Planar Magnetic Drivers Evaluation and Interpretation of Impulse Response, ETC, and Step Response
Planar Magnetic Drivers
Planar Magnetic Drivers Evaluation and Interpretation of Impulse Response, ETC, and Step Response
Planar Magnetic Drivers
Embedded Files
Planar Magnetic Drivers
■Planar Magnetic Drivers
At first glance, the planar magnetic drivers appears similar to the VCD (Voice Coil Diaphragm) method, as both employ a structure that drives the entire vibrating surface. However, their physical behavior is fundamentally different, resulting in significant differences in both measured characteristics and sound quality.
In this section, to clarify the origin of these differences, both systems are examined from the perspectives of structure and operating principles, and it is systematically explained how these differences manifest in time-domain responses such as Impulse, ETC, and STEP.
In planar magnetic drivers, a wide range of structural variations exists depending on the configuration of magnets, conductor patterns, diaphragm shape, and conductor layout. The primary design parameters are as follows:
1. Magnet configuration: magnets placed only on the rear side of the diaphragm, or on both front and rear sides
2. Conductor configuration: conductors arranged on one side or on both sides of the supporting structure
3. Diaphragm shape: circular or rectangular
4. Conductor pattern: single-direction (radial or circular) or folded (serpentine) structures
In contrast to the VCD method, where the coil pattern is directly formed on the diaphragm to integrate the vibrating element and driving element, planar magnetic drivers employ a structure in which thin aluminum ribbon conductors are arranged on the diaphragm (film) with a certain spacing.
As a result, the conductor coverage is limited, and the proportion of the diaphragm that contributes directly to driving force becomes relatively small compared to the total diaphragm mass. Consequently, the system tends toward a structure in which a comparatively heavy diaphragm is driven by relatively low driving force.
To address this, various combinations of magnet configuration, conductor arrangement, diaphragm shape, and conductor pattern are employed in the design. However, each of these approaches introduces new issues, and in practice, planar magnetic drivers are realized as a set of trade-offs among these competing factors.
The main issues are as follows:
1) Insufficient driving force density (BL/M)
Because the conductors on the diaphragm must be thin and spaced at regular intervals, it is difficult to obtain sufficient driving force relative to the total mass and radiating area of the diaphragm. Even with various design measures, the tendency for a low BL/M ratio is hard to avoid, and this appears as slower initial transients and reduced energy concentration.
Furthermore, simply increasing the conductor thickness to improve BL/M tends to induce undesirable vibrations in the lateral direction. Suppressing these vibrations requires increasing the thickness of the supporting structure, which in turn leads to new issues such as increased stiffness and added diaphragm mass.
2) Lack of separation between diaphragm and suspension structure
In planar magnetic drivers, the regions where no conductors are present also function as the movable support (i.e., suspension). As a result, the diaphragm and the supporting structure are not clearly separated.
･ The support regions between conductors require a certain level of stiffness to hold the conductors in place. Consequently, vibration propagation across the diaphragm cannot be sufficiently suppressed, leading to in-plane propagation and modal (break-up) vibration.
･ Meanwhile, rear radiation is emitted and reflected with time delay, causing phase and waveform distortion. Suppressing the leakage of this rear radiation to the front requires adding mass to the support structure; however, this increases the overall moving mass and consequently reduces BL/M.
3) Mismatch between driving force distribution and support distribution
When folded (serpentine) conductor patterns or double-sided conductor configurations are employed, the spatial distributions of conductors and support regions become non-uniform. As a result, both the driving force distribution and the mechanical support distribution become non-uniform, leading to localized variations in driving force and time delays. This contributes to increased temporal dispersion of the response and enhanced modal vibration.
As described above, in planar magnetic drivers, the elements of driving force, support mechanism, and vibration control are not clearly separated in structure. Instead, they interact with one another and operate as a distributed system. This structural characteristic is the fundamental cause of the temporally dispersed response observed in Impulse, ETC, and STEP measurements.
For details on how these issues are addressed by the VCD method, please refer to “Introduction to the VCD Speaker.”
■Headphones for Characteristic Measurement
Based on measurement results obtained using REW, the characteristics of the planar magnetic drivers are presented. First, the unit used in this evaluation is introduced.
The unit used for this measurement is the planar magnetic headphone OPPO PM-3.
■OPPO PM-3
The OPPO PM-3 is a closed-back planar magnetic headphone developed by OPPO.
As a planar magnetic design, it is notable for being a lightweight and high-efficiency model specifically intended for portable use. It features relatively high sensitivity and low impedance, making it easy to drive even with devices such as smartphones.
