AWR1843BOOST: Disable doppler processing

Hi Yaron,

Are you running the AWR1843 SDK demo application? What are the high level changes you have made to perform Range Processing only?

Regards,

Samhitha

Hi Samhitha
Yes, I am running the out-of-the-box demo, I made changes in the CLI (added some commands), changed the high speed UART frame structure, and made some calculations over the radarCube 1FFT results. This all works fine.
Now I am facing two issues:
- I would like to disable the processing done after the 1FFT (doppler, AoA) since I don't need it and would like to reduce the minimal frame rate
- Find a way to get both 1FFT and log amplitude on 1FFT results (not on doppler) fron the HWA
Any help will be appreciated
Regards
Yaron

Hi Samhitha
Thank you for your reply
Your suggestion for the first question works fine.
As to the second, I am still trying to figure out how to execute a second HWA process just after the 1st FFT so it can use the same ADC input and generate the log output in a different location in the memory. If you happen to have examples for that it will be helpful
Best regards
Yaron

Hi
I followed the above instructions:
Increased DPU_RANGEPROCHWA_NUM_HWA_PARAM_SETS to 6
Added this code in rangeprochwa.c after the section that handles the ping (line 683) and similar (with the needed modifications) after the code that handles the pong :

paramsetIdx++;
    hwParamsetIdx++;
    hwaParamCfg[paramsetIdx] = hwaParamCfg[pingParamSetIdx];
    hwaParamCfg[paramsetIdx].accelModeArgs.fftMode.magLogEn = HWA_FFT_MODE_MAGNITUDE_LOG2_ENABLED;
    hwaParamCfg[paramsetIdx].dest.dstAddr = hwaMemDestPongOffset;//+128*4;
    hwaParamCfg[paramsetIdx].triggerMode = HWA_TRIG_MODE_IMMEDIATE;

    errCode = HWA_configParamSet(hwaHandle,
        hwParamsetIdx,
        &hwaParamCfg[paramsetIdx], NULL);
    if (errCode != 0)
    {
        goto exit;
    }

I understand that using hte same des address will make it override the fft data - but jest trying to see that it will run.

I am getting exception on next frame processing -   DPC_Objdet_Assert(objDetObj->dpmHandle, (objDetObj->interSubFrameProcToken == 0));

What am I doing wrong?
Thanks
Yaron

Hi Yaron,

Let me check your response and get back to you by Wednesday.

Regards,

Samhitha

Hi Samhitha
This is not the case, I am using a very long frame (40msec) during this development, and the chrips takes only 270usec in total, I am not getting this eror unless I try to add the two extra stages in the HWA (so I can get both the FFT1 and log magnitude
Best regards
Yaron

Hi Yaron,

Can you trigger a single frame instead of infinite frames? This will help in understanding if the paramsets are executed successfully for the first frame or not. 

Samhitha Bobba said:

I suggest you to create a unit test and verify the paramsets and then integrate in the demo.

Did you create a unit test and verify if the paramsets are configured properly?

Regards,

Samhitha

Sure, can you please advise how to trigger a single frame?

Hi Yaron,

Change number of frames parameter in frameCfg CLI command to 1.

"frameCfg 0 1 32 1 100 1 0"

Refer to "C:\ti\mmwave_sdk_03_06_02_00-LTS\docs\mmwave_sdk_user_guide.pdf" for detailed description of each CLI command.

Regards,

Samhitha

Hi
Setting the frame parameter to 1 eliminated the exception but the application stooped responding to CLI (until power cycle).
So:
With the original code (4 steps in HWA) - works with single frame and multiple frames.
With the additional two HWA sets (total of 6), if frame set to 1 - get stuck, if frame is continuous get an exception:

Texas Instruments XDS110 USB Debug Probe_0/C674X_0 (Suspended)    
    abort() at exit.c:99 0x007FF160    
    xdc_runtime_Error_policyDefault__E(struct xdc_runtime_Error_Block , unsigned short, char , int, unsigned int, int, int)() at Gate.c:34 0x007EE34C    
    xdc_runtime_Assert_raise__I(unsigned short, char *, int, unsigned int)() at Error.c:119 0x007F9818    
    _DebugP_assert(int, char *, int)() at DebugP_tirtos.c:52 0x007FF076    
    DPC_ObjectDetection_frameStart(void *)() at objectdetection.c:435 0x007F8458    
    SOC_SystemISR(unsigned int)() at soc.c:426 0x007FA5F4    
    ti_sysbios_family_c64p_EventCombiner_dispatch__E(unsigned int)() at EventCombiner.c:202 0x007ECB74    
    ti_sysbios_family_c64p_Hwi_dispatchCore__I(int)() at Hwi.c:809 0x007F9150    
    ti_sysbios_family_c64p_Hwi_switchAndDispatch__I() at Hwi_asm_switch.s62:101 0x007FD314  (ti_sysbios_family_c64p_Hwi_switchAndDispatch__I does not contain frame information)   

line 435 in objectdetection.c is :

 / Check if previous frame (sub-frame) processing has completed /
    DPC_Objdet_Assert(objDetObj->dpmHandle, (objDetObj->interSubFrameProcToken == 0));
    objDetObj->interSubFrameProcToken++;

Hi
I didn't change the DMA settings since as a start all I want is for this not to crash so I don't mind that the additional stage will override the previous stage results
If I got you right it should be:
    /* Prepare internal hardware resouces = trigger source matchs its  paramset index */
    rangeProcObj->dataInTrigger[0]      = 2U + pConfigIn->hwRes.hwaCfg.paramSetStartIdx;
    rangeProcObj->dataInTrigger[1]      = 5U + pConfigIn->hwRes.hwaCfg.paramSetStartIdx;
    rangeProcObj->dataOutTrigger[0]     = 0U + pConfigIn->hwRes.hwaCfg.paramSetStartIdx;
    rangeProcObj->dataOutTrigger[1]     = 3U + pConfigIn->hwRes.hwaCfg.paramSetStartIdx;
(Since I added one step to Ping and one step to pong, overall 6 steps instead of 4)
Same error - checking if previous frame processing has completed
BTW - do you offer (or can endorse someone who does) - consultancy services for fixing a problem like this?
Best regards
Yaron

Hi Yaron,

Please find the corrections in the below screenshot -

Yaron Soffer said:

BTW - do you offer (or can endorse someone who does) - consultancy services for fixing a problem like this?

For queries related to the existing SW provided by TI, you can create an E2E ticket and an engineer will reply to you. We can provide some of our inputs but supporting custom changes is out of the scope of E2Es.

Regards,

Samhitha