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.

DRV10975EVM: BEMF Abnormal reported - closed loop, open loop

Part Number: DRV10975EVM
Other Parts Discussed in Thread: DRV11873, DRV10975

Greetings. I am evaluating the DRV10975 as a replacement for the DRV11873 due to the false rotor lock issue associated with the DRV11873. I've worked through the tuning guide, and I appear to have good motor operation at this point . . . but I consistently get a BEMF abnormal reported (#2 fault code). I've confirmed Kt to be ~16 (4V peak to peak, 8ms period) using the finger spin method. I've also confirmed phase to phase resistance and divided by two (2.11 ohms). Odd thing is, if I start the program with the defaults. The motor spins (poorly, can't get out of open loop) with no BEMF abnormal error reported - resistance = 1.17, Kt = 120.

And unfortunately I will need the BEMF error since the motor does not seem to draw enough current in a locked state to trip the rotor lock requirements (more current is drawn during acceleration, etc).

Rapid help appreciated, we are in a bit of a bind with this last minute DRV11873 discovery and so far I'm not feeling too confident in the DRV10975.

-Chris

  • For reference, we are using this Maxon motor (9V winding operating at 12V):

    www.maxonmotorusa.com/.../17-EN-261.pdf
  • I've followed up with some coasting tests of BEMF (after shutdown) and I am measuring a 8V peak to peak at a 4ms period and a 4V peak to peak at a 7.5ms period - further support of the 16mV/Hz Kt. Any feedback regarding this constant BEMF error appreciated - we cannot proceed with this part until it is resolved.

    Thanks,

    Chris

  • Hey Chris,

    It sounds like you're dealing with a false trigger of Lock 2: (Abnormal BEMF/Kt). While this situation is common, it is important to understand when you get this false lock. Does it occur during the open loop process or can you enter closed loop but then trigger the false Lock after achieving some higher speed? You can check this by going on the Basic Settings Tab in the Startup Setting section and marking the CLoopDis box (to disable closed loop) and see if you are still triggering the Lock.

    Assuming that the Lock is falsely triggering in open loop or right at the transition from open to closed loop, this means the Kt that has been programmed is not allowing the driver to spin the motor correctly. I see you've done the hand test, coasting test, and looked at the datasheet so I will offer a different way to check Kt. Make your Open to Closed Loop Threshold sufficiently large (25.6Hz or 51.2Hz) and mark the CLoopDis box as mentioned before using your various Kt's. Then, assuming the Lock did not trigger, read the value in the Motor Velocity Constant (mV/Hz) box and type that in as the new Kt value.

    Also, I have posted a picture of some generic settings that will give you the greatest chance to successfully start up the motor. Just add in your Resistance and Kt into the box as needed and see if that helps.

    Let me know if this helps.

    -Cole

  • Cole-

    It is important to note that, while I am receiving this abnormal BEMF error, the motor is otherwise operating well up to full speed. It spins properly in open loop, transitions well to closed loop, and accelerates quickly.

    When I start the EVM gui, the motor starts up with whatever defaults are burned in the eeprom and no abnormal BEMF error is reported. But, with these settings, I also can't accelerate, move to closed loop, etc.

    When I load my settings that functionally appear to work well, the abnormal BEMF error is reported in open loop (closed loop disabled) and in closed loop too.

    Working through your recommendations above - embarrassingly enough, I previously did not see the driver's report of Kt - but Kt is reported as being between 10 and 17 mV/Hz (varying a bit between running and coasting). Updating Kt does not clear the abnormal bemf error, even once stopped and re-started.

    Curiously enough, when I use the settings you listed above and remain in open loop, the reported Kt value is much higher. If I update the Kt setting to match, the reported value climbs higher still. I "chased" Kt up to over 100 - abnormal BEMF still reported, and current draw in open loop is excessive. Note that, when in closed loop, if I increase Kt from ~16 to over 100, I get a "speed abnormal" error.

    Any additional ideas? Needless to say, moving from a part that issued a "false rotor lock" error (DRV11873) to a part that issues a "false abnormal bemf" error is more than frustrating. I'm hopeful this is just a tuning issue, but I'm not sure what else I should be adjusting to avoid this error.

    Again, rapid help appreciated!

    -Chris
  • In addition, when operating in closed loop, Kt is reported as being 7mV/hz - not sure if this helps
  • Chris,

    If I read the datasheet for the motor correctly:

    Kt = 9.5mNm/A = Ke = 9.5 mV/rad/s = 9.5 V/rad/s (2 pi rad) = 60 mv/Hz

    Curious that it doesn't line up with your measurement.

    I would next take a look at the t-control advance. If this isn't close to aligning the BEMF with the current then the device could report a false lock.

    Thanks,
    Brian
  • Yup, or backtracking out from the speed constant in the datasheet yields 0.059V/Hz. But, I've measured the same Kt a few times, and beyond that, setting Kt to 60mV/Hz also does not clear up the abnormal BEMF error.

    T control advance does not impact open loop, correct? The abnormal BEMF error is received when operating in open loop as well.
  • Hey Chris,

    Yeah, an overwhelming-majority of the time, the abnormal Kt false trigger is a tuning issue. I'm sorry that you've gone from one issue into another! Definitely frustrating.

    In addition, yes, Tadv will only help in closed loop and in the transition from open to closed loop.

    While I try to compile everything we've learned so far, can you disable the Abnormal Kt lock and see if motor operation will trigger any other Lock? I know you would like to keep it in the final implementation but sometimes checking if another Lock occurs can tell us more with what is happening with the motor and its behavior.

