• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 1 answer
  • Subscribers 31 subscribers
  • Views 1547 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.

DRV8402 Problems

Chris B
Prodigy 120 points

I am having some troubles with the DRV8402.  I am using it to drive a very low inductance pancake motor, and have the switching frequency set at 200khz.  The first time, the part worked fairly well, until I started to place a load on the motor.  The current hit about 3 amps, and the DRV8402 failed, and pulls down the 24v supply.  The second DRV8402 failed after 10 seconds with no load on the motor, and exhibited the same problem.  The documentation indicates that external snubber diodes are not needed, but should they be added?

 

 

  • 0
    TI__Guru**** 189401 points

    Hi, Chris,

    Are you using our EVM or your PCB layout?

    -d2

  • 0
    TI__Mastermind 18510 points

    Chris, I will handle this question.  The failures are surprising, since DRV8402 is a 50V 10A device.  I don't recommend snubber diodes at this point.  Instead maybe we can analyze this with some information about the device in your circuit.  Right now I'm guessing that output overvoltage might be the problem, but we shall see.

    - Do you have any scope waveforms of output voltage and current and power supply voltage before the failure?  If you can send these, it might be very helpful in diagnosing the problem.  These can be BMP, GIF, JPG, etc.

    - Do you know the actual resistance and inductance of your motor?  That also could be helpful to know.

     

    Like Don Dapkus, I'd like to know if you're using EVMs or your own circuit and PCB layout.  If you're using your own circuit and PCB layout, it might help for me to look at it.

    - Can you send me PDFs of your schematic and PCB layout layers, if you're using your own?

     

    I'll look forward to hearing from you.

    Best regards,

    Steve Crump.

  • 0 in reply to SteveCrump
    Prodigy 120 points

    Hello Steve,

     

      I am using my own layout, which is somewhat patterned after the EVM.  The motor is spec'd at approx 1ohm resistance, and <.01mH inductance.  The DRV8402 is being fed a 200khz signal from the motor controller, and the DRV8402 was set for 8.2A current limit using a 39k resistor.  Unfortunately I was unable to capture any waveforms, since the second part failed pretty quick.  I started by using the 4uH inductors as shown int he datasheet, but have ordered some high inductors to try as well.  After examining the failed part, it appears that the B output is shorted, and VDD is shorted to ground in the part.  I have kept the power supply bypassing as close as possible to the chip, and the inductors fairly close as well.  I have ordered more of the DRV8402, so I can continue test with it tomorrow.  I was thinking of increasing the inductance, and also changing the VDD power.  Currently the VDD power is supplied by an isolated dc-dc converter, with its ground ties to the power ground.  The dc-dc converter supplies 132ma at 12v, which I thought should be sufficient.

  • 0 in reply to Chris B
    Prodigy 120 points

    Hello Steve,

     

    Attached is the schematic that I am using.

     

    2783.drv8402-schematic.pdf

  • 0 in reply to Chris B
    TI__Mastermind 18510 points

    Chris, it appears that your power supplies are fine, but I'm concerned about your power decoupling and output filter.

    - Power supply decoupling is significantly different from the EVM, which should be the guide.  You've put only 100nF on each GVDD pin, while the EVM uses 1uF.  What may be a greater problem is that the EVM includes 100nF right at each PVDD pin, and these are not in your design.  They are placed very close to the PVDD and PGND pins of the device in the EVM, and their purpose is to minimize switching transients at the PVDD pins.

              Without these you may have very large switching spikes that can damage the device.  I encourage you to duplicate the decoupling in the EVM.

    - Bulk decoupling in the EVM includes 1000uF at the PVDD input to the EVM plus 330uF at each full bridge PVDD pair, totaling 1660uF.  You use only 400uF.

              This may not be a problem - your power level seems to be well below what the EVM can do.  I suggest making sure you don't have significant PVDD ripple.

    - Your output inductors may seem to conform OK to the data sheet specification, but the "Minimum output inductance under short-circuit condition and parallel mode" in Recommended Operating Conditions on page 2 of the data sheet is a minimum value, not a rating.  This means that core saturation must not drive your inductance below 4uH even when output current reaches the OCP level, 8.2A.  If it does, the current may ramp far enough above the device rating too quickly for the device to protect itself.

              Please make sure your inductors can't saturate to less than 4uH at your OCP level of 8.2A.

    - Your output filter characteristic frequencies are 11.4kHz differential mode (eg, across the load) and 62kHz common mode (eg, each output to ground).  The 1ohm load damps the Q of the differential output to less than 4, which is still pretty high.  There is no signficant damping of the common mode resonance at 62kHz.  Each of these modes includes an impedance minimum at its characteristic frequency, and the impedance there is just the load Z for the differential mode and the parasitic resistances in the filter circuit for the common mode.  So the filter can draw heavy currents if it's driven there.

              I encourage you to use the filter configuration in the EVM.  This includes parallel 0.68uF in parallel bridge mode, and you could simply use 1.5uF on each side.

    I don't fully understand the cause of the failures you've seen, but I suspect it relates to these issues.  Let's see what happens when you make the changes I've suggested.

    Best regards,

    Steve.

  • 0 in reply to SteveCrump
    Prodigy 120 points

    Hi Steve,

     

    Thank you for the suggestions, I will try them and see if they cure the problems.

     

     

  • 0 in reply to SteveCrump
    Prodigy 120 points

    Hello Steve,

     

    I have made the changes that you suggested, and this seems to have eliminated the problems.  I have one last question though.  With the current limiting, is it possible to set the current to less than 8.2A?  Ideally I would like to set for 6A if possible.

     

     

  • 0 in reply to Chris B
    TI__Mastermind 18510 points

    Chris, I will check this and let you know.  The data sheet does not allow for it, but let me see what I might find.

    Best regards,

    Steve.

  • 0 in reply to SteveCrump
    TI__Mastermind 18510 points

    Chris, the Rocp pin of DRV8402 does not work with resistances outside the 27k to 39k range, so its Iocp cannot be reduced below 8.2A.  But maybe you can consider using DRV8432, which is pin compatible and has a current limit range from 11.6 to 1.4 A (43k for 6.3A).  There are slight differences in the operation of the chips, including some mode setting differences.  I think none of the differences will affect your application.

    Best regards,

    Steve.

  • 0 in reply to SteveCrump
    Prodigy 120 points

    I have another question about the DRV8402/8432.  When using the device in parallel mode, is the current setting for the total output or only one side?  If set with a 39K resistor, would the total output current be 6.9A or 13.8A?

     

    Thanks

     

    Chris

     

  • 0 in reply to Chris B
    TI__Expert 4320 points

    Chris

     

    the current would be 2x in parallel mode, e.i. 13.8A

     

    rgds,

     

    K. N. Madsen