• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 82 subscribers
  • Views 903 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

SVM gain

Chuck Sampson
Genius 3295 points


Hello,

I am new to motor controls. I am working through the Insta-Spin FOC labs. They are very nice and all the hardware and software is working very well. This is a great learning tool. I have question about Lab 5a. This lab gives a step by step analysis of the PI gains and how to determine them for your motor loop gain. I noticed in the block diagram in Lab 2a the output of the PI block goes to the INV park and then space vector modulator and then to the motor -the way I would expect. However in lab 5a, the output of the PI block produces a motor voltage and skips the SVM gain. Does anyone understand why the SVM gain is omitted? Is it not needed? I know in DC DCs and in sine wave inverters the inverter gain is required to determine the loop gain. Is this not true for SVM.

thanks,

Chuck 

  • 0
    TI__Guru** 119670 points
    chuck,
    "in lab 5a, the output of the PI block produces a motor voltage and skips the SVM gain. "

    why do you think this? perhaps a figure is incorrect. No, the output of the INVPARK goes to the SVM (with 1.0 modulation limit) which then produces the Ta,b,c duty cycles.
  • 0 in reply to ChrisClearman
    Genius 3295 points
    Chris,

    Thanks for the quick response. I agree the block diagram in lab 2a shows exactly what you are saying. However in Lab 5a, where you have to determine the PI gains for the current loop, there is a block diagram shown in Figure 18, Page 76 that shows the PI output going directly to the motor input. The inverter modulator gain is omitted. This may be correct as far as I know. Although I am new to motor controls, I have experience in DC DC and single phase AC inverter design. In both instances, you have to include the modulator gain in the loop response. Maybe that does not apply here or the modulator gain is trivial, i.e., close 1. For instance in a single phase AC inverter design using a sine reference, the modulator gain is equal to 2 VDC. The modulator gain is from the PI controller output to the load admittance. I have read up on the SVM in "Digital Control in Power Electronics" by Buso and Mattavelli. Excellent reference by the way. And they say that the SVM algorithm de-couples the phases so that the per phase analysis applies. Do you agree? Or does this not apply with a motor load?

    Chuck
  • 0 in reply to Chuck Sampson
    TI__Guru** 119670 points
    figure 18 is just showing the concept of a PI control of a plant. This isn't a representation of the software flow, just a general concept of PI control. I agree they should have made it more generic or more specific (actual SW blocks).
  • 0 in reply to ChrisClearman
    TI__Expert 4760 points
    Hi Chuck,
    I am the originator of the diagram and write-up on the PI tuning process. The bottom line is that this section is outdated and needs to be rewritten.
    I originally wrote a procedure for how to tune PI loops in general. In this writeup, the translation between the PI outputs and the motor voltage was treated as a simple generic gain factor. This writeup was then included in the user's manual and lab exercises for InstaSPIN.
    After I wrote this, I realized that the writeup did not include the the velocity signal filter which is used with most control systems. Even in InstaSPIN, the velocity signal is filtered before being supplied to the velocity loop as the velocity feedback signal. So I rewrote the entire procedure and posted it to our Motor Control blog site:
    e2e.ti.com/.../motor
    This is a 10 part series which goes into a lot more detail about tuning PI loops, both current and velocity. It even includes text on how to apply this procedure to InstaSPIN. But even in this series, the gain of the SVM block, PWM block, and bus voltage are treated as a generic gain. However, the lab exercises and user's manual were never updated with this revised procedure.
    Please read the updated procedure on our Motor Control blog site, and then let me know if you have any questions.
    Thanks,Dave
  • 0 in reply to Dave (Wisconsin) Wilson
    Genius 3295 points
    Thanks for the information Dave, I will check out the procedure on the Motor Control blog as you suggest.