I design, build, and test hardware that flies and drives.

Fourteen motors for sub-250 g, 3-inch FPV aircraft (twelve 1404-class, one 1504, one 1505), all tested on the same HQProp T3x3x3 propeller. The goal was to find useful performance balance, not just pick the motor with the highest advertised KV or peak thrust.

Method

• Measured KV: no-load RPM at 7 V, read optically. Measured values came in up to 10.4% below the advertised rating.
• Thrust sweep: 0–100% throttle at about 16 V, logging RPM, thrust, torque, voltage, current, and power.
• Efficiency: mechanical output (τω) divided by electrical input, compared at a common 50 W output.
• Torque: an inertial flywheel dynamometer spun from 5,000 to 20,000 RPM, with torque inferred from angular acceleration.
• Responsiveness: repeated 10→50→10% throttle steps, averaging acceleration and braking rates.
• Scoring: cohort-relative scores (100 = average), plus a weight-normalized score for mass-critical builds.

Key findings

• Advertised KV didn't predict loaded performance. Measured KV ranged from 2,915 to 4,428.
• The iFlight XING 1504 had the highest raw score (117.9), but its 11.2 g mass dropped it to 98.5 once weight was counted.
• The T-Motor F1404 4600KV was the best all-around motor: above average in every category and lighter than average.

A static bench comparison of 16 open propellers (thirteen 3-inch and three 3.5-inch) on a 1504 motor, with a 4S battery backed by an external supply to reduce sag. The stand ran in a pusher configuration with optical RPM, load-cell thrust and reaction torque, and ESC-side voltage and current.

Electrical power depends on the motor, so I compared propellers by mechanical shaft power, P_m = τω = τ(2πN/60), with efficiency taken as grams of thrust per shaft watt at a common 300 gf point.

Key findings

• The 3.5-inch group averaged 22% higher efficiency (3.95 vs 3.23 g/W) and 72% more thrust at 25,000 RPM (253 vs 147 gf), at the cost of more mass, torque demand, and frame clearance.
• The EMAX Avan Mini 3X2.4X3 was the best 3-inch package (score 127.7), with by far the lowest vibration.
• The GemFan Hurricane 3525-3 led the cohort in efficiency (4.10 g/W). The EMAX Avan Scimitar 3.5 had the most thrust (299.7 gf at 25k RPM).
• I flagged a mismatch in the vibration reference (200 vs 300 gf) between the notes and the workbook, and treated the workbook as the source of record.

For SDSU Formula SAE Electric Racing, I wrote a planning report for a regenerative, four-quadrant electric dynamometer to characterize the team's EMRAX 208 motor driving a 4.7:1 reduction under steady and transient loads. A second electric machine absorbs or drives the device under test, with measured shaft torque and speed. The report sets the architecture, sizing basis, measurement channels, risks, and staged commissioning before any hardware is bought.

Engineering decisions

• Output torque sensor: I checked the provisional 800 N·m transducer against a 705 N·m ideal peak. That leaves only 11.9% headroom, or 16.3% at an assumed 95% drivetrain efficiency, and those are static margins that inertia, backlash, and torque reversal can use up.
• Motor-side measurement: I recommended a separate 200–300 N·m transducer rather than relocating the output sensor (rated 6,000 rpm against a 7,000 rpm peak), so drivetrain efficiency can be isolated during transients.
• Datasheet envelope: I showed that the headline 150 N·m, 7,000 rpm, and 80 kW figures are independent maxima, not one rectangular operating envelope, so commissioning has to stay inside verified curves.
• Structure: a heavy welding table as the common foundation, with modular motor, gearbox, sensor, and load-machine plates that have locating features and alignment adjustment.

I'm developing an in-house small UAS platform and the autonomy stack it runs, aimed at reliable flight without GPS. I work across the airframe and the flight software, optimizing mechanical properties while integrating autopilot and companion-computer software.

• Configure PX4 autopilot parameters and filters, and tune PID controllers for GPS-denied autonomous flight.
• Program and integrate ROS 2 software with the flight controller.
• Develop the airframe toward NDAA compliance, a 2 kg payload, and ±5 cm positioning accuracy.
• Fly under my FAA Part 107 Remote Pilot Certificate.

Drivetrain structure

• Designed and fabricated motor mounts and drivetrain brackets in SolidWorks, validated with ANSYS FEA.
• Cut assembly mass by 22% while holding a minimum 2.0 factor of safety.

Cooling system

• Redesigned cooling-system packaging and flow paths, cutting power consumption by 42% and mass by 18%.
• Built an 8-channel temperature and flow data-acquisition system to validate the thermal design.

Alongside my drivetrain and cooling work, I'm on the aerodynamics sub-team, which develops the car's wings and underbody to add cornering grip while managing drag and aero balance. Because I sit on both teams, I also coordinate packaging where the aero surfaces meet the cooling ducts and drivetrain.

• Model and iterate aerodynamic surfaces, including wing elements, endplates, and undertray and diffuser geometry, in SolidWorks.
• Run ANSYS CFD studies to compare configurations on downforce, drag, and front-to-rear balance.
• Support fabrication of composite aero components and design their mounting hardware.
• Help plan on-track validation so CFD predictions can be checked against measured data.

• Led an 18-member mechanical team through the design and integration of a 48 kg rover.
• Designed 38+ components in SolidWorks and ANSYS FEA, cutting structural mass by 17%.
• Ran fabrication and tolerance reviews to reach 95% first-pass assembly fit.
• Drove 40 hours of field testing to find and resolve mechanical issues before competition.