Structurally, it employs a diaphragm with aluminum conductor patterns formed on its surface, which are driven across the entire radiating area by Lorentz force within a magnetic field. In general, planar magnetic drivers are described as using a double-sided magnetic circuit, with magnets placed on both sides of the diaphragm to achieve uniform driving force.
However, actual disassembly observations reveal that the magnets are located only on the rear side (cylindrical magnets with outer diameters of φ37–inner φ26 and φ11–inner φ2, each with a thickness of 2 mm, arranged concentrically). Magnetic materials are placed on both sides of the diaphragm, including the magnet plate, to distribute the magnetic field.
In addition, the aluminum conductor patterns are arranged on both sides of the diaphragm in a single-direction layout. However, because the winding directions are opposite on each side, the regions without conductors (support regions) are not uniformly distributed, resulting in a structurally non-uniform configuration.
This structure achieves a certain level of efficiency and portability; however, non-uniformity in magnetic flux distribution and conductor arrangement tends to remain. As a result, the design prioritizes listening comfort over precise transient reproduction.
In the market, it has been recognized as a “portable planar magnetic headphone.”
■Impulse Response
Impulse represents the time-domain characteristic showing how a loudspeaker responds to an instantaneous input signal.
The ideal response exhibits a sharp single peak, followed by a rapid decay of subsequent vibrations.
If significant subsequent oscillations remain, this indicates the presence of unwanted reflections or resonances within the system. These appear as components delayed relative to the direct sound (“delayed sound”), and as their amount increases, they can lead to blurred image outlines and reduced clarity of the soundstage.
【Measurement Conditions 】
●Acoustic Measurement Software: REW (Room EQ Wizard)
●Analysis Items: Impulse Response
●Measurement Distance: 10 cm
●Bandwidth: 3 kHz – 96 kHz (Butterworth HPF, 2nd order ×2; no LPF applied)
●Sampling Frequency: 192 kHz
●Normalization: Peak Normalization
The characteristics of the impulse response waveform of the planar magnetic headphone OPPO PM-3 are described below. For comparison, the characteristics of the Mundorf AMT21CM2.1-C and the VCD-type VCD-DT63 are overlaid.
■Initial Region (0–50 µs) = High to Ultra-High Frequencies
The sharpness of the initial rise is an important indicator of initial transient reproduction.
In the PM-3, the rise itself is not significantly degraded; however, the temporal width of the main peak is relatively broad, and the subsequent re-rising components and residual vibrations are more pronounced. This indicates that the energy is not concentrated within a short time, but is instead distributed and released over time.
Possible contributing factors include inherent constraints in driving force density (BL/M) characteristic of planar magnetic drivers, non-uniform in-plane velocity distribution across the diaphragm, and the superposition of multiple vibration modes, all of which may reduce the concentration of the main response.
Acoustically, these characteristics may manifest as softer attack contours, reduced extension in the high-frequency range, and diminished reproduction of fine micro-transients.
■First Valley (Inversion) (50–100 µs) = Upper-Mid to High Frequencies
The arrival at the first valley is an important indicator of the post-peak energy inversion behavior and the coherence of the transient response.
In the PM-3, a slight delay is observed in both the timing of the first valley and the transition from the main peak to the first valley compared to other drive types; however, the difference is limited, and the overall coherence of the initial response is largely maintained.
However, beyond the first valley, the re-rising components and subsequent vibration components are relatively large, and a superposition of multiple vibration components becomes evident. As a result, the energy does not concentrate at a single point within one cycle, but instead tends to be dispersed over time.
This behavior suggests that, due to non-uniform velocity distribution across the diaphragm and the superposition of multiple vibration modes, the vibration does not appear as a single piston-like response but rather as a distributed response.
Acoustically, these characteristics may manifest as reduced post-attack coherence and a tendency for the sound image contours to become slightly blurred.
■Second Peak (100–150 µs) (Most Critical)
In the PM-3, the second peak appears most clearly in this region and is significantly larger compared to other drive types, indicating that energy becomes concentrated again after the main response. This suggests that the vibration does not occur as a single concentrated response, but rather as multiple response components with time delays.
Such behavior may be attributed to non-uniform velocity distribution across the diaphragm and the superposition of multiple vibration modes, which can lead to the emergence of re-rising components after the main response.
Acoustically, these characteristics may manifest as a bloated sound image and reduced separation.
It should be noted that the 100–200 µs region is a critical time window that strongly influences perceived sound quality, and high-performance tweeters tend to suppress the energy in this region to very low levels.
■150–300 µs
In the PM-3, relatively large vibration components continue to be observed, and the waveform shows weaker coherence compared to other drive types. The amplitude is also relatively large, and a superposition of multiple vibration components is evident.
