CCS/IWR6843ISK: Capturing Data from 2 TX Antennas

HI, 

You mentioned that "The spectra of the virtual antennas 1-4 are identical to the spectra of the virtual antennas 5-8".   Instead of looking at the spectra, can you look at the raw ADC data, and see whether TX1-RX1 and TX2-RX1 data are on top of each other?

Or you can program to use TX1 only and program TX2 only to see whether your raw ADC data kind of matching your two set of ADC data with the TDM-MIMO configure.

chirpCfg 0 0 0 0 0 0 0 1
frameCfg 0 0 1 0 50 1 0

chirpCfg 0 0 0 0 0 0 0 4
frameCfg 0 0 1 0 50 1 0

Best,

Zigang

Hi Zigang,

the comparison of the raw data shows the same result. The ADC samples of virtual antenna 1-4 are identical to the samples of virtual antenna 5-8.

When I use TX1 and TX3 separately it seems to work correct.

Best regards

Lukas

Hi, Lukas:

Can you check in your software, when you copy the raw ADC data to the L3, have you overwrite them, or you have copy them to the right memory offset. 

Best,

Zigang

Hi Zigang,

I am using the DPC_ObjectDetection_execute function of the mmwave_sdk_68xx_hwa out of the box demo for copying the data from ADCBUF to L3 memory. The code I implemented is as follows.

    uint8_t *ptrDest = (uint8_t*)(&gMmwL3[0] + (gChirpCount * subFrmObj->staticCfg.ADCBufData.dataProperty.numAdcSamples * subFrmObj->staticCfg.ADCBufData.dataProperty.numRxAntennas * 2 * 2));
    uint8_t* adcptr = subFrmObj->staticCfg.ADCBufData.data;

    for(i = 0; i < subFrmObj->staticCfg.ADCBufData.dataProperty.numAdcSamples * subFrmObj->staticCfg.ADCBufData.dataProperty.numRxAntennas * 2 * 2; i++)
    {
        *ptrDest = *adcptr;
        adcptr++;
        ptrDest++;
    }

    gChirpCount++;
    if (gChirpCount * subFrmObj->staticCfg.ADCBufData.dataProperty.numAdcSamples * subFrmObj->staticCfg.ADCBufData.dataProperty.numRxAntennas * 2 * 2
        >= SOC_L3RAM_SIZE/2)
    {
        gChirpCount = 0;
    }

Besides I changed the memory location of the processed data so I can use the half  of the L3 memory for storing my raw data.

    /* Memory related config */
    objDetInitParams.L3ramCfg.addr = (void *)&gMmwL3[0x48000];
    objDetInitParams.L3ramCfg.size = sizeof(gMmwL3)/2;
    objDetInitParams.CoreLocalRamCfg.addr = &gDPC_ObjDetTCM[0];
    objDetInitParams.CoreLocalRamCfg.size = sizeof(gDPC_ObjDetTCM);

So as I understand the code copys the ADC raw data with size

numRxAntennas*numADCsamples*2(IQ)*2(2Byte per Sample)

for every chirp.

Best regards

Lukas

We can not debug and review the code for you.   

Just remember the data comes in the order of:

1st chirp of TX1,  1st chirp of TX3, 2nd chirp of TX1, 2nd chirp of TX3,

3rd chirp of TX1,  3rd chirp of TX3, 4th chirp of TX1, 4th chirp of TX3 

and total size of one chirp is:

4*numADCsamples*2(IQ)*2(2Byte per Sample)

Hope it helps,

Zigang

Sure, but can you please tell me if it is the right approach. The order of the data is clear to me but am I using the right function for saving the data? Older posts in this forum regarding raw data capturing recommend using the chirp interrupt event but I am using the newest SDK version with the data path chain structure so there is no chirp interrupt event. Can you please give a brief overview how to proceed?

Best regards

Lukas

Hi, Lukas:

I am not familiar with this part of the code.   Since you were able to get one TX data capture right, 

and TX1 data and TX3 data chirp are coming at different time, I am guessing maybe you are re-write the 

second chirp (TX3 data) to the same location of first chirp (TX1 chirp).    If I were you, I will make a debug with a breakpoint to check whether the destination address are offset correctly for the first and second chirp.

Best,

Zigang