Brief Analysis of the Network Principle of Power Supply Filter

　　In systems such as communication switching power supplies or switches, most of the conducted EMI is caused by the AC power supply shared by different devices or systems. The interference source can be injected into the broadband conducted emission through the wires and pollute the distribution lines. At the same time, the wires supply power to other sensitive devices. At this time, the common mode interference is connected across the distribution power supply and impedance, thus affecting the normal operation of all devices on the power supply. In addition, when two or more circuits share a common loop such as the ground level, common impedance coupling will occur, and some transmission signals will be coupled to the line and polluted by EMI, which will also cause line pollution. EMI power supply filters can effectively suppress EMI and effectively play a role in decoupling.

　　1. Structure of EMI power filter

　　The network structure of the EMI power filter is mainly composed of differential mode (NM, DM) capacitors, common mode (CM) capacitors, differential mode inductors, common mode inductors, resistors, etc. From the general EMI power filter circuit diagram in Figure 1, you can see the combination of each part.

　　There are many types of network structures for EMI power filters. The following is a brief description of several commonly used ones.

　　The filter network shown in Figure 2 is quite common. Both common mode and differential mode filter networks have certain characteristics: they have few structural components, only two inductors L1 and L2, three capacitors, two CY, one CX and one resistor R. The series arm is the high impedance port of the common mode and differential mode filter network, and its parallel arm is the low impedance port. In order to better control EMI, the maximum source impedance and load impedance should be generated. That is, any circuit with low impedance (such as grounded Y capacitor) is connected in series, and the parallel arm is connected to a high impedance circuit.

　　The typical and differential mode insertion losses measured in the 50 system increase monotonically in the range of 0.01 to 1 MHz, and the slope of the curve depends on the component parameters. See the difference between Figure 2 (AB). In addition to the different parameters, the position of the resistor R is also the same. The inductance of L1 and L2 is about 0.3 to 24 niH. The inductance of the power filter L1 and L2 with a large rated current is smaller. The capacitance of Cx is 0.015 to 10, and the capacitance of 1 CY is 0.

　　The capacity is limited by the maximum leakage current, and the value range is about 1000PF ~ 10000PF. Generally speaking, when the inductance of L1, L2 and the capacitance of CY are large, the common mode insertion loss in the low frequency band is high, the capacitance of CX is small, and the differential mode insertion loss in the high frequency band is better than when CY is large, but the insertion loss at low frequency is poor.

　2. Attenuation performance and plug-in performance of EMI power filter

　　Measurement of input loss ratio q

　　The most common method of specifying the spectral performance of EMI power filters is to use the attenuation that varies with frequency within a specified frequency range. The attenuation of EMI power filter as a function of frequency refers to the ratio of the output voltage before and after its insertion. The attenuation AdB expressed in 6 is derived from the following formula AdB=10Logl0: Where k: power conducted to the load after the filter is inserted Pb: power conducted to the load before the filter is inserted The impedance and load impedance are closely related to the performance of EMI power filters. One or more stages of the filter can be used as a "sacrificial" element to establish a false source or load impedance.

　　This results in different equivalent filter stages n, as shown in Table 1. Figures 1 and 2 show the structure of basic filters applied to circuits with different source and load impedance combinations. The filters given are all low-pass (i.e., they use series inductors and parallel capacitors). The purpose is: A. Either connect the filter series inductor to a low impedance source R or connect the parallel capacitor to a high impedance source, in short, make the impedance of the source and filter elements approximately equal at the required cut-off frequency. Similarly, the series inductor faces the low impedance load, while the parallel capacitor faces the high impedance load, which ensures the best use of the filter elements and partially compensates for some source or load impedances of typical transmission lines that vary over a wide frequency range starting from about 100 times the power frequency. Figure 2 shows the common-mode insertion loss curve of the ZYH-ER-30A EMI source filter of Zhongzihao Electric Co., Ltd.

　　Table 1. Cut-off frequency of equivalent filter order n=1 when the installed impedance is different from the designed source and load impedance

　　It is worth noting that in the case of open circuits, where neither the connectors nor the filters are shielded, direct input-to-output couplings of the order of 40-60 dB are not uncommon, especially in small circuits and integrated circuits. Unless special precautions are taken in the design and manufacture of the filter, the filter may provide little or no attenuation at frequencies two or more times higher than the cutoff.

　　EMI power filter is a source-to-source network with complex conduction characteristics, which is completely dependent on the power supply and load impedance, and the filter's attenuation performance is represented by the complex transmission characteristics. However, in the power line environment, the power supply and load impedance are not limited. Therefore, the industry uses the zero resistance power supply and terminal load attenuation method as the filter's insertion loss IL

　　Where PL(Rd) is the power transmitted from the power supply to the load when there is no filter, and PL is the power transmitted when the filter is inserted between the power supply and the load. Insertion loss can be expressed in terms of voltage or power ratio as follows:

　　Where VL (R6f) is the maximum value measured when there is no filter, IL (R6f) is the maximum value measured when the filter is inserted, and VL and L are the measured values ​​when the filter is inserted. The so-called insertion loss is the ratio of the signal and voltage transmitted from the power supply to the load when there is no filter to the signal voltage transmitted from the power supply to the load when the filter is inserted (expressed in 6 days).

　　Insertion loss measurements are meaningful only when the terminal impedance is standardized, and the results obtained are only applicable to the same circuit. The most common configuration is a resistive configuration with 500 Ω source and load impedances. See Figure 4 for common mode insertion loss measurements. In common mode, the power circulates between the live and neutral terminals and the common (ground) lead. The common mode insertion loss can be measured by coupling the live wires at both ends of the filter to the neutral terminal. Figure 4. Common mode insertion loss.

　　Figure 5 Differential mode insertion loss diagram

　　In differential mode, the signal at the live and neutral terminals is the same, and the current with opposite phases circulates between the live and neutral wires. The differential mode insertion loss can be measured using a 50° 180°" power divider.