• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 635 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.

Original question:

DRV8323: High current consumption while operating without load

DRV8323: High Power Consumption With No Load

Erez Krimsky
Prodigy 10 points
Part Number: DRV8323
Other Parts Discussed in Thread: DRV8302

I'm using a commercial motor with an embedded controller using a DRV8323 (T-motor AK80-6 module) and I am trying to understand why my power consumption at no load, with no speed and no commanded torque, is so high. The driver board is powered from 24 V, when the motor is enabled (enabling the DRV8323) the current draw measured from the power supply jumps from 30 mA to 100 mA (there is some current being used to power an STM32 on the board). The driver board has current sensing and reports zero torque (and I can feel there is no torque), so where is this extra 1.7 W of power consumption going? 

The supply current for DRV8323 from the data sheet is about 12 mA this plus gate charge charge and buck converter losses would still only account for a few hundred mW. Although the peak currents can be ~10 A, the motor is going to frequently be used to provide very small torques so this extra 1.7 W makes a huge difference on system power consumption. I believe the FET control is operating at ~40 kHz. 

A very similar behavior is described here (https://e2e.ti.com/support/motor-drivers-group/motor-drivers/f/motor-drivers-forum/829628/drv8323-high-current-consumption-while-operating-without-load) but with no explanation. 

Any help would be appreciated! 

  • 0
    TI__Mastermind 26826 points

    Hi Erez,

    Thank you for posting your question on the E2E MD forum!

    Would you be able to provide your schematic for me to look at? I want to see what type of load is connected to the buck as well as to the DVDD pin. Are you able to disconnect the motor from the board to see if that makes a difference in the amount of current drawn from the power supply? When you pull the enable pin high from the DRV8323, does the microcontroller immediately apply input PWM signals? If so, could you test it with enabling the DRV but not applying any PWM signals?

    Regards,

    Anthony Lodi

  • 0 in reply to Anthony Lodi
    Prodigy 10 points

    Thanks for the quick reply! This is a commercial board so there is no provided schematic and the firmware isn't open sourced but the component layout matches the DRV8323 recommended layout from the datasheet well. The design is almost an exact copy of the board layout from this thesis (https://dspace.mit.edu/handle/1721.1/118671), for which I have attached the schematic. The initial design was for a DRV8302 but I don't think much else has changed at all.

    The firmware from the above thesis project is open sourced (https://os.mbed.com/users/benkatz/code/HKC_MiniCheetah/), but the exact firmware on the motor is not, so it is possible some pin assignments might be different and flashing my own firmware could fry things. Reading through the code here, it looks like the telling the motor to enter control mode pulls the enable pin on drv high, resets the FOC controller, and sets the q-axis and d-axis desired currents to zero. 

    The motor leads are soldered straight to the board so disconnecting them might be challenging but I could try if its useful. I can scope the voltage at various points of interest and report back if you tell me what to probe. The driver implements FOC control at 40 KHz, so I think it will be applying PWM as soon as it is enabled. 

    The power consumption is also higher than I would expect when the motor is running and applying torques. The power consumption is always about 2 W above what I would get using the standard DC motor model, integrating P = IV with I the effective current measured by the sense resistors, and setting V = IR + k*omega, with k being the torque constant.

    Also the coil resistance from the datasheet is ~ 170 mOhm. 

    Let me know what other values or measurements would be useful. Thanks for the help!

     

    Cheetah_Driver_schematic.pdf

  • 0 in reply to Erez Krimsky
    TI__Mastermind 26826 points

    Hi Erez,

    Let me look more into this on Monday.

    Regards,

    Anthony Lodi

  • 0 in reply to Erez Krimsky
    TI__Mastermind 26826 points

    Hi Erez,

    Thanks for your patience in waiting for my response!

    The fact that the power consumption from the supply is always about 2W higher than expected at all operating points after the DRV is enabled seems that there could be a possibility of some additional circuitry that is powered on at the same time that could be consuming more current, or the buck has a significant load connected to it. It is hard to debug something like this without having access to the schematic or being able to adjust parameters in the microcontroller. Since the access power consumption doesn't vary with the torque applied to the motor, then it seems that this may be related to some other circuitry, though it is difficult to be for sure without the schematic.

