Digital active EMI suppression method based on noise harmonic prediction

Abstract

The invention belongs to the technical field of electromagnetic interference (EMI) filtering, and particularly relates to a digital active EMI suppression method based on noise harmonic prediction. The method particularly aims at a system applying a high-voltage, high-frequency and high-power DC / DC converter, and comprises the following steps: firstly, establishing a CM behavior model of the power converter, then determining a relational expression between CM noise and a noise source, predicting a reverse EMI noise signal for counteracting based on a digital active EMI filter circuit of noise harmonic prediction, and further injecting the reverse EMI noise signal into the converter to realize noise suppression. Through experimental test verification, a thought and a theoretical model are provided for noise suppression of the converter without a sampling circuit, and the method has important engineering value for development of DC / DC high-power and high-frequency application and optimization of EMC characteristics and EMI suppression of a converter system.

Description

Technical Field

[0001] The present invention belongs to the technical field of electromagnetic interference (EMI) filtering, and particularly relates to a digital active EMI suppression method based on noise harmonic prediction. Background Art

[0002] With the wide application of third-generation semiconductor materials such as silicon carbide (SiC) and gallium nitride (GaN) MOSFETs, power converters have made significant progress in improving efficiency and power density. However, the high-frequency switching characteristics, fast rising edges, and high voltage and current waveforms of these new semiconductor materials have also brought more prominent electromagnetic interference (EMI) problems, especially conducted EMI. Conducted EMI is usually divided into two types: common mode (CM) and differential mode (DM), and the impact of common mode EMI on the system is particularly significant. Common mode noise usually propagates through the power line and the grounding system, which may lead to a decline in the performance of the internal circuits of the system, signal distortion, and even equipment failure, seriously affecting the reliability and stability of the system. Specifically, the CM EMI generated by the power converter not only interferes with the normal operation of other electronic devices, but also may cause voltage fluctuations on the transmission line, thereby affecting the dynamic response and accuracy of the system. Therefore, accurately predicting and evaluating the characteristics of CM noise becomes particularly important. This not only helps to optimize the system design, reduce EMI interference, but also serves as an early warning for potential EMI faults, helping to take corresponding suppression measures to ensure that the system can maintain a low electromagnetic radiation level while operating efficiently, thereby improving the overall performance and reliability of the system.

[0003] In practical applications, CM EMI prediction and suppression are usually carried out separately. Specifically, EMI prediction focuses on identifying and evaluating the characteristics of electromagnetic interference that may be generated in the system by establishing a model, so as to provide a basis for subsequent design optimization and suppression measures. EMI suppression, on the other hand, relies more on actual electromagnetic compatibility (EMC) design, including the use of filters, shielding, grounding, etc. to reduce or eliminate the impact of interference. Although these two are separate, they play a complementary role in improving system performance and electromagnetic compatibility.

[0004] In a digital active EMI filter, the harmonic prediction and suppression of periodic CM EMI can be coupled. That is, by predicting the noise harmonic changes of EMI in real time and combining with the dynamic adjustment function of the filter, more efficient interference suppression can be achieved. At present, not much research has delved deeply into this direction. In this coupling method, CM EMI prediction is not only used to identify potential interference sources in advance, but also can adjust the operating state or control parameters of the filter in real time, enabling it to respond dynamically according to the predicted interference characteristics, thereby achieving the best suppression effect. Specifically, a digital active EMI filter can use digital signal processing technology to monitor the electromagnetic interference signals in the system in real time, use a prediction model to identify the interference frequency, amplitude, and change trend in advance, and adjust the parameters of the filter (such as cut-off frequency, gain, etc.) based on this information, so as to achieve precise suppression of periodic EMI. This coupling strategy has great potential, especially when dealing with interference sources with complex and variable frequency characteristics, it can achieve more efficient and intelligent suppression. However, the current research in this field is still scarce. The main reason is that the real-time coupling of CM EMI prediction and suppression involves complex dynamic modeling, signal processing technology, and filter control strategies. Problems such as how to achieve high-precision real-time prediction, how to design an active filter that can adaptively adjust, and how to balance the real-time performance and accuracy between prediction and suppression are still technical problems that need to be solved urgently. In short, although the coupling of prediction and suppression in digital active EMI filters has not been widely studied, it provides a new idea for future electromagnetic compatibility design. By deeply exploring this field and combining advanced signal processing technology and intelligent control methods, it can provide strong support for efficient and precise EMI suppression. Summary of the Invention

