• State
  • Locked Locked
  • Replies 9 replies
  • Subscribers 29 subscribers
  • Views 248 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.

Original question:

DRV8711: DRV8711 FET driver damage

DRV8711: Design to respect Absolute Maximum ratings

Allen Blaylock
Intellectual 310 points
Part Number: DRV8711
Other Parts Discussed in Thread: , CSD18531Q5A, CSD88537ND,

Hello,
I am trying to figure out how to design a circuit with the DRV8711 that will respect the absolute maximum ratings of the DRV8711.

If I focus on the ratings for the "Phase node pin voltage (AOUT1, AOUT2, BOUT1, BOUT2)" I see the rating for the pins are between –0.6 and VM + 0.6 V. My concern is when the system is in a high impedance state where all of the bridges are disabled, for instance during a dead time cycle, the body diodes of the MOSFETs will conduct if there is current flowing through an inductance. When the current is being sourced from a low side fet through the body diode as well as the low side sense resistor. It seems very difficult to stay within the -0.6 rating on these pins given the forward voltage of the body diode on top of the sense resistor voltage drop.


To work an example lets use the DRV8711EVM:
Assume a modest operating current of 1A.

From the CSD18531Q5A datasheet we see the forward diode voltage is:

0.68V

Next we calculate the voltage drop over the sense resistor:
0.033 Ohms * 1.0 A = 0.033V

Now summing the voltage over both the sense resistor and the body diode we have:
0.68 V + 0.033 V = 0.713 V.

Ooops, we have violated the maximum ratings of the DRV8711 with the body diode forward voltage alone.

Are the absolute ratings correct for those pins? If so, how does TI recommend we stay within the absolute ratings when the vast majority of body diodes have a forward voltage of 0.6 V at 1A alone?

Thanks,
Allen

  • 0
    TI__Guru 56650 points

    Allen,

    Yes. "When the current is being sourced from a low side fet through the body diode as well as the low side sense resistor. It seems very difficult to stay within the -0.6 rating on these pins given the forward voltage of the body diode on top of the sense resistor voltage drop."

    I would think a paralleling schottky diode on low side FET could help in this situation.

  • 0 in reply to Wang5577
    Intellectual 310 points

    Wang,

    This appears to an oversight, neither reference design implement parallel schottky diodes and I would expect almost all applications to be using at least half an amp which exceeds 0.6V forward voltage on the CSD18531Q5A diodes chosen for the DRV8711EVM. The CSD88537ND FETs on the BOOST-DRV8711 reach 0.6V at an even lower current of 0.006A. None of the other designs I have seen referenced on this forum or available on the web utilize schottky diodes.

    Is this something that could be clarified with the DRV8711 design team?

    Thanks,

    Allen

  • 0 in reply to Allen Blaylock
    TI__Guru 56650 points

    Allen,

    DRV8711 has been released since 2013. At that time, customers accepted the application and -0.6V spec without the low side FET paralleling Schottky diode.

    Since the minimum ISENSEx pin voltage is -0.7V, the -0.6V phase node pin voltage doesn't mean too much. It means the IC should be safe when the phase node pin voltage is one body diode voltage below the ISENSEx.

  • 0 in reply to Wang5577
    Intellectual 310 points

    So what you are saying is the real absolute rating of the AOUT/BOUT pins should be -1.3V as the sum of -0.6 and -0.7?

     does this make sense?  did you end up implementing schottky diodes to resolve your issue?

  • 0 in reply to Allen Blaylock
    TI__Guru 56650 points

    Allen,

    -1.3V is not right. One body diode voltage below the ISENSEx pin voltage is correct.

    For example: if ISENSEx pin is only -0.2V, I don't want the phase node voltage running into -1.3V. Just one body diode voltage below the ISENSEx pin voltage.

  • 0 in reply to Wang5577
    Intellectual 310 points

    To be clear if the ISENSEx pin is at -0.6V as allowed by the maximum ratings then one additional body diode would make the voltage -1.2V at the phase node right?

  • 0 in reply to Wang5577
    Intellectual 310 points

    Wang,

    To go back to the previous suggestion of paralleling a schottky diode with the low side FET what is the better solution, connecting the anode to the source side of the sense resistor or to the ground side of the sense resistor? With the diode anode connected to ground  the AOUTx/BOUTx pins shouldn't ever see a voltage greater than the forward voltage of the schottky. If the anode is connected to the source side of the sense resistor the AOUTx/BOUTx pins will see the voltage across the sense resistor + the schottky diode forward voltage which may technically violate the absolute maximum ratings.

  • 0 in reply to Allen Blaylock
    TI__Guru 56650 points

    Allen,

    "With the diode anode connected to ground  the AOUTx/BOUTx pins shouldn't ever see a voltage greater than the forward voltage of the schottky"

    Yes. So, I would like to put the Schottky diode in parallel with the power FET.

    "If the anode is connected to the source side of the sense resistor the AOUTx/BOUTx pins will see the voltage across the sense resistor + the schottky diode forward voltage which may technically violate the absolute maximum ratings."

    To me, "Since the minimum ISENSEx pin voltage is -0.7V, the -0.6V phase node pin voltage doesn't mean too much. It means the IC should be safe when the phase node pin voltage is one body diode voltage below the ISENSEx." If you have to make the AOUTx/BOUTx above -0.7V, you may add more clamp circuit (another diode). But, I don't think it is realistic.

    So, I would put a Schottky in parallel with the power FET. And put two small package resistor for the current sensing, instead of one big package sense resistor. In such way, even one component has a bad soldering, the winding current still has a current path and won't generate a big negative voltage spike on ISENSEx pin and switching node.