Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed from November 25 through December 2. Thank you for your patience.

AWR2544LOPEVM: AWR2544 : reduce number of RX channel using MMwave Demo

Part Number: AWR2544LOPEVM
Other Parts Discussed in Thread: AWR2544

Tool/software:

Dear TI experts, 

For a particular test, I am creating a wave form with only 1 tx and Rx channel. I want to to have a 2048 2x real sample on this RX channel, meaning end up with 1024 range gates. As I am reducing by 4 the number of Rx channel and increasing by 4 number of samples per chirps I was expected no L3 ram issues. How ever, when trying this I see that gMMwL3Ram is kind of frozen only few header values are changed (see picture attached).

I have searched through the code, however I am a bit lost. 

Do you have any advice to make it work please ?

Thank you.

Regards , 

Jean-François

  • Hi Jean,

    Do you mean the radarcube is not updating? Can you halt the core to see if there is any error?

    Regards,

    Samhitha

  • Hi Samhitha, 

    I mean when I am trying to use below configuration for 1 RX / 1 TX , radarcube in L3 seems not updated as it does when using "profile_3d_3Azim_1ElevTx_TDM_awr2544_2Xmode.cfg" provided config file. It seems frozen. 

        "sensorStop \r\n",
        "flushCfg \r\n",
        "dfeDataOutputMode 1 \r\n",
        "channelCfg 1 1 0 0 0 \r\n",
        "adcCfg 2 0 \r\n",
        "adcbufCfg -1 1 0 1 1 \r\n",
        "lowPower 0 0 \r\n",
        "profileCfg 0 77 3.5 3.5 60 0 0 19.53125 0 2048 40000 0 0 164 \r\n",
        "chirpCfg 0 0 0 0 0 0 0 1 \r\n",
        "chirpCfg 1 1 0 0 0 0 0 1 \r\n",
        "frameCfg 0 1 128 0 2048 50 1 0 \r\n",
        "compressionCfg -1 1 0 0.5 8 \r\n",
        "intfMitigCfg -1 15 18 \r\n",
        "procChainCfg 0 1 200 0 1 \r\n",
        "adcDataDitherCfg 0 0 1 \r\n",
        "analogMonitor 0 0 0 \r\n",
        "calibData 0 0 0 \r\n",
        "sensorStart \r\n",

    What do you mean by "halt the core"? As for now it seems that there are no issue as udp packet are sent to PC and are the image of "frozen" L3 radar cube.

    Thank you.

    Regards, 

    Jean-François

  • Hi Jean,

    What do you mean by "halt the core

    I meant to halt the execution of the core by clicking on the following button in CCS. Can you please halt the core and check if there is any error?

    As for now it seems that there are no issue as udp packet are sent to PC and are the image of "frozen" L3 radar cube.

    Since you are receiving the UDP packets, there might not be any assert. Can you select continuous refresh in the memory browser toolbar to check if the data is updating? 7.1. Debug Overview — Code Composer Studio 10.1.0 Documentation

    I will check if enabling 1Tx, 1Rx is supported in the demo and get back to you on Tuesday.

    Regards,

    Samhitha

  • Hi Samhitha, 

    I looked deeper in the code and it seems that everything is fine until the EGE data compression part. I mean all data are leaving until this part.

    For example from ADC buffer ( adress 0xa5000000) to FFT out (ping adress 0x82000000) I see data living when refreshing the memory.

    As we can see in below picture:

    But for compression out (ping adress out = 0x82008000), data are frozen as shown below:

    Would you have an EGE compression setting suggestiion for 1 RX / 1 TX? Or maybe how change edma settings to make it work?

    Thank you!

    Kind regards, 

    Jean-François

  • Hi Jean,

    I want to to have a 2048 2x real sample on this RX channel, meaning end up with 1024 range gates.

    I am assuming that you are referring to 2x mode in HWA FFT engine. 

    Would you have an EGE compression setting suggestiion for 1 RX / 1 TX? Or maybe how change edma settings to make it work?

    Additional changes are not required on top of the SDK demo. Can you try enabling 1Tx and 1Rx in the configuration provided in the SDK and check if the radar cube is updating?

    Regards,

    Samhitha

  • Hi Samitha,

    I have spent some time trying to figure out how to make this 1 tx / 1 rx configuration working. 

    What I observe is that edma transfer from ADC buffer  (adress 0xa5000000) to HWA buffer is working well. But then something is not working in HWA.

    For example for we see data living in address 0x82000000 but no data is living (data freezed)  in adress 0x82008000. As if DC est and/or DC sub can not manage this configuration. 

    At this end, radar cube is not updated at all ...

    Here is the configuration I am trying to use:

    "sensorStop \r\n",
    "flushCfg \r\n",
    "dfeDataOutputMode 1 \r\n",
    "channelCfg 1 1 0 0 0 \r\n",
    "adcCfg 2 0 \r\n",
    "adcbufCfg -1 1 0 1 1 \r\n",
    "lowPower 0 0 \r\n",
    "profileCfg 0 77 3.5 3.5 60 0 0 19.53125 0 2048 40000 0 0 164 \r\n",
    "chirpCfg 0 0 0 0 0 0 0 1 \r\n",
    "chirpCfg 1 1 0 0 0 0 0 1 \r\n",
    "chirpCfg 2 2 0 0 0 0 0 1 \r\n",
    "chirpCfg 3 3 0 0 0 0 0 1 \r\n",
    "frameCfg 0 3 32 0 2048 50 1 0 \r\n",
    "compressionCfg -1 1 0 0.5 8 \r\n",
    "intfMitigCfg -1 15 18 \r\n",
    "procChainCfg 0 1 200 0 1 \r\n",
    "adcDataDitherCfg 0 0 1 \r\n",
    "analogMonitor 0 0 0 \r\n",
    "calibData 0 0 0 \r\n",
    "sensorStart \r\n",
    Do you have any idea ? Or hint to debug HWA issue ?
    Many thanks,
    Jean-François Oberle