What is not yet so widespread though is the integrated use of the various CAE approaches. When several load cases (such as mechanical and thermal loads) come together, and perhaps fluids are involved as well then things become very complex, and a high level of experience is needed to make full use of CAE.

One of the more complex areas in engine design is the interface between the cylinder head and the crankcase. Granted, we all know we need to keep sufficient clamping load on the cylinder head gasket to secure sealing under all conditions, but what really happens to the cylinder head, crankcase, liners, pistons and so on under varying load conditions – and being able to predict what happens – is a different thing altogether. Trying to predict in which circumstances this interface is most critical requires a structured approach and, as well as skilled CAE engineers, sufficient computational power to get to realistic results.

First though we should clarify what we are trying to achieve. The main goal is to determine the deformations working on the interface, see which stresses will arise because of this and how the interface reacts to the loads under certain operating conditions. These conditions include not only firing (temperature and gas load) but assembly loads such as the interference fit of the valve seats and guides, as well as cylinder head bolt pre-load – and, if applicable, injector clamping loads. All of these will lead to cylinder head gasket deformation and stress.

Some typical load cases can be described which, in certain combinations, will be the most critical load case. Take assembly loads for example. The engine itself is in cold conditions and all components have been assembled with the correct bolt pre-loads and interference fits. This will lead to an initial stress in the components.

Then there is the temperature load, which depends on the temperature of the combustion gases as well as that of fluids such as oil and coolant. Certain deformations and stresses also occur here.

The third case, which is related to some extent to the assembly loads, is the load coming from external system effects – in this case, cylinder head deformation as a result of cyclic injector load transferred through the injector clamp onto the head. With gasoline fuel injectors these loads can be neglected, but with diesel injectors they can play a major role when combined with the other loads.

As a next step, the combined load cases that would lead to the most critical situation should be defined. Something that is often not considered is that one of the most critical operating conditions for the cylinder head gasket to perform its sealing function is when an engine is warming up. Everything is cold, so overall clamping load of the bolts onto the gasket is minimal, and combustion load is trying to lift the head from the gasket.

So, in these CAE simulations the related components are modelled, mostly in a CAD system, after which the components will need to be meshed (the creation of small elements within the components) for further use with CAE simulation tools. Meshing will need some attention, where typically the shape and size of the mesh is based on the engineer’s experience and will be tailored to specific analysis needs (smaller in areas of interest and somewhat bigger in other areas, all to reduce simulation duration).

The distribution of the temperature throughout the components, especially within the cylinder head and top end of the crankcase, then needs to be determined. Together with the heat transfer coefficients of the various materials in use, this will allow the stresses in the components to be simulated, as well as the predicted component and surface temperatures under the chosen – and most critical – running conditions.

When adding in the external system loads, which lead mostly to local stresses in certain areas of the cylinder head – as with high-pressure diesel injection – the overall stress levels can be simulated. The only thing left to do now is to determine the crank angle (position) at which the combined loads are critical. Here, some rather straightforward analysis will provide this type of information rather easily. Then the computational power comes into play, calculating the statuses of the separate elements in the mesh using finite element methodology.

Then, as a last step, post-processing the results allows us to determine which information is relevant. Mostly we would be interested in verifying the design and predicting the lifetime of some of the parts. This includes gasket pressure but also, for example, the liner deformation in order to predict oil consumption. Apart from this, critical temperatures could be subject to attention as well, such as the maximum temperature on the liner and piston top ring, cylinder head flame deck and many other areas.

I hope from this short summary that integrated simulation techniques have come a long way in being able to predict with greater accuracy what is happening inside an engine at critical interfaces. Some things are now becoming possible with hybrid systems, in areas perhaps even more complex than ‘just the internal combustion engine’ – a design we have been studying for more than 100 years now.

Fig. 1 - Cylinder head gasket pressure shown rotated over 360º of cylinder angle

Fig. 2 - FEA result of local gasket pressure used to quantify sealing quality of the cylinder head gasket, in this case sealing of the combustion chamber

Written by Dieter van der Put

Other than these, and mainly for manufacturing and serviceability reasons, the major engine parts are designed as we know them today, with separate components connected using bolts and sealed with one or more gaskets. As these designs became standard practice, the development of gasket materials started. Initially, the loading on the gaskets was rather low because of moderate power output and limited internal pressures, so gasket design was focused on sealing two or three mating components with only small machining defects, porosities and other imperfections. This is why early gaskets consisted of materials that yielded to some extent, enabling them to deform and fill those imperfections.

As engine temperatures and pressures increased though, these gaskets showed significant failure rates, typically leading to blown head gaskets and/or leaking exhaust manifolds. As early adopters of new concepts, Formula One engineers switched to copper or steel ring head gaskets, which showed to be a significant improvement for sealing the combustion chamber. Soon after that, these gaskets were introduced for production cars as well, where additionally, for ease of assembly, they became integrated into cylinder head gaskets.

Later, the head gasket material changed to single-layer steel. This was a major improvement over earlier materials, which often contained asbestos, a now forbidden material. Single-layer steel gaskets provide significant better mechanical strength as well as the ability to cope better with micro-movements of the mating surfaces. With earlier materials these movements could lead to gasket damage and possibly cracks in the gasket material.

As one of the later steps in the development, multi-layer steel (MLS) gaskets were introduced, with so-called ‘beads’ to increase local sealing capability. To get a feel for their dimensions, the bead shapes are typically up to about 0.4 mm high, depending on size and application, where the actual layer thickness is even less (about 0.25 mm). The top and bottom layers are often coated to reduce friction, in case micro-movements occur.

For  exhausts in particular, given their extremely high temperatures, MLS gaskets have vastly improved the sealing capability, although the actual sealing geometry – as in the bead-deformed zones – remains the sensitive area of the gasket due to relaxation of the material under high temperatures.

The basis of an MLS gasket are multiple layers of steel, of which one or more layers have pressed beads, being positioned on top of each other. When the layers are being assembled and the bolts are being torqued, the beads deform, leading to elevated local pressures in the beads, the actual sealing location. These beads will deform partly plastic and partly elastic, meaning they can adapt to temperature and load fluctuations (as long as the fluctuations are not larger than the elastic compression of the bead).

What was occasionally seen however was that the beads were compressed more plastically than intended, reducing the remaining elasticity in the gasket beads, leading to too little reserve against temperature and load fluctuations. To improve this, a so-called ‘stopper’ layer was introduced. The thickness of this layer – which is also the third or middle layer of the MLS gasket – determines the amount of maximum deformation of the beads of the upper and lower layers, leading to better control over the elastic deformation of the beads. This can be controlled by the thickness of this middle layer.

If a further, more local optimisation of the bead deformation needs to be achieved, the stopper or middle layer can also be designed with a certain topography rather than just being flat. For example, it might be desirable that the beads have more elastic deformation between two bolts, giving a lower clamping load with increased bolt spacing.

In general there are no limits to how this stopper or middle layer can be shaped. That means a very homogeneous contact pressure distribution can be achieved, even with flat gaskets and significant bolt spans.

Fig. 1 - Multi-layer steel gasket with stopper element seen between the two outer layers

Written by Dieter van der Put

When considering coolant jacket height, the mid- or bottom stop liner, which has its mounting surface located low on the liner, gives the design freedom to achieve an open-deck design and therefore to get the coolant highest up and as near as possible to the flame deck. The main aim is to have the top ring in top dead centre, as close to the coolant as possible. With modern pistons, with very little compression height, the top ring is positioned very near to the top deck and therefore the jacket design around the liner is of major importance.

An open-deck solution like the bottom or mid-stop would therefore be preferable. From the point of view of overall length, however, this design is not really preferred.

But what if your engine has a top stop liner? First, you could ask a specialist company to convert your engine block to mid- or bottom stop liners. Make sure you find a company with excellent references though, because this is a very delicate job, and if it’s not done properly it will cost you your engine.

Another option would be to check in detail how your liner has been positioned. Most engines have their liners positioned just under the liner's flange and typically have a tight radial clearance between liner and block. With this design the coolant cannot fully reach the lower side of the flange, leading to decrease in cooling of the top ring.

One solution would be to position the flange on the top diameter (flange outer diameter). This way the diameter of the liner can be decreased just below the flange, which enables coolant entry into this region. Make sure the clamping force between the head and block, which is being transferred through the flange, can still be withstood by the engine block balcony below the flange.

By creating this additional coolant ring below the flange the balcony is obviously reduced, so we need to ensure that the strength remains sufficient. One possibility to achieve the required sealing of the coolant between liner and block would be to include an O-ring groove in the outer diameter of the flange, and install a small O-ring here. This is to ensure proper sealing of the coolant to prevent leakages under the head gasket.

What should this coolant ring look like? For cooling, flow is required, so the ring should be big enough to enable flow. The target should be to achieve a ring of at least 2 mm, depending of course on the pressure drop over the ring. It might therefore be a good thing to connect the ring on one side to the highest possible pressure (directly after the coolant pump) and on the other side to the suction side of the coolant pump. That way the highest pressure drop will be achieved and the ring can be kept as small as possible, limiting the overall length of the engine.

Coming back to the sealing, one might also choose to make use of a line contact between the underside of the liner flange and the block. The benefit here is that no O-ring is needed; the downside though is that this feature might carve itself into the block structure under thermal and combustion loads (leading to micro-movements between liner and block).

If there are still cooling difficulties after these conversions, another option is to drill holes through the flange of the liner to enable flow through (and therefore also just below) the liner flange. This might require additional modifications to the head gasket and cylinder head, but that is not always possible in existing engines.

In the end though, if cooling issues remain in the top end region then it might not have been the best thing to choose that particular engine for the increased performance you had in mind.

Fig. 1 - Cylinder liner concept and alignment/mating surfaces. You can see the difference in coolant height in relation to the piston top end ring

Written by Dieter van der Put

Another significant engine component that has a direct relationship with these increased loads is the cylinder head, so let us take a closer look at the consequences for the head, and which design modifications would be needed to overcome them.

In principle, every cylinder head has three major loads working on its structure – assembly (bolt loads), thermal and gas loads. In the previous article we saw that the maximum gas load has risen significantly, given the fact that the peak firing pressure went from 100 bar with the naturally aspirated V8 engines to more than 200 bar with the current turbo engines. Because of this we estimated an increase in bolt size from M11 to M14. The corresponding load on the cylinder head increases as well, and therefore the material thickness under each bolt head will need to be increased.

Since the bore of the engine has fallen from about 98 to 80 mm we already have a little less room for the coolant jacket, and because of the increase in material thickness another couple of millimetres is taken from the coolant jacket.

The other two loads have a direct relationship with each other and need to be balanced. On the one hand, the increase in peak firing pressure would require the flame deck of the cylinder head to be thicker, simply to withstand the bending of the head under these increased gas loads. Unfortunately, a thicker flame deck means the coolant is further away from the combustion temperatures, leading to higher material surface temperatures. And we mustn’t forget the fact that we were facing increased specific power, again a reason for higher flame deck temperature.

The most critical issue here is thermo-mechanical fatigue (TMF), which is damage of the material based on the differential temperature between hot and cold areas in relation to the material’s inability to expand and shrink freely under these conditions. The question now is which of the measures – increasing or decreasing the thickness of the flame deck – will be more important.

In this context the reduction of the bore by about 18 mm is helpful, simply because the increased gas load works on a smaller area. The resulting average load increase will be about 33%, a negative factor.

So what works in our favour then? Since the wall thicknesses of the ports are constrained by the casting process rather than design, there would be no reason to assume that these have decreased with the bore reduction to 80 mm. Therefore the structural function of these ports in the cylinder head (creating a kind of pillar between the flame deck and the upper deck of the head) has increased relatively.

On the other hand, cooling of the cylinder head needs to be improved because of the higher specific power (from about 225 to 300 kW/l). This increase will result in higher temperatures of the area around the exhaust valves and higher strain in the valve bridges between the exhaust and intake valves.

Another factor is the valve orientation, which is such that the exhaust valves are on one side of the engine and opposite to the intake valves. Thermal growth is more difficult to control in the longitudinal direction of the engines because the cylinders are next to each other. The internal material damage will therefore be greater because in this direction the head provides more resistance during expansion under temperature load, putting the highest strain in the intake-to-exhaust valve bridges. This, in combination with the fall-off in aluminium’s material properties under higher temperatures, means that reducing temperature in these areas is key.

Since the valve area is filling almost the entire bore, there are not many areas where the flame deck wall thickness can really be kept at its minimum. Looking at the very few pictures of Formula One cylinder heads that are available, I would estimate a wall thickness of around 8 mm as a minimum.

Not having any design experience specifically with Formula One engines, I am going to make an educated guess at estimating a reduction of up to 30-40 C flame deck temperature per 1 mm wall thickness reduction (getting the coolant closer to the combustion). These levels of temperature reduction will significantly improve resistance to TMF.

This is the main reason for me coming to the conclusion that the maximum effort, going from naturally aspirated V8s to the current turbo engines, must have been in reducing the flame deck wall thickness instead of increasing it due to gas loads, and I think the area between the valves has probably received the most attention.

We should not forget though that achieving these kind of reductions is not only how well one can design these critical areas but, even more important, how they can be produced – the cylinder head core packages must have been more detailed and more accurate than ever before. Hopefully it will soon be possible to write an article that focuses on these production processes.

Fig. 1 - In red is shown the wall thicknesses between coolant and flame deck. As can be seen, there is no real homogeneous wall thickness, given the complex structure of the head (Courtesy of Toyota Motorsport)

Written by Dieter van der Put

Let us start with the peak firing pressure (PFP), which is one of the most decisive factors in crankcase design. With the introduction of the turbo engines, the PFP increased from just over 100 bar to around 200 bar. This dictates the strength concept of the engine and therefore the major dimensions, such as cylinder distance, main bearing and cylinder head bolt sizes and crankshaft dimensions.

The Formula One V8 engines of 2013 were regulated to a 98 mm maximum bore with a cylinder spacing dictated at 106.5 mm, leading to a distance-to-bore ratio of 1.10. The 2014 Formula One engines, however, do not have a fixed bore spacing but a mandated bore of 80 mm. Given the higher load on the piston system, and subsequently on the block-head interface, let us estimate that this ratio will be increased slightly, to 1.15. This would lead to a cylinder distance of 92 mm. Calculating the overall block length, this would then lead to a reduction of about 14 mm per cylinder. Adding in the reduction from eight cylinders to 6, this would lead to an estimated reduction in length of 150 mm. This of course is only the ICE; all other power unit systems will need to be taken into account as well.

Looking further at the crankcase and the dimensions, we can compare bore sizes and the respective maximum gas loads. With the V8s, the maximum pressure of 100 bar would work on an effective diameter of a little over 98 mm (the gasket diameter is always somewhat larger than the bore, let us say 102 mm). With the 2014 V6 turbo though, the maximum pressure of 200 bar is working on a diameter of 88 mm (again, gasket diameter). This means an increase in load on the piston (and therefore on the head as well) of about 25%. We’re talking here only about gas load, without taking into account the extra thermal loading because of increased specific power (from about 225 kW/l to about 300 kW/l).

To keep the engine from falling apart under these loads, the cylinder head bolt size would therefore have to be increased, which I estimate would be from something like an M11 to an M14.

Then, looking at the height, the effect of the longer stroke (about 40 mm to about 53 mm) and the subsequent increase of the con rod length needs to be taken into account. In addition, the higher PFP would require an increased compression height of the piston as well (which is of course related to piston material, which might even be steel nowadays). I would estimate an overall increase from crankcase centreline to the top deck of the crankcase of around 40 mm.

Earlier we saw that the thermal loading of the piston system and cylinder head has also significantly increased. Together with the gas force on the piston, this would mean a definite requirement for a stronger piston and increased dimensions on the bearing system of piston-con rod interface, as well as the crankshaft.

All in all, the introduction of the turbo engines has led to a significantly different crankcase. Where the length could be reduced considerably, the height went the other way. Owing to the enormous increase in PFP, the whole crankcase structure – including the cranktrain – has had to be engineered to withstand these loads, in the end leading to the regulatory increase in minimum engine mass from the 90 kg with the V8 Formula One engines to the 145 kg now. This in turn has led to a totally different power-to-weight ratio, a shocking decrease from 0.18 kg/kW to 0.31 kg/kW. What is also interesting to see – and this is a nice example of the fact that engines all look alike – is that the engine mass per litre engine size, now at about 90 kg/l against 40 kg/l for the V8s, are in the same range as current production engines.

Fig. 1 - Detail of the Toyota RVX9 engine, with the cylinder head bolt spacing visible in the lower left corner. Also, several coolant ports between head and block (blue-coloured cavities) can be seen (Courtesy of Toyota Motorsport)

Fig 2 - Renault Sport F1’s engine for 2014, showing the ICE in the middle. Although not a very detailed image, one could imagine from the view that there is still a four-bolt connection between head and block

Written by Dieter van der Put

When I read about developments of some kind in gaskets or seals, I sometimes also wonder if the developments really work and if they will keep working in all kinds of circumstances. And ‘circumstances’ can be read here in the broadest sense – do they function under all kinds of ambient conditions, for example, but also do they work with all their surrounding parts with all their deviations. In other words: how robust is the gasket or seal in its environment?

There all kinds of seals and gaskets, but in general you can differentiate between flat seals and form seals, both of which have their various different shapes and designs. Flat seals can for example be made out of materials like organic compounds and steel sheet, while form seals can consist of many materials, like EPDM, Viton and others, and can be shaped like a simple O-ring or as complex shape to seal more intricate components.

Given this broad range of applications, how can we be certain that these gaskets will remain intact throughout their design life. Most suppliers will say their product is the best there is, and most of the time they are directing the statement towards racing in particular by adding that the products are the most robust ones available, that they are made especially for racing purposes. I always ask myself what that means. To my mind, every component is designed to withstand the boundary conditions it has to meet. So, racing or not, robustness is just a matter of what was written in the technical requirement specification, nothing more, nothing less.

And let us look at the initial use of the part and not bother, for now, about durability over its lifetime. Why? Because part of the issue is related to processes, so it appears in very early life/use of the part.

For example, many tests during the development process are being done with early development components, which are often taken very good care of at the parts’ suppliers. These parts are so good that most of the specifications are being met, and are often made to be nominal parts – which of course they won’t be when they are getting more into their production phase. And it really does not matter that much whether we are talking high volumes (production race series engines, for example) or low volumes (prototype racing parts). Suppliers always have a different viewpoint from engine manufacturers: when the full specification range can be used, more parts can be approved, which brings the most money to the company. And that is what every company needs to survive and make a profit – and that is how it should be, otherwise no business would survive.

During development therefore, near-limit parts are not being tested regularly, simply because they are not available (it is really very difficult to produce near-limit parts).

So, depending on the engine manufacturer, near-limit testing will be done to a lesser or greater extent. Specific parts are often distorted in certain areas – local damage on gasket surfaces, roughness steps at the sealing areas, and so on – in order to predict possible deviations in later ‘production’ parts, and tested accordingly. Of course, these test specimens should not fully destroy the testing result, since testing time and equipment is too valuable for that. This type of testing is often scheduled as a smaller part of a broader test, in order to be as efficient as possible with test facilities.

To minimise the influence of gaskets and seals on the robustness of an engine, as far as possible gasket and seal types are being chosen that have little sensitivity to deviations in mating parts, which is why you can often see the O-ring type of form seals being used. These have the best track record when it comes to surface defects, due to the fact that their sealing performance depends on compression of the material. Often the material is being compressed to about 75% of its nominal thickness, on which small surface defects have very little influence.

Typical flat gaskets are ‘desensitised’ by adding so-called beads to them to create greater compression, to try to approach the behaviour of form seals. The most sensitive gaskets as regards surface defects are steel sheet flat types, and these are therefore tested near limit as the structural part of development. Also, the mating parts with these kinds of gaskets – often for cylinder heads and blocks – are checked in great detail for surface quality and defects. Only in this way can the design be made and kept robust, from the very first test sample to the later, more production-related components.

Written by Dieter van der Put

We humans are a very conservative species, especially when it comes to engines. What we have been doing on internal combustion engines for the past 100-plus years must be good, we all tell ourselves, and all the money spent on developing them cannot be thrown away just because a new concept could (or would) provide added value. So let's stick with the piston engine and focus on the area most related to efficiency – the contact zone between liner and piston rings.

Efficiency is the magic word when it comes to engine power and fuel efficiency, both of which are important factors in motorsport. The principle is simple: get as much energy as possible out of the fuel and air being burned in the combustion chamber and transfer it to the driven wheel(s).

As engineers, we learn that the delta between power and efficiency is the friction in the various stages between the combustion chamber and the wheels, as well as the energy lost in the exhaust gas and the coolant. The main area of friction is in the piston package, mainly the friction between piston skirt and liner, as well as between piston rings and liner.

The friction between the rings and the liner depends on many factors. Some are to do with by the chosen engine concept, others are related more to the chosen material specification and geometry of the piston rings and liner. Let us look at both. 

Engine concept

The piston rings need to follow the liner in all circumstances. In doing so, they need to have as little blow-by as possible to maintain the compression and performance. Piston ring life will be much longer when the liner’s roundness is very accurate, and this is influenced mainly by how homogeneous the engine concept works out on initial liner deformation – the liner’s deformation after assembly of the cylinder head and in hot running conditions.

An even cylinder head bolt spacing, acceptable cylinder head bolt forces and homogeneous cylinder head gasket loading are key. That way the piston rings can be designed to have lower radial force on the liner without losing contact, all leading to lower friction. 

Material specification and geometry

Assuming the 'right' engine concept has been chosen, there is still a contact zone between piston rings and liner to be defined in more detail. On the ring it is desirable to have as little contact as possible without actually losing contact, and this can be achieved by properly shaping the contact zone (the so-called barrel shape) and choosing the correct surface finish and treatment. In modern engines almost all rings are treated with different types of coating (PVD, DLC, nitrided and so on) in order to reduce friction and wear. As soon as the ring starts to wear, the contact surface increases and so does friction.

On the liner there are also a couple of parameters we can influence to improve the contact area. From the very beginning, engineers have worked on the surface finish of the liners, and through various steps we have developed state-of-the-art honing procedures. Where at first the roughness was described simply as maximum/minimum values of the profile, we have since learned that the ratio between deep and shallow grooves can also be used to optimise the honing profile, leading to lower wear rates and lower friction and blow-by. Other, more apparent factors such as  bore cylindricity and liner straightness remain of equal importance – all very obvious parameters, but still difficult to maintain in all circumstances.

Of course this is a very simplified view of this topic, which describes only the main influences of the contact area between piston rings and liner on friction. And as long as we continue to use internal combustion (piston) engines, developments will continue in this area to further improve the efficiency of our engines, making it difficult for other engine concepts to come to prominence.

Written by Dieter van der Put

Therefore, during the concept design phase of the engine a lot of energy and focus is directed at keeping it as short as possible. A number of considerations on this issue have been discussed in earlier articles under this RET-Monitor keyword, but this article will look more closely at what actually determines the length with rising maximum combustion pressures.

In general, two design areas determine length – the upper side of the engine housing the cylinders, and the lower side of the engine housing the crankshaft. Both have their own issues in relation to the engine performance, and with increasing performance (and peak combustion pressures) the balance between them has shifted.

Looking closer at the upper side of the engine, the length is determined mainly by the bore and the distance between the cylinders, including the wall thickness and the coolant jacket between adjacent cylinders. Depending on engine type, the ratio between bore and stroke determines the bore.

For naturally aspirated engines this is typically as large as possible, since large intake and exhaust valves can be used in order to get as much air as possible into the cylinders and therefore maximum power output. Given the peak combustion pressures of around 90 bar max, the load from the piston and con rod on the main bearings is rather limited, which means bearing width is not an issue in relation to cylinder distance, taking the crank web thickness into account as well.

Looking at engines with forced induction, this changes things a little. From a certain bore-to-stroke ratio, a given bore size is derived, and since we are forcing air into the cylinder, the valve diameter is less of an issue. However, because the peak compression values are far higher (at around 130-140 bar) for a spark-ignited turbo engine, the resulting load through the piston and con rod will require much wider bearing shells. These, combined with the higher strength crank webs, will lead to wider main bearing distances.

Next we have the compression ignition engines, where peak combustion pressures have been exceeding 200 bar for some time now. Here the combustion loads on the piston are at such a level that the con rod and crankshaft bearing shells need greater material properties and dimensions, in order to withstand the loads on them by the combustion. Apart from material upgrades and detailed shape modifications, the width of these bearings has indeed grown and, in combination with the crank web widths, require more engine length than the minimum required piston distances at the upper side of the engine do. So, with these engines the crankshaft length is actually dominating overall engine length these days.

Exactly where the balance lies needs to be determined for each individual engine, but when moving up in terms of combustion pressures – as in turbodiesel engines – it needs to be taken into account.

The cases described here show that with the combustion pressures these days, especially with the entrance of highly loaded diesel engines into the racing scene, there will be no let-up in striving for higher efficiencies, and will continue for as long as internal combustion engines are raced. It is only when the new ‘green’ fuel initiatives fully take over or when we will run out of oil that this will probably no longer be a problem, but until that day... .

Written by Dieter van der Put

Apart from these two processes, you might have heard of the plastic engine, and although this is not a race engine, the materials and processes used are so unique that I did not want to overlook this as a production method. But this article will cover another manufacturing process which, although not unique is certainly not common in crankcase design and production. It is welding.

For me there is one simple question here, and concerns the issues as to why an engine fabricated in this way would not be competitive – what are the strengths and weaknesses?

First the strengths. In general an engine design is pretty straightforward, and many of the engine features – cylinder liners, crankshaft main bearings and so on – can be oriented in the three common axes, x, y and z. This lends itself pretty well to the fabrication process, since all the construction elements can be fixed before welding starts. Also, the machining could be reduced or at least kept to a minimum compared to a cast engine block, since for example oil drillings could be made from tubes.

Next the weaknesses. Welding several separate pieces into a single engine block requires a significant amount of time, and there’s the likelihood of introducing stresses into the piece. Welding will introduce heat-influenced zones into the part that might lead to crack initiation and distortion. With the proper approach this could be reduced to a minimum, although the question remains as to whether this will be enough. Given the fact that welds are very difficult to be controlled fully, the risk for leakages will remain.

But perhaps the most obvious disadvantage is that, to provide the best and most consistent welding quality, the various wall thicknesses should be as homogeneous as possible. Add to that the principle that the weld itself represents a rather small radius between the welded parts, which would lead to high loads in the weld. Combine that with the fact that the heat-influenced zone, also just alongside the welded area, has poorer mechanical properties than the base material, which would lead to an increase in wall thickness to withstand the loads. And as we all know, wall thickness is mass, and mass is not wanted in our race engines.

Despite all this I still believe that this production technique should not be forgotten, although a number of disadvantages need to be taken into account. Or, and this would be even better, solutions/improvements should be developed in order to overcome these disadvantages.

In writing this article my aim has been to generate some response or feedback from readers who could provide me with some more insight or examples of engines that have used this manufacturing technique. To my mind the opportunities are obvious, and simplicity deserves a chance. Although the fabrication of a welded engine would be significant in working hours, nevertheless, in relation to the creation of moulds and casting tools to create an engine block casting it would still be cost-effective, albeit based on small volumes or one-offs.

I’d be interested to know your opinion on this topic.

Written by Dieter van der Put

I say ‘unfortunately’ because every gasket or seal is a weak spot in the engine. Seals are always made from a material with lower mechanical properties then the base block or head material, and the two mating surfaces always carry some risk of defects (distortions, machining errors and so on), which can lead to leakages. On the other hand, it does not really make sense to talk about something we cannot prevent, so let us see which kinds of seals we could choose from. In this article we will look specifically at sealing rotating parts.

In an engine and gearbox there are a lot of rotating parts, from lower to very high rpm. Many of these do not need sealing since they are fully encapsulated in the housing. Some though, for example the crankshaft, require a connection to the outside world, in this case to transfer torque, so at the point where the axle goes through the housing, some sort of sealing ring is required. Typically a seal consists of some kind of sealing lip as the actual seal, and a carrier such as a metal ring that can be pressed into a machined chamber, or a flange that can be bolted to the flat block surface. In order to function properly, the sealing surface needs to be controlled very tightly regarding diameter and roughness. In motorsport it is mostly PTFE that are used, owing to their tolerance of high temperatures, chemical resistance and relatively low friction.

With the increasing revs of modern race engines, however, PTFE seals no longer have the best sealing capabilities, mainly because of the rotational speed and crankshaft vibrations, which can lead to leakages between the seal and the shaft itself. In addition, the frictional losses rise exponentially with rpm, and lip-type seals require a certain level of pre-tension (albeit very low) leading to initial friction.

This is one reason why carbon-face seals were introduced to seal crankshafts in housings that run in the higher rpm ranges. Before they were brought in for race engines they were already in use in turbochargers, for sealing turbine shafts running at very high revs.

A carbon-face seal exits from a retainer that incorporates a carbon ring and a feature (spring, O-ring, bellows) that presses the carbon ring outwards, sealing against the sealing surface of the rotating part, in this case the crankshaft flange. Carbon-face seals can withstand far higher speeds than sealing lips and show lower friction. But given their design principle, in general being nothing more than a carbon ring sealing to a rotating crankshaft surface, these seals are more sensitive to angular misalignments, something against which the lip of a traditional sealing ring is very robust.

The carbon-face seal is known to show a higher degree of wear earlier then a lip seal, however, but in a race engine this is not a major disadvantage, since the running hours are not very high compared with other applications for them. And thanks to their production techniques, materials and surface finish are of such a high quality level that surface wear can be minimised.

Written by Dieter van der Put

Where the influence on the cooling system can be seen is on the outside of the liner or sleeve. We are talking wet liner design here, where the coolant is in direct contact with the liner. The interaction between cooling system and dry liner, where the liner is pressed into the block material, is less significant.

So what can we see when looking at a given wet liner’s outer shape? The liner’s outside diameter does not need to be a clean cylindrical shape; it can be specified in numerous different ways to be of influence on the cooling capability. Let us discuss some of these in more detail.

Outer diameter

The outer diameter of the liner can be used to increase or decrease coolant flow velocity around the liner by increasing or decreasing the outer diameter of the liner. Increasing the diameter makes the gap between block and liner smaller, which increases the velocity, and vice versa of course.

This also gives us the freedom to specify the outer diameter in different sizes at different liner heights, which will influence the coolant flow velocity in the coolant jacket around the cylinder. Of course, it is not simply a case of modifying this gap, the coolant system as a whole needs to take account of the overall pressure drop across the coolant jacket. Also, the location of the coolant feed and return ports need to be chosen in such a way that they provide sufficient flow and pressure levels. 

Dynamic seal

One of the more significant cooling optimisations with a liner’s outer diameter is the creation of a so-called dynamic seal, which is when, at a given height of the liner, the liner diameter is chosen such that it almost touches the wall of the block. In a sense this gives two coolant chambers – one above this very small local gap and one below.

This can be very useful for increasing velocity near the top of the liner, cooling the combustion as much as possible. In the lower region of the coolant jacket (lower piston positions) the flow does not have to be so high.

Of course, this type of design will not work with a single simple coolant feed port, since the coolant flow would take the path of lowest pressure drop, leading to almost no flow in the upper, high velocity, region. To enable coolant flow in that region, a supply from the highest pressure (just after the coolant pump) is required, and the return should be connected near the lowest coolant pressure (as close to the pump suction side as possible). Naturally, coolant temperature also remains a very important factor, it needing to be as cool (relatively speaking) as possible. 

Local geometries on the liner outer surface

More often one can see local geometries on the liner outside surface, which are specified in order to guide of steer the coolant flow and doing so, influencing the cooling capacity by increasing flow and/or pressure locally. In issue 72 (June/July 2013) of Race Engine Technology magazine there are some very good visuals of how this could look.

One interesting idea is where the top flange provides openings (often drillings) where the coolant can pass to the top deck and into the cylinder head. This is mostly used with dry-to-wet liner transitions, still enabling coolant flow from block to head, through the liner flange. Another advantage is that the coolant flow can be optimised by modifying the dimensions and location of these ports. 

Material

Although this article has looked specifically at geometrical possibilities, one should of course always link geometry to material. Given a certain combustion load (maximum firing pressure) the liner must be able to withstand these loads mechanically and thermally. There will always be a balance or optimum between strength and thermal conductivity. For additional information on the materials though, I recommend that you read the article on Liners in the above issue of RET.

As I mentioned at the start, liners did not at first really have any other role than to guide the piston up and down, and keep the coolant on the outside and the gases on the inside. Since then, however, the liner has taken on a more active role in the overall coolant system of race engines, just by being not too simple and not too perfectly round on the outside. Cool, isn't it? 

Fig. 1 - ‘Open deck’ configuration with external features for heat transfer 

Written by Dieter van der Put

To be honest it took me a lot of time before I had any insights into crankcase design. The main reason was the fact that every design feature in a crankcase has multiple reasons for being the way it is – in fact, engineering a block is rather uncomplicated, as long as you know where to start. And of course you need a group of enthusiastic people working with you on all the different aspects, such as casting, machining and so on.

This article will look at how a structural approach can lead to a block structure required for the specific application intended for the engine. Typically the bore and stroke of an engine, as well as the type of engine (inline engine, vee engine and so on) is already known when the task begins, so for this article let’s assume it is.

