Abstract: This paper proposes a method for directly detecting short-circuit faults in IGBTs. Based on a detailed analysis of the IGBT short-circuit detection principle, a corresponding IGBT short-circuit protection circuit is presented. Simulation and experimental results demonstrate that the circuit operates stably and reliably, effectively protecting the IGBT. Keywords: IGBT, short-circuit protection circuit design. The fundamental task of a solid-state power supply is to safely and reliably provide the required electrical energy to the load. For electronic equipment, the power supply is its core component. Besides requiring the power supply to provide a high-quality output voltage, loads also place higher demands on the reliability of the power supply system. IGBTs are widely used devices with self-turn-off capability and high switching frequency, and are widely used in various solid-state power supplies. However, if not properly controlled, they are easily damaged. It is generally believed that there are two main reasons for IGBT damage: one is that the IGBT exits the saturation region and enters the amplification region, increasing switching losses; the other is that a short circuit occurs in the IGBT, generating a large transient current, thus damaging the IGBT. IGBT protection typically employs a fast self-protection method. This means that when a fault occurs, the IGBT drive circuit is shut down, implementing desaturation protection within the drive circuit; or, when a short circuit occurs, the IGBT is quickly shut down. Depending on the monitored object, IGBT short-circuit protection can be divided into Uge monitoring and Uce monitoring methods. Both are based on similar principles, utilizing the phenomenon that Uge or Uce increases when the collector current IC rises. When Uge or Uce exceeds Uge_sat or Uce_sat, the IGBT drive circuit is automatically shut down. Since Uge remains relatively constant during a fault, while Uce varies considerably, and the change in Uge during desaturation is small and difficult to predict, Uce monitoring technology is generally used for IGBT protection in practice. The IGBT protection circuit studied in this paper protects the IGBT by monitoring the voltage drop Uce when the IGBT is on. Using the IGBT short-circuit protection circuit described in this paper achieves fast protection while saving the need for a Hall current sensor to detect the short-circuit current, thus reducing the overall system cost. Practice has proven that this circuit has significant practical value, especially in applications with low DC bus voltage, where it has broad application prospects. This circuit has been successfully applied in a certain type of high-frequency inverter. 1. Working Principle of Short Circuit Protection Figure 1(a) shows the H-bridge PWM converter circuit operating in PWM rectification mode (this figure is the equivalent circuit under the positive half-wave input of a sine wave; the two IGBTs of the upper half-bridge are not shown). Figure 1(b) shows the drive signals and related device waveforms of the two high-power devices in the lower half-bridge. The analysis will now be based on the positive half-wave operation process (for a three-phase PWM circuit, the analysis process and conclusions of the PWM circuit are basically similar in rectification, inversion, or single-phase DC/DC operation). In the circuit shown in Figure 1, during the positive half-cycle of the mains power supply Us, the high-frequency drive signal shown in Ug2.4 is applied to the gates of the two IGBTs in the lower half-bridge, resulting in the tube voltage drop waveform UT2D. The working process is analyzed as follows: At times t1 to t2, under the action of the driving signal, T2 and T4 are turned on (actually, T2 is turned on and T4 is in freewheeling state). Under the action of Us, the current through the inductor LS increases, forming a transistor voltage drop waveform that rises exponentially on transistor T2 as shown in UT2D in Figure 1(b). This voltage drop is the voltage drop generated by the on-state current across the body resistance when the IGBT is turned on. At times t2 to t3, T2 and T4 are turned off. Since there is stored energy in the inductor LS, diodes D2 and D4 freewheel under the action of the inductor LS, forming the transistor voltage drop waveform shown in the shaded part of UT2.D in Figure 1(b), and so on. The analysis shows that in order to detect the value of the transistor voltage drop when the IGBT is turned on, the transistor voltage drop at times t1 to t2 should be retained, while the value of the IGBT voltage drop detected at times t2 to t3 should be discarded, that is, the transistor voltage drop waveform shown in the shaded part of UT2.D in Figure 1(b) should be discarded. Because IGBTs have a relatively high switching frequency and significant switching noise, this should be adequately considered when designing the sampling circuit. Based on the above analysis, under normal conditions, the voltage drop Uce(sat) of the IGBT when it is turned on is relatively low, usually less than the rated value of Uce(sat) given in the device datasheet. However, if the H-bridge converter circuit malfunctions (such as a "shoot-through" phenomenon where both IGBTs on one side of the bridge arm are simultaneously turned on), a much larger voltage drop than normal will be generated across the collector-emitter terminals of the lower IGBT. If this voltage drop during the fault can be quickly detected, it can serve as a basis for protecting the IGBT, thus providing effective protection. 