CN115360987A - class-D amplifier with nonlinear compensation function - Google Patents

Abstract

The invention discloses a D-class amplifier with nonlinear compensation function, which is characterized in that a compensation circuit is arranged in a feedback path to simulate the nonlinear distortion characteristic of a final-stage MOS tube, and further, the compensation is output to the input end of a modulator through table lookup, so that the nonlinear characteristic is compensated, the distortion of the output waveform of the amplifier is avoided, and the available SNR dynamic range of the output waveform is reduced.

Description

A Class D Amplifier with Nonlinear Compensation Function

technical field

The invention relates to a class D amplifier with nonlinear compensation function, belonging to the field of class D amplifiers.

Background technique

Figure 1 shows a Class D amplifier including a SIGMA-DELTA modulator (hereinafter referred to as "modulator"). The final power stage uses PMOS and NMOS as the final amplifier. In some cases, the final MOS has a very ideal switching characteristics. Non-ideal characteristics include linear distortion and nonlinear distortion, wherein the nonlinear distortion characteristic distorts the output waveform, reducing the available SNR dynamic range of the output waveform.

Contents of the invention

The invention provides a class D amplifier with nonlinear compensation function, which solves the problems disclosed in the background technology.

In order to solve the problems of the technologies described above, the technical solution adopted in the present invention is:

A class D amplifier with a nonlinear compensation function, comprising a SIGMA-DELTA modulator, a final MOS transistor and a filter connected in series in sequence, and a compensation circuit is arranged in the feedback path of the SIGMA-DELTA modulator;

The input end of the compensation circuit is connected to the output end of the modulator, and the output end of the compensation circuit is connected to the input end of the modulator; a compensation table is pre-stored in the compensation circuit, and the nonlinearity established according to the input and output characteristics of the final MOS tube is pre-stored in the compensation table. Distortion compensation data; the compensation circuit looks up the compensation table according to the output of the modulator, simulates the nonlinear distortion of the final MOS transistor, and outputs the corresponding nonlinear distortion compensation data.

The compensation circuit includes a unit delay circuit string and a look-up circuit, the input end of the look-up circuit is connected to the output end of the modulator, the output end of the look-up circuit is connected to the input end of the modulator, and the first unit delay of the unit delay circuit string The input end of the time circuit is connected with the output end of the modulator, and the output end of each unit delay circuit is also connected with the input end of the table look-up circuit.

The number of stages of the unit delay circuit in the unit delay circuit string matches the memory characteristic and time delay of the final stage MOS transistor and the performance requirements of the class D amplifier.

The compensation table is pre-stored in the look-up circuit, and the look-up circuit looks up the compensation table according to the output of the modulator and the output of each unit delay circuit, simulates the nonlinear distortion of the final MOS transistor, and outputs the corresponding nonlinear distortion compensation data.

The compensation table pre-stores the nonlinear distortion compensation data established according to the nonlinear distortion in the input and output characteristics of the final MOS tube. The look-up circuit looks up the compensation table according to the output of the modulator and the output of each unit delay circuit, and simulates the final Non-linear distortion of the stage MOS tube, and output corresponding non-linear distortion compensation data.

The look-up table circuit is a memory without latch input and output functions.

The beneficial effects achieved by the present invention: the present invention sets a compensation circuit in the feedback path, simulates the nonlinear distortion characteristics of the final MOS tube, and outputs compensation to the input terminal of the modulator through a look-up table, so that the nonlinear characteristics are compensated, avoiding This reduces the available SNR dynamic range of the output waveform, which distorts the output waveform of the amplifier.

Description of drawings

Figure 1 is a structural block diagram of a traditional class D amplifier;

Fig. 2 is the structural block diagram of class D amplifier of the present invention;

Fig. 3 is the structural block diagram of compensation circuit;

Fig. 4 is the schematic diagram of class D amplifier of the present invention;

Figure 5(a) is a rising edge with a rising distortion of 0 (ideal rising edge);

Figure 5(b) shows the rising edge with a rising distortion of 0.125;

Figure 5(c) is the rising edge with a rising distortion of 0.25;

Figure 6(a) is the falling edge (ideal falling edge) with a falling distortion of 0;

Figure 6(b) is the falling edge with a falling distortion of 0.125;

Figure 6(c) is the falling edge with a falling distortion of 0.25;

Figure 7 is a simplified MOS tube output and output of the compensation circuit

Figure 8 is the simulated output spectrum of the amplifier system using an ideal MOS tube;

Figure 9 shows the simulated output spectrum without compensation for an amplifier using a non-ideal MOS transistor;

Figure 10 shows the simulated output spectrum after compensation for an amplifier using a non-ideal MOS tube.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings. The following examples are only used to illustrate the technical solution of the present invention more clearly, but not to limit the protection scope of the present invention.

