Hey I'm Djordje Kokot

I carried both boards from schematic capture and component selection through layout and routing in KiCad. Epsilon amplifies a raw strain-gauge signal so OZYS can measure the rocket's relative deformation in flight, and the isolated CAN interface means a failed board can't take the rest of the avionics bus down with it.

The firmware is written in Rust on the Embassy async framework. The DMA keeps the four-channel 22 kHz sampling off the CPU, and a USB CDC (Full-Speed) interface streams logged data straight to a host PC. Before flight, I simulated the analog front-end in Simulink to verify the design performed as intended.

I was going to run closed loop torque control with SimpleFOC. The idea was to feed the encoder's angle and phase currents into SimpleFOC and use field oriented control to command torque straight from the current, letting the board hold a steady tension on the winch line instead of stepping open loop.

All the hardware was built and populated, and everything I've tested works. The STM32G431 drivers, DRV8844 open loop control, and MT6816 encoder over SPI are confirmed working on the bench. The project got cancelled before I even got to write the code, so the SimpleFOC torque loop and RS485 interface were left to be implemented, while the hardware and tested parts are confirmed working.

⤢