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.

IWR6843AOP: Use full DSP data path chain on IWR6843AOP

Part Number: IWR6843AOP

Hi!

Thanks for taking time to look into my question.

I am trying to integrate the high accuracy demo into the out-of-box demo on IWR6843AOP. It seems that I need to use DSP to do the range FFT. However, the provided demos for IWR6843AOP all use HWA to do rang FFT. The only demo that uses full DSP DPC is for xWR1642.

In https://e2e.ti.com/support/sensors-group/sensors/f/sensors-forum/952466/ccs-awr1843boost-dpc-of-dsp Aayush described a quit complicated method to modify the code to run full DSP DPC on xwr18xx 2 years ago. But this is quit hard to follow.

I am wondering whether there is a more up to date instruction on how to implement full DSP DPC on IWR6843AOP.

Thanks,

piao

  • Hi,

    Can you provide a bit of background on what you are looking to achieve by merging the OOB with the High Accuracy Level Sensing demo? The processing chain used in the High Accuracy lab is a bit unique, so it may be more challenging than usual to find something that works well here. I'd like to understand your goals first to ensure that this is the easiest path to achieve them.

    Best Regards,
    Alec

  • Hi Alec,

    Thanks for looking into my question!

    I have tried the high accuray level sensing demo, it seems that it does not work well when the water surface is unstable,  e.g. when the water is flowing. It also lacks the ability to specify an angle to measure. In my application, I need to measure the water level that may not directly in front of the radar both when the water surface is stable and unstable.

    I am thinking that I can use the algorithm from OOB to find the distance of the water surface roughly, then I use the zoom FFT algorithm from high accuracy demo to process the input data again to find the percise distance. Thus, I wish to use rangeProcDSP to save the ADC buffer data before the range FFT is done, then use it for the later zoom FFT.

    Thanks,

    Piao

  • Piao,

    Understood. So this certainly require a bit of effort to make such a custom processing chain, but I will do my best to provide you some guidance to begin working on. 

    Firstly, I believe that even in the AOP OOB demo (which uses HWA for Range FFT), I believe that the results of the Range FFT are stored in shared memory. This means that you could likely make a DSS program for the DSP which grabs the RangeProc results which you are looking for from shared memory, and then do the zoomFFT all in the DSP. 

    If you still want to transition the RangeProc to the DSP, I would recommend that you take a look at the OOB for the xWR1642 device. This device does not have an HWA, and does the majority of its processing on the DSP, so you can likely look at its implementation to modify your AOP OOB accordingly. If you also need guidance on how to configure the DSS project to properly synchronize across both the MSS and DSS, I would recommend that you take a look at the OOB for the traditional xWR6843 device, which uses both cores (however does RangeProc on HWA).

    Take a look at some of these resources to get started and let me know what further questions that you have.

    Best Regards,

    Alec

  • Thanks for your advice!

    But I think high accuracy demo requires raw ADC buffer data to do zoom FFT, which is destroyed when HWA perform range FFT. So the result of range FFT from HWA may not help.

    I have gone through the xWR6843 demo and will go through xWR1642 demo. Thanks for your advise!

    Best,

    Piao