    I looked over the MiniCheetah schematic and there was nothing that stuck out to me as problematic, but we don't know how close the T-Motor design is to the MiniCheetah design.

    I wish I could help out more, but it is difficult to root cause this since the system has already been designed and we don't have access to the exact schematic or software.

    Regards,

    Anthony Lodi

  • 0 in reply to Anthony Lodi
    Prodigy 10 points

    Hi Anthony,

    I don't see anything else on the board that could account for the load. Other than supporting circuitry for the buck converter and the DRV8323 I think it is just an STM32 and a CAN transceiver. I don't think the power consumption is directly related to the buck converter because the load on this shouldn't change significantly when the DRV is enabled. I will reach out to T-motor to see if I can get any more information. In the meantime, can you address the following:

    1) The power consumption I'm seeing is very similar to what's given in the post here (https://e2e.ti.com/support/motor-drivers-group/motor-drivers/f/motor-drivers-forum/829628/drv8323-high-current-consumption-while-operating-without-load) which was never resolved. Do you have any thoughts on what might be going on here? 

    2) How would I go about estimating the total expected power consumption at very low loads (a few mA of rms q-axis current) taking into account gate charge for the FETs and the PWM frequency? 

    3) If there is an issue with the FOC control, could there be d-axis currents that lead to this higher power consumption even when there is almost no q-axis current? The controller should try to drive d-axis currents to zero but maybe there is an issue here.

    Thanks again for all the help,

    Erez 

  • 0 in reply to Erez Krimsky
    Prodigy 10 points

    After digging more into the firmware, there is a field weakening portion of the controller which may lead to power losses from the d-axis currents...

  • 0 in reply to Erez Krimsky
    TI__Mastermind 26826 points

    Hi Erez,

    You mention that there is a field weakening portion of the controller, is this engaged immediately after enable is pulled high? You mentioned that the torque vector and speed is 0 

    Regards,

    Anthony Lodi

  • 0 in reply to Anthony Lodi
    Prodigy 10 points

    The torque and speed is definitely zero. The controller code seems to set a "reference" q-axis and d-axis current to zero, but it looks like these values get adjusted by the field weakening portion of the controller which appears to be active as soon as the enable is pulled high. There may be a way to adjust some field weakening parameters over UART which I will try next week. 

    If the field weakening is running, do you think this could explain the higher power consumption? 

  • 0 in reply to Erez Krimsky
    TI__Mastermind 26826 points

    Hi Erez,

    I am not as familiar with field weakening, but if the field weakening portion adjusts the q-axis and d-axis current no a non-zero value, then this does seem to make sense to correspond to higher power consumption. Looking forward to seeing if adjusting those parameters will effect the current consumption.

    Regards,

    Anthony Lodi 

  • 0 in reply to Anthony Lodi
    Prodigy 10 points

    Finally was able to test this and it seems that the power consumption is not related to field weakening. Even with the maximum field weakening current set to zero, I still see the same power consumption. Any ideas? What current consumption should I expect? 

  • 0 in reply to Erez Krimsky
    TI__Mastermind 26826 points

    Hi Erez,

    I discussed this with the team and was able to get some more insight into some possibilities that can factor into why you are seeing the higher current. 

    1. Circuitry that is powered by the DVDD LDO of the DRV8323 can result in an additional up to 30mA of current

    2. If the buck is set to be enabled at the same time as the DRV, then there can be additional current consumed by the buck that is dependent on the load of the buck. 

    3. Some designs will use an NMOS as a high side switch for some other circuitry on the board, and some people will use the charge pump of the DRV to provide 12V above VM to this NMOS to turn it on. This will result in an increase in current of the supply once the DRV is enabled since the charge pump will also be enabled which will turn on the NMOS somewhere else in the design which then results in another path opening up for current to be drawn from the supply. 

    There are several variables, and unfortunately we will need to have the schematic in order to analyze this any further. The only other option I can think of to further analyze this without the schematic is to try to cut specific traces on the PCB to measure specific currents. I wish I could give you an exact answer, but there are several variables that could effect this and it is not something that is feasible without the schematic.

    Regards,

    Anthony Lodi