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.

IWR1843BOOST: AWR1843AOPEVM:Azimuth FFT size: n-point FFT

Part Number: IWR1843BOOST

Hello Forum,

I wanted correct my understanding,To find N-point FFT for any given lab,the value of n depends on adc sample.If adc sample equals 64 then my value of n is 2 power 6 so N=6.

Is this logic correct to find Azimuth FFT size 

Does the same logic apply for elevation FFT size.

Any refernce to this undertanding is aperciated

Warm regards

Ayush Shetty

  • HI, Ayush:

    Which parameter you are talking about here?  Is a input configuration parameter or internal parameter? can you be more specific?

    Best,

    Zigang

  • We give Number of ADC sample as parmater in the radar configuration.

  • You can share your understanding to find the N ,for N-point FFT for Azimuth and Elevation.

  • Hi, Ayush:

    In the range FFT, FFT size is the power 2 round up to the number of ADC samples.  For example if numADCSample = 64, then rangeFFTSize is also 64.  But if numADCSample = 50, then rangeFFTSize is also 64. 

    params->rangeFftSize = mathUtils_pow2roundup(params->ADCBufData.dataProperty.numAdcSamples);

    The same rule applied to Doppler FFT, angle FFT. 

    Regarding n = 6, probably you are talking about FFT engine inside HWA, the fftsize inside hwaParamCfg is in log2 format. 

    hwaParamCfg.accelModeArgs.fftMode.fftSize = mathUtils_ceilLog2(pDPParams->rangeFftSize);

    Best,

    Zigang

  • Hello Zigang,

    This gives a lot of clarity,One last question,What is the maximum FFT size limit for TI radars in general?

    Warm regards

    Ayush Shetty

  • HI, there:

    In the FFT engine inside radar HWA, the FFT sizes supported are all powers of 2 until 1024, for example, FFT sizes of 2, 4, 8, 16, … 512 and 1024 are supported.  

    Note that FFT stitching is a feature that enables large FFT sizes, specifically, 2048 and 4096, using a two- step process.

    Best,

    Zigang