Everyone has seen it: an old VW Rabbit heading into a turn with the inside rear wheel a foot in the air, or a Porsche 911 coming out of a turn picking up the inside front wheel. A car on three wheels always makes for a good photograph, but what’s really happening when the rubber quits meeting the road, and how can kinematics and compliance testing help you understand it?
To understand what happens when a car lifts a wheel in a turn, it’s important to know why it occurs in the first place. The goal of suspension tuning on a race car is to provide the best handling possible, whether it’s in a corner, braking or the all-important “straightaway handling”. Wheel lift during acceleration (and in rare cases, braking) is a sign you’re doing things right, so in this discussion we’ll focus on cornering behavior.
In a turn the car generates lateral force at the tire contact patch, which is transferred to the chassis and causes the lateral acceleration through the turn. Since the center of gravity (CG) of the vehicle is above the ground, this creates a roll moment which must be resisted by the suspension in the form of roll stiffness: the resistance to roll motion of the suspension. This stiffness can be expressed as degrees of roll per G of lateral acceleration. In order to corner most effectively each tire must generate the most lateral force it can. The ratio of lateral force versus vertical force of a tire decreases as vertical load increases, and so it is desirable to keep the vertical loads on all four tires as equal as possible. The total amount of weight transfer for a given vehicle at a given lateral acceleration is constant, dictated by the track width and the height of the CG. While we don’t have a choice about the total weight transfer, for a vehicle on four wheels we have a choice how much of it each axle resists, as it is not statically determinate. By increasing the roll stiffness on one axle of the car we can create a greater imbalance in the vertical loads on those wheels and decrease the overall grip on that axle. A large part of suspension tuning is focused on finding the right combination of front and rear roll stiffness that gives the car the desired handling characteristics. This is usually described as a car that is fairly neutral, without an excessive imbalance between front and rear grip.
There are as many different suspension setups as there are cars in racing. What makes them work? Consider a vehicle with a 50/50 weight distribution, so that the front and rear of the car support an equal amount of vertical load. In order to generate the most lateral force possible the wheel loads should be kept as equal as possible as the car resists the roll moment. On our theoretical car this would be achieved with equal roll stiffness front and rear, so that when the car reaches equilibrium in the turn the front and rear axles have contributed equally to the roll resistance. In reality this is rarely the case because we are rarely in a pure steady-state cornering condition. In most cases there is some longitudinal acceleration thrown into the mix in the form of braking or traction. This means we have to modify our idealized cornering setup to give the desired handling under these combined states. In order to generate longitudinal force the tire must borrow some grip from the lateral force used in cornering, so we want to have a bit on reserve when we need it. On a rear wheel drive vehicle, we can make the front roll resistance a little stiffer. When we do this the rear axle is left doing less of the total roll resistance and so is capable of generating longitudinal force even when the front is fully saturated in a turn. This helps when exiting the corner on the throttle, giving the rear tires some traction so they don’t send the back of the car sliding out of the turn as soon as power is applied. It helps when braking as well, as the front of the car will tend to lose grip before the rear and the car can corner while braking without the rear of the car trying to pass the front. We now have a car that has a little understeer built into it while still behaving well under race driving conditions. No one likes a car that tries to kill them.
In reality, very few cars have a true 50/50 weight distribution. Whether it’s a rear-engine Porsche, a mid-engine formula car, a front-engine Civic or a Mustang with a heavy V8 on the nose, most cars have a heavy end and a light end. When choosing front and rear roll stiffness of a vehicle, it becomes a game of adjusting the work done at each end to make the car handle well. Since this is a discussion of wheel lift we’ll focus on the primary culprit: front wheel drive race cars.
A front wheel drive (FWD) vehicle naturally has a higher front weight percentage: typically around 60% of the total vehicle weight rests on the front wheels statically. If a FWD race car was set up with roll resistance proportional to the weight on each axle this would result in lots of understeer as well as traction issues when cornering, as the weight transfer would be greater on the front axle than the rear. To combat these tendencies FWD race cars are set up with a higher roll resistance on the rear axle than the front. This causes the rear axle to support more of the roll moment than the front axle, leaving the front tires with extra grip that can be used for acceleration. An obvious downside of this type of setup is that it limits the amount of braking force that can be applied while turning, as anyone who has raced a front wheel drive car can attest to.
