Dragonfly V0 booster
Dragonfly V0 compute can
Dragonfly V0 PCB
-
Project Summary
Dragonfly V0 is a waypoint-guided model rocket that utilizes a four-fin aerodynamic actuation system for active trajectory stabilization. At 1 meter long with a 75 mm diameter 3D-printed airframe, the composite-propelled rocket weighs 993 g in total. With a simulated apogee of ~380 meters providing substantial control authority reserve, the guidance system is tested to operate within a nominal target envelope of 30 m of lateral and 100 m of vertical displacement.
-
Guidance & Control
The control architecture utilizes a dual-rate cascaded control loop to decouple 5 Hz GPS waypoint guidance from 20 Hz IMU heading stabilization. The rocket uses proportional navigation coupled with an aerodynamic physics model following every GPS position update to calculate the heading required to intercept the preprogrammed static waypoint’s relative coordinates. The control loop’s second layer samples IMU orientation and calculates the torque required to match the previously derived heading command. An aerodynamic model of the fins then calculates the required fin deflection angles to achieve the required torques. Finally, the deflection angles are clamped and sent to the servos.
-
System Architecture & Electronics
The airframe is made up of four modular, friction-fit blocks designed with tight tolerances for convenient servicing: propulsion/actuation, raceway/power, avionics, and recovery. The engine/fin block has one AeroTech G80T-14A 29 mm solid rocket motor for propulsion and four metal-geared servos arranged in a 90° radial layout to actuate the 3D-printed control fins. The raceway/battery block contains a wall-mounted LiPo battery and routes four 3-wire PWM cables connecting the control surface actuation servos to the avionics. The avionics block contains a removable cylindrical exoskeleton sled, informally termed the “compute can,” that has mounting points for the custom PCB and the Arduino Mega 2560. Replacing the breadboard prototyping setup used in earlier versions, the custom two-layer PCB increases spatial efficiency, eliminates the high volume of jumper wires, and allows for strong soldered connections. The recovery block contains a spring-loaded parachute ejection system with an open-ended parachute canister kept under tension by elastic bands. On electronic command, a servo disengages a mechanical latch from a retaining notch on the canister to release the mechanism and forcefully remove the friction-fit nosecone on the way out, deploying and opening the parachute as it catches the airstream. The recovery block also features four fixed canards parallel to the main control surfaces for additional passive stability.
-
Evaluation and Analysis
While bench testing verified that all electronic actuation systems and heading stabilization loops were fully functional, active guidance was never fully validated. A flight simulation program was developed in the early days of the project to verify this exact system, featuring 3D kinematics, aerodynamic calculations, actuator dynamics, turbulence modeling, and custom mass & acceleration profiles for an exceptionally accurate flight simulation. While passive unguided flight behaved as expected, active guidance simulations much preferred to dive into the ground. It was never determined if this was an issue with the guidance program or the simulation program, as the project was facing more important issues in its last days—namely, the rocket being made out of foam. 3D-printed foaming PLA made up 100% of all structural components (except for the PLA compute can exoskeleton) which resulted in the rocket being exceptionally weak and brittle, for the advantage of being under the weight budget. Furthermore, all structural components that weren’t printed in one piece or friction fit (i.e., the fins) were hot glued on. At 60 cm² per fin, each fin would generate an immense amount of aerodynamic torque with even the smallest angles of attack, immediately detaching from the body before being able to modulate the heading of the greater system.
-
Next Steps
I plan to resume this project in the near future with a full system redesign using my accumulated knowledge. I’d like to integrate better components such as a 25 Hz GPS module (with possible overclocking to 100 Hz) and a magnetometer for eliminating orientation drift and overall more accurate orientation estimation. I’d like to make a fully custom microcontroller PCB with integrated components and a high-performance processor, enabling a tighter power budget and a much more aggressive 100 / 500 Hz dual-rate cascaded loop mimicking the original control architecture. The airframe would be made of cardboard with custom structural components being made of PETG for a much more structurally robust rocket, while maintaining a modular sliding design on the inside.