This past summer, I designed the front upright for the JH21XT, Blue Jay Racing's 2024 vehicle. This was an extremely fun challenge, because while I had to connect four points in space (the upper A, lower A, and tie rod mounting points and the wheel bearing location) to a high degree of accuracy, I had significant freedom in how I went about it. 

Because of this freedom in form, I was able to design an upright in SOLIDWORKS and use Ansys to optimize it to be less than half the weight of the previous year's upright, all while making sure it didn't interfere with the rest of the front suspension, was able to take the loads applied to it by a Baja competition, and could be manufactured out of billet aluminum using a 5-axis CNC.

I also learned a lot about proper CAD practices and making the upright as parametric as possible during this design cycle, as whenever the suspension hardpoints were updated, the upright had to change. So, to minimize friction in these updates, almost everything critical on the upright is defined relative to the complex 3D sketch that is the suspension hardpoints.

On last year's vehicle, the JH20XT, we expanded our custom data acquisition system significantly. One data point that we wanted to track closely to improve future cars was the torque applied to the drive axles. So, we decided that the best way to collect this data was to apply strain gages to the axles and measure it directly. However, this meant that we somehow had to get data off of a live, spinning axle.

To solve this problem, I designed a custom watertight enclosure to securely hold a custom PCB and battery and securely clamp to the axle. While these requirements seem simple, coming up with an enclosure that seals for hours at thousands of RPM was much harder than expected. For example, the plan from the beginning was the 3d print the enclosure, but when we submersion tested an FDM part, it leaked straight through the part, so the final version had to be SLA to eliminate this failure

I learned a lot about proper testing procedures during this project because of this.  I went through several iterations to ensure that the enclosure would perform as needed, and that it would clear the suspension assembly as the drive axle rotated. 

In addition to wireless strain gauges, we also had linear potentiometers (linpots) on the 20XT. They were mounted between the frame and the upper control arm, and let us track the position of all four suspension linkages over the course of an entire competition, giving us more information on the state of the car and the loads it sees across the entire suspension, allowing future designs to be lighter and more optimized. 

Because much of our data acquisition system is custom, we had to develop firmware and validate sensors ourselves. This rig that I designed and manufactured allowed us to develop this firmware off of the car, speeding up the development process and freeing up the car for maintenance or other testing. 

The system was designed to closely mirror a suspension linkage. While it looks completely different, all the parts that matter for the sensor are there. It mounts one end of the sensor rigidly, and the other end to an arm with the same maximum and minimum angles as the linkage, making it look fairly similar to a full linkage to the linpot, while using significantly less material and greatly simplified construction methods.

Another type of sensor that we had on the 20XT were hall effect sensors that we used to measure speed. To accomplish this, we cut ridges into the outer profile of all of our brake rotors, then were able to use the frequency that they passed the sensor to directly measure speed of each of our four wheels and collect information on the state of our drivetrain over the course of a competition.

Similar to the linear potentiometer development rig, this allowed for firmware development off of the car. It uses a mock brake rotor with the exact same outer profile as the actual rotors, but with a different inner profile, allowing it to be spun with a shaft that can either be spun by hand or with a cordless drill instead of a full drive axle assembly. This simplified what was needed to test the sensors, making firmware development faster and cheaper.