[0005] Aiming at the problems existing in the prior art, the present invention realizes a digital active EMI suppression method based on noise harmonic prediction by using a linear stable impedance network (LISN) and a field programmable gate array (FPGA).

[0006] The technical solution of the present invention is as follows:

[0007] A digital active EMI suppression method based on noise harmonic prediction, used for a DC / DC converter, includes:

[0008] Connect a linear stable impedance network between the DC / DC converter and the input power supply;

[0009] Connect a digital active EMI filter circuit between the linear stable impedance network and the converter. The filter circuit includes an active part and a passive part. The active part includes a field programmable gate array (FPGA), an injection capacitor C i and an injection resistor R i, the passive part includes a common-mode choke L D and a Y capacitor C y ;

[0010] Use the noise source Vds to replace the switching transistor of the DC / DC converter. The noise source Vds is the voltage change generated during the switching process of the replaced device. Determine the relationship between the CM noise and the noise source Vds as

[0011]

[0012] where, Z LISN,CM is the CM impedance of the linear stable impedance network, Z Lwire = Z Lwire1 CZ Lwire2 Z Lwire1 is the impedance of the wire at the voltage input terminal of the DC / DC converter, Z Lwire2 is the impedance of the wire at the ground terminal of the DC / DC converter, Z Ci is the impedance of the injection capacitor C i , Z Ri is the impedance of the injection resistor R i , Z LD is the impedance of the choke L D , Z CY is the impedance of the decoupling capacitor C y ;

[0013] Replace the noise source with a trapezoidal signal, substitute the trapezoidal signal into the relationship between the CM noise and the noise source, and then solve the CM noise to determine the digital active EMI filtering signal VANTI(f):

[0014]

[0015] where, n is the harmonic multiple, f s is the switching frequency, S(n) AT is the trapezoidal signal spectrum,

[0016] In the FPGA, use the inverse Fourier transform to calculate the required digital active EMI filtering time-domain signal VANTI(t):

[0017]

[0018] Inject the filtered time-domain signal VANTI(t) into the DC / DC converter through the digital active EMI filtering circuit to achieve digital active EMI suppression

[0019] The beneficial effects of the present invention are as follows. A digital active EMI suppression method based on noise harmonic prediction of the present invention can achieve EMI suppression without a noise sampling circuit, with a good suppression effect, and is applicable to stable working conditions. A digital active EMI suppression method based on noise harmonic prediction of the present invention can, by determining in advance the relationship between the control mode and the CM noise, track the change of the CM noise in real time, and can also be used as an auxiliary tool to observe the common-mode noise level of the converter in real time. A digital active EMI suppression method based on noise harmonic prediction of the present invention can be applied to DC / DC power converters in stable working conditions. If an FPGA is used as the processor, it can also be used as a PWM generator, optimizing the control architecture of the converter, improving the overall power density, and having good engineering application value.

[0020] The present invention specifically aims at applying high-voltage, high-frequency, and high-power DC / DC converter systems. First, it studies the CM EMI mechanism of the converter system, establishes a CM behavior model of the power converter, and then determines the relational expression between the CM noise and the noise source. It establishes a digital active EMI filter circuit based on noise harmonic prediction, predicts the noise signal and the reverse noise signal for cancellation, and then injects them into the converter to achieve noise suppression. Through experimental tests and verifications, it provides ideas and theoretical models for noise suppression of the converter without a sampling circuit, and has important engineering value for the development of DC / DC high-power and high-frequency applications, as well as the optimization of the EMC characteristics and EMI suppression of the converter system. Brief Description of the Drawings

[0021] Figure 1 It is a schematic diagram of a test platform for a Buck DC / DC converter based on LISN provided by an embodiment of the present invention.