The engineer will start with the CAD system by feeding the basic shapes into the computer, mostly related to the crankshaft axis and the cylinder centre lines. After this the cylinder distance will need to be chosen, in order to determine overall block length. Typically the major factor is the type of cylinder cooling – depending on casting wall thickness and/or cylinder wall thickness, the minimum distance can be chosen.

The only real major issue left is how the drive from crankshaft to camshaft is chosen. For high-speed or highly loaded engines, the use of a gear train is most common, and the gear width will be determined according to loading. Adding all these dimensions together determines the overall block length. The height of the block, measured from crankshaft centre line as the reference, depends only on stroke, con rod length and compression height of the piston. These values together determine the top deck of the engine; for vee engines the slant height also needs to be put into the equation.

Another major decision issue is the way the various loads are transferred through the engine structure. An example is the number of main bearing and cylinder head bolts, their locations and how they are connected. Drawing lines between these main engine bolts (from main bearing bolt to corresponding cylinder head bolt) and one can see how the structure is designed into the engine. Typically one would see thicker portions (cylindrical cross-section) between both bolt types – the so-called main bearing wall structure.

At this very moment the crankcase can be considered complete from a conceptual point of view. Rather simple isn’t it?

This reflects the engine as it functions on its own, however, not within the integrated design of it and the vehicle. In most passenger car applications the engine is mounted using mounts in the chassis, decoupling it in order to avoid vibrations for the driver, thereby increasing comfort. In the high-performance world though comfort is less of an issue, and often the engine is mounted as a structural part of the vehicle. Integration of functions leads to less overall mass, which is the goal.

When thinking about engine design and how these mechanical loads can be engineered for, taking a good look will reveal how the design has to account for these loads and what the load paths are into the crankcase. Often ribs or other strengthened areas can be seen where the pick-up points to the chassis are located. The engine block itself and its standard wall thickness are normally sufficient to withstand these structural loads coming out of the vehicle. 

By the way, the total amount of crankcase area where really minimal wall thickness can be achieved is not very high. The high degree of integrated functionality (pump housings, support structures and so on) means many areas have thicker walls; only around the coolant jackets can some areas be found where the wall thickness is minimal.

So the engine structure is not really anything more than a structurally designed engine, and taking the logical steps in the correct order will automatically lead to the correct structure. Therefore, designing an engine block is not too difficult. One might ask: what is the most difficult design then? The answer is, all the easy stuff like pipes, lines and so on. Where a block gets priority, it is these smaller items that are typically not paid structural attention.

Fig. 1 - Overview of the structural features of an engine block/crankcase

Fig. 2 - View of the main bearing wall and structure between main bearing and cylinder head bolts

Written by Dieter van der Put

Designers of complex castings are familiar with internal stresses in engine blocks and heads. The reason for these stresses is the fact that the solidification of the material is not homogeneous throughout the part, leading to internal differences in the material where one area is ‘pulling’ on others. Normally this is not really a risk, since the stresses will remain below what is permissible in the material being used.

Sometimes the internal stresses can be observed when a component is sawn through in order to check its internal dimensional stability. As it is being sawn, the separate parts sometimes seem to open, or the saw becomes stuck, typically when the component tends to close the sawing gap. But in general the cast component does not show any signs of internal stresses, and will be shipped to the machinist. And it is not until there that the firsts signs of stress can be observed.

The reason why errors relating to internal stresses often emerge at the machinist stage is quite simple. The internal stresses in the casting will not cause issues as long as they remain below the critical level. The fact that the unit of stress is Newtons per square millimetre – that is, force divided by cross-sectional area – means that when either the stress is increased or the area is decreased, issues can be expected. The machinist normally reduces the cross-sectional area in the component by machining away material, and this can cause the stresses to become too high – especially when the machined feature has sharp edges and therefore creates a ‘notch’ effect, which quickly leads to defects due to cracks.

A known sensitive area in an engine block is the region between two cylinders, particularly when using wet liners. I have often seen cracked blocks in this region, even when they were not loaded, just machined. When you discover such a crack for the first time, the casting company is the first to be asked questions. Typically the initial analysis from its side does not show any defects, at least not until the first casting simulations are examined.

But it is the design engineer to whom the first questions should be put when something like this occurs, as it is he or she who is responsible in the first place for coming up with a design that is insensitive to high internal stresses. And even if it is difficult sometimes to design the component in such a way that risks are avoided, there is always scope for improvement, even if it takes a number of design iterations. In most cases, the foundry process or the machinist can minimise the risk, but removing it completely cannot be done during the fabrication process.

A second structural part where internal stresses regularly exist is the cylinder head. The design typically has a thick flame plate (which is the surface facing the block) and a robust upper surface, combined with quite thin inner geometries (such as intake and exhaust ports, for example). And with these ports often being siamesed, stiffness transitions regularly come into play a well. And even when the internal stresses can be kept low, the combination of these as well as thermal and mechanical stresses will lead to a very highly loaded part.

If you can keep the internal stresses under control, or even just influence them so that compression can be achieved in highly loaded area, then the component can be made even more robust. In the case of the cylinder head though, while understanding the problem is one thing, being able to influence it is something else entirely.

I expect that on short notice we will be able to predict and influence complex castings by controlling internal stresses. When we eventually reach this stage then nobody will bother with the topic of internal stresses any more, once new and highly sophisticated simulation tools become available to the few people who have access to them and really know how to use and understand the results generated with them.

Fig. 1 - Crack as a result of internal stress. Crack seen after machining of the O-ring feature

Written by Dieter van der Put

Since the O-ring was first developed, a long time ago now, enormous strides have been made in its engineering. Ongoing material development has increased its range of applications, while better production toolings and processes have improved its dimensional stability and mechanical robustness.

An O-ring is one of the most robust sealing concepts, but is rather limited from an application point of view. Most O-rings are used for radial sealing of cylindrical geometries, so to make use of the sealing characteristics of the O-ring in more complex geometries, so-called form seals were developed. We will look closer at these seals, to see why they can so often be found in modern engine designs.

The major advantage with using elastomeric form seals is their wide variety of applications – the temperature range they need to function in, the ability to withstand different kinds of fluids and gases and the freedom of form and shape are the most apparent ones. The term ‘form seal’ can also point to different kinds of sealing geometries:

• O-ring type seals without an O-ring cross-section. Examples include D-shaped seals for a greater compression width, and X-shaped seals such as those used in some motorcycle chain sets.
• Shaped seals with an O-shaped cross-section. These were known to be used in, for example, earlier types of valve covers for air cleaner housings.
• Shaped seals without O-shaped cross-sections. These are the most common types in modern engines. Typical applications for these seals are in oil and fuel filtration modules.

The degree of complexity in contemporary engine design is growing all the time. One reason for this is that the vehicle layout, and therefore engine installation and integration, needs to be packaged in a smaller engine bay space. With the fast-growing use of hybrid systems within the same envelope, this means less room for what was already there – the engine. This in turn has led to a higher level of integration of engine functions and components, and several engine parts are being engineered together, such as coolant, fuel and oil pumps and their respective filters.

Filtration itself has also become more complex. Where originally we had only mainstream filters, now we see more and more spinner filters for even higher filtration efficiencies. And these additional components need packaging and connections to the engine block somewhere on the engine.

With so many connections of different fluids and gases, the engineering of the sealing types for these connections is more difficult than before. Increased sizes and sealing areas of components have highlighted the limitations of the flat seal types, which need closely spaced bolts and a high-tolerance flatness for both mating components. These demands have proved very difficult to meet, although elastomeric form seals are a significant step forward.

Where flat seals need an absolutely flat surface, the form seal is assembled in a tooled groove, meaning a very stable tolerance can be achieved. Groove depth can be measured before final machining of the corresponding flange – a best-fit machining operation – leading to a very tight tolerance over the total sealing area.

Another design advantage with elastomeric form seals is the ability to create topography (local differences in shape), enabling more freedom in critical sealing areas. This means slight adaptations can be made to the seal height, enabling small differences in seal compression where required. Anything that can be adapted dimensionally in the production tool can be used to increase sealing capability.

Apart from this, the risk of damaging the groove is far less than damaging a sealing surface for a flat gasket, leading to lower scrap during production. Also, the chances of damaging the seal itself are significantly lower with form seals. Since they are pre-assembled in a groove of the ‘to be sealed’ component, again the chance of damage is minimal. And we’ve all known flat seals that have refused to stay where we wanted them to during assembly, and very often they have had to be kept in place using some kind of liquid seal or even glue.

So, with the increasing complexity of engineering solutions, the O-ring seal has evolved into elastomeric form seals of larger area, and are now able to seal larger and more complex geometries – essential in the modern high-performance automotive and racing industries.

Fig. 1 - Example of a large and complex sealing geometry using an elastomeric form seal 

Written by Dieter van der Put

In principle, cylinder liner design does not look too difficult. It is more or less a machined cylindrical component, often with a flange somewhere of somewhat bigger diameter to keep the liner at its correct height. But on closer inspection there are a lot of things that might lead to distortions of either the liner itself or the surrounding crankcase structure. Most of the time when this occurs, an engine will be totally lost, so the liner is a critical component in the engine.

Starting with its function, a cylinder liner guides the piston in its path going up and down, and also plays a role as part of the combustion chamber, keeping combustion inside and coolant and oil outside. Seems rather simple, doesn’t it?

In doing so, however, the liner will interact with a number of other components, each of which has specific interface requirements. Let us start with the obvious: sealing the combustion chamber between the cylinder head and the crankcase. In this particular example the liner, a top-stop type, needs to seal at two locations – the first between the liner’s upper side and the cylinder head, and the second between liner flange and crankcase. And where the first, almost without exception, uses a gasket between both parts, the latter almost never does.

In general, the liner flange and head gasket are clamped together between cylinder head and crankcase by the cylinder head bolts. These bolts introduce significant forces , which can be influenced by following design parameters:

• Number of bolts. The more bolts that can be used in the design, the lower the load per bolt will act on the liner, leading to fewer distortions over the circumference of the liner flange.
• Flange protrusion. The more the liner flange protrudes above the crankcase top surface, the more bending each cylinder of the cylinder head suffers from, leading to high local loads on cylinder head and liner.
• The diameter on which the head gasket seals to the liner top surface. This is one of the most critical design parameters. Most head gaskets have a region called the compression diameter, where the gasket has its sealing function; often this is a specific shape in the gasket. When this diameter is equal or smaller than the smallest balcony diameter of the liner flange, a bending force will occur on the radius below the flange – one of the most sensitive areas of the liner. Since the compression diameter is kept as small as possible, because combustion pressure acting on this diameter might otherwise be too high, there will always be a minimal difference in diameter.
• The radius below the liner flange. Where a large radius is definitely favoured, this will automatically lead to a larger flange diameter, since the mating face between liner and crankcase needs to be of a certain minimal area, keeping contact pressures within their mechanical limits. Smaller radii at the flange area could lead to liner cracks just below the flange, in exactly that same radius.
• Flange radius surface treatment. A surface treatment is often added to the radius of the flange. With shot-peening, relatively small objects – often ball shapes – are impacted on the surface, where they cause plastic deformation of the surface, albeit minimally, leading to compression stresses in the base material. This in itself increases the basic strength of the material locally, leading to improved mechanical properties. Caution should be exercised over the peening materials and pressures, however, since overdoing it might also lead to earlier failure of the component.

All in all, liner design has a lot of simple and somewhat less simple areas of interest. Some areas are straightforward, while others require more focus and depth of knowledge. Since the liner, as a separate part in the engine, is still regularly used, engineers will continue to try to improve them, making them lighter but without reducing their mechanical properties too much.

This is what makes engine design so interesting – a rather simple component with a lot of very interesting and critical engineering areas.

Fig. 1 - Section view of a liner design with the significant areas of interest

Written by Dieter van der Put

Looking at a broad range of engines, not just high-performance race engines, one sees a wide variety in the number of cylinder head bolts in the different designs. Look a little further the reasons for this are obvious. Typically, the number of bolts is a consequence of some other, higher-priority decisions.

Let us start from the beginning. The reason cylinder bolts exist is purely because the cylinder head and block need to remain together under combustion load. Apart from that they have no added value whatsoever to a running engine. Given the size of these bolts in relation to the bore, one can see that significant amounts of material needs to be reserved for them. That means these areas cannot be used for other functions of the cylinder head, such as the intake and exhaust ports, coolant cavities, injector and spark plug bores. Also, the interfaces between head and block need to be located somewhere outside the bolt areas.

Historically, the number of cylinder head bolts used in engine design has varied between four and eight per cylinder, although of course adjacent cylinders make use of the bolts in between. So, when one calculates the number of bolts, for example to calculate bolt loading, one should correct for this and reduce by two bolts (four halves) per cylinder. To achieve optimum roundness of the cylinder bore, the advantage would go towards using more, smaller diameter bolts. The overall clamping load in this case will be divided more evenly over the circumference, leading to a lower deformation of the bore itself, enabling improved - as in more homogenous - contact between piston rings and liner, meaning less friction and an increase of effective power.

Bolt size and material specification are chosen according to bore size and maximum cylinder peak combustion pressures. The intake and exhaust ports are often the largest features that need to be incorporated in between the cylinder head bolts. Engine engineers have two choices - either one bolt in between the two intake ports, or a bolt on either side of both ports.

When the engine is a derivative of a production unit, all bolts need to be accessible from the top. The background here is of course mass production assembly, where process time is of major importance. Some engines, specifically designed for racing purposes, are known for making use of cylinder head bolts entering the cylinder head from below, mostly the bolt located nearest to the intake ports. In this case the intake ports can be designed without taking this bolt into account, leading to a more optimum port shape. However, with this type of bolt location, accessibility to these bolts will be unfavourable, to say the least.

On the exhaust side the engineer faces a similar challenge, although the extreme heat loading of the exhaust gases makes life even more difficult here. It’s impossible to engineer a bolt in between the exhaust ports, since the material around it cannot be cooled sufficiently. The limited area available does not allow coolant to get near this bolt, and this can lead to temperature cracks and other thermal-related issues.

The last possible interference with cylinder head bolts are fuel drillings, in the case of in-cylinder injection, either direct gasoline(-like) injection or diesel injectors. Mostly with diesel-powered engines, a (high pressure) fuel line drilling needs to be incorporated into the cylinder head. Where an intake or exhaust port could be designed to be slightly curved, a drilling requires straight-line access to the injector.

In summary, fewer cylinder head bolts give more design freedom in relation to the other features inside the cylinder head. However, more bolts will provide lower distortion of the bore, leading to other, more friction-related advantages.

And so this brings us to the engineering conclusion we have shared many times already: engine design does not know ideals, only well thought-out compromises.

Fig. 1 - Top view of the cylinder head of a truck engine: plenty of features in a relatively small area

Written by Dieter van der Put

As we all know, an engine cannot be made seal-less, although that would make our engineering lives a lot easier. And I'm sure we can agree that sealing the combustion chamber, including the surrounding passages - coolant jacket, oil drillings, oil return cavities and so on - is one of the most complex tasks in high-performance engine design. Because of the efforts being spent in this area, we have learned to understand the specifics of head gaskets. In addition to the head gasket though, there are many other locations in and on our engines that need sealing, such as intake and exhaust, coolant, oil and fuel system interfaces.

Looking at the installation of the engine in the vehicle, quick connectors are often used in order to enable easy exchange of the engine. These connectors are designed in such a way that, as soon as they are opened, the individual connector ends shut themselves, enabling clean disassembly. All further connections that do not require opening or closing during engine installation can be separated roughly into two groups - flat gaskets and compressible seals, mostly elastomers (O-rings, for example). Flat gaskets are considered to be one of the simpler sealing concepts, but when there is a leak it is often with these rather simple seals that they occur.

A flat gasket is meant to seal two components where flat mating surfaces are being brought together. Remember, a gasket is only required when one or both surfaces shows some irregularity, such as scratches or pitting for example, whereas perfectly flat surfaces can easily be sealed without the need for a gasket. Reality, however, has shown that such a perfect sealing surface cannot be achieved easily. So, when designing the engine, the design engineers need to make sure the gaskets are taken into account, both in terms of their material spec and their dimensions. Flat gaskets can be made from a broad range of sealing materials, depending on the sealing medium and (non-) operating conditions, and can be made in various thicknesses.

Flat gaskets are typically found on connections such as those between coolant pump and engine block (at the least complex end of the spectrum) and between exhaust manifold and turbocharger(s), the latter being one of the more complex gaskets, given the severe environment in which they need to function. The cylinder head gasket is a third example of a flat gasket, in this particular case a gasket typically sealing more than one medium.

Apart from the gasket itself, in order to establish a proper sealing function, the relevant components need to provide sufficient stiffness to achieve an even contact pressure over the gasket. Looking at the cylinder head gasket again, the sealing of the combustion chamber is achieved between the very stiff cylinder liner and cylinder head structure. The areas outside this region have far lower stiffness due to internal differences in stiffness near the coolant and oil cavities. Since it is very difficult to achieve a robust seal in these areas, elastomer seals or local shape adaptations (beads/pressings) are often integrated into the head gasket.

Another well-known critical area where flat seals are used is in the exhaust system. Turbocharger flanges in particular (with their non-ideal rectangular shape and non-uniform bolt distances) are known for their multi-layer steel gaskets with integrated beads. These beads, which can be single or double beads, are designed to achieve and maintain sufficient local pressure between gasket and components in these kinds of harsh environments. Specific attention needs to be paid to high stiffness gradients over the sealing interface, which might lead not only to a leaking seal but to temperature-related cracks in the surrounding structure. A wise lesson I've learned over many years is that it is not maximum stiffness that is required, but useful stiffness. This means one should try to achieve sufficient stiffness with moderate stiffness gradients over the total sealing structure, including the flat gasket - and, absolutely not to forget, proper specification bolts, maintaining pre-tension when cycling between maximum and minimum temperatures.

In summary, sealing interfaces come in two flavours - complex and less complex. Simple ones do not exist.

Fig. 1 - Gasket detail

Fig. 2 - Gasket pressure graph

Written by Dieter van der Put

Let us take a closer look at some of those decision areas:
• Sealing ring position/coolant jacket length
• Sealing ring geometry
• Location of the sealing rings
• Liner material
• Radial location of the liner

The first three all concern the actual shape, location and orientation of the non-combustion side sealing rings (assuming formed engineering plastics sealing rings are used, in whatever section). In general, the coolant will be located as close to the combustion gases as possible, in order to achieve maximum cooling at the desired location. This dictates the required top level of the coolant jacket. The lower level is typically based on the top level minus the piston stroke. The goal is to have coolant in proximity to the piston rings throughout the total travel of the piston, although the lower level is taken somewhat more flexible than the top level. When the block structure prevents having a full coolant jacket, the accepted compromise is to have piston rings below the coolant jacket. Since the heat transfer is much higher at the high temperature area this is an acceptable situation.

As regards sealing ring geometry, some people choose simple O-rings while others opt for non-symmetric sectioned sealing rings or even X-type section rings. The choice remains a personal thing, I guess.

The location of the sealing rings, in the block or on the liner, can be more significant for the engine design. With the sealing rings in the block, proper assembly of the rings is more difficult than on the liner, but because the base of the liner can be smaller (in diameter) then the top flange can be kept smaller, leading to a more compact block.

Another significant parameter is the material from which the liner is made, the principal choices being steel or cast iron. The choice will be based mainly on maximising heat transfer and geometrical stability. In order to keep the engines as compact as possible, thin walls are preferable, for which a material with a high E modulus, like steel, will be chosen. But in order to maximise cooling, a material with a high thermal conductivity is preferable. As usual, not many materials combine the both of best worlds, so a balance will need to be found between strength/wall thickness and thermal behaviour.

One of the final significant parameters in determining the liner design is the decision about where the liner will be positioned radially. A liner needs a radial fit, and if it hasn't been engineered carefully, this can result in excessive oil consumption or even major engine damage.

At the top end there are in principle two possibilities for radially supporting the liner - either on the circumference of the flange or just below it. When choosing a radial fit just below the flange, one must make sure that the fit does not cause deformation of the liner, which would lead to contact issues with the piston rings. The upper flange solution is less risky, since typically the flange height is lower than the distance between the top piston ring and the piston crown. In this case, the piston rings will not get into the area where the fit could lead to deformations.

When choosing between a bottom or middle radial fit, there is still the risk of piston skirt scuffing or increased oil consumption. It has proven to be very difficult to maintain liner cylindricity in such situations. All possible engineering measures to increase robustness in this area will lead to a reduction of overall efficiency. One possible solution - increased ring pressure for the piston rings to be able to follow the liner shape - will lead to higher friction losses, while another option, increased piston skirt clearance, will lead to more piston slap and increased wear rate of the piston system.

The design areas covered here can all be simulated pretty well using FEA, and will lead to well-founded decisions regarding the liner concept. Since most production engine blocks these days feature linerless concepts with coated bores, the number of engineers needing to make decisions on wet-liner concepts will be limited. However, having some knowledge about these kinds of basic engine design will lead to increased understanding of the pros and cons of the liner concepts and their influence on overall engine design.

Fig. 1 - Top-stop liner concept, including lower sealing ring

Written by Dieter van der Put

In deciding on the best liner design for an engine, a number of boundary conditions need to be evaluated. First, a conceptual decision needs to be made between whether the engine will use parent bore (bore in the base block or cylinder material), dry liner (liner assembled into base block or cylinder, without direct contact between coolant and liner) or wet liner (liner assembled into base block or cylinder, with direct contact between coolant and liner).

When a race engine has been derived from a production engine, this decision has already been made for us. While production units will sometimes be re-machined to convert to wet liner, this is not done very often, so I'll assume here that the decision has been made to opt for the wet liner concept. But one question remains: top or bottom?

When evaluating the concept of a wet liner design, a number of layout considerations need to be into account. For example, the chosen liner concept can determine the length (and therefore mass) of the engine. In order to understand this better, let us first look at the three major liner concepts - top, mid and bottom stop. It's assumed that the cylinder head sealing surface is called the top end of the engine.

The top-stop liner concept is so called because it has a flange on the top of the liner with which it is located into the cylinder block. The mid-stop has a similar flange at or near the middle of the liner, and the bottom stop has its locating flange near the lower end of the liner.

First, let us take a closer look at the main differences between these three concepts. In principle, the assembly process requires the top to have a larger diameter than lower down, so for a given liner wall thickness (depending on peak firing pressures) and liner locating surface width, the top-stop design requires the least space. This means the overall length is influenced by the chosen liner concept. Every additional millimetre means additional engine mass, so the liner concept is seen as an important decision parameter in engine design.

Apart from geometrical effects, engine cooling needs to be considered, in this case the area where liner cooling is required most - the top end of the liner, where combustion takes place and temperatures are highest. To get the coolant as near as possible to the combustion chamber, the wall thickness needs to be as small as possible, which would favour the mid- or bottom-stop concept.

A third important point is deformation of the liner over its length, in relation to the oil consumption and friction losses of the piston system. The lower the deformation, the lower the piston ring pre-tension that can be chosen (lowest friction losses). To have as little deformation as possible coming from the assembly forces of the cylinder liner onto the liner (by pre-tightening the cylinder head bolts) the top-stop concept is favourable.

With this design, the liner is not being pre-tensioned at all and will therefore not deform due to the loading of the clamping pressure. However, due to the proximity of the cylinder head bolts to the liner flange in this design concept, one should be careful not to distort the roundness of the cylinder bore. With a liner concept with no direct contact between the top surface of the block and the liner flange, this radial distortion will be limited.

Given the points above, it's easy to see that deciding on the liner design concept requires some prior work in order to balance the pros and cons of the several options.

After the conceptual decision has been taken, there are a number of secondary decisions, such as:

" Sealing ring position or coolant jacket length - this being in relation to coolant system volume

" Sealing ring geometry - simple O-ring of formed sealing ring; related to ease of assembly of the rings and assembly of the liner into the engine block

" Location of the sealing rings - engine block-sided or liner-sided; related to overall required space and assembly of the liner

" Sealing of the top end of the liner, in the case of the top-stop sealing concept - radial sealing of the flange or liner portion below the flange, axial sealing below the flange, and of course

" Radial location of the liner - radial on the flange, or directly below the flange

And apart from these geometrical items, the liner's material is also open to some more thinking in order to decide on the best compromise.

Covering these secondary items is beyond the scope of this article, so I'll focus on them next month, in order to provide the whole spectrum of decision-making on cylinder liner concepts.

As can be seen, there is no single best cylinder liner design concept, it all depends on which of the parameters are given which priority. And, of course, whether you can start from a clean sheet of paper or a production engine will be used as a basis. As mentioned many times by many engine designers, the engine, in all its complexity, will always be a compromise; in this case - top or bottom?

Fig. 1 - Top-stop liner concept, including lower sealing ring

Fig. 2 - Detail of machined engine block for top-stop liner

Fig. 3 - Detail (bottom) view of machined liner interface in engine block

Written by Dieter van der Put

The engine that is the focus in this article has already been described to an extent in the article referenced above, the self-engineered hybrid 1300 cc by John Ellwood from Sweden. Although the swept volume of 1300 cc does not comply with the rule book, the maximum permitted being 800 cc single-cylinder engines, the Nordic Supermono organisation has granted it permission to race in the Nordic Supermono Racing series next year. This is good news but at the same time it means a lot hard work to get the engine ready for the first race, which is in less than a year from now - remember, the engine is still at the early manufacturing stage.

What is it that is so interesting on this engine? Apart from the fact that it can be seen as a combination of a two-stroke and four-stroke engine - pumping its crankcase volume, as a two-stroke engine does, into the combustion chamber - I believe the reason is more to do with the story behind its creation, as it is a perfect example of persistence. When an average man without any background in engineering can create an unusual engine (the existing 500 cc hybrid) and then has the idea to make a somewhat bigger version, no less than 1300 cc, and then believes fully that this engine will work, you could call that madness. And that might prove to be the case when the bike fires up for the first time - but still, what an adventure.

Let's look at some engine data. With a bore of 118.5 mm and a stroke of 121.5 mm, the swept volume is a gigantic 1339 cc. On the earlier, 500 cc hybrid, the intake plenum pressure was measured at 6 psi, meaning it could almost be considered a 2.0 liter engine. With a compression ratio of 13-15:1, the health of the piston system itself will be the first challenge. For this engine a one-off piston is on order from a well-known supplier, to whom Ellwood has provided his thoughts on the piston, and which for its size alone will be an interesting project for the company.

In order to keep charge temperatures down as much as possible, water spray will be injected into the crankcase, together with the required two-stroke lubrication, and the charge will also need to pass the intercooler on its way to the combustion chamber. The trade-off between temperature and pressure loss will not be too difficult in this case; pressure loss will need to suffer in order to keep temperatures as low as possible. The reason why fuel is not injected early on as well is that the small-end bearing, which is nothing more than a bronze spec bush, can't survive it, so the fuel is being injected at the traditional location, just before the intake to the cylinder head.

Also, the cylinder head is a piece of craftsmanship, produced out of billet aluminium. Since the cylinder head intake and exhaust are of the rotary type, a two-piece head could be designed, giving easy access to machine the coolant jacket into the two cylinder head pieces. Sealing is achieved by a machined groove, combined with a fluid sealant, simple as possible. Sealing between cylinder head and cylinder liner is done by a flat copper ring.

In talking about this engine, Ellwood told me in all honesty that he had no background in engine design whatsoever, meaning that a couple of dimensions on this engine have room for improvement. For example, with a con rod length of 200 mm, the length-to-stroke ratio is a minimal 1.65:1, something that could have been chosen more optimally to a ratio of 2:1. Since the height - perhaps in this case it's better called length - of the engine is limited by the front wheel, future changes could include decreasing the stroke and increasing the con rod length. For a ratio of 2:1, this would lead to a swept volume of about 1150 cc, which is still massive for a single cylinder.

The engine has a horizontal cylinder, which will surely lead to interesting driveability, especially since the bike will not have a gearbox. The idea is that the massive torque of the engine, combined with a robust but sensitive clutch (which Ellwood is still searching for, by the way) should provide sufficient driveability between minimum and maximum speeds on the racetrack. Again this confidence comes from the earlier 500 cc bike, which Ellwood has also ridden without a gearbox. Quite a challenge.

Is it sane, to create an engine this different? Probably not, but sometimes the challenge of the attempt provides enough food for thought for further steps - perhaps not by the designer as such, but by others. It is ideas and initiatives like this that have enabled the engineering world to get where it is today, and I think Ellwood deserves the opportunity to ride his creation in the Nordic racing series next year. Until then, I wish him success with the challenge of getting there.

Fig. 1 - Right-hand side view of the racing motorcycle, horizontal cylinder and no gearbox, just the clutch for driveability
Fig. 2 - View of the cylinder head, using a rotary-type intake/exhaust valve
Fig. 3 - Left-hand side of the engine, belt-driven rotary valve

Written by Dieter van der Put

Just before the series announced the 800 cc restriction in swept volume, two brothers from the Netherlands decided that, on the basis that "bigger is better", a big race engine would have a significant advantage. A larger displacement would provide a significant gain in torque and therefore increased acceleration out of the corners, and this again would lead to a higher speed on the straights - in itself all very promising for getting as many wins during the season as possible.

Since all Supermono engines from that era were roughly 650 cc in displacement, the idea was to make an engine of around 900 cc. And since an existing Rotax crankshaft (stroke 81 mm) was chosen, the bore size was set at an enormous 120mm! In comparison to usual bore sizes of about 102-105mm this was a significant increase, which would later become a fragile area of the engine.

Most of the parts were made from scratch. The crankcases were horizontally split machined (from castings) parts with wall thicknesses up to 35 mm. The dimensions were chosen not only to withstand the combustion forces but also to enable the crankcase structure to be used as structural part in the motorcycle chassis.

Fig. 1 - Side view of the crankcase, horizontally split and wall thickness up to 35 mm

Later, when it became obvious that the chosen cranktrain components could not quite handle the brutal forces of the 120 mm piston going up and down 9000 times per minute, the choice was made to create these components from scratch as well. Where initially the con rod and crankshaft were borrowed from Rotax, now these components were made in-house, and in the end the con rod was made from aluminium. The only production parts used were gearbox shafts and gears (from a Suzuki GSX-R 750), coolant and oil pump, and some smaller bits and pieces elsewhere.

The cylinder liner, which started life thinking it would be fitted in a truck engine, was machined down to (press) fit into a made-from-solid aluminium cylinder, acting as a wet liner with a top stop. This way the wall thickness of the cast-iron liner could be kept to a minimum, without the risk of deforming under the clamping load of the cylinder head onto the liner (typically leading to higher oil consumption due to piston ring flattering).

Fig. 2 - Side view of the engine, with forward-pointing intake

Another very interesting piece of craftsmanship on this engine was the cylinder head. This component was oriented with the intake pointing to the front of the motorcycle, in order to create a 'logical' airflow through the engine. During its fabrication, coolant channels were machined into the structure that were sealed later on to create a cooling cavity. Sealing was done by a pressed-in tube at the location of the spark plug and welding on the outsides.

Fig. 3 - Bottom view of the cylinder head

Inside this purpose-made cylinder head, initially Opel passenger car valves were operated by camshafts from a Honda Bol D'Or. To be exact, it was a portion of this four-cylinder camshaft, as it needed to fit inside the cylinder head of this single-cylinder engine.

The engine has raced in a number of events and has been praised for its torque and ease of driving. On the other hand, the engine remained fragile, and although excellent lap times could be recorded, the engine had difficulties with finishing races.

In the battle to make this huge single-cylinder engine work as it should, lots of lessons have been learned. These were put into a second engine, which unfortunately never raced. However, after this engine, a number of others have been developed by the same people, among others a diesel race engine.

A number of years after this engine was built, the next generation has stepped in, and at the moment a young man who's in his final year of studying automotive engineering is building his own single-cylinder race engine. Again, a couple of parts came from production models but most of the structural components were made from scratch. It's an interesting study topic, and it might be a good idea to focus on this second-generation single-cylinder race engine in the next article.

Written by Dieter van der Put

At first the flatness of the cylinder head fire face was established using a granite surface plate as a reference and lever indicator, and was found to be both damage-free and within tolerance. Attention then turned to the cylinder liners. Upon arrival into the stores, each liner was checked for roundness and cylindricity, and the lower and upper faces of the locating flange checked for axial run-out. The lower face of the flange was then 'blued' and the liners placed into their respective positions inside the cylinder block (which had been previously checked for any burrs or machining errors), rotated and removed for inspection.

Once all was well, the liners were replaced, carefully noting their exact position and measuring the precise liner 'stand-off' from the top of the cylinder block. Satisfying himself that the gas-filled fire rings were of uniform size all round, the cylinder head was dropped onto the locating dowels, the bolts inserted and the tightening procedure commenced.

Working from the middle cylinders outwards, the bolts were progressively tightened in turn to minimise distortion then finally brought up to the design torque followed by turning a specific angle (often known as 'angle' tightening) to bring the bolt into yield. This was considered to be a more accurate way of establishing an accurate tension in the bolt (and hence clamp load) than using the torque value alone.

Having completed the assembly and sitting back to congratulate himself, one thing was immediately obvious - something the young man was amazed to see. This engine, you see, was only a concept unit intended for rig testing, with little attention having been paid to enclosing the bolts within the cylinder block, as is normally the case. So when the young man looked at the bolts he could see that they had actually twisted by a considerable amount.

