H.Y.D.R.A. Hardened Quadcopter

All parts ordered; dev environment installed and verified on bench laptop

Place every component order immediately to account for lead times. The u-blox NEO-M9N-00B bare module and Parker Chomerics copper shielding tape are the longest-lead items (often 2-4 weeks from distributors like Mouser/DigiKey); order these first. Simultaneously install ArduPilot Mission Planner, PX4 Autopilot toolchain, Arduino IDE, and KiCad so the software environment is ready before hardware arrives.

Custom PCB submitted to fab; airframe received and dimensioned

Front-load the PCB order now to absorb the 2-3 week fabrication lead time. The custom PCB will carry the u-blox NEO-M9N-00B backup GNSS chip, Rubycon bulk capacitors, AVATEC MLCCs, Murata ferrite beads, and the LM7805CT regulator. While the PCB order is placed, perform a physical inventory of the airframe hardware and lay out a preliminary center-of-gravity diagram to confirm component placement before anything is bolted down.

Airframe fully assembled; motors, ESCs, and props spun up on bench safely

Phase 1 begins: assemble the quadcopter airframe, mount brushless motors, wire ESCs to the power distribution board, and perform a propeller-off bench spin test to confirm motor direction and ESC calibration. This comes before any flight controller work because a mechanically sound, correctly spinning propulsion system is the prerequisite for all tuning that follows. Remove props during every bench electrical test.

Flight controller configured; stable hover achieved in Stabilize mode

Mount and configure the flight controller running ArduPilot Copter, calibrate both IMUs, barometer, and magnetometer, then perform first outdoor hover in Stabilize mode with a safety pilot. This step establishes the baseline flight envelope before any hardening layers are added, making later A/B comparisons meaningful. Log all sensor data during hover for the Phase 6 baseline dataset.

Autonomous loiter, RTL, and waypoint mission verified; baseline logs captured

Enable the SparkFun NEO-M9N Qwiic GPS, tune Loiter and AltHold modes, and fly a simple autonomous waypoint mission followed by a Return-to-Launch sequence. Capturing clean baseline telemetry logs now gives the team a quantified reference point — position error, attitude deviation, packet-loss rate — against which every hardening phase will be measured in Phase 6 testing.

Custom PCB received, soldered, and bench-tested; power rails verified

The fabricated PCB returns this week. Solder the LM7805CT 5V regulator, all 100 AVATEC 100nF MLCC decoupling caps, all 50 Murata BLM18PG471SN1D ferrite beads, the six Rubycon 1000µF bulk capacitors, and the u-blox NEO-M9N-00B backup GNSS chip. Verify all power rails under load with a multimeter and oscilloscope before installing the board in the airframe, because a rework on the bench is far cheaper than a mid-flight power failure.

Copper shielding enclosure built; EMI attenuation measured on bench with SDR

Phase 2 begins: wrap the electronics enclosure with Parker Chomerics copper-foil shielding tape, install ferrite beads on all power and signal lines, apply RF-absorbing foam to interior surfaces, and route shielded cables with proper grounding. Use a software-defined radio (SDR dongle) and a signal generator or known interference source to make before/after field-strength measurements inside the enclosure — this produces the first quantitative shielding effectiveness data point required by the project spec.

FHSS RC link configured; redundant telemetry via Arduino Nano Every operational

Phase 3 begins: configure the frequency-hopping spread-spectrum RC link (e.g., ExpressLRS or TBS Crossfire on 900 MHz / 2.4 GHz) and integrate the Arduino Nano Every as a secondary telemetry and fault-log node that mirrors critical state over a separate serial channel. Program automatic failover logic so the Nano Every can command RTL if the primary link drops for more than a configurable threshold (default: 3 seconds). Test link switchover on the bench by physically blocking the primary antenna.

Arduino Uno R4 Minima logging firmware complete; dual-GPS failover bench-tested

Phase 4 begins: program the Arduino Uno R4 Minima as the companion logging MCU — it reads MAVLink telemetry, timestamps interference events, and logs signal-quality metrics to onboard flash/SD. Simultaneously configure ArduPilot's dual-GPS blending between the SparkFun NEO-M9N Qwiic (primary) and the NEO-M9N-00B on the custom PCB (secondary), then simulate primary GPS dropout on the bench by unplugging the Qwiic module and confirming the flight controller seamlessly switches to the backup. Dual-GPS failover must be verified before any GPS-degradation flight tests.

Fault-detection state machine validated; all failsafe modes bench-tested

Phase 5 begins: implement and test the full software state machine covering normal flight, degraded comms, GPS uncertainty, sensor disagreement, component failure, RTH, emergency land, and mission abort. Use ArduPilot's scripting (Lua) or companion-computer logic on the Uno R4 Minima to cross-check dual-IMU outputs, flag disagreements beyond a configurable threshold, and command the appropriate failsafe action. Simulate each fault condition on the bench using sensor disconnects and injected bad data before any of these states are trusted in the air.

HYDRA completes autonomous RTH after simulated GPS and RC dropout

Fly the fully hardened HYDRA airframe in a controlled outdoor area, deliberately inducing GPS dropout (by commanding GNSS_TYPE off via MAVLink) and RC link loss (by powering off the transmitter) while the aircraft is in Loiter, confirming autonomous RTH executes within the target recovery time. This is the first end-to-end integration flight and is deliberately done before formal interference testing so any integration bugs can be fixed in a zero-interference environment.

All bugs triaged and resolved; shielding effectiveness re-confirmed on bench

Dedicated mid-project testing and debugging week. Review all data logs from Weeks 4–11, triage every open bug (ESC noise, GPS re-acquisition lag, failsafe timing, shielding gaps), and close or formally defer each item. Re-run the EMI attenuation bench test with the fully integrated enclosure to confirm shielding effectiveness hasn't degraded after integration. No new features are added this week — only fixes, tuning, and documentation of known limitations before Phase 6 interference testing begins.

Controlled EMI tests complete; all performance metrics quantified and logged

Phase 6 controlled interference testing: using only legally compliant, Part 15-limited signal sources (e.g., a shielded anechoic box, attenuated signal generator, or campus EMC lab), progressively expose HYDRA to 2.4 GHz, 900 MHz, and broadband noise. Measure packet-loss rate, control latency, position error, attitude deviation, RTH trigger time, and shielding attenuation. All intentional RF emission must remain within FCC and campus safety limits; coordinate approvals with campus safety and FAA before this week begins.

Final demo flown; full technical report and hazard analysis submitted

Final two-week sprint concludes with a live demonstration flight showing normal operation, simulated link loss, autonomous RTH, and recovery — all in front of the sponsor and faculty. The deliverable package includes the full technical report (problem statement, architecture, BOM, shielding strategy, test results, and hazard analysis), all firmware source code in a version-controlled repository, wiring and grounding diagrams, a video recording of the demo, and a lessons-learned document with sponsor clarification items answered or escalated. All regulatory confirmations (FAA LAANC, campus UAS, FCC) must be on file before the demo flight.