As shown in Figure 2, a class-D amplifier with nonlinear compensation function includes SIGMA-DELTA modulators in series (the DAC in the feedback path of the SIGMA-DELTA modulator in the figure is an option, depending on whether the amplifier is an analog input or Digital input, if it is a digital input, there is no DAC, it is a direct feedback, if it is an analog input, the feedback path has a DAC), the final MOS tube and a filter, and a compensation circuit is set in the feedback path of the SIGMA-DELTA modulator .

The input end of the compensation circuit is connected to the output end of the modulator, and the output end of the compensation circuit is connected to the input end of the modulator; a compensation table is pre-stored in the compensation circuit, and the nonlinearity established according to the input and output characteristics of the final MOS tube is pre-stored in the compensation table. Distorted compensation data; the compensation circuit looks up the compensation table according to the output of the modulator, simulates the nonlinear distortion of the final MOS transistor, and outputs the corresponding nonlinear distortion compensation data.

As shown in Figure 3, the compensation circuit includes a unit delay circuit string and a look-up table circuit.

The number of stages of the unit delay circuit in the unit delay circuit string matches the memory characteristics and time delay of the final stage MOS transistor and the performance requirements of the class D amplifier, and the first unit delay circuit input terminal of the unit delay circuit string The output end of the modulator is connected, and the output end of each unit delay circuit is also connected with the input end of the look-up table circuit.

The look-up circuit is usually a memory without latch input and output functions. The input of the look-up circuit is connected to the output of the modulator, the output of the look-up circuit is connected to the input of the modulator, and the compensation is prestored in the look-up circuit. The table and the look-up circuit look up the compensation table according to the output of the modulator and the output of each unit delay circuit, simulate the nonlinear distortion of the final stage MOS tube, and output the corresponding nonlinear distortion compensation.

Since the compensation circuit in the feedback path only needs to compensate for nonlinear distortion, there is no need to store the complete input and output characteristics of the MOS transistor, only the nonlinear part of the input and output characteristics of the final MOS transistor needs to be stored. Nonlinear distortion compensation data established for nonlinear distortion in the tube input-output characteristics. The table look-up circuit looks up the compensation table according to the output of the modulator and the output of each unit delay circuit, simulates the nonlinear distortion of the final MOS tube, and outputs the corresponding nonlinear distortion compensation data.

It can be seen from Figure 3 that after the signal enters the compensation circuit, it is divided into two paths, one path directly drives the look-up table circuit, and the other path passes through the unit delay circuit accordingly, the output of the unit delay circuit drives the look-up table circuit, and the output of the look-up table circuit is used For compensating nonlinear MOS transistors with memory effects, the number of bits of output data of the compensation circuit (how many bits of precision are used in the design) depends on the target performance requirements of the class D amplifier (how many SNR performances are achieved).

FIG. 4 is a simplified schematic diagram of a class D amplifier of the present invention. In the figure, U is the signal input; H(z ^-1 ) is a low-pass filter, usually including a first-order or advanced integrator, which has a very high gain in the signal frequency band; E is the quantization noise; Y is the modulator The output of (SIGMA-DELTA modulator) is also the input of the MOS tube; F(.) is the response of the MOS tube, which is a non-ideal switching device, and W is its output; if the MOS tube has ideal switching characteristics, then F (.)=1 or a constant; G(z ⁻¹ ) is a low-pass filter for filtering out-of-band noise, improving the dynamic range of the amplifier, and increasing the output SNR. In order to compensate the damage of F(.) to system performance, a compensation circuit is added in the feedback circuit, which has the system characteristics of MOS tube. Although the characteristics of this compensation circuit do not need to be exactly the same as those of the MOS tube, here we use F'(.)≈F(.) to illustrate the working principle of the whole system.

In fact, in order to reduce the complexity of the system and increase the stability of the system, the characteristics of the compensation circuit in the feedback circuit should be simplified appropriately, only the part that compensates the nonlinear characteristics; especially, if the characteristics of the MOS tube have memory effects, the compensation circuit in the feedback circuit The characteristics of the circuit should shorten the length of its memory effect.

The compensation principle of the system is explained below by studying the signal transfer function STF and the noise transfer function NTF of the system. Definition STF is defined as W/U ("/" means division), which has nothing to do with E, while NTF is defined as W/E, which has nothing to do with U.

