In our reproduction system, priority is placed on extracting the inherent performance of the speaker as faithfully as possible. For this reason, instead of employing a conventional LC network (passive crossover), a channel divider (active crossover) configuration is adopted in order to minimize any adverse impact on sound quality.
In general, two primary concerns are often raised regarding passive LC networks:
(1) the DC resistance (DCR) of the inductor (L) reduces the effective damping factor, potentially weakening low-frequency control; and
(2) LC elements introduce frequency-dependent phase rotation.
The damping factor is defined, based on the circuit conditions as seen from the amplifier side, as the ratio of the amplifier’s output impedance to the load impedance. However, the actual damping force is determined by how the current generated by the back electromotive force—produced by the motion of the speaker—flows under the circuit conditions as seen from the speaker side, which is the source of the damping force. Therefore, by evaluating the system on the basis of the impedance as seen from the speaker side, it becomes possible to more accurately understand the actual behavior of the damping current.
In this page, the circuit conditions viewed from the speaker side are examined.
The influence of LC networks on damping and phase characteristics is then evaluated.
The figure shows an example circuit in which a 12 dB/Oct low-pass LC network with a crossover frequency of 400 Hz is connected to an 8 Ω woofer.
In this circuit, we determine the total impedance that influences the damping current generated by the back electromotive force produced by the motion of the woofer—that is, the total impedance of the damping circuit as seen from the woofer side.
In this measurement circuit, a signal is applied to the terminals indicated in red in the figure, and by measuring the resulting current, the total impedance of the damping circuit as seen from the woofer side can be determined.
This figure shows the result of reproducing the previously described measurement circuit in simulation and calculating and plotting the total impedance of the damping circuit as seen from the woofer side.
For verification, the measurement circuit was physically constructed using a 5.9 mH inductor, a 47 μF capacitor, and an 8 Ω resistor. The graph shows the total impedance obtained through actual measurement. The measured characteristics closely match the simulation results.
As demonstrated, the total impedance of an LC network can be calculated through simulation with results that closely approximate the behavior of the actual circuit.
To clarify the effect of inserting the LC network, only the increase in total impedance caused by the network insertion was extracted and plotted through simulation. In other words, the characteristic shown here is obtained by subtracting the woofer’s equivalent resistance of 8 Ω from the previously presented total impedance characteristic.
From the graph, it can be seen that the impedance increases by approximately 1 Ω around 100 Hz. At higher frequencies, the increase becomes even greater, rising significantly near the crossover frequency.
Thus, when an LC network is inserted, the impedance component added in series to the damping circuit increases. As a result, the damping current generated by the back electromotive force tends to decrease.
These increases in impedance are all attributable to the LC elements. Since inductors (L) and capacitors (C) inherently possess frequency-dependent reactance components, the impedance increase identified here affects not only the amplitude but also the phase of the damping current. As a result, the component that is in phase with velocity—responsible for effective damping—decreases, and the pure damping force tends to be reduced.
In other words, when an LC network is adopted, the complex impedance introduced by the LC circuit alters both the amplitude and phase of the damping current. Consequently, the damping mechanism becomes more complex compared to simple resistive damping.
Furthermore, below approximately 100 Hz, as discussed later, the impedance of the speaker unit itself rises as it approaches the resonance frequency of the driver mounted in the cabinet. As a result, when a 12 dB/Oct LC network is applied to the woofer, the effective impedance increases across a wide frequency range, creating conditions under which the damping current flows less readily.
For comparison, a simulation was performed for a 6 dB/Oct configuration consisting of a single 3.2 mH inductor.
From the graph, the impedance increase reaches approximately 1 Ω at around 200 Hz, and at the crossover frequency of 400 Hz, the increase is approximately 3.3 Ω. Compared to the 12 dB/Oct configuration, this indicates that the impedance increase near the crossover frequency is smaller.
However, because the 6 dB/Oct configuration has a gentler attenuation slope, its influence extends over a wider frequency range. For example, at approximately 600 Hz, where the output is about −6 dB, the impedance increase reaches approximately 6.5 Ω.
Traditionally, significant emphasis has been placed on the output impedance of the power amplifier and the resistance of the speaker cable, as they are considered to have a major influence on the damping current. For example, an amplifier with a damping factor of 100 (referenced to 8 Ω) has an output impedance of approximately 0.08 Ω. Likewise, a 10-meter round-trip length of speaker cable with a diameter of 2 mm has a DC resistance of about 0.26 Ω.
However, when compared with these values, the impedance increase introduced by an LC network can be orders of magnitude larger at certain frequencies. In other words, while fractions of an ohm in the amplifier or cable are often scrutinized, several ohms of impedance may be added in series immediately afterward by the LC circuit.
