Problem
Can a toroidal prop reduce perceived noise while maintaining or improving efficiency vs a standard 3-blade?
Approach
Sensorized thrust stand: thrust (load cell), V/I/Power (INA219), RPM (laser tach), SPL (dBA). MATLAB post-processing.
Result
Toroidal shows better efficiency and lower noise at low RPM, but trends reverse at higher RPM (efficiency falls, SPL rises).
Methods (Rig & Sensors)
- Actuation: A2212 1000KV BLDC + 30A ESC, GPS-2303 bench PSU.
- Thrust: 1 kg strain-gauge load cell + HX711 ADC.
- Electrical: INA219 for voltage, current, and power.
- Speed: NEIKO 20713A laser tachometer with reflective tape.
- Acoustics: dBA via Decibel X app (iPhone SE), off-axis measurement.
- Procedure: Step RPM bands, log thrust/V/I/SPL, compute thrust/W; repeat for 3-blade and toroidal props.
Load Cell Calibration
Linear fit between known masses and readings: y = 0.98854·x − 0.019826 ± 1.0385 g (95% interval). Calibration was done with 10–82 g references and validated over 10 trials.
Interpretation: scale is near 1:1 with small offset and ~±1.04 g prediction interval across the span.
Representative Results
| 3-Blade: RPM | Noise (dB) | Toroidal: RPM | Noise (dB) |
|---|---|---|---|
| 1855 | 66 | 1115 | 65.5 |
| 2636 | 71.5 | 1500 | 68.3 |
| 3288 | 72.5 | 1805 | 77.0 |
| 3779 | 78.3 | 2087 | 74.7 |
Trend summary: toroidal is most competitive at lower RPM bands (higher thrust/W and lower dBA). As RPM climbs, the toroidal’s more aggressive ramp angle increases acoustic disruption and reduces efficiency; the standard 3-blade overtakes it at higher thrust/RPM.
Notes: values reflect one test rig and specific prints; results vary with prop size, pitch, motor, and stand geometry.
Theory (Brief)
Blade-element theory integrates sectional lift along the radius to estimate thrust; with our stationary setup (V≈0), angle of attack aligns with blade pitch. Toroidal geometry effectively eliminates a discrete wingtip, redistributing vortices and potentially lowering induced drag and perceived noise at low RPM. Model trends captured thrust ∝ n² but absolute magnitudes deviated from experiment due to chord/angle distribution and stationary test assumptions.