, very expensive and is banned in some racing series, notably Formula One.

If we were to ask a valvetrain designer for a list of properties he would find look for in an ideal valve material, he would certainly mention low density, high elastic modulus, good high-temperature capability and good wear behaviour. All these attributes are found in the aforementioned materials but, perhaps with the exception of titanium aluminide, they are not all to be found in one material.

There is a class of materials though that might offer the valvetrain engineer exactly what he wants - ceramics. These offer an excellent combination of properties.

A typical silicon nitride material used for valvetrain component evaluation has an elastic modulus of more than 300GPa (about 50% higher than steel) and a density which is 25% less than that of a typical titanium alloy - less even than titanium aluminide. Silicon nitride has been used successfully for poppet valves in heavy-duty diesel engines, and has passed qualification tests for military use, indicating the durability of the material.

A stated advantage of using ceramic valves with regard to the diesel engine tests is reduced heat rejection, owing to the material's low thermal conductivity. For a race vehicle, lower heat rejection means less vehicle drag and a smaller radiator area.

Silicon nitride has found use in engine valvetrains for components other than valves, in applications for rollers on pushrod valvetrain rockers; ceramic rollers on pushrod lifters would be another worthy application. Moreover, the car companies have taken a keen interest in these materials for valvetrains, with successful dyno and vehicle studies being undertaken more than a decade ago by European and Japanese manufacturers.

With high stiffness and low density, valve control is made much easier using a material such as silicon nitride. Race engine valvetrain designers would naturally put this to good use with more aggressive valve lift profiles, although an equally valid strategy would be to run with lower spring pressures and so reduce valvetrain friction. In the current era of increasing fuel costs and emissions legislation, any gains due to reduction of friction are not to be lightly dismissed.

For this very reason the Japanese have for a long time shown particular interest in the use of ceramics in valvetrains. The 1998 SAE paper, "Advantage of Lightweight Valve Train Component on Engines" (980573) highlights some of the benefits to be had here in terms of fuel economy.

Fig. 1 - Silicon nitride has been widely used in rolling-element bearings for many years. Could it be the ideal valve material?

Written by Wayne Ward

The specifics of the connection between the either hollow or solid centre part of the pushrod has been briefly touched on. In this article the different concepts of connecting the pushrod ends to the centre part are explored further.

In general, there are three concepts - welding, pressing-in or forming from one piece - that have appeared in more or less chronological order in the pushrod's development history. We will take a closer look at the main differences between them.

The welded solution is the most widespread concept, and has been around for many years, originating from the roadcar market of the early 1900s. Here, the ball or cup ends are welded to the rod, which is almost always a tube and either straight of tapered.

The weld is typically a friction weld, mainly because of the low-cost nature of the process and man independent process quality, making the process stable and reliable - at least for OEM purposes.

Unfortunately, the rod ends, which consist of hardened steel, are not ideal for welding. Over time, lots of solutions have been developed to improve the quality of the weld, and significant improvements have been made. In principle, however, the welding process remains less than ideal.

The heat generated by the welding process degrades the material properties of the rod ends and may lead to cracks near the welded zone, which can result in the ball or cup breaking off. This goes even more so for performance engines, running at higher engine speeds and peak firing pressures, increasing the load on the valvetrain system.

In order to minimise these risks, a thorough knowledge of the welding design and process, as well as stringent quality checks on the welded areas, are of major importance.

Pushrods that use 'pressed-in' rod ends are considered a feasible alternative for the welded concept, because there is no heat-influenced zone and the rod ends can be designed in different shapes and specs to accommodate different customer requirements. These can vary from the type of pushrod ends (ball or cup), and of course adjustable length ends can be included as well.

One disadvantage of pressed-in rod ends though is the stress region introduced by the press fit between rod end and tube. At the depth where the pressed-in rod ends stops, a stress concentration will occur, but by detailed engineering it should be able to reduce these stresses to below the critical level. The pressed-in design is a very cost-effective and flexible solution, which can provide a wide range of combinations of tube diameter, shape (straight or tapered), materials and combination of rod ends, without the process uncertainties of the welding process.

Both this and the welding concept have the advantage that drillings through the rod ends can provide quite a simple means to transport oil through the pushrod to the critical contact areas between ball and cup, which is a lot more difficult to achieve with the third concept, the one-piece formed pushrod.

These solid pushrods might become standard practice in extremely loaded drag race engines, but the added value for other race classes, specifically longer-duration race series, is not (yet) apparent. The advantage of these solid pushrods with machined ends is the absence of the external stresses introduced by pressing or welding, as explained above. This enables better predictability and understanding of the mechanical behaviour of the pushrod.

Apart from the increase in mass over tube-style pushrods, which, on the pushrod side of the valve train, is not that significant for the performance of the engine, the machining of the ends is somewhat more extensive in comparison to machining separate rod ends. The requirement to secure oil transportation to the critical contact areas in the valve train might be overcome by using a small diameter, long drilling through the length of the pushrod. But this drilling remains a critical process operation, leading to a cost disadvantage.

Summarising the different pushrod concepts, it will be a matter of customer requirements as to which concept is most suitable for the application. Currently, the pressed concept seems to have the better predictability over the welded design, where the solid pushrod design seems to be feasible only for extremely high loaded applications. And although the pushrod market will not significantly grow - in fact a slight decrease is more likely - race engineers will continue to balance their requirements against what can feasibly be produced.

Based on the multi-race requirement for the engines, one could expect that the reliability of the valvetrain system, together with more stringent cost directives, will play a more significant role in decision-making in the future. That might just lead to a shift in the choice of pushrod concepts.

Fig. 1 - Pressed-in hardened rod end in aluminium pushrod

Written by Dieter van der Put

Because its engines are not based on pushrod designs, Dan Esslinger, president, says there are not very many valve-spring issues. "It's just the design of the engine and the design of the camshaft," he says. "Even though we turn a lot of rpms (in the neighborhood of 10,000), we are not cycling the kind of weight that a pushrod engine would, so the valve springs haven't been nearly the issue they have been for some other builders."

Esslinger uses an off-the-shelf Manley Performance Products tool-steel dual valve spring without inserted damper. "It's their polished spring - it almost looks like it's chromed. This valve spring takes a lot of stress out of the material and we have had excellent luck with them. It's nice and light too, about a third of the weight of other springs we've used," he says.

The heat-treated spring is part of the NexTek grouping of oval track endurance valve springs with a maximum valve lift of .750 in. It has an outside diameter of 1.580 in and an inside diameter of 0.812 in. With chrome silicon of premium grade, this particular valve spring is kept free of impurities, according to Manley Performance.

Esslinger admits, "Depending on the customer and how hard they run, usually, if a guy goes 20 races, we'll change out the valve springs. It just doesn't feel like the spring owes us anything at that point.
"Every spring is the same that way. There isn't a lot to them - it all depends on the harmonics - so we try to keep them out of the 'unhappy spot' to make them last longer."

The company does not use a Spintron to test its products; it has its own methods to determine the 'unhappy spot' for a valve spring. "We kind of chase problems. You have a problem, you move something around and find what it likes," Esslinger says. "Sometimes you can tell by adjusting something or by results as to what's going on. We're more results-based than anything."

Esslinger chose to use Manley valve springs as his standard for midget engines about three years ago, tending to that direction because of the spring's lighter weight. "Weight is always an issue, but we still need the strength for our compression (just over 15:1) and high rpm level. You come to find out there are very few spring manufacturers out there that can do the job we need."

Esslinger Engineering has been able to stay fairly stable with its camshaft design and the balance of its valvetrain items - since it started the midget programme it has used four camshaft designs. "That is a pretty low number. We've tried to stay away from those kinds of changes if we can avoid them," Esslinger says.

"We have just come out with a new camshaft and we're pretty happy with that. We got it at the end of last season and it was more a design to try and soften the loading. Because of the design of our engine, we can't just use anybody's cam design. It costs a pretty penny just to try one - and we've had pretty reasonable success with Brian Crower and Web Cams."

The camshaft stability and the new, lighter valve spring have helped Esslinger Engineering to become the dominant engine builder in the USAC National Midget championship. "We're not having any wear issues, and with our strength in numbers - at O'Reilly Raceway Park the Night Before the 500, there were 40 cars and we had 24-25 of them - those adjustments have been good for the company's overall success.

"This is to our advantage, not having to make big sweeping changes to find that golden BB (the magic spot). Whatever it is that's going to get you where you want to be, we're able to make small changes and therefore smaller mistakes."

Producing about three engines a week, Esslinger relies on the five sets of valve springs he keeps on the shelf in constant rotation. "We can get them next day if needed," he said.

Fig. 1 - Manley valve springs work for Esslinger Engineering, as they are both light and strong (Photo: Anne Proffit)

Written by Anne Proffit

Valves made from these materials are expensive for various reasons, some to do with the price of the raw material, some due to the immaturity of the technology and some due to the extra processing time required to produce the valves. While it is always interesting to take note of the latest materials technology, there are many people in our sport who can afford neither the cost nor the risk of running these materials, and would derive little real benefit from doing so anyway.

For these people, their requirements are that the parts should be of consistent quality, that they are inexpensive (compared to the options above) and reliable. For many in this group (which extends into many forms of world championship racing) the material of choice is austenitic valve steel.

Austenitic valve steels are so named because the material structure remains austenitic even down to room temperature and below. For many steels, the austenitic phase is stable only at elevated temperatures, and on cooling transforms to another structure. Usually, we can quench from the austenitic phase to form martensite, which is very hard and can be tempered to give the required combination of strength and ductility.

Austenitic valve steels, which are commonly used for valves, are characterised by containing large percentages of chromium and nickel. The austenitic valve steels are neither very hard nor particularly strong at room temperature, but they retain a high proportion of their strength at the temperatures at which racing valves operate.

One difference between these steels and standard 'stainless' steels - which are also austenitic steels - is the amounts of other alloying elements, with the valve materials often having a higher percentage of carbon than a common stainless steel such as 316. Most stainless steels contain less than 0.1% of carbon, but austenitic valve steels commonly contain between 0.3% and 1% of carbon.

There are cheaper materials that contain less of the expensive alloying elements, although they are not as strong as austenitic steels at higher temperatures and are not as corrosion-resistant as austenitic steels. For racing purposes then, where temperatures are higher than in normal series-production applications, austenitic valves are common. Their fatigue strength and tribological behaviour can be enhanced by treatments such as nitrocarburising, and such treatments are particularly effective in materials containing high percentages of chromium.

The high percentage of chromium also makes the materials suited to nitriding processes, and plasma nitriding is used by some valve manufacturers to excellent effect here. The surface hardness of the nitrided layers can be in excess of 70 HRc on these materials.

For older engines that still run cast-iron (rather than bronze) valve guides, it may be necessary to chrome-plate the material to prevent wear.

Fig. 1 - These austenitic steel valves are for a road application, but the material remains popular for racing

Written by Wayne Ward

Fine adjustment will always be needed in the valvetrain to allow for manufacturing and assembly tolerances of such a long chain of interacting components. Typically for pushrods, much more adjustment than that is needed.

In all but the most expensive forms of motorsport, components such as cylinder heads and blocks have to be re-used many times, and it is typical to skim firefaces in order to maintain good cylinder seal.

This can be a big issue in NASCAR, for example, with heads and blocks that are used a number of times in their service life. Every time these surfaces are skimmed flat - even though be only a few thousandths of an inch may be removed - that changes the distance from the crank (located by the bottom of the crankcase) and the rocker pads at the top of the head.

Consequently, there has to be more adjustability in the valvetrain than just to allow for the machining and assembly tolerances of new parts. Plus, they also need to be light and very stiff, so it's a problem!