Moreover, an LC network does not merely add resistance. It acts as a frequency-dependent complex impedance, affecting not only the amplitude but also the phase of the damping current. As a result, the component that effectively contributes to damping is further reduced, and the overall damping force is weakened.
There is ongoing discussion regarding coil selection—whether to use an air-core or iron-core inductor, and how large the wire diameter should be. For example, a 5.9 mH inductor with a wire diameter of 1.4 mm may have a DC resistance of approximately 0.9 Ω in the case of an air-core design, and approximately 0.3 Ω in the case of a core-type design. While the higher DC resistance of an air-core inductor is often considered a concern, in frequency ranges above 100 Hz the impedance component introduced by the LC network itself—due to its frequency-dependent behavior—can be significantly larger.
In essence, the fundamental issue is not the magnitude of the coil’s DCR, but the increase in complex impedance that arises from inserting the LC network into the damping circuit. Focusing solely on amplifier and cable specifications is insufficient; unless the influence of the LC network placed immediately afterward is taken into account, the true damping conditions cannot be accurately evaluated.
Next, using the same method, we examine an 8 Ω midrange driver. As shown in the figure, we consider the case in which a 12 dB/Oct, 400 Hz crossover High-Pass LC network and a 12 dB/Oct, 4 kHz crossover Low-Pass LC network are connected. We then evaluate the component of the total impedance that influences the damping current of the midrange driver, specifically the portion of the impedance increase caused by the LC networks.
The red line in the graph represents the increase in total impedance introduced by the LC networks.
Within the reproduction band of 400 Hz to 4 kHz, the impedance increase caused by the LC networks exceeds the nominal 8 Ω impedance of the midrange driver over most of the band, except in the range from approximately 736 Hz to 2,191 Hz.
Finally, using the same method, we examine an 8 Ω tweeter. As shown in the figure, we consider the case in which a 12 dB/Oct, 4 kHz crossover High-Pass LC network is connected. We then evaluate the component of the total impedance that influences the damping current of the tweeter, specifically the portion of the impedance increase caused by the LC network.
The red line represents the increase in total impedance introduced by the LC network. In the frequency range from approximately 3,541 Hz to 7,794 Hz, this impedance increase exceeds the nominal 8 Ω impedance of the tweeter. This band is critical for tweeter reproduction, indicating that the additional impedance introduced by the LC network has a significant impact.
In the high-frequency range, it has long been pointed out that LC elements introduce frequency-dependent phase rotation, and the impact of this phase shift on sound quality has been a subject of concern.
As described earlier, when the total impedance of the damping circuit is evaluated based on the circuit conditions as seen from the speaker side, it becomes clear that the damping current is also affected by this total impedance, with both its amplitude and phase being altered. In other words, LC elements not only cause phase rotation in the signal path but also influence the phase of the damping current generated by the back electromotive force. As a result, the damping component that is in phase with velocity is modified, and the damping mechanism becomes more complex.
Such phase changes in the damping current may affect the rise and settling characteristics of sound. Furthermore, since the LC network alters the phase characteristics of the load, the phase of the load impedance as seen from the amplifier also changes. Therefore, it is necessary to examine the impact on amplifier stability, including phase margin considerations.
Here, the impedance increase of the damping circuit is presented across the entire frequency range of the speaker system.
So far, we have primarily discussed the influence of the LC network; however, it is not the only factor contributing to impedance increase. The impedance rise caused by the resonance of the woofer mounted in the cabinet must also be considered. In particular, below approximately 200 Hz, this effect becomes dominant. In this graph, the Fostex FW208HS is used as an example woofer, and the characteristics shown are based on the specifications provided in the unit’s manual for installation in a bass-reflex enclosure.
Furthermore, LC networks often incorporate an attenuator for level adjustment. In such cases, additional series resistance is introduced, resulting in a further increase in impedance.
As described above, when an LC network is employed in a speaker system, the impedance of the damping circuit increases over nearly the entire frequency range, and the phase characteristics become more complex. A reduction in damping current delays the settling of vibration, resulting in a loss of clarity and definition. In addition, phase shifts directly affect transient sharpness and the contour of the sound image, and may be perceived as blurring or softened outlines.
Furthermore, LC networks present additional concerns, including the influence of capacitor materials on sound quality and the increase in impedance due to the DC resistance of coils and attenuators. These factors act as constraints on the reproduction of sound and soundstage, and their influence becomes increasingly apparent as the performance of the employed speaker unit improves. Even if the unit is refined, the overall system may struggle to exceed a certain performance threshold.
Many of these issues arise from fundamental principles, and due to inherent structural limitations, substantial improvement is difficult to achieve. For these reasons, if further improvement in sound and soundstage reproduction is the objective, LC networks should not be used.