Some Problems of Class D Amplifier

shaorc

[Ask if you don't understand] Class D amplifiers are digital. The principle is to compare the input signal with a triangle wave or a sawtooth wave to generate a modulated wave, which is then amplified by the amplifier circuit, low-pass filtered, and finally connected to the speaker. There are several questions as follows [1] What is the input of an amplifier board connected to? Although I know it is a sound signal, I don't know how to implement the input. For example, in the bus stop announcement voice, the output is amplified and clear sound, so how is its input provided? [2] As shown in the figure below, a typical single-ended output Class D amplifier, it is clear here that it uses a push-pull topology, but it is often seen that Class D uses an integrated amplifier chip (such as the TPA series). At this time, does the amplifier chip already include the comparator and push-pull circuit? [3] Why is low-pass filtering used? High-frequency signals will damage the load (speaker)? [4] Are R1 and C2 in the figure used for impedance matching?

maychang

"What is the input of an amplifier board connected to?" The amplifier board receives the sound signal, which is an analog signal. It may be a differential input or a single-ended input, which depends on the board's manual.

maychang

"For example, the bus stop announcement outputs amplified clear sound, so how is its input provided?" The bus stop announcement is the voice data (digital quantity) stored in the chip, which is converted into analog signals through DA conversion and then output.

maychang

"But we often see that Class D uses integrated amplifier chips (such as the TPA series). Do the amplifier chips already include comparators and push-pull circuits?" The comparators and push-pull circuits are already built into the chip. Just read the instructions carefully.

maychang

"Why does a high-frequency signal damage the load (speaker) when using a low-pass filter?" That's not true. Who told you this?

maychang

“图中的R1和C2是阻抗匹配用吗？”

R1和C2用来使得负载阻抗变化比较小。扬声器通常是所谓“动圈式”，其构造是个线圈置于磁间隙中。有音频电流通过线圈时，线圈振动带动纸盆发声。因为是线圈，所以具有电感性，其阻抗随频率升高而升高。R1C2特性是阻抗随频率升高而降低，R1C2和扬声器并联作为负载，使得负载阻抗随频率变化比较小。此部分电路往往被称为“茹贝尔电路”。
其实，此电路中已经存在电容C1，故R1C2作用不显著，去掉也关系不大。

shaorc

maychang posted on 2019-3-15 16:42 "Why does the high-frequency signal damage the load (speaker) when low-pass filtering is used?" That's not true. Who told you this?

Because LC filtering is used in the figure, which means that all high frequencies are filtered out, so I asked whether low-pass filtering is used in all cases.

shaorc

maychang posted on 2019-3-15 16:49 "Are R1 and C2 in the figure used for impedance matching?" R1 and C2 are used to make the load impedance change relatively small. Speakers are usually so-called "dynamic coil" type, which...

This is the second time I have heard you talk about the Joubert circuit. This time I understand its function. It is to compensate for the increase in load impedance with the increase in frequency. There are three derivative questions [1] If you do not use a compensation circuit like Joubert, take the speaker as an example. Even if the load impedance changes with frequency, its input signal frequency must be within a certain range. Is it not allowed for the load impedance to fluctuate within this corresponding range? Or does it mean that the sound effect is poor? [2] Why not directly use a capacitor in parallel with the speaker? Wouldn't the compensation effect be better? 【3】As shown in the following figure, I don’t know if it is a corresponding schematic diagram. The coil is in the magnetic gap, the current signal flows through the coil, and the coil vibrates and vibrates the paper cone. Does the coil refer to the voice coil in the figure? Shouldn’t it be fixed? Why is it vibrating?

maychang

shaorc posted on 2019-3-15 17:26 Because LC filtering is used in the figure, all high frequencies are filtered out, so I asked whether low-pass filtering is used in all of them

"Because LC filtering is used in the figure, all high frequencies are filtered out" Low-pass filtering is used, but low-pass filtering will not damage the speakers.

maychang

shaorc posted on 2019-3-15 17:49 This is the second time I have heard you talk about the Joubert circuit. This time I understand its function. It is to compensate for the load increasing with the increase of frequency. There are three derivative problems [ ...

If you don't use a compensation circuit like Joubert, take the speaker as an example. Even if the load impedance changes with the frequency, but its input signal frequency is definitely within a certain range, is it not allowed for the load impedance to fluctuate within this corresponding range? Or does it mean that the sound effect is poor? Impedance fluctuations within a certain range are allowed. This kind of Joubert circuit is not very effective.

maychang

shaorc posted on 2019-3-15 17:49 This is the second time I've heard you talk about the Joubert circuit. This time I understand its function, which is to compensate for the increase in load as the frequency increases. There are three derivative problems […

"Why not directly use a capacitor in parallel with the speaker, wouldn't the compensation effect be better?" In the picture in the first post, C1 is directly connected in parallel with the speaker.

maychang

shaorc published on 2019-3-15 17:49 This is the second time I have heard you talk about the Joubert circuit. This time I understand its function. It is to compensate for the increase in load as the frequency increases. There are three derivative problems […

"As shown in the figure below, I don’t know if it is a corresponding schematic diagram. The coil is in the magnetic gap, the current signal flows through the coil, and the coil vibrates the paper cone. Does the coil refer to the voice coil in the figure? Shouldn’t it be fixed? Why does it vibrate?" It is the voice coil. The voice coil is glued together with the paper cone (diaphragm), and is kept in a position that is not in contact with the "soft iron core column" and the magnet by the "centering support sheet". The vibration of the voice coil (left and right in the figure) drives the paper cone to vibrate and make a sound.

xunke

R1 and C2 feel redundant.

maychang

xunke posted on 2019-3-16 21:52 R1 and C2 are redundant.