That means the cam-to-rocker distance will vary much more than typical manufacturing tolerances of a single chain of components and, in some cases, this adjustment is taken up with a screw adjustment in the rocker. Still, for higher performance parts, the extra weight and reduction in stiffness at the rocker for such an arrangement can't be tolerated.

The typical alternative is to produce a number of different length pushrods for the coarse adjustment, then to have the final fine adjustment in the tappet. This solution requires many multiples of sets of pushrods for an engine-build operation, produced in a range of sizes. There is an obvious logistics and cost trade-off inherent in the plan.

An engine builder working in motorsport can't be delayed because the correct length pushrod is not at hand, but doesn't wish to buy, say, 20 pushrods for every one the builder intends to use.

While larger operations can stand having plenty of pushrod stock at hand, a smaller engine build concern can't justify such money being spent. They may, therefore, take a hit on the optimum design (having varying height inserts perhaps) or they may take the hit on the engine build time and order pushrods to specific length for a specific build.

There are manufacturers that have developed their businesses along these routes and have very fast turnround times on specific-length pushrod orders; the downside of this type of manufacture is the cost implication.

Therefore, just considering the major design constraints of low mass, high stiffness, small diameter, precision, high-surface stress ends, together with high volumes and graduated but rapid manufacture, there has to be a huge amount of effort and resulting technology in the manufacture and fitment of the 'simple' pushrod.

Is the pushrod a 'blast from the past'? In some sectors of the racing world, perhaps, it is, but low-tech and crude? Those words simply don't apply.

Figs. - Manton Pushrods

Written by Anne Proffit

"With the economy and money as tight as it is, we're trying to save as much as we can. The cost difference of $1000 is a big deal. You can buy a lot of sets of steel valve springs for the difference in cost."

While some of his compatriots believe there is a penalty to be paid in weight, Stewart says he hasn't noticed any difference. "We changed the valves at the same time we changed the springs, going to a heavier valve. We went to the steel valve spring, and my car still revs up the same as it did prior to that," he says.

With the more expensive titanium springs, Stewart was getting about 30 runs down the 23, 1000-foot dragstrips the NHRA visits in 2010, With the steel valve springs, he says, "We probably make 25 runs at $500 per set maximum, so we're saving a tidy sum.

"We have the most wear - and the most trouble we have is when we get down to the finish line - it hits that rev limiter, takes the mag out and then it starts floating everything. Then the engine wants to kill the valve spring. To counterbalance our problems we just try not to smoke the tyres, which ultimately stops forward momentum," he says.

Stewart is using a triple spring from PAC. He says the most wear on his valve springs is on the outer spring, "because that is where it gets most of the heat.

"The double springs inside aren't that stout a spring, but the engine picks on the ends of the steel ones, too," he explains. "It'll break the ends off a lot. The valve spring still checks out all right (at the end of a run) but you have to get it out of there because on the next lap it's going to break and you're going to drop an air intake and ruin the whole engine. If you drop an intake, it could cost you $60 grand, and it could blow the blower out and knock the rods out."

Stewart doesn't relish those kinds of expenses for his team and makes sure his crew maintains and checks the valve springs as an essential part of every run. "Every single run, we take them apart and make sure there is no wear to the ends and check the seat pressure of about 425lb.

"We ran the same pressure with the titanium valve springs - there is no difference," he says. "The only thing that has changed is we are using the heavier steel spring that is maybe a little less durable than the titanium - by about five runs - but the spring costs substantially less.

If there is a physical difference, it's simply in colour, as there is more of a silver colour to the titanium and the steel is darker. The weight difference is minimal."

Stewart's team changed valve springs before the season started. "We went to the steel and sold off our titanium stock for way less than it was worth - but at least we got something for it, which was better than leaving them on the racetrack! And we switched out to a flat valve, rather than the dished valve we used to run," he says.

"In good air, we couldn't get enough gasket with as much cc's as I had," Stewart reveals. "So I took some cc's away by taking the dish out of the valve, and that made me run a thinner gasket. That way I could make enough power to run a flat valve with the combination I like," along with the steel valve spring.

The combination appears to be working well as the NHRA heads toward its Countdown to the Championship at Indianapolis in September. Langdon is eighth in points and poised to be a title contender.

Fig. 1 - PAC steel valve spring

Written by Anne Proffiit

If we are constrained to using the existing valve design, we may be able to cope with increased forces and stresses due to increased acceleration. But we will need to increase the valve spring load if we are to maintain the same factor of safety against loss of control at the cam nose. This may require a new spring design.

If we can lower the mass of the valve and its associated reciprocating components, we might be able to achieve our aims using the existing valve spring. There are many ways to do this, and one of the most common has been to have domed or dished valve heads.

This solution has been common in both racing and series production engines for some time, and there is little doubt that this measure is effective in reducing valve mass. There are a couple of very good reasons why we might want to reduce valve mass at the head.

The first is that it is the valve head - specifically the seat surface - that we are looking to control. When we design a new cam profile, what we are really looking to do is control the lift of the seat surface of the valve. Decreasing the mass of the head, of which the seat surface forms a part, is therefore an effective way of improving control.

The second reason is that the valve head is the largest part of the valve, and therefore has the greatest scope for mass reduction.

Many in racing, however, have now turned away from domed valves and are looking to other solutions for lighter valves. So what are their reasons for looking for other solutions?

In the case where compression ratio is hard to find, as in Formula One, where the bore-to-stroke ratio is extreme, losing volume from the combustion chamber due to a dished valve means there is often some consequential effect in having to reclaim this volume. Often - and this has also been the case in road applications - the solution is to design the piston to have a raised portion in the centre of the valve pocket.

This makes the piston slightly heavier and leads to higher stresses in the piston. There is also a consequent increase in the stresses in the con rod and crankshaft, albeit a very small one.

If we provide both a dish in the valve and a lump on the piston, we have maintained the combustion chamber volume but increased its surface area. However, this is known to be bad for engine performance, leading to lower efficiency due to higher heat transfer from the combustion chamber.

There is also some evidence from combustion simulation studies to suggest that the combustion in the dish of the valve head is not complete, and combustion efficiency can therefore be diminished by the use of a dished valve head.

Fig. 1 - A high-performance two-valve head, recently rebuilt with dished valves

Written by Wayne Ward

Fig. 1 - Pushrod display

Written by Anne Proffit

HPD has tried different designs to get the spring right, looking at the right length of the spring and getting its diameter right, so there has been a lot of development going into the choice of its geometry, shape and materials.

Incredibly, the engine builders have been using the same specification of spring since 2006, even as the Indy cars have gone from 3.5 litre engines to a 3-litre mill, and traded methanol fuel for ethanol. The cam profile hasn't changed since 2006 so lift has been consistent. As with all of its suppliers - who remain unnamed for confidentiality reasons - HPD relies on their expertise during design and ongoing development.

"It's a two-way process with our suppliers," says Roger Griffiths, manager of HPD's development division. "You have to respect the fact that your supplier makes so many of them (valve springs). And they'll quite often have an idea. So we have those kinds of conversations all the time concerning recommendations. Sometimes their solution might be on the expensive side, however, so we ask if they have something else."

When HPD does make changes, it normally tests in-house before applying to any race engine. "On valve springs, even though we have the same design and the same supplier, whenever we get a new batch we test some to make sure they are durable enough to go the distance," Griffiths says. "We never put a new component or batch in an engine without signing off on it, using our durability processes."

The steel valve spring used by HPD for the Honda Indy V8 has some coating on it, he says, "And we replace valve springs with each engine rebuild. It's a progressive curve with valve springs.

"Obviously, the revs haven't changed, so that is a plus. When we did begin the Honda Button Push to Pass that gives a couple hundred extra revs (last year), we were concerned that the springs would be okay at 10,500 rpm as opposed to 10,300 (maximum revs)."

In order to verify there would be no issues with the 200-rpm change, where drivers can have the extra power for as much as 20 seconds at a time, HPD took what Griffiths describes as "a whole bunch" of used springs and tested them at higher engine speeds. "We didn't have any problems; it wasn't an issue, but it was something we wanted to make sure, in introducing the Honda Button, that with the higher revs we weren't going to compromise the reliability of the valve spring - or the entire engine. So we did the testing on the valve springs and we were happy with the result," he says.

The value in having more than one supplier - which HPD has for its valve springs - is the ability to find the right vendor for the right application. Also, in some cases, HPD might be concerned by the vulnerability of a supplier, so it would not be sourcing springs from a single source.

"We have found with both our suppliers - we have used US-based and foreign vendors - that they can make the components we need and are happy to do so," Griffiths says. "We've had no issues with our suppliers, and this is what it comes back to - we pay a lot of attention to selection and how we develop components with those suppliers."

Fig. 1 - Close-up of generic double valve spring

Written by Anne Proffit

These are nickel-based alloys which, similar to Inconel materials, maintain a high proportion of their room-temperature strengths at elevated temperatures. As with Inconel alloys they are commonly used in turbocharged or supercharged applications, where the high operating

temperatures in these engines call for high-temperature materials for the exhaust valves.

The main constituents of typical Nimonic valve alloys are nickel and chromium. A typical exhaust valve material will contain 65-70% nickel with about 20% chromium, the other main metallic ingredients being aluminium, titanium and iron. The ultimate tensile strength and proof strength properties of a Nimonic valve alloy are almost undiminished at 600ºC compared with ambient temperature.

Turbocharged and supercharged engines have a higher power density than their naturally aspirated counterparts, with the same power being available from a lighter base engine. Supercharged engines are not common in many circuit racing series, owing to poor fuel economy, although their performance and response make them ideal for other forms of competition, especially drag racing.

Turbocharged engines are found in WRC rallying, where some very clever anti-lag strategies have been used to counter their perceived disadvantage of 'turbo-lag'. Equally clever are strategies and systems applied by some engine manufacturers to circuit racing applications. Turbocharged engines commonly race in endurance racing; the heavily funded factory Audi R8 cars with their turbocharged gasoline engines, for example, had great success at Le Mans and on the circuits of Europe and the US. Audi has since switched to diesel-powered cars and found continuing success.

One leading supplier of turbocharged racing engines to Le Mans and the US and European Le Mans Series continues to compete successfully using its turbocharged gasoline engines in the LMP1 and LMP2 categories. Advanced Engine Research (AER) of Basildon, UK, has supplied teams competing in these events with turbocharged engines for almost a decade, and it uses Nimonic exhaust valves in all its turbocharged engines, including the current works-backed Mazda LMP2 engines.

The photo here (figure 1) shows the company's turbocharged P32 V8 LMP1 engine, which is currently competing in the US Le Mans series.

This engine, and the latest P70 Mazda LMP2 engine, both use a similar design of Nimonic exhaust valve, as shown in figure 2. AER's Mark Ellis said of its valves, which are made to their own design but from an undisclosed supplier, "The Nimonic exhaust valves have proved reliable at Le Mans race distance and beyond". This was proven in 2009 when a car using its LMP2 engine claimed third place after a gruelling 24 hours at Le Mans.

As the company develops both turbocharged and normally aspirated engines we have no reason to doubt Ellis's word. Given that the 2 litre LMP2 engines all produce about the same power, and that the normally aspirated V8s enjoy a 70% larger capacity, we can expect significantly higher combustion temperatures and pressures.

Fig. 1 - AER's turbocharged V8 endurance engine uses Nimonic exhaust valves (Courtesy of Advanced Engine Research)

Fig. 2 - These Nimonic exhaust valves are typical of those used in turbocharged endurance engines (Courtesy of Advanced Engine Research)

Written by Wayne Ward

For a long time, Kenny Bernstein Racing used only hollow pushrods - and purchased those exclusively from Manton Pushrods in Lake Elsinore, California. While the Top Fuel team continues to work with Terry Manton for its pushrods, the specification has changed. "We've gone to stiffer pushrods using tool steel in 7/16-inch size, using intakes at 10/700 and exhausts at 11/600.

