Camshaft basics.
A few members have contacted me recently regarding performance cams for their cars, asking what the process involves. Here is a brief look at what goes into selecting cam profiles for a given application.
A good camshaft profile is one which satisfies many different criteria.
The cam must allow the engine to pull from the desired rpm at the bottom end through to the maximum desired rpm up at the top end of proceedings. This can often be difficult to achieve. A good starting point is to look at the current ability of the stock camshaft with its current timing (the placement of valve events relative to crankshaft position expressed in crankshaft degrees), working forward from this point. Porsche has the clever Variocam and Variocam plus systems, which advance and retard cam timing relative to rpm and load request, along with altering the cam profile entirely on the 'plus'. All good stuff but the system has implications on profile design as we will see later.
Assuming we are dealing with engines that are currently in a relatively high state of tune, say 90hp/litre and up, the final design will be matched to the total airflow characteristics of the intake and exhaust systems, bearing in mind that the factory will have already run through this process at the design stage. So the system in place will already be a well-balanced one, making the stock cams a good starting point. In the case where engine size is increased, or airflow characteristics are improved, the cams will need to be altered to match the new specifications.
A larger engine can tolerate a larger cam with little or no ill-effect on low end running, but with much better cylinder filling at high rpm, the nett result of which is no drop off in power as the cam runs out of puff, just a climbing power curve. As mentioned in another post, any time cylinder volumes increase underneath a standard head, the mean port velocities (mpv) will raise at any given rpm. There will come a point where the original cam and port flow will reach its maximum, after which the torque/cylinder filling will drop off a cliff. To counter the fact that the ports may now be too small for the new cylinder size, a larger cam profile can be installed, giving the mixture more time to fill the cylinder. This is a good halfway-house arrangement to stop the inevitable drop off in toque/cylinder filling at high rpm, but not the best solution.
The answer is to evaluate the standard ports and intake system as a whole, as opening port runners to optimise the mpv will lead to greater area under the toque curve IF the rest of the intake runners/manifold/plenum/throttle body and air box allow a greater throughput of air. Put simply, it is no good uncorking the heads if there is a restriction upstream, or down for that matter.
Getting back to the cam profile design, there are many steps to producing a reliable and high performing camshaft. Firstly, all of the components in the valvetrain are weighed so that accelerative forces can be calculated. The spring is also weighed, the total number of coils recorded along with 'active'coils (those that physically move when the valve motion takes place). The springs are evaluated for seat load, the value in lbs or newtons that the spring exerts statically between the spring retainer and cylinder head, along with the 'over the nose' forces at full lift, then the rate is calculated, expressed as lbsf/inch or N/mm generally. Springs are complex things that exhibit natural frequencies, along with 'excitation' frequencies. The latter is the frequency imparted by the profile design.
Once the cam designer has the necessary data he will begin designing the profile to match the engine builder's request and also balancing the airflow characteristics. For example, if a head flows huge amounts of air at low valve lifts but brick-walls at 8mm of lift, the designer will optimise the lift profile below that number. Cylinder heads often reverse flow at high lift numbers in standard form, ie, go backwards and flow less than they did at lower lifts. Some do the opposite, usually as a result of shrouding within a chamber.
There are four main areas he will concentrate on; Displacement (lift), velocity (rate of change of displacement with respect to time), acceleration (rate of change of velocity with respect to time) and the wonderfully named 'jerk' (rate of change of acceleration with respect to time). There is a fifth term known as 'quirk' but nobody utilises this as far as I know.
Of those terms, acceleration and jerk are the ones that have the ability to be destructive. Jerk in particular can put a lot of energy into a spring, causing issues such as coil breakage, whereas high acceleration values can lead to 'float' (separation of components) and high stab torques being fed back to the chain drive. There is a way to envisage the effect jerk has; most of us will be familiar with the little 'kick-back' we get when braking to a standstill in a road car, this is as close an analogy as I can think of to describe what it is, the sudden rate of change of deceleration gives this feeling. Variocam can complicate this aspect of the design, as it has the ability to accelerate the already accelerating valvetrain, so some leeway will be built into the numbers to allow for this.
Many designers will use Fourier Analysis of the spring package, which looks at the excitation frequencies imparted by the profile to the spring, the idea being to use a profile which has low excitation values to enable the spring to have a long life, as in general, the spring is undamped. If anybody fancies looking at a really well designed valetrain at speed, search BMW S1000RR valvetrain in youtube, it is mind blowing! I have some software by Prof Blair and associates which demonstrates the fourier process but do not have the permission to publish it.
The designer will asses the stiffness of the components before embarking on the design, as weak components or design will require a much less aggressive profile to cater for those inadequacies. Finger followers are the stiffest components, direct acting are next followed by wobbly old pushrod and rocker! Which is exactly why the 991.2 GT3 runs finger followers, made even better than before by eliminating the hydraulic element of the 991.1 version. Racing does improve the breed it seems....
Hopefully I haven't missed anything, though this was typed in a hurry, if I have, just jump in and let me know.
