0
\$\begingroup\$

As stated in the heading, I'm currently having a problem with driving a BLDC motor with VCC being 30 V. I'm using the DRV8320S motor driver IC from Texas Instruments. It is configured in 1x PWM mode, i.e. it only needs PWM, the desired direction of rotation as well as the hall sensor values. It will then drive the three external NMOS half bridges accordingly.

Given below is a simplified schematics of the board. For simplicity, I removed the two current sensor ICs and their shunts as well as the MCU, which actually gets the hall sensor values and delegates them to the motor controller. It also generates the remaining signals for INHx / INLx. The PWM frequency is 24 kHz. Attached motor is a maxon 402685.

enter image description here

Most of the time, the motor behaves just as expected. Hall sensor mapping is also correct, as other mappings will result in high pitch and frequency noises or rapid vibration of the motor. But if the motor is in standstill, most of the time it won't start up by itself. More precisely, I got this behavior in the following situations:

  • After powering up the board (firmware is configured to drive motor instantly after initialization)
  • If motor PWM is a sawtooth (after reaching zero, the direction of rotation will alternate), the motor sometimes won't turn on after PWM passed zero
  • If you stop motor by applying a high enough external torque

In these situations, the motor driver still generates opposite PWM signals at two of its outputs as expected (active freewheeling). If it helps, current differs slightly if the motor turns clockwise or counterclockwise. Besides that, I didn't measure anything unexpected at all.

A guess from my side is, that the motor reaches a steady state acc. to the hall sensors, i.e. the rotor is perfectly aligned with the stators so that there is no effective torque. Maybe I have to adjust the hall sensor values in the MCU before I delegate them to the motor drivers, not by permuting them but by choosing the next / last state in the commutation table. But doesn't that presumption imply that the motor driver drives the motor always so that it reaches steady state, which IMHO is a contradiction to the fact that it can drive the motor clockwise / counterclockwise?

I searched a lot on the Internet, but only found some solved the problem but won't tell you how posts. Hope you can help me with this issue. If you need any further information, just ask me.

\$\endgroup\$
  • \$\begingroup\$ Are you sure the mapping is correct? There are connection possibilities that put the phase current far out of phase to ideal, that will still generate positive torque overall, but leave some positions where there is zero torque generated, so it won't start from that point. If you look at the phase voltages, they should look like figure 55 in the datasheet, with the back EMF more or less symmetrical with the applied voltage. \$\endgroup\$ – Phil G Dec 10 '18 at 23:04
  • \$\begingroup\$ We tested all six possible ways of delegating the three hall sensor values to the motor controller. Two of the combinations worked, i.e. the motor rotated. Four of them resulted in the above mentioned high pitch noises, rapid vibrations and a lot of current. The second combination that worked was asymmetric: In one direction, the motor drew more but acceptable amount of current compared to my normal commutation (but it rotated faster). In the other it drew a lot of current while rotating very slow. Hence, the only commutation which seemed to work fine was the one I already had. \$\endgroup\$ – Seradir Dec 11 '18 at 0:03
  • \$\begingroup\$ Or do you mean the arrangement of hall sensor value tripplets per revolution? \$\endgroup\$ – Seradir Dec 11 '18 at 0:06
  • \$\begingroup\$ Yellow GLB, Blue GHB: i.stack.imgur.com/hWg1S.png and i.stack.imgur.com/P5Sl8.png \$\endgroup\$ – Seradir Dec 11 '18 at 0:51
  • \$\begingroup\$ Yellow SHB, Blue GHB: i.stack.imgur.com/aBZGk.png \$\endgroup\$ – Seradir Dec 11 '18 at 0:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.