"We haven't always used solids but we've gone that route to make everything stiffer and stronger," said co-crew chief Rob Flynn, working with Mike Guger and Todd Smith on Brandon Bernstein's Copart-sponsored machine. "Everybody's choice of length is a little bit different," Smith chimed in. "It depends on the combination you're running."

While the solid, tool steel pushrod is a brand new combination for the Bernstein team in 2010, "I couldn't tell you that we have noticed much difference because we don't have enough runs on them yet," Flynn said before Gainesville, the third of 23 races. "We didn't have to change rocker arm configurations at all. It's the same length and everything; it's just a different pushrod, that's all."

The Bernstein team was not having any issues with its hollow pushrods; the reason they switched was for strength and reliability. "Even though we were not having reliability problems, we felt we might be able to make more power with this unit and, thus far, it has been fine," Flynn allowed.

Flynn, Guger and Smith make sure the crew checks their pushrods after every run. "If there is some issue with run-up, then we will change them, but really," Flynn noted, "it is a pretty reliable part. Easily, our pushrods will go through a race weekend. Last year, for example, we hardly bought any pushrods and we were running hollow rods."

The Bernstein team changed to the tool steel, solid pushrod for the 2010 season. "A lot of the people that were using tool steel last year would send them back to get new cups put on them, but that was about it. That pretty much speaks about how reliable they are." While some components are chosen for lighter weight, Flynn believes the weight of a pushrod is not a huge thing.

According to Manton, the change from chrome moly to tool steel "made a strength increase of 22 percent, which calculates into a stiffer pushrod tube that is the same outside diameter (as a hollow piece). The stiffer pushrod means greater lift at the valve and an improved valve action in relation to cam position."

With reference to this improvement in stiffness and strength, "At this point we don't have any issues with the pushrod," Smith admitted. "Unless we bend them - and that usually happens on the exhaust side because cylinder pressures are so high it doesn't want to open so you get deflection on the shaft," he pointed out.

No one in the Top Fuel ranks really knows their cylinder pressure. Flynn said that most teams have no way to measure it. "I'm sure there is someone you could bug and they'd give you some number, but all of us don't know. It's all a guesstimate."

Tuning for the new pushrod combination is not a black art. "It is all based on what you have done and what you know works. If you get something that works, then you learn how to get more out of your engine and find the weak link, build that part up to make it stronger," Smith said. "It is all a basic evolution."

Fig. 1 - Kenny Bernstein Racing went from hollow to tool steel solid Manton pushrods to gain strength and reliability

Words and photo by Anne Proffit

"I've been using PSI springs for over a year and I have to give those guys, particularly Larry and Steve, a lot of credit. They want to make the valve springs fit our engines and that is good; they don't just want to sell me valve springs. We are working with different cams and setting the pressure over the nose - that's what we pretty much work off."

He will admit to running 450 pounds of pressure on the seat and 1300 over the nose. "It's not higher than normal. We are working on different cam profiles and we are going to need a little more over the nose pressure. There is so much to be gained here in the valvetrain - and that is how we do it over here," he explained.

Morgan admitted to trying 7-8 springs since he started development work on the Ford programme last year. "It's taken a lot of trial-and-error on the dyno, and once we figure it out, we go on the track with it to see how it works there. To be honest, these things will lose 30 pounds and take a seat and they stay good.

The Max-Life PSI triple spring has flat heads. It's a light spring, but Morgan isn't exactly sure of the weight. "On the water side the weight is 0.239, I can tell you that. Another manufacturer's stuff is 0.252 so in comparison it is quite a bit different," he confirmed.

What makes Morgan happiest about the valve springs he is using on his new Ford engine is the service life and reliability. It is very, very reliable. I keep them in for six runs and then I take them off, whether they show wear or not. I don't want to take any chances, because if you break one, you destroy a lot of parts. I don't like taking chances so that is what we choose to do," he told me.

Three races into the season - and with one semi-final round in his pocket - Morgan has tested half a dozen iterations of his valve spring on the track, but he continues to "rely on what the dyno tells us. We have to do that. After we figure what we want on the dyno, we make sure it's good on the track."

In this development year, Morgan expects to change cam profiles fairly regularly. "As we change the cam profiles, we anticipate another valve spring change. I hate that part but that's what you have to do. If we can pick up two horsepower on the cam, you got to go for it. You can't turn down two horsepower," he laughed. He expected to have the next change in time for Charlotte, at the end of March.

Fig. 1 - Larry Morgan is on his eighth different valve spring as he develops the new Ford Pro Stock engine

Words and photo by Anne Proffit

One part of the discussion with Managing Director Willy Tagliavini concerned manufacture of Inconel valves, and how this differs from a more normal austenitic valve material. The cost of an Inconel valve is around 40 - 50% more than an austenitic valve of the same design, and I asked where this premium comes from. It was claimed that the material itself costs between three and four times as much as a typical stainless grade and that the machining is more complex, both during the turning and grinding stages, requiring special tooling and different techniques. "It is necessary to use special tool inserts for high temperature alloys with a special composition for higher durability and slower machining speeds" said Tagliavini. It was clear from this response that their typical Inconel valve, which is made of Inconel Alloy 751, takes longer to make than a stainless valve. Coupled with the premium that they pay for the material, we can see where the extra cost comes from.

The company uses a basic rule of thumb when advising it's customers on choosing a high temperature material over a stainless item. When the output reaches or exceeds 130hp per litre of displacement, then the recommendation is for the exhausts to be made of a high-temperature material. Although they make valves from both Inconel and Nimonic alloys, it is claimed that "due to commercial reasons and availability" the company have settled on Inconel 751 as their high temperature material of choice. The 130hp per litre rule is not by any means set in stone, as there are many factors which can influence the requirement for a high-temperature valve material. One important factor is the rate of heat transfer between the valve and the seat, where typically 75% of the heat from the valve is released through the contact with the seat, so the contact area and time of contact have a big influence.

Owing to the ability of Inconel materials to run at higher temperatures than a stainless alloy, some customers take advantage of this and use Inconel valves with a narrower seat contact. The narrower seat means that the valve will run at a higher temperature which the material has within its capability, but more importantly allows a better flowing exhaust valve which may be a very significant advantage in some applications.

The above company has had a number of customers who have had bad experiences with sodium-filled OEM hollow valves, and that some of their trade in Inconel valves is supplying replacements for the hollow parts with customers willing to accept the weight penalty for the prize of increased reliability.

Fig. 1 - Inconel valves for extreme applications (Courtesy Supertech Performance Inc)

Written by Wayne Ward

Although the service life of intake and exhaust pushrods is, conceivably, longer than the 700 miles that make up each Sprint Cup race weekend, TRD elects to replace them after each race. "Generally they are in a condition where we probably could reuse them, but we just don't," Currier told me. "We'll use them again for something else, like in development or performance engines or in dyno engines for testing other components, because they are in good shape."

The pushrods used by TRD are part of the TP (Trend Performance) line of double-tapered rods, using 7/16-inch diameter and a 165 wall. The manufacturer places a 100-thousandth wall down the centre, which is, according to Trend's John Williams, a standard hole. Cases are hardened to Rc 60 and there is a black oxide finish with laser-etched length. The pushrod, 7/16-inch at its widest point, tapers to 3/8-inch. Lengths on both intake and exhaust pushrods runs in the 7-8-inch range, Currier said.

The material for TRD's pushrods is 4130 chrome moly and the 5/16-inch ball end is CNC-lathe machined onto the rod during manufacturing; it is a single-piece item. "There used to be some tip wear issues, but we don't have any of those issues anymore," Currier said.

"Configurations are changed when we do something different to the valve train; then we change the length of the pushrod; sometimes we'll do a little other work with it, but it just fits into the process. We might change the diameter or something like that," Currier explained.

The challenge, he noted, "is to fit them into the head so it's a little bit tricky because we have to work to make it fit. We're trying to get the biggest pushrod we can fit into the head and to accommodate the bigger rod depends on what kind of machining we can do around the head for clearance where the pushrod goes. This depends on what we're doing with the rocker arm geometry. That can affect where the rod is placed, relative to the head."

Of course there are times when you can't compromise on placement, so Currier might need to go with a different size pushrod in order to get it to work. "That is where we have to compromise," he shrugged. "Sometimes we might have to go with a less hefty pushrod for fitment purposes, but we try to go with a stiffer, more responsive pushrod in all instances."

Fig. 1 - Toyota Racing Development uses double-tapered pushrods from Trend Performance (Photo courtesy of Toyota Racing Development)

Written by Anne Proffit

"Our biggest challenge with valve springs is making them survive with more lift and extra cam work, trying to make them survive. We are constantly working on different designs and lots of testing to improve the valve spring," Currier told me. "It is basically a straight spring with flat tops; any taper would be unnoticeable to the human eye."

Because NASCAR rules determine valve spring material of simple steel spring wire, Currier can't work with exotic materials for his engine's needs. "We do a lot of FEA analysis and we drive development on our valve springs because we end up doing most of the testing and most of the analysis."

TRD drives development of its valve springs, but their suppliers have a lot of input into it. "For example a stiffer spring design might be more driven by them than by us: they'll say a reason for doing it and we'll say okay, we'll try it. Or it will be integrated into something we are doing, so it's kind of a mutual thing most of the time."

Although it's not always necessary, TRD does exchange its valve springs after every Cup race, which is a life cycle of about 700 miles on a standard race weekend. "We don't necessarily trickle those down even though they could last a bit longer," Currier said.

The configuration for an intake valve spring could be different from that of an exhaust and TRD makes changes "because the dynamics of the intakes versus the exhausts are a little bit different. The demands on each side are a little bit different; because of the interaction between them sometimes it's helpful to have them be a little bit different," he noted.

For TRD, specification updates on Cup valve springs is part of a whole package of things. "We run the changes about the same period for all of our internal parts and we might have three different valve spring options per year. We might use up some of our specs. For example, at some of the short tracks where top end power is not so critical, when we go to a new valve spring spec, we may use up some of the engines of the older spec at those particular races where it's not so critical or doesn't matter as much.

"Even though we're on and off the pedal a lot at that type of short track, it affects the wear of the spring somewhat, but when we qualify an open spec package, it's generally either for a particular track like a high-speed track (such as Atlanta or Fontana) and then it's only going to be for that. Or it's qualified for both the short tracks and fast tracks, which makes it less of a problem to switch over between one or the other."

The difference between Plate and Open springs is due to the duty cycle. At a Plate race, "It's such a constant speed; the RPM is lower and the throttle is always open, so the stresses are reduced and we can take advantage of that and do more with the spring," Currier said. "We can try and reduce friction or make it stressed more - put in more lift or a more aggressive cam - because it's not using that amount of stress capability, so we can lean on it a little harder. But then, when you lean on it harder, it usually fatigues and breaks and we end up with a milkshake," he laughed.

Fig. 1 - TRD prefers a double valve spring without insert (Photo courtesy of Toyota Racing Development)

Written by Anne Proffit

In this article we shall begin to look at other valve materials, beginning with those materials used for supercharged or turbocharged applications. Highly charged engines extract impressive bmep and high 'power density' by forcing much greater amounts of air into the combustion chamber than is possible with even the most highly optimised naturally aspirated engines. To this greater quantity of air is added fuel in the correct ratio and much greater energy is released in a combustion chamber which is not significantly larger than found in a naturally aspirated engine. Consequently combustion pressures and temperatures are very high and the exhaust valve in particular is required to run at extremely high temperatures, having high-speed burned gases flowing over it at high velocity, with the consequent high levels of heat transfer to the valve head and stem.