[0022] Figure 2 It is a schematic diagram of a circuit model of a Buck DC / DC converter based on LISN provided by an embodiment of the present invention.

[0023] Figure 3 It is a schematic diagram of an equivalent circuit model of a Buck DC / DC converter provided by an embodiment of the present invention.

[0024] Figure 4 It is a schematic diagram of a common-mode equivalent circuit model of a Buck DC / DC converter based on a current source provided by an embodiment of the present invention.

[0025] Figure 5 It is a schematic diagram of a common-mode equivalent circuit model of a Buck DC / DC converter based on a voltage source provided by an embodiment of the present invention.

[0026] Figure 6Schematic diagram of a simplified circuit model of a digital active EMI filter without a detection circuit provided by an embodiment of the present invention.

[0027] Figure 7 Schematic diagram of a common - mode equivalent circuit model based on a noise source and a complementary noise source provided by an embodiment of the present invention.

[0028] Figure 8 CM noise amplitude and phase curves for experimental measurement and prediction.

[0029] Figure 9 Spectrum comparison diagram before and after digital active EMI suppression based on noise harmonic prediction. Specific implementation manners

[0030] The present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0031] A digital active EMI suppression method based on noise harmonic prediction proposed by the present invention has the following principle: First, analyze the generation and conduction mechanism of the common - mode EMI of the converter, determine the converter control mode and the high - frequency parasitic parameters contained in its circuit model, obtain the common - mode EMI model of the power converter system, and realize CM EMI prediction; use the CM EMI spectrum (including phase and amplitude) predicted in advance by a signal processor (such as an FPGA), and through the inverse Fourier transform (IFFT), reverse the predicted signal into CM EMI, and then inject it into the power converter through a digital EMI filter to achieve EMI suppression. This process does not require sampling of EMI.

[0032] For a Buck DC / DC converter, the analysis and EMI suppression process are as follows:

[0033] Step 1: Analyze the generation and conduction mechanism of the EMI of the Buck DC / DC converter, and determine the main high - frequency parasitic parameters in the power converter. In addition to the basic circuit components, its main high - frequency parasitic parameters are composed of the high - frequency parasitic inductance L of the input wire wire and the parasitic capacitance C of the converter to the ground SG These parasitic parameters have a major impact on CM noise, and these parameters can be measured using an impedance network analyzer.

[0034] Step 2: Determine a DC / DC converter test platform based on a linear impedance network (LISN). For example, an actual measurement model of a Buck DC / DC converter, including a DC source V DC , LISN, Buck DC / DC converter, and the parasitic capacitance C of the mid - point of the converter half - bridge to the ground SG(The dv / dt at the midpoint of the half-bridge is relatively high), the output load RL, and the ground. The LISN includes the L1-L2 isolation circuit and the CN-RN-CL-RL high-frequency noise measurement branch. The Buck DC / DC converter includes the output filter inductor L out , the output capacitor C out , and the output resistor R load , the input wire inductance L wire1 -L wire2 , the input capacitor Cin, the control transistor Q1, the freewheeling transistor Q2, where C SG is the lumped parasitic capacitance of the midpoint of the Buck DC / DC converter half-bridge to the ground (the dv / dt at the midpoint of the half-bridge is relatively high).

[0035] Step 3: Analyze the EMI mechanism in the Buck DC / DC converter based on LISN, and establish an equivalent CM behavior model based on the substitution theorem.

[0036] Step 4: Analyze the effects of current sources and voltage sources on the CM noise of the power converter respectively, and determine the relationship between CM noise and noise sources (including current sources and voltage sources).