It was at this point that a magical thing happened. For no apparent reason, and sitting in the middle of a spotlessly clean and air-conditioned room, he heard what could only be described as a 'pinging' noise, then another nine similar noises, reminiscent of a machine gun. Taking another close look he could see that the bolts had actually unwound themselves and were once again untwisted.

Now you could ask what all this has to do with sealing. The short answer is 'everything', for that pinging noise was indeed the bolts unwinding themselves in response to some slight disturbance and releasing the clamp load of the cylinder head against the fire ring seal. The culprit, it would appear, was the high levels of friction under the heads of the cylinder head bolts, which had prevented the full load to be transferred into the bolt, even though the threads had been carefully oiled before assembly.

The moral here is to ensure that the full load is applied through to the fastener and hence into the seal, and that the friction under the head of the fastener has to be minimised. The seal is part of a system. If it is not loaded correctly, how can you expect it to work?

Here endeth the lesson.

Fig. 1 - The twisted bolt

Written by John Coxon

When combustion loads permit, aluminium alloy is the material of choice for the cylinder block, since as in motorsport, weight is a critical parameter in the modern road-going passenger car. But unlike race engines, where ceramic coatings on aluminium bores are still preferred, road transport OEM products often prefer cast-iron liners.

For best durability and minimum costs, as well as other factors unrelated to motorsport (such as cold-start emissions), these may be cast on the factory floor during the block manufacturing process, and under normal use should give more than 150,000 miles of trouble-free motoring. But as in motorsport, OE engine manufacturers are downsizing and adding turbochargers, and with space limited in the engine bay, to make engines more compact the spacing between the cylinder bores (sometimes referred to as the bore pitch) is getting less, to the exclusion of the coolant, particularly towards the top of the liner.

One solution is to have what is called 'open deck' designs. Here the liners, located towards their lower end in the cylinder block casting, are effectively clamped against the cylinder head fire face and surrounded with coolant. While excellent for cooling, liners will shuffle under the high gas loadings of a race engine, to the detriment of the head gasket. The alternative is to have a 'closed deck' design, which anchors both the top and bottom of the liner, but doing so limits the cooling at the all too critical position at the 'bridge' between the bores.

In some designs, holes can be conveniently cross-drilled through the bridge area to improve the coolant flow across the block and keep inter-bore metal temperatures to acceptable levels. This is fine if you are machining only a handful of components but for OEM-type volumes, long drills and the risks associated with the drill bit 'wandering' during the process make such measures unpopular. And anyway, these small holes may be easily clogged or difficult to get close to the top deck.

Casting small slots between the bores is perhaps the only elegant solution, but doing so using sand cores is not really practical either. Replacing the cores with small oval glass tubes - something like 4 mm long by 1.3 mm wide but only 0.2 mm thick - is a technique that has been used for small-volume castings in the past but has now been perfected for volume manufacture by at least one OEM. Embedded in the sand cores, the glass is blown out using high-pressure compressed air once the casting has cooled.

Better coolant flow at lower coolant pressures, lower 'bridge' temperatures and reduced thermal gradients all sounds like a good place to start for the next generation of OEM-derived race engines.

Fig. 1 - Glass tube cross-section

Written by John Coxon

A few people though will not want to accept horsepower limits, and will choose to develop their own engine. Since this is definitely not the fastest and easiest route, we should give them our compliments now. As all of us with experience of engine development know all too well, the actual development begins with the first engine start-up, not before, which is why taking the initiative to build one's own engine deserves the compliment.

There are a some basics though that need to be understood before such an initiative can be realised. For example, one should know whether to use existing components or to make them yourself. In all honesty, I have not seen many amateur racing projects where all the parts were made from scratch. In most projects, some components are sourced from production units, and are typically components that did not fail in the first place.

In the example I want to use here - single-cylinder competition engines - the builder uses mostly production parts such as gearboxes and crankshafts, as well as smaller parts such as valves, valve springs and so on. Parts designed from scratch and produced using prototyping processes are crankcase housings, covers and cylinders. The remaining parts are often those based on standardised parts with engine-specific modifications, such as pistons where the piston head and valve pockets are often made to spec.

I don't want to give too many details about the engine concept (yet), but I can say that the engine is destined for a racing series in northern Europe. Normally the regulations for the single-cylinder racing series, the Supermono, permit engine capacities of up to 800cc. This engine, however, throws in an enormous swept volume of well over 1000 cc - from a single cylinder!

Where a rule of thumb says that any racing piston over 100 mm in diameter will lead to disaster, one can imagine how big this piston is. This of course also affects the crankcase design. As can be seen from the photos here, the design of the engine is quite basic.

The parts shown in the photos are machined castings, all produced by the designer/racer himself, with some advice from a number of friends and interested followers. To give an idea of size, the visible bolt heads are M8, which means the cooling ribs for example will not damage easily.

A closer look to the top end, both with and without the cylinder head installed, reveals a similar approach - keep it simple and use appropriate dimensions.

Some readers may be very interested to read more about this engine, but because the aim of this article is to provide an insight into the kind of initiatives involved in designing and making your own blocks and heads, more in-depth information will have to wait until a future article.

During my search among amateur racers and engineers, I have seen a number of fascinating engine projects, not just in terms of the different levels of design maturity but also the drive which these engineers have in realising their ideas. Creating these kinds of parts forces us to think about their specific functions in the engine, something which has kick-started careers as race engine engineers.

There is plenty of interesting information to share on privately built amateur race engines, so look out for future articles on the subject.

Fig. 1 - Side view of a self-made horizontal single-cylinder race engine. Cylinder and cylinder head cast and machined, crankcase machined from solid
Fig. 2 - Top view of cylinder head and support to the frame
Fig. 3 - Top view into the cylinder, showing the basic approach and shape of the self-made engine components

Written by Dieter van der Put

In this article I would like to show the inventiveness that small/low-level racing teams need for them to progress in their racing classes. Professional teams, with professional-level budgets, have many facilities and resources available for developing their vehicles. A major portion of these budgets is transferred into detailed knowledge, not only as regards the performance and handling but on the reliability of the used components. This is often a very difficult area for smaller constructors.

In the late 1980s and early '90s, a very interesting motorcycle racing class started to get more and more attention from true four-stroke enthusiasts around the world. This class, called Supermono, consisted of race bikes using single-cylinder, four-stroke race engines. Technical rules were almost unrestricted; the most limiting factor was that you could use only naturally aspirated internal combustion engines, so no pressurised charging was allowed. At the time, the only single-cylinder bikes on the market were enduro bikes and one or two street bikes, which were not designed to be used for serious road racing. This is also why the start of this racing class was dominated by major engine failures - broken pistons, cracked con rods and destroyed crankcases were just some of the repeated failures; not really the best way to remain friends with circuit directors.

At the beginning of Supermono, many bikes used Rotax engines, owing to their availability and reasonable cost. As power levels rose, these engines became known for their enormous power potential. The only downside was that they did not last very long with higher power outputs, and the failure that kept occurring at around 75 hp was a cracked crankcase (Fig. 1).

When something like this happens, the root cause of the failure needs to be found. Without access to sophisticated finite element models and software, basic engineering knowledge is often used to overcome these issues. Some people call this the trial-and-error method.

The crankcase material was aluminium, and because steel is stronger it was thought that this crack could be solved by clamping two metal plates on either side. Indeed, in the following race the crankcase did not crack between the bearings; unfortunately, this time the crack occurred just above the steel plates, the next weakest link (Fig. 2).

Normally, entering the second trial-and-error sequence, one thinks rather deeper about the real root cause and takes a more principles-based approach. Factors like tension, notch, compressions, load path, E-modulo and crack initiation are taken into account, all in order to find real solutions. Discussions revolve more around the function of the parts, rather than the parts themselves, and conceptual ideas are tailored to the specific design of the crankcase.

This resulted in a far more logical design modification. In order to increase overall strength in the load path direction, a steel insert was added into a machined hole in the crankcase. An extended cylinder mounting stud was assembled from the top and screwed into the insert, and an additional steel fastener was assembled from the bottom end of the crankcase, introducing compression into the structure. The most important improvement was that the total gas and inertia loads were withstood by this steel pillar (Fig. 3).

This modification meant that the 75 hp failure hurdle could be overcome, which allowed the Rotax engine to remain one of the most powerful engines in the field. A couple of years later the performance of newly developed race engines rose to more than 90 hp, and although this was of course a major achievement, it also was the beginning of the end of the Supermono class, as the level of spending needed to race at these performances with a certain reliability could not be met by the true enthusiasts.

Although small constructors and private racing enthusiasts do not have the facilities and/or budgets that professional (factory) racing teams have, there is always scope to improve. It might take a bit more time and a few more broken bits and pieces, but in the end solutions will be found. The real requirement, equal to all competitors, is engineering experience and logic thinking. Like the Britten V1000, one man's dream… .

Fig. 1 - Cracked crankcase of a Rotax 605 single-cylinder race engine
Fig. 2 - First trial-and-error solution, using two parallel plates on either side of the crankcase
Fig. 3 - Final solution with steel insert connecting the cylinder stud with a further bolt from below in the crankcase

Written by Dieter van der Put

In reviewing wear therefore it is perhaps necessary to distinguish between what might be termed normal and abnormal events. Abnormal wear might be considered the result of a foreign component, a piece of grit for instance, ingested into the cylinder and trapped between piston and bore, scoring the liner. Contact between piston ring top land and the bore would also, I maintain, be classified thus. Normal wear, however, the subject I intend (briefly) to discuss here, is inevitable and can only ever be minimised, never eliminated. The most obvious case for the piston ring and bore is the wear manifesting itself at the top ring reversal point.

For older readers, the 'step' in the bore at the highest point of the top ring travel will have been a familiar sight in some slightly higher mileage (usually gasoline) engines. Most noticeable on the thrust side of the cylinder, the combination of high combustion loads and lack of relative movement between piston ring and bore causes the 'wedge' of oil generated in the lubricant to break down, allowing metal-to-metal contact and subsequent wear.

It's much less of an issue these days, as improvements to cylinder surface finish and lubricant technology have mitigated the problem such that engines now can expect to reach 150,000 miles and more without any deterioration in performance. Boundary layer additives, such as esters, in early lubricant formulations gave way to more effective ZDDPs (zinc dithiophosphates), which for many years have provided a cheap and highly effective source of anti-wear chemistry. Working by way of polar attraction to metal surfaces, this is still the most popular anti-wear chemistry used in competition oils where engine exhaust catalysts are not used. In catalyst-equipped vehicles where limits to sulphur and phosphorus in engine oil exist, ZDDPs are steadily being replaced by other proprietary anti-wear chemistries.

Another form of cylinder bore wear not generally seen but which may be more apparent in future years is that of 'bore polish'. An issue more associated with highly pressure-charged diesel engines or engines where cylinder distortion is apparent, in the former, bore polish is a result of the combustion deposits becoming trapped around the piston ring assemblies and then acting as a mild abrasive, steadily wearing away the top layer of the cylinder surface. With each successive pass, the carefully crafted honing marks containing the lubricant are progressively 'machined' away to produce a polished-type surface, which is often hard to see, even to the most experienced of viewers. And as the cylinder wears the amount of oil trapped in the peaks and valleys of the surface finish, even further bore wear results.

In diesel applications the remedy is down to the dispersant in the oil and how it interacts with the detergents to prevent these soot deposits from coalescing and becoming rather similar to a grinding paste. This is why diesel engine oils tend to have much higher levels of detergents and dispersants over those of oils intended only for gasoline engines.

Although bore wear is rarely directly associated with the choice of lubricant, it is often as well to be reminded of the role it actually takes in preserving your engine.

Fig. 1 - As the peaks and valleys of the liner honing are worn away, a shiny polished patch remains

Written by John Coxon

It was only when I came across the acronym 'T-A-M-P' that my confidence rose, and while it certainly didn't solve the problems, the structured approach it offered ensured that they were tackled in a clear and logical way. The letters of course stand for Temperature, Application, Media and Performance, and may be somewhat dated today what with all the complex solutions available, but as a starting point it should surely begin to focus the issues.

Temperature

The first parameter to consider is that of the maximum temperature to be experienced. Tackling this aspect in the first instance will quickly eliminate many of the possible candidate materials, especially if temperatures above 130 C are to be experienced.

However, when doing so, account must be taken of other sources of heat, for instance friction in a rotating lip seal, which could add as much as 20-30 C to the bulk metal temperature in the region. The close proximity of a turbocharger turbine housing might also need some thinking about. In some environments it might be wise also to consider the minimum temperature likely to be experienced, and the chance of that happening. Fluoroelastomers like FFKM for example perform exceptionally well at high temperatures. Unfortunately their low-temperature performance may be lacking when less expensive compounds could do a better job.

Application

Having ruled out a number of elastomer technologies, attention can now be focused on the application and any particular characteristics required. Rotating lip seals may need a certain amount of damping if the lip is to remain in contact with the (slight) elliptical orbit of the crank surface. Hoses, water oil or turbo air may need to have some flex resistance. On the other hand, an O-ring or moulded seal might need to deform significantly if it is to fill the groove designed for it, and in such cases an elastomer with a slightly lower Shore number may be required.

Media

The fluid to be sealed may at first appear simple, but when one considers that gasoline can consist of 400 or so different hydrocarbon compounds, in addition to any number of aggressive (to elastomers) additives, it may not be quite so easy. Add to that the move towards bio-components such as ethanol or butanol, and the problem begins to escalate.

Lubricants also have their concerns. Mineral oils, for instance, may be very easily absorbed into some elastomers, causing them to swell, while other components (PAO) in oil can have quite the opposite effect. Some oil suppliers are beginning to include more ester technology into their blends, so this may also need to be considered.

Performance

In some versions of the acronym, some might include P for pressure. In automotive applications pressure is perhaps of less concern, whereas performance - and in particular, things like shaft surface speed - are of more importance. In the engine world, 'high pressure' generally means high temperature and the use of a mechanical joint. For all others some form of elastomer technology would seem to rule. Nevertheless, these days, with the ever-increasing complexity of materials and fillers, designs and coatings, at this point you'd do well to consult the experts; at least you will have a clearer idea of what is required.

And who knows, rather just specifying the most expensive material, using this approach may be more suitable and save you some money as well.

Fig. 1 - Oil pump seal

Fig. 2 - Simple oil seals

Written by John Coxon

One of the simplest approaches has to be the straightforward flange. Simple to machine, with its surface at right angles to the cylinder centre line, the mating flange of the liner is clamped against it by the reaction of the cylinder head against its retaining bolts. In the past, sealing has been achieved using what is essentially a large, flat copper washer placed around the liner. Trapped between liner and block, this type of seal can be highly effective, rather like a 'fire' ring only midway down the liner and sealing the lower gap.

Unfortunately though, pure copper begins to soften at not much beyond 150 C, and the ensuing creep can very quickly destroy the clamping load at the cylinder head. When such simple seal arrangements are still used, copper has invariably been replaced by medium strength, anaerobic methacrylate sealants, which are not only cheaper but more forgiving and much easier to assemble to retain the correct amount of liner 'stand-off'.

However, the most common method these days is to use O-ring technology which, although comparatively expensive in terms of machining costs, has proven to be far more reliable, if a little fiddly at the assembly stage. Modern O-ring elastomers are also far less prone to the effects of aggressive oil basestock technology or hardening at the temperatures experienced, but despite the levels of reliability generally achieved, manufactures still often opt to put two or even three rings where arguably one should suffice. Whether the retaining grooves are machined into the block or the liner is down to the designer's preference, but in most race applications the former is more common, and since the whole of the liner is not loaded in compression, its thickness can often be reduced.

At one time I well remember these separate O-rings being replaced by a two-pack fluid silicone elastomer. Injected into a hole or gallery in the side of the block that led into the O-ring groove, the thixotropic mixture (one that becomes less viscous when subjected to an applied stress) flowed around the outside of the liner and into another gallery on the opposite side of the block. A close fitting between the block and liner at this position ensured that little or no elastomer escaped, and that when the sealant appeared at the other side a continuous sealing ring was assumed to be in place. Once the elastomer had set (after about 20 minutes), however, removal of the liner was difficult since, rather than relying on contact pressure - as in the case of an O-ring - the seal was a result of a bond between the elastomer and the materials surrounding it.

While the trials with this approach were successful for the most part, new rig test equipment and techniques had to be developed to seek out any weakness well away from the engine test cell. For, as we know, once inside a running engine, finding the root of any leak becomes so much harder!

It may often be the smallest of seals in the most obvious of places, but finding out what is truly happening if things go wrong is never that simple.

Fig. 1 - Lower liner sealing

Written by John Coxon

Open-deck cylinder blocks, where the engine coolant is in close proximity to the liner top, are unpopular in race engines. Insufficient lateral location invariably creates cylinder head sealing issues. While ingenious solutions for increasing the wetted surface area have been used successfully, as heat fluxes increase then attention surely has to be paid to the cooling medium used, with the liner and heat transport fluid surrounding it being viewed as a system.

For most applications, water is the perfect heat transport fluid. Good specific heat characteristics, low viscosity - and, above all, its easy availability - have made it the most obvious candidate. Various additives will reduce its tendency to corrode while safeguarding against freezing and boosting its boiling temperature; ethylene glycol, for example, can be introduced generally at concentrations up to 50%. But the presence of water around the top of the liner can have both good and bad consequences.

In the cylinder head and around the top of the cylinder liner where large heat fluxes occur, the process of nucleate or instantaneous boiling next to the hot surfaces helps to conduct the heat into the body of the fluid. Heat is removed from the surface using the latent heat effect, and the fluid later condenses and offers up that heat when away from the liner wall. So long as there is a constant stream of coolant, the process will continue efficiently.

However, if the flow becomes stagnant for any reason, the steam produced will eventually act as a thermal barrier, resulting in local overheating. Known as film boiling, even if it doesn't harm the contact surface then the instantaneous pressure perturbations may cause fatigue damage and consequent pitting, eventually leading to liner failure. The presence of water in a high-performance cooling system, where heat fluxes are high and the flow uncertain, is surely therefore questionable.

Pressurising the system raises the boiling point of pure water to 121 C at 15 psig, but increasing the pressure brings an invisible cost - that of requiring better sealing, stronger hoses and a greater risk of cooling system failure at some time. The alternative of a 50% mix of ethylene glycol and water is only a partial cure because any boiling of water in the mix can still cause film boiling at critical parts. But since ethylene glycol is toxic, surely this is undesirable anyway.

One solution alternative to using ethylene glycol (C2H6O2) is to use a close cousin, propylene glycol (C3H8O2). It has a very similar boiling point, at around 180 C (at atmospheric pressure), and although the specific heat is around two-thirds that of water, when used as a heat transfer fluid it gives more consistent cooling at high heat flux and virtually eliminates film boiling.

Since propylene-based fluids have much greater viscosity, engines will have to be preheated, but since this is common practice anyway, it shouldn't cause any extra concern.

Above all, however, the non-toxic nature of propylene glycol means that any accidental spills are not particularly hazardous - and the workshop cat, useful in the keeping the mice under control, will be spared serious harm should it accidentally drink any.

Fig. 1 - Nucleate and film boiling

Written by John Coxon

In general, an engine block casting consists of a number of main areas when looking at its production. These areas are:

" Crankshaft cavities, often used as the reference of the component at the start of the machining process

" Front and rear side (also often the top side), often acting as stable carriers of the total sand core structure

" Coolant and lubrication cavities, with a focus mainly on stable wall thicknesses

" The need to keep all internals on the inside, aside from providing support for all external bolted-on components

If we take a somewhat simplified look at producing an engine casing as part of the engine's development, flexibility is mostly required on the outside of the engine block and on the various coolant and lubrication passages. After the first prototype batch has been produced, functionality will be tested and vibrations measured. After analysing the results, design changes are often needed to improve engine behaviour. Typically the crankshaft cavities do not require many updates during development, and are therefore directly tooled from the beginning. This can also be said of the top end of the engine block. The head gasket side depends purely on cylinder distance, which is fixed from the beginning, and can therefore be tooled straightaway.

In most development projects, however, the same cannot be said for the coolant and lubrication cavities, as the probability that modifications are required during the development process is significant. It is therefore in these areas that the use of rapid prototyping processes has risen significantly over the years, in order to have parts that are available as quickly as possible. Laser sintering is often used for these kinds of cores, mainly because of their typically complex geometries. A machined core would have significant more risk of breakage.

The outer geometries of the casting, being the front and rear of the block, often have a 1:1 relation to the overall core structure of the part, and although these might require (often minor) modifications during development, they are sometimes rapid prototypes in order to decrease lead time.

Both engine sides are prone to modifications. As stated earlier, vibrations are one of the most important issues in a race engine, and often lead to improvements in bolt positions to reduce the vibration levels of external components. Therefore the patterns used for the outside of the engine blocks are well suited to being machined from a big brick of bonded casting sand.

In this way, the highest degree of freedom and therefore maximum flexibility can be achieved during the initial build of an engine for development purposes. The further the development goes, the more definite the engine block geometry will become, and investing in definite tooling can then become sensible. Remember though that it might still be economically more feasible to use rapid prototyping processes when production numbers are low, as such processes do not require specific and often costly tooling.

One should of course not underestimate the use of these processes in combination with each other. Every specific rapid prototyping process has its own peculiarities, but a number of rapid prototype casting companies have the required knowledge and experience to handle any such challenges.

So, when thinking of producing a rapid prototype component, find a partner company that can support you throughout the entire casting process - the 'one-stop shop'. All too often, nicely produced rapid prototype cores fail to survive the trip to the foundry. Remember that a core remains nothing more than a sandcastle, so be gentle with it.

Fig. 1 - Complex casting geometry, initially produced by using rapid prototyping and machined down in order to check stability

Written by Dieter van der Put

I am reminded of the time many years ago when I was looking at the possibility of building the ultimate engine - one that could be built up fully using robotic technology and, because of its construction, never be dismantled again. If the unit could be made fully reliable over its design life then it should never need to be dismantled, and therefore all the time and effort (not to mention the cost) in using mechanical fasteners where they weren't strictly needed could be saved.

Effectively it was a 'throwaway' engine, before the era of recycling and, but for the occasional oil change, would be more or less 'fit and forget'. With near 100% reliability, if anything did go wrong it was certainly cheaper and quicker to replace the whole unit rather than attempt any repair. Central to the concept was the one of replacing the gaskets along the various flanges with adhesives rather than sealants and, if not always deleting the mechanical fasteners, reducing them in size and cost.

Key to the initial part of the programme was an understanding of the anaerobic products being developed at the time and how the design of the power unit might change to use their properties more effectively. It would also provide feedback to their use in the more traditional designs being contemplated at the time.

Whether used as sealant or adhesive, the way in which these products act is often quite complex, but research indicates that they generally work either by mechanical or chemical bonding. In the simpler and more common form of adhesion - that of mechanical bonding - the bond formed between the adhesive and the surface occurs when the adhesive works its way into the small pores of the surface. Mechanically interlocking with the lattice of the metal substrate, the strength of the bond is down to the physical shape of the surfaces.

In the less common chemical bond, the adhesive or sealant bonds chemically with the surface of the substrate, and the forces involved are due to chemical attraction at a molecular level. In the case of anaerobic products, the basic element tends to be a monomer from the acrylic family, existing in the form of a polymer containing a double bond between two carbon atoms. The active constituent in the process is called a free radical, which under most conditions will react with the carbon atom, but when present much prefers to react with oxygen. In this latter state the polymer remains liquid, but when oxygen is excluded, the free radical reacts with the carbon double bond to produce a chain reaction and a solid cross-linked polymer.

So while the product remains on the flange of a mechanical component it will stay fluid, but as soon as air is excluded and the flanges placed together, the process of setting begins. The rates of cure depend on the metals used, but the good news is that aluminium considerably accelerates the curing time. Capable of coping with gaps up to 0.25 mm and temperatures of up to 230 C without degradation, the initial programme, as far as it went, was deemed a success. But the fact that fully bonded, throwaway engine technology is still not with us today may say something about the overall concept.

However, with so much invested in traditional engine design and technology, can you honestly see any manufacturer throwing this all up and simply sticking engines together?

Fig. 1 - Adhesive or sealant?

Written by John Coxon

A lightweight engine will surely, by definition, have its length as short as possible. At odds with the requirement to keep the bores large, the zone between the adjacent cylinders becomes progressively narrow, to the point when the structural rigidity becomes impaired. As this 'bridge' becomes narrower, so cooling becomes more critical, to the point where the temperature can destroy any heat treatment applied and lead to softening, loss of strength and eventually a loss of clamp load, resulting in head gasket failure.

One way around this problem is to introduce more cooling into the affected zone. This was the topic covered in an earlier RET-Monitor article. Another way is to increase the strength of this 'bridge' area by the addition of ceramics in the form of fibres or discrete particles into a cast-in cylinder of the same basic aluminium as the rest of the block, rather like the way the filaments of carbon or glass support the plastic matrix in a carbon fibre or glass-reinforced composite mix. Surrounded totally by the aluminium matrix of the cylinder, and unlike the filaments in other composites, the ceramic materials used for metal applications are generally alumina (Al2O3), silica (SiO2), silicon carbide (SiC) or boron carbide (B4C).

The result of many years of r&d, these materials have exceptional uniaxial tensile strengths and moduli as diameters reduce. For example, the tensile strength of a single 142-micron (0.0056 in) diameter sample of silicon carbide with a density of 3 g/cc can have a tensile strength of well over 5000 MPa and a tensile (or Young's) modulus of 400-plus GPa. Compare this with a typical casting aluminium, (LM25) having an ultimate tensile strength (UTS) of about 250 MPa, or a high-strength steel with its UTS nearer 750 MPa.

Referred to as metal matrix composites (MMCs), and just like the composites used elsewhere in the vehicle, there is no chemical bonding between the constituents and the aluminium of the matrix. And like carbon or glass fibre, these fibres take the form of whiskers (single crystals), continuous lengths or randomly arranged short lengths similar to the chopped stranded matt in the plastics world. The increased strength thus formed comes solely from the ceramics used, their size and orientation. And because of this, when used as fibres, the resulting properties can be highly non-isotropic.

When used as particles and not fibres, however, high specific strength may not always be the ultimate aim since other requirements, such as reduced scuffing or wear, or improved rigidity, may be the target. Whatever the precise qualities required though, and despite the difficulties associated with manufacture (which are considerable), aluminium MMCs will have improved strength at elevated temperatures, making them ideal for the bridge area between adjacent cylinder bores.

But if you spoke to my wife and asked her what would be needed to make me the perfect husband, she could only wish it were that simple.

Fig. 1 - Metal matrix composite cast-in cylinders

Written by John Coxon

Why is it so important and difficult to position the machining features correctly to the casting or forging? There are a number of reasons, but the main one is to be able to produce a part that has an homogeneous wall thicknesses throughout, without the risk of machining through the component's walls.

Every part has certain tolerances, within which the part can be produced - no part is ideal. Overall tolerances depend very much on the production process. For example, a part that is machined from billet has only the machining tolerances to take into account. These can be maintained very tightly by expert machinists on modern machining equipment. But when the same part is machined from a casting, the situation is more complicated. Considering that the casting process is a very complex one, in which molten material is poured into a mould made from sand, one can imagine that the tolerances achievable will be different from those of the billet, although the demand on the end product is the same.

The cast part itself already has certain tolerances of shape, within which the machining features need to 'fit'. The overall shape of the casting depends, among other things, on how the patterns are positioned, how much the cast part deforms and shrinks during cooling down. Sometimes the component even deforms during cooling down due to its internal stresses; see the latest article on the Heads-Blocks keyword.

Starting from the rough casting, the machining process needs to be initiated from certain reference points, called targeting points, which are often a number of specific cast shapes integrated into the casting. With these points the casting can be positioned for all six degrees of freedom. Typically the internal geometry - the crankshaft main bore and cylinder bores - is of most importance to the function of the end component. Because of this, the reference points are often located within these geometries.

Based on the deformations of the cast part, the crankshaft bore is often taken from the second and last-but-one main bearings. The outermost main bearings tend to show slightly higher tolerances in outward bending than the inner ones; this will centre the crankshaft bore. Taking the longitudinal references (sides) of these outermost main bearings will position the length of the part. After this, the cylinder centre lines are measured and taken as reference. Their positions will be averaged to achieve the full fixing of the casting.

When there's a smaller batch of cast parts to be machined, other geometries are counter-measured against this fixing position, creating a final chance to make slight modifications to achieve the best overall fit.

In this way the machining features will be best positioned into the casting, but it's only when taking a closer look at these reference points that one will start to realise their importance. For example, a pattern in the traditional sense - in comparison to pattern-less parts, created by laser sintered cores, for example - requires the several cores to be deformed out of the pattern. This means that, depending on the split line of the pattern, some of the reference points might not be exactly 'flat' but have be slightly tapered. And although these influences are mostly minor, these need to be taken into account, especially when considering the minimal wall thicknesses used in today's high-performance engines. Every extra millimetre of material in the overall wall thickness needed to compensate for these tolerances will add mass to the component.

This is exactly why the machining, both pre-machining as well as finishing, needs to be done by specialised machinists in well equipped machine shops. From the outside it often seems rather simple, but when exploring the production processes in detail, one will find out that the devil is often in the finest detail.

Fig. 1 - Offset machining due to reference point casting error

Written by Dieter van der Put

Every mechanical engineer has learned during his or her education that one can only load a structure so much, until the maximum permissible stress is reached. This implies that an external load of some kind is being applied to that structure. In the meantime we have all kinds of simulation possibilities to predict the maximum loading. Every so often, simulation results do not correspond with real-life behaviour. In that case things get rather more complicated, especially when the structure fails earlier then expected.

One possible reason, especially when dealing with cast parts, is the presence of residual stresses. Let us take a closer look at what this is.

Residual stress is defined as stresses that remain in a material without the application of an external load such as force, displacement or thermal gradient. Engineers working on structural cast parts, such as engine blocks and cylinder heads, will know what I'm talking about.

The mechanical properties of cast cylinder blocks are not uniform. Differences in wall thickness and overall geometry will result in non-homogeneous cooling, solidification and structure, and therefore the mechanical properties throughout the part. It is these factors that are also related to the residual stresses.

Residual stresses can lead to the initiation of cracks, often in combination with additional external loading. In extreme cases, and even without external loading, residual stresses have led to failure before the part has even been used. On the other hand, when it concerns compressive stresses it can also be beneficial to have these in the part, as it leads to increased 'loadability'.

More often than not the part will not fail, but other unwanted issues, such as deformations during machining of the part, can arise. Since typical values for residual stresses in a cast-iron cylinder block are about 150 MPa, both in compression and in tension, this will not directly lead to major issues. But given today's tolerancing and required accuracies, it does increase the risk significantly of producing scrap. To prevent this from happening, heat treatment can be used to reduce and balance the residual stresses. Given sufficient experience and development of the heat treatment process and its settings, overall stress levels can be reduced.

Perhaps the most important observation looking at the complex array of stresses, internally and externally, is that in today's world we cannot do without integrated simulation techniques any more. As I mentioned above, straightforward simulation results often do not fully correlate with real life, especially in the context of ever increasing performance:weight ratios. When looking at the current complexity of castings used in high-performance race engines, one can imagine that the influence of the production process has to be taken into account in the simulation process, and given the same relevance as all other loads on the component.

In summary, the addition of assembly, thermal and firing loads, combined with the casting stresses, will provide the most accurate safety factor, something every engineer is looking for. Local stress values, which are required to predict the final loadability of the part, can only be provided by detailed casting process simulation, using advanced non-linear elastic and plastic material models.

Fig. 1 - Cracked cast-iron cylinder block due to residual stresses

Written by Dieter van der Put

The statement: "There ain't no substitute for cubic inches" is as true today as ever, so when it comes to extracting more performance the first call is often to increase the size of the bore. Unfortunately, since few of us ever get to design an engine from the proverbial 'clean sheet of paper', the inter-cylinder spacings have invariably been decided for us. Torn between a rock and a hard place, enlarged bores on a fixed-bore spacing means only one thing - a narrower 'bridge' area and the prospect of bore distortion, scuffed pistons and engine failure caused by inadequate cooling. Cooling is the key but the options are limited depending primarily on the original liner installation - either wet, when the liner is replaceable and in contact with the cooling water, or dry, when it is not.

During the combustion phase of a reciprocating engine, the part of the cycle that generates the greatest heat release is the mid-stage of combustion when the piston is towards the top of the cylinder bore. Depending on many factors, the heat release will peak at 20-30º after TDC, with 90% of the combustion having taken place by, say, 40-50º after TDC. As the piston moves down the bore the gas expands and cools, so the critical part of any cylinder liner as far as cooling is concerned is therefore towards the top 30%.

For a wet liner construction based on cooling concerns alone, a mid-supported liner with the upper face clamped against the cylinder head would appear to be the best approach. In such an application a cooling water jacket thickness of less than 1 mm in the zone where two spun-SG iron liners come together has been proved to be more than adequate in the past. The downside to this approach is that while supported at the mid-point and clamped against the cylinder head only at the top, the tops are effectively free to move and present potential cylinder head sealing issues. The alternative to the mid-supported wet liner is that of the top-supported version which, although having poorer cooling, produces a much robust surface for sealing and minimal distortion.

