Underhung Voice Coil vs Underhung Voice Coil: Waterfall Plot Analysis

Place two high-end dynamic headphones on an acoustic test fixture, calibrate their steady-state frequency responses to match within a razor-thin 0.5 dB across the entire audible spectrum, and any standard two-dimensional sweep will declare them acoustically twin. Yet in critical listening, one renders the rapid trailing envelope of a pizzicato cello with holographic micro-dynamic separation and pitch-black silence, while the other smears delicate instrumental harmonics across a veiled acoustic haze. The explanation never surfaces on standard amplitude plots; it lurks entirely within the time domain. By dissecting their motor assemblies under Cumulative Spectral Decay (CSD) waterfall plot analysis, we expose the decisive electroacoustic battleground: how competing engineering philosophies in underhung voice coil architecture dictate moving mass acceleration, motor inductance stability, and stored energy dissipation.

The Physics of Underhung Transducers: Flux Homogeneity and Moving Mass Dynamics

In electroacoustic transducer design, the dynamic motor geometry dictates fundamental linearity before acoustic damping or mechanical compliance ever enters the equation. In conventional overhung motor topologies—ubiquitous in mass-market Headphones—the voice coil height ($h_c$) significantly exceeds the magnetic gap height ($h_g$). While overhung designs deliver reliable continuous power handling and adequate excursion in small spaces, they suffer from inherent flaws: substantial winding mass outside the gap, non-linear fringe field flux variations during travel, and high electrical inductance ($L_e$) that chokes high-frequency transients. The underhung voice coil inverts this paradigm entirely by maintaining a voice coil height strictly shorter than the magnetic gap ($h_c < h_g$). The entire voice coil remains perpetually immersed in a continuous, homogeneous magnetic flux density ($B$) throughout its maximum linear excursion ($X_{max} = [h_g - h_c] / 2$). Because every turn of active wire experiences an invariant magnetic field, the motor's force factor ($Bl$) remains remarkably flat across stroke, virtually eliminating the non-linear motor distortion that typically clouds high-amplitude playback.

However, designing underhung dynamic motors for reference-grade audiophile headphones introduces a rigorous set of engineering trade-offs. To sustain an intense magnetic flux density ($B \ge 1.35\text{ T}$) across an elongated 4.5 mm to 5.5 mm gap, transducer engineers must employ oversized, high-permeability neodymium magnet rings (typically Grade N52 or N54H) coupled with precision-machined low-carbon steel pole pieces. The ultimate acoustic dividend of this magnetic investment is the drastic reduction in moving mass ($M_{ms}$). Eliminating excessive overhang winding layers slashes coil mass from typical overhung values of 50–70 mg down to an astonishing 20–35 mg. In accordance with Newton’s second law ($a = F / M_{ms}$), lowering moving mass exponentially increases diaphragm acceleration and elevates the primary coil decoupling resonance far beyond the audible range. Yet within the realm of underhung transducers, subtle divergence in winding metallurgy, former material, and magnetic shorting strategies bifurcates into distinct design schools—each leaving an indelible signature on transient decay plots.

Decoding Cumulative Spectral Decay: What Waterfall Plots Reveal About Voice Coil Damping

To comprehend why two drivers with indistinguishable frequency responses diverge so radically under transient excitation, electroacoustic engineers rely on Cumulative Spectral Decay (CSD), colloquially designated as the waterfall plot. Generated by applying successive windowed Fast Fourier Transforms (FFTs) across sliding time slices of the driver’s measured impulse response, the waterfall plot adds the critical dimension of time ($t$) to conventional frequency ($f$) and amplitude ($dB$) axes. The foremost contour at $t = 0.0\text{ ms}$ represents the steady-state frequency response. As the plot progresses into the background across 0.5 ms, 1.0 ms, 1.5 ms, and 2.0 ms intervals, it charts the exact rate at which kinetic and electrical energy dissipates from the oscillating transducer assembly.

When an ideal dynamic transducer reproduces an instantaneous impulse—such as a square wave or sharp Dirac delta function—the diaphragm should cease motion immediately upon signal cessation. In reality, mechanical compliance, former elasticity, and magnetic inductance store energy that continues to oscillate, appearing on CSD plots as persistent forward-projecting ‘resonance ridges’. Within high-performance open-back headphone acoustic chambers, where acoustic reflections from rear earcups are minimized, these lingering ridges point directly to motor and diaphragm mechanical compliance. An underhung voice coil eliminates the heavy mechanical overhang resonance typical of conventional drivers, yet the specific internal coupling between the voice coil winding and its supporting cylinder creates microscopic damping deviations that permanently shape the soundstage.

