This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LMG3525R030-Q1: LDO5V fails when operating in a totem-pole PFC

Part Number: LMG3525R030-Q1

While testing the LMG3525R030-Q1 in a totem-pole PFC at low voltages (around 36V at DC side), the low side GaN FET fails. It seems that the LDO5V fails, the resistance between LDO5V and GND is only a few ohms and the GaN FET is not able to deliver 5V when powered.

We started testing at an even lower voltage (8V) to reduce the risk of device failure and we found a voltage spike at the switch node after a zero crossing in the inductor current indicated in the figure below. This spike occurred every line cycle. 

C1 is the gate signal of the low side GaN FET measured at the microcontroller

C2 is the voltage at the switch node relative from DC-, measured with a differential probe 

C3 is the gate signal of the high side GaN FET, measured at the microcontroller

C4 is the inductor current

We suspect this has something to do with the ideal diode mode, but we're not exactly sure how this could happen. Could this be the cause of the device's failure?

  • While testing today, we saw a similar spike in the switch node voltage at the beginning of each line cycle. The FLT trigger was too short for the microcontroller to act, causing the low side GaN FET to fail eventually. Again it seems like the LDO5V fails, when looking at the FLT signal, this is not a surprise. In these figures:

    C1 (yellow) is the gate signal of the low side GaN FET, measured at the microcontroller

    C2 (red) is the FLT signal from the low side GaN FET, measured at the microcontroller

    C3 (blue) is the voltage at the switch node relative from DC-, measured with a differential probe 

    C4 (green) is the inductor current (positive current is current flowing into the AC side of the circuit) (The first figure shows the average inductor current)

    Do you maybe have an idea how we can prevent this from happening? I will add a detailed description of the circuit and some more measurements tomorrow.

  • Hi Thijs,

    This seems to be an issue with the LDO5V and not IDM. Could you please try de-soldering the LDO5V and adding a resistor from LDO5V to the digital isolator LDO5V?

    If this does not work, try a slower slew rate.

    Thanks,

    Noah

  • Dear Noah, 

    Thank you for your quick reply. We added a 180k resistance between RDRV and GND to lower the slew rate and a 47E resistance between the LDO5V and VDD1 of the digital isolator. Furthermore, we replaced the 12V isolated DC-DC converter with an alternative with a lower isolation capacitance (2.5pF instead of 40pF). 

    Even though these changes improved the situation, they did not solve the problem. The totem-pole rectifier was able to operate without glitches at 14.1VAC/25VDC, but there were again some glitches with operation at 28.3VAC/50VDC. This led to failure of the low side FET after a few line cycles. 

    Did we use the correct resistor values for the changes? Is there something we are missing? If needed, I can provide some more information about the control scheme and the hardware design. 

  • Hi Thijs,

    Thank you for the update with your issue. Let me review this with the team and I will get back to you soon.

    Thanks,

    Noah

  • Hi Thijs,

    I would recommend simplifying debugging by using the LMG3522R030, in case it may be an issuing concerning the IDM version of this device. I would also make sure to read through IDM section (page 28) in the datasheet for the design precautions laid out there. Again, I would use the LMG3522R030 to reduce complexity until you are able to get stable power up.

    https://www.ti.com/lit/ds/symlink/lmg3425r030.pdf?ts=1649695716669&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FLMG3425R030

    Thanks,

    Noah

  • Dear Noah,

    Thank you again for your reply. The datasheet of the LMG3425R030 makes it clearer how the device behaves in IDM. 

    I see that the LMG3522R030 is not in stock anymore on the website. Would you be able to provide us with some samples for development? We would need around 10 pieces.

  • Hi Thijs,

    Could you please send an email to gan_marketing@list.ti.com with your sample request?

    Thanks,

    Noah