Common errors encountered in the use of power supply filters

　　During the experimental testing process, we often encounter such a situation: although the design engineer connected a power filter to the power line of the equipment, the equipment still cannot pass the "conducted interference voltage emission" test. The engineer suspects that the filtering effect of the filter is not good, and keeps replacing the filter, but still cannot get the ideal effect.

　　The reasons for analyzing the equipment exceeding the standard are nothing more than the following two aspects:

　　1) The disturbance generated by the equipment is too strong;

　　2) Insufficient filtering of the equipment.

　　For the first case, we can solve it by taking measures at the disturbance source to reduce the intensity of the disturbance, or by increasing the order of the power filter to improve the filter's ability to suppress disturbances. For the second case, in addition to the poor performance of the filter itself, the installation method of the filter has an impact on its

　　This is often overlooked by design engineers. In many tests, we can change the installation method of the filter to achieve the desired effect.

　　The following are some common examples of how incorrect filter installation affects filter performance.

　　1 Input line is too long

　　After the power cord of many devices enters the chassis, it is connected to the input end of the filter through a long wire. For example, the power cord enters from the rear panel of the chassis, runs to the power switch on the front panel, and then returns to the rear panel to connect to the filter. Or the installation location of the filter is far from the power cord entrance, resulting in a too long lead. As shown in Figure 1.

　　Figure 1 Schematic diagram of a power cord that is too long

　　Because the lead from the power inlet to the filter input is too long, the electromagnetic interference generated by the equipment is re-coupled to the power line through capacitive or inductive coupling. The higher the frequency of the interference signal, the stronger the coupling, causing the experiment to fail.

　　2. Filter input and output lines are routed in parallel

　　In order to make the wiring inside the chassis beautiful, some engineers often bundle the cables together, which is not allowed for power cables. If the input and output lines of the power filter are parallel routed or bundled together, due to the distributed capacitance between the parallel transmission lines, this routing method is equivalent to connecting a capacitor between the input and output lines of the filter, providing a path for the interference signal to bypass the filter, resulting in a significant decrease in the performance of the filter, and even failure at high frequencies (as shown in Figure 2). The size of the equivalent capacitance is inversely proportional to the distance between the wires and directly proportional to the length of the parallel routing. The larger the equivalent capacitance, the greater the impact on the filter performance.

　　Figure 2 Effect of parallel routing on the filter

　　3 The filter is not grounded well. The filter housing is not connected to the metal.

　　Good chassis connection

　　This situation is also quite common. When many engineers install filters, the overlap between the filter housing and the chassis is poor (there is insulating paint); at the same time, the grounding wire used is long, which will cause the high-frequency characteristics of the filter to deteriorate and reduce the filtering performance. Due to the long grounding wire, the distribution of the wire at high frequencies

　　Inductance cannot be ignored. If the filter is well bonded, the interference signal can be directly grounded through the housing.

　　Poor, equivalent to a distributed capacitance between the filter shell (ground) and the chassis, which will lead to a large ground impedance of the filter at high frequencies, especially

　　Near the frequency of the distributed inductance and distributed capacitance resonance, the ground impedance tends to infinity. The effect of poor filter grounding on filter performance is shown in Figure 3.

　　As can be seen from Figure 3, due to the poor grounding of the filter and the large grounding impedance, some interference signals can pass through the filter. In order to solve the problem of poor overlap, the insulating paint on the chassis should be scraped off to ensure that the filter housing and the chassis have a good electrical connection.

　　Figure 3 Effect of poor filter grounding on filter performance

　　Figure 4 shows an example of an ideal installation of a power filter for reference by design engineers.

　　Figure 4 Power filter installation example

　　In this installation method, the filter shell and the casing are in good contact, blocking the opening of the power cord on the chassis, improving the shielding performance of the chassis; in addition, the input and output lines of the filter are isolated by the chassis shielding, eliminating the interference coupling between the input and output lines and ensuring the filtering performance of the filter.

　　The installation method of the filter directly affects the filtering effect of the filter. In order to give full play to the performance of the filter, the following principles should be followed when installing the filter:

　　1) Install near the power inlet, preferably covered with a filter housing

　　Power cord entry hole on the chassis;

　　2) The shorter the ground wire, the better;

　　3) The filter housing and chassis are well overlapped;

　　4) The filter input and output lines are separated and cannot be parallel or crossed;

　　5) Avoid strong interference sources near the filter.

　　4 Conclusion

　　This paper mainly introduces the calibration method of electromagnetic pulse sensor under strong field strength. The device under test uses analog optical fiber transmission system to transmit pulse signal, which has the characteristics of low noise, small nonlinear distortion and large dynamic range. The peak response sensitivity of the sensor system is obtained by measurement. A comparison between a group of laboratories shows that this calibration method has good consistency.

　　The experimental results show that: on the one hand, based on the existing electromagnetic pulse simulator, a pulse electric field that meets the calibration requirements can be generated. The waveform measured by the electric field sensor in the experiment is consistent with the waveform measured by the simulator load voltage. The voltage measured by the voltage divider can obtain the electric field in the standard device; on the other hand, the calibration data obtained in different calibration devices have good consistency. The effect of sensor placement on the uniformity of the field was not considered in this study, and will be further evaluated in the next step of the study.