Carry the wheel and transfer its loads into the wheel bearings, while surviving the fatigue that a rotating shaft under bending accumulates.
A rotating shaft running in two deep groove ball bearings inside the upright, with a flange the wheel bolts to. Because it rotates under a bending load, every revolution is a fully reversed stress cycle, which makes fatigue the governing failure mode rather than yield. Design Philosophies and Processes: Determine the minimum shaft diameter for the worst load case considering fatigue life and factor of safety, then size the shaft up to an appropriate bearing and remove as much inner material as possible. Material chosen on fatigue strength: 7075-T6 at 73 ksi yield, 83 ksi ultimate and 23 ksi fatigue strength at 500 million cycles, against 2024-T3 (42.1 / 63.8 / 20) and 6061-T6 (40 / 45 / 14). Criteria: fatigue life greater than 600 miles, and minimum system yield safety factor. From the design reports and one-pagers: The CR22 Wheel Hub System Design Report sets out the design considerations that must be answered before spindle design begins: 1. Surviving required loads. The spindles see the same loads the tires see, and must survive them over many cycles. Fatigue is the thing to research: the spindles cannot fail, because if they fail the car fails. 2. Brake rotor integration. Requires working directly with the brake rotor designer. The team uses a flower petal design on both parts with a floater connection so the rotor can float independent of the spindle and interact correctly with the caliper, and was moving toward a floater/circlip design to reduce weight and improve workability. 3. Bearings. Bearing selection must be finished before spindle design starts, by analysing the given loads. Deep groove ball bearings have been used because they are cheap, sealed and simple to design around, but they are poor at high axial forces and require a spacer in the spindle to handle high moment loads. 4. Rear spindle drivetrain connection, which drivetrain decides and the wheel hub designer accommodates. 5. Interference fit of upright and spindle, assembled using thermal expansion, needing tolerances that give a proper fit without excessive interference increasing power losses. Modelling order: revolve the main body (which puts internal geometry in the same operation and shortens the feature tree), cut the wheel stud and brake rotor connections, make split lines to outline the bearings (invaluable for simulations and assemblies), then make left and right threaded versions as configurations of an unthreaded spindle, using the thread function rather than revolving a cut.
Bore diameter needed to fit over other components. Bearing spacing and its effect on bearing loads counteracting moments. Interfaces that must be considered: wheels, brake rotors, CV inserts or other power transmission, wheel and wheel nuts, and spindle nuts. Positive retention of the wheel and spindle nuts is required.
Analysis flow: wheel loads from lateral acceleration, axial stress from bending, alternating and mean stress, equivalent reversible stress by modified Goodman, then S-N fatigue life. Peak stress at the outboard bearing shoulder, which is where CR22 failed. Torsion from braking does not affect stresses at that location. Calculated axial stress: maximum top 13.4641 ksi, maximum bottom -15.9714 ksi; FEA gives a less rigid reaction because of bearing stiffness. Braking results: peak stress 28.8 ksi, safety factor 2.5. For comparison, CR22 hand calculations gave top axial stress 90.5 ksi and bottom -97.7 ksi. The hand calculation for the spindles matched the full-corner model result most closely; single-body analysis showed singularities. From the design reports and one-pagers: The hand calculation is a simple shaft analysis using the inner and outer diameter of the bearing support area, with a stress concentration factor from Roark's (around page 791) giving a solid baseline. For the simulation: apply remote loads at the tire contact patch onto the faces the wheel sits on, for loads whose moments are resisted by the spindles. Braking and accelerating forces are excluded because the bearings mean they impart no moment on the spindle. Apply cylindrical fixtures around the bearings, with the inboard bearing also restricting rotation.
CAM in Autodesk Fusion 360, turned on the HAAS TL1 lathe. From the design reports and one-pagers: Mostly lathe work, described in the report as straightforward but time-consuming and stressful. The two most important things are the flip between each operation and the bearing tolerances. Two specific lessons recorded: CR22's bearing surfaces came out slightly larger than preferred, and the recommendation going forward is to follow SKF's guidance of very slight clearances on the spindle side of the shaft. When flipping the stock, the stock length entered into Fusion must be exactly the actual stock length. When threading, verify that the pitch in Fusion is actually the pitch desired.
Al 7075-T6. From the design reports and one-pagers: 7075-T6, chosen for high yield stress, high machinability and availability, on the reasoning that the spindle is rotating unsprung mass and therefore one of the most weight-sensitive parts on the car while also being manufacturing-intensive. 2024 was identified as a possible future choice if temperature data shows the spindles getting hot enough to warrant the lower nominal yield stress. Steels were disregarded because of the increase in manufacturing time and difficulty, and a lack of design time to explore what a steel spindle would look like.
Fatigue failure at the outboard bearing shoulder on CR22, which is why the spindle design process is fatigue-first rather than yield-first. Fault Vault, 2023-07-16, CR22: the front left spindle was taken out of service after the wheel nut thread failure and CR21 spindles were installed in its place, because the team had run out of steel to remake it.
Nothing in CrimsonStorage covers these yet.