Post Go back to editing

LTC7890 gets hot with no load

Category: Hardware
Product Number: LTC7890

Im currently working on a DC/DC board in a two rail configuration. The schematic is identical to the schematic in the datasheet for the 12V dual rail example except there is a trim poti to set the output voltage. Currently the LTC7890 gets hot while the EPC2361 GaN-FETs themself stay cool. When not enabled (RUN=0V) the circuit draws about 11mA (with INTVCC=0V) and when enabled it draws about 15mA (with INTVCC=5.5V) and the output voltage is at 2.7V. Even in Off state the high side FETs get 0.8V on their gate while the Low Side Fet gets 0V on its gate. In RUN mode the high side fet has 2.7V on its gate.
When loaded the output drops to 800mV and the high side fet now has 800mV on its gate. The gate has no short circuit to drain or source.

Do you have any ideas what could be wrong?

  • Thank you for sharing the schematic. A couple things stand out that I would like to check first before assuming anything.

     

    Could you please capture the high-side VGS and SW node waveforms? I'd like to verify the gate drive operation.
    Also, when RUN is held low, is the LTC7890 section the only circuitry connected to the input supply, or are there other circuits powered from Vin as well?

     

    We can start debugging with these scope shots and go on from there.

  • This is the schematic of the whole circuit. There is just one other DC/DC Buck Converter for a 5V supply but that was not soldered during the tests so the 11mA is from the LTC7890.

    While testing the switching Node and VGS on the FETs i noticed that VGS is always slightly negative and the FETs are not switching. Whats also remarkable is that the bootstrap capacitors have Vsw on one side and while loading it with the probe the voltage decreases to zero on the bootstrap node.

    Is it possible that a soldering error damaged the IC? Ive noticed that on the first run some ground nodes maybe had a bad connection.

  • Thanks for the additional information. 
    A damaged LTC7890 is possible, especially if there were ground connection issues at power-up, but I would not conclude that yet.

     

    Since the FETs are not switching, could you please measure:
    INTVCC
    DRVCC
    BOOST1-to-SW1
    BOOST2-to-SW2

     

    Let me know if you have any questions or anything else comes up.

Before You Switch


Switching languages will make ADI Explorer unavailable. Resume your session by switching back to English and reopening ADI Explorer.