Module Failures Look Like System Problems

Most IGBT module failures in the field trace back to the gate circuit, the thermal path or the layout, yet the symptoms appear at the system level as overheating, EMI or a random shutdown. The temptation is to replace the module and restart, but the module is usually the victim rather than the cause. This article presents a systematic diagnosis for the most common problems in Semikron SEMITRANS and SEMiX designs, from gate ringing to overvoltage and overheating.

Problem 1: Gate Ringing

Ringing on the gate waveform at tens to hundreds of megahertz is parasitic oscillation between the gate capacitance and the stray inductance of the gate loop. The first fix is geometric: shorten the loop from driver to gate resistor to gate and back to emitter. Then increase the gate resistance in small steps until the ringing falls below roughly twenty percent of the drive voltage, or add a ferrite bead close to the package to damp the high-frequency oscillation without slowing the transition much. Mount a matched Semikron driver close to the module so the loop starts small.

Problem 2: False Turn-On in a Half Bridge

In a half bridge, the fast dv/dt of one switch can couple through the Miller capacitance of the other and lift its gate above threshold, causing cross-conduction and a possible failure. The classic fixes are a negative gate supply, a lower-impedance pull-down or a lower off-transition gate resistance. Measure the off-state gate voltage during a switching event: if the spike crosses the threshold, apply the negative rail. This problem is common when the gate loop is long, so the first fix is often the layout.

Problem 3: Turn-Off Overvoltage

Collector overvoltage at turn-off comes from the energy stored in the commutation-loop inductance. The fix is to reduce the loop, then tune the turn-off gate resistor or add a clamp. Always measure at the module terminals, because the bus measurement hides the spike the device actually sees. A module with low internal inductance, such as SEMiX or SEMITRANS 3+, reduces the overshoot, but the external busbar still dominates, so a laminated busbar and a close DC-link capacitor are the real levers.

Problem 4: Overheating and Thermal Failure

If the module overheats, separate a device problem from a system problem. Measure the gate waveform and the switching losses first; if they are normal, check the thermal path: the interface material, the mounting torque and the heatsink. A poor interface or a loose mount raises the case-to-heatsink resistance and derates the design severely. Estimate the losses at the actual operating point and confirm the junction temperature at the worst-case ambient, because a design that passes at light load can still fail at full load. The BeiLuo FAE team supports this diagnosis and can validate a fix on the bench.

A Systematic Procedure

Work from the obvious to the subtle: confirm the load and the wiring, then inspect the gate waveform, then the protection thresholds, then the thermal path. Measure at the module terminals, not the bus, and log the operating point for every event. Because a module failure usually has a system cause, finding the cause before replacing the module is what prevents a repeat failure and protects the next part.