🔬 applied physics

Non-intrusive MEMS microphone sensing of acoustic field state in resonant acoustic levitators

This paper demonstrates that transducer-mounted MEMS microphones can serve as effective, non-intrusive external sensors for accurately assessing the acoustic field state in resonant levitators, enabling precise identification of optimal operating conditions and providing a basis for compact closed-loop feedback control.

Original authors: Jan H. Dörsam, Maximilian L. Amberg, Sven Suppelt, Sören Soennecken, Chuanchao Xu, Alexander A. Altmann, Tomislav Maric, Dieter Bothe, Mario Kupnik

Published 2026-07-15

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Jan H. Dörsam, Maximilian L. Amberg, Sven Suppelt, Sören Soennecken, Chuanchao Xu, Alexander A. Altmann, Tomislav Maric, Dieter Bothe, Mario Kupnik

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a magical, invisible hand made of sound that can float a tiny object in mid-air without touching it. This is acoustic levitation. To make this invisible hand strong enough to hold the object, you need to tune the distance between a sound speaker (the transducer) and a mirror (the reflector) with extreme precision. If you get the distance wrong by even a tiny bit, the hand vanishes, and your floating object drops.

The problem is that the "perfect" distance changes constantly. It shifts if the room gets hotter, if the air pressure changes, or if you drop a new object into the floating zone. Traditionally, scientists have had to guess and check, or use bulky, slow scales placed under the mirror to feel how hard the sound is pushing. But you can't put a scale under the mirror if you want to build a small, portable levitation device, and you certainly can't put a scale inside the floating zone without blocking the view or messing up the magic.

The Big Idea: Listening from the Sidelines
In this study, the researchers asked a simple question: Can we listen to the sound leaking out from the side to know if the invisible hand is strong, without sticking a sensor inside the magic zone?

They attached tiny, super-small microphones (called MEMS microphones) directly onto the speaker itself. These microphones sit just outside the floating area, like a security guard standing in the hallway listening to the party inside. They don't block the door, and they don't get in the way.

What They Found (The Good News)
The team tested this setup by slowly moving the speaker closer to and further from the mirror, sweeping through different "resonance modes" (think of these as different musical notes the system can sing). They compared what the side microphones heard against two things:

A super-precise scale under the mirror (the "gold standard" for how strong the force is).
The electrical current going into the speaker (a common way to guess what's happening).

Here is the exciting part: The microphones were surprisingly good at finding the sweet spot.

When the researchers looked for the point where the floating force was strongest, the microphone signals peaked almost at the exact same moment.

The microphone's "peak" was found within 30 µm (that's 0.03 millimeters, or less than the width of a human hair) of the actual strongest force.
At the exact spot where the microphones said "Here it is!", the actual force was still at least 98.3% of its maximum possible strength.

This means the microphones could tell the system, "Stop moving! You're right here!" with incredible accuracy. In fact, in this specific setup, the microphones were actually sharper at finding the peak than just watching the electrical current in the speaker wires.

The "Direction" Trick
Knowing where the peak is great, but what if you miss it? What if you need to know which way to move to get back to the peak?
The researchers tried changing the sound frequency (like tuning a radio slightly) to shift the "perfect" spot. They found that the phase (the timing of the sound wave) from the microphones changed in a predictable way.

If the sound wave shifted one way, the microphones told them to move the speaker one way.
If it shifted the other way, they knew to move the other way.
This wasn't a universal rule for all situations, but it worked as a "proof of principle" for a specific experiment, showing that the microphones could give a hint about which direction to nudge the system to get back on track.

The "Wiggly Object" Test
They also dropped a tiny piece of foam into the floating zone. As the object wiggled and danced, the sound field around it changed. The microphones picked up these rapid, tiny changes in the sound envelope (the shape of the sound wave) that the slow scale under the mirror couldn't see. This suggests the microphones could one day help detect if a floating object is unstable or wobbling too much.

What They Explicitly Ruled Out (The "Not-So-Good" News)
It's important to know what this method isn't yet:

It is not a perfect tilt sensor. The researchers tried using a ring of microphones around the speaker to detect if the speaker and mirror were tilted (slanted) relative to each other. While the microphones did react differently depending on the tilt, the signals were messy. They could not yet calculate a precise, unique angle of tilt just from the microphone data. The paper argues that you cannot rely on this ring setup alone to tell you exactly how crooked your system is.
It is not a calibrated pressure gauge. Because the sound inside the levitator is so incredibly loud (over 145 dB), the microphones are technically being pushed beyond their normal limits. They aren't measuring the exact pressure in Pascals; they are measuring relative changes. You can't use them to say "The pressure is exactly 150 dB," but you can use them to say "The pressure is at its highest point right now."
It doesn't work for every electrical signal. While the microphone beat the "peak-to-peak current" measurement in this study, the authors are careful to say they haven't proven it beats every other electrical method (like impedance or admittance).

How Sure Are They?
The authors are very confident about the amplitude (loudness) of the microphone signal. They measured it across four different resonance modes and found the match with the force scale to be consistent and precise (within 30 µm).

They are more cautious about the phase (timing) and the tilt detection. They describe the phase result as a "proof of principle" for a specific case, meaning it works as a concept but needs more testing to become a standard rule. The tilt detection is described as "exploratory," meaning they found interesting patterns, but the system isn't ready to be a calibrated tilt sensor yet.

The Bottom Line
This paper shows that you can stick tiny microphones on the outside of a sound levitator to "listen" to the invisible hand. They can tell you when the hand is strongest and give you a hint about which way to move if you drift off. It's a clever, non-intrusive way to keep the magic floating without needing a giant scale or sticking a sensor right in the middle of the action. While it's not a perfect solution for every problem (like measuring exact tilt angles yet), it opens the door for building smaller, smarter, and more self-correcting levitation devices in the future.

