Hello,

We are currently conducting a research project in the field of motor control using your IGBT gate drivers (type 2SC0115T). The drivers are used to drive 650V-IGBT gates with a gate charge of 1200 nC in a 3-level configuration using a base switching frequency of 15 kHz. You can find the IGBT configuration in the attached image. Two driver cores (2SC0115T) are used to drive the four IGBTs. Driver core one drives IGBTs T1 and T2, driver core two drives IGBTs T3 and T4. During operation, we had multiple failures of these drivers. The defects manifest in form of one or both of their high side voltage potentials (-8.5V and +15V) failing, so that the gate either switches from 0V to +24V (-8.5V-potential failed) or being stuck at 0V (both potentials failed). Further investigation showed, that the IC on the driver (type 2114SC1) failed and shorted one or both potentials. This can be measured by measuring the resistance at the according supply output pins (60 ohms on a failed module, >600 ohms on a new module). Desoldering the mentioned IC removes the short at these pins. Measureing directly at the desoldered IC (e.g. between pins 13, 16) shows the low resistance of 60 ohms again.

We made sure to supply the drivers with the adequate supply voltage and to not exceed the voltage limits at the IGBT side.

Another (possibly related) problem we face is, that the occurance of a very short switching impulse of one driver channel (about 1 µs) results in the other channel to switch off its gate. This gate then stays switched off until the next switching impulse is given at the drivers input. According to the application manual no external capacitance is needed to drive a 1200 nC gate, so no additional capacitors are used. Furthermore, in none of these failure modes the failure is reported on the status output pin (SO), so the cutoffs are not caused by a recongnized fault condition.

Do you have any advice on how to avoid the failures of the driver outputs and to avoid one channel cutting out during short switching pulses? Is this a known problem with these drivers?

Kind regards.