• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 30 subscribers
  • Views 486 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.

DRV8305: nFAULT asserted with no faults or warning in registers, increased current draw

Freddy Gauss
Intellectual 300 points
Part Number: DRV8305

This is my first time using the DRV8305 and so far it's been a great part to work with. During development I found that two parts had failed. The parts showed increased current draw and stopped responding to inputs. At the time I wasn't able to thoroughly investigate the failures as I should have. Then for a long time I saw no failures but a couple just popped up again. This time I was able to investigate the failures further and found the following.

The failure presents as before, the parts have an increased current draw and the nFAULT line is asserted. The increased current draw suggests that there is physical hardware damage. As expected, the high-side gate driver charge pump is not running and PWM inputs to the chip are ignored. When reading the warning and fault registers (addresses 0x1, 0x2, 0x3, 0x4) all reads return 0. Reading out the rest of the configuration registers reflects the configuration with which the part is initialized, which are default values for the most part. Reading the warning and fault registers does not clear the fault condition, neither does toggling the EN_GATE line, the warning and fault registers read 0 as before.

As I was unable to find any hints to what might have caused the fault and potential damage to the IC I started reviewing the datasheet to see if I made a mistake in the design that could have caused the problem. So far the only thing I've found is that the datasheet mentions that DVDD and AVDD should be bypassed with a 1uF capacitor, I've got that. Additionally it mentions that DVDD and AVDD should _not_ be supplied externally. It appears that I had missed this note (pg. 25, section 7.3.6) and added a zero-ohm resistor to power DVDD and AVDD from 3.3V and 5.0V, respectively. So far this is the only mistake I've found.

Could having supplied DVDD and AVDD externally (with 3.3V and 5V, respective) have cause such a failure? I understand that it might be difficult to answer with any reasonable amount of certainty, I was just wondering if anyone had any insight. Currently my best option is to remove the 0-ohm resistors which allow for DVDD and AVDD to be supplied externally and run some long term tests and see if the parts survive. This is a difficult way to prove (or even suggest) that the root cause has been found and solved but it's the best I've got at the moment.

  • 0
    TI__Guru 65250 points
    Freddy,

    AVDD and DVDD are internal supplies that are self supplied and only need external decoupling.

    Powering these supplies internally can cause issue as the internal power sequencing and any protection feature that relies on these could malfunction.

    Please remove the 0 ohm resistors and see if it resolves the issues.

    Regards,

    -Adam
  • 0 in reply to Adam Sidelsky
    Intellectual 300 points

    I've removed the 0 ohm resistors on two failed ICs and they do behave differently.

    With the 0 ohm resistors the parts that had asserted the nFAULT line read 0's from all of the warning and fault registers (addr: 0x1, 0x2, 0x3, 0x4). The configuration registers read back exactly as configured on both chips.

    Without the 0 ohm resistors (DVDD and AVDD powered internally):
    On chip 1, the DVDD voltage reads 2.4V and AVDD is 0.056V. On initialization all of the registers (warning, fault, control) read back 0x0000. The initialization firmware reports an error since the registers don't read back what was programmed but the part doesn't actually indicate a fault in the registers. The nFAULT line is asserted continually.
    On chip 2, the DVDD is about 0.56V and AVDD is 0V. This chip reads back 0x07FF from all registers upon initialization (the MISO line has a pull-up resistor and the lower 11 bits are masked in the firmware). It looks like the chip is not driving the MISO line. The nFAULT line is _not_ asserted on this chip. This makes some sense if the part is not at all powered up (DVDD = 0V).

    It appears as though the internal power supplies were damaged, possibly because they were powered externally under operation. Is this realistic damage if powered externally? Would frequent power cycles make this problem more likely to appear? In other words, is supplying external power to DVDD and AVDD a problem because the IC doesn't have control over its own power sequencing?

    On future designs I will definitely be removing the 0-ohm resistors to DVDD and AVDD, I just wanted to find out the root cause of these particular failures.

  • 0 in reply to Freddy Gauss
    TI__Guru 65250 points
    Freddy,

    We don't test for that so it's hard to say exactly but I can say that the power sequencing is important and is likely to cause an issue.

    Regards,

    -Adam
  • 0 in reply to Adam Sidelsky
    Intellectual 300 points
    Haha, makes sense! I see no reason to test such a case, just wondering.

    I've tested the part for thousands of hours while powered up with no problems but when the device experienced a few hundred power cycles in another test, it failed. While correlation does not equal causation, this is currently my only hypothesis. It's a reasonable to assume that connecting DVDD and AVDD to external power (against the datasheet recommendations) resulted in failure of the part during power cycling.

    Going forward I'll be sure to follow the datasheet recommendation. Thank you for your help!