In the case of the power amplifier stage approximating the current source output, R1C2 is useful, but at the output of the class D amplifier, R1C2 is somewhat redundant.

shaorc

maychang posted on 2019-3-15 18:11 "As shown in the figure below, I don't know if it corresponds to the schematic diagram. The coil is in the magnetic gap, the current signal flows through the coil, the coil shakes and vibrates the paper cone, the coil refers to the figure...

So that's how it is. I would like to understand two more questions [1] Some audio amplifiers have very good effects (such as subwoofers), while others have obviously noisy sound quality. Is this because the (sinusoidal) signal output from the amplifier circuit (push-pull) is not well matched in the process of being transmitted to the speaker? The speaker load is often said to be 8 ohms. If the sound quality is good, is the entire output signal loaded on it? 【2】On the train, each carriage has a station announcer, which is equivalent to a series connection of many loudspeakers. In addition, the transmission line is longer, and the output energy of the power amplifier is greatly lost during transmission, and the load impedance also becomes complicated. If a Class D power amplifier is used, the output signal of the power amplifier circuit is passed through a step-up transformer to increase the output signal voltage. After passing through the transmission line, the signal voltage is reduced by a step-down transformer to drive all the speaker loads. I saw in the information that this is called a constant voltage or fixed resistance Class D power amplifier. Can the hardware structure of this type of power amplifier be viewed as the following figure?

maychang

shaorc posted on 2019-3-17 13:54 So that's the case. I would like to know more about two issues. [1] Some audio amplifiers have very good sound effects (such as subwoofers), while others have obviously very low sound quality. ...

"Some audio amplifiers have very good sound effects (such as subwoofers), while others have obviously very low sound quality. Is this because the impedance matching of the (sinusoidal) signal output from the amplifier circuit (push-pull) is done well when it is transmitted to the speaker?" This question is very complicated and can be filled with the content of almost half a book. Generally speaking, the nonlinear distortion and frequency distortion of modern audio amplifiers are good enough (early transformer-coupled audio amplifiers were relatively poor). "Obviously low sound quality" mainly occurs in speakers. In modern audio systems, "poor sound quality" usually has little to do with "impedance matching".

maychang

shaorc posted on 2019-3-17 13:54 So that's how it is. I would like to learn more about two issues [1] Some audio amplifiers have very good effects (such as subwoofers), while others have obviously very noisy sound quality, ...

"On the train, each carriage has a station announcer, which is equivalent to a series connection of many speakers. In addition, the transmission line is longer, and the output energy of the power amplifier is greatly lost in transmission, and the load impedance becomes complicated. If a class D power amplifier is used, the output signal of the power amplifier circuit is passed through a step-up transformer to increase the output signal voltage. After passing through the transmission line, the signal voltage is reduced by a step-down transformer to drive all speaker loads. I saw that this is called a constant voltage or fixed resistance class D power amplifier. Can the hardware structure of this type of power amplifier be seen as the following figure?" It can be represented by this figure. It is a very common method to transmit power audio signals over a slightly longer distance (hundreds of meters or even thousands of meters) by using a transformer to step up the voltage before transmission. This method is mainly to reduce power loss during transmission. This method has an impact on the sound quality (assuming the speakers are the same), and the main impact comes from the transformer.

chunyang

shaorc posted on 2019-3-17 13:54 So that's the case. I would like to learn more about two issues [1] Some audio amplifiers have very good effects (such as subwoofers), while others obviously have very low sound quality.

The quality of sound involves many aspects, and the final quality also conforms to the "bucket principle", that is, the part with the highest distortion will drag down the entire system. As for impedance matching, this is only one of the many circuit qualities and does not constitute a necessary and sufficient condition for high fidelity. A good amplifier circuit must be impedance matched, but simply achieving impedance matching can be a far cry from high fidelity. Therefore, there is no such thing as good sound quality when all output signals are added to the speaker. Good sound quality generally has higher signal restoration, but sometimes it requires special distortion because the human ear is nonlinear.

maychang

shaorc posted on 2019-3-17 13:54 So that's the case. I would like to know more about two issues. [1] Some audio amplifiers have very good sound effects (such as subwoofers), while others have obvious noise quality.

The frequency characteristics of transformers are quite poor, especially when the frequency range is quite wide. If the ratio of the highest and lowest frequencies is 10 times, then the transformer can transmit the signal quite satisfactorily. If it is 100 times, special winding methods and a relatively large core must be considered. It is almost impossible to require a flat frequency characteristic for 1000 times. Normally we say that the frequency audible to the human ear is 20Hz to 20kHz, which is 1000 times higher. It is simply impossible for a transformer to transmit such a wide frequency range and require the frequency characteristics to be flat and the nonlinear distortion to be small enough.

zhuyebb

Thank you for sharing