[Walt Fricke]

Burgermeister

My modest understand of levers (or arms) says that no matter what the shape of the lever arm is, its effective length (assuming rigid materials) is the distance between where the load is applied (tire contact patch center) and where the lever articulates. In the case of a 911 front end, this is the axis of the A arm through the chassis mounts. Which for all practical purposes is a perpendicular from the patch to the centerline of the torsion bar if bushings are reasonably stiff.

So if you move the patch out (widen the track of the car), you are using a longer lever. Easier to twist the T bar. Softer.

Now what this ignores is the fact that the spindle is not a rigid part of this total "lever" length. We know from scrub effects that as the tire moves up and down it can move in and out, changing track. And this can mess up handling if the geometry is bad and there is much of it, and so on.

Still, is it a reasonable first approximation to treat the ball joint and strut side loading as if they were frictionless? So, other than accounting for track changes with suspension movement, you could take the perpendicular length from the A arm mount axis to the center of the contact patch as your lever length when computing wheel rate based on a given torsion bar diameter and length?

Your long ago (so you say) learning has to be a couple of steps ahead of my lack of formal engineering education. Free body diagram? Gee, I've seen engineers use the term.

Walt