• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 31 subscribers
  • Views 1767 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.

AWR1642: Questions about the EDMA configuration

Ming Yang
Prodigy 120 points
Part Number: AWR1642

Hi:

There are some questions about the function of MmwDemo_dataPathConfigEdma in the demo->mmw_dss_16xx->dss_data_path code confused me.

1. LINE 1381: Why dose the srcBIdx is (obj->numAdcSamples * BYTES_PER_SAMP_1D * 2) * obj->numChirpsPerChirpEvent.

I think the srcBIdx should be numAdcSamples * BYTES_PER_SAMP_1D, because the acount of an array is numAdcSamples * BYTES_PER_SAMP_1D, so the beginning of the N array to the         N+1 array is numAdcSamples * BYTES_PER_SAMP_1D.

2. What's the difference between the numAdcSamples and numRangeBins?

I do not know if  the equation of numRangeBins=numAdcSamples *numChirpsPerChirpEvent is right.

3. What's the difference between the numChirpsPerChirpEvent and numChirpsPerFrame?

I think they are equal. But it may be wrong.

4.Is there a detailed flow chart of the signal preocessing for this demo?

 

THANK YOU !

 

  • 0
    TI__Guru 50240 points
    Hi Ming,

    Someone should have an answer for you tomorrow.


    Cheers,
    Akash
  • 0 in reply to Akash Gondalia
    TI__Guru*** 148334 points
    Hi,

    Yes there is detailed documentation abou the mmWave SDK demo. There are also many threads related to the demo.

    The documentation is located in

    C:\ti\mmwave_sdk_02_00_00_03\packages\ti\demo\xwr16xx\mmw\docs


    Please search the forum as well
  • 0 in reply to Cesar
    Prodigy 120 points

    Hi,

            I have found the detailed documentation, but there are still some problems make me confused.

            I have some questions about  the diagram of 1st Dimension FFT Processing as the fig.1 shown,

            1. Does the ADC Buffer Ping store the chirps of TX1 antenna while the Pong stores the chirps of TX2 antenna as the fig.2 shown?

             2. Dose the diagram only shows the processing of chirps of Ping?

            2. If the answer of questio1 and 2  are right, why are there chirps of TX2 in the L3:radarcube, and where are the odd chirps of TX1?


    fig.1

                                                                                                                                                                                        fig.2

  • 0 in reply to Ming Yang
    TI__Guru*** 148334 points
    The diagram is not correct.

    Please check mmWave SDK 2.0 documentation for updated diagram. The Pong grey line should be directed to Pong ADC buffer.

    TX1 and TX2 chrips are processed alternatively

    After processing, the output of L1 FFT is stored in L3 for both TX1 and TX2 chirps


    C:\ti\mmwave_sdk_02_00_00_03\packages\ti\demo\xwr16xx\mmw\docs\doxygen\html\index.html

    Thank you
    Cesar
  • 0 in reply to Ming Yang
    TI__Expert 5235 points
    In the diagram you quoted, the pong arrow needs to point to the yellow boxes labelled 2 corresponding to chirp1 rx ant2, the change that was done in 2.0 release that Cesar quoted is also wrong (we misapplied the correction) and we will correct in next release. Basically the ping and pong processing on the input side (adcDataIn) happens on even (0,2,..) and odd (1,3,..) receive antennas of the same physical chirp (which is from a single tx antenna). The processed chirp result is being collected in fft1Dout i.e while adcDataIn is storing only one antenna worth of the chirp, the fft1Dout is storing full (all rx antennas) chirp before it is stored out into the radar cube. The fft1Dout buffer's ping as shown in the picture corresponds to even chirp (tx1) and pong corresponds to odd chirp (tx2). The storage in radar cube is organized as all chirps corresponding to same transmit antenna are contiguous in memory as clearly indicated by the labels on the left of the boxes (text written in vertical orientation).

    The input samples in ADCdataBuf arrive in the format indicated in the picture. The example shows a case where multi-chirp threshold is set to 2 (which translates to numChirpsPerChirpEvent) i.e there are two chirps being collected in ADCdatabuf before we get an interrupt from ADCBUF to start processing. The purpose of multi-chirping is explained in e2e.ti.com/.../681025.