> To a first approximation, rollover is a simple matter of loading geometry. If
> the roll torque is more than the stability emparted by gravity acting on a
> wide-based (i.e. the track width) platform, then the car will roll. This is
> true even for solid-axle cars, but a swing-axle car has the added disadvantage
> that if the wheel does tuck under, the platform's width decreases! The simple
> forces are thus:
>
> lateral force L
> <--------* (Center Of Mass, assuming that front and rear are alike)
> a / |
> / b | - gravity G
> / |
> W V
>
> W is the wheel. If L * sin(a) > G * sin(b), then the car will go over! Any
> second-order suspension-change effects will modify this slightly. Consider
>the
> limiting case of a car that slides sideways into a curb! This is nothing more
> that the lateral force going very high, since the tire can't ride up over the
> curb. This is guaranteed to roll almost any car if it hits hard enough.
Nice picture, but it does leave out a potentially important factor:
Roll inertia. If your moment of inertia about the roll axis is high,
and the car develops significant angular velocity during suspension
travel, when you run out of travel the bump stops may not stop the roll.
There are several ways to counteract this effect:
i) the JSY method: smOOOthly initiate and terminate your turnin, so
that trasient torques along the roll axis are minimized.
ii) BIG anti-sway bars.
iii) small moment of iniertia along the roll axis.
i) and ii) above are the oens that generally apply to a vehicle that
is already in hand, as opposed to "on the drawing board".
--
---
John R. Lupien
lupienj@hpwarq.hp.com
|