For dry liners and blocks using plated parent metal bores, the lack of cooling in the zone between the bores will be a major issue as those bores become larger. In a cast-iron block I once recorded temperatures in excess of 360 C at the top of this inter-bore region, so some form of cooling is normally an absolute must. In the end the only practical solution was to drill a series of holes of progressively smaller diameter to pass through the siamesed region and ensure that a pressure difference across the block would ensure the coolant would flow through it. With only what must be a very small flow of coolant, temperatures dropped substantially.

Easy to do? Certainly not, for the long drill bits used readily wandered during the process and the angle necessary to avoid the drill chuck hitting the face of the block was quite steep. But after two or more scrapped blocks (actually I forget how many) the modification was eventually completed.

Smaller than it otherwise would have had to be? Yes. Better? You bet!! And cheaper? Well, two out of three isn't bad.

Fig. 1 - Cooling the siamesed block

Written by John Coxon

But when it comes to sealing the gap between a rotation crankshaft and the engine crankcase, a wholly different set of rules apply. In this case not only does the seal have to cope with variations in temperature, the shaft will also physically move up and down in the bearing, and in and out, and the seal will be required to maintain a constant contact against a moving surface for many hours. And but a single drop of oil constitutes a failure. No wonder then, for those in the know, that the most sensible thing to do with any competition engine is to use seals specifically designed for the task.

The vast majority of race engines are often modified versions of mass-production OE units. But while racers will use any number of specially designed high-performance head gaskets, for instance, they will almost certainly still insist on using the stock crank oil seal. This will have been designed by the OE manufacturer down to a cost, and will rarely have the same performance capability as one specifically designed for the much higher crankshaft speeds generally encountered in a competition unit.

At these much higher speeds the rear crank seal will often experience a great deal of distress when the friction present increases the temperature of the oil film at the lip. Degrading the oil, the abrasive carbon particles thus formed will therefore eventually grind a groove in the shaft. At times, and with a stock seal, this under-lip oil temperature may be as much as 60 C higher than that of the bulk oil. A specialised low-friction competition seal will help alleviate the problem -especially if you include a dirt lip in the design.

In designing a crankshaft seal, sometimes referred to as a dynamic seal, in addition to the space envelope available, the application and perhaps equally important the environment has to be taken into account. A transmission oil seal will be different from, say, a crankshaft seal, and the seal design at the front of the crankshaft could, as well as the difference in diameters, be different from that at the rear.

In particular, if an engine is to be run in a dirty environment an additional dirt lip will be considered essential. Designed to protect the main lip on the seal, the dirt seal is a second protuberance just in front of it and pointing downwards and outwards towards the shaft. Intended just to skim the surface of the shaft acting as a shield to the inner lip, the idea is to prevent the ingress of dirt but not actually add to the total friction. However, for those engines generating higher crankcase vacuums this dust lip can be encouraged to constitute the main vacuum-retaining lip, pulling the edge tighter onto the shaft.

If this is a simple rubber or NBR (nitrile butyl rubber) elastomer then the temperature created will burn the seal away. But in using low-friction PTFE-coated technology such as that used on specialised competition seals, as the vacuum increases the friction doesn't increase at the same rate, pulling more power at less friction.

A dirt seal may at first glance appear to be another unnecessary source of friction but, designed carefully, the crankcase vacuum can be increased to the point where significant increases in performance can be achieved.

Fig. 1 - Cross-section of a competition oil seal

Written by John Coxon

One of the most interesting things about an engine is that its various systems need to cooperate in order to achieve the engine's overall function - providing power to the vehicles we race in. These systems all have their own integrated function, for which the engine provides their respective infrastructures. Looking at the fluids in the engine - coolant, oil and fuel - this infrastructure means all of the connections, ports, channels and cavities creating the system's boundaries.

Traditionally, transporting fluids throughout the engine block has been via drillings and cast cavities - sometimes combined with lines, pipes or tubes - and their connections. Due to their simplicity, drillings have been the most favoured method, assuming accessibility for the drilling tool. When this accessibility is not available, however, or the geometry gets too complex, cast cavities are often used, and these are widely used in coolant jackets in cylinder heads for example.

It gets more difficult though when the geometries get too small and/or too complex to cast. For example, it's often not possible to produce oil channels using sand cores. The reason for this is the stability of the sand core, where certain minimum dimensions are required in order to prevent core breakages during the production process.

So, wouldn't it be worth it investigating other solutions to transporting fluids through the engine block casting? Another typical issue requiring a possible alternative solution is stress concentration at locations where drillings cross through highly loaded areas of the block, such as the bulkhead areas.

As a general engineering rule of thumb, every engine designer will try to design all drillings away from these areas, but sometimes this isn't possible. For inline dohc engines, typically no drillings are located in the bulkhead region other then those from the main oil rifle to the crankshaft bore. Typically these need to stay away from the threads of the main bearing bolt, so as not to increase stress levels. Vertical drillings from the main oil rifle feed the cylinder head with its camshafts, through holes in the head gasket.

With inline engines using in-block camshafts though it's a different story, as it is for vee engines with the oil pressure pumps located alongside the crankcase. In these designs the pressure drillings need to cross the bulkhead at least once, to get the oil from pump to main oil rifle.

Investigations show that a drilling at right angles through the bulkhead increases the local stress level in the bulkhead by about a factor of nine. Angled drillings will reduce this to a factor of three compared to no drillings at all.

If relocating these drillings does not seem possible, one could take several measures to reduce the stresses to acceptable levels, but unfortunately in almost all cases these will lead to increased wall thicknesses around the drilling and therefore additional mass and engine length. Are there no other possibilities?

Some years ago I was involved in a development project to cast a structure of pipes, the so-called pressurised oil feed structure. This consisted of a number of curved steel pipes to create the oil system for the crankcase. The block was an inline, six-cylinder cast-iron item with an in-block camshaft. The main oil rifle was on the exhaust side of the engine and the single camshaft at the intake side, both supported by seven plain bearings, each requiring their own oil feed. This meant the task was to install seven pipes into a sand core package that could withstand the casting process itself.

It was decided to drill the main oil rifle, due to the risk of leaks between the main oil rifle and the other seven pipes when integrating them. The oil system needed to be fully open and cleanable after the block was cast, and the existing machining operation was not allowed to be significantly changed. This concept was brought to the stage where a couple of engines ran successfully in the test bed, but did not make it to series production because the concept was not fully developed.

Let me describe on some of the lessons learned with this concept, which should be able to be overcome given sufficient time and attention.

• Effective closure of the pipes before the casting process. The pipes were extended significantly into the crankshaft bore sand cores (see image) to prevent them from filling up with fluid iron during the casting process.

• Fusion between pipe and cast material. To prevent leaks and uncontrolled cavities between pipe and cast base material, a secure connection between both materials during the pouring phase was required.

As can be seen in the image here, the oil feed makes a slight bend in the path from the main bearing bore (lower right corner) to the main oil rifle (left middle), in order to maintain sufficient clearance with the main crankshaft bore bolts. From the main oil rifle the oil feed heads up, through the bulkhead, to the camshaft bore (not visible in the picture).

The engines that ran with this concept did run successfully, providing the confidence that the concept is feasible, and therefore able to create some additional freedom for crankcase oil system design. Again, the major goal is to feed the oil system through low loaded areas of the crankcase, not strictly limited to the geometry of the cast material itself. If required, this concept would allow the pipes to exit the cast structure where applicable.

Together with the hybrid casting concept, this could open new thinking about engine block design.

Fig. 1 - Cast-in oil system concept in a cast-iron engine block's main bulkhead

Written by Dieter van der Put

For me, discarding anything with the value of something like a cylinder block is total anathema, and in these environmentally aware times it is perhaps the last thing I would want to do. And since recycling and repair is now very much back in fashion, when it comes to cylinder bores, reclaiming them seems the most sensible thing.

A few months ago we looked at the thermal spraying of cylinder liners but here I would like to examine one of the processes mentioned in more detail, the Plasma Transferred Wire Arc (PTWA) method.

The process is divided into four stages. The first involves machining the bore to remove the damaged surface before pre-treating the surface to accept the new material. Following deposition, the bore will be honed using recognised materials and techniques depending on the qualities of the surface required.

Developed initially by a US-based company, the plasma-generating gun consists of a tungsten cathode, an air-cooled pilot nozzle made from copper and a consumable wire acting as the anode. Attached to a rotating spindle, the gun rotates at up to 600 rpm while simultaneously moving up and down the bore. The plasma gas is introduced through tangential bore holes in the cathode holder while the wire is fed down into the central orifice of the nozzle.

When a high voltage is applied across the electrodes, the gas mixture is ionised and dissociated, whereupon it is forced to exit the nozzle at hypersonic speeds. Transferred to the consumable electrode, the plasma melts the tip of the wire and, together with another atomising gas, creates a finely atomised stream of particles targeted at the surface of the cylinder bore. Used in conjunction with low carbon alloy steel wire and compressed air, finely divided iron oxides (FeO) can be created within a softer iron matrix which, once suitably bored and honed, creates an ideal surface for a piston ring to cross. Interestingly, despite the particle temperatures reaching 2000 C, the plasma torch needs no external cooling.

Since the material is not actually bonded to the liner surface, preferring instead to be mechanically locked to the substrate, the method of pre-conditioning the liner surface is critical. The normal method for roughening such surfaces is to grit-blast. Highly effective in creating a surface full of undercuts into which the plasma-sprayed metal can be forced, the presence of carborundum inside any engine requires meticulous cleaning afterwards, and is therefore not necessarily the best. Perhaps better alternatives are high-pressure water jet blasting methods or a fine machining process giving a dovetail-type surface finish similar to that seen in the figure here. Giving a very good bond strength - twice that generally reckoned to be recommended for aluminium liners - this latter process was developed by Braunschweig University of Technology in Germany.

A technique perhaps more suited to the needs of an OE engine supplier reclaiming damaged cylinder blocks, and not so far as I am aware used in competition engines, the processes described produce a reliable and low-cost option for reclaiming cylinder liners.

Fig. 1 - Thermally sprayed coating locked into the mechanically roughened substrate

Written by John Coxon

engine architectures, will recognise the issue, and the tell-tale signs of the oil slick in the driveway. It may be annoying for the house-proud enthusiast but for the racer the situation is potentially much more serious.

The real problem is not necessarily the seal itself but the task it is asked to do lurking behind the engine flywheel. In attaching the flywheel to the crankshaft, the rear flange invariably has to be bigger than the main journal to avoid overstressing it. Big enough to accept the six or eight bolts clamping the assembly together, the diameter of a one-piece rear crank oil seal is of necessity therefore smaller than this flange. Consequently, to enable assembly, this rear seal historically has to be installed in two halves.

And that's where the problem starts - two semicircular 'moons', one in the block and the other in the main bearing cap, means two potential zones for leaks. While the seal itself may be (just) adequate in dynamic terms when the crankshaft is rotating, once it has stopped, oil will in time inevitably leak along the two-piece seal split zones.

However, I have recently seen one design of seal that offers a degree of hope for all big-block Chevy 'sufferers'. The new dual-lip seal uses modern fluoroelastomer technology for both seal lips, and incorporates a low-friction PTFE coating on the key contact surface. When assembled correctly, it should significantly reduce the incidence of these leaks and possibly eliminate them altogether. Conceptually the seal is designed to be split in one position only.

Since the seal is flexible it can be manipulated around the crankshaft before the crank is assembled back into the engine and the bearing housings replaced. The two edges of the seal will match up inside the recess and no leakage path should result. However, to minimise any chance of a leak, the split end of the seal is assembled into the engine so that it is uppermost and therefore less likely to see too much oil. Reducing the number of potential leak paths to only one, positioning it in a less aggravating position and using much better sealing technology will, if the seal is assembled correctly, minimise the chances of leaking and possibly provide an oil-tight solution.

A relatively recent innovation and by no means unique, early test bed running on one particular prototype seal not only appears to have solved the static oil leak problem but also, according to dyno tests, gives improved torque and a useful additional 12 bhp. While this in no way can be attributed to the PTFE coatings, the general opinion is that the increase in bmep is more likely the result of the higher crankcase vacuum created under test.

Creating an extra 12 bhp and building a leak-free Chevy for only a few extra dollars? Sounds like good value for money to me.

Fig. 1 - Big-block Chevy rear crankshaft main oil seal (Courtesy of GST Racing Seals)

Fig. 2 - CAD detail (Courtesy of GST Racing Seals)

Written by John Coxon

Looking for alternatives, on the one hand possibilities for improving the existing material could be considered, such as shot-peening of critical areas like valve bridges. The goal is to put compression in the material, providing a higher load margin before critical tension occurs. As long as the temperature loading remains modest, this is a well known and popular technology.

At the other, more extreme end, a totally different cylinder head material could be thought of, combined with sophisticated machining capabilities in order to take away the overall traditional material limitations. One could think of fully or partly machined cooling jackets, and the requirement to gain machining access to the individual parts of the assembled cylinder head.

In between these conceptual ideas, there is a world of possibilities which, for many reasons, are not directly used on broader scale in current high-performance engines - reasons such as product cost, manufacturing lead times of components and the significant development time and cost required to bring a early concept to the required maturity level with which to go racing. That is why the concepts, which are not that far removed from common practice, are being developed for engine and component testing in order to be used in competition racing.

There are always two sides to looking at conceptual ideas. One should be able to think openly without being limited by manufacturability or whatever. Only this will lead to thinking 'outside the box', often generating further ideas. On the other hand, there is always the need to 'come back to earth', in order to get the real component on the test bench.

Would there be a concept for cylinder head cooling that outperforms current known water jacket designs? Let's take a look.

Although in the comment placed on the website by a reader named Terry, the idea was expressed to get rid of the coolant by a different material choice, I have taken a somewhat less extreme approach in this case, recognising the absolute must for a liquid-cooled engine.

When considering cooling of the cylinder head, the main area of interest is the valve bridge area, more specifically the exhaust valve bridges. The second area of interest is the exhaust port area, although design-wise the port length in the cylinder head could be shortened significantly, reducing thermal loading and therefore reducing the need for cooling this area. A neat example of a minimalistic design of cylinder head port lengths is the Cosworth CA cylinder head.

Focusing on the exhaust valve bridge area, conceptually the distance between coolant and flame deck is of critical importance. To achieve the shortest distance from coolant to valve bridge, and maintain sufficient wall thickness and strength, a possible solution would be to include drillings through (and parallel to) the flame deck exactly between the valve seats. This design is used in very large engines, where obviously there is more room for drillings.

In order to investigate the potential on cooling, concept CAE simulations have been carried out on a very simplified cylinder head model - just a square box, including four valves and ports. These simulations have shown the potential of this concept on cooling, with increased structural behaviour, due to the fact that the flame plate thickness can be increased as it no longer needs to be as thin as possible for cooling purposes. The temperature of the exhaust valve bridge could be significantly reduced, and the temperature on the flame deck could be kept far more homogeneous, resulting in less distortion and stress.

Analysing the possibilities of drilling these kind of coolant channels into the flame deck of the cylinder head led to a couple of 'challenges'. One major issue is the available wall thickness and clearance between the valve seats to drill through. Nowadays the seats are located so near to each other (maximum valve size:bore ratio) that hardly any material is left to drill through. To overcome this, one could think of getting rid of valve seats as separate inserts, and creating a wear-resistant seat in the cylinder head parent material.

A second area of attention is the spark plug dome. A specific detailed design would be required not to have the spark plug in direct contact with coolant, leading to possible leakage paths. Both these challenges, however, look as they could be overcome by smart engineering.

I have described a possible conceptual approach here that could lead to a simpler, more efficient cylinder head design, again from a cooling perspective. Would the conceptual advantage be of sufficient magnitude in order to replace current thinking? I would be interested in your ideas.

Fig. 1 - Conceptual cylinder head design: deformation

Written by Dieter van der Put

Until that time, cast-iron cylinder liner technology in race engines was pretty ubiquitous. But whereas cast iron was safe and generally well understood, the stronger or lighter materials such as steel or aluminium on their own were nowhere near as satisfactory. The solution, at least at the time, was a form of chromium plating referred to in engineering circles as 'hard chrome'.

Used widely in many engineering steels where corrosion is a concern, pure chromium on its own is a sort of bluish-white, not too shiny but brittle metal notable for its resistance to wear. Electroplated on a surface as a thin film, so-called 'chromium plating' can sometimes be a mixture of nickel and chromium with the nickel deposited first, the brightness of the decorative finish being a result of the nickel showing through the very thin chromium layer.

However, to make use of its engineering properties, the material has to be deposited in very much thicker layers (from 0.2mm upwards) that take on the colour of the metal itself, a much duller finish. This 'hard chrome' can be deposited onto almost any metal surface, steel or aluminium alloy liners being the most obvious for cylinder liners, and as well as having a hardness exceeding 900-1100 Vickers harness, it has excellent wear resistance - and, more important, a low coefficient of friction.

During the electrolytic process, hard chrome can develop small, pinprick voids of porosity or even micro-cracks. By altering the composition of the electrolytic bath and changing its temperature, the number of these cracks can increase, which will reduce the corrosion resistance and fatigue strength of the layer but increase its ability to retain lubricating oil. This ability to trap some form of lubricant is essential in cylinder bore surfaces, so an optimum balance has to be found. However as the surface steadily wears away these cracks may be steadily diminish in size, reducing the volume of oil available for lubrication. Increased wear will therefore inevitably result. Chrome bores therefore still require honing to finish.

The process of honing is designed to introduce more channels for this oil to reside and so the optimum surface coating of hard chrome therefore may seem to involve a combination of careful honing together with a set amount of porosity. Initially, with least bearing area but most trapped oil, wear rates may seem initially high; this however will reduce after a while when, after bedding-in, the compromise between bearing area and trapped oil is more favourable. Finally, as wear proceeds and the trapped oil diminishes, wear rates will begin to climb again. When specifying hard-chrome finishes, therefore, if contact pressures are high then experience dictates that the less porosity the better.

As for piston ring compatibility, chrome-surfaced rings against hard-chrome bores is not to be recommended. Molybdenum inlays or even plain cast iron might be best.

In 1969, early Porsche 917 engines were fitted with individual cylinder liners made from high-silicon (hypereutectic) aluminium alloy with chrome-plated walls. Once into the 1970s, however - when Led Zeppelin and Deep Purple were strutting their stuff, and nickel silicon carbide coatings were more commonly available - hard chrome on cylinder bores was slowly relegated to historic units and large diesel engines, where the combination of high levels of durability and comparative low cost are perhaps more important.

Fig. 1 - Comparison of material hardness of hard chrome

Written by John Coxon

Today we take it pretty much for granted that gas filled 'wills' rings or beryllium copper crush rings will solve the problem. But these require large studs feeding through the cylinder head, compromising not only the design of the water jacket therein but also producing high tensile loads where they locate in the cylinder block. One method of getting around all this is to discard these sealing rings and put the steel or cast-iron cylinder liner into compression, forcing it against the soft aluminium of the fire face on the cylinder head and dispensing with the seal. This has been achieved over the years in a number of ways.

One method is to use long cylinder head bolts. Passing through the cylinder head and then block, and anchoring into the main crankshaft bearings, they put the whole of the cylinder block into compression and clamp the liner flange at the top into compression. Successfully achieved on a number of prototype engines, the design still requires design compromises in the cylinder head water jacket to accommodate the long bolts as well as dictating the position of the bolts around the bore by the position of the main bearings. The Alfa Romeo Tipo B Grand Prix engine of 1932 used a slightly different method by clamping the liner into compression against the bottom face of the cylinder block of the monobloc construction.

Reasonably satisfactory but not necessarily foolproof, the technique used in later Grand Prix Alfas, used a different sealing method. Starting with the Type 158 of 1938, a single-stage Roots-type supercharged straight eight, the thin-walled steel liners were screwed into the aluminium alloy combined block and head, monobloc casting. Restorers of these classic engines tell interesting tales. With a 'twelve thou' (0.012 in, 0.3 mm) interference fit, the liners have to be cooled in liquid nitrogen complete with tooling, and the monobloc warmed in an oven before assembly can begin. A two-man job, assembly is fairly straightforward, but should things go wrong or the liner jamb part way in, the whole casting risks being scrapped. Ouch!

Used on Grand Prix Alfas until the early 1950s, screwed-in liners can also be found on the unsupercharged, Lampredi-designed 4.5 litre V12 Ferrari engines of the same period. Once again, combining the cylinder head and block water jackets, the thin-wall steel liners, threaded at their upper end, passed up through the open bores of the bottom of the water jackets and screwed into matching threaded bosses surrounding the combustion chamber. Projecting out of the base of the water jacket/cylinder head casting and sealed by a pair of simple O-rings, these liners located directly into the crankcase.

Much nearer to the present day, John Judd used screwed-in liners in the 1996 Grand Prix OX11A Judd-Yamahas. Claiming reduced bore centres and a simple but stiff water jacket, free of stud bosses, the engine weighed 99 kg without wiring loom and ECU.

With a maximum engine weight limit in Formula One, improved parent metal bore coating technology and the many difficulties associated with casting, to my knowledge screwed-in liners have once again been confined to the realms of history. Pity.

Fig. 1 - Screwed-in liner

Written by John Coxon

The unceasing demand for higher specific power has enforced higher operating temperatures on engine parts, in particular cylinder heads, which has a significant effect on thermo-mechanical loading and consequently on mechanical damage. The cylinder head is one of the critical components of a high-performance race engine that is subjected to complicated thermal and mechanical loadings. Temperature distribution and gradients lead to thermal stress, and could eventually lead in turn to thermo-mechanical fatigue of the cylinder head.

As was briefly touched upon in an article by John Coxon in the May 2010 issue of RET-Monitor, the cylinder head coolant system relies on a mechanism called nucleate boiling. John mentioned that the localised temperature of the surface exceeds the localised boiling point of the transport fluid, so nucleate boiling occurs at the boundary. "This is good, and can actually increase the heat dissipation at that point," he said.

First, let me explain that boiling is the formation of vapour bubbles at the heating surface. The boiling heat transfer is sensitive to the temperature delta between surface and liquid. In addition, the heat transfer coefficient is affected by the local vapour-liquid mixture ratios and flow velocities.

A typical boiling point curve is shown in the attached figure. Up to point A, heat transfer occurs by natural convection, so no boiling occurs. From point A, the surface temperature is high enough to activate nucleation sites, usually around the combustion chamber and exhaust valve area, and vapour bubbles are formed. Heat transfer increases due to the very rapid, almost explosive, formation of the bubbles which causes a very strong local velocity within the liquid film. In this phase, large numbers of bubbles form on the hot surfaces and travel through the bulk of the coolant, later condensing as they move to a lower temperature region, distributing the heat into the coolant.

The increased heat transfer reduces the cylinder head surface temperature and increases energy added to coolant. In the region from A to C, more bubble nucleation sites are activated - this is the region of nucleate boiling.

As you can probably understand, the design of the coolant system in total - and specifically the coolant system in the cylinder head - requires significant development time, effort and experience. In order to predict heat transfer in the cylinder head, 3D CFD and FE analysis is used, in combination with 1D modelling. 1D models are typically used for well-established flow without swirl and measured over certain lengths with constant cross-sections.

In this case, however, these models have to be applied to swirling, accelerated and undeveloped flows in non-constant cross-sections, as in cylinder heads. Here, flow behaviour may differ significantly from that in straight sections. Therefore, 3D simulations are used to investigate local behaviour and flow effects. Based on these 3D simulation results, 1D empirical models can be developed to improve accuracy, while still achieving the required short process lead times, leading to cylinder head coolant system predictions being available early in the development process.

The nucleate boiling mechanism is very sensitive to a number of parameters, such as coolant system pressure and temperature, coolant flow velocity and cylinder head material. Working on high-performance engines, the deeply rooted drive to achieve the highest efficiencies is what makes us tick, so we use high-temperature, high-pressure coolant systems - all in order to keep coolant heat exchangers as small as possible, and packaging space as aerodynamic as possible.

With these parameters already on the engineering extremes, might there be an opportunity for looking at the cooling fluid composition?

Fig. 1 - Typical Boiling Curve (Courtesy of The McGraw-Hill Companies)

Written by Dieter van der Put

Low-silicon (around 7-8%) aluminium alloys, the type favoured for casting, have poor tribological characteristics when used against any piston ring material. On the other hand, high-silicon aluminium alloys, those around 17% silicon, are difficult to machine. The traditional approach with the low-silicon alloys therefore has been either to use heavy cast-iron cylinder liners, either cast-in or free standing and pressed into place, or an electroplated surface consisting of hard silicon-carbide particles in a nickel matrix commonly referred to by the one of many trade names ending in the letters 'sil'.

All these approaches have disadvantages; cast-iron liners impart as much as a 30% weight penalty while electroplated surfaces encounter environmental issues and have been linked to corrosion due to certain impurities in fuel.

One option that avoids all these issues and which is becoming increasingly popular because of its low cost is the use of a thermal spray process to spray atomised molten metal onto the surface of the bore via a plasma arc. In this way a hard, wear-resistant coating can be applied to just about any engineering surface.

To date, a number of these processes have been developed, each one with slight variances to circumvent many of the patents that exist. However, most are characterised by the use of either metallic powder or wire as the feedstock coating material, a rotating spindle travelling up and down the axis of the bore and a gun head, which creates the plasma arc spraying the atomised metal towards the bore surface.

Since the mechanism of bonding of the layer to the substrate is predominantly mechanical, the surface finish before the deposition is critical. One technique is to roughen the surface of the substrate using traditional grit-blasting techniques, taking care to ensure all the grit is removed afterwards. Used now, for more than 10 years in Formula One, GP2, MotoGP and many other categories, this would seem the most popular approach, and is sometimes referred to as surface activation.

Another option, developed in the US and claimed to be more applicable to volume manufacture, is to machine the surface using special tooling to give the necessary undercuts in the substrate surface for the molten metal to 'lock' into. The use of this system to hold the coating to the wall of the cylinder is consequently claimed to be more reliable, and enables material to be built up on otherwise incompatible materials, but with more than three million bores coated using the grit-blasting technique, this is often disputed. Typically the process will deposit a thickness of 200-250 µm, which after final-stage honing will be somewhere around 100-150 µm depending on the material and the precise thermal spraying technique used.

There are three main systems currently in use or under development.

• Rotary powder plasma spray, sometimes referred to as Air Plasma Spray (APS)
• A rotating twin wire arc system (TWAS)
• A plasma-transferred wire arc system or PTWA

With research in this area very much ongoing, many of the coatings used are based around low-carbon (0.1-1.1%) steels, although with powder systems a broader range of materials can be used - including ceramics. Using a suitable powder feedstock and compressed air to atomise and spray the metal, hard, self-lubricating oxide particles of FeO are formed within a softer iron matrix on the surface of the liner. This performs rather like the silicon carbide in the nickel matrix of a Nikasil coating. And since the deposited material is porous, any cavities revealed after honing will act as reservoirs for lubricating oil. More important perhaps, you won't need exotic coatings for your piston rings, unlike many of the nickel silicon carbide-based coatings.

Fig. 1 - Schematic of a thermal spraying process (Courtesy of Sulzer Metco)

Written by John Coxon

Hidden out of sight somewhere between the rearmost main bearing on the crankshaft and the engine flywheel, the zone is rarely awash with oil but neither is it completely free of it, so designers in the past have come up with solutions that worked tolerably well at the time but are hardly acceptable for today's drip-free expectations.

The principal issue relates back to the practice of attaching the flywheel to the crankshaft and ensuring that the fastener loads are distributed evenly around the mating flange. In most situations this flange will have a diameter greater than that of the rear bearing and consequently, the usual method of sealing - fitting a lip seal - wasn't possible. At this time the surface speeds tolerated by the elastomer technologies available were much less.

Advances in elastomer and lip-coating technology eventually cured the problem, enabling the manufacture of much larger diameter seals. But engines designed much before the 1960s or '70s didn't have that as an option. For them, fortunately, oil leaks were an accepted fact of automotive life, possibly a carry-over from the steam engine tradition when a wipe with an oily rag was the most practical solution - and anyway, the presence of a liberal quantity of oil served to protect the underside of the car from the ravages of rust.

At the time, a number of possible solutions existed and, apart from efforts to direct the oil spraying from the rear bearing away from the area, other options included a metal disc or flinger and/or the machining of some form of scroll device on the crankshaft immediately after the journal. Acting as an Archimedes screw, these spiral grooves are a close fit between the surfaces of the crankcase, above and sump, below. As the crank rotates, the action is to pump any oil trying to escape back towards the rear bearing again and return it to the sump.

In practice, this system worked very well when the engine was running; it was only when the engine came to a halt that slight leakage occurred, slowly dripping down through the flywheel/clutch bellhousing and eventually ending up on the garage floor or driveway. And in time, through wear in the bearings or movement of the sump, this leak would increase.

Because no contact took place, however, these mechanical seals are in some ways superior to lip seals, since (unlike the lip seal) no frictional forces are involved. But today's demands for drip-free driveways and the availability of fluoro-elastomer seals that can tolerate much higher surface velocities have virtually outlawed the Archimedian scroll, when even on classic engines these are being replaced by more modern (if less mechanically efficient) technology.

R. I. P. the DRIP.

Fig. 1 - The rear crankshaft oil flinger and Archimedian scroll

Written by John Coxon

A visit to the Race Retro exhibition at Stoneleigh in the UK brought this home to me recently when examining the handiwork on one of the historic engine exhibits. At this, the Mecca of the historic racing world, vintage engine manufacturer Jim Stokes showed his latest racing recreation - a 2.7 litre star straight eight of the type fitted to the Alfa Romeo Tipo B, otherwise known as the famous Alfa Romeo P3.

An inline eight-cylinder unit made from two, four-cylinder cast-iron blocks with camshaft and supercharger drive coming from a central gear train, the most noticeable aspect of this engine was the combined structure of the cylinder heads and cylinder block in a single casting. Splitting the engine between cylinder block and crankcase, rather than between cylinder head and cylinder block, allows the complicated and therefore potentially unreliable head gasket to be replaced by a simpler sealing arrangement at the crankcase cylinder block interface. In this particular design the cylinder walls were formed by dry steel liners inserted from the bottom, shrunk into position and retained via a small flange against the top of the crankcase.

Apart from the deletion of the head gasket from the bill of materials, one major design advantage should be better cooling towards the top of the cylinder liner, as well as the complete absence of distortion as a result of the omission of clamping bolts and the high strains that would otherwise be introduced.

The downside, apart from the difficulty of manufacture in the first place, is the inability to service the valves and valve seats easily. At the time, this would surely have been a major handicap, involving the removal of the engine and a complete strip-down to remove crankshaft and pistons assemblies. As so very often happens in engine design, overcoming one particular hazard is likely to introduce many more of a different type. And even now, simply retouching the 16 valves (this engine was only two valves per cylinder) each inclined at 52º to the vertical is reputed to take somewhere near 40 man-hours of work!

But of course the Alfa wasn't the last engine to use this type of construction. Indeed, apart from many early Grand Prix engines (for example Peugeot, Bugatti and Delage), integral head-block designs were often used well into the 1950s, at which time more reliable mechanical seals were developed. Even in the 1980s, engine designer Brian Hart introduced such a design with a 1.5 litre turbo unit used in Formula One.
It may be the cause of many a headache for some over the years but eliminating the head gasket at the design stage can unleash a whole new set of migraine pain for others.

Fig. 1 - Alfa monobloc casting

Fig. 2 - Alfa monobloc - combustion chamber viewed through the cylinder block

Written by John Coxon

Where the external loads are strictly mechanical, these deformations typically remain in the elastic area, meaning that the deformation will disappear when the external load is removed. These kinds of deformations are so-called linear deformations. As explained in earlier RET-Monitor articles, for example here, plastic deformations also occur in cylinder heads, mainly as a result of heat loading.

When mechanical loading (combustion, assembly) is combined with thermal loading (heating up/cooling down), things get more complicated. But there is some pretty straightforward analysis on the relation between the deformation of the cylinder head (due to heat) and cylinder head design, in this case valve pattern orientation.

The starting point of this analysis is the fact that the exhaust valve area is the most heat-influenced zone, and is more or less shaped as an ellipse around both exhaust valves. The result of the difference in heat loading between intake and exhaust is that the flame deck on the exhaust side will deform into the combustion chamber significantly more than the cooler intake side. When considering a straightforward four-valve DOHC race cylinder head with a parallel valve pattern, this deformation will be quite symmetric. Most of us engineers are symmetrically focused, and therefore we will find rather easy the required design solutions in order to withstand these loads.

But, considering that some race engines also still use other valve patterns, such as slightly rotated, diamond-like valve orientation or even the older two- or three-valve heads with asymmetrical valve orientation, the deformations start to look far more complex and not so straightforward.

Some years ago I had two students doing their practical assignment under my supervision. I gave them the task of investigating the deformation effects of heat loading on several cylinder head concepts (existing competitor cylinder heads were taken as input), using simple-concept CAD and CAE modelling. A couple of design variables were defined, where valve pattern orientation was the primary one. Heat loading was kept constant to get clear comparison.

