• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 462 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.

DRV5012: Duty Cycle of Hall sensor is not stable at high speed

Sayali Shinde
Intellectual 705 points
Part Number: DRV5012
Other Parts Discussed in Thread: DRV5011

Hello Team,

  We are using 3 qty DRV5012AEDMRT hall sensor  in our flat BLDC motor for commutation purpose. There are placed 120 degree apart. Motor is of 16 pole magnet (8 pole pair).

distance between hall sensor surface and magnet is 1.6mm

when we drive motor using sensored driver we observed it is taking very high current and when we driver same motor with sensorless driver it consume very less current.

With sensored driver we observed hall sensor waveform on oscilloscope and found that at low speed , the duty cycle and phase shift is with in spec i.e 50% and 120 degree respectively.

As we increases speed the duty varies and phase shift also.

for reference refer below snapshot of hall sensor at different speed. Can you please help us to understand what will be root cause for this and how can we improve duty cycle at higher speed.

1. speed -1300 rpm

2. speed -5300 rpm

  • 0
    TI__Expert 7950 points

    Hi,

    Thanks for posting your question to the Sensors forum.

    With 5300 rpm, the number of poles passing the sensor per second is approaching the sampling frequency of the device -> 5300 rpm * 16 poles / 60 seconds = 1413 poles per second. The device sampling frequency with SEL pin high is 1665 Hz / 2500 Hz / 4700 Hz (min / typ / max). Generally, the device sampling rate should be at least faster than 2 times the the number of poles per second, but ideally this number is 3 times, 4 times, or even more than the number of poles per second.

    Additionally, at ~5300 rpm, which is ~88 revolutions per second, each revolution has 16 poles (8 electrical cycles), so the electrical frequency is ~704 Hz, a period of 1.4 ms. The DRV5012 in 2.5 kHz mode has a typical sampling period of 0.4 ms, which is ~28 % of the electrical period. Generally, the maximum timing error should be kept under 10% to ensure the BLDC motor spins, and timing error can reduce motor efficiency.

    More information on the calculations above can be found in the Application and Implementation section of the DRV5012 datasheet. For a device with a higher sensing bandwidth in a similar package, you could try the DRV5011 (30 kHz bandwidth).

    Best regards,

    Jesse

  • 0 in reply to Jesse Baker
    Intellectual 705 points

    Hello Jesse,

      Thank you  for detail explanation. please give calculation in our case on " how the device sampling rate should be at least faster than 2 times the the number of poles per second, but ideally this number is 3 times, 4 times, or even more than the number of poles per second".

    can DRV5011 can solve our issue or we need to validate first.

    can you please provide above theortical calculation with DRV5011 just to reconfirm that it will work at higher speed.

    Regards,

    Sayali'

  • 0 in reply to Sayali Shinde
    Intellectual 705 points

    Hello Jesse,

       Please verify my below calculation if I use DRV5011 in my motor

    motor speed - 6310
    motor poles - 16 poles (8 pole pair)

      DRV5011
    Motor pole 16
    Motor speed 6310
    Freq (poles per sec) 1682.666667

    , at ~6310 rpm, which is ~105 revolutions per second, each revolution has 16 poles, so the electrical frequency is ~841 Hz, a period of 1.1ms. 

     from above calculation can this sensor will work or give proper output at high speed.

    please provide approximate working  speed range  for both sensor  DRV5011 and DRV5012.

    I think for DRV5012 speed it can support is upto 2000 rpm  with 16pole i.e  poles per sec should be equal or less than 533 right?

    similarly for DRV5012 poles per sec should be equal or less than 6000 right?

    please confirm

    Regards.

    Sayali

  • 0 in reply to Sayali Shinde
    TI__Expert 7950 points

    Hi Sayali,

    The rule of thumb is that the frequency bandwidth of the sensor should be at least two, but ideally three times greater than the number of poles per second.

    Poles per second = (RPM * number of poles) / 60

    The sensing bandwidth of the DRV5011 is 30 kHz. Assuming the rule of the bandwidth being ideally 3 times greater than the poles per second, the max poles per second for the DRV5011 would be 10 kHz. If we solve for RPM in the equation above assuming 16 poles, the max RPM for the DRV5011 would be 37,500 RPM.

    Following similar calculations for the DRV5012, with a bandwidth of 2.5 kHz, the max poles per second would be 833 Hz, yielding a max RPM of 3,125 RPM.

    A final thing to consider here is that the B-field strength is sufficiently strong enough to meet the sensitivity parameter for the device. The Bop max specification for the DRV5011 is 3.8 mT. We have a recently released online simulation tool called TI Magnetic Sense Simulator (TIMSS) where you can configure magnet and device location and parameters and simulate the expected magnetic field strength and device output. You can access TIMSS here: https://webench.ti.com/timss/ 

    Best regards,

    Jesse

  • 0 in reply to Jesse Baker
    Intellectual 705 points

    Hello Jesse,

      thank you for detail explanation

    Regards,

    Sayali