Development

Testing First Prototype

The testing data was collected with an on board data box which recorded suspension motion data from potentiometers and linear/lateral accelerometers.  Onboard and track side camera recordings as well as testing pictures were taken.  Common processes of testing and tuning such as variation of tire pressures, tire temperatures, camber settings and spring rates were also done during testing.

dash 1st prototype and roll shicks 1st prototype

The results for the rear suspension behavior were as we expected.  The simplified equations (How it Works - Equations) we used during design seemed to be describing the response of the suspension pretty well.  These equations worked both in terms of geometry, overall roll and vehicle response to different combination of spring rates for the roll and dive suspensions.

However, the results for the front suspension seemed quite different as the front wheels lost camber during cornering.  The simulation couldn’t predict this because it didn’t include the steering, castor and yaw factors in its model.

The inner roll suspension rolled towards the center of the turn with respect to the sprung mass.  This reverse roll of the roll suspension meant that it was jacking instead of rolling.  The reverse roll also meant instead of recovering camber the roll suspension was losing camber.  Roll center for the roll suspension wasn’t high enough to cause such jacking on its own and the vehicle didn’t lift enough to cause that much camber loss.  The roll suspensions interaction with dive suspension made the roll suspension jack over the dive suspension.  The roll center for the roll suspension was higher than to the roll center for the dive suspension.  The strut design of the dive suspension had its roll center slightly underground.  Thus, with lateral load dive suspension compressed (rolled out) at the outer wheel, as roll suspension extended (rolled into the turn) and appeared to be rolling reverse, while the vehicle wasn’t.

Unlike conventional suspension systems, increasing the castor angle didn’t help recover the lost camber with steering input.  Adding more negative camber for initial set up didn’t help either.  These approaches were simply not addressing the source of the problem, the roll center location.   

The rear suspension had a lower roll center for its roll suspension; its dive suspension roll center was on flat ground with its initial set up and didn’t move with steering inputs.  In addition the rear suspension wasn't subject to lateral loads that the front was while it was steering/turning the vehicle into the turn.  This final lateral loading related to yaw also increased jacking forces for the front suspension.

Lowering the front roll suspension roll center wasn’t possible due to the existing packaging problems.  Reducing scrub radius or changing the kingpin inclination without changing the dive suspension strut axis could have helped as well, but such modifications weren’t an option either.

roll hoop weightsSince we couldn’t modify geometry of the suspension to help the problem, we decided to raise the CG of the car.  We were able to adjust the CG of the car without disturbing its weight bias by mounting weights on top of the roll hoop.  This solved the problem for the front suspension and once the CG was raised it started to perform well.  Front wheels were able to stay up right during both jounce and roll.

This problem with the front suspension showed us the importance of the interactions of the roll centers of roll and dive suspension systems with each other.  The unique solutions of this problem showed us the importance of the ratio between jacking forces and roll moment on the interactions of roll and dive suspension systems with each other.  In Sacli suspensions the ratio of the jacking forces over the roll moment should never be able reach to a value that would allow the jacking forces to dominate over the roll moment. 

The roll center height and track width determine the direction of the jacking vector while the lateral loads determine the magnitude of the jacking vector.  The roll center to CG distance determines the roll moment arm length and roll moment arm combined with cornering loads determine the roll moment.  A good design should have proper roll center locations and migrations that lead to the desired moment arm lengths while keeping the jacking forces under control.

The lateral loads that come from cornering motion affect both the roll moment and the jacking forces and don’t upset the ratio in between.  While the lateral loads that come from yaw motion only affect the jacking forces, and not the roll moment.  Thus with yaw, the ratio between the jacking forces and roll moment could upset suspension response.

cornering 1st prototypeOur first prototype front suspension had a high roll center for its roll suspension, low roll center for its dive suspension and a low CG.  All leaving very little room for an increase in jacking force without an increase in roll moment.  As the front suspension steered around the corner it turned the vehicle and the lateral loads acting on it increased due to this yaw motion.  In addition, the rear wheels steering into the turn required the front suspension to work even harder to turn the vehicle around, thus further increasing the lateral load on the front suspension due to yaw.  Yaw related lateral loads led to higher jacking forces on the front suspension, yet there was no increase in the roll moment.  As a result the ratio was so off that the jacking forces dominated over the roll suspension roll moment.  The solution of increasing CG height, thus the roll moment, made the roll moment to be higher than the jacking forces and allowed room for an increase in jacking forces.  This allowed the front suspension respond properly even with increase in lateral load due to yaw.