[0037] Step 5: Replace the noise source with a trapezoidal signal to obtain the harmonic amplitude and phase signals of the noise source. Substitute the trapezoidal signal into the relationship between CM noise and noise sources to predict the CM noise signal.

[0038] Step 6: Build a digital active EMI filter circuit platform and determine the filter circuit architecture.

[0039] Step 7: Based on the digital active EMI filter circuit platform, re-analyze the relationship between CM noise and noise sources (including current sources and voltage sources) under this platform, and then determine the digital active EMI filter signal V ANTI (f).

[0040] Step 8: Use the inverse Fourier transform (IFFT) in the FPGA to obtain the required digital active EMI filter time-domain signal V ANTI (t), and inject it into the power converter through the digital active EMI filter circuit. Then use LISN to measure the CM noise waveforms before and after digital active EMI suppression based on noise harmonic prediction, and determine the suppressed EMI level.

[0041] Based on the above theory, the EMI suppression process of the DC / DC converter can be summarized as the following steps:

[0042] 1) Replacing the control transistor Q1 with a voltage source and the freewheeling diode Q2 with a current source Id is actually a method for simplifying and equivalently modeling the power converter circuit. In this replacement model, the voltage source represents the operating state of the control transistor Q1, that is, it simulates the switching control behavior of Q1 and can generate similar voltage variations to effectively control the current flow; while the current source Id is used to simulate the function of the freewheeling diode Q2, simulating the freewheeling function provided by Q2 during the current on-off process. Specifically, when the control transistor Q1 performs a switching action, it will affect the voltage variation in the circuit, and the replacement method of the voltage source can more intuitively reflect the driving characteristics of Q1; while the main function of the freewheeling diode Q2 is to ensure the smooth flow of current in the converter by providing a continuous current path, and the replacement method of the current source Id enables the current variation to be more accurately simulated and controlled.

[0043] 2) Based on the theory of circuit analysis, considering the influences of the current source and voltage source respectively, a decoupled model of the current source and voltage source is obtained. After obtaining the decoupled model of the current source and voltage source, we further optimize the entire circuit model. By combining the influencing factors of the current source and voltage source, a more accurate equivalent common-mode (CM) electromagnetic interference (EMI) model of the Buck DC / DC converter is obtained. This optimization process involves comprehensive considerations of various factors such as the parasitic effects of components in the circuit, the interaction between current and voltage, and filtering and suppression measures. Finally, through this equivalent CM EMI model, we can clearly establish the relationship between the CM noise source and the CM noise, providing a theoretical basis for the location, isolation, and suppression of the noise source, and guiding the subsequent electromagnetic compatibility design and optimization.

[0044] 3) Replace the noise source with a trapezoidal signal. The spectrum of the trapezoidal signal can be given in advance. The common-mode noise is predicted through the CM EMI model, and its accuracy is verified. Replacing the noise source with a trapezoidal signal is a commonly used simplification method, aiming to approximate the noise source in the power converter through the known trapezoidal signal model. In this process, the spectral characteristics of the trapezoidal signal can be given in advance or obtained through experiments, and used as input parameters to analyze its contribution to the system's common-mode noise. The trapezoidal signal usually has a clear spectral composition, mainly composed of the fundamental frequency and its harmonic components, so it can effectively simulate the high-frequency noise generated in periodic switching operations. By substituting the trapezoidal signal into the CM EMI model, we can use this model to predict the common-mode noise of the system. The CM EMI model can analyze how these factors work together to finally generate common-mode noise by modeling the coupling effects of various elements such as voltage sources, current sources, parasitic capacitors, and inductors in the circuit. Specifically, the trapezoidal signal as a noise source can help us simulate the high-frequency voltage waveform caused by the switching action of the control tube and determine its impact on the system noise through spectral analysis. Once the common-mode noise is predicted through the CM EMI model, we need to further verify the accuracy of the prediction result. This is usually done by comparing it with the actual measurement results. By measuring the output common-mode noise of the converter in the experimental environment and comparing it with the model prediction result, the accuracy and reliability of the model under different working conditions can be evaluated. If the prediction result is close to the actual noise, it indicates that the model is effective; otherwise, the model needs to be further adjusted and optimized, which may include re-evaluating the spectral characteristics of the trapezoidal signal, correcting the parameters of parasitic elements, or modeling the noise source in more detail. This method not only provides an effective means for predicting common-mode noise but also enables the early identification of possible noise problems in the design stage, providing strong support for electromagnetic compatibility (EMC) optimization.

