• State TI Thinks Resolved
  • Locked Locked
  • Replies 1 reply
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 1519 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.

LM5113-Q1: Gate Drive Compensation for Large Gate loop inductance/resonance

Joel Barker
Prodigy 100 points
Part Number: LM5113-Q1
Other Parts Discussed in Thread: LM5113

We are currently using the LM5113 to drive parallel half-bridge GaN eHEMT FETs which are located inside of a separate module. The  issue is that the large gate loop inductance (which isn't possible to lower) is causing a lot of oscillation and false turn-ons of the devices. As a result, our switching losses are much too high for our high-temperature application. (Circuit shown below). 

The schematic is below. Damping and resonant compensation (circled in blue on the schematic) has been applied and adjusted, but only has minimal effect. The graph below shows the issue in the response of the gate drive output relative to the actual Vgs voltage. Are there inherent ways to improve the LM5113 to compensate for this gate loop inductance/resonance, or are there any solutions you could recommend? 

  • 0
    TI__Expert 4420 points
    Hi Joel,
    the best way to avoid ringing is to reduce the loop inductance, which you mention is not possible for your application.

    The next best thing is to split the damping resistors on each of the four outputs into two: one close to the IC, one close to the FET, this should create a RC filter close to the pins that could somewhat help the situation.

    Increasing the value of R will help dampen the ringing further, but you will also reduce the gate current and slow down the transitions.

    Also if your noise has common mode quality to it, adding a signal common mode choke will help mitigate it.

    The compensation network approach although theoretically sound, is harder to implement in practice because of all the parasitics and the components variation. This would have to be fine-tuned for every board.

    I hope this addresses your question, but please let me know if you require further clarification.

    Best regards,
    Alberto