Owing to the temperatures involved in such applications, those materials normally used for the manufacture of valves, such as steels or titanium are often not sufficiently strong, or may offer insufficient durability. For the majority of those racing every weekend, the option of rebuilding engines and replacing valves after each race is not an appealing one. Some very highly optimised turbocharged racing engines have successfully been able to exploit titanium valves, but these items have a short service life and are extremely carefully designed and managed.

The materials that valve manufacturers and engine designers turn to for these applications are those developed specifically for high temperature use and which find wide application in the gas turbine engine sector. They have been developed for their ability to retain high levels of strength at high service temperatures and also to resist creep. Creep is the phenomenon whereby materials will continue to stretch under constant load over time, and this is particularly apparent with increasing temperature. Interested readers will find that the particularly dry book Stress-Rupture Parameters by Conway, contains much data on the subject.

The materials which maintain high strength at high service temperatures and which are creep resistant are often expensive, as they contain large amounts of expensive elements such as nickel and cobalt. They are often described as 'superalloys'.

One of the most popular classes of materials for poppet valves on turbocharged or supercharged engines are the 'Nimonic' alloys, which are based on Nickel, but which also contain significant percentages of other expensive alloying elements. A typical material often used for valves is described as having a service temperature of around 800 degrees C, or 1500 degrees F, and which maintains around 90% of its room temperature strength at 600 degrees C (1100 degrees F). The valves shown in the accompanying picture are Nimonic parts for a motorcycle application.

Another suitable class of alloy are known as Inconel alloys. These alloys are again based on Nickel, but are not as widely used as the Nimonic alloys.

Fig. 1 - Nimonic exhaust valves.

Written by Wayne Ward

The coatings that these companies offer are generally quite thin ceramic coatings which offer very low thermal conductivity compared to the valve material. We should at this point note that a titanium valve has very much lower thermal conductivity than a steel valve.

So, what are the claimed advantages for these coatings? The most commonly quoted is that there will be less heat lost from the combustion chamber as a result of combustion through the face of the valve. The vast majority of any heat flow through the valve goes into the cylinder head via the valve’s contact with the valve seat, with some also being conducted into the cylinder head via the valve guide. Naturally the valve also becomes heated to a certain extent, and the claim is that this loss of heat from the combustion chamber is reduced, thus increasing the amount of work which can be extracted from the fuel. We are often told, not only by the authors of text-books but by politicians of all persuasions, how much energy is wasted from an engine in the form of heat, and the heat lost to coolant (be it air, oil, or water) is an extremely significant part of this lost energy. Lower heat rejection from the combustion chamber should mitigate this loss of energy, providing that we make some use of this hotter gas and don’t simply let it out of the exhaust port. In terms of a racing car, an engine with lower heat rejection to the water circuit in theory means the ability to run less cooling air flow and therefore smaller radiators.

The thermal barrier coating itself becomes very hot as it cannot conduct heat away to the seat very efficiently, but owing to its low mass, it doesn’t impart much heat to the incoming charge, which quickly cools the coating. Another positive side-effect of the lower level of conduction through the valve seat is that the valve and the valve seat should be cooler. Heat transfer to the incoming charge from hot valves and valve seats has been shown to be significant and to have a measurable and detrimental effect on engine performance, and so any detrimental heating of the incoming charge will be reduced. This benefit is seldom stated by those advertising such coatings. If more heat is transferred to the incoming charge, the pressure of the charge in the cylinder is raised, thus causing the pressure differential between the inlet and cylinder sides of the chamber entry to be less favourable to further air flow into the cylinder. We can see, therefore, that lower heat transfer to the incoming charge should result in greater volumetric efficiency.

I don’t know of any ‘serious’ application of this coating in bespoke racing engines, and enquired further with someone who has wide-ranging experience. ‘Valve Expert’ has excellent knowledge of valve applications across the whole range of racing formulae, and knows of nobody using these types of coatings on bespoke racing valves. His opinion was that the lack of ductility would affect the reliability of the coatings, and that reduced fatigue strength was also a concern.

Written by Wayne Ward.

Building for the American Power Boat Association's (APBA) Super Cat two-time World champion and 11-time Western Division titleholder 36-foot The Renegade, driven by Garden Grove's Craig Ferguson, Stewart Van Dyne II and Stewart Van Dyne III (Tres) specify Tool Room valve spring material by Isky Racing Cams from nearby Gardena, CA.

Because they have to use a spec head and manifolds on their near-510 cubic inch Chevy that makes around 830 horsepower, the engines are limited to 7600 rpm to keep speeds under 140 mph, the Van Dyne's fit a double spring with flat insert damper.

"These springs hold their own over the rebuild schedule of four races, running without any problems. The springs usually go down about 10 pounds from where they started and we normally replace them when we rebuild the engine," father and son agreed.

"The only time we ever broke a valve spring is when we had a driveline failure, but that doesn't occur very often," they said. "Actually, that happened before we had a rev limit. If you have a failure like that, it causes the engine to hit really high rpms" in excess of the decreed redline.

In this series, it's not possible to port and polish and the carburetor size and compression ratio (12:1) are fixed. "There was a time when they didn't have rev limiters and they ran to 8500 rpm the whole race without missing a beat," they reminisced. "Everything was perfect; ran like a champ. We would check the valve lash after the races and a couple of valves would change about a thousandth or so. Everybody was happy!"

The Isky Tool Room valve springs feature full-spectrum harmonic vibration control to suppress surge and avoid harmonic convergence. These springs are the "RAD" style, which is a special process that makes them nearly bullet-proof. They have endured Gold Stripe test certification through a 1000-mile certification test for each batch.

The springs have an outside diameter of 1.625-inches and inside diameter of 0.770-inch on the inner spring. The seat pressure is 250 pounds at 2.020 inches installed height and 675 pounds at 1.270 inches when new. "We actually shorten the installed height a little and bump the pressure up a bit." These valve springs are used on hardcore roller tappet engines. With these spring pressures, the builders recommend using a roller bearing camshaft.

According to Richard Iskenderian at Isky Racing Cams, this heat-treated super high-grade spring uses a super-clean high tensile chrome silicon material. The manufacturer uses a special process in constructing this spring that helps it hold pressure better, last longer and manage breakage better. It is, he told me, best suited for higher rpm applications.

The Renegade races about four times per year in contests of one hour duration. There are customarily three 20-minute practice sessions the day before each race, giving the valve springs a life span of about 10 hours on the water. They are replaced at each rebuild but could be returned to service a second time if needed.

Fig. 1 - The Isky Racing Cams Tool Room valve spring is strong enough to last on the water.

Words and photo by Anne Proffit

Pushrod sizing is important for the 830-horsepower, 510-cubic-inch big block Chevrolet (with spec heads and manifold) that the Van Dyne father-and-son combo prepare for the American Power Boat Association (APBA) Super Cat used by two-time World champion Craig Ferguson's 36-foot Skater Catamaran "The Renegade" of nearby Garden Grove.

"The length and diameter of the pushrods we use depends on the cylinder head, valve length, rocker arms, tappet length and base circle of the cam lobe," they told me. "The diameter is also controlled by the position of the pushrod cup in the tappets and the pushrod cup in the rocker arm." To meet the challenge of keeping the intake ports out of the way, "When possible, we run offset rockers and offset tappets to help the problem.
We try to get the biggest pushrod that will fit the application," the Van Dyne's agree.

The pushrods chosen for this application are of one-piece 4130 chrome moly tubing with 0.120 wall and swedged ends made of 4130. These pushrods are heat-treated and black oxide coated. The swedged formed ends have exact radii for low friction and long life in the cups on the tappet and rocker arm, the Van Dyne's said. The pushrods normally last a full season of running on the water and are not affected by any difference in water condition.

With a 7600 rpm redline dictated by the American Power Boat Association (boats had been running in excess of 140 mph with the previous 8600 rpm redline), "We wanted to keep the RPMs up to come off the corners better. But we had to go to more aggressive cam profiles to pick up the torque at a little lower rpm, to jump off the corner harder and still run down the straights hard," they said.

When the Van Dyne Engineering brain trust went to the 0.120 wall pushrod, it added some weight to the mix. The 3/8-inch diameter, 8.2-inch long 0.120 wall Manley pushrod weighs 96.5 grams, compared to the same length, 3/8-inch diameter 0.080-wall unit, that weighs 73.5 grams. Another example is an 8.2-inch long, 5/16-inch diameter 0.080 wall pushrod Van Dyne currently uses that weighs 61.5 grams.

"The difference in weight on these pushrods really doesn't change what we do. It has been found that the weight on the pushrod side of the rocker arm isn't quite as critical as on the valve side. It doesn't matter what it weighs; if it's a short fuse that won't let you run as hard as you need to," they told me.

In addition to its work building championship APBA engines, Van Dyne Engineering manufactures the Offenhauser "Offy" engines, prepares stock blocks for racing purpose and has its own line of racing water pumps.

Fig. 1 - The larger, 3/8-inch diameter 0.120 wall 8.2-inch long Manley pushrod works best in the big block Chevrolet used in American Power BoatAssociation Super Cat racing.

Words and photo by Anne Proffit

This particular Manton intake pushrod was installed for eliminations at the penultimate NHRA round (of 24) on The Strip at Las Vegas Motor Speedway. It was being used in the Don Schumacher Racing Matco Tools Top Fuel rail of Antron Brown, who finished third in the season-long standings after leading the regular season point chase. Brown’s engine had a problem during eliminations that exposed the pushrod to a head gasket failure – or maybe a tune-up failure, according to co-tuner Rob Wendland.

The engine “dropped a hole [cylinder] and normally, in qualifying you would have gotten out of it,” said Wendland, “but in eliminations you can’t do that. It picked the hole back up and basically lifted the cylinder head. It was really ugly,” he told me.

Each intake and exhaust pushrod is thoroughly checked after every single pass down the 1000-foot dragstrip. “We look at the ball end; we look at the cup end and make sure there are no cracks to the cup or the ball,” Wendland said. “We have a fixture that we put the cup and ball ends into, then tighten it up a little to hold it perfectly centred. By twisting the pressure on the pushrod, if the indicator moves at all, like 8-9 thousandths out of round, we take it out of service – but that’s not usually the case.”

The Matco Tools team has moved its usage to a completely solid, billet pushrod that can last as many as 200 runs, Manton confirmed. The pushrod shown here is a solid tool steel intake pushrod and it is a bit thicker than 475 thousandths in diameter.

“We used to worry about a couple of things. Weight was an issue and moving up and down at 8500 rpm, a part that is heavier, it’s hard to move that pushrod up and down. You have masses going this way and masses going the opposite way. We learned that it’s not weight – the biggest issue is the correct opening and the correct opening speed that keeps the longevity of the part,” he said.

The cup configuration has stayed the same but the material for these pushrods has changed a little bit over the years. “The thinner part is billet tool steel and the other pieces are actually put onto this particular pushrod,” according to Wendland.

The pushrod tips are press fit into the centre section with a specific, very accurate pressure fitting. This manufacturing process allows a pushrod producer to use different materials for a more properly manufactured pushrod.

Manton is able to push both ends to this billet piece and make it stable. Or relatively stable within the realm of its punishing environment.

Words and photo by Anne Proffit.

That change will likely result in changes to parts distribution, as DSR campaigns two other T/F dragsters for [now] six-time consecutive champion Tony Schumacher and 2009 fourth placed contestant Cory McClenathan – Brown finished third in the title chase.

“There is a bunch of testing going on with a smaller diameter spring – like a small block Chevy – and they are having excellent luck with them,” noted Rob Wendland, co-tuner for Brown’s rail. “Our understanding is that DSR are doing a lot of valve spring testing in December so we might see changes.”