[0045] 4) Build a digital active EMI filter without a detection circuit. Based on this platform, re-predict the common-mode noise, and then calculate the required digital active EMI compensation signal V ANTI , and then inject it into the converter through the digital active EMI filter circuit to achieve EMI attenuation.

[0046] Example:

[0047] In this example, digital active EMI suppression is carried out for a Buck DC / DC converter, specifically including:

[0048] Based on the LISN test platform, analyze the generation and conduction path of the common-mode (CM) noise of the Buck DC / DC converter, as Figure 1As shown. During the operation of a Buck DC / DC converter, the switching device (such as a field-effect transistor) generates a high dv / dt (rate of voltage change) during the switching operation. This high dv / dt induces a common-mode current (I SG ) on the lumped parasitic capacitance (C CM ) between the Buck DC / DC converter and ground. Specifically, this common-mode current flows from the converter to the ground, then returns through the grounded terminal of the LISN, and finally flows back to the Buck DC / DC converter through the LISN. During this process, the generation of the common-mode current (I CM ) is mainly caused by the voltage change during the switching of the field-effect transistor (FET). Especially at the moment when the switching device turns on and off, the change in dv / dt causes the capacitor to charge and discharge, thereby forming a common-mode current in the circuit. The conduction path of this common-mode current is: converter → C SG → ground → LISN → converter. Here, C SG represents the parasitic capacitance between the control terminal and the ground. The common-mode current ICM flows through this path during the switching process. Since the current needs a loop to close, the entire coupling path forms a closed-loop circuit. Each part in this conduction path plays a specific role:

[0049] 1) Converter: The switching action of the switching device in the Buck DC / DC converter induces the generation of the common-mode current.

[0050] 2) C SG : The parasitic capacitance in the converter (such as the gate-source capacitance of the switching device) is an important coupling medium between the common-mode current and the ground.

[0051] 3) Ground: The common-mode current flows to the ground, serving as a reference and return path.

[0052] 4) LISN: The LISN (Line Impedance Stabilization Network) is used to isolate high-frequency noise between the power grid and the device and monitor and measure the current and voltage of the noise. The role of the LISN is to provide a path for the common-mode current to flow back to the converter, and its grounded terminal also provides a function of noise isolation and measurement.

[0053] Establishing a CM behavior model for the power converter: As shown in Figure 2 and Figure 3 , Figure 2 is the circuit model of the Buck DC / DC converter, Figure 3This is its equivalent circuit model. To further analyze the common-mode (CM) noise behavior in a power converter, we use the substitution theorem to equivalently model the CM noise sources and coupling paths. The substitution theorem helps us simplify circuit analysis and more accurately predict the generation and propagation paths of CM noise by replacing complex components (such as field effect transistors Q1 and Q2) in the actual circuit with simpler equivalent components. In a Buck DC / DC converter, the field effect transistor Q1 is the switching element for the control signal, and its switching process causes voltage changes, resulting in a high dv / dt and generating noise. Figure 3 Figure 1 shows the equivalent circuit model of the Buck DC / DC converter, in which the field effect transistors Q1 and Q2 are replaced by the voltage source Vds and the current source Id, respectively. This equivalent circuit can help us simplify complex current and voltage paths and focus on analyzing the generation and propagation mechanisms of CM noise. The voltage source Vds and the current source Id form noise sources and conduction paths by coupling with parasitic capacitors, inductors, and other components in the circuit. The equivalent circuit model provides a clear framework for subsequent EMI analysis and noise suppression.

