Projects / Rocket Airbrakes

Rocket Airbrakes

Active drag control for Borealis · Dec 2024

Designed the cam mechanism that drives three panels from one servo, and co-wrote the firmware that steers Borealis to its target apogee.

Designed the cam-slot mechanism that drives all three panels from one servo
Co-wrote the STM32 firmware: state estimation, apogee prediction and PI control
Designed the 4-layer avionics PCB for the canards, raising power efficiency from 66% to 95%

Airbrakes for Borealis, Canada's first liquid-fuelled student rocket. A modified cam-slot spiral drives three panels in perfect sync, with extension linear in servo angle.

Ansys CFD gave C<sub>D</sub> against extension and checked the centre-of-pressure margin. STM32 / FreeRTOS firmware estimates state with a Madgwick filter, predicts apogee with RK4 and closes the loop with a PI controller. The 4-layer KiCad flight computer raised power efficiency from 66% to 95%.

Why airbrakes

The competition rewards hitting the apogee you declare before launch. In 2023, without active control, the team came within 1,500 ft of its 30,000 ft target: 4th closest of 32 teams and best in the hybrid and liquid category. For 2024 the team added airbrakes to steer the rocket onto its target in flight. I worked across the whole subsystem: mechanical, simulation and firmware.

The mechanism

After researching other teams' designs and talking with mechanical engineering professors at Waterloo, I set four requirements:

One relatively low-power servo drives everything
All three panels extend together and are always in sync
Panel extension is linear in servo rotation
As compact and light as possible

A cam-and-slot mechanism met the first two, but a standard slot isn't linear. I modified the slot profile so extension is proportional to rotation, put three of these spirals on one spinning plate, one per panel, and designed the rest of the assembly around it. I did the mechanical design independently.

Simulation

With a teammate I ran Ansys CFD on the panels to check that deploying them keeps the centre of pressure safely behind the centre of gravity, and to map drag coefficient against extension for the trajectory model.

Firmware

With about ten others on the subteam I wrote embedded C on an STM32 under FreeRTOS: state estimation, fourth-order Runge-Kutta apogee prediction, and closed-loop control of panel extension to hit the target apogee.

The flight

The rocket flew, but the propellant was underfilled by more than the airbrakes could correct for, so they correctly stayed retracted the whole flight.

The next board

For 2025 the team moved to roll-control canards, and I designed the new 4-layer central avionics board in KiCad. It raised power efficiency from 66% to 95%, with an onboard IMU over I²C, an external AHRS over UART and a second external IMU over I²C.