• State
  • Locked Locked
  • Replies 7 replies
  • Subscribers 29 subscribers
  • Views 132 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.

DRV8711: Pre-driver faults

Robert Karpinski
Prodigy 20 points
Part Number: DRV8711
Other Parts Discussed in Thread: , , CSD18531Q5A

Tool/software:

Hi,

My company has ~800 stepper driver boards built with DRV8711s & CSD88537NDs.
They have worked pretty well for us at 24V and 1-2A with one pre-driver fault every 100 or so activations.

Recently we have been wanting to speed up out stepper motors so have been increasing the voltage to 28V and 48V as well as increasing the current to 4A.
This then leads to a massive increase in the number of pre-driver faults to the point the drives cannot be used at all.

Here's the schematic:

DRV8711 and CSD88537ND stepper power stage

We've mimicked the BOOST-DRV8711 as much as we can.
We've added the suggested 75Ω resistors to the low side MOSFET gates (and the firmware sets the deadtime to 850ns). I've also added 0Ω resistors to the high side MOSFET gates in case I need to add anything there in future.
I've also added output ferrites (BLM18KG300TN1D) and capacitors (470pF 100V) to minimise noise as we do have up to 30cm of wire to the motors and the end device needs to be CE marked.

The register settings are:
Register 0 = 0x0C23
Register 1 = 0x012E
Register 2 = 0x001E
Register 3 = 0x0028
Register 4 = 0x0419
Register 5 = 0x0040
Register 6 = 0x0A59
Register 7 = 0x00C0
This configuration causes pre-driver faults every hour or so, but when we increase the torque register to 4A it happens within 5 minutes.
I've tried many combinations of drive current and drive duration but they've not helped.


Are there are any other suggestions for running this at high voltage & current to minimise pre-driver faults?
Is it worth adding some capacitance on the gate pins to filter noise?
Should I add a filter to the current sense resistors?
Are there any registers that may also need changing?
Thanks.

  • 0
    TI__Mastermind 31115 points

    Hi Robert,

    You can try the following:

    1. Increase the LS-FET series gate resistors to up to 120 ohms each.

    2. Register 6 = 0xFF9. This will increase the IDRIVEP and N as well as extend TDRIVEP and N. If this solves the xPDF faults you can then optimize for trying the next lower TDRIVEx settings. 

    Regards, Murugavel 

  • 0 in reply to Murugavel4637
    Prodigy 20 points

    I've not been able to solder the 120Ω gate resistors yet but I have tried setting register 6 to 0xFF9 and that has not helped. I still get pre-driver faults. Is there any way of disabling the pre-driver fault detector? 

  • 0
    Prodigy 20 points

    I managed to get a scope capture of the A1LS and A2LS as the FAULTn trips:

    Channel 1 is FAULTn
    Channels 3 & 4 are A1LS and A2LS
    The STATUS register was 0x00C8 which means I have an APDF fault.

    My DRIVE register is set to 0x0A59 which means I have TDRIVEN & TDRIVEP set to 500ns and IDRIVEN is 300mA and IDRIVEP is 150mA.
    According to the scope capture it's appearing to drive one of the A*LS for ~2000ns which is 4 times the register setting.

  • 0 in reply to Robert Karpinski
    Prodigy 20 points

    Here's another scope capture:

    This time I had doubled the TORQUE current as well as setting IDRIVEN to 400mA, IDRIVEP to 200mA and TDRIVEN & TDRIVEP to 2us.
    Again the STATUS register was 0x0A59 which means I have an APDF fault.

    As before these look alright to me. You see a steep rise, a wiggle around 5V as you overcome the Miller effect and then a final rise to 10V. So I'm a bit bemused as to why I get pre-driver faults.

  • 0 in reply to Robert Karpinski
    TI__Mastermind 31115 points

    Hi Robert,

    Robert Karpinski said:
    Is there any way of disabling the pre-driver fault detector? 

    Unfortunately this is not possible with this device.

    Robert Karpinski said:
    As before these look alright to me. You see a steep rise, a wiggle around 5V as you overcome the Miller effect and then a final rise to 10V. So I'm a bit bemused as to why I get pre-driver faults.

    The reason for suggesting a 120Ω gate resistors for LSFET was to mitigate noise leading to a false xPDF detect. It seems like you were running into a false detect issue. Did you observe this with multiple units? Can you try with another unit of the DRV8711 to see if it is a specific unit related issue? 

    Sorry you're facing this issue. Are these output FETs same as the ones in the TI DRV8711EVM? 

    Regards, Murugavel 

  • 0 in reply to Murugavel4637
    Prodigy 20 points

    Being able to disable the pre-driver fault detector when it seems to trigger erroneously would be pretty useful for us.

    I've ordered some 120Ω resistors and will try to fit them once they arrive, this isn't my preferred solution as I have ~800 boards out in the wild and I wouldn't be able to get them all modified.

    I do have another board physically right next to this one so it should be very easy to just swap them over. However I already know that this issue affects all my boards and that it's very likely to still be a false pre-driver fault detect.

    The output FETs are CSD88537NDs like the BOOST-DRV8711. The DRV8711EVM uses the CSD18531Q5A instead.

  • 0 in reply to Robert Karpinski
    TI__Mastermind 31115 points

    Hi Robert,

    We understand. Unfortunately this device is hard wired. I hope the 120 ohms fixes the issue for you.

    Regards, Murugavel