E-Bike Hub Motors Don’t Have to Be Loud—We Proved It
SHARE: ON
Adam
Journalist with over 7 years of experience covering the intersection of technology and transportation
The Problem We Are Trying To Solve
Walk through any city bike lane in Amsterdam, Berlin, or San Francisco, and you’ll hear it—that persistent, high-pitched whine from e-bike hub motors. It’s not loud enough to violate noise regulations, but it’s there. Always there.
At TO7 Motor in Suzhou, we started hearing the same feedback from OEM partners and riders: “Can you make it quieter?”
The challenge is this: Direct-drive hub motors are already among the quietest propulsion systems in urban mobility. We eliminated gears years ago, which removed the primary source of mechanical noise. Yet that residual acoustic signature—what acoustic engineers call the “electromagnetic harmonic signature”—persists.
So we decided to stop guessing and start measuring. What follows is an honest look at our ongoing research into hub motor acoustics, including what we’ve learned, where we’re headed, and the engineering challenges we’re still working to solve.
Why Motors “Sing”(And not in a good way)
Here’s what we discovered early in our simulations: A hub motor casing doesn’t just sit there passively.
It behaves like a cylindrical shell—essentially, a very precisely machined bell.
When electromagnetic forces from the rotating magnetic field interact with the stator (the stationary part of the motor), they create what’s called Maxwell stress—tiny pressure waves that ripple through the metal structure.
Even though the rotation itself is smooth, these forces contain “harmonics”—repeating patterns that occur at specific frequencies.
Think of it like plucking a guitar string.
You hear the fundamental note, but there are also overtones—higher frequencies that color the sound. In a hub motor, these electromagnetic harmonics can excite specific vibration patterns in the motor housing.
The “Bell Effect”: Structural Resonance
Finite Element Analysis of a hub motor casing showing the four-lobed vibration pattern that generates acoustic noise. The deformation is measured in micrometers and occurs at specific operational speeds
Our finite element analysis revealed something fascinating: At 300 RPM, the dominant noise source wasn’t the rotation itself, but a specific vibration mode we labeled “circumferential mode n=4.”
What does that mean in plain English?
The motor casing was deforming into a four-lobed shape—like a flower with four petals—vibrating too fast for the eye to see but perfectly audible to the ear. This particular mode shape is incredibly efficient at radiating sound into the air.
This explained why simply adding more damping material (foam, rubber mounts) wasn’t enough. We weren’t just dealing with vibration amplitude—we were dealing with modal radiation efficiency.
Some vibration patterns radiate sound far more effectively than others, even when the physical movement is tiny.
What the Simulations Told Us
Using COMSOL Multiphysics and boundary-element acoustic modeling, we ran hundreds of simulations to determine which design parameters actually matter for noise reduction.
We tested:
Shell thickness variations (+10%, +20%, +30%)
Material damping ratios (0.01 to 0.05)
Elastic modulus changes (stiffness modifications)
Different mounting configurations
Here’s what we found:
Sensitivity Analysis Results
Parameter
Modification
Noise Reduction
Shell thickness
+20%
-4.3 dB
Damping ratio
0.01 → 0.03
-6.8 dB
Elastic modulus
+20%
-3.1 dB
The damping ratio had the greatest impact, reducing it by more than 6 dB simply by optimizing the material’s ability to absorb vibrational energy.
But here’s the catch: These are simulation results under ideal conditions. The predicted baseline in a perfectly isolated, anechoic environment was 38 dB at 300 RPM. In real-world applications—with mounting systems, frame coupling, and manufacturing tolerances—we expect to add 8-15 dB to that figure.
That means realistic operational levels would be in the 46-53 dB range, which is still competitive with the best direct-drive motors on the market but not the dramatic breakthrough the raw simulation suggests.
The Harmonic Fingerprint
One of the most useful findings came from our harmonic force decomposition analysis. Not all electromagnetic harmonics contribute equally to noise.
We found that the 6th and 12th temporal harmonics were the dominant contributors, aligning with structural resonance frequencies in the 600-1200 Hz range.
Meanwhile, lower harmonics (2nd and 4th) produced negligible acoustic radiation because their corresponding mode shapes had low radiation efficiency.
This gave us a clear design target: If we can shift the structural resonances away from those critical harmonics—through material selection, thickness optimization, or pole-slot combinations—we can achieve meaningful noise reduction without sacrificing torque or efficiency.
Where We Are Now: The Experimental Phase
Schematic of the planned ISO 3744 acoustic test setup for the hub motor, showing microphone positions around the unit under test.
Test Domain
Measurement Method
Target Accuracy
Structural Vibration
Tri-axial accelerometers + impact-hammer testing for frequency response functions (FRFs)
±3 dB vs. simulated natural frequencies
Acoustic Radiation
10-point free-field microphone array in semi-anechoic chamber (ISO 3744)
±1 dB accuracy with background subtraction
Signal Processing
Short-time Fourier transform (STFT) for order-tracked spectral analysis
Simulations are valuable, but they’re not the full story. We’re currently preparing an experimental validation framework based on ISO 3744 acoustic measurement standards to verify our predictions.
Here’s what the testing protocol looks like:
Test Motor:
TO7 Motor ZM05-175DL —our 3000W-rated, 4000W peak direct-drive hub motor with 120 Nm torque and 82% efficiency.