Something perhaps already known to some RET-Monitor readers, but which was an eye-opener at the time for us, was the simple fact that either a zero or 90° valve pattern was pretty easy to control from a deformation point of view, but valve orientations in between these values led to significant asymmetrical behaviour when looking at the area where the exhaust ports connect to the cylinder head and the main head geometries.

The exhaust valve area deformations are transferred through the port structures to the upper regions in the cylinder head, leading to significant stress levels at the locations where these connect to the main structure or port-to-port connections. Making geometry modifications to several areas of the cylinder head structure - such as flame deck thickness, profile, port geometry, port wall thickness, coolant jacket shape and so on - provided a conceptual understanding of creating the geometry in such a way that the stresses could be absorbed in an homogeneous area, in order to prevent stress concentrations, possibly leading to cracks.

As mentioned above, the goal has been to provide some basic insight into the different aspects of cylinder head loading, and the consequences. Where the thermo-mechanical fatigue was described and discussed in earlier articles on the Heads-Blocks keyword, this article has focused more on the additional mechanical loading and load paths due to thermal deformation of the flame deck. When investigating these areas using rather simple conceptual designs, a basic engineering feel and experience can be developed that helps in understanding the complex behaviour of cylinder heads.

Engineering cylinder heads in these times requires the use of high-end CAD and CAE packages, which cannot be overlooked any more - certainly when designing these components on the limit between loading and mass. But engineering knowledge and conceptual understanding of cylinder head behaviour is required before these tools will deliver added value to the design.

Fig. 1 - Deformation of the cylinder head

Written by Dieter van der Put

In the case of a purpose-designed race engine the answer is simple - new liners, new pistons and rings and she'll be like new. However, in the case of our 'spec' unit and where rebuild costs may be included in the lease agreement, such luxury is out of the question, and anyway, the liners are most likely cast into the cylinder block. The chances are that the standard production rings, designed for up to 200,000 miles of sedentary road use, will have hardly worn at all and it could be that the increased leak-down is all in the valve seating anyway. Replacing the rings will be no great hardship since the engine will need some form of initial bedding-in on the dyno, but in order to assist this using new rings do we need to do something to the surface of the cylinder bore as well?

This is a dilemma faced by many an engine builder, particularly those, say, supplying Formula Ford engines where the regulations are quite strict. Should we 'glaze bust' the engine or should we just fit new rings (assuming no obvious bore damage or excessive wear) and hope for the best?

To start with, let's get our terminology correct. Although the title to this article is 'glaze busting', bore glazing is a phenomenon normally observed only on diesel engines that run a light load for long periods of time. The 'glaze' is in fact a lacquer coming from the oil or products of combustion that condense on the cold surface of the cylinder bore and fill in all the ridges and furrows (however invisible to the naked eye) that represent the surface topology. Hard and almost impossible to remove, the ring will pass directly over it, producing excessive exhaust gas blow-by and high oil consumption.

In the case of race engines, however, what I am referring to is 'bore polishing' - when the peaks and valleys are worn away gradually by the passage of the piston ring. In diesel engines this can be caused by the use of high-detergent oils that produce ash which acts as a grinding paste in the ring bore contact zone. In race gasoline engines, however, when these detergent additives are present in lower concentrations, this is attributed to just general wear and tear.

In the case of our Formula Ford engine, when new rings have been fitted there are a number of possible approaches. One option may be to do absolutely nothing. After only less than 1000 miles the amount of wear against the standard Ford ring pack would, it is assumed, be almost negligible, so the plateau finish produced at the Ford factory under closely controlled procedures would be hardly affected. At the other extreme, re-honing is surely out of the question. This would remove too much material, increasing the piston-to-bore clearance, which would increase even further after bedding in.

The middle option, and one used by some engine builders known to me, is to use what is known as a 'Flex hone' or 'bog brush'. A resilient flexible cutting tool with a soft cutting action, the Flex hone consists of a number of abrasive globules bonded to the ends of nylon filaments, mounted spirally around the axis of the tool. Placed in a simple pillar drill or even handheld electric drill, the tool is run in and out of the bore at about 350-500 rpm and 30-200 strokes per minute for no more than 10-20 s. Depending on the abrasives selected, this produces a surface finish that is less harsh than that of the traditional hone and, done correctly, once the engine has been run-in it can restore leak-down test results to their original values.

Fig. 1 - The Flex hone

Written by John Coxon

Sometimes it isn't the fault of the component but perhaps the environment into which it was placed. Sometimes it may have worked well in engines over many years and then suddenly as a result of a single, slight change to the rest of the engine it goes on to fail. And other times we get the strangely bizarre - the failure we could have never predicted, but for our experience, for 100 years. Given a long list of engine failures over time you will therefore not be surprised that the cylinder liner features quite prominently.

An example of the first type of failure is in the machining and preparation of the upper flange and its location in the cylinder block. Poorly machined liner seating or a small particle of dirt trapped between the liner flange and its seat can lead to sudden failure at the assembly stage when the cylinder head is finally tightened. Bending loads introduced into the brittle iron liner by the action of increasing the clamp load can introduce excessive shear stresses, which cause the liner to crack at an angle of 45º all the way around. Characterised by a noticeable 'ping' at the final stage of assembly, the heart - and wallet - drop.

Manufacturers, especially engine manufacturers, tend to be a cautious bunch. Components that have served their masters well will often be carried over into new designs, sometimes untouched or with only slight modification. A major change in liner material has, perhaps understandingly, caught out at least two luxury high-performance vehicle manufacturers. Another, however, their liner technology having been proved and apparently unchanged over many years, is starting to alarm owners with low-mileage examples, fastidiously looked after since new.

The reasons for the failure may or may not be understood by the one-time manufacturer, but with engines well out of warranty the aftermarket is trying to understand if there is a common root to the problem. Scuffed liners as a result of piston seizure would seem to be the outcome, but is this down to marginal lubrication, piston ring sticking or marginal cooling? The cause is not very clear.

Early examples are apparently rarely affected but when increasing the stroke in later versions the failures increased. In some cases, classic D-shaped fatigue cracks also appear in the liner, propagating from the coolant side, suggesting perhaps some form of cooling problem. But is this one problem or two - cooling or lubrication, or perhaps even a combination of the two? At this point the engineers among you will be wondering, and will possibly already have their own thoughts or ideas.

Finally, one of the most bizarre failures I have heard of occurred in a large marine diesel, but it could happen just as easily in a race engine. A shock wave from the combustion of a slow-speed diesel unit traversed the cylinder liner and passed into the cooling water. High coolant temperatures and low cooling water flow rates produced a condition where the repeated pressure waves coming through from the combustion chamber created what is known as cavitation in the liquid next to the liner wall. Alternating pressure spikes and then rarefactions in the liquid caused vapour bubbles to form and then implode again, eventually eroding the external wall of the liner, until eventual failure.

Bizarre or not, a one-off occurrence or a common problem, it is difficult to say, but whenever you have high thermal throughput, low coolant flow rates and rapid changes in pressure, the destructive forces of cavitation may not be very far away.

The cylinder liner may be a familiar component but it still has the ability to surprise and confuse in equal measure.

Fig. 1 - Liner installation failures
Fig. 2 - Classic D-shaped crack in the cylinder wall

Written by John Coxon

In my most recent article, a number of options were mentioned to reduce cylinder head flame-deck temperatures - achieving good flow conditions and velocity of the coolant through the cylinder head, transferring heat as quickly as possible from the heat source to the cooling medium and getting coolant as near as possible to the hot areas. In this article I would like to focus on the first point, coolant flow conditions and velocity, and provide information on the advantages and disadvantages of different cylinder head coolant flow concepts.

In general, one can distinguish between two basic coolant flow concepts. The first is longitudinal coolant flow, which can be described as a coolant flow that enters the cylinder head at one side of the engine and flows through the cylinder head to the last cylinder in a longitudinal direction along the engine block, where the coolant's exit is located. This means that the heat of the first cylinder will heat up the coolant before it enters the following cylinder.

The second is transverse coolant flow, also called parallel flow. Transverse is the opposite to longitudinal, and is where the coolant flow enters the cylinder head at the side, and where the return is oriented at the opposite side of the head. In this case, the flow direction is perpendicular to the longitudinal axis of the engine block. In other words, all cylinders will receive the same temperature coolant.

What are the major differences between both concepts when focusing on cooling performance? In the longitudinal concept, the coolant is fed to the cylinder head through an entrance at the front of the first cylinder - or at the rear of the last cylinder - where it will flow through all cylinders to the opposite side. The consequence is that, at every cylinder, heat is transferred to the coolant, leading to a heat-up per cylinder, which is not very pleasant for the last cylinder.

The parallel concept typically requires a coolant gallery, mostly integrated into the crankcase, from which coolant feeds per cylinder are designed towards the cylinder head. For maximum cooling, the entrance of the coolant is often located just between the exhaust valves, where the cooling efficiency on the exhaust valve bridge is highest. Looking from the top side of the cylinder head, the coolant has three ways to flow, through the other three valve bridges. The goal is to design the coolant jacket such that all valve bridges receive just the amount of flow in order to achieve sufficient cooling. This is where Computational Fluid Dynamics (CFD) analysis is an absolute must in order to predict the flow pattern.

For both concepts a difficult area to predict and influence flow is around the centre of each cylinder. Typically, a dome is located just in the middle for either the spark plug or the injector. For the spark plug, material is required for its thread, where for an injector either a sleeve is assembled to keep coolant away from the injector, or the injector is assembled in a casted geometry, both restricting the coolant flow significantly. Besides being a major obstruction, the coolant is also forced upwards at this centre dome, meaning that the coolant is not at the location where it needs to be - at the flame deck! These are typical issues that design engineers spend many hours and iterations on to get the flows and speeds just where they are required.

From a production (casting) perspective, the longitudinal concept is less complex. This is due to the fact that the casting of the cylinder head coolant jacket can exist out of a single solid structure, without having to keep the cylinders separated. With the parallel flow pattern, the cylinders need to be separated from each other, in order to force the flow from exhaust to intake side and prevent the coolant flowing to other cylinders, and reducing intensity.

And as with all other conceptual choices, these concepts can be combined. There are designs using an upper and lower coolant jacket, where the lower one is used to create a parallel flow pattern and the upper jacket as coolant transport gallery to a single coolant return gallery. This enables a more compact design as a true parallel flow concept, and is not much more difficult than the longitudinal coolant concept.

Nevertheless, it will not be difficult to understand that purpose designed high-performance race engines will be based on the parallel or transverse cooling concept, making it possible to achieve the highest cooling capacity by individual adaptations to the coolant flow per cylinder.

Fig. 1 - CFD image of cylinder head coolant flow

Written by Dieter van der Put

But how often does that simple little gasket, costing little more than loose change, let us down and, seeing a gradual weep of oil, the unit has to be removed from the vehicle and put back on the bench?

At this point I expect to find I'm probably the only person to do this, that everyone out there uses brand-new seals and gaskets at every rebuild and that, consequently, nothing ever leaks. But somehow, on reflection, I think not. So when I met someone at a recent exhibition who talked about reusable gaskets in motorsports applications then I was immediately interested.

I have to start by saying that the idea of reusable gaskets is nothing new. Your grandfather, perhaps even great-grandfather, may have used solid copper head gaskets, which could be reused many times. Placing them on a flat steel surface and heating to cherry red with a propane blowtorch would, upon cooling slowly, soften and make them malleable enough to compress and seal the cylinder head once again - 'annealing' is the term metallurgists use here. Still used in some forms of drag racing, the galvanic potential between copper and aluminium (at 1.645 V) in the presence of the engine coolant can create severe corrosion.

However, the particular gasket design I was looking at was one specifically for a gearbox housing which, I admit, may not be the most arduous of applications. But when grubbing around in the paddock after, say, changing a set of gear ratios and trying to find a replacement gasket in the back of the trailer with oily hands, anything reusable is a godsend.

Consisting of a thin aluminium sheet laser-cut to shape onto which is bonded on either side a foamed NBR (nitrile butyl rubber) elastomer, the gasket, I was told, can be used and reused up to four or five times. When chemically blown onto the aluminium substrate and bonded to it, NBR has a temperature limit of up to 120 C. That's not particularly high as far as some elastomers are concerned, when some polyacrylates are good for 150 C and others much more, but if limited to the less arduous environment of a transmission casing then some of its other characteristics can be exploited.

In this case the property is one of recovery when, after removal of the clamping load, the elastomer rebounds back to 85% of its original thickness. Not only that but the design retains a strong torque retention in the clamping fasteners, therefore not requiring any re-torque after initial assembly.

While not always a good idea to keep on using gaskets again and again, a strong and reliable seal that does not depend on messy liquid sealants is attractive when oil is dripping down your arms.

Fig. 1 - A brace of motorcycle engine gaskets. While the copper one may be reusable, the other certainly won't be

Written by John Coxon

On one of these items, the valve seat area, will be the focus of this article, and not so much the valve seat ring as a component, but as it is installed in the cylinder head and the relation to the cooling of the firing deck.

Assuming a certain experience and familiarity with valve seat assembly, readers will be aware that valve seats rely on a certain level of press-fit in order not to fall out of the cylinder head when the engine is running. Depending on the material of seat ring and cylinder head, a certain oversize of the rings is determined.

To be able to withstand the stresses introduced by the pressed-in seat rings, the cylinder head is designed such that a certain wall thickness is available around the valve seat. Where the valve seats are closest to each other, these walls will 'grow' together. Logic says that coolant cannot be in the same place as material, which means that at these locations the coolant will cool the flame deck less effectively.

In a previous article, I noted that the areas between the valves/valve seats, the so-called valve bridges, are highly sensitive to TMF. Because TMF depends heavily on the temperature delta between cold (engine off) and hot (engine full load), where the 'cold' situation cannot really be influenced, development should focus on reducing the valve bridge material temperatures as much as possible.

There are several options to reduce temperatures, but in general this means:
" Achieving good flow conditions and speeds of the coolant through the cylinder head
" Transferring heat as quickly as possible from the heat source to the cooling medium
" Getting coolant as near as possible to the hot areas.

The first point will be discussed in a later article. The second depends mainly on material choice, where aluminium, with its thermal conductivity value of around 250 W/mK, will remain the material of choice. The third point is the topic of interest here, and contradicts the situation described above concerning the material between the valve seat rings.

A typical high-performance, four-valve cylinder head has a maximum area of intake valves for improved breathing, leading to seat rings nearly touching each other. In this case, the only way to get coolant as close as possible to the flame deck is to lower the height of the seat rings to the minimum, while still achieving sufficient press-fit. It is in this area of the cylinder head design where development effort and simulation techniques are used to their utmost to achieve the best compromise between port size, design robustness and cooling.

In race engines where valve diameters are restricted, other possible improvements could be achieved by designing the coolant jacket in between the valve seats, as far as the casting tolerance and accuracy of pattern making allows.

The image here shows a scanned image of a cylinder head (exhaust ports in blue) with the coolant jacket in between the valve seats. On the opposite side (intake in red), clearly there is no room to design in a coolant jacket. Given the thermal loading profile of the fire deck, the significance of coolant on the exhaust side is apparent.

One other interesting 'solution' needs to be mentioned, albeit only to illustrate the drive towards cooling of hot areas. Down the years, several designs have seen daylight - some even patented - which allowed for direct cooling of the valve seats. The general idea was to design the valve seat pocket in such a manner that openings were created, either by clever casting or machining, where coolant could directly reach the seat ring. Without having personal experience of these types of designs, most probably the design/manufacturing complexity and lack of stability of the seat ring area explain why these developments have not become common practice.

Fig. 1 - Scanned image of cylinder head design

Written by Dieter van der Put

So too are the differences between hypoeutectic alloys of aluminium and silicon and their hypereutectic alloys.

Aluminium casting alloys used in many race engine cylinder blocks are generally a mixture of aluminium and silicon with smaller amounts of copper, magnesium or nickel. As we add more and more silicon to the molten mix the silicon dissolves, but once the amount added exceeds 11.7% - the eutectic point - upon cooling, the silicon will separate out at the crystal lattice boundaries as hard particles in a soft aluminium matrix.

Aluminium containing amounts of silicon less than this level are known as hypoeutectic alloys, while those with amounts greater than 11.7% are known as hypereutectic. And while these hypereutectic alloys have lower thermal expansion coefficients with slightly greater strength, until relatively recently suitable examples of use were hard to find.

Adding silicon to aluminium improves its fluidity in the molten state. This makes high-silicon alloys relatively easy to cast, and the hard silicon crystals precipitated out provide a source of excellent wear resistance, enabling pistons to run directly in an aluminium block. Low-pressure die casting techniques to control the flow of the molten material in the mould are not cheap but do allow the use of sand cores for water jackets and thus produce structurally stiff closed-deck cylinder blocks.

Furthermore, a carefully designed cylinder sleeve cooling system can control the silicon crystal precipitation process to produce small uniform crystals around the cylinder bore to give optimum surface characteristics once fully machined and honed. This honing process has to be augmented with an additional 'etching' process to expose the hard silicon-based particles from the 'sea' of soft aluminium, but altogether the resultant surface is highly wear-resistant, unlike hypoeutectic materials.

Previously, low-silicon hypoeutectic bores (LM 25 for instance) would need to have been coated with silicon carbon particles in a nickel matrix either plasma sprayed or electrolytically deposited on the aluminium bore. Producing a coating of 0.003-0.006 in thick, this can make subsequent machining of the top deck or around the ports on a two-stroke unit very difficult when the plating can break away if not treated carefully.

Also, since nickel in the workplace on health grounds is now considered to be less than desirable, manufacturers are looking towards alloys of aluminium that contain the silicon particles but without the nickel. In this instance, a 17% silicon alloy with such silicon added in the form of sand (silica) would seem to offer a good overall prospect for the future.

Hypo or hyper, the difference may seem very slight, but when referring to the surface of the cylinder bore, the way to achieve a wear-resistant surface is so very different. However, it is surely when at the repair stage that the problems of identifying the correct material, and hence the correct surface treatment method, may raise more than just a little titter.

Fig. 1 - Typical crystalline structure of a hypereutectic aluminium-silicon alloy (Courtesy KS Aluminium-Technologie)

Written by John Coxon

To call it a ring is perhaps not totally correct. For although most are circular in both section and plan view, 'O' rings are perhaps better referenced as a torus, a shape formed by rotating an enclosed section about a line in the same plane but not intercepting it. And although generally circular in cross-section and made from elastomer-type materials, 'O' rings can be used to seal circular holes as well as many irregular but planar apertures.

'O' rings can be described by three main characteristics - outside diameter; inside diameter which, assuming a circular cross-section, specifies the cross-sectional diameter; and the material from which it is made. Held in a groove of specified dimensions, sealing is achieved by the compression of the section against the fourth surface, which is clamped to the other by mechanical means. The circular cross-section therefore distorts, ensuring the maximum sealing pressure with the minimum of clamping load. With an 'O' ring the maximum sealing pressure is equivalent to the pressure created by the seal deformation at the contact point.

Whereas the external and internal diameter of the seal is normally dictated by the size of the hole to be sealed, the thickness of the ring is to some extent quite arbitrary. A smaller cross-section will be more compact and lighter in weight and, being smaller, inevitably less expensive, particularly if the more expensive elastomer materials are used. On the other hand, a larger 'O' cross-section will be less prone to compression 'set', less prone to swell characteristics of the elastomer and will allow for larger machining tolerances on the ring groove. So, for a race engine, smaller is not always necessarily better.

That other characteristic, the elastomer, is defined as either a synthetic or natural product that has enough resilience to return to its original shape after some level of distortion. It is this property alone that makes them work as seals. The design parameters that reflect this ability are referred to as compression squeeze and compression ratio.

Compression squeeze is the difference between the compressed thickness and the original cross-sectional diameter of the seal, while the compression ratio is compression squeeze expressed as a ratio with the original cross-section, and is essentially the elastomer compression expressed as a percentage of the seal cross-section. Compression squeeze is expressed as a minimum dimension (in mm or in) while compression ratio is generally desired to be between 5% and 20% at all times. The next major decision is the choice of elastomer material.

Elastomers have to be selected according to their operating environment. In an engine, therefore, the ability to withstand temperatures of up to 200 C and be unaffected by modern engine oils, fuels (ethanol in particular) and water-glycol mixtures are paramount. At high temperatures, rings can harden and crack, while engine oils and hydrocarbon fuels can create some level of seal swell in most elastomers. Some elastomers are highly resistant to hydrocarbons, while others are easily attacked by acids or alcohols (methanol/ethanol). Some types are also affected by the glycols used in cooling systems.

Selecting the correct elastomer is therefore critical. Fluoroelastomers of the FPM category (of which the trademarked name Viton is an example) would seem to have the best overall properties at a reasonable price. For greater temperature performance FFKM (Perfluoroelastomer) elastomers might be slightly better but cost more.

Simple it may be, but the 'O' ring is not without its own share of design issues and potential pitfalls.

Fig. 1 - The 'O' ring
Fig. 2 - Compression squeeze

Written by John Coxon

As mentioned previously, a cylinder head is loaded in many ways. In general, there are three different loads - assembly load, (peak) firing pressure load and thermal load. A combination of these can lead to failure of the cylinder head, TMF.

What usually happens is that the heat loading of the cylinder head leads to expansion, which cannot take place fully due to the constraining cylinder head bolts (or studs) with which the head is mounted to the block. Constraining the expansion will lead to internal damage of the material (plastification by compression) which, when the engine cools down, is transferred into tensional stress.

When this cyclic loading of compression and tensional stresses exceeds the maximum permissible tensile stress limit of the cylinder head material, cylinder head cracks form. Most cracks occur in the valve bridge area between the intake and exhaust valve seat rings, partly because of the minimal material width of the valve bridge but mainly because of the high degree of constraint in the longitudinal direction of the cylinder head. Typically the transverse direction has less constraint due to the lack of adjacent cylinders in that direction. This, in summary, is the failure mechanism behind TMF.

Besides the design freedom concerning cylinder head material, there some other areas of freedom in designing for robustness against TMF. In this article the influence of the design of the cylinder head coolant jacket will be further clarified.

Let us look at how a coolant jacket is designed in the first place. A coolant jacket in the cylinder head is nothing more than a consequence of the rest of the component arrangement in the cylinder head.

Take the example of a common four-valve DOHC cylinder head. The double overhead camshaft automatically leads to a parallel valve pattern, meaning the intake and exhaust valves are oriented in the longitudinal direction of the cylinder head. Mostly the in- and exhaust ports are in this case are designed opposite each other, and a logical choice for the number of cylinder head bolts is four or six, depending for the most part on bore size and block and head structure.

Other components of interest are the valve guides and, even more so, the valve seat rings, in order to have a robust valve seat area. And last, but not least, the spark plug needs its place, almost always located centrally between the four valves and protruding perpendicularly into the combustion chamber.

Take these structural boundary conditions into account and one can see that the available space left can be used as coolant jacket. Therefore, the only real design freedom is where to locate the intake and return of the coolant jacket, the choice between a single or double water jacket and the choice between a cross-flow or longitudinal flow of the coolant. Due to the heat rejection of highly loaded race engines, most designs incorporate a cross-coolant flow, meaning flow from exhaust side to intake side or vice versa.

Having said this then, what is the relation between coolant jacket and the mechanism of TMF? To answer this, an additional constraint needs to be explained, that of the flame deck of the cylinder head.

The flame deck is the side of the cylinder head side that is in direct contact with the combustion chamber. It needs a certain thickness (strength) to resist the combustion loading, or so-called peak firing pressure. Roughly speaking, the peak firing pressures of race engines will lie in the range between about 100 and 250 bar, varying between naturally aspirated and turbocharged engines, and both diesel and gasoline.

The combustion loading will try to push the flame deck upwards. In order to withstand this deformation, the deck needs to have a certain minimum thickness. However, to reduce the risk of TMF, the temperature of the material needs to be kept as low as possible, meaning as little material as possible between combustion gas and coolant jacket.

This is exactly where the conflicting demands on the flame deck can be seen. Typically the flame deck thickness is a consequence of the diameter and height of the valve seat rings, requiring sufficient base material around them to achieve the required press fit. Between the seat rings in the tightest section - the valve bridges - there is often insufficient coolant available to reduce the risk of cracks. As long as these cracks do not progress into the water jacket, there is no problem. Otherwise, however, coolant will leak into the combustion chamber.

In the next article I will take a more detailed look at TMF and cylinder head design.

Fig. 1 - RVX-09 coolant jacket

Written by Dieter van der Put

In order to design and develop a new unit, as well as working out the best options in any semblance of order, two years is perhaps pushing things a little, so 2013 seems rather optimistic now. Gone are the days when the designer can scheme out an entirely new engine in his front room in six months and expect it to perform straight out of the box, as in the case of the DFV.

But setting aside the difficulties of getting an engine up and running in time, the chatter on some Formula One websites has for some reason been all about diesels. Let me say at this point that I doubt very much whether Bernie would agree to anything resembling a diesel engine anywhere near the grid of a Grand Prix but the presence of Porsche/Audi throughout the technical discussions does instil some degree of doubt.

It is of course most likely that the Porsche brand may be seen back on the grid in future years, but if we consider that the Formula One engine of the future is all about technology then I can't really see why Audi might not want to be there either. Manufacturers of diesel engines they may be, but the technology of the diesel engine has come on a long way since the mid-1990s.

Common-rail diesel combustion pressure of more than 200 bar has led to the development of thin-wall cylinder blocks carrying parent metal liners which are not only much lighter than previous cast-iron cylinder blocks but, as a result of improved FEA techniques, are much stronger too. The cast iron in question is called compacted graphite iron (CGI) and has been so successful in modern diesel units that its use - or so I am reliably told - is now being examined for gasoline engines.

For race engines the material of choice is usually aluminium. Cast as an open block with plated aluminium liners, the structure loses some of its stiffness. Cast with parent metal bores in situ, some of this stiffness can be reclaimed but the strength/stiffness of aluminium, particularly at elevated temperatures, can still leave much to be desired. CGI blocks, however, especially with integral cast liners - and especially when cast down to 3.5 mm section width or less - can be so much stiffer, retaining the shape of the bearings and the cylinder bores better than aluminium ever could.

The particles of graphite appearing as individual 'worm-shaped' or vermicular particles in the iron along the cylinder bore surface can also be run directly against most ring plating technologies to give excellent wear characteristics. For the ultimate, skeletal blocks using CGI for the load paths cast into aluminium 'skins' are also being investigated apparently.

In the1980s, BMW used an iron block for its 1500 cc Formula One turbo unit. Producing a reputed 1150-1200 bhp at a little over 5 bar qualifying boost, these units added a spectacle to the grid rarely seen before.

If, as we are led to believe, the shape of Formula One is to be four-cylinder 1600 cc turbos running at something around 3 bar boost, then the strength in the cylinder block will count for a lot. While it remains to be seen if CGI is the optimum route, developments would at least be relevant to the roadcar industry.

Fig. 1 - Inline CGI block

Written by John Coxon

Large leaks - whether oil or coolant - are, funnily enough, usually the easiest to solve, where a loose bolt or a forgotten hose clip are generally quickly dealt with. Worse by far though are the slight 'weeps' - the leaks that appear out of nowhere, travel down to a joint, run across the engine and down to the sump before finally spoiling your spotlessly clean drip tray underneath. Apart from the risk of fire, such a leak is annoying in a test cell, and unacceptable in a vehicle.

To the inexperienced it may look as though a gasket is not seating properly or that an 'O' ring has become dislodged. By now, however, the more experienced will be far less optimistic and will be suspecting a range of other causes - of which the worst possible is casting porosity.

Thanks to modern CAD methods and mould-flow casting programs, the integrity of metal castings is now better than ever. Low-pressure casting techniques, improved metallurgy and better understanding of the processes involved have combined to produce components that are approaching the quality in some cases of that of billet machined parts. For production parts, the skills and efforts required, as well as the costs - together with some trial and error - will eventually produce a casting free of porosity.

But for prototypes with limited budgets, for most of us the only sensible approach is one of caution and to seal any potential porosity from new. One such process used for almost 50 years is that provided by Impregnation Services of Lancashire, England.

The process starts with gently heating the bare casting - iron, aluminium or whatever -to about 120 C for several hours, to drive out any moisture. The dried components are then placed in a large autoclave and the pressure reduced to 10 mbar absolute or below. This has the effect of evacuating all the voids within the material in readiness for impregnation with a very low-viscosity (less than that of water) methacrylate liquid resin.

With no air present the resin has little option but to fill any voids in the casting, although in certain cases a pressure cycle of up to 7 bar can be introduced as well to consolidate it. Drained and centrifuged to recover as much of the surplus resin as possible, the components then go through a two-stage washing process to emulsify any remaining surface product.

After this, the resin retained in the casting undergoes a thermal cure in a hot bath at 95 C. Referred to as a thermosetting resin, the temperature triggers the catalyst in the resin to cross-link the polymer chains, producing a hard, glass-like substance with excellent sealing properties.

To be 100% effective, however, the casting should be treated from new. While there are techniques to remove contaminants such as oils or other fluids, successfully purging them cannot always be guaranteed. Nor can the process bind cracks together. In this case the component should be welded first and then returned for impregnation afterwards just to be sure.

When you consider how much time and effort goes into producing prototype parts, however, it is surely a false economy not to impregnate such items as a matter of course - even if only to let the test engineer sleep soundly at night.

Fig. 1 - Casting porosity

Written by John Coxon

First let's look at some of the key criteria of a cylinder head. It has a number of functions, such as keeping gas and coolant inside, and providing the structure for assembling or bolting on other components - valve seats and guides, for example, but also camshafts. To do so, the cylinder head needs to withstand a number of different loads. During tensioning of the cylinder head bolts, assembly loads are put into the head structure. When the engine is fired up, the cyclic loading of the combustion forces (peak pressures) are added.

And last, but certainly not least, another significant load is the thermal loading due to the heating up and cooling down of the engine. The severity of this last load is often underestimated.

It doesn't need further explanation to say that the flame deck of the cylinder head will see the highest combined loading. The flame deck temperature will force the head to expand and deform into the combustion chamber. The higher temperature in the area of the exhaust valves leads to non-homogeneous deformation. Combustion will deform the head in exactly the opposite direction, leading to a constant cyclic deformation.

These loads can result in cylinder head flame deck cracks, especially in the valve bridges - the area between the ports which, because of their relatively small cross-sections, are most sensitive to damage.

So how does this mechanism work?

The difficulty with cylinder heads is that the head needs to be mounted as rigidly as possible to the crankcase to be able to withstand the combustion peak firing pressures, and prevent leakage of combustion gases or fluids. By doing so, the material is being limited in its ability to expand freely under thermal conditions.

Constraining this material when it's heating up will lead to thermal stresses in it. Up to a certain limit, this will lead to elastic strain, but when exceeding this limit, compressive plastic strain will occur. When the engine cools down the same happens, only in the opposite direction, leading to tensional stresses in the already plasticised areas. These stresses could again exceed the tensional yield limit, again creating plastic strains.

The material gets damaged from cyclic plastification, called thermo-mechanical fatigue (TMF). A number of factors directly influence the amount of plastic strain per cycle, namely the level of constraint, the highest material temperature, the time history of the thermal cycle and of course material properties. These plastic strains will increase from cycle to cycle. Due to its location and shape, the highest value will occur mostly in the intake - exhaust valve bridges, locally at the exhaust port side.

Some of these can be influenced by the design. The thermal conductivity, in combination with the coolant jacket layout, will determine the maximum temperature at the flame deck. In most race applications, aluminium is the material of choice, mainly because of its excellent thermal conductivity and the ease with which complex geometries can be cast in the water jackets.

The level of constraint is more difficult to control. Typically, a cylinder head has intake and exhaust ports, a water jacket (single or dual) and a number of bolts to mount the head to the block. This leaves almost no freedom to influence these constraints. Port orientation is normally the consequence of the valve train layout. The typical layout for race engines is a longitudinal port orientation, meaning intake valve on one side and the exhaust ports on the other side of the cylinder head's longitudinal axis, especially when using overhead camshafts.

In highly loaded motorsports applications in particular, with their many temperature cycles (engine on/off) - as in diesel road race engines - the loading can become too high for aluminium, leading to cracks in the flame deck. As long as these cracks do not reach the coolant jacket it will not be considered a problem, but if it does then a DNF cannot be avoided. In these cases, cast iron is often chosen as alternative. Having higher strength, grey cast iron is more robust against thermal fatigue cracks, although at the penalty of higher mass and worse heat conductivity.

Although not widely used at the moment, CGI can also be considered a very feasible solution. It has a significant higher mechanical stress than normal grey cast iron, and somewhat lower thermal conductivity.

In some cases, this might lead to a positive balance between these two major TMF-influencing parameters. In the early days, when CGI was far more expensive then normal grey cast iron, its use was not usually considered feasible. Nowadays, the costs of producing CGI have come down considerably due to the increased demand for this material.

CGI as cylinder head material is still something of an exception in race engine technology due to the fact that the combustion loading on cylinder heads is still rather modest, especially on gasoline engines, which means that aluminium will remain the material of choice. Diesel race engines, however, running significantly higher peak firing pressures than their gasoline equivalents, will most probably drive the developments for stronger head materials such as CGI.

And although, for example, the Audi Le Mans race diesels are running aluminium heads, for the lower-level race classes this development will be unaffordable. I therefore expect Compacted Graphite Iron to become one of the key cylinder head materials for high-performance (race) diesels in the near future.

