• State TI Thinks Resolved
  • Locked Locked
  • Replies 14 replies
  • Answers 4 answers
  • Subscribers 33 subscribers
  • Views 1701 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.

DRV8308EVM: PWM Duty Commend

Brian Wang0928
Genius 14030 points
Part Number: DRV8308

We're trying to fine tune the customer fan with DRV8308EVM. However, we are confusing the installation step 5 in the user guide of DRV8308EVM. It says that "a 2.1 x factor is applied to the input duty cycle, 40% duty cycle is treated like 84%" for the 180 SINE commutation control. Why it's not linear control? I meant 10% input duty should have 10% duty cycle output, or 100% input is 100% duty output. what's the purpose?

When the customer fan is controlled with DRV8308EVM and GUI, the max. speed is around 18,000rpm at GUI 60% PWM duty cycle. however, when the PWM duty cycle is increased from 60% to even 70%, 80%,90%, the speed is still around 18,000rpm.  

Regards,

Brian

  • 0
    TI__Expert 4435 points
    Hi Brian,
    MOD120 will change depending on what PWM mode customer is using. The section described in step 5 of the user's guide is for open-loop PWM.
    In the case of 120 degree commutation, then yes, MOD120 can be set to 2048 and will get in 1 to 1 input to output duty cycle .
    What is the PWM mode customer is using? If closed loop Clock frequency mode, datasheet states to set MOD120 to 3970. Thank you.
  • 0
    TI__Intellectual 900 points

    Part Number: DRV8308

    Hi, Team

    One customer reported one issue on DRV8308.

    When they set MOD120 to 3970 and achieved sin wave,

    output duty should be 2.1 times of input duty.

    So motor should arrive full speed before 50% input duty.

    But test showed 50%-17000rpm, 60%-18000rpm.

    LOCKn low confirmed.

    Any thoughts to explain this.

    Thank you.

  • 0
    TI__Genius 14487 points
    Hi Alvin,

    It looks like the device uses internal scaling values from a lookup table to choose the duty cycle. The quote below comes from page 33 of the datasheet.

    "During sine commutation, the input duty cycle is multiplied by the modulation values for each phase (MOD_U,
    MOD_V, and MOD_W) to generate a 12-bit value that determines the output PWM duty cycle of each phase."
  • 0 in reply to James Lockridge19
    TI__Intellectual 900 points
    Hi, James

    True. But in sine wave driving, MOD120 is fixed as 3970. The input duty should be multiplied by 2.1 per my understanding.
  • 0 in reply to Luis Riveros
    TI__Genius 14030 points

    What PWM mode we using is the open loop PWM duty cycle mode. if so, how to use it for 180 commutation SINE control?

    and also, i attached the file which shows the U phase current waveform, duty cycle and FG. In which you can see the current is unbalancing and speed is NOT significant increased when the PWM duty cycle > 60% on UCC8308EVM. Why? how to turning the parameters well for fan control?Please advice? Very thanks.  

    DRV8308 180degree Spin test.pdf

    Regards,

    Brian

  • 0 in reply to Alvin Zheng
    TI__Genius 14487 points

    Alvin,

    Are you debugging the same issue as this post? If so, I would like to combine these posts.

  • 0 in reply to James Lockridge19
    TI__Genius 14487 points
    If these issues are the same, the customer may be running into some kind of performance limit with this device based on their speed and motor parameters. We may be able to find a workaround.
  • 0 in reply to James Lockridge19
    TI__Intellectual 900 points
    Yes. Same issue.
    From remote support point of view, any ideas to suggest what needs to be checked?
  • 0 in reply to Alvin Zheng
    TI__Genius 14487 points

    Hi Brian and Alvin,

    I combined your E2E posts on this issue.

    To help debug, I want to separate your questions into separate items.

    1) Current loses regulation: At high speeds the back EMF generated by the motor may be significant. This may also contribute to the nonlinear PWM/speed relationship as well as the strange current waveforms. If possible, increase the VM supply voltage. Back EMF limits motor speed as the magnitude of the back EMF reaches the supply voltage. Increasing the supply voltage allows for larger back EMF values, which means higher speed.

