Tool/software:
I've connected 2 drivers to IGBTs in a half-bridge configuration as you would to perform a double pulse test. I'm using the circuit to test the desat protection.
I've captured the waveforms for V_DESAT, V_CE (low side switch), V_GE (low side switch) and I_L (current through the inductor).
I've set up the desat protection to activate at 3V on the DESAT pin. Table on page 11 says that with a programmed value of 3V the threshold will be set between 2.7V and 3.3V. I've highlighted the point on the plot where V_DESAST crosses 3.3V. Follwing the table, on page 12 it says that the time between fault detection and action is 160ns + t_DESFLT which is combined a maximum of ~500ns. As seen from the plot, from the moment v_DESAT crosses 3.3V it takes another 100us for the pulse on the gate to go down, and that is because the maximum pulse length on the MCU generating the pulse is set to ~400us.
I've confirmed that the desat fault is signaled by the driver. The MCU reads back the status registers at the end of the pulse and it reads back DESAT and OV faults (OV is expected because the end of the pulse).
In the configuration registers I've confirmed that:
CFG4
- DESAT_EN pin is set to enable
- DESAT_DEGLITCH time is set to 316ns
CFG5
- TLTOFF_STO_EN is set to STO for SC, DESAT nad OC
- STO_CURR is set to 0.3A (also tried with 1.2A)
- DESAT_DSCH_EN is enabled
- DESAST_CHAR_CURR is 0.8mA
- DESATTH is set to 3V
CFG8
- GD_2LOFF_STO_EN is enabled
CFG10
- FS_STATE_DESAT_SCP is PULLED_LOW
I've also tried other combinations, using 2LTOFF and normal turn off, but I've saved the settings for this particular measurement.
Is my understanding of how the desast fault protection works wrong or should I be looking for a hardware/software fault?
Best regards,
Šimun