Solution Research — Design & engineering

Quick read

The physics of noise dictated three design levers.

Before drawing anything, we asked where the noise of a desk fan actually comes from: half of it is born in the interaction between air and blades, a quarter in the motor, the rest in the housing and structural vibrations. That map let us set aside the factors that are purchasing decisions — motor quality, bearings — and focus the research on three factors of pure geometry.

Each lever was then grounded in its sources: biomimetic serrations validated by Nature Communications, a bell-mouth housing documented by ASHRAE, nose cone and diffuser drawn from the fan-engineering literature (NASA, ISO 5801). The outcome: three selected solutions and a combined potential of up to −16.5 dB — silence is earned through geometry, not budget.

50% of the noise is born in the air-blade interaction
−5.5 dB for the biomimetic 3D-SC propeller against the industry benchmark
−16.5 dB combined potential of the three selected levers
+20% propulsive efficiency gained in the same move

Learn more — click a sub-step

01Noise sources

You cannot silence what you cannot name. The deliverable first breaks desk-fan noise into four families: the air-blade interaction dominates — about 50%, a broadband rush caused by blade-tip vortices and air sliding over the surfaces — ahead of the motor (25%: bearings and current chopping), the air-housing interaction (20%: the periodic tone of blades "striking" the air in front of the struts) and vibrations transmitted to the structure (5%).

That hierarchy has a direct consequence: most of the gain lies in aerodynamics, not electronics. It also sets the reading grid for everything that follows — every decibel gained must be justified by the physics of its source.

The source map: air/blades 50%, motor 25%, air/housing 20%, mechanical vibrations 5%.

02Factors and selection

Twelve influencing factors were then mapped, from the obvious (rotation speed, size) to the subtle (surface roughness, current-chopping frequency). Some are economic choices rather than research topics: motor quality or bearings are bought, not designed.

Three factors of pure geometry were selected for research: blade geometry, housing geometry, and the front/rear hub profile. Three grounds where engineering creates value — and which would go on to structure the CAD, simulation, and printing stages.

Twelve factors sorted by noise source; the three boxed ones — pure geometry — were selected for research.
What each selected factor can deliver: energy efficiency, vortex suppression, flow smoothing, fewer dead zones.

03Biomimetic blades

The chosen blade geometry comes from cutting-edge research: the "3D-SC" propeller published in Nature Communications combines the outline of a cicada wing with 3D sinusoidal serrations inspired by owl feathers, on a NACA 8412 airfoil pitched at 15°. Bench-tested against the industry benchmark, it gains 5.5 dB while delivering 20% more propulsive efficiency.

The mechanism is as elegant as the result: the wavy topography creates coherent vortex structures that damp both tonal and broadband noise. The authors swept sixteen amplitude × wavelength combinations; we will tune ours to the Ventilencieux's operating speed.

Five printed propellers on the bench: the 3D-SC against four benchmarks, including the DJI industry propeller (Wei et al., 2024, CC BY 4.0).
Sixteen amplitude × wavelength combinations swept to map the noise/thrust trade-off (Wei et al., 2024, CC BY 4.0).
The measurements: −5.5 dB at the best angle against the industry benchmark, across the whole thrust range (Wei et al., 2024, CC BY 4.0).

04Bell-mouth housing

On the housing side, two parameters concentrate most of the gain: a flared "bell-mouth" inlet, which smooths the flow before the blades and is worth −2 to −4 dB for a curvature radius of about 0.12 × D (ASHRAE 2020), and the blade-tip clearance, kept between 0.5 and 1% of the diameter — beyond 2%, the penalty reaches +3 to +5 dB (NACA TN 2495).

Architecture rules complete the picture: struts placed downstream rather than upstream (a 3 to 5 dB difference), in an odd count different from the blade count to avoid coinciding harmonics of the blade-passing frequency. The flared housing's material overcost — +8 to 12% — was accepted in the specification from the start.

The bell-mouth inlet is worth −2 to −4 dB, and tip clearance — the critical parameter — lives between 0.5 and 1% of the diameter (ASHRAE 2020, NACA TN 2495).

05Nose cone and diffuser

The hub gets the same aerodynamic treatment: a half-ellipse nose cone at the front removes the dead zones that whistle at intake (−1 to −3 dB) and cuts inlet pressure losses by 5 to 8%; at the rear, a diffuser opened at 5-7° recovers 15 to 20% of static pressure — that many fewer revolutions per minute for the same airflow.

Completed by a 3 to 5 mm acoustic liner on the rear face, this treatment of the central profile stacks up to −7 dB against a flat hub. Combined with the blades and the housing, the theoretical potential reaches −16.5 dB: the specification for silence is set, design can begin.

Hub profiles compared: the nose cone + 6° diffuser wins — up to −7 dB combined against a flat profile.