Study on power electronic converter electromagnetic interference experiment method

Effective measurement is the foundation for predicting and suppressing EMI. This article implements an effective EMI experimental method to conduct conducted EMI testing on high-power three-phase converter power cables. Detailed analysis and discussion are conducted on the experimental instruments, experimental layout, installation details, and other issues during the experimental setup process. The experimental method has improved the EMI testing technology of the converter and has a certain universality, which has practical significance for predicting and analyzing the conducted electromagnetic sensitivity of power electronic equipment.


Introduction
In order to reduce the impact of electromagnetic interference between power electronic devices, improve their adaptability in complex electromagnetic environments, and enable various devices to work normally in the same electromagnetic environment, it is essential to study the electromagnetic compatibility of power semiconductor devices [1], and this problem is particularly evident in high-power converters.
Although there is relatively abundant research literature works in this field [2]- [8], there is relatively little research work considering the impact of changes in the position of power cables during the testing process.In [1]- [6], meaningful research on transformer interference testing methods and suppression techniques was conducted, but they were all based on the testing situation of low-power equipment in operation.Overall, there is relatively little research on testing methods for conducted interference in power cables of high-power three-phase converters.Based on this, this article focuses on the electromagnetic interference testing technology of high-power three-phase power converter power cables.

Test instrument
The conducted interference test is mainly to measure the CM noise and DM noise in the wire at the beginning of the power line or the socket.The instruments commonly used include EMI electromagnetic disturbance measurement receiver, linear impedance stabilization network (LISN), voltage probe, current probe, etc. [4].

EMI receiver
Most receivers use the heterodyne operating principle, which makes it easier to obtain a higher frequency spectrum and faster measurement results than the traditional Fourier analyzer, and the frequency spectrum of the input signal is calculated directly from the frequency domain analysis.Its composition is shown in Figure 1.The working principle of the EMI receiver is that: first, the input RF signal processed by the attenuator, preselector, and amplifier is neutralized with the mixer; By adjusting the local oscillator frequency, a fixed intermediate frequency can be obtained [5]; Then the resolution of the analyzer is obtained after the IF filter processing; Then, the IF signal is converted into video signal by the envelope detector and weighted by the detector; Finally, the measured data that meet the requirements of CISPR will be displayed [6].As the core instrument of the EMI test, the receiver measures the output voltage of the equipment to be tested (EUT).

Linear impedance stabilization network (LISN)
As the necessary instrument for conducting EMI measurement of EUT, as shown in Figure 2, LISN has a low-pass filter composed of L1 and C1, which can isolate the high-frequency interference generated on the grid side.The DC isolation capacitor C2 transmits the high-frequency interference generated on the EUT side to the receiver, so as to ensure that the interference measured by the receiver is actually generated by the EUT.In addition to filtering unwanted input EMI on the grid side and providing "pure" power for EUT, LISN can also provide a standard impedance for EUT power input terminals to isolate the impact of grid impedance caused by load switches and other factors on EUT interference.
In the low-frequency band (50~60 Hz), the inductance in LISN shows low impedance, which can be regarded as a short circuit; Capacitance shows high impedance, which can be regarded as an open circuit; AC power can pass through the EUT power supply smoothly.
In the high-frequency band (≥10 kHz), the inductance is like an open circuit, and the capacitance is like a short circuit [7].The interference on the AC power supply side will be shorted to ground by C1, while the interference generated by EUT will first reach the output end of LISN through C2, and then be measured by the receiver.

Voltage probe
The voltage probe is a converter used to measure the transmission line voltage, and the basic quantity measured is the common mode voltage.When LISN's current bearing capacity is insufficient, or the test site cannot meet the complex test configuration, the voltage probe can be used to replace LISN for testing.

Current Probe
Since it is not necessary to directly contact the current probe with the measured power line, the measured interference current can be converted into interference voltage.Therefore, the current probe is usually used as an circular transformer in the low-frequency conducted interference test of GJB152A-97.The current probe is a weak coupling transformer.The primary winding of the coil is generally connected to the coaxial connector, and the power line around it can be regarded as the secondary coil.

Conducted interference test items
The reliable experimental environment of this paper is provided by the EMC laboratory of the school.The laboratory has not only successfully passed the CNAS and CMA on-site review, but also passed the acceptance of the Beijing Radio Measurement and Testing Institute, thus initially possessing the national-level testing qualification.In addition, it is equipped with a signal generator, impedance analyzer, receiver, and other instruments and equipment, as well as a variety of electromagnetic and circuit design simulation software.Based on the requirements of GJB152A, the CE101 and CE102 projects are measured to check the interference on the AC side of the three-phase converter.It is confirmed that the CM/DM interference generated by RF current enters the AC power supply section of the laboratory.To maintain generality, a detailed analysis of the testing method is conducted and useful conclusions are drawn.

