After the responses I got my original post on this topic I decided to send a
letter to Carroll Smith (of "...To Win" book fame) and ask his opinion. He
called a little while ago and we talked a bit. Recall the two points of
view on this subject:
1. An anti-roll bar is just a linear rate torsion spring. Only the center
part of the anti-roll bar transfers the load and it could care less about
the lengths of each end. The center part of the anti-roll bar sees both bar
ends as one lever (which is the combined lengths of the two bar ends).
Each end is totally dependent on the other because you can't disconnect one
end and have the bar still work on the other.
2. Consider that both wheels are to move the same distance when our perfect
car encounters our perfect speed bump. If both ends of the anti-roll bar
are identically connected to the suspension, both end points of the
anti-roll bar will move exactly the same distance. When this happens, the
net effect of our perfectly frictionless anti-roll bar is now exactly zero.
If a bar is asymmetrically adjusted in the manner implied in the original
post, here is what would happen: Assume that the lever arm on the left side
is 9 inches. Assume that the right end was stiffened by reducing the lever
arm to 8 inches. Say the suspension is to deflect three inches when we hit
the speed bump. Each end of the swaybar would theoretically move through
the same arc having a chord distance of say three inches. The right end
with the 8-inch lever arm moves through 21.24 degrees. The left end with
the 9-inch lever arm moves though only 18.92 degrees. The result of this is
a twist of the working portion of the sway bar of 2.32 degrees. Therefore
there would be a resultant force upward on the left side and downward on the
right side. This means that the net effect of the anti-roll bar in the
speed-bump condition is NOT zero.
Well... Carroll said both points of view are correct in their specific
examples! Its never clear cut is it?
The length of the individual anti-roll bar ends comes into play only if
there is an extreme difference in their adjustment ranges. On most normal
adjustable anti-roll bars the difference between the two adjustment extremes
(one end at one adjustment range limit and the opposite end at the other
adjustment range limit) is not large enough to show any appreciable effect
except, maybe, in the two wheels over the speed bump example.
Carroll said that for all intents and purposes in chassis tuning, you can
adjust one end of the anti-roll bar and not end up with noticeably
asymmetric handling (in vehicle roll). But you can induce some level of
asymmetry for two wheel bumps given an extreme enough adjustment.
When I asked him why that answer appeared contradictory, he just laughed.
----------
John Coffey
johncof@ibm.net
|