This year I am a suspension engineer for RIT's FSAE team, where I designed and installed subassemblies of the suspension system, including the front anti-roll. The main tradeoffs in this design were clearance, weight, adjustability/serviceability, and manufacturing. The previous year's car had a front anti-roll which was difficult to adjust (softer or stiffer) so adjustability was a large focus of the design. The main things I learned from this project includes strengths of materials, vehicle dynamics, and FEA using ANSYS.
 |
| Complete system installed on the car |
The anti-roll's purpose is to provide resistance to the car rolling (tipping side to side) while cornering, to maintain a mostly even load distribution between each side of the car. In the front, I opted for a Z-bar type design because of its compactness and ease of adjustability. Working with the teams' vehicle dynamics group, I determined the required bar stiffness to achieve a specific roll stiffness for the front of the car.
 |
| Given physical geometry and stiffness goals |
The stiffness goal is a range with adjustability through that range. Part of the design's goal is to make those stiffness adjustments with ease and be able to make repeatable adjustments. To do this, I opted for a set of holes on each end which makes each adjustment discrete and repeatable. After discussing with the VD engineers, five settings in the range are enough.
 |
| CAD screenshot of system |
The team models in Solidworks where I created the subassembly and clearance checker. I mated all of the relevant moving parts and went through each setting and position to ensure the components won't collide with each other or the chassis in operation. The main constraint with this design was the required stiffness and the space given, so I ended up making the bar as long as possible without contacting the chassis, to make the bar as soft as possible.
 |
| Worst-case bar end to chassis clearance in CAD |
 |
| Worst-case drop link to chassis clearance in CAD |
Previous to the design of this system myself and the other suspension engineer on the team tested many of our materials using a tensile tester in a materials lab on campus. We machined and tested specimens to the ASTM E8-04 standard, from several different large pieces of aluminum stock. The data we collected came in very handy during this design as I used it to compare the strength and stiffnesses of the available materials. All of the tested materials are 7075 Aluminum, so the stiffnesses are the same, letting me choose the material with the highest yield strength.
 |
| Test specimens on MTS Insight tensile testing machine and graphed results |
I compared the material properties of the best aluminum I had to other materials, namely magnesium and titanium. The FSAE team is not allowed to machine these materials in the machine shop, and they are also much more expensive as aluminum, but it was still worth it to satisfy my curiosity (our aluminum is pretty good):
 |
| Comparison of ratio of strength:stiffness between materials |
After basic geometry and material selection, the last part of the design is to make the flexure geometry. I knew the length of the bar from doing clearances, and I knew the applied load. From there I calculated the desired deflection for soft and stiff, and the bending stiffnesses.
 |
| Required deflection for a given 125lbf load |
I did a hand calculation to find the moment area of inertia for a beam in bending, given a load P, deflection d, length L, distance from x-axis x, and Young's Modulus E.
 |
| Hand calcs in terms of I |
I set this up such that I had all the inputs to the equation, and could output a moment area of inertia of any cross section at any distance from the horizontal axis in the diagram. The important part is that this beam will have a constant stress due to bending along its entire length, letting me optimize the aluminum to make the softest bar possible in the given space without the bar yielding.
I did just that by assuming a rectangular cross section with a constant height and variable width, letting me establish the relationship between the actual cross section and its distance from the pivot point.
 |
| Calcs relating an arbitrary x distance from the pivot to a thickness value for the cross-section |
I then put a point at each end into CAD to make each side of the bar a straight line. Although not completely precise, the shape was close enough to make a good model.
 |
| Top view sketch with thickness dimensions at each end |
The last part of the design was FEA which I used ANSYS for, with a model of only half the bar for faster solve time. I applied a test force to each of the holes at the end and measured its deflection for comparison to the original goals.
 |
Half-bar ANSYS setup
|
 |
Soft bar deflection in ANSYS
 | | Evenly-distributed bending stress in ANSYS |
|
I fine-tuned the bar thickness slightly, but it only changed by less than 0.01" from the original calculated value! This was my favorite part, seeing the calcs work. The tuning was required to get rid of stress concentrations at each end, and to put the holes in the right places.
From here I calculated how close to the goal I got and discussed these results with the Vehicle Dynamics engineers, which are satisfactory. The bar itself is slightly too soft on its stiffest setting and slightly too hard on its softest setting, however this does not include the other system components like drop-links or fasteners which also contribute (although presumably much less) to the flex of the system.
 |
| System deliverables and margins to goal |
 |
| Finished system being flexed by hand |
 |
| System on completed car |