Structural Measurements:
Tri-axial accelerometers mounted at multiple locations on the motor housing. We’ll use impact-hammer testing to measure frequency response functions (FRFs) and compare the measured natural frequencies with our simulated values.
Acoustic Measurements:
Free-field microphone array in a semi-anechoic chamber, measuring sound pressure levels at ten points around the motor following ISO 3744 protocols. Background noise will be subtracted to ensure ±1 dB accuracy.
Signal Processing:
Short-time Fourier transform (STFT) to extract order-tracked spectra—basically, separating motor-generated noise from environmental noise.
We’re targeting ±3 dB correlation between simulation and measurement across the electromagnetic, structural, and acoustic domains.
Even before completing physical testing, the simulation work has already informed several design decisions in our next-generation motor development.
The insights from FEM analysis and harmonic decomposition aren’t just theoretical—they’re directly shaping how we’re approaching the ZM05-175DL’s successor.
Based on the sensitivity analysis and modal behavior we observed, we’ve identified four key acoustic design strategies that will be implemented in the next iteration.
Each targets a specific noise generation mechanism we identified in the simulations:
Design Strategy
What We’re Doing
Expected Impact
Harmonic alignment avoidance
Selecting pole-slot combinations that avoid exciting high-radiation modes
Reduces 6th and 12th harmonic contribution
Structural damping optimization
Evaluating constrained-layer damping and alternative housing materials
Potential 6.8 dB reduction (simulation)
Stiffness tuning
Optimizing housing geometry to shift resonances to lower-radiation frequencies
Moves critical modes away from 600-1200 Hz range
Mounting isolation
Designing compliant motor mounts to minimize frame coupling
Reduces structural noise transmission path
Why This Matters (Beyond The Specs)
Acoustic comfort isn’t just about meeting noise regulations. It’s about the riding experience.
In shared urban spaces, a quieter motor means:
Less noise pollution in residential areas
Better integration with public transit (folding e-bikes on trains)
Reduced rider fatigue on long commutes
Improved perception of e-bikes as “premium” rather than “utilitarian.”
For OEM partners, lower acoustic emissions mean:
Competitive differentiation in crowded markets
Fewer warranty claims related to “motor noise.”
Better integration into high-end bike frames
Compliance with emerging urban noise standards
The Honest Challenges We’re Facing
Not everything in the lab translates cleanly to production. Here are the real-world constraints we’re navigating:
Challenge
The Problem
Our Trade-off
🏭
Manufacturing Tolerances
Simulations assume perfect geometry; real machining introduces micro-variations
Accept ±0.5mm tolerance, validate with physical testing
💰
Cost vs. Performance
Increasing shell thickness by 20% means 20% more material cost
Target 10-15% cost increase for premium motor lines
Focus on high-efficiency damping materials, not thickness
🔩
Mounting Variability
Can’t control how OEMs mount motor to frames; poor mounting undoes optimization
Provide installation guidelines + optional isolation kit
What’s Next: Our 12-Month Development Roadmap
We’re currently in the prototype phase, with four critical milestones ahead. This isn’t a press release promising breakthrough technology—it’s a realistic timeline for validating and refining what we’ve learned so far.
Q2 2026: Physical Testing
ISO 3744 acoustic measurements in a semi-anechoic chamber, tri-axial accelerometer validation of structural modes, and order-tracked spectral analysis. We’re targeting ±3 dB correlation between simulation and measurement. If we hit that, the models are trustworthy. If we don’t, we iterate.
Q3 2026: Iterative Refinement
Adjust FEM/BEM models based on test discrepancies. Optimize damping treatments, refine pole-slot selection, and re-run sensitivity analysis with validated parameters. This is where simulation meets manufacturing reality.
Q4 2026: Production Integration
Work with our Suzhou manufacturing team to implement feasible acoustic optimizations. Not everything that works in simulation survives cost analysis or production tolerances. We’ll document what makes it through and what doesn’t.
Q1 2027: Field Validation
Partner with 3-5 OEM customers for real-world testing in production e-bikes. Different frame geometries, mounting systems, and riding conditions. This tells us if our lab results translate to actual rider experience.
Our commitment: The goal isn’t to claim “the quietest motor ever made.” The goal is continuous improvement—measurable, verifiable, and honest. We’ll publish results as they come in—including the failures.
Conclusion:
This article isn’t a press release announcing a finished product. It’s a progress report on work that’s still evolving.
The truth is, acoustic optimization in hub motors is a balancing act between competing priorities: weight, cost, performance, manufacturability, and, yes, noise. There’s no single “solution,” only better compromises.
What we can say with confidence: We’re measuring, we’re modeling, and we’re testing. Preliminary simulations indicate significant potential for 4-6 dB reductions through structural optimization. Whether that potential survives contact with manufacturing reality—that’s the experiment we’re running right now.
C. J. Hansen and S. Snyder, Active Control of Noise and Vibration, Taylor Francis, 2012. https://www.taylorfrancis.com/books/mono/10.1201/b15923/active-control-noise-vibration-scott-snyder-laura-brooks-danielle-moreau-xiaojun-qiu-colin-hansen