This behavior suggests that, due to non-uniform velocity distribution across the diaphragm and the overlap of multiple vibration modes, the energy does not converge into a single temporal component but is instead dispersed over time.
Acoustically, these characteristics may manifest as a lack of core definition, increased muddiness, and reduced resolution.
■300～600µs
In the PM-3, vibration components persist for a relatively long duration, and the decay is observed to be more gradual compared to other drive types. Although the amplitude gradually decreases, periodic components continue to be observed, indicating that the energy is dispersed over time.
Such characteristics suggest that, due to in-plane vibration of the diaphragm, the superposition of multiple vibration modes, and the redistribution of energy within the internal structure, the energy does not readily converge into a single temporal component.
Acoustically, these behaviors may be perceived as extended decay, while also manifesting as smearing of the sound image and a lack of tightness.
■After 600 µs
In the PM-3, minute vibration components are observed to persist over an extended period. Although the amplitude is small, periodic components continue to be present, indicating that the energy does not fully converge but remains temporally dispersed.
Such characteristics suggest that, due to the superposition of multiple vibration modes and the redistribution of energy within the diaphragm, the energy decays in a stepwise manner.
Acoustically, these residual components may manifest as diffusion of the sound image and a reduction in clarity.
■Overall Evaluation (Impulse Characteristics)
In the PM-3, the initial rise itself is not significantly delayed; however, the temporal width of the main peak is relatively broad, and both the second peak and subsequent residual components appear comparatively large. As a result, the energy does not concentrate at a single point in time, but instead exhibits a response that is dispersed along the time axis.
In addition, the waveform is not characterized by a single peak but rather by a multi-peak structure, with re-rising energy and vibration components continuing to be observed after the main response. This suggests that, due to multiple vibration modes and non-uniform energy distribution across the diaphragm, the energy may be released in a temporally segmented manner.
The primary contributing factors to these characteristics include insufficient driving force density (BL/M), the lack of clear separation between the diaphragm and the support structure, and non-uniformity in both the driving force distribution and the support distribution.
Acoustically, these characteristics may manifest as a smooth and non-fatiguing sound with a sense of spaciousness, while also appearing as slight ambiguity in sound image contours and reduced reproduction of fine micro-transients.
■ETC（Energy Time Curve）
ETC is a metric derived from the impulse response that shows how the energy of the input signal decays over time.
The faster the unwanted energy converges, the less blurring occurs in the sound image, and the clearer the localization and spatial reproduction become.
Differences in sound quality and soundstage reproduction are determined by the amount of components that arrive later than the direct sound (“delayed sound”). In the ETC, the energy components that appear after the direct sound are observed as sounds arriving with delay (“delayed sound”).
【Measurement Conditions 】
●Acoustic Measurement Software: REW (Room EQ Wizard)
●Analysis Items: Impulse Response / ETC (Energy Time Curve)
●Measurement Distance: 10 cm
●Bandwidth: 3 kHz – 96 kHz (Butterworth HPF, 2nd order ×2; no LPF applied)
●Sampling Frequency: 192 kHz
●Normalization: Peak Normalization
Similarly, for the ETC (Energy Time Curve) of the OPPO PM-3, the characteristics of the Mundorf AMT21CM2.1-C and the VCD-type VCD-DT63 are overlaid, and their features are described below.
■Initial Decay (0–100 µs)
In the PM-3, the decay immediately after the initial rise is relatively gradual, and a temporal spread of energy is observed from the early stage. As a result, the initial energy does not concentrate at a single point in time, but instead exhibits a response that is slightly dispersed along the time axis.
■Approximately 100–300 µs (Most Critical Region)
In the PM-3, energy remains in the −15 to −25 dB range within this region and persists at a clearly higher level compared to other drive types. This corresponds to the magnitude of the second peak in the impulse response, indicating that energy remains concentrated in this time region.
■Approximately 300–800 µs (Midrange-Dominant Region)
In the PM-3, the response does not exhibit a monotonic decay; instead, multiple peaks continue to be observed, indicating that energy is released in a temporally segmented manner.
This behavior suggests that the diaphragm is not operating in complete synchrony; rather, the driven vibrations propagate across the diaphragm with time delays.
Possible contributing factors include non-uniform driving force distribution arising from the conductor layout, as well as residual and re-radiated local vibrations due to a structure in which the support regions also function as part of the vibrating element.
As a result, the response does not converge as a single vibration, but instead appears as a superposition of multiple temporally separated components.
Acoustically, these characteristics may manifest as a sense of spatial spread or diffusion, reduced separation, decreased resolution, and midrange muddiness.
