• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 84 subscribers
  • Views 723 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.

TMS320F28035 IQmath limitations (2P2Z)

RR_DSP
Prodigy 30 points
Other Parts Discussed in Thread: CONTROLSUITE, TMS320F28032

Dear All,

We are facing problems related to dynamic range and resolution of 2P2Z controller. Using the standard 2P2Z controller we cannot reach enough dynamic range neither resolution, because we need a high dynamic range and high resolution at the same time, and it isn't possible using the standard solution. We tried to improve the dynamic range change from iq26 to iq19 but it was not enough to solve the dynamic range problem and brought low resolution problem (we need the iq26 resolution and iq16 dynamic range). Do you have any suggestion to implement this combination (iq26 resolution and iq16 dynamic range)? Do you have 2P2Z controller that works with 64 bit result of multiply operation instead of 32 bit result?

Thanks,

RR.

 

 

  • 0
    TI__Guru** 119670 points

    I'm curious what kind of application using 2P2Z needs this much dynamic range?

    Is there another way you could scale numbers in your system as you go?

    At the end of June we will have a controlSUITE release for an updated Power Library. Included in this will be a CLA version of the 2P2Z.  The CLA on F28035 is native single precision floating point, so this will solve your issue (24-bit resolution and almost unlimited dynamic range). You'll still need to interface this back to anything else in your system using IQ values.

     

     

  • 0 in reply to ChrisClearman
    Prodigy 30 points

    Thanks for your reply.

    We are developing an Half-Bridge DC-DC converter with high frequency (300KHz) and extremely low output ripple.

    Because of it that we need this combination.

    Regards.

     

  • 0 in reply to RR_DSP
    TI__Guru** 119670 points

    So do you think using the CLA will be a good fit?

  • 0 in reply to ChrisClearman
    Prodigy 30 points

    We will try to use TMS320F28032 without CLA in order to reduce the cost.

    If you find something about this subject, please send us.

    When TMS320F28032 will be available?

     

    Regards.

  • 0 in reply to ChrisClearman
    Prodigy 110 points

    ------------

     

  • 0
    Genius 5170 points

    Although it is not my area of work but I found your requirements very interesting .. basically your system requires the best of both worlds !!! fixed point accuracy and floating dynamic range all over the accuracy!!! You need iq26 resolution and therefore a single precision floating point implementation will not be enough (I believe a single precision float is kind of equavalent of iq24 from resolution point of view),  on the other hand you need a dynamic range of iq16!!!! Is this a completely new development? How did your previous products address this issue? If possible for you to share then I am interested to know the compromises you made in previous products/implementations.