• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Subscribers 30 subscribers
  • Views 2390 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.

DRV8332: need help estimating power usage

Robbie Valentine86
Expert 1550 points
Part Number: DRV8332

Hello,

We are using the DRV8332 motor driver in a design that is getting ready to go into small quantity production.  We need to do a thermal analysis on the design and therefore I need to figure out how much power this chip uses under load. 

I'm assuming the 3 phase BLDC motor draws 3A at 24V when being driven under full load.  I need some guidance as to calculating the power dissipated in the DRV8332.  

My initial idea was to use I^2 * R where R is the Rds_on of the FETs in the driver.  I got 80 milli ohms from the data sheet which specifies a single FET so I doubled that to 0.160 Ohms.  So, my initial power estimate is 9A*0.160 = 1.44W.  This estimate alone doesn't seem to match with our bench testing when analyzing how hot the chip gets and when it over heats and shuts off.

Does anyone have any guidance as to how I could more accurately calculate a power estimate in this situation?  Does the PWM waveform/gate driver switching play a role?  What other things should I take into account and how would I go about doing that?

  • 0
    TI__Guru** 114905 points
    Hi Robbie,

    Have you measured the current in the motor to determine you have 3A?
    How far off do you think you are?

    There are a few items that can help improve the calculation or improve the performance:
    1) Please note Figure 3 of the datasheet, which will help you estimate the RDS(on) as the junction temperature increases.
    2) How long will the body diodes conduct current? This can be several W (Vf * I).
    3) How long is the high side on without refreshing the boot capacitor? This can lower the Vgs, increasing the RDS(on).
  • 0 in reply to Rick Duncan
    Expert 1550 points

    Hi Rick,

    Thank you very much for the response!  I will do some further research based on your guidance.  I estimated 3A which was a little high.  The current was actually measured at 2.5A while running at a constant rpm with a constant load.  The issue we're having is that the drivers overheat much faster than our thermal model expects when using the power number I came up with in my estimation.  I know I'm missing some things so I'll go back and further review the data sheet with your notes/numbered items and see if I can get a better number.  

    How do I figure out how long the body diodes conduct?

    How do I know whether the boot capacitor is being refreshed or not?  Does this affect the Rds_on because without the boot cap the power fets don't turn on as "hard" and therefore experience high Rds_on values?

    Thanks for your time!

    Rob

  • 0 in reply to Robbie Valentine86
    TI__Guru** 114905 points
    Hi Rob,

    How do I figure out how long the body diodes conduct?

    You can determine when the body diodes are active by monitoring the voltage at the phase pins. When the voltage is approximately VM+0.7, the high side body diode is conducting. When the voltage is less than -0.7V with respect to GND, the low side body diode is conducting.

    How do I know whether the boot capacitor is being refreshed or not?

    The boot capacitor is either being refreshed by your firmware setting the PHASE low or by the internal recharge sequences. Please refer to section 8.3.2.1 for details.

    Does this affect the Rds_on because without the boot cap the power fets don't turn on as "hard" and therefore experience high Rds_on values?

    As the voltage in the bootstrap cap droops, the RDS(on) will increase some. The recharge cycles described above are intended to minimize the RDS(on).
  • 0 in reply to Rick Duncan
    Expert 1550 points

    So the current through the body diodes....Is that just the current delivered to each motor winding?

  • 0 in reply to Robbie Valentine86
    TI__Guru** 114905 points
    Hi Rob,

    Yes it is. Please refer to Figures 6 and 7. If you are using cycle by cycle and high current, the power could be significant.
  • 0 in reply to Rick Duncan
    Expert 1550 points
    From reading through that section of the datasheet and looking at the figures, it looks like figure 6/7 are only applicable to an over current event. I am not experiencing an over current event while running. For example, my 3 phase BLDC motor draws 2.5A under load. My current limit is set to 8A currently. Does this mean figure 6/7 are not applicable to my specific situation?

    Why are the body diodes conducting during normal operation? I thought the diodes would be effectively shorted out when the FETs are on/enabled.
  • 0 in reply to Robbie Valentine86
    TI__Guru** 114905 points
    Hi Rob,

    That is applicable to normal operation also if you are using M3=1, and M2=M1=0 (cycle by cycle current limit). In this case, if the current reaches the current limit set by the OC_ADJ resistor, the outputs will be disabled until the next PWM cycle.

    A more detailed description of cycle by cycle operation can be found in section 8.3.2.1.1 of the datasheet.
  • 0 in reply to Rick Duncan
    Expert 1550 points
    So would it be accurate to say the power dissipated in the body diodes (with a 2.5A load) are: Pd = (Vf*I) = 1V * 2.5A = 2.5W?

    I could then add that to the power calculated using Rds_on to get a reasonable power number?
  • 0 in reply to Robbie Valentine86
    TI__Guru** 114905 points
    Hi Rob,

    Yes, that would be a good approximation of the power in the body diodes.
  • 0 in reply to Rick Duncan
    Expert 1550 points

    Hi Rick,

    You're advice was helpful as we are now seeing numbers that make more sense.  Am I correct in seeing that the body diodes dissipate much more power than the FETs due to the Rds_on resistance?

    If the motor is pulling an average 3A I get a much larger number for the body diodes than the Rds_on.

    I am curious, why do the body diodes conduct so much?  From the data sheet it looks like the diodes should be shorted out by the FET conducting when the FET is on.  What is the circumstance that causes the body doides to conduct?

  • 0 in reply to Robbie Valentine86
    TI__Guru** 114905 points
    Hi Rob,

    The power through the diode is considerably higher than through the FETs. At 3A, the FET power is ~.72W (I2R) while the body diode power is 3W (VI).

    The diodes should only conduct when the FETs are off. This is typically during dead time, but can be extended by the cycle by cycle current limit. If you monitor the voltages and see the voltage switch high or low prior to the end of the PWM cycle, you are probably entering cycle by cycle current limit.

    If the cycle by cycle current limit is activated, the remainder of the cycle will use the body diodes. Ideally this can be avoided by monitoring the current and switching to fast or slow decay prior to reaching the cycle by cycle cutoff. The current in fast decay (also referred to as coast) flows through the opposite high side and low side FET. The current in slow decay (also referred to as brake) recirculates the current in the low side FETs.