    2) Speed does not increase linearly with duty cycle: This may be due to the torque-speed performance of the motor. As the speed increases, motor torque decreases. Additionally, if the fan is connected to the motor shaft, the load torque requirement will increase significantly as speed increases at high RPM. The load torque requirement is nonlinear, so the torque-speed operating point that the motor reaches also may not be linear with PWM duty cycle.

    3) PWM duty cycle is not 2.1x the input duty cycle: Can you please confirm this with oscilloscope screenshots for a few different duty cycles? If this is an issue, I will ask the design team for more details on why the part might operate this way.

  • 0 in reply to James Lockridge19
    TI__Genius 14030 points
    Thanks. In terms of your answer #3, in the previous mail, we also attache a measurment for different duty cycle with 19.51%, 26.82%, 48.72%, 70.72% and 97.53%. However, as a results, when the duty cycle is higher than 60% duty cycle around, the speed is not increased signeficant and the current is unbalancing. Is this measurement good enough? or what specific measurements you need to know?
    We want to know why the output PWM duty needs to be scale with 2.1x of the input PWM duty cycle. and what's the purpose? and how to explain the current & dutycycle we measured for 180 SINE commutation? We'll also check if increased VM can improve the unbalancing current or not? We're requested to answer it to my customer. Very Thanks.
  • 0 in reply to Brian Wang0928
    TI__Intellectual 900 points
    Hi, Brian

    As James mentioned, we would like your help to check if this issue is either related to BEMF or motor torque-current requirement. Test with higher VM will help to see if the issue comes from high BEMF of motor. Test with smaller load (small fan leaf etc.) may help to clear the torque-current requirement. For the 2.1 factor, this IC is designed to be that way.
  • 0 in reply to Alvin Zheng
    Expert 1875 points

    Dear Alvin,

    I have done  the VM-changed experiment,

    Please refer to attach file and the phase U current wave form is still unbalance.

    DRV8308 Test with VM Change.pdf

    best regards,

    Simen

  • 0 in reply to Simen Biao
    TI__Genius 14487 points
    Hi Simen and Brian,

    Alvin mentioned to me that the device drives properly at high speeds with trapezoidal control. I thought this was interesting. I looked at the sinusoidal and trapezoidal commutation tables in the datasheet. One particular difference is that the sinusoidal driving technique tries to ramp the current in the form of a sinusoid, while trapezoidal control drives the power stage at the input PWM frequency without ramping it. I think this allows the phases to reach the desired voltage faster. At slow speeds, the current can respond fast enough for sinusoidal control because there is more time for the coils to charge and for the current to track the ramp properly. I think the higher speeds make it more difficult for the current to track the sinusoidal reference, especially when back EMF is high.

    I still think that this issue is due to the significant back EMF that is created at high speeds. Your scope shots show that the distortion becomes worse as duty cycle increases, and because duty cycle relates to speed, this makes me think that the back EMF is causing worse distortion. This distortion from back EMF may limit the speed.

    When you increase the VM voltage, the 80% duty cycle creates a different voltage on the coil. For a good comparison, run a test at VM = 15 V with 80% duty cycle, and a second test at VM = 30 V with 40% duty cycle. This will keep the average voltage across the motor phases the same, but the higher voltage should have faster charging and discharging times within the PWM.

    If this still does not show an improvement, you may need to use a flux-weakening technique at higher speeds. Unfortunately, this feature is not available on the DRV8308. This is a higher-performance driving technique used by the C2000 controller and I think it is available on InstaSpin. To show if the back EMF effect is the root cause of this issue, you can test the motor with flux weakening control. I think flux weakening control will help with sinusoidal driving at high speeds. Alternatively, you can use a motor with a weaker magnetic field on the rotor.
  • 0 in reply to James Lockridge19
    TI__Expert 8985 points
    Hi Brian,

    Did James' suggestion resolve your question?

    If so, please close this post by accepting the post is resolved in James' highlighted in yellow answer above me.

    Thanks!