• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 83 subscribers
  • Views 397 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.

On digital PI control using f2808

billstrong
Intellectual 390 points

Hi all,

what kind of problems will occur when change from analog PI control to digital PI control?

Here is my situation:

1.I have recently using the analog PI control to control the current, the analog PI output is sampled via the ADC inside f2808 and the result is very satisfying, current ripple only 0.7A out of 10A

2.Based on the analog control scheme, I used the digital PI controller from TI and the current ripple increases to 1.4A out of 10A

3.I reduced the system's bandwidth trying to bring the current ripple down but without success

3.Since the current ripple limitation has to be below 1A, can anyone share your experience when using digital PI control before? Will the performance really be reduced when changed to digital control?

Thanks!

Bill

  • 0
    TI__Mastermind 43730 points

    What's the application you are using this for?

    I am assuming Digital Power, what topology? Buck, Boost?

    What is the switching frequency? and what frequency (for the digital implementation) are you running your control loop at?

  • 0 in reply to Manish Bhardwaj
    Intellectual 390 points

    Hi,

    the topology is buck with switching frequency of 40KHz

    since there are only few things to be done in the control loop it runs every 2.5us, means duty cycle is updated every 2.5us

    the epwm update scheme is configured to be update twice in a period with symetrical up down counting mode, and the CMP register is configured to be shadow mode

    I really wonder why this configuration is so successful for analog PI control but not so really for digital PI control, I am sure the control scheme from TI is correct!

    Thanks for sharing your experience!

    Bill


     

  • 0 in reply to billstrong
    Genius 5170 points

    May be I am missing the setup but if you are tied with PWM to update compare values then with 40kHz PWM you should only be able to update maximum at 80kHz rate assuming you are updating both on period and on zero with an up/down counting mode. You can calculate at a faster rate (in this case you are calculating at 2.5uS rate/400kHz!!) but will not be able to update the compare values if you use shadow mode (which you should)!! May be that is causing the excess error .... in general 40KHz operation is not fast switching for these controllers and your error should be less than 10% ...

  • 0 in reply to Arefeen
    Intellectual 390 points

    Hi,

    Attached is my control scheme

    I really wonder why with the same switching frequency and update mode, it would be so successful for the analog PI control scheme while for the digital PI control, its performance will deteriorate?

    thanks for your suggestions!

    Bill

    Control Scheme.pdf
  • 0 in reply to billstrong
    TI__Mastermind 43730 points

    Bill,

    are you making sure that you sample at the mid point of the switching cycle to measure the inductor current of the buck ?

    Also you should not update the CMPA value any faster than 40Khz i.e. switching frequency..

    Also make sure that you subtract the offset current when the measured current is fed back to the the PI, 

    How is your current sense configured? are you measuring the switch leg current or the inductor current directly?

    Regards

    Manish Bhardwaj