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.

DRV103H burning out

Other Parts Discussed in Thread: DRV103

I recently had a DRV103H burn up ("let the smoke out") around the pin 4 quadrant of the chip.  My schematic is similar to the application circuit shown in the datasheet.  The event occurred when the chip was being powered, but the input level was low and there was no load at the output.  I have attached a screenshot of my circuit.  The 24V is being supplied by an isolated DC to DC converter.  Any thoughts?

Also, I gathered from the datasheet that the power pad is to be directly soldered to the board, but to some thermal plane and not to ground.  Is this correct? ("though the metal pad of the PowerPADTM SO-8 (H) package is electrically connected to ground (pin 4), no current should flow in this pad. Do NOT use the exposed metal pad as a power ground connection or erratic operation will result. For lowest overall thermal resistance, it is best to solder the PowerPADTM directly to a circuit board")

Thanks.

  • Hello TIm Sweet,

    Your schematic seems fine except for the connection of pin 5 to the ground labeled “VALVE_RTN1” which would render the circuit not very useful. I assume the schematic is only for illustration purposes and that your application will actually have a load connected between VALVE24V and pin 5 (with no direct connection to VALVE_RTN1).

    In your board, what is the voltage difference between PUMP_RTN and VALVE_RTN1?

    Is this failure repeatable? If so, could you please acquire and post the voltage on pins 5, 6, and 8 with respect to pin 4 of the DRV103? I will need to see and compare the waveforms for a board that does not fail with those for a board that fails for the indicated pins.

    The datasheet note you quoted in your post is indeed correct.

    Best regards,

    Jose

  • You are correct, the load will be connected between VALVE24V and pin 5, sorry for the confusion.  VALVE_RTN1 is not really a ground, only a return path to pin 5 for this load only - it was just easier to label it this way so I could have a connector on another sheet.  The voltage difference between PUMP_RTN and VALVE_RTN would be then whatever the voltage drop across the internal DMOS transistor is.

    I have not seen this failure again.  The chip worked fine for several weeks, but then one day just popped during operation.  I have since pulled off the burned chip and replaced it with a new DRV103H and it seems to work fine.  Could be ESD related?

    I have included the requested voltages on pins 5, 6 and 8 WRT pin 4 when the anomaly occurred (Pin 8 was low, meaning that the chip was disabled.  All figures taken during this time have "disabled").  I also included the same voltages when the chip is enabled (Pin 8 is high.  All figures taken during this time have"enabled").  All of these were taken with no load - I have not loaded the chip yet on the rev of the board, but have loaded it on past revs with no issues.

    Note:  the 100mV signal on pin 8 in the disabled state is because of the voltage divider of the 6.98k and the transistor resistance in the opto-isolator (ILQ615-4X009).

    Thanks,

    Tim

  • Forgot the attachments.

    Voltages.zip
  • Hello Tim,

    Thank you for the images. I do not see any abnormal behavior that could explain the failure you observed.

    From your description (temperature high enough for the package to melt) I doubt the failure was caused by an ESD event; such level of damage is more likely to be cause by a serious electrical over-stress to the part. A significant overvoltage applied to pins 5, 6 or 8 was the first potential cause that comes to mind but given the scope captures you sent I cannot identify any evidence of problem.

    When you mention

    TIm Sweet
    The event occurred when the chip was being powered, but the input level was low and there was no load at the output.

    Have you monitored the 24 V supply during startup to see if there are any overvoltages in the supply bus? Sudden voltage spikes in the 24 V during startup could potentially have damaged the part by exceeding the specified voltage range on pins 6 or 8.

    Aside from that the only idea that comes to mind would be to ty to reproduce the failure and capture the voltages in the nodes mentioned above for an event where the DRV103 fails.

    Best regards,

    Jose

  • I just monitored the 24V rail during startup.  Scoped it for more than 10 power cycles, watching the voltage never creep above 25.6V at startup (which should be fine for the DRV103).  

    I will scope the 24V rail and run for an extended period of time hoping to capture something if the problem occurs again.  I'll post back after a couple of days to keep you updated.

    Thanks.