I am trying to control a BLDC motor using space vector modulation using the attached document. I am making a lot of effort to understand the document. I think I came to certain level to understand but struck up to go further. I am not able to understand the space vector modulation. UOUT is the desired resultant vector which we want to represent using two adjacent vectors U0 and U60.


The values for T1 and T2 are taken from a look up table containing 172 fractional sinusoidal values for 0 to 60 electrical degrees. The 172 fractional values are arrived as (1024*60/360) = 172.

I really don't understand the below statement
As a result, for each of the six segments one axis is exactly opposite that segment, and the other two axes symmetrically bound the segment.
The values of the vector components along those two bounding axis are equal to T1 and T2.

And then the Fig14. Whether this figure is for only UOUT vector. Why T1/2 and T2/2 are equally spaced. I mean to say their timings are same. In case if I take another vector say UOUT1 which is closer to U60 will the timings of T1/2 and T2/2 will be different? The PWM for period will keep varying for each resultant (UOUT vectors)? Thank you in advance.

 

I think this is a lot of Krapity Kraps with fancy words to say something much simpler. In this case, what they mean is that whenever you switch to one of the 60 degree segments, you are in essence switching one phase and leaving the other two the same. For example, U0 is basically 001 which if you look at phase information as in CBA, it means C = 0, B = 0 and A = 1. Then you go to segment U60 which is 011. In this case, C = 0, B = 1 and A = 1. Notice only one phase output changed, in this case B. C and A stayed the same. If you look at the other 4 segments you will notice that only one output change at a time, whereas the other two remain the same. This is intrinsic to 3 phase BLDC systems, but their wording here is super hard to follow!

I too would love to understand SVM theory but to be honest all I do is a lookup table from 0 to 360 where the first 120 degrees are a sine wave from 0 to 119, the second 120 degrees (120 to 239) is a sine wave from 119 to 0 and then the last 120 degrees (240 to 359) are just zero. With this you get something like a sine wave commutation output.