Double Pulse testing with FlexGui and RDGD31603PSMKEVM Eval Board (GD3160)

[AdrianJoerressen]

Hello,

I'm trying to use the FlexGui for a Double Pulse test setup for SiC MOSFET half bridges. If testing HS or LS simply using one set of pulses it is working fine. If I enable the "Complementary PWM", so that also the other MOSFET of the half bridge is switched complementary to the DUT a "DTFLT" error is triggered. According to data sheet this is triggered if the PWM deadtime is violated. Is this a violation because HS and LS are triggered without enough deadtime in between (short circuit protection)? If yes, how can I avoid this, since only the main pulse timings (break >10 µs) for the DUT can be configured and the complementary seem to be fixed based on register for deadtime (which I tried configure to minimum).

GD3160 

[AdrianJoerressen]

Hello Tomas,

thanks for the fast replay! In the meantime I did some more investigation on the topic. To trigger a measurement on the oscilloscope I'm using the rising edge of the DUT (high side) gate voltage at 7V (VCCREG default 15V). Like mentioned with the single setup it always works. With the complementary setup, there is no trigger coming. Meaning the gate of the DUT / HS is never getting any voltage (above 7V...). This looks more like the  "If PWMALT goes high while PWM is already high, then the gate turns off immediately, without inserting tmin_DT." Currently I can't explain this, since at the beginning of the first pulse LS should be high and HS low, thus HS should go high and maybe LS turned off immediately - or deadtime of HS enforced, but gate still turned on. Pulse lenght I'm simply using 30-30-30 µs to avoid any problems with overlap.

During investigation I found another issue: Seems I cannot write on registers. I tried is with several via Daisy Chain Config and other tabs. After writing I did a readout and the old value was still in. Is there anything I need to consider?

[TomasVaverka]

Hello Adrian,

Regarding the complementary PWM question, it would be helpful to have a plot where you show the complete DP pattern with the following signals captured:

- PWML & PWMH

- GateL & GateH

Regarding the registers not being written, did you check that you are in configuration mode when you do these changes (i.e. the CONFIG_EN bit in the MODE2 register is set)?

BR, Tomas

[TomasVaverka · Accepted solution]

Hello Adrian,

The "Complementary PWM" option in FlexGUI doesn't generate any deadtime on the PWML/H pins.

Therefore, the deadtime protection feature of the GD3160 is activating, which has 2 effects:

- enforce a minimum deadtime, as programmed in the DEADT field of the CONFIG5 register, avoiding any short circuit condition

- generate a DTFLT event, to warn you that the MCU didn't provide the required minimum deadtime.

To use complementary PWM mode with FlexGUI, you can therefore use the DEADT field to program the deadtime you actually want to be used (range is 100ns to 6.5µs).

If you do not want the DTFLT appearing, you can mask it with the DTFLTM bit in the MSK2 register, the fault bit will not appear again but the minimum time will still be enforced.

See section 11.2.5 of the datasheet as well.

Alternatively, you can use an external signal generator for the PWML/H signals, providing external deadtime, if you want finer control than the DEADT selections offered by the GD.

BR, Tomas

[AdrianJoerressen]

Hello,

thanks for the hint. I did this measurement and saw, that the FRDM-KL25Z MCU was generating a wrong signal - the complemntary side was always high without any change.

I made the recommended firmware update from 0.8.3 to 0.8.8 and the proper signals were generated and the DUT also behaved as expected.

Unfortunately after some trial double pulse tests I got a short circuit on the board (supply current went into the limit of 500 mA and voltage dropped to 4V). Maybe I have to create another threat for this issue if I can't find the problem.