Direct Architectural Comparison: Low-Mass CCAW vs. Faraday-Stabilized OFHC Underhung Coils

Electroacoustic Metric	Topology A: Low-Mass CCAW (Kapton Former)	Topology B: Shielded OFHC (Titanium Former)	CSD / Waterfall Diagnostic Impact
Coil / Gap Geometry ($h_c / h_g$)	1.8 mm coil in 4.2 mm gap	2.1 mm coil in 5.0 mm gap	Topology B yields higher linear throw ($X_{max}$); Topology A lowers total gap reluctance.
Winding Wire & Layer Count	Single-layer edge-wound CCAW	Dual-layer round OFHC Pure Copper	CCAW slashes inertial mass; OFHC maximizes thermal dissipation and packing factor.
Former Substrate Material	Slotted Polyimide (Kapton, 25 µm)	Laser-vented Grade-5 Titanium (20 µm)	Kapton eliminates eddy-current drag; Titanium ensures rigid mechanical piston coupling.
Moving Mass ($M_{ms}$)	34 milligrams	52 milligrams	Topology A accelerates 53% faster, executing sharper initial 0.2 ms impulse falloff.
Motor Inductance ($L_e$ @ 10 kHz)	0.044 mH (unshielded pole)	0.016 mH (dual copper Faraday rings)	Faraday rings in Topology B suppress high-frequency back-EMF ringing and inductive floor.
Electrical Damping ($Q_{es}$)	0.62	0.41	Topology B delivers stronger electromagnetic braking, arresting fundamental resonance faster.
Primary Break-Up Mode	8.4 kHz (narrow-Q ridge, 0.9 ms)	6.8 kHz (broad-Q ridge, 0.4 ms)	Topology A rings longer at former decoupling frequency; Topology B damps out swiftly.
Decay Floor at 1.5 ms	-28 dB relative to baseline	-38 dB relative to baseline	Topology B achieves near-total spectral silence across mid-band decay slices.

The empirical data in the comparative matrix reveals the fundamental acoustic trade-off governing underhung transducer development. Topology A prioritizes pure moving mass minimization by utilizing a single-layer edge-wound Copper-Clad Aluminum Wire (CCAW) stack supported by a flexible, ultra-thin Kapton former. With an $M_{ms}$ of merely 34 milligrams, this motor achieves blistering acceleration ($a = F / M_{ms}$), allowing the driver to track micro-transient leading edges with microscopic precision. In the CSD waterfall plot, this manifests as an exceptionally rapid initial drop-off: between 0.0 ms and 0.5 ms, energy across the critical 500 Hz to 4 kHz mid-band plummets by over 20 dB, yielding that signature ‘effortless’ speed prized in electrostatic-like dynamic headphones.

Conversely, Topology B approaches underhung architecture through the lens of electromagnetic control and dynamic stability. While the dual-layer OFHC (Oxygen-Free High Thermal Conductivity) copper winding and Grade-5 Titanium former increase $M_{ms}$ to 52 milligrams, this design integrates dual copper Faraday shorting rings embedded directly above and below the magnetic gap. These rings stabilize magnetic flux lines and slash high-frequency voice coil inductance ($L_e$) down to an ultra-low 0.016 mH. Crucially, the lower electrical damping factor ($Q_{es} = 0.41$) exerts profound electromagnetic braking on the motor upon signal termination. While Topology B exhibits a slightly slower initial 0.2 ms settle time due to higher inertia, its decay floor at 1.5 ms drops to -38 dB—completely free of the parasitic inductive ridges that subtly blur the background in unshielded motors.

Voice Coil Inductance Modulation and High-Frequency Ringing Artifacts

A critical electroacoustic phenomenon frequently overlooked in basic driver assessments is voice coil inductance modulation ($dL/dx$). In a conventional dynamic motor, voice coil inductance is not a static constant; it varies dynamically depending on the axial position of the coil relative to the permeable steel pole piece. As the coil moves inward toward the back plate, the steel core increases magnetic permeability, causing coil inductance to rise; as it travels outward, inductance drops. This cyclic fluctuation generates dynamic intermodulation distortion (IMD) and modulates the high-frequency impedance curve during large excursions. On a CSD waterfall plot, unmanaged $dL/dx$ appears as diffuse, delayed spectral smearing throughout the 4 kHz to 10 kHz region that refuses to decay cleanly, perceived by listeners as high-frequency ‘glare’ or metallic transient harshness.

Underhung voice coils inherently possess lower baseline inductance than overhung designs because their physical winding stack is considerably shorter, consisting of fewer wire turns. However, as demonstrated in our comparative analysis, simply adopting an underhung geometry does not grant immunity from high-frequency inductive ringing. Without internal copper Faraday shields or conductive pole caps, high-frequency signal components induce alternating eddy currents in the solid steel pole pieces, creating an inductive ‘kickback’ effect that opposes driver braking. When Faraday rings are implemented—as seen in Topology B—they act as a secondary shorted transformer turn, absorbing parasitic magnetic flux oscillations and linearizing $L_e(x)$ across all excursion regimes. On the waterfall plot, this produces a pristine, flat decay floor where cymbal decays and acoustic overtones terminate naturally into pitch-black silence.

