AWR1642BOOST: AWR1642 raw data acquisition with DCA1000EVM with disabled tx channel

Edgar,

Can you please elaborate so more on the specific problem or error that you are encountering when attempting to use the DCA1000EVM with the AWR1642 to measure the noise floor?

I was able to run a complete capture and basic MATLAB analysis using my own DCA1000 EVM and AWR1642BOOST this afternoon. I did not run into any issues when I disabled all of the TX antennas.

Please upload your mmWave Studio log file to assist in this debug.

Regards,
Kyle

Thank you for your reply Kyle,

I am not using mmwave studio as I am working in a linux environment, I am just establishing a udp connection using matlab to get the data and I am successful when using "normal" configurations.

When using a configuration with the tx antennas disabled, there is no data acquisition between the DCA1000EVM and the awr1642 - I press the "CAPTURE_START" button but there is no data stream (no data received in the computer and the "DATA_TRAN_PRG" led does not blink).

In the awr1642 board I am using the sdk 2.0.0.4 and the changes I made to the demo source code are the following:

-> in mss_main.c, lines 1705 through 1707 commented:

if(memcmp((void *)&gMmwMssMCB.cfg.openCfg, (void *)&openCfg,
                          sizeof(MMWave_OpenCfg)) != 0)
        {
            /* Post event to CLI task that start failed so that CLI/GUI can be notified */
            //Event_post(gMmwMssMCB.eventHandleNotify, MMWDEMO_DSS_START_FAILED_EVT);

            //MmwDemo_mssAssert(0);*/
        }

-> in dss_main.c, lines 2245 through 2249 commented:

       /* Below configurations are to be reconfigured every sub-frame so do only for first one */
            if (subFrameIndx == 0)
            {
                retVal = MmwDemo_dssDataPathReconfig(dataPathObj);
                if (retVal < 0)
                {
                    return -1;
                }
            }
        }
        //else
        //{
            /* no valid profile found - assert! */
        //    MmwDemo_dssAssert(0);
        //}

With this changes made, I had no problems disabling the tx antenna in the configuration file. I did so simply by changing the last field in the commands "chirpCfg" to 0 but perhaps that is not the best way to do it.

Hello Kyle,

Yes the switch2 was already set this way:
SW2.1 ... SW2.5
LVDS_CAPT; ETH_STREAM; AR1642_MODE; RAW_MODE; HW_CONFIG

So I don't think that was the problem.

Regards,
Edgar

Hello Kyle,

Thank you for the suggestion, I managed to operate the boards with the mmwave studio, (however had to use the version 1.0.0.0 instead of the latest but it's ok). I managed to confirm that it is possible to send configurations to disable the tx channels, I wonder why it cannot be done with the binaries that are supplied with the SDK demo.

Regards,

Edgar

Kyle,

I am using the XA1642ABSABL ES2.0 with the SDK 2.0 binaries flashed (found in ~/ti/mmwave_sdk_02_00_00_04/packages/ti/demo/xwr16xx/mmw/xwr16xx_mmw_demo.bin) connected to the DCA1000EVM, with the configuration file with the following line introduced: "lvdsStreamCfg -1 0 1 0" that will enable the LVDS functionality. This method works without using mmWave Studio and I am able to receive udp packets with raw data through the ethernet connection and so far so good.
My problem lies with the fact that when I disable the TX antennas, this firmware (SDK 2.0.0.4 Demo) does not accepted that kind of changes. Now are you telling me that the firmware demo version that comes with the SDK 2.1 will allow TX antenna disable natively without any problem?

Regards,

Edgar

Edgar,

How are you disabling the TX antennas? What are the steps that you are taking? Are you attempting to do this in mmWave Studio or in the configuration file (.cfg)?

Regards,
Kyle

Kyle,
Using the mmWave Studio I managed to obtain some conclusions so thank you for your suggestion to use it.
However I am implementing a solution that does not rely on mmWave Studio or Windows.
I am using Ubuntu 18.04 and the steps I took to perform a regular capture are the following:

-Connect the hardware: XA1642ABSABL ES2.0 board to the DCA1000EVM according to instructions on DCA1000EVM, connect each to its power supply, USB from XA1642ABSABL to the laptop and Ethernet cable from DCA1000EVM to the laptop.
-Shunt the SOP2 pins and flash the demo pre-built binaries that come with the SDK (~/ti/mmwave_sdk_02_00_00_04/packages/ti/demo/xwr16xx/mmw/xwr16xx_mmw_demo.bin) to the XA1642ABSABL, then release the SOP2 pins.
-Guarantee that the SW2 in the DCA1000EVM board is as follows: SW2.1 - LVDS_CAPT; SW2.2 - ETH_STREAM; SW2.3 - AR1642_MODE;SW2.4 - RAW_MODE; SW2.5 - HW_CONFIG
-Start the XA1642ABSABL with a regular configuration file, with the exception that the line refered to the LVDS is as follows: lvdsStreamCfg -1 0 1 0
-press the CAPTURE_START push button on the DCA1000EVM board, the DATA_TRAN_PRG led starts flashing signalizing that the board is streaming data to the laptop through the ethernet cable.
-Use wireshark to check the data stream, it is possible to observe several UDP packets in this connection, all of them coming from the source 192.168.33.180 (the IP set by default in the DCA1000EVM)
-Using MatLab I open the previous UDP connection with MatLab's UDP objeject and stream the raw data coming from the DCA1000EVM.

After observing some successful results, I tried to disable the TX channels, changing the following configuration file lines:
chirpCfg 0 0 0 0 0 0 0 0
chirpCfg 1 1 0 0 0 0 0 0
Setting the last value to 0 as it can be seen.
This configuration launched an error in the firmware, so I changed the source code using CCS and commented the code shown in a previous reply: e2e.ti.com/.../2830467
I built this changed version of the demo and flashed this new binaries to the board. I was now able to disable the TX channels and could check it using the demo visualizer, however when I press the CAPTURE_START button in the DCA1000EVM there is no data stream.

I hope this is clear enough to understand and I know this seems to be a quite troublesome alternative to simply using the mmWave Studio but it would really be beneficial for me to use linux so that is why I am exploring this workaround.
Best regards,
Edgar