Space Vector PWM and its Implementation 1

[Grollor33]

Hi all,

Figured the VSD guys would have some good answers for me...

I am using a component level time domain analysis suite as a simulation package to verify some of my customer's submittals (done in Matlab). I'm only after the correct top level behavior, not anything too specific. My main issue is with implementation of their control algorithm (or a clone of it).

The design has three-phase three-level inverters using Space vector PWM (with a common mode voltage cancellation scheme). After much digging and reading about SVPWM, I found many good articles explaining what it is...but none that explain how to use it...a rotating vector in a "honeycomb" is fine in theory, but i'm just missing how to implement the algorithm in a microcontroller...or (what I'm really after) 'fake' it in a simulation...

Can anyone give me a 10 cent lecture on how to actually make this work or refer a textbook where I'll find it?


But now the real question:

Because I haven't yet been able to implement the space vector approach, I figured I could get a ROM approxmiation of my inverter behavior using (easier) sinusoidal PWM - the old fashioned sine-triangle. There's a very good IEEE paper called "Multi-Level Inverter Modulation Schemes to Eliminate Common Mode Voltage" which outlines both types of control. I'm comfortable implementing the 'classical' SPWM for a three-level inverter...but implementing a carrier-based PWM scheme to eliminate common mode voltage (i.e. only use the switching states where the instananeous voltages of my inverter bridge phases add to zero) is giving me heartburn. Unfortunately, its not explained very well in the paper...yet another roadblock!

Has anyone used this carrier-based PWM method for common mode voltage elimination? Able to point to the theory behind it or to a resource that has a good explanation?

Thanks a lot!

 

[dpc]

Maybe here:

http://ww1.microchip.com/downloads/en/AppNotes/00955a.pdf

 

[Grollor33]

Thanks dpc,

I had actually read that one a while back...It is indeed one of the better top-level descriptions of the SVPWM...Its tough to come by something that actually has swithing waveforms...

Another good reference I've found is "Space-Vector PWM Voltage Control with Optimized Switching Strategy" by Stefanovic and Vukosavic.

 

[sreid]

I think the fundamental feature of SVM is that one of the six transistors is fully on at any given time, i.e., one of the six transistors is either sourcing or sinking all the current.

 

[sreid]

Take a look at

http://amazon.com/o/ASIN/0471208140...&pf_rd_t=101&pf_rd_p=265623401&pf_rd_i=507846

 

[Grollor33]

I paged through the book at Amazon.com...

Seems like a good fit! Should be arriving shortly.

Thanks!