There's no doubt that the valvetrain is the most complex part of any racing engine. It's also pushed the hardest against some incredible stresses. Nextel Cup teams and some manufacturers are putting tremendous amounts of resources and some pretty innovative research systems into place to advance the science of the racing valvetrain, and the things they are learning are quite interesting.

One of the companies leading the development front is Comp Cams, which not only works with many Nextel Cup teams on their engine programs but also has its own R&D program that it uses to advance its products used by the Saturday night racer. Comp's engineers have pushed the use of the Spintron machine and complex mathematical models to better understand what's going on with all the various parts of the valvetrain when the rpm starts creeping into the upper limits. This story isn't a rehash of all the stuff you slept through in physics class in high school. Instead, there's a lot of practical knowledge that can be learned-and used right away-by the computer modeling and Spintron testing done behind the scenes at Comp's R&D lab.

Thomas Griffin, who is one of Comp's lead design engineers, says that one of the keys to finding ways to improve the valvetrain is to be able to look at it as an entire system. He points out that an engine is comprised of many interrelated systems. For example, the ignition system is made up of the distributor drive gear, the distributor shaft, rotor, cap, wires, plugs, and coil. Each piece is required to ignite and burn the air/fuel charge in the combustion chamber. If one piece is missing, the entire system fails to work. Likewise, if one piece of the system is functioning poorly, the performance of the entire system will suffer.

The same holds true for the valvetrain, which is made up of more components than any other system. In a pushrod-style engine (typical for oval track racing), this system is comprised of the drivebelt or chain, camshaft, cam lobe profile, lifters, pushrods, rocker arms, valvesprings, valves, retainers, and locks. Just like with the ignition system, all the parts of the valvetrain must work individually and together as a complete system. This duality of roles can cause problems that are often missed. Griffin says that it isn't uncommon to make a change to a component-say a valvespring retainer-that is an improvement (maybe a reduction in weight, or an increase in strength). But when that "improved" component is added to the system, overall performance suffers.

This is because our new "improved" piece doesn't play well with others. This doesn't mean that the new part is a poor choice for every valvetrain system, just this specific one it was tested with. Because there are so many variables involved (everything from lash to the stiffness of the pushrods to the weight of the valves), you can never assume that identical changes to similar valvetrains will produce identical outcomes.

As an example, Griffin points to a test in which lighter valves were used in a Spintron test to try to achieve better valve control at high rpm levels. This is proof that sometimes when an individual component is changed, the system result is something other than what is desired. This is because it can often be hard to predict how the new combination of components will react with one another.

Individual Components versus the System

Reducing component weight is one of the Holy Grails of most engine builders because it often allows increased engine speed. Unfortunately, one of the side effects that often comes from reducing the weight of a component is decreased rigidity. This means it is more easily bent or twisted, and it also changes that object's natural frequency. The natural frequency of an object is the rate at which it vibrates. We'll delve deeper into the importance of natural frequency a little later.

In the following two illustrations (Figures 1 and 2), Griffin tracked and compared the amount of valve bounce of two different valves at different engine speeds. In the first figure Griffin compareid two steel valves, one weighing 115 grams and the other weighing just 99 grams. At first, you would think that the 99-gram valve would behave better because its lighter weight should mean that it would be more easily controlled by the spring. This, however, isn't the case. The lighter valve exhibited significantly more bounce between 5,600 and 7,200 rpm. There can be many reasons for this, but the most probable is that the thinner head of the lighter valve provided a "springing" action when it was pulled against the seat, creating a bouncing effect.

In Figure 2, the same test has been performed with two titanium valves, in this case weighing 97 and 85 grams. Here, Griffin explains that the amount of bounce changes with engine speed in two ways. First, the bounce is reduced almost everywhere. Second, the speeds where the bounce is greatest is different. In this case, the lighter valve seems to help where, in Figure 1, this isn't the case across the entire rpm range.

The same type of surprises can also be found with valvesprings. Comp Cams has tested an untold number of different valvespring combinations, and the results show that a group of springs that measure the same on the seat and over the nose don't produce the same results.

In racing, the double spring is most commonly used because it allows for greater combined spring pressures. What many people don't realize, however, is that different springs can use different combinations of inner and outer springs to achieve the same total pressures. In other words, two sets of springs may measure the same when checked as assemblies, but when you break them down into individual components and compare the two inner springs or the two outer springs, they may be completely different. The weights, stiffness, and natural frequencies may all be different. Griffin provided this graph (Figure 3) plotting bounce with two different springs that check out the same as complete assemblies but are quite different as individual components. This is why a spring set from one manufacturer may have all the same specs as one from another manufacturer but behaves completely different than the other set once it is installed in an engine that's in a race car or put on a dyno. It's not until you break everything down to the individual component level that the real reason for the performance difference becomes visible.

Finally, Griffin says the same thing holds true for camshafts. A cam profile that produces horsepower or rpm gains in one engine may reduce power in another with a different valvetrain setup.

"For instance, a little more peak positive acceleration on the profile in one engine might pick up the power over the entire operating speed range. Applying the same philosophy to a cam profile in another engine might hurt the power all over the board and make the engine lose a few hundred rpm," says Griffin.