2. Design of the Short-Circuit Protection Circuit [align=left] From the analysis of the circuit shown in Figure 1, the schematic diagram of the IGBT short-circuit protection circuit can be obtained. IC4 and its peripheral components constitute the selection logic circuit, IC5 and its peripheral components constitute the filtering and amplification circuit, IC2 and its peripheral components constitute the threshold comparison circuit, and IC1 and its peripheral components constitute the holding circuit. Under normal circumstances, the outputs of IC2D, IC2C, and CD4011 connected to the cathodes of D1, D2, and D3 are all high, and the output state of IC1 remains unchanged. Suppose that for some reason, when a drive signal is sent to T2, the voltage drop across the lower transistor T2 in the left half of the H-bridge PWM converter circuit abnormally increases (let's say the level is "high"), i.e., the voltage at the UT2-d terminal abnormally increases. This high level UT2-d is applied to the cathode of D8 through R2; simultaneously, the high-level drive signal sent to T2 is also applied to the cathode of diode D5. For IC2C, its inverting input is high. If this level is greater than the threshold level of the non-inverting input, then IC2C outputs "low". This "low" level is applied to the R input of the RS flip-flop IC1 through D2, causing its output Q to flip and sending an IGBT fault alarm signal to the control system. If the low output of IC2D is caused by an abnormally high voltage drop across transistor T4 in the right half-bridge, this low level is applied to the R input of RS flip-flop IC1 via D5, causing its output Q to flip and sending an IGBT fault alarm signal to the control system. The filtering and amplification circuit, composed of IC5A, IC5C, and their peripheral components, preprocesses the voltage signal describing the IGBT voltage drop from the gating circuit and sends it to the adder composed of IC5B for calculation. If the adder's output level is greater than the threshold level determined by R22 and R32, the third input of RS flip-flop IC1 will be low, also sending an IGBT fault alarm signal to the control system. By changing the threshold level determined by R22 and R32, the physical meaning of this third alarm signal can be flexibly changed, allowing for flexible design of the protection circuit. Terminals T4-d and T2-d are connected to the collectors of T4 and T2 respectively, and T4-G and T2-G are connected to the drive signals of IGBT devices T4 and T2 respectively. Special attention should be paid to the fact that D8, D5, D9, and D4 must be fast recovery diodes during circuit design. 3. Simulation and Experimental Results When the PWM converter shown in Figure 1 operates in single-phase high-frequency rectification mode, the circuit is simulated using PSPICE simulation software. The simulated waveform is equivalent to the signal waveform observed at pin 7 of IC5B in the circuit. The simulation results show that the detection circuit can quickly and effectively detect the voltage drop across the lower transistor of the PWM converter when it is turned on. The waveform shown in Figure 3 is the relevant waveform detected during actual circuit operation. In the figure, channel 1 shows the given waveform of the single-phase high-frequency rectifier inductor current, and channel 2 shows the actual waveform detected at pin 7 of IC5B in the circuit. Comparing Figures 2 and 3, it can be concluded that this detection circuit can quickly and effectively detect the voltage drop across the IGBT when it is conducting, thus providing effective protection for the IGBT. Figure 4 shows the actual current flowing through the PFC inductor during IGBT overcurrent and the waveform of the protection circuit operation. Actual circuit operation results prove that the IGBT short-circuit protection circuit described in this paper can effectively protect the IGBT, is low-cost, and reliable. Practice has shown that this circuit has significant practical value, especially in applications with low DC bus voltage, and has broad application prospects. This circuit has been successfully applied in a certain type of 3KVA high-frequency inverter.