There is a limit to this game of stiffening the rear axle however. Under constant lateral acceleration each axle can only transfer as much weight as it has available statically. In other words, the total axle load remains constant as the vehicle corners. Obviously the rear axle has less load to transfer than the front, around 40% of the vehicles weight. Once the rear axle has transferred all the weight it has available the inside rear wheel lifts off the ground and any additional weight must be transferred at the front axle. Since the car was set up with a much lower front roll stiffness, it is left with the challenge of having to resist the roll moment of the same vehicle mass but with a fraction of the roll stiffness available. Remember that due to the much higher roll stiffness of the rear axle the remaining front roll stiffness is not half of the initial stiffness. Instead, it is the percentage of the front axle stiffness relative to the total stiffness, which may be as little as a third or less. With the full mass of the car only being supported by a third of the roll stiffness, the roll angle per G of lateral load is much higher and any small increase in lateral acceleration results in a huge amount of body roll.
So what does this mean to the behavior of the car? Well for one, the increased roll motion translates into positive camber gain on the outside tires, which is a sure-fire way to lose grip. Along with the roll motion comes the problem of chassis heave. Because the rear axle has no more weight to transfer the car cannot roll further onto the outside rear wheel, and instead just pivots about it. Anyone who has seen a VW Rabbit teetering through a turn is seeing just that: the car is rocking along the outside rear and inside front wheels and the chassis is bouncing upward and pitching forward as it does so. This increase in bounce travel raises the CG, further increasing the weight transfer and exacerbating the problem. Luckily this feedback loop is interrupted by the outer edge of the outside front tire grinding away from the positive camber gain, and the front of the car begins to slide as the driver complains of understeer. And that’s when things can get a bit out of hand.
A car on three wheels is statically determinate. This is an important point to note because it means we have lost our ability to tune the balance of the car with roll stiffness. But without realizing the situation they are in, many people try anyway. They may increase the stiffness of the rear axle in roll, which only serves to make the problem worse as it runs out of weight to transfer even sooner. Remember that the rear is already at a disadvantage, as it is the lighter axle and so has less weight to transfer in the first place. They may try some form of droop limiting or preload, hoping that it will “hold the inside rear wheel down”, but again physics isn’t on their side. Preload and droop limiting just reduce the amount of travel available before the inside wheel rear runs out of vertical load, and with no mechanism for the tire to pull the car downward it is perfectly happy hanging out in the air watching its other three friends do all the work. A droop limiter is just a much higher wheel rate and only serves to cause a lot of weight transfer on the rear in a tiny amount of travel while the front doesn’t contribute much. This can also create a very non-linear behavior in the suspension which can be hard to predict and tune around.
If so many things can make this effect worse, what can be done to fix it? Fundamentally we need to keep a four wheel vehicle on four wheels, even if the inside rear is only touching the ground as a token gesture. The first thing that will make all subsequent tuning easier is to eliminate any preload or droop limiting of the rear wheels, so that when the tire achieves zero load the spring is fully unseated and has run through its full range of linear travel. Moving as much weight to the rear as is practical will also help, giving the rear more weight to transfer in a corner. Once that is done, if the inside rear is still lifting it is actually time to make the front stiffer. While it may seem counter-intuitive based on driver feedback of the car plowing across the track, it will serve to reduce the roll angle of the car and correct the camber without changing the load distribution. The goal is to sync up the inside rear wheel running out of load to transfer with the entire car running out of lateral grip, so that no more load transfer will occur and the car is left on all four wheels. In a FWD car the driver can buy a bit of rear load by applying the throttle. This causes forward acceleration, transferring load to the rear tires and helping the inside rear stay touching the ground. With enough power it might be desirable to compromise corner entry by lifting the inside rear in order to have the car in the optimum spot on exit when the driver gets back on the throttle. By tuning the car to get the inside rear just touching under corner exit acceleration you have maximized the inside front load under acceleration, which can aid traction.
