Fabrication of Multi-channel Electronic Frequency Division Amplifier

Since the introduction of digital technology into the audio field, the sound quality of the sound source and input system has been greatly improved, and the preamplifier has become almost a simple thing with a sound source selection switch and a volume potentiometer. On the contrary, the output system has not changed much from the analog era, mainly because the principle of the speaker has not changed much. Since the audio frequency range is as wide as nine to ten octaves, it is very difficult to make the vibration system of the speaker vibrate completely linearly according to the electrical signal within such a wide frequency range, and it is almost impossible to require linear sound radiation characteristics.

One solution is to divide the audio frequency range into several segments and then use several speakers to play the sound in segments. This is a multi-speaker system, and the most common ones are two-unit and three-unit systems. However, dividing the frequency band requires a frequency division network, which is generally inserted between the power amplifier and the speaker. L and C filters are inserted between the power amplifier and the speaker. Since the speaker is not a pure resistance component, it is difficult to design the crossover and it is not easy to get good performance; and a high-quality crossover requires the use of high-quality inductors and capacitors, which are expensive.

In addition, due to the different efficiencies of various speakers (the tweeter is about 6 decibels higher than the woofer), in order to balance the sound pressure of the entire frequency band, it is necessary to insert an attenuator in the crossover to reduce the level of the high-efficiency speaker. The result is that the entire speaker system becomes a combination of several lowest-efficiency speakers.

In order to change this situation, a multi-channel amplifier method was created.

After the preamplifier, an active filter is used to divide the frequency band. Each frequency band has its own power amplifier and speaker. The level of each frequency band is adjusted by a potentiometer before each power amplifier. The advantages of this method are obvious. It eliminates the aforementioned LC network and can effectively utilize the efficiency of each speaker. At the same time, it also reduces the frequency requirements for the power amplifier, and the output power can also be smaller. This structure is shown in Figure 1, and its key circuit is the active filter.

There are low-pass, high-pass, band-pass and band-stop filters. Low-pass filters allow components from zero frequency to its cut-off frequency to pass through, while blocking components above the cut-off frequency; high-pass filters block components below its cut-off frequency, while allowing components above it to pass through: band-pass filters allow frequency components between its low cut-off frequency and high cut-off frequency to pass through, while blocking all frequency components outside this frequency range.

Active filters using operational amplifiers can eliminate inductance components and obtain voltage or current gain. According to the cutoff characteristics of the filter, it can be divided into Bessel type, Chebyshev type and Butterworth type. Its characteristic curve is shown in Figure 2, which is mainly manifested near the cutoff frequency. The Bessel type decreases slowly, the Chebyshev type decreases steeply, and the Butterworth type is between the two. The cutoff characteristic is usually expressed in decibels of attenuation per octave. The attenuation per octave of the second-order filter is 12 decibels, and the attenuation per octave of the third-order filter is 18 decibels. Figure 3 is a standard Butterworth second-order active filter. Figure 3a is a low-pass filter, and its calculation formula is as follows:

The actual value of R1 is 18kΩ, the actual value of R2 is 9.1kΩ, and the actual value of C is 2200pF and 270pF in parallel.

Figure 4 is a schematic diagram of a 12 dB three-channel electronic crossover for audio. Using a multi-channel pre-stage crossover can achieve better sound quality than dividing after the power amplifier. The frequency range of the three-channel crossover is low frequency ~ 500Hz:

The intermediate frequency is 500Hz~5kHz; the high frequency is 5kHz. Their combined frequency characteristics are shown in Figure 5.

The low-frequency filter and high-frequency filter are the design examples above: the intermediate frequency uses a bandpass filter, which is composed of a high-pass filter and a low-pass filter. The calculation of R and C is the same as the design example. Here, the low-pass filter is set after the high-pass filter to reduce residual noise. A buffer is set before the filter to facilitate matching with the sound source. The 1kΩ and 150pF at the input end are used to limit the bandwidth of the input signal: the output end of each filter uses a 1kΩ 10-turn wire-wound potentiometer for output level adjustment.

The output signals of the three-way filter are connected to the same three power amplifiers respectively, and the circuit is shown in Figure 6. First, the LF357 operational amplifier with FET input stage is used as current buffer, and the final power amplifier tube adopts MOSFET with good high-frequency characteristics. The bias circuit is composed of diodes and resistors. The semi-variable resistor VR2 is used to set the quiescent current. The quiescent current can be measured by measuring the voltage across the source resistor (0.47 Ω) when there is no signal. Then it is calculated using the formula I="U" / R. The final negative feedback is added from the source of the MOSFET to the inverting terminal of the operational amplifier. Since the power supply voltage of the operational amplifier used as the driver cannot be too high, the maximum output of the power amplifier is limited. If the power supply voltage of the operational amplifier is ± 15V, the maximum output voltage of the driver stage is 12V = 24V, and the speaker impedance RL = "8" Ω, then the maximum output power of the final stage is P = "Vcc" × (VCC / 8RL) = 24 × 24 / 64 = 9W. This power seems to be small, but in fact it is only the output power of one frequency band. Combined with the output power of the other two frequency bands, it is fully applicable.

In Figure 6, Rx, Cx, LY, and RY at the output of the power amplifier are designed to stabilize the circuit. Since the speaker is not a pure resistance component, its inductance component will increase when the frequency increases, which is equivalent to a lighter high-frequency load and higher high-frequency gain, which may cause circuit oscillation; adding Rx, which is equivalent to a high-frequency load, can avoid oscillation.

When a long cable is used to connect the power amplifier and the speaker, the high-frequency load will be increased due to the presence of cable capacitance, making the power amplifier unstable. Adding LY and RY can avoid this situation. LY and RY are made of 1mm diameter enameled copper wire tightly wound 10 turns on a 101Ω 5W carbon film resistor.

In order to protect the speakers, a 2A fuse should be connected in series at the output end of each power amplifier. In the high-frequency channel, a 2.5 μF polypropylene capacitor should be connected in series between the power amplifier and the speaker to protect the high-frequency speaker.

As long as the resistance and capacitance of each channel filter are accurate, no debugging is generally required. Adjustment of the power amplifier: when there is no signal input, adjust VR1 to make the output voltage 0V, and then adjust VR2 to make the voltage across the source resistor 0.47Ω to 0.1V (about 200mA).