Taking a Supersonic Selfie

Building a rocket that can reach 30,000 feet is already a serious engineering accomplishment. But Joe Barnard wanted to take his project even further by launching a camera alongside his rocket to capture a cinematic, third-person view of its parachute deployment. The plan involved reaching supersonic speeds, ejecting a stabilized camera pod near the peak of the flight, and recording the action in slow motion. Unfortunately, none of that worked out exactly as planned.

The rocket was designed to reach Mach 2 before deploying its camera payload at apogee. To withstand the forces involved, Barnard constructed the upper airframe from nine layers of fiberglass, providing enough stiffness to accommodate a large opening for the camera deployment mechanism. The lower booster section was built around a four-inch-diameter carbon fiber tube, powered by an N2300 rocket motor loaded with his custom propellant mixture.

Four carbon fiber fins, each equipped with a servo-controlled trailing-edge tab, provided active flight stabilization. These were managed by AVA, a custom flight computer Barnard developed in 2020. The system uses accelerometer measurements to determine the rocket's orientation before launch, then adjusts the fins during flight to control pitch, yaw, and roll.

Separate avionics handled payload deployment and recovery. A flight computer called Fluctus monitored velocity, barometric pressure, and orientation to determine when the rocket had reached its highest point. Multiple GPS trackers and telemetry radios made it possible to locate the rocket after the flight.

The most interesting part of the project, however, is the camera pod. Barnard disassembled a GoPro HERO12 and installed its components inside a 3D-printed enclosure. A sensor extension cable allowed the camera lens to be positioned correctly, and custom GoPro Labs firmware enabled recording at 240 frames per second.

To keep the camera pointed toward the rocket after separation, the pod includes a motor-driven brass flywheel that acted as a gyroscope. A small battery powered the stabilization system, and a pull-pin mechanism activated the camera at liftoff. A streamer and GPS tracker were also included to assist with recovery.

Unfortunately, a combination of seemingly unrelated problems brought the mission to an early end. Post-flight analysis showed that an earlier drop test had shifted the calibration of AVA's accelerometer, causing the guidance system to introduce a persistent 10-degree steering error. Meanwhile, a software change that removed a smoothing filter allowed the control system to excite vibrations in the rocket's structure.

Those vibrations were especially troublesome for Fluctus. Its inertial measurement system sampled data at 50 Hz, matching a strong vibration frequency in the airframe. This produced an aliasing error that convinced the computer the rocket had tilted more than 100 degrees, even though it was nearly upright.

Believing the rocket had tipped over, Fluctus enabled an emergency apogee-detection mode. Pressure fluctuations during transonic flight then fooled its barometer into detecting a descent, triggering the camera ejection and parachute charges while the rocket was still traveling at around 700 mph.

The aerodynamic forces shredded the parachutes and tore the booster away from the upper airframe. Although the mission failed to deliver the expected footage, the flight provided valuable data about the interaction between structural vibrations, sensor sampling, and flight control software. It's an expensive lesson, but one that could help Barnard get the shot he's after on a future attempt.