2 vs. 3 current sensors / OPA for InstaSPIN

Hos,

We added the 3 external opamps because:

• we wanted to allow a best 2 of 3 current sensing approach to enable 100% trapezoidal over-modulation. this increases the available DC bus / duty cycle for significantly higher speeds. If you only use 2 current sensors you will have to limit yourself to a maximum SVPWM modulation.
• it is usually best to balance these sense circuits, so we used three external. however, we have a new inverter board we'll release in a couple months using DRV8301 and in this case we made the trade-off and are using the 2 internal PGAs on DRV8301 and a single external OPA for the third phase.  This is working quite well so far.

Hi Chris,

I can't understand what you mean "Best 2 of 3" ?

what is the difference if using 2 or current sensors ?

anyhow they always need to sum to zero, so from 2 you can calculate the third...

"anyhow they always need to sum to zero, so from 2 you can calculate the third..."

AhHa! That's the gist.  Yes, your A+B+C=0, but can you ALWAYS get measurements of A and B?  Think about where and when you are measuring A and B in a bottom leg shunt system.

I'll let you ponder on that one for a bit... :)

Hi Chris,

Yeah, I see your point.. but if you're using LEMs (like I usually do), placed on the motor phases, then you don't need to worry about which FET is conducting... and then they must sum to zero (which can also be argued further...)

mojo,

if you are using LEMs then you don't have to worry about sampling windows through the shunt. You can use 2 phases (select this in your user.h) and you can still go into ovemodulation.

Hi Chris,

you said that you need to take into account which lower FET is conducting when taking the current measurement.

however, I looked into the code and could not see such a decision making.

could you please refer me to the code section where the decision about which phase current to use is being conducted ?

by the way, something from older versions - you're still configuring TZ6 in the PWM module, however it is not defined in the GPIOs (I'm using the latest ver 09). the other two-TZ2,3 are configured ok...

thanks

quick white board example

now why is that showing upside down? It shows rightside up on every other application I open it with on my PC.  Strange.

Regardint the TZ.

The trip zones are set per board per target in the drv.c file.  For DRV8301 and F2806x this is the set-up:

  // Configure Trip Mechanism for the Motor control software
  // -Cycle by cycle trip on CPU halt
  // -One shot fault trip zone
  // These trips need to be repeated for EPWM1 ,2 & 3
  for(cnt=0;cnt<3;cnt++)
    {
      PWM_enableTripZoneSrc(obj->pwmHandle[cnt],PWM_TripZoneSrc_CycleByCycle_TZ6_NOT);

      PWM_enableTripZoneSrc(obj->pwmHandle[cnt],PWM_TripZoneSrc_CycleByCycle_TZ3_NOT);

      PWM_enableTripZoneSrc(obj->pwmHandle[cnt],PWM_TripZoneSrc_CycleByCycle_TZ2_NOT);

      // What do we want the OST/CBC events to do?
      // TZA events can force EPWMxA
      // TZB events can force EPWMxB

      PWM_setTripZoneState_TZA(obj->pwmHandle[cnt],PWM_TripZoneState_EPWM_Low);
      PWM_setTripZoneState_TZB(obj->pwmHandle[cnt],PWM_TripZoneState_EPWM_Low);

      // Clear any spurious fault
      PWM_clearOneShotTrip(obj->pwmHandle[cnt]);
    }

and elswhere in the drv.c file we set-up the GPIO pins which are TZ2 and TZ3

  // OCTWn
  GPIO_setMode(obj->gpioHandle,GPIO_Number_13,GPIO_13_Mode_TZ2_NOT);

  // FAULTn
  GPIO_setMode(obj->gpioHandle,GPIO_Number_14,GPIO_14_Mode_TZ3_NOT);

TZ6 (as are TZ4 and TZ5) is an internal Trip Zone which does not require a GPIO set-up.