Fig. 1 - Valve bridge crack between exhaust valve seats

Written by Dieter van der Put

For many years it has been accepted practice to locate wet liners via a recess in the cylinder block top deck and seal between the outer liner surface and the cylinder block casting at the base of the liner using a system of 'O' rings under compression. Clamped at the top by the cylinder head and fire ring, the liner is free to expand both vertically and diametrically, and yet still maintain a positive seal with the water jacket at both top and bottom. The seal at the top is maintained by the clamping load of the cylinder head pressing the lip of the liner into the softer surface of the aluminium block, while that at the bottom is classic 'O' ring practice.

The downside with this method is the restricted level of cooling around the top of the liner and the effect that this can have on combustion. Years ago, however, respected Formula One engine supplier Coventry Climax used another approach - at least for a short time.

In order to obtain maximum cooling around the top of the liner on its prototype FWMV 1.5 litre V8s, and create what is known as an 'open deck' design, the company located the cylinder liner in a spigot some way down, about 3in (75 mm) inside the cylinder block. Clamped against a ledge at this lower location and sealed at this point using traditional 'O' ring practice, the liner was clamped at the top by the cylinder head, combustion being sealed with a Cooper's ring. Using cast-iron liner technology within an aluminium cylinder block, all went well with the engine on the dyno.

The normal testing procedure with prototype engines of this type is for a progressive warm-up phase followed by lots of wide-open throttle testing at a constant temperature. The procedure gave no hint of the problems to come, however, and it was only when the engine was installed in a car that the issues began to show.

Although it ran perfectly well on the dyno, once installed in a car the engine would soon blow all its coolant out of the radiator header tank. Stripping the engine revealed no obvious failure, but in time the real reason for the phenomenon became apparent.

Locating the liner at the lower position allowed the effects of the differential expansion to affect the clamp load at the fire face. While the gradual warming of the engine on the dyno kept this clamp load under control, once installed in a car the rapidly changing thermal loads, the different expansion coefficients and thermal conductivity soon provided the exact conditions to take away this clamping pressure. Combustion gas then quickly found its way into the coolant system, pushing the coolant out of the header tank.

In this particular case, after much deliberation the eventual solution was to move towards a dry liner system using aluminium sleeves into which the turned-down iron liners were pressed. Effectively located at the top of the cylinder block now, coolant loss through the header tank was a thing of the past.

It was a hard lesson to learn, and one I am sure the engineers at Coventry Climax were never to forget.

Fig. 1 - The flawed design

Written by John Coxon

In the past, when oil was cheap and the environment was simply a word in the dictionary - or indeed, if we just liked the aroma of Castrol 'R' - stem seals, particularly those of the exhaust valves, could be removed and the car would circulate just in front of an oil haze. Punctuated by an additional blast of blueness immediately after every gear change, we all thought the friction saved at the seal-stem surface was worth real power and was evidenced by the fact that once in front, no-one else could catch us. Little did we understand at the time that the real reason that no one could pass was simply because no-one could get anywhere near because of the oil haze!

In reality, however, the power consumed by a typical valve stem seal is measured in Watts rather than kW, and friction values in the region of 10 W per seal for some of the latest OE engine manufacturer designs are typical. And while reduced friction (maybe up to 10%) can be achieved using PTFE coatings around the lip of the seal, the real benefits are more to do with oil consumption, particularly at high turbo manifold pressures or vacuum levels at elevated temperature.

But no matter how effective they might be, the barrier to using more bespoke competition valve stem seals is often cost. Roadcar engines now use smaller valve stem diameters, and when oil consumption is as critical (as it is these days) to achieving zero oil top-up between servicing in passenger cars, emission-critical components like the stem seal have to be of the highest quality. And when made in thousands to the strictest quality control, it is easy to see why many race engine designers opt for this route.

It is therefore only when running conditions are outside traditional roadcar operating envelopes that bespoke seals may be needed. Roadcar oil seals may be made from any of the common elastomer compounds - nitrile (NBR), polyacrylic (ACM) or silicone (VMQ) - but are more likely these days to be made from any of the modern FKM/FFKM types of fluoroelastomer. Bespoke race engine stem seals, however, are more likely to be made from any of the PTFE products, which accounts for their additional stiffness and higher temperature performance characteristics.

And as for that oil haze of the past, much of it could have also come from the missing oil control rings, which were removed just to see how much more friction could be saved. If a two-stroke could get away with only one ring, then why not a four-stroke?

Well, you simply have to try these things!

Fig. 1 - Cross-section of a valve stem seal

Written by John Coxon

What I did not mention though - and this is where this article connects to the previous one - is the fact that fracture splitting cannot be done with every type of cast iron. Based on the process-specific requirements, fracture splitting is possible when using Compacted Graphite Iron (CGI) material. So would there be more to gain using this material, and what are the limitations?

CGI is a cast material that sits between ductile and gray cast iron, and which has better thermal conductivity and fatigue properties than gray cast iron. In particular the higher temperature fatigue limit of CGI, being higher than aluminium, makes it the 'perfect' material for blocks and heads.

For highly loaded engine blocks, the main advantage is the strength of the material, making it well suited for optimising block geometry, given the focus on reducing wall thicknesses. Some studies say a mass decrease of up to 20% should be achievable. Whether this number is correct or not is not the issue here, but it shows the potential of the material.

To achieve the maximum gain in weight, the design engineer will always try to reduce the wall thickness as much as possible, until either strength or process stability prevents him from going any further. Based on the mechanical properties of CGI, which has a tensile strength typically in the region of 450N/mm2, the wall thickness is not normally limited by the material but by the casting process.

Currently, nominal wall thicknesses can be cast at about 3mm for a typical crankcase size, while still allowing some tolerances on the pattern side and to provide enough 'path' for the melted iron to flow through, without the risk of cold shut. In non-structural parts of the crankcase and other low-stress areas, such as engine skirts, it's desirable to reduce the wall thickness to less than 3mm. Compared with aluminium, this would lead to a competitive crankcase weight. Unfortunately though the casting process prevents the engineer from achieving this.

Would it not be ideal then if CGI could be used only in the areas where maximum strength is required, and where the casting process can achieve the required geometries - as in bulkheads, the cylinder liner and top deck? All other areas could be made in a more suitable material, such as magnesium or aluminium, perhaps even engineering plastic. There are some examples of these hybrid designs, such as BMW with its aluminium-magnesium hybrid block casting, and Volkswagen with the hybrid cast-iron structure integrated into an aluminium crankcase casting.

A known area of concern with these hybrid castings, however, is the connection between the cast iron and aluminium materials internally in the casting, where contact corrosion could occur locally. But let's assume that in the end this should be solvable, especially for non-long term racing blocks, making it a very mass-optimum concept.

Combining such hybrid engine block technology with the fracture-split main bearing cap and its ability to reduce overall length of the engine (see former article) would have the potential to achieve maximum weight reduction. Given the developments and trends towards an increase in diesel race series, where peak firing pressure is the key to high performance, these series would benefit the most from the hybrid CGI application. The current regulatory discussions around downsizing and the use of turbocharging could also prompt a move towards these kind of hybrid crankcase designs, using CGI as skeleton within the crankcase.

But the question of course will be if developing such technologies is affordable in a time where costs need to be maintained at acceptable levels.

Fig. 1 - Structure of Compacted Graphite Iron

Written by Dieter van der Put

There are exceptions though. Hypereutectic aluminium alloys (aluminium alloys containing more than 12 % silicon) for instance can be used successfully in cast-iron blocks where the lower rate of expansion is comparable, and of course cast-iron liners can be used successfully in aluminium blocks. But there is one material that is regularly used in many engines, particularly diesels, and which is only seldom used in competition engines - steel.

In general, cylinder liners don't fail from lack of strength. So when you look at the properties of a typical liner steel it is often difficult to understand why you might want to use it. Grey cast iron has a tensile strength somewhere around 400 N/mm2, and very rarely have I seen a cylinder liner break. Certainly I've seen them shatter when hit by the occasional stray connecting rod but I can't say I've ever seen them fail through overload.

In terms of stiffness, ductile iron is getting on for almost twice that of grey cast and is only slightly less than that of the steels specified. So if stiffness is the benefit, why doesn't everyone else use it?

On closer analysis, however, and when seeking advice from those in the know, the real benefit would seem to be the hardness of the steels and the uniformity of the grain structure when compared to that at the surface of cast or ductile iron. Essentially low-carbon tool steels they consist of small amounts of chromium, nickel and molybdenum uniformly distributed throughout an iron matrix. The surface of the cylinder bore is therefore much better at accepting wear resistance coatings than grey or ductile iron it replaces.

So when lubrication is marginal, usually around the top ring reversal point in highly loaded engines, the nickel silicon carbide, Nikasil-type coating will be more coherent and less likely to fail than a similar coating on a grey cast or ductile iron liner. The harder the surface, the more wear-resistant it will be. This could allow the use of more aggressive piston ring technology without the horrific wear that would otherwise accompany it.

Used only in cast-iron blocks, for applications such as drag racing, steel cylinder liners made from chromium nickel molybdenum alloy tool steels do seem to make sense. Having made a considerable investment in a cylinder block in terms of detailed preparation, the last thing you might want to do is scrap it because of worn or damaged bores. Reclaiming it using a low-carbon steel liner can bring it back to life again and create a reliable low-friction surface at the same time.

Steel liners have been used for a long time and are nothing new. Used in heavy-duty diesel engines as well as many piston aircraft, when fully developed they can give thousands of hours of reliable use. But in competition engines they could be viewed either as a sign of desperation or the pinnacle of cast-iron block technology.

Fig. 1 - Comparison of the properties of iron and steel liners

Written by John Coxon

But of all the areas in a typical engine cooling circuit, the water pump bearing is probably the most difficult to seal reliably. Almost universally consisting of some form of mechanical face seal, designs of this type can be made up of up to ten separate components.

Although they are complex and often bulky designs, the costs associated with components such as the pump body, the springs, the elastomer bellows, the ceramic faces and clips to hold it all together cease to be an issue when mass-produced, as they can be amortised across the numbers sold. But when bespoke solutions for motorsports applications are needed, mechanical seal designs - despite performing very successfully - become an expensive option.

An alternative explored by one bespoke seal supplier was to use a design based around the traditional radial lip seal, which is more usually found sealing crankshafts or camshafts. The requirements for preventing coolant from escaping from the water pump are in some ways very similar, however, although in others totally different.

So while the elastomers could take care of shaft misalignment - however slight - the PTFE radial lips were found to have a limited life in this new-found environment. In the case of a Formula One seal design, a typical pump speed of up to 10,000 rpm and a coolant pressure of about 7 bar would see the PTFE coating on the lip completely worn through after no more than ten hours on the test rig. With engines having to last four races plus Saturday practice and qualifying, this didn't bode well.

The major problem was that, at the underlip temperatures experienced, the action of the glycol in the coolant and the silicones in the elastomers caused silicates to be formed on the sealing lip. Very hard and highly abrasive, the seal lip rapidly deteriorated as soon as these silicates were formed, a situation compounded by the fact that water itself is a poor lubricant.

After 18 months of development the resulting radial lip design is now surviving for more than 100 hours on the rig running at 14 m/s and a pressure of 5 bar. The manufacturer is of course refusing to give too much detail but it is understood that this twin-lip design uses no form of garter spring and seems to rely wholly on the elastomer for its sealing pressure. As well as being lower in cost and giving a marginal reduction in parasitic drag, with a radial section of only 4.5 mm and an axial width of 6 mm, it is also much smaller than the mechanical seal it replaces.

In preventing leaks it may not be the most dramatic way but when it comes to eggs I now prefer mine on toast.

Fig. 1 - Water pump lip seal

Written by John Coxon

The reasons for this are the higher mechanical strength of the material in relation to increasing combustion loads, and a broader availability of CGI combined with the ability to cast thin-wall sections. Examples of the application of CGI are Toyota, where CGI is used for its NASCAR engine blocks, and Hyundai, where in the late 1990s CGI was used as block material in its World Rally Cars.
What cannot directly be found in the available information sources though is the fact that CGI has further advantages over more traditional materials, such as aluminum or grey cast iron. One such advantage is the fact that CGI is perfectly usable as a way to optimise the main bearing cap design.

Look at existing CGI race engine blocks and you'll see that the main bearing caps have a traditionally machined split line. This split-line design has some consequences that are less than ideal. Typically the main bearing caps are made from a different material than the block.

One critical parameter in the area of the main bearings is the crankshaft bore. This needs to be highly accurate regarding roundness and straightness, so that it doesn't suffer from possible bearing damage due to misalignment. With a machined split line, therefore, a positioning feature needs to be foreseen between cap and block, where typically positioning pins or bushes are used.
Although this is a well established method, a slight clearance will remain, on which the bearing shell geometry needs to be adapted. Also, not only does the bolt force of the main bearing bolts need to carry the firing and inertia loads, it needs to provide enough clamping load to achieve the required friction in the split line to prevent micro-movements between bearing cap and block.

This is where the advantage of CGI its mechanical properties comes in. CGI's elongation properties mean it can very easily be fracture split, resulting in a 'perfect fit' between cap and block. For con rods this has become standard procedure, but not for blocks. There are some examples of fracture-split main bearing caps in the automotive industry, but apparently this has not yet crossed over into the race engine industry.

The form fit in the fracture prevents any movement, which enables lower bolt loading and therefore offers the possibility of reducing bolt size. And the material's strength has already made it possible to reduce general wall thicknesses. So with the additional advantage of a reduced bolt size, it now becomes possible to reduce main bearing cap width. This allows some additional improvement possibilities for the crankshaft web's strength (bore size, and therefore cylinder pitch, will not change).

The fractured-split line makes the use of a positioning feature obsolete, which enables better overall roundness of the main bearing bore without the risk of a possible 'step' at the split-line location. This gives better robustness against bearing wear in this horizontal plane of the engine where, depending on cylinder number and balancing level, typically transverse inertia loading occurs.
I am certain that some design improvements in this area will be seen in upcoming race block designs.

Fig.1 - Fracture split main bearing

Written by Dieter van der Put

Before we go on to talk about shape it is perhaps useful to look at it from a theoretical perspective. Students of metrology will be familiar with the terms roundness and cylindricity and, along with that of the diameter, these describe the geometrical condition of any cylinder bore. A highly magnified trace of the surface through a section of the bore, and at right angles to its axis, reveals that what might look perfectly round to the casual eye is in fact composed of a number of peaks and troughs of varying size and undulation.

If we accept that the mean radius of this is denoted by the symbol R0 then mathematically either dynamically or statically, the actual surface radius Ra, can be given by the expression:

Ra = R0 + where is the polar angle, k is the order and A and are constants.

The mathematically fluent among you will recognise the above as a Fourier series where the 'out-of-roundness' term, is the sum of all the Acos(k( +?)) parts from k = 1, 2, 3 and so on up to the value of integer n.

When k = 1 the bore shape is perfectly round. When k = 2 this becomes oval (picture (a) in the graphic). At k = 3 the figure moves towards a tri-lobe shape reminiscent of a cloverleaf (see picture (b)), while when k = 4 four lobes are now present (picture (c))

In theory, as n increases and all these increments are added together, the resultant shape (using the appropriate constants in each term) will describe precisely any cylinder bore you care to think of. In practice, however, although some experts suggest a limit of 5 or even 6, the more usual approach is to have a cut-off around k = 4.

At this point you might think these shapes are all very nice and produce pretty pictures. But when it comes to an understanding of them, each has a special significance.

When k = 2, for instance, referred to as second order, the ovality of the cylinder has much to do with the point at which the piston ring is touching the bore as it travels up and down. This reflects issues such as the thickness of the liner and how it is located (top, bottom or pressed fit).

When k = 3 (third order), excessive 'lobing' of the bore is a function of the imperfections as a result of the boring method and the equipment used. And when k = 4, experts tell us that shapes of this nature are caused by tensions in the cylinder block caused by fasteners and the like.

Thus when boring any block/liner, the representative loads caused by cylinder head studs or any other bolts need to be present. As a very minimum the main bearings should be torqued to the appropriate value while a boring plate of representative stiffness needs to be fastened to the top face.

So while bores can be measured with two- or three-point micrometers, to understand so much more requires a special machine called an incometer and some computer software. But we'll look at those at a future date.

Fig. 1 - Elements of the Fourier series representing a cylinder bore

Written by John Coxon

A crankshaft oil seal for instance needs to be 100% reliable under all conditions. This is because in the mind of the customer, it takes but a single drop of oil to constitute a failure. In the case of a turbocharger however, the shaft seals are altogether a different proposition.

"Turbochargers are uncommonly leaky devices," claims one turbocharger manufacturer. Rather than the compressor intake air or exhaust turbine side of things, he was talking in particular about the bearing housing and the passage of air or exhaust gas into the bearing housing or lube oil in the other directions.

When a turbocharger is working as it should, the gas pressure immediately behind the turbine wheel or behind the turbo compressor should be higher than anything inside the bearing housing. Seals are therefore generally required at each end of the bearing housing, and normally take the form of simple steel piston rings in a grove at both the hot turbine end and cold compressor end of the shaft.

These rings expand outwards against the bearing housing and with such a snug fit do not rotate. Since they form a very close fit in the groove in the shaft, hot gas from the exhaust or cool intake air from the compressor is therefore prevented from entering the bearing housing.

In high-boost applications, air may leak into the bearing housing, but provided this is not too great it will mix with the oil-air mixture already present and drain back into the sump. At the hot turbine end, however, excessive gas leakage into the bearing housing will rapidly degrade the lube oil and result eventually in increased crankcase/sump pressure if not adequately vented.

Under other, often transient conditions, oil can leak into the turbine housing resulting in smoke, high oil consumption and generally messy oil streaks. The seal at the turbine end is therefore considered critical.

In the normal piston ring design a gap of 0.008-0.013 in (0.20-0.33 mm) is produced to allow thermal expansion at the turbine side as the temperature increases. Over time and as the ring wears, this gap will steadily increase, allowing oil to pass through the void where the two ends of the ring meet.

To combat this, two new configurations of this simple ring design have been produced. The first, a so-called stagger-gap arrangement, effectively replaces the standard ring with a version, which closes this gap using the stepped-end approach (see figure). A simple idea but presumably difficult to manufacture, this is sometimes called a 360º turbine oil seal.

The second, an even simpler idea in some ways, however, is to double the width of the ring groove and introduce two normal piston rings with ring gaps phased at 180º to each other. Supplied for high-performance turbines, where high engine oil pressures could introduce oil pullover, this approach is surprisingly effective and uses standard low-cost piston rings.

So while turbochargers might still be regarded as uncommonly 'leaky' in some ways, it is not for the lack of sealing options.

Fig. 1 - The stagger-gap oil seal

Written by John Coxon

Ever since the invention of the 'heat' engine, limits to the properties of the metals used mean they have necessitated some kind of cooling. And since water is all around us, it makes sense to use this is a transport fluid to remove heat from the cylinder head and disperse it into the atmosphere.

But water brings with it certain problems. When the temperature of the metal being cooled is lower than the boiling point of the transport fluid, the dissipation of heat is relatively straightforward, consisting mainly of conduction aided by convention in the fluid. So the greater the temperature difference the greater the heat flow.

As the heat flux rises, however, and the localised temperature of the surface exceeds the localised boiling point of the transport fluid, then a phenomenon known as nucleate boiling occurs at the boundary. This is good, and can actually increase the heat dissipation at that point.

A full account of the situation is quite complex, but essentially the peaks of the surface at a microscopic level intensify the heat flux and promote a change of state of the transport fluid from liquid to gas. At a very localised level, this process absorbs heat in the form of latent heat and makes the heat extraction process even more efficient. So long as the coolant flows quickly and is replaced by a fresh supply, the heat flows through the boundary at an increased rate. The gas thus created moves away from the surface and condenses again dispersing the heat into the bulk of the fluid.

But there may be times when these small bubbles of nucleate boiling are not transported away quite so easily, for instance when the heat flux is far greater or when the flow rate of coolant is restricted in some way. In such cases these bubbles will build up and form a continuous layer across the boundary, effectively insulating it in a 'blanket' of vapour. This can cause hot spots, erosion and possibly even detonation in extreme cases.

Compounds such as ethylene glycol that are added to coolant protect the water jacket under freezing conditions, and which are added at up to 50% of the total mixture, can raise the boiling point of the coolant slightly, as can pressurising the system. But when really high temperatures are generated by high levels of heat flux, the water in the system will have to be completely removed and replaced by something with a much higher boiling point.

In Advanced Gas Cooled nuclear reactors, for example, this cooling medium is pressurised carbon dioxide, although the inert gas helium has also been tried. For our purposes however, that's a bit excessive, even for the most extreme engine applications.

No, the way to go is to use propylene glycol, another of the non-aqueous glycol products sometimes found in industrial process plants. With a similar density to water, propylene glycol has a boiling point of 187º C and a latent heat from liquid to gas 30% greater than that of water.

This is offset though by a 30% reduction in its heat-bearing capacity (specific heat) compared to water and its high viscosity, particularly at low (about 10º C) ambient temperatures. Adding other components to form an homogenous mixture can mitigate this, to the detriment of some of propylene glycol's other properties, but as engineers we are well aware that in order to get one characteristic we may well have to compromise with another.

So, if water can be the 'kiss of death' for our racing engine, propylene glycol and its additives could be the 'elixir of life'.

Fig. 1 - Comparison of coolant properties

Written by John Coxon

But are we making it as round as it could be? Take for instance the typical case of a replacement dry liner. We'll measure its external diameter in at least three positions around its axis, and then again in another three places up and down the bore. Averaging these and allowing something like 0.001 in per inch diameter of interference, we'll happily bore away, taking care of squareness and roundness of the boring head until we finally have any number of liners firmly seated in the block. This, however, is only the starting point.

The thin metal liner may have been distorted from careless handling (such as leaving it lying on its side for a long time instead of upright) but no matter what its initial shape, in the end it will conform to the shape of the hole into which it was pushed. Too little interference and the thing will just rattle around in use; too much and the surface could end up being convoluted and even more distorted. Ever tried to get one of those plastic wiring grommets through a hole that was too small for it? Well the situation is much the same here, except that the forces are so great that in the process the edges of the liner could even be damaged.

Often it's a case of 'suck it and see' and for each engine the machine shop will find a way that works for them. But whichever way you look at it, further machining/honing will be necessary.

For anyone concerned about the clamp loads from the cylinder head, a boring plate will be needed. This consists of a metal plate of equivalent stiffness to the actual cylinder head, with holes drilled through it above the bores to enable access by the boring bar/honing tool, and should be clamped into place - ideally using the similar bolts to those used in service - and tightened into yield using the same torque and angle values. This should simulate the loads introduced at build.

In addition, and to get even closer, a similar gasket/fire ring arranged should be inserted to replicate the real geometry at the head face. While many good quality machine shops will use a simple 1 in-thick boring plate, it is always difficult to know how representative such a plate might be. In using similar geometries and actual components as far as is practicable, the loads should be closer to those encountered in use.

But what about the thermal input, I hear you ask? Won't that cause distortion as well? Unsurprisingly, the best machine shops have thought about this and it's not uncommon for them to use hot oil circulating through the cylinder block during final machining. The oil will allow temperatures of up to 150º C, perhaps more if you really want, but I would have thought 100-120º C would be enough of a problem - after all, some poor machinist may have get close to it at some time.

But despite all this, are we not forgetting the effect of combustion pressure on all this distortion? Perhaps we need to introduce nitrogen at high pressure to the cylinder bore which, when sealed top and bottom, will use a specialised boring head. I haven't heard of anyone doing this as yet - although that doesn't mean it hasn't been done - and the pressure couldn't be anywhere near that encountered in a fired engine. Hang on a minute though, am I not taking this to the extreme?

Fig. 1 - Bore preparation the traditional way

Written by John Coxon

Chief among these is Federal-Mogul, which has just introduced a 'new approach' to lip seal technology, claiming that it reduces friction by up to 70% compared with conventional technology. Although a slightly revised version of its own-developed fluoroelastomer has been used, friction is solely the result of a new 'dual hinge' concept that ensures the seal tip can more faithfully follow the contours of the shaft at speeds of up to 8,000 rpm.

Larry Brouwer, director of sealing technology and innovation at Federal-Mogul Powertrain Sealing and Bearings, explains, "There are currently two types of seal types for light-duty use. The first uses a sprung elastomer with a helical spring holding the lip against the shaft, while the second is a 'laydown' type seal." The laydown type is favoured in Europe and its construction produces a much lower sealing pressure on the shaft. Already a low-friction component, it is to this latter type that the 'duel hinge' concept applies.

These two hinges are, first, at the point where the elastomer (shown in green in the graphic here) moves away from the steel backing and, second, at the 'U' bend where the elastomer turns back towards the fluted section. The fluted section in the diagram is the scroll, which returns any excess oil that may have reached the outer part of the seal, back to the sump.

In any rotating seal design it is imperative that the lip of the seal conforms to the shape and position of the surface of the journal at all times. Thus the out-of-roundness of the journal, the clearance between journal and bearing, and a host of minor manufacturing imperfections, all have to be accommodated. In compensating for these it is tempting to have the seal grip 'tighter' but as we know this leads to friction and heat - the biggest enemy of all.

So the real task of the designer is to accommodate all these with the minimum of friction but without compromising its performance as a seal. And while all this is difficult enough, the greatest challenge is to accomplish it at high engine speeds. At such speeds the hysteresis in the elastomer is likely to encourage seal and shaft to part company, as well as encourage higher temperatures that can carbonise the oil, forming ash in a critical area. If it's left in place for long this ash will act as a form of grinding paste and quickly destroy both the seal lip and crank surface finish.

But in optimising the friction in the oil seal it would be easy to forget that the same has to be done for the outer dust seal, since although its function is slightly different, the same issues apply. And just like a design of which our race engine designer would be proud, all this has been packaged such that it needs only 5mm of axial space on the crankshaft.

To get to this point, the latest in finite element techniques has been adopted and, using a sophisticated 'design of experiments' approach together with a compliment of 25 different tests, seal friction, according to the company is now probably about as low as it can get.

And when you think that a single drop of oil in 150,000 miles would constitute a component failure, the stakes are pretty high. Often we're so used to thinking that motorsports has the greatest challenge but with the push towards ever greater fuel economy, road transport could be seen to be leading the way here.

Fig. 1 - Crank seal

Written by John Coxon

Take for example, the instance of a poppet valve, its guide and seat. If the seat is not concentric to the guide bore, every time the valve closes a bending load on the valve stem will be created. This will eventually fatigue the valve and its head will break off. And so while everyone is looking at the valve, the real culprit could be that of the concentricity between the guide and seat.

During initial manufacture, the locating bore for the guide and that for the seat insert will most likely be machined off the same centre. When this is the case the concentricity of the guide hole and that for the valve seat insert should not be an issue. However variations in the concentricity between the outer diameter and inner diameter of the guide and the way it is assembled in the head as well as those associated with the valve seat insert, will no doubt, introduce a level of uncertainty, between the bore of the valve guide and that of the seat when the whole lot is assembled. And unless the strictest of geometrical limits have been used, the alignment of the valve seat on the insert and the bore of the valve guide will have suffered to some degree. To correct this, it is normal to finish machine the valve seat using the bore of the valve guide as a datum.

One way, is to individually 'clock' each valve guide bore to the concentricity of the cutter spindle. Machining the valve seat thereafter will produce a high degree of concentricity. Potentially the most accurate method of all, the concentricity here is down to the skill of the operator, but with multi-valve heads this is surely only rarely practical. However most people, especially those in the re-manufacturing business, prefer to use other, much quicker means.

When it comes to re-machining valve seats there are basically two schools of thought and both of these surround the method of identifying the machining 'centre'. The first of these is sometimes called a 'live' pilot and fixed to the machine spindle but floating in the valve guide, the tooling cuts the valve seat while spinning in the valve guide bore. The other way and that patented by machine tool manufacturer, Rottler Manufacturing of Washington USA, is to use what is known as a 'fixed' or dead pilot which, it is claimed gives a greater degree of accuracy over the 'live' equivalent. Essentially this system allows a tungsten carbide centralizing stem to locate in the guide bore but, importantly remain stationary while the valve seat is re-cut. Because the pilot is stationary, the clearance between the centralising stem can be less than that of the 'live' pilot such that the concentricity between the valve seat and the guide can be as low as 0.01mm (0.0004in).

But accuracy invariably costs and at this level, speeds will probably be around 10% or so slower than the 'live' pilot. But in the search for perfection surely this is of little concern?

Fig. 1 - Internationally recognised symbol for concentricity along with the limits allowed with respect to datum A on the drawing

Written by John Coxon

The bore of any port-scavenged two-stroke engine is always vulnerable. With the piston ring sliding over often cavernous exhaust ports, flanked by even more transfer ports, the opportunity for some form of breakdown in the lubrication between the ring and bore is always very likely. Scoring below the exhaust and transfer ports is highly likely while that above them indicates a more serious issue. In both cases however, loss in engine performance will almost certainly result and the only way to correct the situation is to re-plate.

After dismantling and cleaning (using hot immersion, jet spray cleaners or even ultrasonics), the existing plating has to be removed. Since this is only a few thousands of an inch thick machining it out is totally impractical. Although many re-platers prefer to keep this aspect secret, the nickel substrate in a nickel/ceramic surface can be stripped off using either a chemical or an electrolytic technique and sometimes a combination of the two. Thus while the nickel can be stripped using a Laybere solution consisting of three parts concentrated sulphuric acid to two parts water, nitric acid or a solution of aluminium potassium sulphate will do the job as well. Some will strip away the plating quicker than others but all are highly dangerous giving off toxic fumes and so need to be carried out by specialists with the correct equipment and fume extraction apparatus. Hence as the plating is removed, the aluminium casting remaining will be left totally unaffected.

At this point it is as well to remember that if any port re-profiling is to be undertaken, it should be done at this stage. Re-profiling a port after plating risks part of the plating coming away if done incorrectly and so at this stage it is advisable to remove any sharp edges and gently trim away any areas, which might cause the ring to 'pick-up' in later use. If the bore has been scored at some point and the score has gone through the original plating then since the plating process will not cover even the smallest of scores, the surface may need to be welded and then re-machined back to the nominal size. Methods of re-welding and the filler rods used depend very much on the original barrel material composition, but when fully washed and cleaned again, plating can commence. Although the plating of aluminium bores is highly competitive with few, if any organisation swilling to explain in detail, these processes generally consist of a homogenous mixture of nickel salts and finely divided ceramic powder flowing over and through the cylinder liner using a central anode, with the liner/barrel itself acting as the cathode. Deposition rates will vary according to the size of the component and the current drawn. In the end the coating will consist of two components: a hardened metal matrix into which is evenly dispersed (to about 12-15%), an extremely hard ceramic powder. The plating thickness is targeted at around 0.004in but this has to be chosen to suit the piston diameter and the clearance desired.

Because of the hardness of such a coating (apparent hardness - Hv 580-680), finishing will consist of a diamond honing process, configured to give a surface finish of 0.05 Ra.

So the next time you think about making a coffee table out of that old block, maybe a little bit of old-fashioned re-cycling is called for?

Fig. 1 - Plated bore

Written by John Coxon

Throughout history experiments have always taken place questioning why a purpose-designed seal for this part of the engine should ever be needed. Surely the narrow rim of a hard cast iron or steel cylinder liner, abutting the soft aluminium surface of the fire face on a cylinder head should be sufficient to seal the gap? And indeed sometimes this can be the case, but in doing so the damage caused visible only after dismantling can make the cylinder head unserviceable again. Way back, the 4-cylinder 2.5-litre Formula One Vanwall engine, instead of using a normal gasket, solved the problem using a separate sealing component around each cylinder bore. Designed and made by the Coopers Mechanical Joint Ltd in England, Coopers Rings as they were inevitably called, consisted of a pack of narrow steel rings, some flat and some corrugated, each 0.010 in. (0.25 mm) or 0.015 in (0.38 mm) thick, encased within a spun and folded outer casing. Made from stainless steel and Nimonic material, these were inserted into the groove formed at the top of the liner when it was located into the cylinder block. Rectangular in shape, these compressed from their manufactured size of around 0.126 in (3.2 mm) to 0.092-0.094 in (2.34 - 2.39 mm) when the cylinder head was finally fully 'torqued' into position. At the combustion pressures of the time, and apart from when the liners were bottom mounted rather than hung from a spigot at the top of the block, this design was reported to work with 100% reliability. So it is therefore little wonder that seals of this type were used and would continue to be used in many engines of the period. So in 1963 while Johnny Cash was 'ringing' out the first few bars of his song, the Coopers joint was winning all before it.

Nevertheless, the complexity of its manufacture and inevitable extra cost, eventually forced the business to close and made way for the cheaper and some say, technically superior, metal 'O' ring technology. Originally developed by the Wills family in Bridgewater, England and now known irrespective of manufacturer the world over as Wills Rings, these were, and still are essentially coated or uncoated thin tubes formed into circular rings. Made from copper, mild or stainless steel, and pressurized from within with nitrogen, stainless steel versions can withstand up to 800 deg C. For the most arduous of applications, Inconel is considered the best choice. Designed to conform into an oval, or more correctly, 'race track' shape in service, the circular cross-section makes it much more tolerant to the amount of crush necessary to ensure a positive seal around its whole circumference.