The Matco Tools team currently uses a titanium double spring with Teflon insert from Michigan-based PSI. “The Teflon insert keeps the inner spring a little straighter without having it touch the outer spring. Our springs go through a crazy heat cycle as it is, and that Teflon piece is what keeps the two apart,” he told me.

“These springs go through a frequency when they’re running and this frequency can be destructive – there are destructive and good frequency ranges, so you want to be in the ‘good’ frequency range, depending on the spring and how its ground, how it’s heat treated and the materials used in manufacturing that determine the frequency,” Wendland said.

Valve control is what determines the choice of spring. “We want our valve control to take place between 6500-8500 rpm, so we use our Spintron to help us determine that frequency. If you find a product that works well, you darn sure want to know what the frequency is for that spring so that you can build maybe a better spring at that same frequency. These harmonics do change, depending on the rpm changes.”

Wendland explained that some tuners have tried foam filling the driveshaft to have an effect on valvetrain harmonics, or putting weight in varying places on the driveshaft. “Even tuning the exhaust a little bit differently can change harmonics or frequencies,” he said. “It’s all about finding the motor’s happy place and that’s where you want the spring to activate. That is how we get longevity and better control on the valve. When the frequency is off, then it doesn’t have valve control.

Valve springs, like every component in a Top Fuel engine, are checked and replaced as necessary. On the Matco Tools dragster, valve springs can last as few as 12 runs down the 1000-foot dragstrip or as many as 30 passes.

Breakage occurs through heat cycles and poor harmonics. “This spring had a failure in the inner spring. One of the good things about a titanium spring is that, when it loses an inner, it still has a lot of force on the outer spring, which has a bigger, heavier wire diameter.”

This failure occurred when the team was simply turning the motor over – not during a qualifying or elimination pass – it “just went ‘bing’! It’s nice to detect this before it happens on the track. That is why we’re doing so much valve spring development here at Don Schumacher Racing.

Words and photo by Anne Proffit.

Even in the case of those engines where rockers are used, the sliding velocity of the adjacent parts, relative to the valve tip, is low. The consequence of this is that there is little or no chance that a significant and stable oil film can be formed. The oil films associated with hydrodynamic lubrication rely on significant and continued relative movement (or more precisely they depend on there generally being a non-zero entrainment velocity for those of your familiar with the technical terms).

Therefore, what scant quantity of lubricant finds its way into the valve/lash cap contact probably relies to a great degree on squeeze-film lubrication for any useful effect. Where rockers act directly on the valve, there may be some operation within the hydrodynamic lubrication regime, but this will be intermittent and transient, with a fully developed oil film being present only momentarily.

There can be a tendency for the lash cap to shuffle about on the top of the valve and this may cause some wear problems. The poor lubrication situation of engines using rockers is probably of greater concern, as there is a reasonable amount of relative movement with little protection offered by a fully developed oil film.

Whilst the frictional losses here are negligible, the problem of wear can be significant, leading not only to damaged components, but also to increased valve clearances if wear is significant. Where extra clearance is present, the result may be loss of performance, but there is a genuine risk of premature and catastrophic component failure owing to high valve seating velocities.

It is for these reasons that people have over the years tended to specify various processes to prevent damage to the tip of the valve from contact loads. Hardened tip materials such as ‘stellite’ (a cobalt based hard alloy) have been specified for many years, and the process of brazing a tip of a different material from that used for the rest of the part to the tip of the valve has been used with success. For instance, austenitic steel valves have a low level of general hardness and this cannot really be improved greatly by heat treatment. However, it is possible to add a tip of a hardenable steel material to give increased resistance to surface damage. There are many suitable candidates for the tip material. A steel such as 4340/EN24 is a popular choice and can be hardened to a level above 50HRc. Some people prefer even to go a step beyond this level of hardness and specify a high-carbon steel which can be hardened beyond 60HRc.

Written by Wayne Ward.

Honsowetz said the 166-cubic-inch Toyota four-cylinder engine they build for use in USAC’s National midget competition requires a stable valve train – something that’s inherently tough to achieve. For that reason, he has progressed from the 5/16 (with 0.080 wall) to using either of two specifications: 7/16-inch double taper pushrod with 0.165 wall thickness (at a cost of about $16 each) or 9/16 full taper with 0.188 wall thickness, costing him roughly $33 each.

“The ultimate thing we’re looking for is thickness of the pushrod – but we can’t have it too thick that it touches anything.” For that reason, they use the double taper pushrod in most of its engines, a unit that tapers from centre to either end. This specification keeps the material away from other vital parts of the engine.

“The smaller, 7/16-inch pushrod works for the application in many instances and it is the pushrod we have been using since the start of the Toyota programme in 2006. The National midget engine has more stress to it (than, say, the Silver Crown eight-cylinder) and, bottom line, we use both the 7/16 and 9/16 pushrods; it all depends on the track and the stress on the engine.

“For instance,” Honsowetz continued, “With a heavy duty-cycle, we might go for a stronger pushrod,” even though such usage requires cylinder head adaptation. “When we really have to lean on the engine at Phoenix International Raceway, Iowa Speedway and Belleville – those would be the three hardest tracks on the Toyota midget engine – we’d likely adapt the cylinder head to use the stronger, thicker pushrod,” that is, one that has a full taper.

“If we are leaning on the engine and running higher revs (upwards of 9200 rpm), we want to go for the thicker, stronger pushrod. We’ll then make adjustments to the cylinder head wall in order to accommodate it.” This engine builder has been using the 9/16 full taper pushrod for about 2-3 years: “We use a mix of pushrod sizes, but it’s certainly not a 50-50 mix,” Honsowetz told me.

Weight isn’t much of an issue with the pushrods. The older 5/16 weighs all of 55 grams, while the 7/16 goes off at 112 grams and the larger, 9/16 diameter pushrod weighs 160 grams. “We don’t worry about the weight of the pushrod,” Honsowetz allowed. “We are mostly concerned with stiffness.”

The pushrods used in the Toyota engine “get cycled in and out, but they all go through a magnaflux check,” he said. “Each is numbered and checked and we actually use them for quite a long time. They do get changed during rebuilds because there is machining of components associated with the pushrod.” Honsowetz said that, with his stiffer valve train he has no oiling problems. “It’s not something we worry about.”

Written by Anne Proffit.

“We use PSI springs in everything we do, not just in the Toyota midget engine,” he told me. “We have such a long relationship with them and we are able to work together to advance development of our springs.

“I think Steve looks forward to Ed’s input and talking with Ed about different solutions. We’re very cooperative with them because of our long, exclusive relationship. They see all of our parts and see what the springs are doing so they do have a lot of interest in development because they get exposure to our information,” Honsowetz said. The engine builder is on either their sixth or seventh spring iteration since beginning the Toyota project back in 2006.

What they look for in the valve spring specifications for the Toyota USAC National Midget engine is stability and reduction of inherent vibration. “You want everything to be stable. When you take the spring out, you can measure the tension and you want it to look like it’s had a pretty easy life. That is hard to do, very hard to do,” he admitted. “At 9200 rpm, which is where we run these engines, the valve spring opens and closes 76 times a second. That is a very hard life!”

The valve spring in question use a radius rounded at the tips, where the tail wire is squared off a little bit. The spring is a custom unit made specifically for the Toyota EPRE engine; it lasts about 150 miles – or about six shows – and is always replaced on rebuilds, Honsowetz said. “The detail on the ends of the wire is a very important feature of this spring.”

Working with the Toyota 166-cubic-inch USAC mill, “That midget engine needs every drop of power it can get so they do run right on the edge. When you figure in practice, qualifying, heats and feature races, each race is about 25 miles. Each spring gets a general inspection and it might go out for another six,” albeit in another engine.

The double wire spring used in the Toyota USAC National midget engine weighs about 142 grams; a standard 1573D triple spring with a flat damper – one of the more popular off-the-shelf units from this spring supplier – weighs 165.5 grams. In this instance, weight is a factor, one reason why Honsowetz persists with the custom spring from their regular supplier, even though they have tried the 1573D on occasion. The stability and vibration reduction they get with the custom product is superior; the USAC Toyota engine warrants the use of this product.

Written by Anne Proffit.

In terms of surface treatments, we can split these into three basic groups:

• Coatings, which are applied on top of the bulk material
• Surface hardening treatments, which alter the mechanical properties or composition of the valve material itself
• Mechanical treatments, which work the surface of the valve.

If we take these in order, the first and probably most important group is that of coatings. Many of you are probably familiar with the main hard coatings that have become so popular in recent times. Certainly there are many valve manufacturers who offer coated valves with this type of coating. Titanium valves, owing to the properties of the material are ripe candidates for these coatings. As we have discussed in other articles, the surface behaviour of titanium when sliding, even under modest pressure is not good, especially if lubrication is scarce. Chromium Nitride, Titanium Nitride and diamond-like carbon (DLC) coatings are all popular. Prior to their prominence, it was quite common for the valves to have their stems coated with another, more wear resistant metal. Molybdenum, which has other uses in racing engines for sliding applications, is probably the main metal sprayed coating used for this purpose. The valve is prepared by machining the stem slightly in the area required for the metal spraying, and afterwards the stem is re-ground to the correct size along the entire length. Hard surface coatings also have great value on the seat area of the valve, preventing damaging material transfer between valve and seat. This is especially pertinent in the case of titanium, which can suffer from serious problems in this area.

The surface hardening treatments are generally applied to steel valves, which are generally made from austenitic steels. Both nitriding and low temperature nitrocarburising treatments are popular for this application, and both are done at around 500 – 550 degrees C. both of these treatments diffuse nitrogen into the surface of the metal and owing to the high percentage of chromium in the material, both treatments work very well. The chief advantages are an increase in hardness of the surface and greater wear resistance, coupled with an increase in fatigue strength on account of the residual compressive stress associated with both of these processes.

The final category is mechanical treatments, and whilst these can be said to work the surface of the material, it is likely that any significant residual compressive stress is relieved by temperature effects, especially in the exhaust valve. However, the main benefit to both shot-peening and the helical ‘brushed’ finish sometimes used on the back of the valve head and the lower stem is probably in disrupting any remaining machining marks, which would have a benefit in terms of fatigue.

Written by Wayne Ward.

The clearance difficulties inherent with a nitro-burning Funny Car engine define the specification of pushrod. “We have clearance problems in the cylinder head and the block, so making the pushrod strong enough to work under the conditions these cars work under,” is the biggest challenge for Wilkerson.

The intake and exhaust pushrods he employs are of standard chrome moly material with a proprietary heat treat. Wilkerson use a straight pushrod; “they’re not tapered like a lot of them. We were bending them up when we had a 0.130 wall but now that Terry’s gone to a 0.160 wall, it’s helped a lot.”

Wilkerson checks his pushrods after every run to make sure they’re not bent. “On average we probably get about 25 runs out of the pushrods; that’s a lot of runs.” He cites the move to the thicker wall for helping lengthen the life cycle. “We were bending them up when we had a 0.130 wall. We usually only have a problem when the thing smokes the tires and they get bent about 2-3 times when the valvetrain is trying to catch up.”

The result is a blacked pushrod end, “Which means the heat treat wasn’t good on it or something like that. All the ends are heat treated on our pushrods. It helps make them harder. We use Terry’s adjusters, too, because the radiuses are all the same on the adjusters and the cups. That way they mate better.” Wilkerson’s crew are quick to notify him if they find a dark spot on the inside of a cup, “So we throw them away after that.”

As a result of improvements to the wall thickness and heat treatment, “We have no problems with the pushrods at all, once we got the little quirks figured out. I’d say for five years we haven’t had any oiling problems. Right now, knock on wood, that’s the part of the motor we don’t worry about very much.”