Furthermore, these behaviors are better understood not simply as insufficient decay, but as a response characteristic arising from the time-segmented release of energy.
■Overall Evaluation
The PM-3 exhibits a typical distributed-system response in which energy continues to be released with delay over the entire time range. In particular, within the 50–400 µs region, it is characterized by slower decay and a higher level of residual energy compared to other drive types.
This behavior is not merely due to insufficient decay, but is attributed to the time-segmented release of energy. It is reflected in the impulse response as a delayed arrival of the first valley and an increased second peak. In other words, the response does not converge into a single peak but is divided into multiple temporal components.
This is caused by in-plane propagation due to the diaphragm not accelerating and stopping uniformly, non-uniformity in both driving force distribution and support structure, and the redistribution of energy through the excitation of multiple vibration modes.
As a result, in terms of sound quality, this appears as weaker post-attack coherence, blurred sound image contours, reduced separation, and midrange muddiness (a phenomenon in which temporally dispersed components are perceived as a midrange sound image).
■STEP Response
STEP Response represents how the output changes over time when an input signal rises instantaneously and is then maintained at a constant level.
In an ideal response, the waveform rises rapidly and then settles into a stable state without subsequent oscillation.
However, if unwanted reflections or delays exist within the system, oscillations and fluctuations appear in the output. These are observed as the effects of components that occur later than the direct sound (“delayed sound”).
【Measurement Conditions 】
●Acoustic Measurement Software: REW (Room EQ Wizard)
●Analysis Items: Impulse Response / Step Response
●Measurement Distance: 10 cm
●Bandwidth: 3 kHz – 96 kHz (Butterworth HPF, 2nd order ×2; no LPF applied)
●Sampling Frequency: 192 kHz
●Normalization: Peak Normalization
Furthermore, for the STEP response waveform of the OPPO PM-3, the characteristics of the Mundorf AMT21CM2.1-C and the VCD-type VCD-DT63 are overlaid, and their features are described below.
■Rise (around t ≈ 0)
Compared to other drive types, the PM-3 exhibits a slightly slower rise, and a tendency toward delay attributable to differences in initial acceleration and vibrational synchronization can be observed.
■Initial inversion to first response (up to ~100 µs)
In the PM-3, the recovery after the undershoot is somewhat slow, and the waveform exhibits a delayed behavior rather than a smooth return. Rather than naturally converging as a single response, the behavior suggests the involvement of multiple vibration components.
■Two-Stage Structure (100–300 µs) (Most Critical)
In the PM-3, a clear re-rise is observed in this region, and the response does not settle in a single event but exhibits a multi-stage structure. A delayed component appears after the main response, clearly indicating that the behavior does not conclude as a single vibration. This re-rise corresponds to the second peak in the impulse response and indicates that the energy is released in a temporally segmented manner.
■Mid Region (300–800 µs)
In the PM-3, although the amplitude gradually decreases in this region, fine oscillations continue to be observed, and the response does not exhibit a monotonic decay.
This indicates that the vibration does not dissipate in a single event but instead persists over time as multiple components.
■Late Region (after 1 ms)
In the PM-3, small residual vibrations persist for a relatively long duration, indicating a tendency toward a longer settling time. Although the amplitude is low, periodic components remain, confirming the presence of energy tailing.
■Consistency with Impulse Response / ETC
This STEP response corresponds clearly with the results of the Impulse and ETC measurements. In the PM-3, the presence of the second peak in the impulse response, along with the energy retention and re-rise observed in the ETC, directly appear as the re-rise and multi-stage structure in the STEP response.
These can be understood as the same phenomenon observed from different perspectives.
■Correlation with Sound Quality
The characteristics of the PM-3 are reflected in its sound quality, with strengths including a soft, smooth, and spacious presentation, while its weaknesses appear as reduced post-attack coherence, less well-defined sound image contours, and decreased resolution.
■Overall Evaluation
The PM-3 exhibits a slightly slower rise, does not settle in a single event, shows a multi-stage response, and tends to retain residual vibrations over an extended period. This behavior is not simply attributable to differences in damping characteristics, but rather reflects a response in which energy appears in a temporally segmented manner along the time axis. Under the present measurement conditions, features characteristic of a distributed-type response are clearly observed.
Such behavior is consistent with structural characteristics inherent to planar magnetic drivers—namely, non-uniform driving force distribution due to constraints in conductor layout, a structure in which the support mechanism is not clearly separated from the vibrating element, and the resulting in-plane propagation and excitation of multiple vibration modes—and can be regarded as one of the tendencies that may relatively often appear in this type of system.
Page updated
Google Sites
Report abuse