FOC alignment routine

MARK PESACH

Other Parts Discussed in Thread: DRV8301

HI!

I am using the IDDK_PM_SERVO example for sensored FOC motor control.one of the first steps is the rotor alignment for determining rotor flux angle. during this routine the Iq_ref current is set to zero , and the Id_ref current is set to some initial value that will set the motor to 0 degrees with respect to phase a. during this stage the angle of the I_park and the Park macro is 0 degrees, so basicly the alfa beta coordinates are same as the iq coordinates . i checked with some Id_ref values and i always get the voltage vector that the currents pi regulator produces is u60 (011) , i was sure that if i want all of my current to be in the d axis direction it will mean that the qd voltage vector will also be exactly in the d direction also. and the voltage vector will be u0(001).

i will be very helpful i some one can explain this issue

best regards

Mark

over 10 years ago

0 Ramesh Ramamoorthy over 10 years ago

TI__Mastermind 27240 points

If park.angle is zero, then alpha = d and beta = q. Since the q voltage is zero, only alpha should be greater than zero. If single step through the SVGEN code, you will notice that the variable 'v.sector' takes the value 1. For this value, the PWM state should be 001 or 100, depending on how you set the phase sequence.

rgds,

ramesh

0 MARK PESACH over 10 years ago in reply to Ramesh Ramamoorthy

Intellectual 850 points

hi ramesh

thanks for the reply , i am still getting (011) vector instead of 001 or 100.

its strange because i am using the original code of the IDDK_PM_SERVO example without any changes, maybe its have something to do with this code part:

EPwm6Regs.CMPA.bit.CMPA = (INV_PWM_HALF_TBPRD*svgen1.Tc)+INV_PWM_HALF_TBPRD;

EPwm7Regs.CMPA.bit.CMPA = (INV_PWM_HALF_TBPRD*svgen1.Ta)+INV_PWM_HALF_TBPRD;

EPwm8Regs.CMPA.bit.CMPA = (INV_PWM_HALF_TBPRD*svgen1.Tb)+INV_PWM_HALF_TBPRD;

where TC goes to phase a pwm register?

i will try to check it , would be grateful for any suggestion.

mark

0 Ramesh Ramamoorthy over 10 years ago in reply to MARK PESACH

TI__Mastermind 27240 points

How did you find 011?

-ramesh

0 MARK PESACH over 10 years ago in reply to Ramesh Ramamoorthy

Intellectual 850 points

i am using the build levels and debugging mode so i can see the variables during the alignment routine.i see what the pwm output is , and also the output of the pi regulators Qs and Ds is not as i expect it to be: as i understand it should be QS=0 and Ds different from 0 but instead i get values that represent the 011 vector.

0 Ramesh Ramamoorthy over 10 years ago in reply to MARK PESACH

TI__Mastermind 27240 points

But it is not clear how you inferred 011. Btw, where did you see pwm output? On the high side or low side?

0 MARK PESACH over 10 years ago in reply to Ramesh Ramamoorthy

Intellectual 850 points

i inferred 011 from the pwm outputs as i see 2 high signals in the same duty cycle and another one with lower duty cycle , of course the ratio is changing as i decrease \ increase the Id_ref alignment current , but the 2 pwms are always in same duty and are higher then the third . i am pretty sure that this is the high outputs (it now came to me that this could be critical..) but i cant check it right now , promise to check it as soon as i can.
but still , isnt the Qs and Ds values supposed to be different from what i get?

0 Ramesh Ramamoorthy over 10 years ago in reply to MARK PESACH

TI__Mastermind 27240 points

did you check the waveform on the HIGH side IGBT or LOW side?

0 MARK PESACH over 10 years ago in reply to Ramesh Ramamoorthy

Intellectual 850 points

Hi Ramesh

yes i checked the pwm wave forms on the HIGH side of the Mosfets .

any way i made the following changes and now its seems to work very good:

EPwm6Regs.CMPA.bit.CMPA = (INV_PWM_HALF_TBPRD*svgen1.Ta)+INV_PWM_HALF_TBPRD;

EPwm7Regs.CMPA.bit.CMPA = (INV_PWM_HALF_TBPRD*svgen1.Tb)+INV_PWM_HALF_TBPRD;

EPwm8Regs.CMPA.bit.CMPA = (INV_PWM_HALF_TBPRD*svgen1.Tc)+INV_PWM_HALF_TBPRD;

also i changed the sign of my measured currents Ia and Ib .

thanks

0 Ramesh Ramamoorthy over 10 years ago in reply to MARK PESACH

TI__Mastermind 27240 points

Mark,

Glad to know you figured out the matter.

0 Ramesh Ramamoorthy over 10 years ago in reply to Ramesh Ramamoorthy

TI__Mastermind 27240 points

Mark,
The prev post got posted by mistake before completion. Would like to know if you are testing it out an IDDK board or your custom board?

rgds,
ramesh

0 MARK PESACH over 10 years ago in reply to Ramesh Ramamoorthy

Intellectual 850 points

sorry i didnt mentioned it , i am not using the IDDK board .

just delfino launchpad with DRV8301 booster pack.

0 MARK PESACH over 10 years ago in reply to Ramesh Ramamoorthy

Intellectual 850 points

sorry i didnt mentioned it , i am not using the IDDK board .

just delfino launchpad with DRV8301 booster pack.

C2000™︎ microcontrollers

C2000 microcontrollers forum

FOC alignment routine