The exhaust pushrods on Wilkerson’s engines are longer than his intake. “We have three different lengths for the exhaust; we try to keep the oil in the right range because it’s tight. One about 100 thousandths or 125 thousandths. That’s all the range is, so if you get the adjuster too far up into the rocker, it won’t oil very well. The length will depend on the thickness of head gasket we’re running.”

Using an altimeter to determine his choice of head gasket, “You don’t really have too much of a problem with that. The altimeter doesn’t really change more than 100 feet over a race weekend, so we make a choice and stick with that.”

Written by Anne Proffit.

The solution Tartaglia has found in his valve spring composition is going from off-the-shelf componentry to custom-built springs for his high revving engine. “To get our valve springs to live, we’ve had to buy good springs made of good material. That’s basically it. It’s all about going with the best materials and the best processes.”

Tartaglia, at the suggestion of McLaren Performance Technologies, took his business to Associated Springs – with two Michigan locations – and “told them what we were looking for. They made us a valve spring that seems to be working pretty well. I just gave them some space-related dimensions and told them how long the spring could be.”

The biggest challenge is finding room between the roof of the intake port and the bottom of the camshaft: “Everything’s got to fit in there, the bucket, the shim, the valve spring and the base. I told them what kind of pressures I was looking for, the rpm we turn, the valve lift we have and they took it from there.”

The result is an uncoated double spring made of Kobe steel. Tartaglia’s intake valve springs last about 20-25 runs; the exhaust springs will last as many as 50-60 passes down the quarter-mile dragstrip.

That is an increased life, on the intake springs, from the six-eight passes he used to get with commercial product. There is a trade-off: the current spring is a bit heavier than what he used before, weighing in at 60 grams.

“The valve on the intake side is a little bit heavier and the camshaft is more aggressive,” Tartaglia explained. “On the exhaust side the camshaft profile isn’t as aggressive. It doesn’t run as close to coil bind.”

The Suzuki engine turns about 13,500 rpm at full song. “I think some others are a bit higher, maybe closer to 14,000 but we discovered that’s where the engine is happiest, between 13,500-13,600 rpm.”

After running the same specification for three to four years, Tartaglia decided to make the change to custom during the 2008/2009 off-season. “We’d had some updates last year and before that, we ran the same specification all that time. Instead of buying an off-the-shelf spring that doesn’t last, we just decided to have the custom spring made.”

Now that he’s gone custom, Tartaglia is getting offers from other valve spring manufacturers who want his business. “We’ll talk to some of them over the winter.”

Written by Anne Proffit.

There are, of course, lower density materials than titanium, although these generally have not been employed for the manufacture of racing valves. One exception to this is titanium aluminide, once a favourite of the Formula One community but, yet again, banished due to the current FIA regulations. Titanium aluminide is a material under the watchful eye of the motor manufacturers, as well as other industries. Where we might remember, not all that long ago, that titanium was very expensive, there comes a tipping point where production quantities of the material, or one particular alloy means that the price quite quickly becomes realistic. To my pleasant surprise around 10 years ago, I found that it was less expensive to source good quality titanium bolts rather than flanged stainless items. So it is with titanium aluminide – there will inevitably come a point where production quantities become sufficient that it will become economically viable for someone to put valves into a production car or motorcycle, just as happened with titanium for valves and other parts. Thankfully titanium aluminide can be used by other race series for valves, so there is some hope that racing can indeed improve the breed as far as series production engines are concerned.

Titanium aluminide was mentioned as a racing valve material in the mid to late 1990s (and possibly earlier), and was widely reported to have exceptional specific strength and specific elastic modulus combined with good high-temperature mechanical properties (it promises higher tensile strength at 760C or 1400F than it does at room temperature). Apart from commanding a high price, it seemed to offer an ideal combination of properties for a valve material. So, why was it that it took so long to become widely adopted, and why was it that it almost disappeared in short order whilst it was still legal?

The first question is certainly one of money and immaturity of technology. The material was very expensive, and titanium materials were certainly man-enough for the task. However, the properties were such that people did start to experiment with ways to make a valve which didn’t cost a ridiculous amount of money. Some of the processes used were more expensive than others, but gradually the manufacturing and processing technology matured to the point where it was viable, although still expensive, to produce Formula One valves. These valves were widely adopted at a time where Formula One engine speeds, although already high, were heading ever higher and at quite a rate of increase. Valvetrain control can be a major limiting factor in this push for higher engine speeds, and a lighter valve offers obvious advantages in this regard. So, it was that, in the quest for higher engine speeds, that titanium aluminide became widely adopted.

However, the valves made from this material seemed prone, in a small number of cases, to suffering premature failure. Where perhaps 99% of valves would happily last the required engine life, the remaining 1% could fail very early indeed, sometimes before the engine had run at full load. This situation seemed to occur after some time during which failures were rare. The material seemed to act in a brittle manner, and the failures could be quite spectacular, with large sections of the head becoming detached. The failures affected multiple teams, and emergency action was taken to put titanium valves back into action, accompanied by a consequent (and temporary) halt to the race for increased engine speeds. At the end of the same year, titanium aluminide was proscribed under the Formula One engine rules.

Written by Wayne Ward.

He works with the raised port (RP) head and relates, “The RP cylinder head layout is really hard on parts. The intake rocker is offset 3/4-inch and that exacerbates the stress at the adjuster and at the end of the pushrod, because you’ve got a rocking couple across the span of the rocker. This ‘ball on ball’ pushrod has an undercut on the ball and on the rocker end that allows for a better sweet at high lift.”

In building these engines, “The most important thing to keep the pushrods functioning in the engine is lubrication. You need the oil to lubricate and pull the heat away from the pushrod. We gave ourselves a good demonstration of this when I decided to upgrade the W9 engine to bigger lifters. I tightened up the lifter clearance,” he related.

“Without enough clearance, the oil couldn’t get past the lifter to the pushrods. We welded the pushrods to the adjusters in the first 30 seconds on the spin fixture,” he said. After Wirth increased the lifter clearance, put the lifter, new pushrod and adjuster back in the engine, he noted the same pushrod and adjuster were still in the engine – a year later.

Wirth started with a 7/16 x 0.065 wall pushrod. That worked well with the aluminium rockers and the valve spring he had chosen, but as the demand for more rpm increased from 8500 to 9500, the valvetrain parts began to fail. Steel rockers and a new valve spring brought stability back and increased the power output.

However, Wirth noticed a substantial power loss in the sixth harmonic range (about 8800-8900 rpm). “We found that increasing the wall thickness of the pushrod from 0.065 to 0.090 reduced the loss,” he related. “We then went to 0.120 and finally to 0.165 wall with each increase in wall thickness reducing the power loss further.”

According to Wirth, “The pushrods right now are bullet-proof. We’ve gone from using 5/16-inch five years ago to 7/16 and even 9/16- and 5/8-inch in NHRA Pro Stock,” he noted. “I firmly believe that you cannot make a pushrod too stiff.

Smith Bros’ Greg Tanner noted his company is “toying with making some ends out of tool steel (A10) instead of 86L20. A10 is an air hardened tool steel that has fair machining properties and very good durability when heat treated in a vacuum furnace to maximum hardness (58-60 RC). The cost for the material is 2-3 times that of 86L20, which has good machining properties and good durability. Our testing is just about to begin,” he confirmed.

Written by Anne Proffit.

The biggest problem with midget racing – like Pro Stock drag racing – is that the rate of acceleration of the engine can be exceptionally high, above 2000-3000 rpm per second. The valve spring must be able to control the mass of the valvetrain through that acceleration – on pavement and dirt.

Wirth began the process with the valve motion he wanted. Then he purchased samples of all the springs available with his proper dimension for fitment into the cylinder head, entering all the detailed information about the characteristics of these springs onto his spring dynamics software program. “This program gave us the safety margin threshold for each spring, using our valvetrain weights and target rpm of 9500 rpm,” Wirth told me.

At the time, there was only one spring that met adequate safety margins, so that was the one he used. After extended racing, though, Wirth started to experience rocker arm failure because of the cyclic fatigue of the aluminium. “Changing the rocker material to steel stopped the rocker failure, but also changed the dynamics of the valvetrain.”

Unable to figure it out, with his software not having “a provision for modulus of the material so that it could not predict this change or the increased stressed on the valve spring that caused them to fail,” Wirth contacted PSI for help with the problem.

Discussions with the engineers there indicated the types of failure Wirth was experiencing meant the “spring did not have enough rate to control the valvetrain dynamics in the engine,” Wirth said. Settling on a triple spring combination, this was tested on his spin fixture using a high-speed camera and video camera with a triggered strobe, where they watched for surge and spring rotation.

It was found that the triple spring combination had the dynamic control and fatigue cycle life to meet Wirth’s criteria of 460 service laps. “You look at it and you might like to have a threshold below one million cycles.

“When we finally did a production run of this spring,” Wirth said, “the manufacturer used their Max Life surface preparation,” (a micro shot-peening procedure) to further extend the cyclic life of the spring.”

Wirth’s Dodge Mopar is a four-cylinder midget engine of 166 cubic inches with a bore of 4.125 and stroke of 3.100. The triple spring that works for him is made of Kobe steel, “All the high-end chrome silicon wire comes from Kobe.” The Wirth way is to stiffen the valvetrain and control it dynamically with valve spring and pushrod stiffness.

The triple spring combination is heavier than any other spring Wirth uses, weighing in at 180.9 grams. “Using heavier [material] goes against contemporary thinking, but it meets our ultimate test: it works.”

Future racing challenges will also present new opportunities for him to move to new performance levels, Wirth said. “Then the process of matching valvetrain components will start all over again.”

Written by Anne Proffit.

Titanium has very low density compared to the more traditional poppet-valve materials and so is the kind of material that we look to when trying to minimise mass as part of a valvetrain optimisation exercise.

However, as I once explained to someone full of excitement at the prospect of turning some valves from Ti-6Al-4V on the lathe, it isn’t as simple as using any old piece of titanium. The titanium alloys used for racing poppet valves are carefully selected for their properties, and have specific heat-treatments applied for their specific application. They need to have the correct level of fatigue strength at operating temperature (not just room temperature!) and this is where we begin to depart from the common titanium alloys.

A common titanium valve alloy, which many of the manufacturers use, is Ti-6242, which is more highly alloyed than Ti-6Al-4V, and contains the same 6% of aluminium, along with 2% of tin, 4% of zirconium and 2% of molybdenum. There may optionally be small additions of other elements to this alloy.

The heat-treatment of the material, which is a solution treatment followed by aging, is different depending on the final application. Inlet valves and exhaust valves are treated differently according to their operating environment. The inlet valve is much better cooled than an exhaust valve, having a blast of relatively cool air/fuel mixture passing over it (in the case of a carburetted or port injected gasoline engine) each engine cycle, where an exhaust has to withstand the discharge of high temperature combustion products. Possibly, where scavenging is poor, the exhaust valve also has to withstand some cool gas passing over it, increasing the temperature range over which it has to operate. The exhaust valve material is generally treated to a lower level of strength than the inlet material.

It is possible to combine two pieces of the same titanium valve material, but having differing mechanical properties, arrived at by thermal means, or by different work-hardening processes in the same finished component. This is often done by the process of friction welding, where one component is held stationary whilst the other part to be welded is rotated at high speed and then brought into contact with the stationary part. Heat quickly builds, melting the two surfaces and providing a strong welded joint which is perfectly able to withstand the rigours of use as a poppet valve. This type of welding, which is used widely in industry, provided excellent quality across the whole joint face.

The welding, of course, takes place before final machining, and on visual inspection of a valve thus produced, there is no sign of there being two parts welded together.

The accompanying picture shows a friction welding process for an industrial application, and although it is not a racing valve, the process is very similar in principle.

So, we can see that, in addition to its low density and strength, titanium lends itself to manufacturing processes which mean that its properties can be tailored precisely to its intended application.

