Learning SVM - direction of current

1

Hi Guys,
First off, great job for everyone working on the ODrive firmware. It’s written really clearly so that even I can understand most of it; certainly way better than what Microchip or TI has in their examples.

I’ve been studying the BLDCs and FOC for a couple of days now and I get how the math works. I have however encountered a problem with the direction of currents through the windings of the motor. The way the ODrive firmware is written, it seems the direction of currents is the opposite of what I have imagined them to be. The SVM function (utils.cpp) is the one which bothers me.

Let’s call the SVM function with a simple vector alpha = 0.2 and beta = 0.0. The vector is clearly in sector 1, true. The function gets this right. But then the function calculates the PWM duty cycles and returns this:

t1 = 0.4, t2 = 0.6, t3 = 0.6

With these values, the voltage at U will be lower than at V and W and thus, the current will be flowing “from” the winding center point, through the U winding and into the half bridge. Thus the current will have a negative sign. Shouldn’t it be the other way around? I always imagined that this is true:

  • The current is positive, if it is flowing from the half bridge into the winding to which it is connected
  • If alpha is positive and beta is zero (so theta = 0°), then this means positive current through the U winding and negative current through V and W (each one half)

I’ve drawn a simple image to demonstrate this. So to ask plainly, what am I getting wrong?

Thanks for any pointers,
Ivo

bldc
bldc957×881 44.4 KB

2 Likes
2

The convention we use and also what I picked up from standard FOC theory, is:
When alpha is positive (and say beta is 0), the voltage on phase A is higher, and the current wants to flow into phase A and out of B and C.
I think the confusion is from what t1, t2, t3 mean: they refer to when in the cycle that phase should go high. So if the number is small, it goes high earlier, so it stays higher for longer. So if phase A has a low timing value, it goes high early, and is hence higher for longer, so you get more phase A voltage.