CE101: 25~10 kHz power line conducted interference test
CE101 is mainly used to measure the conducted emission on the EUT power input line (including the loop) and is applicable to all power conductors (including the return line, but excluding EUT power output terminal conductors).The test layout of this project is shown in Figure 3.The current in this low-frequency band will be transmitted to the housing of the electrical equipment through the filter capacitor, thus interfering with the low-frequency sensitivity receiver on the equipment.Therefore, if the measurement result exceeds the standard, the method of improving the power factor of the power supply can be used to reduce the impact of conducted interference on the equipment on the power supply line in this frequency band.

CE102: 10 kHz~10 MHz power line conducted interference test
CE102 uses LISN to test the magnitude of interference voltage transmitted to the power grid along the EUT input power line.The test layout of this project is shown in Figure 4, where LISN is used to couple EUT noise, that is, to obtain interference signals.

Configuration and key considerations of experimental facilities
This article uses the test configuration shown in Figure 5. Due to the fact that the experimental object is the AC side cable of the high-power converter, it is more convenient to place it on the ground.However, considering the reliability and accuracy of the experiment, the LISN is directly placed on a sufficiently large test table covered with metal plates, and the EUT is also placed on it to ensure more accurate test results.
It should be noted that EUT is connected to a LISN, and other equipment is powered by another LISN.The distance between EUT and LISN is ≥ 100 cm.The rear of the desktop should exceed 0.6% of the rear of the EUT (including peripherals), 5 m.The excessively long I/O cable shall be bundled in the middle of the cable with a length of ≤ 40 cm and the cable that cannot be bundled shall be arranged in a serpentine routing mode.Meanwhile, the test table shall be placed 40 cm away from the vertical conductive wall lapped with the ground.

Factors to consider and analysis of test results
Firstly, this experiment uses a current probe to preliminarily determine the presence of DM interference and CM interference in the converter wires.When the current probe only surrounds one wire (phase or neutral), the output voltage of the probe will be proportional to the DM or CM current flowing through this wire.When wrapping around two wires (one is the input wire and the other is the return wire), if the output voltage is zero, only DM current exists on the wire.If the output voltage is not zero, CM current must exist on the wire.When surrounding all wires (input and return lines), if the output voltage is zero, then there is CM current on the ground wire; If the output voltage is not zero, then CM current exists on all wires.Especially crucial, although the construction characteristics of the current probe make it only respond to the current surrounding the wire and effectively isolate the external field, in multiple tests in this article, it is found that using the current probe as close as possible to the LISN side can ensure more accurate testing results.Instead of being commonly recognized, when used in low-frequency testing, it is placed between LISN and EUT.For the existing research objects in the laboratory, as shown in Figures 6 and 7, when the test frequency is less than 0.5 MHz, the output voltage waveform changes little under load and no-load conditions; When the test frequency is greater than 0.5 MHz, compared with the voltage stability of the no-load output terminal, the fluctuation of the output voltage of the switching power supply under load is more obvious than that under no-load due to the influence of parasitic parameters in the load resistance; When it reaches 10 MHz, the fluctuation amplitude changes more obviously.With the increase in frequency, the voltage fluctuation at the output end of the switching power supply is abnormal.Interference on L1 measured by the current probe, however, is stable, as shown in Figure 8.
To sum up, when the test level is 10 V and the frequency is below 0.5 MHz, the output terminal of the switching power supply can obtain a stable 5 V voltage value under both no-load and load-connected conditions.With the increase in frequency, the output voltage waveform of the switching power supply fluctuates more and more under no load.Compared with no-load, the output voltage waveform with resistance load also shows an upward trend, but the fluctuation amplitude is more obvious.Last but not least, to avoid the influence of conducted interference on the AC power supply in the low-frequency range, which further affects the continuation of testing, it is best to add a power line filter between the AC power supply and LISN to reduce the impact of conducted EMI on the system by filtering the AC power supply itself.
The interference in the power cables of high-power power electronic converters cannot be ignored and is the basic source of conducted interference in power electronic equipment.Accurate and effective practical testing is the foundation and guarantee for EMI prevention and suppression.Therefore, interference testing on AC side cables is particularly important.

Conclusion
This article designs an experimental method for conducting EMI in high-power three-phase converter power cables and conducts theoretical analysis and experimental system construction.The anti-interference performance verified by the experiment proves the universality of this experimental method, improves the EMI testing technology of the converter, and contributes a beneficial idea to reduce the impact of electromagnetic interference between power electronic devices and improve their adaptability in complex electromagnetic environments.

Figure 5 .
Figure 5. Layout of the test system.

Figure 8 .
Figure 8. Interference on L1 measured by the current probe.