Written by Wayne Ward.

The upshot? Fourteen Pro Stock drivers – some fulltime, others who run the category on a part-time basis – are using the product to great affect. “Nobody has contacted us with any problems so far; not one of them has worn out the pushrod yet!

Normally, Pro Stock racers swap out their pushrods after 15 runs at the most, but there have been zero reports of wear to the pushrod or to the adjuster to date.

The problem with keeping oil flowing through the pushrod is due, in part, to the fact that Pro Stock engines normally use “0” weight oil. “Their oil is very thin and that has a lot to do with it. There is no cushion built in,” Manton told me.

The new pushrod adjuster insert material is a copper blend without beryllium using a high iron content and has to be cut with carbide, Manton said.

The bottom of the tip’s cavity has to be flat, as does the insert, in order to spread out properly on the tip. Tolerances are very, very tight, according to Manton, citing a normal tolerance of 0.0006.

Cost of the pushrod tube and its adjuster tip go up incrementally with this new product, thanks to a three-time increase in production effort to make the unit. While the tooling and machinery are the same, the new tip is more time consuming and more difficult to make, giving it a 35 percent price increase.

At this point, Manton is still experimenting with his new product “Since the original design, we have modified the adjuster tip twice and now have both radiused and V-cut tips available to our clients. As for who is using which ones, Manton is not about to divulge.

Larry Morgan, currently running a Mopar engine, figures he might be happy with the new adjuster on the new Ford Pro Stock mill he is about to start building up. “Our cams are so big we cut off the oil to the adjusters on the pushrod ends, so Terry’s copper insert is making a big difference,” Morgan said. “His insert has a bigger radius” to keep oil flowing through the pushrod.

While he believes the new pushrod insert could be a better product, Morgan is still using his older pushrods. “I fixed my problem without having to go to the newer pushrod end. We found with a stiff valvetrain, we can use a stiffer pushrod as a tuning tool. We control the life of the cam with the pushrod.”

Working with the older specification pushrod, Morgan has concentrated on his cam profile to get more power up high. “It has been a trial-and-error procedure because the camshaft profile has to be changed to get a more accommodating and stiffer valvetrain” that can last a significant period of time.

Manton expects his pushrod with the copper blend adjuster tip insert to increase both longevity and durability. No doubt, should it perform as expected – and once he settles on a distinct specification after experimentation – it will become the standard of the Pro Stock industry.

Written by Anne Proffit.

Jim Oberhofer, responsible for the Top Fuel Kalitta Motorsports rail run by Doug Kalitta on the NHRA Full Throttle Racing Series' 24-race tour, explains what the challenges are for a nitro motor. “We don't have high lift camshafts and high engine RPMs to worry about – like a Pro Stock engine. Our biggest challenges are overcoming blower boost, fuel volume and heat.”

Oberhofer likes to run a PAC 1350 Triple Spring on both the intake and exhaust sides, “while some teams prefer to use a titanium spring,” he related. “On the intake side, we will use a brand new PAC triple spring. We'll set it at about 550 PSI. We normally like to run a new spring on the intake side and break it in for about 10 runs. Then it goes to the exhaust side,” he said.

Oberhofer finds the “normal service life for a PAC triple valve spring around here is about 25-30 runs, which is significantly better than the five runs we used to get about 10 years ago.” While they do tend to wear out a lot faster on the exhaust side, the effect of beginning the usage on the intake side allows the Kalitta group to gain more runs from each spring.

In the early 1990s, most drag racing teams used a double spring with a dampener. “We would set them at about 275 pounds. When we started going to bigger fuel pumps better blowers, better flowing heads and bigger camshafts, the need for a better spring with more PSI became apparent. In 1993 we made the switch to a triple spring. When we first went to the triple spring, we would set them at 325 pounds on both intake and exhaust sides.

“We used to just flat wear springs out in 2001-2002,” Oberhofer continued. “There has been an evolution of the triple spring (since they began using it) and we are just on the latest version, which has been the best so far.”

These days the team will install the intake springs at 550 pounds and, after five runs or so they might lose 5-10 pounds of pressure. “Sometimes we might keep a spring on the intake side for up to 15 runs,” Oberhofer allowed. “It just depends on how much pressure they lose.”

Of course it is important to keep high PSI springs on the intake side because “the intake valve is always fighting against high blower boost and the high amounts of fuel volume we put in a nitro engine. It is important to have a very good and very reliable valve spring – otherwise, you could have a misfire in the manifold, which will lead to an engine explosion,” he explained.

On the exhaust side, Oberhofer runs a triple steel spring installed at about 425 PSI. “We change them every run because the exhaust valve spring sees extreme heat and, with that it will lose about 5-15 pounds of pressure at its installed height on each run. This is significantly better than the old springs we used to run that would lose about 30-40 PSI per run.”

Written by Anne Proffit.

Whilst hollow valves would offer a weight saving over their solid equivalent (given the same stem diameter), they would be less stiff and so would almost certainly be a disadvantage in the case that the solid valve design which they replace was close to optimal. The decreased stiffness would mean less control over the valve head, with the attendant risk of it seating in an unsatisfactory manner, namely at the wrong time and with a higher than desired velocity.

The other option, if we aim to maintain axial stiffness, is to increase the outside diameter whilst maintaining cross-sectional area. Whilst this increases bending stiffness, which might be a worthwhile aim, the increased blockage offered to the flow of gas might be significant and require some alteration of the port profile in order to maintain the desired area schedule.

So, we might conclude, quite reasonably, that there is little real benefit in replacing a previously optimised solid valve with a hollow part of the same axial stiffness, unless there was a need to increase the bending stiffness of the valve. So, what is it about the hollow valve that makes it so valued by those seeking to optimise the performance of their engine? The answer can be given by a single word – cooling.

In general a hollow valve will not be strictly hollow, but will contain a coolant. The coolant is, certainly to the extent of my knowledge, always metallic. Low-melting point metals or alloys are used to remove heat from the valve head. A certain percentage of internal volume of the hollow valve is filled with the cooling material (I’m assured that this is not as easy as it sounds) and, as the liquid is splashed around, it removes heat from the valve head and conducts it more efficiently away from the area from which it is removed (generally through the valve guide).

‘Splashing around’ is not a good description, and we should be more specific - the cooling material, the most popular of which is sodium, oscillates along the stem owing to the acceleration of the valve, removing heat far more efficiently than the phenomenon of conduction. This is especially relevant if the valves are made of titanium, which has very low thermal conductivity. As we already know that titanium is a favoured valve material, we should not be surprised to learn that the use of sodium-cooled hollow valves is widespread, especially for exhaust valves. Sodium-filled inlet valves can be successful, although it is quite common for these simply to be hollow structures.

Later developments in hollow valve technology have concentrated on extracting extra mass by producing a hollow head – until fairly recently most racing valves relied on a simple drilled stem.

Written by Wayne Ward.

Dr Randolph notes that the exotic materials solutions produced some beneficial effects, but to his mind, the cost-to-benefit trade-off just did not make the grade.

“There may be some teams who feel otherwise,” he said.

“The challenge with many of the composites is getting a hole up the centre of the pushrod for oiling the pushrod/rocker interface, and complying with the ‘magnetic’ rule from NASCAR,” Dr Randolph continued. The ‘magnetic’ rule (20.5 8.2, Section F) permits “Only magnetic steel one-piece pushrod assemblies without any moving parts.”

“To the best of my knowledge, most teams found the considerably more expensive pushrods offered very little benefit, because mass savings on the pushrod side of the valvetrain do not translate readily to performance gains,” Dr Randolph recalled. “In fact, the trend over the past several years has been towards thick, stiff, heavy steel pushrods.”

While most race-related parts tend toward the lighter end of the spectrum, in the case of pushrod technology, “Mass on the cam side of the rocker does not impact valve event dynamics nearly as much as mass on the valve side. In some cases, increased mass can improve high-speed performance,” Dr Randolph told me, by generating more loft (dynamic valve lift) past the nose of the cam.”

The only variations have been in single-piece versus three-piece pushrods with pressed-in tips, and straight versus tapered designs, in the realm of recent NASCAR use.

The single piece designs offer “excellent reliability but the roller tip shape is not optimal,” because of the manufacturing processes necessary to compress the material on the ends of the tube to form the spherical lobes that mate with the rocker – or swedge – a single tube.

The manufacturing cost differences, according to Dr Randolph, are miniscule between various designs of steel pushrods as compared with the costs of utilising exotic materials. “Conversely, three-piece designs offer the potential for extremely accurate ends, plus the opportunity to use different alloys for the ends, as opposed to the shaft.

“Unfortunately,” he noted, “the pressed-in tips can compromise reliability. Pressed-in tips create a stress and interface at the ends of the pushrods that do not exist in a single-piece swedged design. However, manufacturing the tips separate from the body of the pushrod allows the option to use different materials and can result in more precise ends. Both types are used commonly” in the NASCAR engine building community.

Because pushrods have a minor impact in engine performance but a more significant impact on valvetrain durability, Dr Randolph explained, the dynamic performance of the valvetrain system is definitely impacted by pushrod design. “This is a valvetrain systems design issue,” he noted. “What ‘works best’ will vary depending on the complete
valvetrain system.”

Dr Randolph does vary the pushrod design to change stiffness and frequency characteristics of the valvetrain. Replacement of units occurs when “we start to see any sign of wear on the ball,” he said. There are different designs for different applications, including plate races, but Dr Randolph is adamant that “they all last forever!”

Written by Anne Proffit.

“In the late 1990s I helped design the current racing valve spring and it really hasn’t changed much over the years. It is still made of an alloy silicon chrome with vanadium added to the mixture,” Rickard told me.

The only real manufacturing changes he’s noted were that the materials have got better over the last four years. Where once all units were shot-peened, there is now a micron beam used in manufacturing with very tiny shot polishing added that reduces fatigue, he said. While the finish is still silver, it’s more of a matte colour.

“We have taken the polish off and this micro-peen adds to the longevity and lack of fatigue,” Rickard said. The micro peen process utilises very small shot media, producing an extremely high compressive stress layer on the surface of the spring which, in turn, reduces the likelihood of surface initiated failures.

The use of symmetrical springs has decreased among racers who use single springs look to the ‘beehive’ valve for their applications. The beehive valve spring gets smaller as it gets taller and the thickness changes, as well, although this shape is more of a fir cone than a simple traffic cone shape. “With the beehive valve spring, you don’t need a damper because there is no harmonic frequency change. The top retainer is smaller and lighter, too.”

Tony Oddo, engine builder for Bill McAnally Racing’s NASCAR Camping World Series West driver Paulie Harraka, is sold on the PSI beehive valve spring, he told me.

“They really seem to hold up well,” he said. “We are able to run smaller retainers and keepers, so there is less weight to the whole valvetrain.” He has been using this product exclusively for the past five years, in the LS2 NASCAR spec engines he builds.

Oddo finds there is a substantial change in diameter from bottom to top of the spring and he attains considerable weight savings, as well. “The savings on the whole valvetrain,” Oddo cited, “are between 20-30 grams and any time you take a gram off the valvetrain, it is a big deal.”

The older type of spring has a larger diameter than the current beehive valve spring and needs a damper and inner spring inside. “They were really heavy,” Oddo said. “We had to work them very hard to get them to push weight.”

The older spring could only be revved to 7500 rpm. “We had to work those springs really hard to get them to push weight. The installation of beehive valve springs “allows us to spin the engine harder. When you have a real light valvetrain, you can spin the motor to 8000-8200 rpm, while still pushing 300-350 pounds.”

Service life of the valve spring has also doubled to between 3-4 races, a cost savings for any engine builder. “We don’t have any problems with them as long as the driver doesn’t overheat the engine,” Oddo explained. “When that happens, it takes about 20-30 pounds right out of the spring,” and lessens valvetrain life.

