• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 712 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LMG3410: Inverting buck-boost converter

Colin Setzke
Prodigy 40 points

Part Number: LMG3410

Hello,

I have two of the LMG3410 implemented in an inverting buck-boost converter. One of them works as a switch. The other one replaces the diode so that the source is connected to the load and the output capacitor. So a part of the time (depending on the Duty-Cycle) the transistor has a voltage drop of Ui+Uo over it and the other time it conducts current from source to drain so it should work as given in figure 3 of the datasheet. 

The maximum input voltage of the converter should be 150V.  When I test the converter with an input of for example 30 V and in boost mode with Uo=1,5*Ui everything works fine. That means when the Transistor conducts current the Voltage drop is about -7V. When I increase the Input Voltage also everything works as it should up to when the voltage drop over the transistor which works as a diode reaches about 200V (80V input, 120V output). If I increase the input voltage now the voltage drop over the transistor changes from about -7V proportional towards 0V and even higher to a positive value. Just before that happens the lower falling edge of Ugs gets rounded and the curve of Uds gets crooked somehow and then the whole curve gets raised. After a short time the LMG3410 shuts off.

To increase or lower the current through the device changes nothing. 

The other Transistor works as it should. The external wiring of both tansistors is identical. For the layout I used an example layout of TI. 

I tried to control the diode-transistor with a PWM and to ground the input pin, but nothing changes, in both cases Uds=-7V and the error occurs. I thougt if I drive the device with a PWM the Uds should be nearly 0V even in 3. Quadrant, but it isn't. Isn't the device capable to be driven in 3. Quadrant easily?

So I hope you understand my problem and can help me to fix it.

Best regards,

Colin Setzke

  • 0
    Prodigy 40 points
    I forgot to mention that changing the frequency of the PWM also changes nothing.
  • 0
    TI__Prodigy 80 points
    Hi Colin, thanks for the interest of our GaN products. Based on your explanation, the questions and possible suggestions that we have:
    1) Firstly, we confirm that LMG3410 is capable of operating in the 3rd quadrant. It shall be helpful if you could inform us how you drive the GaN. GaN can operate in 3rd quadrant (reverse conduction) similar to a “body diode” but with a high voltage drop. Therefore we recommend synchronous switching operation at all times. If you already are doing synchronous switching, what is the dead time? This is actually the 3rd quadrant conduction time which is lossy and potentially cause thermal shutdown depending on your load condition. Assuming 1A of load and 50% 3rd quadrant conduction, you could have 7W loss from the GaN device, so it is suggested to measure the temperature rise during operation. We suspect a thermal shutdown that interrupts the operation. We don’t think that this is 200V issue but I guess you run your converter in open loop and as you increase the input voltage, output voltage rises , drawing more current and increasing 3rd quadrant current.
    2) Besides, could you provide waveforms of Vds shifting (-7V to 0V) and FAULT pin output of LMG3410 in the same scope screen? It is helpful to observe the timing sequence between the fault signal (if any) synced with the Vds changing. This will be useful debugging information.
    3) Are you using our HB EVM board for this setup or have you build your own pcb? If you build your own pcb, are the FAULT outputs of these two GaN FETs logically ORed?

    Hope above addresses your question.