According to Figure 4, we can get:

[UF'(Y)]H(z ^-1 )+E=Y

W=F(Y)

To get STF, ignore E in [UF'(Y)]H(z ^-1 )+E=Y:

UH(z ^-1 )=F'(Y)H(z ^-1 )+Y

Since the SIGMA-DELTA modulator works at a sampling frequency much higher than the signal bandwidth, and H(z ^-1 ) is within the signal bandwidth range, it usually has a very high gain at the low frequency end, that is, in the signal frequency band UH(z- ¹ )>>Y; thus, UH(z ^-1 )≈F'(Y)H(z ^-1 ), that is, Y=F' ^-1 (U);

Since W=F(Y) and F'(.)≈F(.),

Similarly, neglecting U in [UF′(Y)]H(z ^-1 )+E=Y, we get E=Y+F′(Y)H(z ^-1 )≈F(Y)H(z ^-1 )=WH(z ^-1 ); W/E=1/H(z- ¹ )=>>NTF=1/H(z- ¹ ).

可见，在信号频段内，信号无损地通过了系统。由W/E＝1/H(z^-1)＝》NTF＝1/H(z-¹)可见，量化噪声在信号频段，通常是在低频端内呈现剧烈衰减。我们可以看出F(.)没有出现在系统的STF和NTF中，MOS管的非线性特性得到了补偿。by the equation

It can be seen that within the signal frequency band, the signal passes through the system without loss. From W/E=1/H(z ^-1 )=>>NTF=1/H(z- ¹ ), it can be seen that the quantization noise shows severe attenuation in the signal frequency band, usually in the low frequency end. We can see that F(.) does not appear in the STF and NTF of the system, and the nonlinear characteristics of the MOS tube have been compensated.

After the ideal MOS tube output is stable, the ideal low level or ideal high level (high level normalized to 1, low level normalized to -1) is symmetrical; when the input drive voltage changes , its output changes immediately, even if it cannot change immediately, its rising or falling rate is the same. The actual MOS tube has various non-ideal characteristics, such as the existence of a dead zone, different rising and falling rates, and so on. Moreover, even if the output voltage of the MOS tube with memory characteristics also rises, its rising rate is not fixed, but is related to the previous control input sequence of the MOS tube. The rate at which it falls is the same.

In order to determine the feedback value of the compensation channel according to the output characteristics of the MOS tube, we define the rising and falling distortion as the ratio of the voltage area of the waveform distortion part per unit time and the high and low voltage difference per unit time. As shown in Figure 5(a), it is an ideal rising waveform, and the area of the distorted part is 0. We define the degree of distortion in this case as 0; as shown in Figure 5(b), the area of the rising distorted part accounts for 0.125 of the entire waveform. The degree of distortion is 0.125; as shown in Figure 5(c), the degree of rising distortion is 0.25.

Similarly, as shown in Figure 6(a) is an ideal falling waveform, the area of the distortion part is 0, and the degree of distortion is 0; as shown in Figure 6(b), the area of the falling distortion part accounts for 0.125 of the entire waveform area, and its drop The distortion degree is 0.125; as shown in Figure 6(c), the descending distortion degree is 0.25.

The compensation for the above MOS tube can be as follows:

First obtain the input and output characteristics of the MOS tube, which can be obtained by a high-efficiency pseudo-random sequence test, or by a high-precision sine or cosine sequence test, or by other sequence tests; after obtaining the input and output characteristics , only compensate for the non-linear part of the characteristic. For example, only compensate for the unequal portion of the ascent and descent rates. At the same time, when simulating nonlinear distortion in the feedback path, the input and output delay should also be shortened to ensure the convergence of the SIGMA-DELTA modulator.

The output of the compensation circuit in the feedback path is a time-discrete output, which is inconsistent with the analog output characteristics of the MOS tube. For this reason, the present invention proposes a compensation method to keep the distortion degree consistent per unit time. That is, at the same time, if the compensation circuit outputs a positive waveform signal, its output amplitude is 1-rising distortion. If the compensation circuit outputs a negative waveform signal, its output amplitude is -(1-down distortion). Further, in order to improve the performance of the whole system and ensure the stability of SIGMA-DELTA, we can only compensate the difference between the rising distortion degree and the falling distortion degree. If the rising distortion is greater than the falling distortion, no compensation is required when outputting a negative waveform, while the output of a positive waveform is 1-(rising distortion-falling distortion). If the descending distortion is greater than the ascending distortion, no compensation is required when outputting a positive waveform, while the output of a negative waveform is -1-(rising distortion - descending distortion).

Figure 7 is a simplified MOS tube output and output of the compensation circuit. The left side of the figure is a simplified MOS tube output, the distortion degree of the rising edge is 0.2, the distortion degree of the falling edge is 0.3, and the difference between the distortion degree of the upper and lower edges is 0.1. The right side of the figure is the output waveform of the compensation circuit. Since the distortion degree of the falling edge is greater than the distortion degree of the rising edge, no compensation is required when the circuit outputs the rising edge, and the output is 1 (after normalization), while the compensation output of the falling edge is -0.9 (after normalization ). Of course, it is also possible to output 1.1 at the rising edge and -1 at the falling edge; or 1.05 at the rising edge and -0.95 at the falling edge. It is sufficient to ensure that the difference after compensation is 0.2 of the rising edge distortion - 0.3 of the falling edge distortion = -0.1. It depends on which compensation can better ensure the convergence of the SIGMA-DELTA modulator.