Today however, in many engines both historic and contemporary, the preferred material for any fire ring has to be solid copper beryllium, but this is one, I think, we'll leave for another time.

Fig. 1 - Cross-section of the Cooper's Mechanical Joint

Written by John Coxon

A considerable amount of any design and cost in any reciprocation engine goes into the cylinder block/crankcase assembly. Often split at the bearing line horizontally forming the face to which the sump is attached, as it was then, or more commonly now the bearing ladder, the cylinder block is a complex relatively heavy casting using many internal cores. Along with complexity comes the difficulty, not only in casting but also machining such that any traditional cylinder block is a considerable investment. This part of a prototype engine was therefore an attempt to move away from this complex and expensive approach having the main bearing split not horizontal and parallel to the cylinder head face as is normal, but vertical, running up through the cylinder bores and perpendicular to the cylinder head face. Such a concept, nicknamed quite appropriately 'the vertical split', greatly simplified the cylinder block casting although did introduce one or two other development issues.

In addition to the much simpler casting requiring no internal cores, for the first time the sump could be included as well. Using two identical crankcase halves which would be clamped around the pre-assembled cylinder liners/pistons and crankshaft, the engine was fastened together using 10 (it was a four cylinder) transverse bolts, one above and below each of the five main bearings. In between the cylinders the castings were bolted together using lots of tiny bolts while along the base of the sump the castings were clamped using long thin plates and bolts traversing the whole width of the unit. The two halves were sealed using anaerobic sealants, the first recorded use in engines to my knowledge, and the lower cylinder liner seal was injected after assembly using a 2 part specially designed sealant into holes leading to the machined grooves.

After a few development issues, some minor, others quite major, the engine ran for well over 300 very tough hours on the dyno. When stripped for inspection, despite some concern that the main bearing split line occurred close to the maximum main bearing pressure zone, all the bearings, bores and piston rings showed no obvious signs of distress. More importantly, we had proved that there is more than one way to cast a cylinder block!

Fig. 1 - The vertical split engine.

Written by John Coxon

However, inevitably when the glory days are over and the unit is discarded, perhaps left on its own in damp storage or bolted into the back of some historic racing car somewhere, another process is slowly taking over, more silently or deadly than the occasional missed gear change - that of corrosion. And while this can (and does) occur in many parts of the engine, the main area of concern is at the back of the cylinder liner inside the cooling water jacket.

Euphemistically now referred to as environmental stability, the causes of corrosion can be many and varied. We can have general surface corrosion, pitting corrosion, crevice corrosion, stress corrosion, fretting corrosion and some, maybe even all may be found inside a water cooling jacket of an engine. The common link between them all however, is water, but some materials are more prone to this type of activity than others. In the case of that most common of cylinder liner materials, cast iron, this process can be highly sensitive to the presence of water while others, aluminium for instance, apparently not so. The governing factor in all this is normally the position of the metal in the electrochemical series, but this is not always the case. For example, with its position higher in this table you would expect aluminium to be more readily corroded. This is not always so. When exposed to the air, aluminium rapidly reacts with the oxygen in it to produce a thin layer of aluminium oxide. Aluminium oxide has good adherence properties and is by and large impermeable to water and therefore will form a protective coat around the metal. The standard electrode potential of -1.66 volts being greater than that of iron at -0.44 volts therefore has little influence.

Perhaps the most common form of corrosion in the water jacket however, when the cylinder liner is cast iron or steel, is rust. Formed in the presence of both oxygen from the air and water, the iron atoms of the liner outer surface give up their free electrons and become positively charged (Fe2+). Electrons liberated so then migrate through the metal and into the water where they can combine with any dissolved oxygen to create negative hydroxyl (OH-) ions. These negative hydroxyl ions then combine with the positive metal ions to produce hydrated iron oxide or Fe(OH)2. This oxide, referred to as rust is somewhat powdery and easily crumbled and unlike aluminium oxide when coating aluminium, does not protect the metal from attack. If anything it actually promotes it by retaining the water close to the metal surface. In cast irons however consisting of a mixture of iron and carbon in the form of graphite, this effect can be much worse since the graphite can act as the cathode and the iron as the anode. In the presence of water the pair act as a galvanic cell and the iron corrodes producing rust and a phenomenon sometimes referred to as graphitisation.

In general however, the easy way to prevent or certainly reduce the effects of corrosion, is to use the correct inhibitor or anti-freeze solution if ambient conditions are likely to fall to freezing.

Fig. 1 - Corrosion in a cast iron block - the deadliest of enemies.

Written by John Coxon

That modern engine lubricants are a complex cocktail of chemicals, required to perform a range of functions, (lubrication, cooling, cleaning, friction reduction, protection, and so on) is most likely well understood. And while the base oil will constitute between 60 to 90 % or more of the overall volume, the remaining part, that of the additive pack, can have as many issues associated with it as the choice of base oil itself.

Base oils can introduce swelling or shrinkage to a seal according to the type of oil and the elastomer used but it is the additive pack that can sometimes have the most unintended of effects. One of the most common historically was the cross linking of nitrile rubbers by lubricants. The active sulphur, part of the extreme anti-wear agent, zinc dialkyldithiophosphate, can react with the double bonds present in nitrile rubber and other elastomers to appreciably increase this cross linking density. The result is hardening of the rubber and subsequent cracking.

Molybdenum disulphide, a common solid lubricant included in many greases cannot be introduced into traditional engine lubricants because of its insolubility. However, soluble versions of this product, again often used as extreme pressure agents, in the form of molybdenum dithiocarbamates or dithiophosphates can also cause cross linking in nitrile rubbers. In the case of dithiophosphates this can also lead to de-polymerisation of silicone rubbers.

Perhaps the most concerning in recent times, is the effect on fluoroelastomers by free amines present in modern day lubricant dispersants and some friction modifiers. These dispersants are present in the oil to prevent the agglomeration of degradation products caused by oxidation or thermal stressing of the lube, which might otherwise cause harmful deposits. Because of their thermal stability and durability, fluoroelastomers have become a desirable seal material in engines but even a small amount of basic amine present in the oil in the form of a dispersant can result in dehyrofluorination leading to premature failure. Base resistant fluoroelastomers have been developed to offset this issue, but in formulating oils and balancing the needs of long term stability, soot handling capability and overall engine cleanliness, taking into account both new and older seal materials, this is a task not to be taken lightly.

In many cases it has been up to the oil industry to develop solutions to these sealing issues but, in the case of RTVs or Room Temperature Vulcanising sealants, similar to those used commonly in engine build, the converse is true. First generation sealants of this type produced the by-product acetic acid, which leached out into the oil after curing. This reduced the base reserve in the oil designed initially to 'mop up' all the weak acids caused during the combustion process. RTVs developed for use in engines have now addressed this problem. However a secondary issue, the leaching of low molecular weight silicone plasticizers from the sealant into the oil is causing major concern. In the oil industry high molecular weight silicones are used to inhibit foaming in engine oil. Insoluble in the base stock in order to work effectively, they are dispersed through the oil by clever formulation. Silicone plasticizers however are soluble in the base oil and rather than working to prevent the formation of foam, actually stabilise it. In any engine particularly competition engines, the presence of foam because of these silicones can have potentially catastrophic consequences.

Fig. 1 - Highly Saturated Nitrile Rubber seal.

Written by John Coxon

A simple analysis of the rate of heat released of combustion plotted against the piston position will give an indication as to where and when this heat is released. With peak combustion pressures around 7-10 degrees after top dead centre much of this heat is created next to the combustion chamber (hence the name). Later in the expansion stroke the exhaust valve will open and the exhaust 'blow-down' thus created will divert part of this heat into the exhaust port. And while the cylinder bores and piston crown will take some of the heat away, by far the vast majority is conducted through the cylinder head combustion chamber walls. In the attempt to maintain a constant temperature across and around the whole engine the cooling system to minimize thermal stresses, all this has to be taken into account.

From that stated above the cylinder block and liners needs relatively little cooling. A small water jacket towards the top of the liner perhaps may be all that is required. However since there is generally a lot of space available down and behind the bore, designers tend not to skimp in this area and water jackets here are invariable larger than they truly need be. With the water pump positioned generally towards the front of the end, the coolant will be normally pumped from this position and along the cylinder block. Since water jackets are large and heat flux low, the coolant temperature will normally not increase much at all. It is only when it enters the cylinder head that things can get quite critical. If all the flow goes to the back of the block and then into the cylinder head for its return trip, the heat picked up along the return may cause combustion issues - detonation in the front cylinder. Likewise the water flowing down the intake side will be considerably cooler than that travelling down the side of the exhaust. To get round this, manufacturers tend to bleed off coolant from the cylinder block up into the cylinder head at various points down the block and feed the coolant from the cool inlet side of the head across and between the chambers and into a water rail across to the other, exhaust side. In this way temperatures along the cylinder head will be more uniform with the minimum of distortion. With competition engines, where the added complexity and costs can be accommodated, this exhaust side water rail may be external to the cylinder head itself.

Designing an efficient cooling system has never been easy but with modern computational methods it is so much easier to get it nearer the optimum if not perhaps quite so much fun.

Fig. 1 - Cylinder head water rail.

Written by John Coxon

At the outset of any engine design, in the search for efficient use of engineering materials, the designer has to package all the components parts inside the smallest of dimensions. In the case of the crankcase this means getting the largest size piston diameter inside the shortest of cylinder blocks and therefore keeping the bore centres (the distance between adjoining bore centre lines) to the absolute minimum. At first this might sound relatively easy and straightforward and in the early stages of a design when these critical decisions are made, such is the case. However later on, it is an undeniable fact of life that customers will eventually want a larger capacity or a reduction in weight and in either case, one of the options available to us is that of Siamese, or perhaps more politically acceptable these days, conjoined bores.

In the ideal world of engine philosophy, the surface of the cylinder bore should be surrounded by a uniform thickness of supporting material, which in itself will be fully surrounding by the cooling water jacket. Having a uniform thickness should ensure that, with the heat flowing out radially, the strains as a result of temperature variations will be minimised and the metal temperature especially that towards the top and between the cylinders, will be kept within reasonable limits. However, with unchanged bore centres increasing the bore diameter and yet retaining sufficient material around them for strength, a time comes when these two surfaces effectively fuse together forming a figure of eight shape where no cooling water can penetrate them. At this condition the bores are said to be 'siamesed' or conjoined. Whatever the term used, these bores are highly likely to suffer from severe distortion and high temperatures around the upper 'bridge' area which will limit the reliability of the gasket in this zone. In cast iron cylinder blocks where this design route has been taken, the lack of critical cooling in this area has caused temperatures to be as high as 360 deg C in this area at times. It is therefore little wonder that engineers try to avoid such design approaches whenever possible.

But to any rule, there would appear always to be exceptions. In this case it would appear to be the introduction of an alloy block Cosworth BDG in 1972 to the Works Ford rally Escorts. With the cylinder block in this engine based on the Ford 1600 711M component, the maximum safe bore size even based on the most carefully produced cast iron blocks, was 85.6 mm. This gave an overall capacity of 1790cc. Machining the blocks and inserting vacuum brazed iron liners would enable 90 mm bores and bring this up to 2.0 litres but even then distortion was an issue. A change to an aluminium alloy cylinder block brought about a saving of 40 lbs (18 kg) but handicapped with the same bore centres as before and the reduced strength of the cast aluminium, came the necessity to Siamese the bores as described above. However despite the potential issues outlined, the engines were surprisingly reliable.

Quite what the poor Bunker brothers would have thought we can never know. But as showmen to have such a technology named in their honour, I am sure they would have approved.

Fig. 1 - Siamese bores.

Written by John Coxon

In the world of engineering however, if something is continually presenting a problem, the first course of action is to get rid of it. Such was the thinking of Brian Hart in the Formula One turbo days when the monobloc Hart 415T engine was designed. Casting the cylinder head and crankcase in one piece certainly eliminated the head gasket but must have presented a much wider range of engineering issues in the process. History doesn't record how successful this exercise was but it certainly wasn't the first time it had been tried nor to my knowledge was it repeated thereafter. Around the same time and in the much less glorious surroundings, experiments were taking place using conventional aluminium wet liner technology to discard the cylinder head gasket in another way. Increasing the liner 'stand-off' above the top deck to greater than that normally used and at the same time increasing the cylinder head clamping loads, created a successful seal at the top of the liner past which combustion gases could not pass. To take care of the coolants and oil, the gap between the fire face of the cylinder head and the top deck of the block was bonded together using anaerobic adhesives with an 'O' ring seal for the oil feed transfer hole. Successful in prototype engines, the idea was never truly likely to catch on in the real world outside the R&D lab but the potential was always seen to exist even if removing the cylinder head thereafter was likely to cause all manner of issues.

We are perhaps all familiar with the older style metal/soft material gasket technology. A simple central supporting plate onto which is either side a thin layer of soft material, the combustion chamber is sealed by a simple metal eyelet, which also protects the soft material from the combustion gases. In newer versions of these the water jacket sealing is assisted by small beads of elastomer carefully positioned around each aperture such that water leakage is now rarely an issue. Coatings can also be deposited over the whole of the gasket surface to prevent sticking and assist sealing at a micro level. While these are popular for many older engine applications, lightweight engine designs demanding lower clamping loads with less distortion have moved towards single piece metal-elastomer gaskets. Elastomer sealing lips fitted into an alloy or a stainless steel carrier layer with reinforcement around the combustion zone, the clamping loads can be much more accurately controlled.

For the best in high performance cylinder head gasket technology however, multi-piece metal gaskets are now the norm. Consisting of a number of beaded metal carriers with elastomer coatings the exact design is bespoke to the particular application. Designed to pay particular attention to the dynamic events around the combustion chamber seal, improving its fatigue life, these do however generally require a much better surface finish on both the cylinder head fire face and crankcase top deck, than previous designs.

Fig. 1 - Modern multi-piece metal gasket.

Written by John Coxon

A typical case of this, I was reminded recently, was in the casting of cylinder heads and crankcases. In those days and I am only talking about the 1960s here, most of those with casting skills resided in the Pattern shop. A most unusual place and quite unlike everywhere else in the works, the Pattern shop smelt of pine wood forests and the floor had a liberal coating of sawdust rather like the local drinking establishment just around the corner. In place of the milling machine was a rather frightening bandsaw and the lathes seemed to lack a proper saddle and cross-slide so familiar in other parts of the works. You see, this place was used to dealing in wood and when sawn, machined and stuck together with glue, together with a liberal application of polyester filler for the fillet radii, (oh yes, this was technology at its best) a wooden facsimile of the external parts of the cylinder head would evolve. For the internal parts, life was a bit more difficult. Working from a 2D drawing and using a pattern maker’s rule, the patterns to produce the internal cores were produced which looked totally unrecognisable to anything on the drawing. You see, the pattern maker had to think in negatives and produce the patterns to make the cores, which would be placed in the finished mould to create the voids where the metal was to go. To the apprentice this was all rather too confusing! At the time we often doubted if the designer had ever thought about things like draft angles or split lines or runners where the molten metal would flow. The great skill of the pattern maker therefore was in deciding where all these were to go and, in effect, how the thing was to be made in the first place!

But how casting technology has changed in the modern world. Gone is, in effect what I would recognise as the pattern shop. Gone is the evocative smell of wood and polyester filler. Gone too is the pattern maker’s rule, with proportionately larger scales to account for shrinkage upon cooling - one side for aluminium and the other for cast iron.

Today, cylinder head designs are encapsulated in 3D virtual form, and core boxes will be machined directly using 5-axis machines. Indeed using rapid prototyping technology the cores themselves can now be made directly on a laser scanning machine. Sadly, gone also are the cylinder head drawings. Comprising sometimes of up to three or four full size A0 sheets, each one would be detailed with lots of little gems of knowledge, passing on titbits of information to the pattern maker. As a historical statement to the design philosophy, students of the art could pour over these for hours on end and learn much.

However, in its place we have castings that are not only accurate but thanks to modern casting software are of far better quality and free from porosity or blow holes. Only quality patterns can make good castings from which good components flow. As an engineer, I know which one I would sooner have.

Written by John Coxon.

Cast into the aluminium cylinder liner at the foundry, these have proven to be difficult to manufacturer but have given very little trouble over the years.

However, the increasing levels of heat rejection from ever improving performance, the complexity of the manufacturing process and the difficulties associated with recycling cylinder blocks and end-of-life legislation, have made it absolutely essential to research new materials and new methods of manufacturer. The traditional alternative to cast-in steel liners, and that used by some motorcycle manufacturers is ‘plated’ aluminium onto the parent metal bore. Similar to the Mahle trademarked Nikasil coating, where the parent metal bore is deposited with a coating made a soft nickel matrix and finely dispersed silicon carbide particles, these have excellent oil retaining properties and good heat transfer but still retain the relatively poor strength of the block material. In addition, the thermal expansion may not be compatible with current thinking in piston design.

In the end after much research, a new material and a completely new manufacturing technology was established that gave not only the heat dissipation and increased durability required but was also environmentally friendly to manufacture and at mass production levels, cheap too. Configured as a separate liner to the die-cast block, the material would need to have similar heat dissipation properties to aluminium, the strength of steel and a coefficient of expansion nearer to that of the pistons used.

A common addition to the melt of most aluminium alloys is silicon. Widely used in the production of aluminium castings, this is added in small amounts to provide fluidity in the molten phase and low shrinkage upon solidification. Typical aluminium alloys for cylinder blocks may therefore contain somewhere between 5-10% silicon by weight. In general about 5-7% is optimal for gravity sand casting while for pressure die-casting, when speed of the process is more important, the extra fluidity of a 10% mix is more typical. At 11.7% silicon the alloy is said to be eutectic and because of the coarse gain structure, without modification, the mechanical properties are poor. Beyond this point, into what metallurgists refer to as the ‘hypereutectic’ zone, these coarse primary silicon crystals (which eventually will give a highly abrasion resistive surface quality on the cylinder bore) become problematic leading to issues in both casting and machining. However as these problems increase so does both the potential strength of the resulting casting. At the same time the coefficient of thermal expansion is reduced. Alongside wear resistance, these are two very desirable characteristics in any cylinder liner.

Presented many times with problems of this nature, engineers will look towards sintering as a way of controlling the composition of the alloy and minimising any machining. For the revised liner material a chemical composition of 22-26% Si, 0.1-0.3% Cu, 0.05-0.9% Mg, and the balance Al, was therefore selected. The molten alloy was first atomized, and cold isostatic pressed before vacuum sintering. Once consolidated, the billet was hot extruded into a tubular form and heat-treated before casting into the cylinder block.

A complex process, but one which would give the characteristics of high strength, low thermal expansion and excellent durability in a lightweight aluminium block.

Written by John Coxon.

Early lip seals were made using a nitrile rubber. Now referred to, as NBR the application was limited to working temperatures of no more than 90-100 degree C immediately under the lip at the rubbing surface. Changes in seal designs at this stage to use much narrower contact points not only improved its performance, but also reduced the amount of heat generated in the first

place. And along with less heat came improved reliability. With high resistance to the chemical attack of the mineral oils used at the time, these polymers were substantially improved over the years to form hydrogenated nitrile (HNBR) compounds. Using the extra 25° C performance gained over NBR products, these are still to be found where reduced power outputs and minimal costs are the only requirements.

As engine performance and speeds increased in the post war era of the late 40s and early 1950s, so did the need for a new breed of material. Enter the polyacrylic now referred to as ACM. When introduced these had a much higher maximum operating temperature (up to150° C) and were chemically more resistive. But despite all this, as a sealing material they were much inferior having a greater wear rate. Dynamically poor as well, at the high speeds sometimes encountered, the seal lip could not follow the slight eccentricity of the shaft while it was rotating. This led to the phenomenon of ‘stick-slip’ at the seal lip, which was accompanied by a loud squealing noise.

By this time however, the chemical industry was coming to grips with post war economies and a range of new products, those of Silicone rubbers, was now available. Having a better operating temperature range (-50 to 230° C) but poorer wear characteristics and chemical resistance, because of their excellent dynamic behaviour, the use of these in seals at least produced an instant fix for squealing issues at high speed. Although much weaker in terms of mechanical strength, silicone was regularly used in crankshaft seals until the range of fluoroelastomers came along in the mid 70s. Today mainly because of its wide operating temperature, silicone remains an excellent material for static applications and with good resistance to aging, it can still be found in low pressure applications in other parts of the engine.

A spin-off from the aerospace industry and because of their price, fluoroelastomers were introduced initially very much as a problem solver. More of a revolution in seal material than an evolution, these now form the backbone of most rotating seal technology. Sometimes referred to as Viton (which is in fact a registered trade name of DuPont) and generically referred to as FKM elastomers, these products have excellent high temperature properties, are chemically resistive and have exceptional dynamic behaviour. For crankshaft sealing applications where shaft surface speeds don’t exceed 70 m/sec, FKM products are the most commonly found using PTFE (polytetrafluoroethylene) coatings bonded to the lip to minimise frictional losses.

In extreme applications, fully fluorinated perfluoroelastomers would be used. Rated at temperatures up to 320° C, these are very expensive but are totally resistive to all forms of chemical attack, which makes them ideal for use with methanol.

The move to highly synthetic crankcase oils in the past twenty years or so has highlighted a gap in the performance envelope of HBNR products such that a new range of optimised HNBR materials are now available. Maintaining its high temperature capability and extending its low temperature performance, optimised HNBRs with their added extended abrasion resistance are now offering similar levels of performance to FFKM materials but at lesser cost.

In the world of NBR, HNBR, FFKM, CDM and XYZ…., developments will never stand still!

Written by John Coxon.

With a specific gravity of 1.74, magnesium is by far the lightest commercial engineering metal. Compared with the alloys of aluminium (specific gravity 2.7) it is around 30% lighter and with a superior strength to weight ratio, can save weight without the loss of any component stiffness. The metal itself has insufficient strength for use in engineering applications but when alloyed with small amounts of zinc, zirconium or some of the more exotic rare earth metals, the grain refinement induced increases its strength to that of many of the aluminium alloys available, save those of the ultra high tensile versions. This is fine for low pressure, gravity sand casting that will subsequently require some form of heat treatment, but when higher volumes require pressure die casting technology, the addition of aluminium to the melt improves its castability. However the presence of aluminium reduces the high temperature performance of the alloy and so other alloying additions – calcium, strontium and rare earth elements have been found to help.

Designing in magnesium however, is not just about changing the melt. Notch sensitivity and the poor response to impact loads require the designer to take greater care at changes of section to minimise the stress raisers. The modulus of elasticity, at around 45 KN/mm2 and lower to that of most aluminium alloys, will also need additional material to maintain the stiffness in critical areas. But the real issue with using magnesium and its alloys in power units apart from its poor corrosion properties, is its performance at elevated temperature and particularly that of high temperature ‘creep’.

Creep is described as the slow plastic deformation of materials under the actions of a constant stress. Not usually a problem with most metals at ambient temperatures it is often easy to forget that as the working temperature of a component increases, so its ability to retain its elastic behaviour reduces. Thus in iron and steel, creep is only an issue at over 400-450 deg C, while that in aluminium alloys is not normally an issue in power unit applications. However in magnesium and its alloys, this phenomenon is of major concern and makes itself known in two ways – that of dimensional stability, and bolt load retention (BLR). Described as the load retained on the bolt when the component is heated to a certain temperature for a specific period of time and then cooled down again, some magnesium alloys have been known to have as little as 45% load retention after 100 hrs at 150 deg C falling to 38% at 177 deg C. Developments in metallurgy in recent times however have produced alloys which have comparable performance to some of the more common aluminium A319 and A380 versions exhibiting something nearer to 80% BLR at these temperatures.

Selecting, designing and making a cylinder block out of a suitable magnesium alloy, is a complex process. Using information on its performance at temperature (fatigue, elongation,
proof stress and creep properties), as well as some of its other traits (casting properties, corrosion resistance and not forgetting its cost), the process is very exacting and fraught with unknowns. Nevertheless, I will still expect to see more of this material in power units in years to come.

Written by John Coxon.

In recent years and with vehicle emission standards becoming more and more exacting, the surface of the cylinder bore is considered to be an emission critical component. While the technologies developed appear to have more direct benefits to the OE vehicle industry, the spin offs, in this case the bore finish, has helped to reduce engine friction. Modern engines are therefore very much more likely to have what is known as a ‘plateau’ surface finish. Consisting of a finely bored cylinder, bored more or less to size, the surface is then textured to produce the combined characteristics of a smooth, precisely cylindrical surface but with oil retention capability. To generate this texture will require dedicated honing machines when a series of coarse and then fine grit stones will be plunged through the bores in turn and while rotating at the same time produce a series of random grooves at angles varying between 30 and 45 degrees. The rougher stones will gouge out deep grooves in the material while the finer ones in subsequent operations, will effectively ‘top slice’ the surface away producing altogether a surface smoother to the eye but with deep groves spiralling down the bore. It is these deep grooves that are intended to retain the oil, which will subsequently lubricate the piston ring as it passes. When oil is present on the surface and the thickness of the oil film is greater than the height of the surface peaks, then hydrodynamic lubrication will be present and friction will be minimised.

The main concern with all this is that of the initial honing operation. The coarse grit whether it is vitreous stone or diamond rips out the material and can leave small amounts of residue deep in the grooves. While later operations will remove and effectively polish the upper portion, loose or partially removed material can still be retained in the grooves even after a thorough washing. By changing the honing procedure and introducing a laser to remove small but controlled amounts of the surface material later in the cycle, discreet pockets can be eroded which can be deeper than these initial rough grooves, more precisely controlled and can retain much more oil.
Thus the complete cylinder surface is smoother and more homogenous than with conventional practice. Furthermore, these discreet pockets can be machined anywhere on the bore to introduce additional oil to those places most in need. For most engines, that means nearer to the top ring reversal point where lubrication is traditionally difficult as the ring slows down.

With a million engines worldwide claiming to use this sort of technology, maybe it is only a matter of time before we all are using it.

Written by John Coxon.

In designing any rotating shaft seal, the primary concern must be to take into account the degree of misalignment between the surface of the shaft and the lip of the seal. This is the result of two main factors – the eccentricity of the housing holding the seal and the dynamic run-out of the surface of the shaft. For plain journal bearings, where the journal effectively rattles around within the shell bearing, this also needs to be accommodated, however small the clearance.

The eccentricity or perhaps more accurately, the shaft to bore misalignment (often referred to as STBM) is purely down to manufacturing. On older engines which have been rebuilt a number of times it may be advisable to check the concentricity of this against the crankshaft journal if this is possible. A special fixture may be necessary but in terms of the ultimate in reliability and minimising any potential source of unwanted friction, this eccentricity should be kept to a minimum. The more time spent minimising this factor and the less work the seal has to do, the greater the opportunity to reduce the radial loads and hence parasitic losses at the crankshaft.

The dynamic run-out (or DRO) of the shaft is, however, an altogether much more complex issue. Expressed as the ‘total run-out of a position on the outside diameter of the shaft as it is rotating’, this can be caused by a number of crank manufacturing related problems. Apart from the difficulties in machining a journal to be precisely circular, wear in the grinding machine itself, however slight, can introduce unwanted geometric shapes, while the dynamic effects of the crankshaft also have to be considered. Possible whirling and the effects of the firing of individual cylinders can create a ‘shaft shake’ condition. This can create an out of round movement of the journal, which also has to be absorbed by the seal. At speeds upwards of 18,000+ rpm, such effects if not properly controlled can have a significant effect. Furthermore, although the surface finish will have an effect on the sealing performance, the way in which the surface was machined can also make the difference between a seal that works and one that doesn’t. In particular, in order to produce the desired surface finish of between 0.2 to 0.4 Ra but also ensure there is no machining ‘lead’, journals should be finish ground using a ‘plunge’ grinding technique only. Failure to do this could introduce a slight ‘lead’ or scroll into the journal surface and actually pump oil out from behind the seal through the action of the scrolling.

While there are many other aspects to the design of oil seal systems, only if the geometry of the installation is correct, can it ever hope to succeed.

Written by John Coxon.

In any analysis of the weight of an engine, the cylinder block and head will together, and depending upon the material used, account for something around 35 – 40%. It is therefore not surprising that as the rotating and reciprocating parts have been tweaked here and there to remove every last ounce of excess, the great push thereafter is towards the big bits that hold it all together. But weight reduction is not just about selecting cast iron or aluminium alloy; not just about making the cross-section a little bit thinner or reducing the length of the bolts holding the assembly together. The weight reduction exercise can start at any stage from the base design through the material selection right through to the choice of casting processes used.

However selecting the correct grade of metal is just as important as choosing the type. While outlawing cast iron and even some of the newer CGI (compact graphite iron) materials as well as some of the more exotic stuff (for instance magnesium) the only other realistic choice left is that of aluminium.

But there are many grades of cast aluminium and in the selection process there are many factors other than tensile strength of the material that have to be taken into account. Factors such as fatigue strength under cyclic loading, yield strength, ductility and hardness obviously come into it but the ability to retain pressure (in other words porosity) is paramount as well as the fluidity of the molten mixture during casting. Finally the material needs to be readily machined and give a good surface finish if the end product is to be acceptable.

As an example of the evaluation process I have taken three grades of aluminium alloy suitable for casting.

Aluminium alloy 242 T77, a 4%Cu, 2%Ni, 2.5% Mg alloy which, when solution annealed and aged at 350 deg C gives good ductility, hardness and fatigue strength at the expense of tensile strength and fluidity in the molten state;

Aluminium alloy 319 in T5 heated condition, a 6%Si, 3.5%Cu alloy thermally aged at 150 deg C; or

Aluminium alloy A356 T6, a 7% Si, 0.3% Cu alloy with greater strength than either of the above at the expense of ductility / elongation but with superior fluidity and is easily machined.

As a general rule the addition of copper or silicon to the melt reduces the thermal expansion and improves the fluidity during casting but while copper hardens the structure over time (age hardening), silicon improves its abrasion resistance. Small amounts of magnesium or manganese can also improve strength.

Of these three alloys, only A356 satisfies the greater majority of our material requirements of high ultimate tensile and yield stress, suitable hardness and low ductility but with excellent casting fluidity and adequate machinability. Unsuitable ductility can often be compensated in the design process while poor castability, using modern casting software and CAD can sometimes be overcome. And cost, of course, is always a factor!

A356 and is in fact very similar in many ways to the BS 1490 specification LM25 (at 7% Si , 0.1%Cu and 0.4% Mg) used for many years in motorsport and many high quality aluminium castings. This is a version of LM25 in which the iron content is limited to 0.15% and the range of magnesium from 0.20 to 0.45%. While the minimum properties as specified in the standard refers to separately cast test bars it nevertheless becomes very difficult to attain these without very close control of the composition, melt cleaning, degassing and careful temperature control to meet these specifications in the casting.

While the research into new grades of material is ongoing the eventual selection procedure will inevitably be a compromise of properties.

Written by John Coxon.

Aluminium-silicon alloy materials are some of the most widely used for castings of all types. At silicon levels of less than 11.7% (the eutectic point) these hypoeutectic alloys are characterised by excellent fluidity during casting and impressive corrosion resistance as well as being readily machined. Specifications like LM4 ( 5%Si) and LM25 (7% Si) have been common place in the past as cylinder block materials, but when used uncoated as a cylinder liner, they are totally unsuited.

In the mid 1960s, this issue of iron or steel piston rings running against the parent metal bores in aluminium cylinder blocks was partially solved by the introduction of a thin layer of silicon carbide material. Deposited electrolytically, these processes were expensive and were not always robust when viewed against the demands of the engine OE business. Increasing the amount of silicon beyond this eutectic point to somewhere between 15-25% and producing what are now known as hypereutectic alloys, improves the wear characteristics and also reduces thermal expansion but all this is at the expense of machinability which together with the spray compacting technique used during manufacture, make them very expensive.

A different approach and altogether considered to be much cheaper, is the use of metal matrix composites particularly those using a particulate ceramic such as silicon carbide or in the example below, alumina (Al2O3).

Consisting of a metal phase and a ceramic phase, which are not miscible, MMCs in cylinder liners can be designed to have wear resistance similar to that of the cast iron that they replace. However the major issue with aluminium MMCs is ensuring that the ceramic element is in the optimum position within the liner and isn’t all cramped up in one particular place in the component where it’s properties are of little value. The favoured method of manufacturing top quality liners is casting, and like high quality cast iron liners, these can be made using centrifugal casting techniques.

A continuous process and hence much favoured by production engineers, centrifugal casting rotates the heated cylindrical mould while at the same time introducing the molten mix. The centrifugal force generated from the revolving mould forces the molten mass against the wall of the mould, which quickly cools and solidifies. Any low melting point impurities as well as gases that would otherwise cause porosity will quickly migrate to the centre leaving a high quality fine grain structured material at towards the outside of the cylinder with the lower quality, impure metal left towards the centre. Once cooled the poor quality metal in the centre can be machined out leaving a high quality outer liner. When a two-phase molten MMC mix is introduced the more dense phase will gravitate towards the outside and the lighter ceramic will distribute itself according to the relative properties between it and the parent metal depending on such things as relative density, speed of rotation, particle size and rate of cooling. In this way the distribution of the ceramic phase can be engineered to coincide more precisely with the properties desired.

