4 Way Multi Channel System

Whitesunds National Park, NM USA Photo by Yuichi

The four channel system is one of the most complicated system. However, it is worthwhile to challenge.
In this case, a excellent electronic frequency divider equipment is used to separate bandwidth to each frequency channel. The frequency divider equips not only digital filter functions but also digital delay, gain control, phase control and others. In the system, Accuphase DF-55 is used.

Specifications of DF-55 (Digital Frequency Dividing Network)
- Number of Channels : Four (Numbers are expanded by cascade connection of the other DF-55)
- DA Converter : 24bits
- Digital Filter : (see right column)
- Time Delay + 30m to -30m (by every 5mm step)
- Level Adjustment : By every 0.1dB step
- Phase : + or -
- Dynamic Range : 109 to 112dB
- Frequency Range : 2.0 to 44KHz
- Distortion (THD) : 0.001% (20 to 20KHz)
- Input : Digital / Analogue (Balanced / Unbalanced) / Copper / Optics / RJ45
- Output : Digital / Analogue (Balanced / Unbalanced)

(Please refer to Accuphase web site for details)

Digital Filter Specifications
Filter Slope :
6, 12, 18, 24, 48, 96dB/octave (some exception for 10Hz and 20KHz)

Cut-off Frequency :
10, 20, 31.5, 35.5, 40, 45, 50, 56, 63, 71, 80, 90, 100, 112, 125, 140, 160, 180, 200, 224, 250, 280, 290, 315, 355, 400, 500, 560, 630, 710, 800, 900, 1000, 1120, 1250, 1400, 1600, 1800, 2000, 2240, 2500, 2800, 3150, 3550, 4000, 5000, 5600, 6300, 7100, 8000, 9000, 10k, 11.2k, 12.5k 14k, 16k, 18k, 20k, 22.4kHz

The four channel sound system consists of following equipment.

- Bass system: JBL 2245 (46cm Diameter) with 270 liters Enclosure
- Mid-Bass system: EVM-10M (25cm Diameter) with Horn (A-300E). Its Frequency range is 250-700Hz.
- Mid system: JBL 2441 (5cm throat) with A-290 horn
- High system: TAD TD-2001 with A-1000 horn

- Accuphase C-200X Control Amplifier
- Accuphase DF-55 Digital Frequency divider
- Accuphase P-300L Power Amplifier for Bass system
- Accuphase P-300X Power Amplifier for Mid-Bass system
- Accuphase P-266 Power Amplifier for Mid system
- Accuphase P-11 Power Amplifier for High system

You can understand how old some of components are! except DF-55!! However, all of those equipment work perfectly as of today.

Design concept of this system is to draw capability of the Mid horn as possible as I can. Therefore, goal image is maximize bandwidth of the Mid speaker. Other speakers support the Mid speaker.
From physical layout stand point, it is not the ideal one. Here, I can introduce the adjustment method using the impulse response.

Impulse Response method
Frequency range of each channel or speaker is as follows. (Since we can change the cross over frequency quite easily by using DF-55, we can tune up after coarse adjustment is made. )

Cross over frequency is temporarily set as follows.
224Hz, 630Hz and 5000Hz
( Bass: <244Hz, Mid-Bass: 244Hz to 630Hz, Mid: 630 to 5000Hz,Hi: 5000< )

The relation of the diaphragm positions is shown as lower left figure. A mic. is set at 2m in front of the Bass speaker. Unfortunately, there is no perfect adjustment mechanism exists. So I tried several approaches. This is the impulse response method.

Above is a illustration how the input impulse is divided and mixed again in the air space after each source releases responsible band.

Since diaphragms are located in different distance from the listener position, impulses arrive in different timing at single mic. position in this case.

In this approach, adjustment concept is to align output impulse timing by each channel.

In this case, 24dB/oct traditional filter slope is used. Curve is shown in the left figure.
The left figure shows the relation of the Bass output and the Mid-Bass. We can see there are two problems to be resolved in this figure. Red (Mid-Bass) impulse behaves differently from Black (Bass). The Bass output is + polarity and the Mid-Bass is -, and timing of each rise time is different.
( Please note that both waveforms are not filtered in this case. )

So, by using the Phase switch on the DF-55 console, the phase is reversed. This shows in the lower left figure. The polarity is now matched.
(These are both filtered waveforms.)

Now, the timing can be adjusted by using the Delay function and the Value setting knob. The lower right figure is the adjusted the Bass output and the Mid-Bass output waveforms starting up at 14msec.

Same adjustments were made between the Mid-Bass and the Mid speakers and the Mid and the High speakers.
The upper left figure shows "Before the adjustments were made" and the upper right figure shows "After". Those are also adjusted to 14msec point. These mean that all of four impulse outputs match at 14msec line.

In lower frequency domain there exist large Group Delay of the speaker. On the other hand human, ear or brain, I should say, can not recognize the delay so sensitively. So, this subject can be discussed separately. However, the delay will cause phase difference. Phase will be a potential cause of interference between two waves which we have to be careful.
Anyway, frequency response by each channel is shown above. Since measurements were made in the room, many standing waves and reflections are observed, we see many peaks and dips, and these problems have to be discussed separately, too. The lower figures show the Wavelet analysis of total system. The Wavelet analyzes both time domain and frequency domain data on the same chart.
Wavelet Analysis
X-Axis represents passing time by m-seconds. And Y-Axis is for frequency scale. The red area shows high sound pressure zone by dB. Because of the wavelet nature, there exist some uncertainty among frequency and time. And lower frequency is displayed with expanded manner. Therefore, it looks like wide spread horizontally in lower frequency.
Anyway, in this pattern the Mid output timing and the High output timing arrives deterrently as indicated.
This chart shows "After" all time alignment process is made.
We can see that the red zone is all aligned as indicated at about 14msec line.
Finally, all level adjustments were made by using the Level function switch on the DF-55 console.

In general, it is convenient to utilize the RTA (Real Time Analyzer) with the pink noise. Left is the data acquired from adjusted 4 way system as mentioned above.

The adjustment was made relatively fine. But I like to add one thing. Strictly speaking, the impulse of each band does not represent its band delay. This is because of the group delay of the speaker. The group delay depends upon frequency. Ideally it must be independent from the frequency. But it is not. So, we have to consider about total delay including the delay caused by diaphragm locations and the group delay caused by each speaker properties at crossover frequency.

.... Return to System Adjustment Page
Whitesunds National Park NM, USA Photo by Yuichi