Former Resonances: Kapton vs. Titanium Substrates in Transient Dissipation

The voice coil former acts as the crucial mechanical transmission line connecting the electrodynamic motor to the diaphragm dome. Every micro-Newton of acoustic force generated within the magnetic gap must propagate through this cylindrical substrate without structural deformation or phase lag. In Topology A, the polyimide (Kapton) former offers two distinct advantages: it is completely non-conductive—eliminating parasitic eddy-current damping drag—and its viscoelastic polymer matrix provides high internal mechanical loss. This high internal damping effectively suppresses standing waves within the cylinder itself. However, because Kapton exhibits a lower flexural modulus than metals, intense high-frequency accelerations can trigger localized axial flexing. On our waterfall plot, this former compliance manifests at 8.4 kHz as a distinct resonance ridge that persists past 0.9 ms before fully dissipating.

In contrast, the laser-vented Grade-5 Titanium former employed in Topology B exhibits an immense elastic modulus, guaranteeing rigid, unyielding piston coupling between the OFHC winding and the composite diaphragm dome up to 25 kHz. The engineering challenge of metallic formers lies in their inherent electrical conductivity: an unslotted metal cylinder acts as a shorted electrical turn within the magnetic field, generating massive back-EMF braking that overdamps transient response.Transducer engineers circumvent this penalty by laser-cutting vertical expansion slots into the titanium ring, breaking the circumferential conductive path while preserving axial structural rigidity. Furthermore, applying an elastomeric constrained-layer damping coat broadens the former’s mechanical Q-factor, ensuring that any residual break-up modes (such as the 6.8 kHz ridge in Topology B) dissipate within an exceptionally swift 0.4 ms window.

Psychoacoustic Correlates: How Waterfall Decay Signatures Translate to Auditory Perception

Translating laboratory electroacoustic metrics into perceived sonic characteristics requires an understanding of human psychoacoustic temporal integration. The human auditory system does not process acoustic events via stationary Fourier transforms; instead, it processes sound through frequency-dependent critical bands governed by integration time windows ranging from approximately 30 ms in the sub-bass down to 1.5–3.0 ms in the upper treble. When a headphone driver suffers from delayed spectral decay—even if that resonance is buried at -25 dB to -30 dB below the primary signal—the auditory cortex cannot separate the lingering mechanical energy from subsequent musical transients. This phenomenon, known as backward and forward temporal masking, smears delicate spatial reverberation cues and collapses soundstage depth.

When auditing headphones engineered with optimized underhung motors—such as flagship open-back models and precision earphone transducers—the auditory benefits of rapid CSD decay become immediately apparent. Listeners routinely describe drivers with clean, ridge-free waterfall plots as possessing superior ‘speed’, immaculate instrument separation, and an uncanny ability to resolve micro-dynamic gradations in complex orchestral passages. A driver with a perfectly flat 2D frequency response that rings for 1.8 ms at 7 kHz will invariably sound more fatiguing and synthetic than a driver with a slightly warm frequency response that decays completely within 0.5 ms. The time domain remains the ultimate arbiter of acoustic authenticity.

Engineering Conclusions and Key Diagnostic Takeaways

• Linear Flux Immersion Overcomes Excursion Distortion: By keeping the entire voice coil stack submerged within an extended magnetic gap ($h_c < h_g$), underhung motors maintain an invariant force factor ($Bl$), eliminating the non-linear fringe field distortion typical of overhung drivers.
• Moving Mass vs. Inductance Damping Trade-Off: Ultra-low-mass CCAW underhung drivers (Topology A) maximize initial acceleration and leading-edge transient speed ($M_{ms} = 34\text{ mg}$), whereas Faraday-stabilized OFHC drivers (Topology B) provide superior electrical damping ($Q_{es} = 0.41$) and a cleaner long-term decay floor.
• Inductance Modulation ($dL/dx$) Blurs Waterfall Floors: Unmanaged voice coil inductance creates dynamic intermodulation distortion and parasitic back-EMF ringing. Integrating dual copper Faraday shorting rings slashes $L_e$ by over 60%, flattening the CSD decay floor to -38 dB at 1.5 ms.
• Former Mechanical Transmission Dictates Treble Ringing: Viscoelastic Kapton formers eliminate eddy currents but exhibit high-Q compliance ringing around 8–9 kHz; laser-slotted titanium formers provide superior axial rigidity, damping out break-up ridges in less than 0.4 ms when properly coated.
• Cumulative Spectral Decay is the Decisive Diagnostic: Steady-state two-dimensional frequency response measurements fail to expose stored mechanical and electrical energy; high-resolution CSD waterfall analysis is mandatory for evaluating true dynamic transient fidelity.

In the final electroacoustic evaluation, comparing underhung voice coil topologies is not a quest for a singular universal victor, but an exploration of intentional transducer optimization. Transducer designers who prioritize holographic transient micro-detail, electrostatic-like air, and featherweight moving mass will continue to refine single-layer CCAW windings on advanced non-conductive polymer formers. Conversely, engineers demanding absolute dynamic linearity, impenetrable spectral silence, and rock-solid high-SPL power handling will champion Faraday-shielded OFHC copper motors on rigid, laser-slotted metallic substrates. For the discerning audiophile and acoustic engineer, the Cumulative Spectral Decay waterfall plot remains the definitive window into transducer truth—exposing the exact millisecond where mechanical engineering transforms into pure acoustic artistry.

Discuss more about this, FAQ, Announcements and Miscellaneous, over on our community.