Juhazi that info is great to look at. So it seems it only stays linear down to around 20, and with 1000W and 2 drivers I could see it having a massive peaky response, and they just crank it back in eq to remain flat at a lower volume. This seems possible with standard drivers in a small box as long as they have the power handling and the dsp eq can pull back significant db to get rid of spikes.
As the Newport Test Labs KEF KC92 frequency response measurement Graph 3 explanation states:
"The truncated low-frequency extension is due to the anechoic measurement".
An anechoic environment is (should be) equivalent to "free space", no boundaries.
The KEF KC92 DSP EQ is configured for for in room use, six boundaries.
Ideal 1/8th space boundary gain is +18dB at very low frequencies in an enclosed room.
If +18dB were added to the Corner(green trace) response, the measurement would be very near the KC92's specified response of -3dB at 11Hz, certainly within expected measurement deviations due to room dimensions.
This article goes in depth as to how room boundaries affect low frequency response:
https://www.prosoundtraining.com/2011/08/29/how-boundaries-affect-loudspeakers/
The above graph using the Bose MB4 (4x5.25", specified at +/-3dB 45-310Hz) demonstrates how the boundary gain primarily increases low frequency.
Art
I would love to see a full TS parameters list for a singular KC92 driver >
I'm pretty sure it would reflect some interesting design features 🙂
I'm pretty sure it would reflect some interesting design features 🙂
As an adjunct to the above figures, consider a Dayton Audio UM18-22 18" Ultimax DVC 2 Ohms/Coil driver in a 60-litre enclosure. If we apply a Linkwitz Transform filter to this closed-box enclosure, we can achieve the following response. The −3dB point is 11.1Hz, and the SPL in the passband is 95dB. At 10Hz, the required input power is around 800W, and the driver excursion is 11mm. An 18-inch driver has 2.0x the radiating area of two 9-inch drivers. This closed-box alignment is only −0.1dB at 20Hz! At that frequency, it requires 80W of input power to get 95dB SPL. With a 2kW amplifier, this subwoofer would potentially offer a great deal of low-frequency extension at moderate listening levels in a small-to-medium-sized room.
Below is an alternative filter assisted closed-box enclosure designed around the Dayton Audio UM18-22 18" Ultimax DVC 2 ohms/coil driver in a 60-litre enclosure. If we apply a less aggressive Linkwitz Transform filter to this closed-box enclosure than in the previous design, we can achieve the following response. The −3dB point is 15.0Hz, and the SPL in the passband is 100dB. At 15Hz, the required input power is around 525W, and the driver cone excursion is 9.7mm. This closed-box alignment is only −0.4dB at 20Hz, due to the application of +1.0dB of parametric EQ at 18.0Hz with Q=2.0. At 20Hz, this subwoofer requires 243W of input power to get 99.6dB SPL.
If the quite-high power input at infrasonic frequencies is considered a worry, then a 2nd-order Butterworth high-pass filter can be added around 8Hz or so. This is almost an octave below the 15Hz cut-off frequency of the subwoofer. As a result, it shouldn't affect the cut-off frequency very much.
If the quite-high power input at infrasonic frequencies is considered a worry, then a 2nd-order Butterworth high-pass filter can be added around 8Hz or so. This is almost an octave below the 15Hz cut-off frequency of the subwoofer. As a result, it shouldn't affect the cut-off frequency very much.
witwald, how about just a peak filter with high Q? I use them with my sealed dsp-speakers based on in-room measurements, woofers having free-air impedance peak 25-35Hz.
Drivers and classD amps seem to be ok, but I don't play absurdly loud.
Drivers and classD amps seem to be ok, but I don't play absurdly loud.
As you suggested, a 60-litre closed-box enclosure using the Dayton Audio UM18-22 18" Ultimax can be equalized using PEQs to produce a −3dB point of 15Hz. To obtain that cut-off frequency, I needed to use two PEQs: 1) +15.8dB at 16.0Hz with Q=2.0, and 2) −4.8dB at 44.5Hz with Q=1.1. The resulting response is shown below. It is evident that there is a small amount of ripple in the passband, but nothing really to worry about. If another PEQ is available, then the passband response could have been smoothed out a little bit more.
As can be seen, for the same 100dB output SPL, this design has lower input power requirements than the earlier one that used the Linkwitz Transform, although not by much. The main difference is the greatly reduced excursion and power input at frequencies below 16Hz. The trade-off is the initial faster roll-off rate below the −3dB point, which is about 22dB/octave. This results in about 6dB less output capability at 10Hz.
As can be seen, for the same 100dB output SPL, this design has lower input power requirements than the earlier one that used the Linkwitz Transform, although not by much. The main difference is the greatly reduced excursion and power input at frequencies below 16Hz. The trade-off is the initial faster roll-off rate below the −3dB point, which is about 22dB/octave. This results in about 6dB less output capability at 10Hz.