To give the driver accurate, low-effort control of front wheel angle with enough feedback to feel the limit, and to do it inside a cockpit that has almost no room.
The driver's inputs pass through a quick-release into a steering wheel, down a two-piece column (column-to-spline shaft and rack-to-column shaft) supported in a bearing housing, into a NARRco rack mounted between top and bottom rack mounts on the frame. The rack drives the front tie rod assemblies, which are CFRP tubes with bonded aluminium inserts and rod ends at each end, into tie rod mounts on the uprights. Tie rod mount position on the upright sets Ackermann; CR26I provides one nominal and three alternate mounting points giving 16%, 23.64% nominal, 30.57% and an additional 23.65% larger-wheel-steering option. The wheel is a CFRP half-shell on a CFRP structural backing plate, 9.22 in diameter, with an acrylic cutout so the dash is visible through it, TPU 3D printed grips, six buttons and two 8-position rotary switches. Paddle shifters mount behind the wheel and request shifts from the pneumatic system.
Steering effort is a fatigue problem, so it gets designed down rather than lived with. Static steering force fell from 37.53 lbf on CR22 base setup to 22.80 lbf on CR26I base, roughly 43%, which puts it near the NASA anthropometric limits for critical operations (18 lbf arm up, 22 lbf arm down). Wheel diameter is a compromise, not a preference: larger diameter cuts hand force for the same torque, smaller diameter improves packaging and reduces hand travel. 9.22 in was chosen to balance leverage, visibility and feedback, with +/-90 deg travel so no driver ever has to go hand over hand. 1. Kinematics sets steering geometry, Ackermann targets and rack travel. 2. Size the column and shafts for torsional stiffness and for the bearing housing that supports them. 3. Calculate static steering force from the tire and geometry, and check it against anthropometric limits. 4. Design the wheel shell around the button and switch layout and the dash sightline; prototype grips and take driver feedback. 5. Validate steering effort and Ackermann on track, using the alternate tie rod mounts to sweep setups.
- Reduce steering effort without losing feedback. - Provide Ackermann adjustability so steering effort setups can be swept and Ackermann validated. - Keep the wheel and column out of the driver's sightline to the dash. - Keep the quick release fast and reliable for the 5 second egress requirement.
- Steering angle is 18.60 deg against an ergonomic optimum of 2.93 deg, forced by design limitations. Tested and found functional, but it remains a known deviation. - Paddle shifter interference on previous designs. Fixed on CR26I by rapid prototyping in PETG and iterating on driver feedback. - Customised per-driver grips were attempted from hand molds, but driver preference converged on a single design, so the customisation was dropped. Steering is the most fault-logged subsystem in the Fault Vault, with six separate entries across CR22, CR25 and CR26I. They fall into three groups. Geometry and joint phasing (CR22). Steering did not lock evenly on both sides because the apex joints were not phased properly, fixed with holes at a 90 degree offset and flipped joint orientation. Strong friction in the centre and near lock came from the Mark Williams geometry raising the operating angle 2 degrees past the joints' maximum, and putting the joints deliberately out of phase did not help. Steering locked at 45 degrees instead of 90 because the KAZ rack has double the ratio of the Mark Williams rack and that was never communicated, which required an upright, geometry and rack mount redesign. Fit and tolerance (CR25). The column to rack connection did not account for clearance between the rack output shaft OD and the column ID, leaving the column free to wiggle laterally and slide fore and aft; 3D printed sleeves were the trackside mitigation, and a thicker bottom tube lathed to fit is the recorded fix. Oversized, off-centre bolt holes in the column added rotational play, with the fix being to drill all holes on the CNC mill. A saddle washer snapped from overtorquing, which oversized its holes and added more play. Vendor knowledge (CR26I). All NARRCO racks have a different output shaft alignment when centred. A planned rack swap before competition forced last-minute column tube remanufacture, which was not done properly and left the wheel slightly off-centre.
Lotus SHARK and WinGeo3 for steering geometry. Altair HyperWorks MotionView for load paths through the tie rods. SolidWorks for packaging and for the wheel shell. Bosch LWS steering angle sensor for on-track measurement of column angle and velocity. Fish scale and static force calculation for effort.