• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 259 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.

IWRL1432BOOST: 4K FFT Implementation - numOfAdcSamples Configuration Inquiry

JAEHOON KIM
Intellectual 405 points
Part Number: IWRL1432BOOST

Tool/software:

Dear TEXAS INSTRUMENTS

This is a follow-up question related to the URL below:

I am currently working on implementing a 4K FFT using the high_accuracy_xwrl14xx project, referencing the document below:

I have a question regarding the numOfAdcSamples parameter.

In order to implement a 4K FFT with the high_accuracy_xwrl14xx project, should the numOfAdcSamples value be set to 1024 or 4096?

I would appreciate your technical support and clarification on this matter.

Thank you.

  • 0
    TI__Mastermind 19550 points

    Hello.

    We are looking into this and will provide a response within a few days.  Thank you for your patience.

    Sincerely,

    Santosh

  • 0
    TI__Expert 4180 points

    Hi,

    In the demo, we only support a maximum of 2048 ADC samples. Hence you can do a maximum of 2K point FFT with 2048 ADC samples.

    Regards

  • 0 in reply to Sharan S S
    Intellectual 405 points

    Dear [Sharan S S],

    Hello, I would like to ask a few technical questions regarding the IWRL1432BOOST board:

    1. Is it possible to perform a 4K-point FFT by applying zero-padding to 1024 or 2048 samples?

    2. Are there any architectural limitations that prevent the IWRL1432BOOST board from performing a 4K-point FFT using the Hardware Accelerator (HWA)?

    3. Could you recommend any TI evaluation boards (EVBs) operating in the 76–81 GHz band that support 4K-point FFT processing?

    I would appreciate your guidance on these points.

    Best regards,
    Jaehoon Kim

  • 0 in reply to JAEHOON KIM
    Intellectual 405 points

    Dear [Sharan S S],

    I have modified the 2K point FFT parameters based on the 4K point FFT reference. I would appreciate it if you could review the changes at your earliest convenience.​

    Thank you for your time and support.

    Best regards,

    [JAEHOON KIM]

  • 0 in reply to JAEHOON KIM
    Intellectual 405 points

    Hello,

    In the document 'swru526b.pdf', Section 2.4 "FFT Stitching Use-Case Example" (specifically Table 19 and Table 20) describes an example based on directly using the APIs for HWA to implement FFT Stitching.

    However, the 'high_accuracy_xwrl14xx' project for the IWRL1432BOOST board is designed using the RangeProc DPU and external data structures, where the HWA configuration is abstracted. Due to this abstraction, it is difficult to directly control the HWA at the API level within this project structure.

    Because of this difference, I would like to request guidance on how to effectively apply the FFT Stitching example described in 'swru526b.pdf' to the 'high_accuracy_xwrl14xx' project.

    I would appreciate any advice or recommendations from the TI support team.

  • 0 in reply to JAEHOON KIM
    TI__Expert 4180 points

    Hi,

    Integrating into level sensing demo would require a lot of driver changes. 

    You can start with modifying the paramsets in HWA driver example and take it from there.

    Regards

  • 0 in reply to Sharan S S
    Intellectual 405 points

    Hello,

    I still have a few unresolved questions regarding the project.

    1. In the 'high_accuracy_xwrl14xx' project for the IWRL1432BOOST, is it possible to configure the ADC sample size to 1024 and then apply zero-padding to perform a 1D FFT with a size of 2K or 4K?
      If this approach is structurally not feasible, we are considering selecting a different EVM board for continued development.

  • +1 in reply to JAEHOON KIM
    TI__Expert 4180 points

    Hello,

    Yes, it is possible to do a 2K FFT even without zero padding as we support a maximum of 2048 adc samples. But you might be required to change a lot in paramsets, dpu etc to get it working. For 4K FFT, yes you can zero-pad, but again it would require a lot of changes in the SW to get it working. Purely from a technical standpoint, yes it is possible.

    Regards