Here's something I just wrote up for another board, and figured it'd be worthwhile to cut and paste it over here:
There's a very good article on Steering Axis Angle and torque steer in the current issue of Racecar Engineering.
One very interesting point is that torque steer is not caused so much by "unequal length" driveshafts, as it is by "unequal angle" driveshafts.
Now, geometrically, if you have unequal lengths you will of course get unequal angles. That is why most people assume that it's because of the lengths. But the real reason is that if the driveshaft and the steering axis angle of inclination are not at 90*, then some component of the driveshaft torque will create a torque about the steering axis. Normally, the forces on the right and left side of the car cancel eachother out, and you don't have torque steer.
But what happens when you have unequal lengths, is you also get unequal angles. And that unequal angle means that one side or the other has a greater component of driveshaft-torque-to-steering-axis coupling, and thus one side of the car has more torquesteer than the other, resulting in a net torque steer to one side that is felt in the steering wheel.
Most modern FWD cars now employ "jack shafts" or intermediate shafts, that brings the inboard passenger side CV joint to the same relative distance as driver's side which is plugged into the side of the transaxle.
Now, it wasn't mentioned in the article, but I would think unequal length would give some torque steer because of "axle windup". The side with the longer shaft(s) is less stiff, so when you apply a transient torque to the shafts, the longer one winds up and there is a delay in the torque reaching the steering axis.
This would also be greatly reduced with the intermediate shaft arrangement however, because the intermediate shaft is usually large and relatively stiff compared to the halfshafts.
It did go on to mention however, that a TORSEN type LSD will create torque steer, because when one wheel has less grip it feeds more torque to the other wheel. This would again result in a situation where the two driveshaft-torque-to-steering-axis-torques are not equal, resulting in a net torque on the steering wheel.
On the Focus RS for example, this is apparently quite strong, and many drivers find it objectionable.
However, it was made worse by a typical Ford cluster****. The wide track of the Focus RS was a marketting driven requirement, NOT as they suggest an engineering driven requirement. Of course, the Ford beancounters got involved and told the engineers it had to be done with as little money as possible.
So, the chassis engineers were faced with the prospect of being required to increase track for looks, without spending any money actually redesigning the suspension. The result was a less-than-optimal front geometry.
One of the ways they did this, was by decreasing the offset of the wheels. This moved them out, to fill the fenders, but it also destroyed the Focus's excellent Zero-scrub geometry.
This means the steering axis intersection with the contact patch, is not centered. So, any longitudinal (acceleration or braking) force on the contact patch results in a torque about the steering axis.
Again, with a normal diff, these two torques on both sides would cancel eachother out. But once you add a Torsen LSD into the mix, you now have the capability of having more torque on one side than the other, resulting in unequal steering axis torques, giving you a net steering wheel torque.
This situation has created a supposedly large amount of torque steer in the RS, which many drivers find objectionable. However, most magazine writers have blamed it on the RS's "new type of differential". Which is laughable not only because there's nothing new about a Torsen diff, but because it's caused more by wheel offset than the diff itself. It's sort of a chicken and the egg scenario... but I blame the offset not the diff.
I've just installed a Quaife torsen diff in my Focus, and don't have any torque steer to speak of.
There's a very good article on Steering Axis Angle and torque steer in the current issue of Racecar Engineering.
One very interesting point is that torque steer is not caused so much by "unequal length" driveshafts, as it is by "unequal angle" driveshafts.
Now, geometrically, if you have unequal lengths you will of course get unequal angles. That is why most people assume that it's because of the lengths. But the real reason is that if the driveshaft and the steering axis angle of inclination are not at 90*, then some component of the driveshaft torque will create a torque about the steering axis. Normally, the forces on the right and left side of the car cancel eachother out, and you don't have torque steer.
But what happens when you have unequal lengths, is you also get unequal angles. And that unequal angle means that one side or the other has a greater component of driveshaft-torque-to-steering-axis coupling, and thus one side of the car has more torquesteer than the other, resulting in a net torque steer to one side that is felt in the steering wheel.
Most modern FWD cars now employ "jack shafts" or intermediate shafts, that brings the inboard passenger side CV joint to the same relative distance as driver's side which is plugged into the side of the transaxle.
Now, it wasn't mentioned in the article, but I would think unequal length would give some torque steer because of "axle windup". The side with the longer shaft(s) is less stiff, so when you apply a transient torque to the shafts, the longer one winds up and there is a delay in the torque reaching the steering axis.
This would also be greatly reduced with the intermediate shaft arrangement however, because the intermediate shaft is usually large and relatively stiff compared to the halfshafts.
It did go on to mention however, that a TORSEN type LSD will create torque steer, because when one wheel has less grip it feeds more torque to the other wheel. This would again result in a situation where the two driveshaft-torque-to-steering-axis-torques are not equal, resulting in a net torque on the steering wheel.
On the Focus RS for example, this is apparently quite strong, and many drivers find it objectionable.
However, it was made worse by a typical Ford cluster****. The wide track of the Focus RS was a marketting driven requirement, NOT as they suggest an engineering driven requirement. Of course, the Ford beancounters got involved and told the engineers it had to be done with as little money as possible.
So, the chassis engineers were faced with the prospect of being required to increase track for looks, without spending any money actually redesigning the suspension. The result was a less-than-optimal front geometry.
One of the ways they did this, was by decreasing the offset of the wheels. This moved them out, to fill the fenders, but it also destroyed the Focus's excellent Zero-scrub geometry.
This means the steering axis intersection with the contact patch, is not centered. So, any longitudinal (acceleration or braking) force on the contact patch results in a torque about the steering axis.
Again, with a normal diff, these two torques on both sides would cancel eachother out. But once you add a Torsen LSD into the mix, you now have the capability of having more torque on one side than the other, resulting in unequal steering axis torques, giving you a net steering wheel torque.
This situation has created a supposedly large amount of torque steer in the RS, which many drivers find objectionable. However, most magazine writers have blamed it on the RS's "new type of differential". Which is laughable not only because there's nothing new about a Torsen diff, but because it's caused more by wheel offset than the diff itself. It's sort of a chicken and the egg scenario... but I blame the offset not the diff.
I've just installed a Quaife torsen diff in my Focus, and don't have any torque steer to speak of.