Potentially, this is a much cheaper way of producing aluminium ceramic alloys, expect to see more of this metal matrix composite technology in the future.

Written by John Coxon.

Dave Salisbury, Chief Design Engineer at Engine Developments Ltd (EDL), told RET Monitor, “It’s a simple philosophy really… we just work hard to eliminate seals and gaskets wherever we can.”

Whilst that sounds straight forward and much like common sense it is actually something which is very hard to achieve and requires careful attention during the design stage of a project.

An FMEA (Failure Modes and Effects Analysis) of an engine will identify that where there is a joint there is a potential leak path with the inherent risk of a fluid leak, whether it be air, the working fluids (air / fuel charge or spent combustion gasses), coolant, fuel or oil.

The potential effects of fluid leakage are all extremely serious for a manufacturer such as EDL which bases its business on its reputation for performance and reliability in the Le Mans 24 Hours, as Salisbury explains, “With a season based essentially on one race there is very little room for seal or gasket failure in the schedule.”

We asked Salisbury what sealing issues EDL had been through recently on the LMP engine programme, “At Le Man this year we had no leakage problems whatsoever. In 2008 the engine was entirely leak free but we were nearly tripped up by a car installation issue. One of our LM P1 entries suffered from an oil hose coming loose which meant the car lost almost all of its oil. The driver saw the low pressure alarm and very sensibly cruised back round to the pits. Once the hose was re-fitted and the system refilled the engine ran OK and finished the race. It was still very lucky obviously as a major oil leak like that will often cause a crash.”

So leaks can cause fires, crashes and engine failures. They can also lead to disqualification at Le Mans as the ACO still test restrictor legality with the stall test; that is whereby with the engine idling the restrictor inlet is blocked. To pass the test the engine must stall. Salisbury continues, “When we started the LM P1 engine programme in the late 1990’s we looked closely at the sealing of the slide throttles and the airbox, as we had to ensure that designs which had worked well in Formula One would be up to the stall test. You don’t have to be quite so leak proof with these things in Formula One!”

Next month we will look further into EDL design philosophy before we get back on track with our examination of cylinder head gasket technology.

Written by Tom Sharp.

It was while examining a cylinder block prior to rebuilding a classic engine that made me reflect on how far, in terms of technology modern road vehicle engines have come. The crankcase in question was a five bearing four cylinder unit of 1970s origin and having a relatively undersquare bore/stroke ratio would rev regularly up to 9000 rpm. At these speeds the main bearing caps and the block material in the vicinity of the main bearings would regularly be checked for evidence of cracking. This was standard practice at the time and historic racers in particular are still very much aware of this requirement. At the time the five-bearing block was considered a great technical leap over the three-bearing crank, which, I seem to remember always tried to tear out the centre main bearing from the block. But since these engines were reputed to give a little extra power, all that was forgiven in the search for any form of competitive advantage. Eventually all that was outlawed either by mandating the five-bearing block or they simply ran out of three bearing blocks - I forget which, but road car engines used for racing purposes at that time required substantial modification.

The fitting of steel bearing caps and line boring the mains again helped improve reliability and steel cranks, with their extra strength, helped control any potential whirl issues. With cylinder block ending at the centre line of the bearings and the sump invariably made of pressed sheet steel, very little else could be done. In later engines, and if the web thickness tying the bearing to the side of the block was sufficient, additional ‘helper’ bolts could be used to increase the clamp load on specially widened main bearing caps. Designed to reduce cap movement under the increased loading in some heavy-duty castings, these are sometimes referred to as four-bolt main bearing caps when all four bolts act in the same geometric plane.

Another type of four-bolt design is sometimes referred to as cross bolting. Engines with this feature have deep-skirted cylinder blocks and a sump split line well down past the centre line of the main bearing assembly and therefore have the potential to have a very stiff lower crankcase. In many cases the design was principally to simplify the lower block to sump joint and remove what was traditionally a tricky seal at either end around the bearing housings and so, in the interests of economy, ‘ordinary’ two bolt main caps would be used. However in the quest for increased stiffness particularly when these cylinder blocks were modified to take diesel combustion systems, manufacturers tied these into the sides of the deep skirt using bolts at 90 degrees which when bolted through and fully tightened, would produce a significant increase in rigidity. An increase in rigidity, if not quite as much could be achieved using a ‘girdle’, a component which either replaced the main bearing caps attaching to the sump flange or simply picked up off the tops of the existing caps and tied these in to the sump flange. In either of these this extra stiffness was assured at the cost of an extra external joint. This is something that has to be considered very carefully in any engine modification if the engine is to remain oil tight at all times.

Many lightweight production gasoline engines today solve the issue of lower crankcase stiffness using a ladder frame design. Racing practice for many years and sometimes reinforced using steel inserts to control expansion around the bearing assemblies, the principle reason is not necessarily to increase the stiffness, but to reduce structure-born noise.

And so it seems, that one issue in racing can solve another in road transport, many years later.

Written by John Coxon.

The move towards improving cylinder bore technology could only begin once we had the means to accurately quantify the quality of the surface. Until the invention of the electronic surface texture testing machines, surface finish measurement was limited to only a few parameters. The mean Roughness (Ra), the total Roughness (Rt) and the bearing area throughout the various levels in the surface were all that could be reasonably expected. Construction of the Abbot-Firestone bearing area curve (a curve describing the theoretical bearing area at each level throughout the surface) was a complete headache and required much manual graphic endeavour. The subsequent development of electronic surface finish measurement machines automated much of this process and enabled a far greater number of parameters to be established. Ra alone doesn’t reveal that much at all about the surface finish but values like Rpk (peak height), Rvk (depth of the valleys), Rk (the average core roughness based on Rpk and Rvk measurements) and Rz (highest peak to valley height) can be begin to describe the precise surface roughness from which we can get a more complete picture. Thus instead of just specifying a simple Ra value, bore specialists now have to grapple with Ra, Rpk, Rvk, and Rk values as well.

Specialists tend to come up with their own numbers but as a guide, typical values for a cast iron bore suitable for use with chrome-plated rings may be:

Ra 12 to 24
Rpk 2 to 24
Rvk 20 to 80
Rk 28 to 48

Keeping in mind that these values are stated in micro inches (0.000001”), you will appreciate that modern bores are very fine indeed. Established over many years the ideal surface finish should also be perfectly cylindrical, smooth at a micro level and contain a series of grooves in it to retain the lubricant. In practical terms the nearest we can realistically get to this is a finely bored surface with all of the peaks of the asperities removed leaving only the oil retaining grooves behind. Such a surface is described as a ‘plateau’ and will take many honing operations to achieve.

At one time following the boring procedure, the surface finish would consist of a single honing operation. Modern multi-stage honing methods however, use different sizes of grit stone and different grinding materials (silicon carbide, diamond or boron nitride) and may well finish with a nylon brush to remove the last traces of folded metal. This will create a surface much closer to our ideal but with all the tears and folded material from machining removed, together with a correctly angled cross-hatching, piston rings can now be expected to bed-in in a matter of minutes.

When you consider the amount of effort that goes into the design and manufacture of the piston and rings, surely only the most carefully prepared bore surface will suffice?

Written by Jown Coxon.

There are of course many ways of achieving this seal, but perhaps the first question we should ask ourselves is whether or not we need to have a joint which needs sealing in the first place.

Race Engine Technology magazine is soon to feature an in depth review of a historic racing engine which did away with this joint by casting the cylinder liners integrally with the cylinder head. This engine was the Hart 415T four-cylinder Formula One turbo engine of the early 1980s.

Without pre-empting the magazine article, the Hart was produced in this way because the cylinder pressures were so high that it was deemed the most feasible solution despite the inherent difficulties which this solution presents.

If though we assume that we have a separate cylinder head, what choices do we have? Well, we can still ask if we need a gasket. Will two parallel surfaces with a sufficiently good surface finish seal against combustion gasses if the clamping pressure between them is high enough?

The answer of course depends on various factors such as the medium to be sealed, clamping pressure, component stiffness, surface profile (whilst hot and clamped) and surface finish. Certainly it is not unheard of for manufacturers to eliminate conventional gasket technology, a classic example of this would be brass sealing rings. These are typically 8 mm deep and 3 mm in radial thickness and locate into a recess at the top of the cylinder liner. Sealing is then achieved purely by the flat face at the top of the sealing ring pressing against the fire face.

Whilst brass is an obvious material choice due to its ability to conduct heat away from the combustion chamber, it is heavy and difficult to achieve a perfect surface finish with. There is no act of parliament which says sealing rings have to be made from brass; steel is one perfectly viable alternative, and is easier to grind and polish to a low Ra value.

For sealing ring technology to operate well requires a high standard of machining and finishing of both the cylinder head and the sealing ring, and it also requires that the engine is not abused. Overheating the cylinder head for example could warp the fire face to the extent that the seal will no longer operate.

Sealing rings are almost universally used in conjunction with O-rings which seal the oil and water passages.

In future issues we will look at developments in head gasket technology, and examine another innovative sealing method from the recent past.

Written by Tom Sharp.

But, as subscribers to Newton’s third law will testify, to every action there is an equal and opposite reaction and the forces due to combustion / inertia will eventually react all the way back up through the cylinder block and into the cylinder head. At the top of the engine these are retained by the arrangement of cylinder head studs or bolts clamping the casting in place. During this process while the cylinder liner will always be in compression, that portion of the cylinder block between the threaded portion holding the main bearing bolts and that retaining the cylinder head studs/bolts will be in tension. And although, as engineers, we have been transmitting loads into all manner of castings for many hundreds of years, the method of feeding these loads through the cylinder block casting is not always the most reliable or the most efficient. If we could somehow sandwich the cylinder head, cylinder block and lower bearing structure using a long bolt effectively reacting the combustion and inertia loads direct back into the cylinder head, then the cylinder block between the fixings would be in compression rather that tension, less material would be needed and hence a significant weight saving could be achieved. Under these conditions the cylinder block would become more of a cooling water jacket than a structural part of the engine. That was the theory.

In practise however, things weren’t quite that simple. To start off with suppliers of long bolts, or indeed any type of long fastener were hard to find. The problem was that the material to make these bolts comes wound on huge drums, which has to be straightened out in the initial stages of manufacture. Having straightened it, forged the head and rolled the thread however, the bolt would try to regain its former coiled state during final heat treatment. The other problem with this arrangement was one of assembly. In any long fastener, there is inevitably an amount of ‘wind up’ as the cylinder head is gradually tightened and finally torqued into place. The amount of twist depends partly upon the length of the bolt and partly about the amount of friction at the flange just underneath the head and can be quite significant at times. Up to one complete turn was observed at times. I remember also quite vividly once when finish assembling a prototype engine, to hear a ‘pinging’ noise. Only one at first, but then rapidly followed by another nine as they all sprung back releasing their load and causing us all to have to think again. The solution was to minimise the amount of friction under the heads of the bolts either by using special coatings or even washers.

But much as with the long wait, the lesson was learned. In tightening any nut or bolt it is essential to keep the friction immediately under the fastener being turned to an absolute minimum. Failing to do so and you may not get the clamping loads you think.

Written by John Coxon.

As the specific performance increases along with larger diameter pistons, the pressure to move to bore materials

more compatible with the piston becomes more intense. Improved heat transfer and coefficients of thermal expansion similar to that of the piston, together with the quest for lower weight make the solution obvious but for the problems of excessive wear and a tendency for aluminium to gall. In addition to this the problem of poor oil retention made the selection of aluminium as a bore material totally unsuited to the application.

While suitable hypereutectic alloys have been developed since, it is over 40 years ago that engine component supplier, Mahle, developed a suitable coating. That coating was Nikasil and was to pave the way for a range of trademarked nickel ceramic processes designed to supply a hard, wear-resistant coating on an aluminium surface.

Compatibility in expansion coefficients between the aluminium alloy liner and piston, enabled skirt-to-bore clearances to be radically reduced benefiting not only noise when cold but also reducing blow-by losses and hence improved performance when it really mattered. Reportedly used first in the Wankel rotary engine to solve tip sealing problems in the aluminium housing, the process is also claimed to have contributed to the success of the Porsche 917 programme, enabling the 4.5 litre 180 degree V12, an air-cooled engine to be bored out to a full 5.0 litres and still maintain the same bore spacing.

Nikasil is an electrodeposited nickel silicon carbide coating which has the essentially property of being able to absorb engine lubricant. It is therefore said to be oleophilic, retaining the oil ‘like a sponge’ it was once explained to me. Consisting of a layer of nickel plate into which is introduced particles of hard, silicon nitride equally distributed throughout, the exact details of all the various processes used tend to be jealously guarded and are slightly different depending upon the material of the parent metal bore. In most cases individual liners will be placed in a multi-station rig and connected as the cathode (-ve) using a central anode (+ve).

Although there are several proprietary formulations, the electrolyte - a mixture of a Nickel salt and silicon carbide held in suspension together with various plating additives - is pumped up through the cylinder and overflows at the top to fall back into the sump, and then re-circulated. Depending on the exact parent material and the thickness of the coating required, deposition rates vary with the current density but the whole process should take no longer than an hour. After this the coating will be somewhere between 0.002”- 0.006” thick. Any greater and the layer could become brittle and break away from the bore. Any less and the final honing processes could end up by breaking through into the parent metal beneath.

Written by John Coxon.

In that article we focused on water contamination of bio-diesel, something which is to an extent, almost inevitable. Water contamination has a significant detrimental effect on the performance of certain elastomers; in particular the bisphenol cured FKM’s and FVMQ, which showed acceptable performance with pure bio-diesel.

In this article we will focus on the compatibility of typical automotive biofuels for gasoline engines with a variety of fuel system sealing compounds, focusing in particular on the phenomenon of rapid decompression failure in high pressure gasoline applications.
Initially Trelleborg conducted standard laboratory tests, which suggested that typical elastomers of Fluorocarbon (FKM) and Flourosilicone (FVMQ) were compatible with commonly used biofuels.

The following fuels were tested; they represent typical automotive fuels for diesel and gasoline engines:

Figure 1. Fuels Evaluated

The following compounds were evaluated; all are used in automotive fuel systems:

F= Fluorine content
Figure 2. Compounds Evaluated

Figure 3 illustrates the effects of biofuels typically used in gasoline engines, tested for 168 hours at 60°C. It is immediately clear that an ethanol blend ratio of 22% is more aggressive than 85%. Furthermore, FAM Fuel B is more aggressive than E22. For the bisphenol cured compounds fluorine content has the highest effect on volume swell. For certain peroxide cured compounds, in addition to the fluorine content, the substitution of hydrogen by oxygen in the polymer also reduces the volume swell. Although significantly higher property deterioration was noted than in diesel, most of these compounds are used in applications where alcohol containing fuels have been utilised for many years.

Figure 3. Immersion in gasoline based biofuels for 168 hrs at 60°C

Certain gasoline engines operate at high pressures. In an attempt to approximate the effect of pressure on fuel ageing compound 7 was tested in E22 for 168 hours at 60°C at 120 bars. The test was conducted in a pressure vessel and on completion the pressure was reduced at a rate of 1 bar per minute. Once atmospheric pressure was reached testing of properties was conducted within five minutes. Results are shown in Figure 4 and indicate no significant change in the properties after exposure to the different pressures.

Figure 4. Change in properties of compound 7 (64%F P) after immersion in E22 biofuel for 168 hours at 60°C.

Although all of the above test data indicates that the selected biofuels do not cause unexpected deterioration to the compounds, further testing was conducted to represent service conditions. In this section the effect of rapid decompression on O-rings used in high pressure gasoline systems was investigated.

In high pressure gasoline applications there is the risk of a sudden reduction of pressure in the system. To study the effect of this rapid decompression an experiment was set up as per the parameters in Figure 5. An O-ring manufactured from compound 7 was assembled onto test equipment replicating a high pressure fuel injector. The O-ring was in contact with pressurised fuel from one side, and aged for 168 hrs at 60°C. The pressure was then reduced to atmospheric pressure within one second and the O-rings were inspected. The fuels tested were Fuel C, representing regular gasoline, and E22. For each condition five O-rings were tested.

The results indicate that whereas no failure was detected in Fuel C all O-rings tested in E22 exhibited internal cracks, which whilst still sub-surface, are very difficult to detect. The higher polarity and smaller size of the ethanol in E22 causes higher swell and a large enough reduction in mechanical properties to result in O-ring damage during the rapid decompression. In addition to compound 7, O-rings manufactured from compound 8 were also tested and resulted in no rapid decompression failures in E22. Compound 8 is based on the same polymer as compound 7 but has been specifically compounded to withstand a rapid decompression environment. The results show that these tests are the prerequisite for qualified material recommendations for gasoline high pressure applications.

Figure 5. Rapid decompression testing in Fuel C and E22

Conclusions

Although standard laboratory tests suggest that typical FKM’s and FVMQ’s are compatible with the above biofuels, tests designed to replicate service conditions present a different picture.

The addition of ethanol in high pressure gasoline applications increases the possibility of rapid decompression failure. The resultant cracks dramatically reduce the integrity of the seal and are difficult to detect whilst still sub-surface.

These and other similar challenges can be met with the right selection of compounds.

References
[1] Biofuel Systems – New Challenges for Sealing Technology
Gordon Micallef
Trelleborg Sealing Solutions Malta
Axel Weimann
Trelleborg Sealing Solutions Germany

Written by Tom Sharp.

In some designs, especially those with compact combustion chambers, the cam motion is often transmitted to the valves via finger followers. This has the double advantage of reducing the overall height of the cylinder head and removes the restrictions of lift on a given tappet diameter in a direct acting arrangement. And although each design has its difficulties, this is often the preferred route when using such narrow valve angles. However, for ease of manufacture (high quality mechanical tappets can be readily purchased ‘off the shelf’) and therefore maximum reliability, many manufacturers of endurance engines prefer to stick to the tried and tested technology of DAMB. But as the valves become more upright the space available becomes less and more compromises need to be made.

That landmark design in racing engines, the original Cosworth DFV surmounted the issue by introducing a separate cam carrier as part of the ‘sandwich’ construction of the cylinder head. Consisting of a one piece aluminium alloy casting which held both the tappets and the camshaft, this was bolted to the main structure of the cylinder head by long studs which also served to retain the bearing caps and down the two middle rows at least, the shallow cam cover.

But such complication is not to be accepted in the volume engine business. Separate cam carriers and associated fixings mean added cost, while long studs can mean variable clamping loads. At the same time additional joints can provide a source of potential oil leaks all of which are undesirable and so the industry did what it always does in such circumstances – it simply removed the cam carrier from the Bill of Materials designing all future cylinder heads with integral carriers. However the introduction of the integral cam carrier further complicates the design. Since access from above is still required for the cylinder head fasteners and with fixed cylinder head bolt positions and a camshaft centre line dictated by the valve positions, the problem of the location of the cam bearing fixings can be a trial. In the end the traditional approach of positioning the cam bearings above and between the cylinders had to be rejected. The proximity to the cylinder head bolt positions effectively rules this out. The final, albeit compromised, design was to have the cam bearing not between cylinders but between the cams of each cylinder. Although restricted by the diameter of the tappets, which in turn limits the cam profile velocity, this would seem to offer the stiffest solution. Once fully proven by the volume manufacturers, the smaller race engine suppliers followed suit.

And the cam carrier? R.I.P. - Rest in Peace.

Written by John Coxon.

A wet liner, of course, has many advantages. The simplicity of casting the block using the minimum number of cores lends itself very well to low volume production such as that used by most race engine manufacturers. And should the bore suffer any terminal damage then the liner can be simply pressed out and replaced with the minimal amount of effort. The downside to such designs is the necessity to seal the liner to the cylinder block at both the top and the bottom of the liner under all engine-operating conditions of temperature and load. Replacement liners can also be fully finish machined and even honed to suit but to develop a system of locating the liners rigidly in the block and sealing it reliably against the coolant at both top and bottom as well as against combustion pressures at the fire face, can take quite a lot of time and in some ways can be likened to an act of faith.

The dry liner however, is not without its issues either. Favoured by most volume engine manufacturers because it removes a potential source of engine failure – the seal, the resulting components should ultimately be more reliable. Occasionally, engines of this configuration use parent metal bores, but when liners are cast in place from the outset the overall mass of the resulting cylinder block can give rise to a slight weight penalty. To compensate, the assembly is likely to be slightly stiffer but it is when the bore surface is damaged that the dry liner loses its attractiveness. At these times inevitably some level of machining will be required. For most applications the preferred approach is to machine the bores of the block to leave a slight step at the base. A replacement cylindrical can then be pressed home to abut against this step and the top deck machined such the liner is flush with it. To do the job properly the external diameter of the liner should be measured in a minimum of three places, equidistant around the circumference and then averaged. The correct block boring size can accordingly be calculated allowing for an interference fit (for cast iron liner in a cast iron block) of somewhere around 0.0006” per inch bore +/- 0.0005”. The actual fit will depend very much on the block material, the design and in the final analysis, what actually works! As it was explained to me by an engine builder of many years, “There is no substitute for experience in this business.” Some blocks will require a gentle press fit using industrial adhesives for extra security but most small cast iron blocks/liners should work perfectly satisfactorily using around 0.0015- 0.0025” of interference. Since the liner will conform to the shape as machined, it is essential that this is completed to the highest of accuracies running down the axis, although the final bore shape can be corrected during final machining.

Whatever your politics, when it comes to liners either ‘wet’ or ‘dry’, you either believe or you don't.

Written by John Coxon.

Every time an engine builder has to lay a silicon bead on a head gasket, he finds himself muttering, “There has to be a better way to do this!”, as he wipes the goo off his fingers. Excess silicone sealant inevitably squeezes out at the joint, creating a freestanding bead of material that is held in place solely by a very thin film at the seal joint, creating the perfect opportunity for it to fatigue off

and find its way into the return oil flow, eventually blocking a screen at the oil pickup.

The advent of ‘bench top’ CNC routers and similar equipment has allowed the development of inexpensive, precisely controlled dispensing systems. Now a well-established technology, simple, two and three-axis CNC stepper motor gantry machines can be retrofitted with lead screw-controlled syringe injectors in order to create custom programmed gasket and adhesive patterns. Or, one can simply purchase a form-in-place gasket machine, which is essentially an already assembled package exactly as described above. Nor does one have to be an expert to refit a gantry router; a number of companies offer kits for this purpose. The price of these systems has come down to the point where moderate-to-large-volume race engine builders can reasonably consider one of these machines for their shop.

Form-in-place gasket technology is the controlled application of a bead of material onto the desired surface. Virtually 100% of the gasketting material is used, since it is dispensed directly in a specific pattern onto a given component: hence, there is far less material waste than with a conventional die-cut gasket. While the majority of applications are planar 2D patterns, complex 3D joint sealing is readily accomplished. Applying the bead into an already machined O-ring groove results in a permanent seal that can be reassembled several times, ideal for race engine applications. Even better, this technique allows for metal-to-metal contact on the remainder of the surface, as the O-ring needs very little cross-section to accomplish its sealing function. Programming the bead ‘toolpath’ is a relatively simple conversion of DXF drawings, something that the most rudimentary cadcam system can accomplish with ease. The cross sectional area of the bead can be adjusted to account for local features as it is dispensed by simply changing the syringe lead screw feed rate relative to the combined XYZ feed rate.

Turnaround is rapid; no custom tooling, moulds, fixturing, silk-screening, or involved setups are required, and chemical companies offer a tremendous range of dispensable materials, including silicones, natural rubbers, foams, and epoxies. Some of these materials can be UV cured for near-immediate assembly. In some cases, such as a silicon bead on a head gasket, assembly can be with the material still ‘wet’, or a small bead can be allowed to dry, and ‘cold flow’ when assembled.

For smaller shops not interested in doing this work themselves, there are now many finishing and coating companies that provide this service. The low cost of the equipment allows them to do this at a reasonable price, and this could become feasible for multiple applications on major gaskets for the engines that a given builder specialises in.

Written by John Stowe.

At those sorts of power outputs, the amounts needed to be competitive in Division One of the European Rallycross championship, even the best-designed production car engine would begin to feel the strain. At 600+ bhp, the production cylinder head on any turbocharged 2.0 litre engine would be bound to struggle and so the only sensible way around the problem might be to design your own. But setting aside all the other issues of porting and cam drives, at these power levels, cooling, particularly in the cylinder head itself, is a big issue.

Cylinder heads like that of the OE production Duratec need to have a low thermal mass with minimal cooling around the exhaust ports. Strict exhaust emission standards worldwide today mean that engines need to warm up extremely quickly and all subsequent waste heat has to be retained in the exhaust gas and directed towards the exhaust catalyst. To pass exhaust emission legislation, this needs to reach ‘light-off’ temperature within 20 seconds of ‘key-on’ and so for rapid warm up, coolant volumes will need to be very small with minimal cooling around the exhaust ports.

For competition engines, however, with long periods of running at wide open throttle, the overall amount of cooling will need to be much greater and so while overall coolant volumes won’t necessarily need to increase much, the coolant may need to be redirected to the more critical areas, such as the exhaust valve seat or where excessive temperatures can destroy some of the heat treated properties of the aluminium itself. At the same time the coolant flow rate needs to be balanced out to ensure even temperatures around and across the coolant cores and hence minimise any temperature-induced strain.

In optimizing this coolant flow, the designer also has to ensure that high flow velocities and rapid changes of cross-section don’t coincide with high heat flux zones. I remember one engine with this design flaw, which proceeded to erode the internal cooling jacket zone through cavitation in a matter of hours. In this particular case, the cylinder head (which was designed to use only minimal amounts of coolant flowing across the head) introduced a region of low static pressure just where the heat flux was greatest. The high coolant velocity at this point of low pressure, caused instantaneous boiling of the coolant and the sudden energy released eroded the wall of the coolant jacket just above the combustion chamber creating a nice little hole into the combustion chamber by the time the engine was just becoming run-in.

Stagnant flow areas can also lead to similar problems and so to address this problem and minimise any build up of any entrained air/steam, some form of internal or external water rail above the exhaust ports might be advisable. By carefully pumping the coolant along and across the cylinder head not only does this reduce the mechanical stresses but is also likely to reduce the possibility of the rearward cylinder running hotter than the others and running into detonation.

At 600+ BHP getting the air into the engine and the power out is one set of problems. Ensuring that the cylinder head simply doesn’t melt into sort kind of amorphous mass is quite another...

Written by John Coxon.

To be more exact, the material to which I refer is grey cast iron since it is grey in colour, it is iron and manufactured by casting. Easily machined, even if the resulting sward chips and flies all over the machine shop, this type of cast iron has excellent surface properties, which make it ideal for the application. And in case you think it isn’t a high performance material, let me remind you that the original Cosworth DFV engines incorporated this material in its liners. Delivering something like 400+ bhp at 10,000 rpm from the outset, there is still a strong demand for these liners on the historic scene. However for powers or speeds much greater than this, its ability to dissipate the heat generated begins to diminish.

The G1 specification cast iron used today has benefited from a further 40 years of development but these liners are still made using a centrifugal casting technique. The molten iron is poured into a rotating drum where the centrifugal force pushes the liquid mass outwards while the lower density slag, gas bubbles and impurities migrate towards the middle layers and escape such that when the metal solidifies, a fine grained homogeneous thick-walled tube results. While it might be easy to imagine that modern high performance engines would be better using stronger steel liners, cast iron has, despite its brittle nature, some admirable qualities. Principal among theses is the inclusion of graphite between the grain boundaries of the crystalline matrix. Whereas steel consists of iron with a controlled amount of carbon (normally less than 1.2%), cast iron will contain between 2-4%. In controlling the content, the melting point can be lowered making it more fluid and easier to cast but critically, allowing free carbon in the form of graphite to precipitate out. A natural lubricant, this graphite provides an excellent rubbing surface for the piston ring minimising both friction and wear. While other liners, either steel or aluminium alloy may be lighter, unlike cast iron, they all require some form of surface coating to give any form of acceptable durability.

The fitting of liners to cast iron cylinder blocks is a fairly straightforward affair. However the fitment of ‘wet’ cast iron liners into an aluminium block may require more care. In such cases, and apart from the increased level of interference fit required, the assembly may benefit from the use of Austenitic iron which alloyed with nickel and a small amount of either copper or manganese, will give a coefficient of expansion similar to that of the aluminium.

It may not be the most fashionable of materials but when the application can tolerate it, you could do worse than using cast iron in your liners. It might also save you a ‘bob’ or two.

Written by John Coxon.

For many years, automotive seals were manufactured from nitrile rubber because of its excellent resistance to mineral oils and petroleum solvents and fair resistance to heat and ageing. The introduction of smaller, higher revving, engines with elevated running temperatures has driven the switch to hydrogenated nitrile (HNBR) compounds with a higher top temperature range.

Optimised HNBR

Today's modern lubricants are complex products, containing detergents, surfactants and viscosity-improving low-molecular polymer additives. Synthetic oils are widely accepted while transmission and power steering fluids increasingly contain aliphatic amines. These highlight gaps in HNBR’s performance envelope.

A new generation of optimised HNBR compounds addresses these problems, offering many applications in a performance envelope similar to perfluoroelastomers such as FFKM - at a significantly lower cost. These new polymers maintain HNBR’s high temperature performance while extending low temperature performance to –40°C, thereby expanding opportunities in extreme arctic weather conditions.

FFKM

This fully-fluorinated perfluoroelastomer provides exceptional temperature and chemical performance, making it the material of choice for the harshest down-hole well drilling conditions. FFKM is a difficult material to mould but as more rubber moulders learn to master its production parameters, product prices are coming down, enabling increased use in oil and gas applications.

Originally formulated for aerospace applications, Viton soon found applications in automotive engineering because of its high temperature performance, exceptional resistance to mineral oils and chemicals, resistance to climatic factors such as ozone and sunlight, and low gas permeability. However, Viton exhibits long-term problems with high pH chemicals (bases), triggering premature failure.

VITON Extreme

The corrosion inhibitors designed to protect metal components use bases to counteract the long-term effects of acid build-up. Fortunately for designers, there is a new Viton (trade name Viton Extreme) that addresses this challenge and provides other advantages, including reduced manufacturing costs.

In addition to this enhanced base resistance, and Viton’s inherent fluid and chemical resistance, Viton Extreme offers improved tensile strength, superior compression set resistance and lower volume swell for longer seal life and wear resistance, making it ideal for use in the automotive and heavy duty/off-highway markets.

This fluoroelastomer also offers the moulder significant advantages: better mould flow, faster cure rates, improved mould release and reduce mould fouling, all of which should lead to more efficient manufacturing with reduced reject rates - and manufacturing costs.

Waiting in the wings

To respond to the requirements of engineering companies at the cutting edge of their technologies, seal manufacturers should be continuously investigating new rubber compounds through their on-going R&D programmes. These programmes should cover material flows, out gassing, residues, etc., that is, the material's moulding characteristics as well as its design performance.

For example, Teflon-coated rubber has been around for some time but the coating flakes and reduces flexibility; filling fluoroelastomer with nano-particle size Teflon shows immense promise. The resultant seal significantly reduces friction with none of the side effects of coatings, making it ideal for rotating seals in automotive, aerospace, oil and gas applications. It also exhibits reduced permeability, meaning reduced emissions to automotive manufacturers. Early indications are that it moulds well.

One seal manufacturer which specialises in providing customised solutions to customers seeking extreme performance without an extreme price tag, has successfully supplied seals and moulding to leading Formula One, WRC, Indy and Moto GP power plant developers. As a result of their efforts in this high tech market, the production parameters surrounding these exciting new materials have been mastered.

Written by Tom Sharp.

The following fuels were tested – they represent typical automotive fuels for diesel and gasoline engines:

Figure 1: Fuels EvaluatedThe following compounds were evaluated – all are used in automotive fuel systems:

F= Fluorine contentFigure 2: Compounds EvaluatedDiesel and B30

Standard tests involved immersion of the seven diesel compounds in both fuels (diesel and B30) for 336 hours at 150°C.

All compounds seemed to perform acceptably. Volume swell was well under 10% for all compounds, hardness reduction was under 5% for everything other than FVMQ (8%), tensile strength reduction was 25% or under for all compounds and the reduction in elongation was under 20% for all compounds except FVMQ, which was 26% in B30.

Water is a common contaminant in diesel fuel; typically it exists as a discrete phase at the bottom of storage tanks. It is significantly more soluble in biodiesel than in conventional diesel so there is an increased risk of water contamination.

The original tests were then repeated but using B30 contaminated with 1% water. The results showed a large deterioration in properties for all bisphenol cured FKM’s, i.e. compound refs 2, 4 & 5. These suffered volume swells of 77%, 23% and 61% respectively. The reduction in hardness was markedly increased too; FVMQ showed a 32% reduction , whilst compound refs 2, 4 & 5 reduced by 23%, 14% and 22% respectively. Reduction in tensile strength was a similar story. The FVMQ samples cracked before they could be tested, whilst compound refs 2, 4 & 5 showed reductions of 56%, 38% and 54% respectively.

The reduction in properties is due to the formation of carboxylic acids which have a significant degenerative effect on the metal oxide containing FKM’s, such as the bisphenol vulcanized compounds.

What Trelleborg have shown is that it is very important to consider water contamination in diesel fuel and to carry out appropriate material investigations.