TMS320F280023C: BLDC motor taking more current, when number of poles is reduced.

Hi Ratnesh,

Perhaps try higher PWM frequency like 20kHz. We can't expect every motor will require the same drive current. How much harder to turn the 8 poles shaft by fingers to that of 24 pairs, did you really mean to say 48 poles? Typically Trapezoidal modulation will take more startup current for low pole count motors than do higher pole counts.

Regards 

Hello GI,

1. I tried with 20kHz PWM frequency. The results were same.

2. 8 pole motor is easier to rotate than 24 pairs ( 48 pole) motor, because 48 pole motor is geared.

3. It's not only about starting current, overall running time is also huge.

If 8 pole motor is run with Chinese motor controller, which came as a set with the motor. The motor runs very smooth and takes only about 4A at full duty. While in my algorithm, it takes nearly 6-7A current when nearly 30% duty is ON. 

Could it be switching sequence problem ? Or it has to something with inductance of motor winding ?

Regards,

Ratnesh 

Ratnesh Kumar Singh said:

Could it be switching sequence problem ? Or it has to something with inductance of motor winding ?

Sound like that is the issue 8 pole has very low inductance, 30% duty is not unexpected at 6-7A NFETS in slow decay current mode. Does your x23c support SVM modulation via embedded ROM? You could try the FAST estimator universal motor control projects, if x23 has not embedded ROM there is the FAST ROM library file to use. The imported projects require to use the x23 peripherals so there is some conversion time involved.  

Oddly enough, Nidec H25 6 pole motor has 1.3A 18w 20vdc listed but only take 300mA/24vdc via SVM sinewave current modulation. Some posters have reported high current consumption via low inductance large motors. H25 is only 3mOhm p2p and does 1.3A peak, 1A average with Trapezoidal modulation.

Hi,

I understand, you are suggesting to use SVM method to control Motor. Is there any specific advantage of using SVM over trapezoidal control ?

These are the values of Inductance and resistance (phase to phase).

Ls = 67.2uH @10kHz
Rs = 30mohm @ 4Hz

48V / 1.2kW / 8 pole motor 

Are you using sensorless or sensor based algorithm for BLDC with trapezoidal control? Which TI reference lab/project are you working on?

Do you change the control parameters for the different motor? And do you do the offset calibration if you are using the hall sensor based trapezoidal control?

Hi Luo,

I am using sensor based Trapezoidal control. 

Not using TI reference lab.

Also, I would like to inform that problem was with hall sensor wrong connection, which resulted in wrong switching pattern. And hence lots of current was being drawn. Now Motor is being controlled in a fine manner. 

Thanks.

Also, Can you please suggest some TI reference to measure BLDC motor speed using hall sensor signals. Now I am using Hall interrupts and calculating time between rising edge to rising edge to calculate speed. But, using this method it holds the value of speed, when when speed is brought to zero. Please suggest some better method or some TI reference lab.

Ratnesh Kumar Singh said:

But, using this method it holds the value of speed, when when speed is brought to zero

You can add a function ADC tick counter to reset the speed value when EMF edge is not being detected for some period.

I was reviewing the Universal code for BLDC with Hall's and assuming it used trapezoidal modulation. When Hall sensors are used for rotor position how is that considered FOC is good question. How the SVM with Halls has any advantage over FOC is another good question.