[0054] Analyze the common-mode noise propagation paths of the voltage source Vds and the current source Id separately. As Figure 4 shown, when we only consider the current source Id, we can short-circuit the voltage source Vds, so that the influence of the current source Id on the common-mode noise can be analyzed separately. The current source Id will form common-mode noise through components such as capacitors and inductors in the circuit and propagate through specific paths. As Figure 5 shown, when we only consider the voltage source Vds, we can open-circuit the current source Id and ignore its contribution to the propagation of common-mode noise. The voltage source Vds represents the voltage change (high dv / dt) generated by the switching transistor Q1 during the switching process, which will cause common-mode noise. Therefore, the CM noise path of the current source Id is: current source Id - C SG - ground - LISN - input wire - current source Id; the CM noise path of the voltage source Vds is: voltage source Vds - C SG - ground - LISN - input wire - voltage source Vds. At this time, the output terminal Lout - Cout - Rload can be ignored. In the frequency band from 150 kHz to 30 MHz, the impedance of the input capacitor Cin is low, so it can be regarded as a short circuit. In this frequency band, the current source Id will be short-circuited by Cin, thus greatly reducing its influence on the common-mode noise. Therefore, in this frequency band, the contribution of the current source Id to the common-mode noise can be ignored, and the main noise source will come from the voltage source Vds.

[0055] Determine the relationship between the CM noise source and the CM noise of the Buck DC / DC converter: To simplify this common-mode circuit model, simplify the CM impedance of LISN to

[0056] Z LISN,CM = 0.5(Z CL + Z RL ) = 0.5(Z CN + Z RN ) ≈ 25 Ω (1)

[0057] Cin is regarded as short - circuited, so the CM impedance of the Buck DC / DC converter is

[0058] Z CM = (Z Lwire1 + Z Lwire2 )C Z CSG (2)

[0059] where Z Lwire1 is the impedance of Lwire1, Z Lwire2 is the impedance of Lwire2, Z CSG is the impedance of C SG . The voltage source Vds can be replaced by a trapezoidal wave. The common - mode noise can be expressed as

[0060]

[0061] So the common - mode gain of the voltage source is

[0062]

[0063] The spectrum of the trapezoidal wave can be expressed as:

[0064]

[0065] Therefore, without adding a filter circuit, the predicted common - mode harmonic noise spectrum can be expressed as:

[0066] V CM,p (kf s ) = G vds (kf s )S(k) AT (6)

[0067] There are two ways to predict the common - mode noise: measuring Vds or using a trapezoidal wave to replace Vds.

[0068] Build a digital active EMI filter circuit based on noise harmonic prediction, predict the injection signal V ANTI required to cancel the EMI noise, and then verify it through experiments. The built digital active EMI filter circuit includes: a CM injection circuit, a decoupling circuit, an FPGA, and a computer. The role of the computer is to generate the injection signal V ANTI (t) required to cancel the EMI noise in the time domain. The FPGA will use the injection signal VANTI (t) is converted into a digital signal and then injected into the main line through an injection circuit. The CM injection circuit includes: a DAC, an injection resistor Ri, and an injection capacitor Ci, and the decoupling circuit includes a CM choke coil L D and a Y capacitor C y . At the same time, the FPGA gives complementary PWM signals to drive the switching transistors Q1 - Q2. Further, after adding a digital active EMI filtering circuit based on noise harmonic prediction, the CM noise will change at this time. Therefore, it is necessary to re - predict. After building a digital active EMI filtering platform based on noise harmonic prediction, the corrected CM EMI model is obtained as Figure 7 shown. At this time, the common - mode noise change is

[0069]

[0070] where

[0071] At this time, the voltage drop generated by the injected signal Vanti on the LISN is:

[0072]

[0073] where

[0074] Assume that the noise drops to 0 after injecting V ANTI :

[0075]

[0076] Further, V ANTI can be determined as:

[0077]