In order to verify the effectiveness of the present invention, we simulate the non-linear characteristic MOS tube, and to a class D amplifier modulated by a third-order SIGMA-DELTA, its operating frequency is 32 times of the signal frequency, and the input signal is a sine wave. If the MOS tube There is no nonlinear distortion, and it has ideal upper and lower edge characteristics, and its ideal SNR can reach 66.8dB, as shown in Figure 8. If the MOS tube has nonlinear distortion, and the amplifier does not compensate the nonlinear MOS tube, its SNR is 35.8dB, as shown in Figure 9. After compensation, its SNR is 63.8dB, see Figure 10. It can be seen that the above-mentioned compensation method has a remarkable effect.

In the present invention, a compensation circuit is set in the feedback circuit to simulate the nonlinear distortion of the final MOS tube, and then the compensation is output to the input terminal of the modulator through a look-up table, so that the nonlinear characteristics of the amplifier are compensated, and the distortion caused by the final MOS tube is avoided. The output waveform distortion caused by the non-linearity of the output waveform reduces the available SNR dynamic range of the output waveform. The output of the compensation circuit does not need to be exactly the same as the output of the MOS transistor, it only needs to ensure that the difference between the distortion degree of the upper and lower edge outputs is equal to the distortion degree difference of the MOS transistor analog circuit.

The principle of the invention can be used not only in the class D amplifier adopting SIGMA-DELTA modulation, but also in ADC and DAC circuits adopting SIGMA-DELTA modulation, and is used for compensating the non-linear distortion existing in the circuit. The principle of the present invention can not only be used for the non-linear distortion existing in the compensating circuit in the class D amplifier, ADC and DAC circuit of one-bit quantization adopting SIGMA-DELTA modulation, but also can be used for multi-bit quantization adopting SIGMA-DELTA modulation In the class D amplifier, ADC and DAC circuit, the non-linear distortion existing in the compensation circuit.

The above descriptions are only preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, some improvements and modifications can be made without departing from the technical principle of the present invention, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims (6)

1. A kind of D class amplifier with nonlinear compensation function, including SIGMA-DELTA modulator, last MOS tube and electric-wave filter connected in series sequentially, characterized by that, there are compensating circuits in the feedback path of the SIGMA-DELTA modulator;

the input end of the compensation circuit is connected with the output end of the modulator, and the output end of the compensation circuit is connected with the input end of the modulator; a compensation table is prestored in the compensation circuit, and nonlinear distortion compensation data established according to the input and output characteristics of the last-stage MOS tube is prestored in the compensation table; the compensation circuit searches the compensation table according to the output of the modulator, simulates the nonlinear distortion of the last-stage MOS tube and outputs corresponding nonlinear distortion compensation data.

2. The class-D amplifier of claim 1, wherein the compensation circuit comprises a string of unit delay circuits and a look-up table circuit, an input terminal of the look-up table circuit is connected to an output terminal of the modulator, an output terminal of the look-up table circuit is connected to an input terminal of the modulator, an input terminal of a first unit delay circuit of the string of unit delay circuits is connected to an output terminal of the modulator, and an output terminal of each unit delay circuit is further connected to an input terminal of the look-up table circuit.

3. The class-D amplifier with the nonlinear compensation function according to claim 2, wherein the number of the unit delay circuits in the unit delay circuit string is matched with the memory characteristic and the delay of the last stage MOS transistor and the performance requirement of the class-D amplifier.

4. The class-D amplifier of claim 2, wherein the compensation table is pre-stored in a look-up circuit, and the look-up circuit is configured to look up the compensation table according to the output of the modulator and the output of each unit delay circuit, so as to simulate the non-linear distortion of the last stage MOS transistor and output the corresponding non-linear distortion compensation data.

5. The class-D amplifier according to claim 4, wherein the compensation table is pre-stored with non-linear distortion compensation data based on non-linear distortion in the input output characteristics of the final stage MOS transistor, and the table look-up circuit is configured to look up the compensation table based on the output of the modulator and the output of each unit delay circuit to simulate the non-linear distortion of the final stage MOS transistor and output the corresponding non-linear distortion compensation data.

6. The class-D amplifier with nonlinearity compensation function of claim 2, 4 or 5, wherein the LUT circuit is a memory without a latch I/O function.

Images