Mike
A few members have contacted me recently regarding performance cams for their cars, asking what the process involves. Here is a brief look at what goes into selecting cam profiles for a given application.
A good camshaft profile is one which satisfies many different criteria.
The cam must allow the engine to pull from the desired rpm at the bottom end through to the maximum desired rpm up at the top end of proceedings. This can often be difficult to achieve. A good starting point is to look at the current ability of the stock camshaft with its current timing (the placement of valve events relative to crankshaft position expressed in crankshaft degrees), working forward from this point. Porsche has the clever Variocam and Variocam plus systems, which advance and retard cam timing relative to rpm and load request, along with altering the cam profile entirely on the 'plus'. All good stuff but the system has implications on profile design as we will see later.
Assuming we are dealing with engines that are currently in a relatively high state of tune, say 90hp/litre and up, the final design will be matched to the total airflow characteristics of the intake and exhaust systems, bearing in mind that the factory will have already run through this process at the design stage. So the system in place will already be a well-balanced one, making the stock cams a good starting point. In the case where engine size is increased, or airflow characteristics are improved, the cams will need to be altered to match the new specifications.
A larger engine can tolerate a larger cam with little or no ill-effect on low end running, but with much better cylinder filling at high rpm, the nett result of which is no drop off in power as the cam runs out of puff, just a climbing power curve. As mentioned in another post, any time cylinder volumes increase underneath a standard head, the mean port velocities (mpv) will raise at any given rpm. There will come a point where the original cam and port flow will reach its maximum, after which the torque/cylinder filling will drop off a cliff. To counter the fact that the ports may now be too small for the new cylinder size, a larger cam profile can be installed, giving the mixture more time to fill the cylinder. This is a good halfway-house arrangement to stop the inevitable drop off in toque/cylinder filling at high rpm, but not the best solution.
The answer is to evaluate the standard ports and intake system as a whole, as opening port runners to optimise the mpv will lead to greater area under the toque curve IF the rest of the intake runners/manifold/plenum/throttle body and air box allow a greater throughput of air. Put simply, it is no good uncorking the heads if there is a restriction upstream, or down for that matter.
Getting back to the cam profile design, there are many steps to producing a reliable and high performing camshaft. Firstly, all of the components in the valvetrain are weighed so that accelerative forces can be calculated. The spring is also weighed, the total number of coils recorded along with 'active'coils (those that physically move when the valve motion takes place). The springs are evaluated for seat load, the value in lbs or newtons that the spring exerts statically between the spring retainer and cylinder head, along with the 'over the nose' forces at full lift, then the rate is calculated, expressed as lbsf/inch or N/mm generally. Springs are complex things that exhibit natural frequencies, along with 'excitation' frequencies. The latter is the frequency imparted by the profile design.
Once the cam designer has the necessary data he will begin designing the profile to match the engine builder's request and also balancing the airflow characteristics. For example, if a head flows huge amounts of air at low valve lifts but brick-walls at 8mm of lift, the designer will optimise the lift profile below that number. Cylinder heads often reverse flow at high lift numbers in standard form, ie, go backwards and flow less than they did at lower lifts. Some do the opposite, usually as a result of shrouding within a chamber.
There are four main areas he will concentrate on; Displacement (lift), velocity (rate of change of displacement with respect to time), acceleration (rate of change of velocity with respect to time) and the wonderfully named 'jerk' (rate of change of acceleration with respect to time). There is a fifth term known as 'quirk' but nobody utilises this as far as I know.
Of those terms, acceleration and jerk are the ones that have the ability to be destructive. Jerk in particular can put a lot of energy into a spring, causing issues such as coil breakage, whereas high acceleration values can lead to 'float' (separation of components) and high stab torques being fed back to the chain drive. There is a way to envisage the effect jerk has; most of us will be familiar with the little 'kick-back' we get when braking to a standstill in a road car, this is as close an analogy as I can think of to describe what it is, the sudden rate of change of deceleration gives this feeling. Variocam can complicate this aspect of the design, as it has the ability to accelerate the already accelerating valvetrain, so some leeway will be built into the numbers to allow for this.
Many designers will use Fourier Analysis of the spring package, which looks at the excitation frequencies imparted by the profile to the spring, the idea being to use a profile which has low excitation values to enable the spring to have a long life, as in general, the spring is undamped. If anybody fancies looking at a really well designed valetrain at speed, search BMW S1000RR valvetrain in youtube, it is mind blowing! I have some software by Prof Blair and associates which demonstrates the fourier process but do not have the permission to publish it.
The designer will asses the stiffness of the components before embarking on the design, as weak components or design will require a much less aggressive profile to cater for those inadequacies. Finger followers are the stiffest components, direct acting are next followed by wobbly old pushrod and rocker! Which is exactly why the 991.2 GT3 runs finger followers, made even better than before by eliminating the hydraulic element of the 991.1 version. Racing does improve the breed it seems....
Hopefully I haven't missed anything, though this was typed in a hurry, if I have, just jump in and let me know.
Mike