[0078] Use a trapezoidal wave to replace the voltage source. The frequency spectrum of the trapezoidal wave is expressed as:

[0079]

[0080] where j is the imaginary number, V DC is the DC source input value, T is the switching period, τ r represents the time required for the signal to jump from the low level to the high level, τ f represents the time required for the signal to drop from the high level to the low level, and D is the duty cycle of the converter. Then the V ANTI predicted by the processor is:

[0081]

[0082] Further, the time - domain V ANTI (t) can be obtained by using IFFT:

[0083]

[0084] Next, experimental tests are used to verify the effectiveness of the proposed digital active EMI suppression method, which is applicable to power converter systems with stable EMI, especially suitable for Buck DC / DC converters. This method does not require a sampling circuit and realizes EMI suppression by predicting noise harmonics, providing a way for dynamic EMI monitoring and theoretical support for digital active-based EMC optimization and EMI suppression in the application of high-frequency and high-power-density power converters.

[0085] To verify the effectiveness of a digital active EMI suppression method based on noise harmonic prediction provided by the present invention, a digital active EMI filtering platform based on noise harmonic prediction was built on an EMI receiver test platform, and the EMI suppression effects before and after digital active EMI filtering based on noise harmonic prediction were given. Among them, the experimental parameters are shown in Table 1.

[0086] Table 1 Experimental parameters

[0087]

[0088] Figure 8 The measured and predicted values of the CM noise are given. Generally speaking, the analyzed CM EMI model is accurate and reliable in the EMI prediction of power converter systems. The main reasons for the existing errors are as follows:

[0089] (1) The resolution of the sampled signal noise is low, so there will be deviations in the sampled noise signal, especially in the high-frequency range;

[0090] (2) The larger the noise amplitude at a certain frequency point calculated by FFT, the more reliable it is, which means that the smaller the noise amplitude after FFT, the greater the error;

[0091] Figure 9 The measured values of the CM noise before and after digital active EMI suppression based on noise harmonic prediction are given. It can be seen that after suppression, a maximum EMI attenuation of about 40 dB can be achieved, and the low-frequency suppression is mostly above 30 dB. Therefore, a digital active EMI suppression method based on noise harmonic prediction proposed by the present invention is very effective.

Claims

1. A digital active EMI suppression method based on noise harmonic prediction, which is used for a DC / DC converter, is characterized in that Including: Connect a linear stable impedance network between the DC / DC converter and the input power supply; Connect a digital active EMI filter circuit between a linear stable impedance network and a converter. The filter circuit includes an active part and a passive part. The active part includes a field programmable gate array (FPGA), an injection capacitor C i and an injection resistor R i . The passive part includes a common mode choke L D and a Y capacitor C y ; Use the noise source Vds to replace the switching transistor of the DC / DC converter. The noise source Vds is the voltage change generated during the switching process of the replaced device. Determine the relationship between the CM noise and the noise source Vds as Among them, Z LISN,CM is the CM impedance of the linear stable impedance network, Z Lwire = Z Lwire1 CZ Lwire2 , Z Lwire1 is the impedance of the wire at the voltage input terminal of the DC / DC converter, Z Lwire2 is the impedance of the wire at the ground terminal of the DC / DC converter, Z Ci is the impedance of the injection capacitor C i , Z Ri is the impedance of the injection resistor R i , Z LD is the impedance of the choke coil L D , Z CY is the impedance of the decoupling capacitor C y ; Replace the noise source with a trapezoidal signal, substitute the trapezoidal signal into the relationship between the CM noise and the noise source, and then solve the CM noise to determine the digital active EMI filtering signal VANTI(f): where n is the harmonic multiple, f s is the switching frequency, and S(n) AT is the trapezoidal signal spectrum, In the FPGA, the required digital active EMI filter time-domain signal VANTI(t) is calculated using the inverse Fourier transform: Inject the filtered time-domain signal VANTI(t) into the DC / DC converter through a digital active EMI filtering circuit to achieve digital active EMI suppression.

Images