    Best,
    -Cole
  • Cole-

    BEMF abnormal lock detect has not been selected during all of this testing, but the fault code is still issued (indicator illuminated, confirmed with "manual refresh" after configuration changes, etc). In the past, upon enabling the BEMF abnormal radio button, the motor immediately shuts down due to the presence of the fault, but admittedly I haven't tested this in a while. I will go double check. With the BEMF abnormal lock detect radio button deselected, no other faults are issued. As I mentioned before, the motor seems to operate well with the fault issued (BEMF abnormal lock detect deactivated).

    I don't want to jump to this conclusion yet, since it rarely seems to be true, but is it possible that there is an issue with this EVM? ESD zap or something else? I'm assuming that since the driver seems to be working well otherwise, it is pretty unlikely, but figured I would ask.

    Can you also explain how the BEMF abnormal fault is issued/cleared? I'm assuming it is based on the "abnormal Kt lock detect threshold", but why does it not clear when I manually set Kt to be within that threshold? This is the only behavior that makes me think that there may be something wrong with this EVM - but I'm guessing that the fault detect algorithm is more complicated than I am making it out to be.

    -Chris
  • One more thing to add. When I started adjusting Kt up in the 60-100mV/Hz range, I did start to see "speed abnormal" errors reported.

    -Chris
  • Some more information that is hopefully sparks some ideas.

    I can run the motor up to about 34% duty cycle without the abnormal BEMF error being issued - this is starting from dead stop, moving through open loop and closed loop. Good news. At 34%, the reported BEMF is ~9.4mV/Hz whereas the setting is at 16.7. As the duty cycle moves up, the reported BEMF continues to drop. Bad news. Once it is outside of the lock threshold (0.5Kt-2Kt), not surprisingly, the abnormal BEMF error is issued.

    So, why is the reported BEMF deviating so much from the setting? Or why is it deviating so much period? Are there any tuning parameters that affect the reported BEMF?

    -Chris
  • Hey Chris,

    So I'm going to go back a few steps and make some comments about what you've said earlier.

    Good to know you've just been monitoring it with the GUI instead of allowing the device to immediately stop operating and attempt to restart. Unfortunately, since the motor runs well without it selected that usually means the programmed Kt is the problem (which is conflicting with our findings so far).

    Let me explain how Abnormal Kt Lock works: Our device takes the programmed Kt in the GUI and compares it to the estimated Kt (because measured Kt is not completely accurate) by our device. The estimated Kt will vary during operation some amount but our device determines it should stay within some threshold of the programmed Kt. This is shown by the picture below (imagine the estimated Kt is the line). If the variance is outside the threshold for too long, it triggers the lock. In summary, your "manually set Kt" actually determines where the threshold is. You can check out the Lock2: Abnormal Kt section of the datasheet for more information.

    The "reported BEMF" (I called it estimated Kt) deviates because it depends on current throught the phase, voltage applied to the phase, Tadv, programmed motor resistance, and motor speed. It can calculate Kt different based on these factors and compare it against the programmed Kt (we actually do this for the speed abnormal lock1. feel free to check out the datasheet for more information).

    As for ESD, I'm not this is very likely but its a fair question to ask. You can probe the voltages on 3P3V or VREG to make sure they are within the expected range of values. However, if the device was damaged, you probably wouldn't be able to communicate via I2C with the device.

    As for Speed Abnormal, when you increased Kt, did Abnormal Kt Lock go away but Speed Abnormal take its place even? Or if you disable Speed Abnormal and observe it trigger, did you eventually see Abnormal Kt occur again? Also, does aritificially lowering your programmed Phase resistance with a higher Kt help at all?

    As for giving a duty cycle of 34% and in closed loop without error, are you able to change Tadv in a way that will allow you to increase your duty cycle higher? We're trying to make sure the motor some how isn't getting out of phase with the way we are applying current.

    Best,

    -Cole

  • Cole-

    I purchased a new dev board to eliminate "dev board damage" as a possible source of error. The issues seen on the first dev board are replicated on the second. Likewise if I swap the motor. The issue appears to be the driver's interpretation of the Kt of the motor (per your comments above) - it seems to vary significantly from open loop to closed loop and changing further still as the motor speed changes. Is this expected? It seems like the only way I can get around these errors would be to increment speed slowly, adjust parameters, increment, adjust, etc. Doesn't seem like a great solution. Additionally, if I drop "set Kt" down to match "reported Kt" (to avoid the BEMF error), motor speed drops - limiting the output of our blower to an unacceptable level.

    Regarding your questions above, if I tweak Tadv, startup behavior is adversely effected - lower and I can't start, higher and current is excessive along with a "closed loop stuck" error reported.

    -Chris

  • Hey Chris,

    Just to summarize what we did so far I can close out this thread.

    • We verified that your Kt setting (16.7mV/Hz) was correct. Note this was different from our datasheet reading but correct based on measuring the Kt manually
    • An incorrect Tadv was causing the Abnormal Kt Lock2 error
      • This happens because Tadv will control how in phase the BEMF and Current waveforms are. If they are misaligned, that means the motor is not perfectly driving the motor in phase which can lead to the Abnormal Kt Lock2 error
    • Doubling the PWM frequency helped spin the motor at higher speeds

    Please verify the answer if you feel I've answered this sufficiently.

    Best,

    -Cole

  • Cole-

        I feel like that might be slightly marginalizing the effort that it took to get this motor up and spinning reliably - certainly when I was initially working with the motor, changes to Tadv didn't help - but yes, at a high level, doubling the PWM enabled a Tadv value which maintained a reported Kt value that stayed within range of the measured Kt value. I really appreciate all of your and Brian's help on this, certainly I was not making very good headway alone!

    -Chris