Written by Anne Proffit.

As engine speeds have increased, there has been continuous development of the valves. One of the main reasons for this development is to reduce the mass of the valve in an attempt to maintain control over the valve motion with the increase in speed. In order to preserve the good breathing characteristics of the engine, the cam profile imposes ever-increasing accelerations and the consequent forces are mitigated by reductions of valve train mass. Elimination of multiple steel (or titanium) wire springs and the spring retainers by using pneumatic valve-return systems (often call air springs) was a very important step in eliminating mass. In the early days these systems often featured some quite bulky and heavy (by today’s standards) components. However, these systems now use some very lightweight parts, with at least one team using magnesium spring-tops (effectively a piston which carries a pneumatic seal) before this material was banned from use in Formula One engines. Higher system pressures mean that the size of this piston is now pretty small, with an appropriately low mass, owing to very thin walls where possible. The collets are often made from a low-density alloy, although the tappet shim (lash-cap) is still generally made from hardened steel and is often coated.

Therefore, there is only one component remaining where there is significant mass to target: the valve. There was a very significant step in reduction of valve mass with the advent of the lightweight titanium valve, although these were in common use before the appearance of the pneumatic valve-return system. Thus there has been intense pressure to reduce the mass of the valve. This problem was complicated because, as speeds increased, there was a natural tendency for the bore of the engine to increase, and with it the valve sizes – a situation which would normally lead to an increased valve mass. There are two quite obvious measures which can be taken to reduce the component weight for a given valve diameter and length; we can reduce the stem diameter, and flatten the head profile (the ‘nail-head’ profile as described last month is a good description).

In general, if we choose to reduce the stem diameter, we need to make an effort to reduce the valve-head mass; otherwise we will soon get ourselves into trouble with loss of valve control with all of the attendant problems that that can bring. The flattened ‘nail-head’ will, in many cases, bring with it an unwelcome reduction in breathing capacity. Therefore many people have, in the past, used the normal tulip-headed valve, with the valve face in the combustion chamber dished. This too has a complication in that the compression ratio is decreased owing to an increased TDC volume. This can be regained by adding a corresponding volume of material to the face of the piston pockets, or elsewhere in the combustion chamber if this is possible. Adding material to the piston means that extra effort needs to be made to reduce the mass of the piston somewhere else. There is evidence to suggest that, with a dished tulip-headed valve having an equal breathing capacity to an un-dished version, there may be a penalty in combustion efficiency which leads again to lower engine output. As we can see, there is no simple solution to valve-mass reduction.

However, we should not despair; the flattened nail-head valve can be used for its virtue of low mass and there are features which can be introduced to the valve to restore the original flow characteristics and even improve upon them. Machined features on the valve head, which need to be carefully designed which, if we aren’t to require lots of engine tests, can give us everything that we need for our lightweight valve with good breathing. Furthermore, these features can, if ill-considered and poorly designed, lead to premature valve failure.

Written by Wayne Ward.

According Abbett, this type of intake valve significantly lightens the overall valvetrain and is easier on the lifters, camshaft, rockers and valve springs during standout driver Antron Brown’s sub-four-second, 1000-foot trips down the dragstrip. In his second year competing in this category, at the time of writing Brown stood second in the NHRA Top Fuel Full Throttle Racing Series points chase and was a two-time 2009 winner.

When Mike Ashley Racing switched to Alan Johnson cylinder heads toward the end of the 2006 NHRA drag racing season, the transfer of seat material to the valve was a heat-related issue for Abbett. “We tried to lighten it up but there wasn’t a big difference. It was the CrN (chromium nitride) coating (provided by the Xceldyne division of CV Products in North Carolina),that reduced material transfer. Once we had that, we could control valve life, get the valve to last longer and the springs to last longer,” he said.

Through the use of this CrN coating on his Xceldyne hollow-stem titanium valves, Abbett has discovered the life cycle of his springs improved and he gets three times the life from his valves. “We get eight to 12 runs now where we were lucky to get two runs before. We used to go through a lot of intake valves… “ he mused.

The valve in a Top Fuel engine making nearly 8000 horsepower has a tendency to get burned before failure of a coating occurs but as Abbett explains, “That is another part of those tune-up issues we constantly face.”

Not every nitro-powered car in the NHRA uses either hollow-stem valves or CrN coatings. There is a $25 financial penalty for the coating, Abbett told me. Some of the Funny Car teams are reluctant to try hollow-stem titanium valves because those machines are more susceptible to fire when the intake valves fail. The backfire can go on to the blower (supercharger) and ignite ferociously from there, Abbett explained (Mike Ashley Racing competed solely in the Funny Car arena until this season).

Another good technology boost for the Top Fuel team is Xceldyne’s change from square to round keeper grooves for the intake valves. “There is less tendency to pop the head off with these round keeper grooves,” Abbett said. “It, too, has certainly helped with valve longevity.”

Written by Anne Proffit.

Every month or so, Johnson gets new springs to try; often the updates come quicker. There might be two or three varieties and maybe, once in a while, he might find one that works for his high revving 500-cubic-inch mill. “First we’ll test them on the Spintron and if that doesn’t tear it up, we’ll put the spring in one of our dyno engines,” Johnson explained.

Johnson always uses a down-on-horsepower engine on the dyno to test a newer specification valve spring and, if it works to his satisfaction, Johnson will place the spring in a newer motor. If the updated spring tests successfully on the newer mill, it will then go to the track for further testing. If a new-spec valve spring lives through further torture, it might get fitted to a race engine for the first qualifying session – of four – during the Full Throttle Drag Racing Series race weekend.

“Every day that we work on the dyno we are looking to develop a different size wire and a different diameter spring, trying to get it to live, but most of the time the camshaft beats the heck out of it!”

Steve Bown of PSI Springs confirmed that Johnson “either breaks them or wears the springs out pretty quickly,” he said. “Our new Triple drag racing spring seems to be working for Roy and (some of) his competitors pretty well right now, but we continue to research and upgrade our products to help these guys.”

During the Bristol weekend in mid-May, another Pro Stock team was trying PSI’s latest spring configuration, the DR1263RML and was happy with it following an initial run. This triple spring, along with a couple of dual spring options are available to the Pro Stock community.

“Most significantly,” Bown told me, “We will soon be introducing a new line of springs for drag racing that we have developed. These utilize a new manufacturing process – signified by an ‘R’ following the part number – that improves the strength of the material, thus improving the life of the springs dramatically.”

Written by Anne Proffit.

For two years, Terry Manton of Manton Pushrods in Lake Elsinore, California has been working to find a solution to stop these burn-up difficulties. “It’s been two years worth of study and testing and nothing solved the problem until now. Those pushrod tips were just burning up continuously,” he said.

Manton believes he has now found the solution with a new, very high-grade hybrid copper that does not include beryllium material, he told us. “We’ve got a new copper insert for the top of the pushrod cap that allows the Pro Stock teams to use a 281 thou ball adjusting screw, without burning up the pushrod top and the adjusting screw. That’s always been a big, big problem with these machines,” Manton said.

“One other attached component to this pushrod solution for the Pro Stock racers is the new tool steel adjusting screws that complement the copper cup insert,” Manton explained. Of course, the final surface hardening treatment applied to this adjusting screw is a compound he’s keeping close to the vest. The new part has only been in use since the beginning of April and was first tested on a dynamometer by Kurt Johnson, son of Warren Johnson. “It looks better than anything we’ve used before,” Johnson said. Kurt and Warren Johnson are using the piece, as is Greg Stanfield, David Nickens and V Gaines.

Because they can use dynos to test and measure their engines, unlike the nitro-burning cars in Top Fuel and Funny Car, Pro Stock teams are constantly trying new options between races. “We really don’t know yet how long these pushrods will last because they still haven’t burnt them up, even with all the testing these guys do between races,” according to Manton. “It’s all about the copper insert in the pushrod tip; that is what makes the biggest difference,” Manton said. “The teams are reporting increases of five to seven horsepower and less friction at the adjusting screw contact area.”

Seven drivers have been using the new product, which fits all 3/8-9/16-inch pushrod diameters in use for the Pro Stock category. “It is still so new that we do not yet have a good amount of information on the product, but so far, everything appears to be working extremely well and we are very excited about it.”

Written by Anne Proffit.

The following is a bite-sized synopsis of a few of those engineering challenges.Sources of loads experienced in a pushrod-rocker system:

Spring-related 

Whenever a rocker arm is holding a valve off its seat, that rocker arm experiences contact, bending (tensile and compressive), and shear stresses as a result of valve spring loads. Pushrods must bear compression, contact, shear, and even bending (resulting from column-deformation) stresses. The compressive load applied to the tip of the pushrod is the instantaneous spring force multiplied by the instantaneous rocker ratio. (Note that spring forces on the rocker and pushrod go to zero during lofting.

Acceleration-related 

In an effort to maximize the area under the valve lift versus rotation curve, race cam lobes incorporate the highest levels of acceleration (increasing the velocity away from the seat and decreasing the velocity moving toward the seat) the lobe designer thinks the valvetrain will survive. The acceleration and jerk (impact) forces applied to the rocker and pushrod as a result of the opening and closing ramps of the cam lobes increase with the square of rpm.  Those loads apply the same types of stresses to the rocker and pushrod as do the valve springs, but at much higher levels.  Spring loads prevail (over the nose) at lower speeds.  Acceleration and jerk forces prevail at higher rpm, and the over-the-nose load the springs apply to the rocker and pushrod approach zero as rpm increases. It is also interesting to note that these acceleration forces are transmitted across the rocker arm to the tip of the pushrod at the square of the rocker ratio.

Impact 

A poorly designed or manufactured cam lobe can have high rates of change of acceleration (jerk) which imparts impact loads to the entire valvetrain. Valve lofting causes impact loads on the entire valvetrain when the separated valvetrain components reconnect. (I know of one particular helicopter engine in which the cam lobe acceleration curves rise nicely then go flat to a constant acceleration value, then suddenly ramp downward toward zero on both the opening and closing side. The resulting jerk forces typically destroy the valvetrain in about 200 hours of operation at 4200 rpm.)

Vibration 

Intermittent forces applied to a mechanical system at a regular interval (excitation) will cause components to vibrate in any number of ways if the excitation or a harmonic of the excitation is close to a natural frequency of a given component. The acceleration profile of an aggressive cam lobe can contain several high-amplitude harmonics, any of which could coincide with a component’s natural frequency. It is important to try and design valvetrain components (pushrod, rocker and springs especially) so that their natural frequencies are as far away from high excitation harmonics as possible. That is often much easier said than done, especially in engines having a wide operating band. (Cup engines at Martinsville go from about 5500 to 9300 rpm twice per lap.)

Cylinder pressure acting on the valve area

For a NASCAR Cup engine at full tilt, Professor Blair noted that the cylinder pressure early in exhaust valve opening (EVO) is sufficient to keep the valve seated for 8-10 cam degrees while the pushrod-rocker valvetrain flexes to accommodate the motion imparted by the cam ramp. That situation is even worse in a blown nitromethane NHRA motor, which retained so much cylinder pressure at EVO that exhaust pushrods can be bent, and until the advent of some recent technology, the use of roller-tipped exhaust rockers was often avoided.

Friction in the rocker bearings and rocker-valve interface

For the sake of completeness, friction needs to be considered. In a non-roller-tip rocker arm, the friction at the valve stem interface can be high, and the resulting lateral forces applied to the valve can generate some rapid valve guide wear. In some cases, friction can be beneficial in that it can apply damping forces to various components in the system and thereby help to control the amplitude of vibrations being driven by an excitation frequency in the vicinity of resonance.