Up to this point the discussion has focused on front wheel drive cars, but this is akin to having a higher front roll resistance on a rear wheel drive (RWD) car in order to reduce the weight transfer on the driven wheels. Similar to the FWD scenario, a front-stiff setup on a RWD car can be taken too far as well. This is especially true on mid and rear engine cars which don’t have a lot of weight on the front. The bottom line is that a car on three wheels probably isn’t doing what you hope it will, and it can be a long and frustrating day getting it sorted out.
How does kinematics and compliance (K&C) testing fit into this equation? By performing basic bounce and roll tests, and using the unique capabilities of the Cornering Simulation test, some very useful information can be uncovered. In the bounce test the chassis is moved straight up and down, usually to the point of the wheels lifting off the pads. This allows you to instantly see the amount of droop travel available, as well as the amount of preload in the suspension, if any. As an added bonus, if the springs unseat without any preload you get to see the unsprung mass of the suspension. A roll test starts to look more directly at the behavior of the suspension in a turn. By performing a natural axis roll test, which keeps the axle loads constant as the vehicle rolls, we can begin to see some key characteristics of the suspension.
Of obvious importance is the roll stiffness of the front and rear axles. This gives an indication of the amount of weight transfer each axle will provide as the vehicle rolls. The plot of Roll Moment on the right comes from a roll test run on the K&C rig. It shows the roll stiffness of a vehicle and the effect that lifting the inside rear wheel has on the rear and total roll stiffnesses. As the vehicle begins to roll to the right the roll moments of the front and rear increase together. At point A the rear roll moment starts increasing at a higher rate, which indicates the droop limiting has started picking up the left rear wheel. The total roll stiffness, which is the sum of the front and rear, also reflects this higher rate. But at point B the left rear wheel lifts off the ground, causing the rear roll stiffness to completely flatten out. When this occurs, the total roll stiffness also gets much softer. The slope of the total roll stiffness is seen to be the same as the front roll stiffness from point B onward.
There are several other ways to examine roll stiffness, including graphs of weight transfer per degree of roll or the wheel rates in roll. A secondary calculation is the Static Roll Weight Transfer Coefficient (SRWTC). This is a fancy name for a pretty simple metric: the roll stiffness of each axle normalized by the amount of weight on that axle. A natural axis roll test will also show the dramatic change in chassis motion when a wheel lifts, but there is another test that gives an even better simulation of a cornering event.
A Cornering Simulation test is just that: simulating a cornering event on the K&C machine. By applying the lateral force seen in a turn combined with the roll moment the chassis must react, it creates a very accurate depiction of a steady-state cornering event. By including lateral forces as well as chassis roll, the effect of jacking forces from the instant centers is included in the measured roll stiffness. These can be a significant contribution by changing the chassis motion from that predicted by the springs and anti-roll bars alone. As with the natural axis roll test, the Cornering Simulation will show the chassis motion as the car rolls. If a wheel lifts the rate of chassis roll per G of lateral acceleration will rapidly increase, as seen in the plot of roll angle versus lateral acceleration measured in a Cornering Simulation test on the K&C rig. As the lateral acceleration starts to increase the roll angle steadily increases. At point A the left rear droop limiter picks up and the rate of roll decreases from the momentary increase in roll stiffness. Then at point B the left rear tire picks up and the chassis roll angle increases dramatically. This illustrates the teetering motion of the car as it falls over on the outside front tire.
As the rate of chassis roll increases, the rate of camber loss on the outside wheels increases as well. Through different combinations of springs, anti-roll bars, bumpstops, and instant center settings, the motion of the chassis can be tuned to achieve the desired condition: using all of the stiff axle’s weight transfer just as the whole car runs out of grip. At the same time, extremely accurate measurements of all the parameters of the car are recorded, giving a wealth of information that can be used when matching the car setup to the track conditions. What’s not to like about repeatable, objective data that is almost impossible to obtain anywhere else, measured faster and cheaper than track testing?
When tuning a car we must work with physics, as we don’t have any other choice. Trying to limit chassis roll by further stiffening the stiff end of the car will only magnify the problem, and using preload and droop limiting to try to “hold that wheel down” also makes the issue worse. Understanding the true causes behind suspension behavior is critical in making the right setup and tuning decisions, and K&C testing offers some helpful shortcuts and insights in the process. Making a car go as fast as possible isn’t an easy job, so why not grab any advantage you can?