From 2019 to 2022 I held several roles at KTH Formula Student, an 80-person student team developing DeV17: a four-wheel-drive electric car built around a carbon-fibre monocoque, hub motors and driverless capability. I joined as a trainee split between powertrain and electronics and mechanical design, then concentrated on braking before becoming Head of Mechanical Design.
brake-system constraints
In spring 2020, I worked with two other mechanical engineers on a pedalbox and master cylinder that could blend regenerative and hydraulic braking without giving the driver an unpredictable pedal. The 2020 Formula Student rules made this a constrained mechanical problem: the car needed two independent hydraulic circuits, the pedalbox had to survive 2 kN, brake-by-wire was prohibited in manual mode, and only the first 90% of pedal travel could be regen-only. The final travel had to actuate the hydraulic brakes directly.
Driverless mode added a separate emergency brake system. It stored energy pneumatically so that loss of electrical power produced a brake manoeuvre rather than disabling one. The official EBS reference architecture shows the same safety boundary: software may supervise the system, but redundant actuators and non-programmable shutdown logic must still bring the vehicle to a safe state.
modelling pedal feel
For an ideal closed hydraulic chamber, piston displacement changes pressure through the fluid bulk modulus. The first-order stiffness is
where is the brake fluid's bulk modulus, the master-cylinder piston area and the compressed fluid volume. This was useful for understanding design direction: reducing bore area lowers stiffness, but it also leaves less radial space for springs inside the cylinder.
Model note. The equation is not a complete brake model. Hose expansion, seal motion, caliper and mounting compliance, and the compensation port all affect the measured response. We therefore used pedal-force measurements for the effective stiffness and the equation mainly as a sensitivity model.
Our April 2020 design review recorded an effective hydraulic stiffness of 117 kN/m per cylinder. The first concept needed springs of 23 and 137 kN/m to keep the pedal curve approximately linear through the regen-to-hydraulic transition. The 137 kN/m spring was too large to package, and reducing hydraulic stiffness made the available spring volume even smaller.
revised master cylinder
We replaced the pre-compressed-spring concept with a low-stiffness spring and an initial gap between two pistons. A rotary potentiometer measured the first piston movement and requested regenerative torque; once the gap closed, the second piston built hydraulic pressure. The force-displacement diagram reconstructs that sequence from the original review.
Spring seating created a safety trade-off. Referencing the spring against the cylinder wall gave a cleaner regen-only region but could mask a failed brake line. Referencing it against piston 2 made line failure easier to detect, at the cost of coupling the hydraulic stage earlier. We also kept the design modular: a longer tie rod could bypass the regen mechanism and return the car to a conventional hydraulic configuration.
I documented the calculations and presented the review while the team iterated pedal-plate stiffness, mounting and mass. Moving from four to six plate mounts reduced the simulated plate displacement from 0.26 to 0.089 mm and the corresponding pedal displacement from 0.48 to 0.16 mm. The complete early brake-system estimate was 1.27 kg before the reservoir and EBS masses were known, and we asked Swedish manufacturer ISR Brakes to review the master-cylinder concept for manufacturability.
leading mechanical design
In 2020 I became Head of Mechanical Design. Together with Edvin Hagberg, I led a team of 15 student engineers responsible for the brakes, cooling, gearbox, steering, suspension, uprights and impact attenuator. With the 2020 competition cancelled by COVID, we continued developing DeV17 rather than beginning another car.
I recruited senior and junior engineers in pairs so each project had both ownership and mentorship. Designs lived in Siemens NX and Teamcenter, but being finished in CAD was not the same as being ready to build. I ran manufacture-readiness reviews covering rule compliance, weight, budget, main-assembly packaging, mounting, serviceability, drawings, raw material and manufacturing route.
Those reviews caught steering collisions with the monocoque, incomplete cooling mounts, missing gearbox stock and components that could not yet be assembled. Formula Student taught me that engineering a car is as much about interfaces, manufacturing and communication as it is about getting an individual calculation right.