Quote:
The sum of the power to both wheels is the same, but the split is constantly varied to maintain constant speed around corners.
The only differentials that apply the "same torque" to both wheels are LSD's that have locked up, or "full lockers" that mechanically lock axles together to ensure that all wheels turn at the same rate. This is the only way to get the same torque applied to both or all wheels .
This is still backwards. Open differentials apply the same torque to both outputs, by definition. This can result in different angular velocity of the two outputs. Locked differentials apply the same angular velocity, which can result in different torques. I'll work through the ice example again to explain it.
Let's say you have an open differential, and your right wheel is on ice, and it's spinning at a constant velocity. From Newton's laws, we know F=m*a (or actually, T=I*
a). At constant velocity,
a=0, so we know that the sum of all torques on that wheel is zero. For simplicity, let's say this is really slick ice, and so there is no friction between the tire and the ground. The only torque acting on the wheel is the right output of the differential, so it must be zero. That means that the left output must also be zero, and that explains why the left wheel is just sitting still, not driving your car at all.
Now let's consider the case of an LSD. The torque at the left output is equal to the torque at the right output
plus some torque based on the difference between the velocities:
T_left = T_right + C(T_right-T_left)
So there is some torque applied on the right side, and the right wheel drives you forward off the ice.
In any case, back to the issue of torque steer. By the way, an LSD could actually create torque steer if you partially lost traction in one wheel: the other wheel would be driven with more torque, causing the vehicle to steer towards the wheel with reduced traction. But we're mostly interested in other causes of torque steer, so I'll keep my discussion to open differentials from here on.
Quote:
The popular engineering consensus appears to be that unequal drive axle lengths results in torque steer. As desribed, the longer drive axle twists more (or flexes or perhaps even offers greater resistance due to greater total mass) momentarily with the effect that the wheel on the shorter drive axle actually sees the engine torque a split second sooner and takes off while the wheel on the longer axle is still waiting for the slack to take up in the longer axle assembly. Depending on steering geometry and other factors, this will tend to turn the steering wheel (and the other drive wheel) toward the side of the car with the longer shaft; which serves to amplify the problem because now you are making a right (or left) turn.
Torque steer is caused not by the wheels directly steering the car, but by the force of the drive wheel against the ground creating a torquing moment about the steering axis of the wheel. This is why it is an issue only on FWD cars, since RWD cars do not have pivot points that the drive wheels turn around. The issues you described with a RWD burnout probably have as much to do with unequal friction between the two tires and unequal weight distribution left to right as they do with the principles behind FWD torque steer.
Here's what's happening in torque steer: The torque on your steering wheel is the sum of two steering torques (call them Ts_left and Ts_right) on the two front wheels. These steering torques are the product of the force the tire is applying to the ground (F_left and F_right) and the distance between the center of the contact patch and the axis about which that wheel turns (call that distance d_left and d_right):
Torque Steer = Ts_left + Ts_right = F_left*d_left + F_right*d_right
(note that d_left and d_right are usually opposite in sign, so the sum, in a perfect world, would be zero).
You experience torque steer when one of two things happens: either Ts_left does not equal Ts_right, or d_left is not exactly the negative of d_right. How can this happen? The case I experience most (my car is simply a less-powerful version of that mighty torque-steerer the Viggen) is when I get on a road that has some crowning or some rutting. Let's say that my left wheels are on flat road, but my right wheels are on a patch of asphalt that slopes down from left to right. The effect of this is to shift the contact patch on the right wheel towards the left, or towards the center of the car. All other things equal, F_left and F_right (the differential output torques divided by the radius of my wheels) are equal, but now d_right has gotten smaller, while d_left is unchanged. So the left wheel applies a greater torque about its steering axis than the right one, which makes the steering wheel turn to the right.
Other things can cause differences between d_left and d_right: uneven flexing of bushings, unequal camber or caster of the wheels, and any geometrical imperfection. F_left and F_right might also be different if there is internal friction that is different between the two wheels: an imperfect differential, a bad bearing, or the effects you mentioned from unequal drive shafts. There are so many possible causes for torque steer that what is the biggest contributor in any given car is probably different from one car to the next. The point is that any imperfections in manufacture or in road surface in a FWD car can cause torque steer, so the best way to avoid it is to use RWD. Torque steer in a FWD car can be reduced by using suspension geometries that minimize the possible differences between d_left and d_right. One easy-to-visualize way to do this is to use narrow tires: the contact patch can't move as far on a narrow tire as it can on a wide one. Of course, narrow tires don't look cool
Stiff bushings that keep suspension components where they belong can also help.
Hope that helps some people understand torque steer. Keep two hands on the wheel if you're driving a powerful FWD car hard, but if you're not pushing it, you can generally ignore the nit-picking that Car & Driver and the other mags do.
Matt
'99 Saab 9³
