• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Subscribers 31 subscribers
  • Views 546 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

BOOSTXL-DRV8323RH: Burnt EVM gate drivers

biter terbi
Prodigy 160 points
Part Number: BOOSTXL-DRV8323RH
Other Parts Discussed in Thread: DRV8323, , BOOSTXL-DRV8323RS

Hello TI people.

Sadly I recently burnt an evaluation board while doing a test in a 250W BLDC motor. The test was designed to stress the system forcing it to do repeated heavy accelerations.

The board has the following damages.
    - GLC shorted to GND

    - GHB shorted to GND

    - GHA shorted to GND

    - The mosfet temperature was 95°C when the evm broke down.

    - No visible damage

The half bridges test all ok. So most probably the gate driver is burnt.

I'm driving the motor with the driver in 6xPWM mode, sending 6 SVMPWM with a dead time of 200ns.

The driver is set to 100ns dead-time and the default IDRIVE setting (120/240mA).

I'm supplying 40Apeak, 20A nominal current from the battery at 42V.

So my question is how come I burnt the evaluation board? What happens if the MCU deadtime is different than the one the driver is putting? Why did these gates get shorted?

Please help me try to figure out the issue.

Thank you in advance.

  • +1
    TI__Genius 14010 points

    Hi Biter,

    Thank you for posting to the Motor Drivers forum!

    There are a couple possible reasons for why the issue that you are seeing is occurring:

    • MOSFET Shoot-through aka Cross-Conduction
      • During normal motor commutation, phase current will conduct through either low-side (LS) or high-side (HS) MOSFET in a half-bridge configuration. If both HS and LS are conducting partially or fully at the same time, then a shoot-through condition occurs and creates a low-resistance high-current path between high-voltage and ground. This can potentially damage the PCB system and its components including the MOSFETS.
    •  Excessive Heating 

    More information as well as well as debugging tips can be found at the following FAQ: Quick Guide to Debugging Common Issues in BLDC motor drivers

    Best,

    ~Alicia

  • 0 in reply to Alicia Rosenberger
    Prodigy 160 points

    Hello Alicia.

    I accidentally marked this as resolved, but I wanted to continue this question.

    There shouldn't be any cross-conduction in the evaluation board, given we're using it with the default settings in pwmx6 mode and 100ns unmovable dead time. Also, no fets were damaged in the board, only the driver itself.

    Excessive heating shouldn't be a issue here, because the highest point of the board (the fets) were at 95°C.

    Regarding my second question:


    We're adding deadtime between INHx and INLx. Is this an issue? What would happen to the TDRIVE protection in this case?



    Thank you very much!


  • 0 in reply to biter terbi
    TI__Genius 16700 points

    Biter,

    I'd like to add a third possibility for the issue you are seeing; something I observed with multiple customers of mine. If you are indeed experimenting with heavy acceleration, there is a possibility that the deceleration is causing excessive back EMF, to the extent of which the bus voltage can violate the limit of the device. This could easily explain why the MOSFET's survived, while the motor driver didn't. To identify the problem, you need to closely monitor the bus voltage - small and repeated transients could cause a failure. To fix it, many options exist. The simplest is to increase the bulk capacitance at the bus. You could also add a resistor to dissipate the excess of energy, brake the motor by switching all low or high side MOSFET's, and finally add a comparator to switch a power resistor when the bus voltage crosses a pre-determined voltage level. 

  • 0 in reply to Lenio Cacula
    Prodigy 160 points

    Hi Lenio.

    I measured the battery voltage while hard accelerating and deaccelerating, there's hardly any noise in that signal in the EVM.

    The motor I'm using cannot regen brake.

    Thank you for your comment!

  • 0 in reply to biter terbi
    TI__Genius 14010 points

    Hi Biter,

    Regarding your question about adding deadtime between INHx and INLx, the TDRIVE component provides automatic deadtime insertion, though an additional digital deadtime (tDEAD) can be inserted. For hardware interfaced devices, this time is fixed at 100-ns as shown below:

    Keep in mind that if your deadtime is too long, the efficiency of the motor driver can decrease due to diode conduction losses. More detailed information on this can be found in section 2.2.1 MOSFET Handshaking of the following Application Report: Understanding Smart Gate Drive.

    Best,

    ~Alicia

  • 0 in reply to Alicia Rosenberger
    Prodigy 160 points

    Thank you Alicia,

    Can the TDRIVE component work with my external deadtime?

    Thanks

  • 0 in reply to biter terbi
    TI__Genius 14010 points

    Hi Biter,

    The DRV8323 provides automatic deadtime insertion with a fixed deadtime at 100-ns, for H/W interfaces, so an external deadtime should not be needed. Would you mind explaining the reason behind the external deadtime?

    Best,

    ~Alicia

  • 0 in reply to Alicia Rosenberger
    Prodigy 160 points

    Hi Alicia.

    It's just to have a little more control over that variable.

    Thank you

  • 0 in reply to biter terbi
    TI__Genius 14010 points

    Hi Biter,

    Do you know how much deadtime you are planning on adding?

    I am going to check with the team internally about whether the added deadtime will affect the TDRIVE and will aim to provide a response before the end of the week. 

    Best,

    ~Alicia

  • 0 in reply to Alicia Rosenberger
    Prodigy 160 points

    Hi Alicia,

    Around 100ns of aditional deadtime.

    Thank you, I'll be waiting for that response.

    Best regards

  • +1 in reply to biter terbi
    TI__Genius 14010 points

    Hi Biter,

    Adding 100ns of deadtime should be okay. I know you are currently the BOOSTXL-DRV8323RH, which is the hardware interface variant, but would you be willing to use the BOOSTXL-DRV8323RS which will give you more control over tDEAD?

    Best,

    ~Alicia