Technical Summary: Non-intrusive MEMS microphone sensing of acoustic field state in resonant acoustic levitators

Problem Statement
Resonant acoustic levitators require precise knowledge of the acoustic field state to maintain stable trapping, as the optimal transducer–reflector distance shifts with wavelength, temperature, object insertion, and mechanical alignment. Existing sensing approaches face complementary limitations:

In-cavity sensors (e.g., microphone probes) perturb the acoustic field, obstruct sample access, and often cannot withstand the high sound pressure levels (SPLs) required for levitation.
Optical methods (e.g., schlieren, interferometry) are non-intrusive but require extensive instrumentation and are difficult to integrate into compact systems.
Electrical observables (transducer current, impedance) are readily available but describe the coupled transducer–electronics system rather than directly quantifying the acoustic field's levitation capability.
Reflector-side force sensing (precision balances) provides a direct cavity-level reference but is slow, stationary, and incompatible with compact or active transducer–transducer architectures lacking a passive reflector.

Consequently, there is a need for non-intrusive, cavity-external sensors that can provide relative observables of the acoustic field state without obstructing the workspace or requiring a reflector-side mount.

Methodology
The authors investigated the use of transducer-mounted Microelectromechanical System (MEMS) microphones as off-axis external acoustic sensors. The study utilized a single-transducer resonant levitator with a reflector mounted on a precision balance to serve as a ground-truth reference for acoustic radiation force ( $F_{ARF}$ ).

Sensor Configuration: Two microphone arrangements were tested:
- Linear Configuration: Three microphones aligned axially along the acoustic axis to evaluate distance-dependent responses and average spatial variations.
- Ring Configuration: Three microphones distributed at 120° intervals around the transducer to evaluate spatial sensitivity to non-axisymmetric conditions (e.g., transducer–reflector tilt).
Hardware: SPV01C8LR5H-1 (JAMILA) MEMS microphones were used. To manage high SPLs (145–165 dB) in the levitation cavity, the acoustic ports were intentionally reduced below manufacturer recommendations to mitigate saturation and harmonic distortion.
Experimental Protocol:
- Distance Sweeps: The transducer–reflector distance ( $H$ ) was swept across resonance modes $n=5$ to $n=8$ .
- Comparisons: Microphone amplitude ( $V_{pp}$ ) and phase ( $\phi_{mic}$ ) were compared against the balance-derived $F_{ARF}$ and the peak-to-peak transducer current ( $I_{pp}$ ).
- Dynamic Tests: Experiments included a frequency-shift test to evaluate phase-based correction direction, object insertion to test static shifts, and object oscillation to test dynamic envelope modulation.
- Tilt Tests: The transducer–reflector angle was varied to assess the ring configuration's ability to detect asymmetric field states.

Key Contributions and Results

Amplitude as a Resonance Indicator:
- The channel-mean microphone voltage ( $V_{pp}$ ) exhibited local maxima that coincided closely with the balance-derived force maxima ( $F_{ARF}$ ).
- Across four resonance modes, the microphone-derived peak positions were within two sampled distance increments (at most 30 µm) of the force maxima.
- At these microphone-derived peaks, the retained force was at least 98.3% of the maximum.
- In the tested setup, microphone amplitude localized the force maximum more sharply than the peak-to-peak transducer current magnitude, which remained elevated beyond the force peak.
Phase as a Directional Observable:
- In a frequency-shift experiment, the microphone phase relative to the transducer current provided a proof of principle for estimating the direction of correction.
- The phase gradient increased significantly near resonance. By comparing the wrapped phase displacement from a known reference point, the system could infer whether the transducer–reflector distance should be increased or decreased to restore the previous operating state.
- The authors note that absolute phase is not a universal resonance marker due to wavelength dependence but serves as a locally initialized directional cue.
Dynamic and Spatial Sensitivity:
- Object Oscillation: During the levitation of an oscillating object, the microphone voltage envelopes showed channel-dependent modulation (up to 11.5% peak-to-peak). This demonstrated that external microphones can capture fast, dynamic acoustic field changes that are averaged out by the slow balance readout.
- Tilt Detection: The ring configuration showed channel-dependent phase responses when the transducer–reflector tilt was varied. However, the response was not unique or calibrated; phase variations occurred even off-resonance, preventing a definitive tilt estimate with the current setup.
High-SPL Operation:
- The study confirmed that MEMS microphones can operate in high-SPL environments if acoustic ports are reduced and signal quality (harmonic content) is monitored. Saturation and distortion were treated as part of the sensing problem, with harmonic analysis used to identify nonlinear regimes.

Significance and Claims
The paper claims that external MEMS microphones serve as effective relative observables for assessing resonance-related acoustic field states. The primary contribution is the demonstration that:

External microphone amplitude can localize the acoustic radiation force maximum with high precision (within ~30 µm) without placing sensors inside the levitation cavity.
Microphone phase can provide complementary, locally initialized information regarding the direction of required corrections after a resonance shift.
These sensors offer a pathway for compact, transducer-side feedback in architectures where reflector-side sensing (balances) is impossible, such as in transducer–transducer or array-based systems.

The authors maintain a modest scope, explicitly stating that the microphones are not calibrated pressure or force probes and do not provide a unique reconstruction of the complete acoustic field. The results are presented as a basis for future compact feedback systems, with the sensing principle potentially transferable to other array-based architectures pending further experimental validation.

More like this