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.

Linux/IWR1642BOOST: Device disconnects while running mmw ros demo

Part Number: AWR1642BOOST

Tool/software: Linux

Hello,

I'm trying to run the ros demo on Ubuntu 16.04, but I'm having a problem with the device disconnecting after receiving a sensor start command. The error I receive is below:

[ INFO] [1511461205.137585004]: mmWaveQuickConfig: Sending command: 'multiObjBeamForming 1 0.5'
[ INFO] [1511461205.144097471]: mmWaveCommSrv: Sending command to sensor: 'multiObjBeamForming 1 0.5'
[ INFO] [1511461205.151121949]: mmWaveCommSrv: Received response from sensor: 'multiObjBeamForming 1 0.5
Done
mmwDemo:/>'
[ INFO] [1511461205.168349544]: mmWaveQuickConfig: Command successful (mmWave sensor responded with 'Done')
[ INFO] [1511461205.169156636]: mmWaveQuickConfig: Sending command: 'sensorStart'
[ INFO] [1511461205.172085242]: mmWaveCommSrv: Sending command to sensor: 'sensorStart'
terminate called after throwing an instance of 'serial::SerialException'
  what():  SerialException device reports readiness to read but returned no data (device disconnected?) failed.
[ERROR] [1511461206.155143781]: mmWaveQuickConfig: Failed to call service mmWaveCLI
[ERROR] [1511461206.155243027]: sensorStart
[mmWave_Manager-2] process has died [pid 16547, exit code -6, cmd /home/david/catkin_ws/devel/lib/ti_mmwave_rospkg/ti_mmwave_rospkg __name:=mmWave_Manager __log:=/home/david/.ros/log/ec4d9786-d07a-11e7-a08a-00215d3ebae0/mmWave_Manager-2.log].
log file: /home/david/.ros/log/ec4d9786-d07a-11e7-a08a-00215d3ebae0/mmWave_Manager-2*.log
[mmWaveQuickConfig-3] process has died [pid 16548, exit code 1, cmd /home/david/catkin_ws/devel/lib/ti_mmwave_rospkg/mmWaveQuickConfig /home/david/catkin_ws/src/ti_mmwave_rospkg/cfg/1642_2d.cfg __name:=mmWaveQuickConfig __log:=/home/david/.ros/log/ec4d9786-d07a-11e7-a08a-00215d3ebae0/mmWaveQuickConfig-3.log].
log file: /home/david/.ros/log/ec4d9786-d07a-11e7-a08a-00215d3ebae0/mmWaveQuickConfig-3*.log

I'm using the clean mmw demo firmware from the 1.0 sdk, and the ROS package is unmodified. I followed the procedure specified in the user guide for the ROS package.

Any suggestions

Thanks!

  • Hello David,

    I believe this error is due to some configuration that is passed to the device before the sensorstart.

    We will look into the issue and respond ASAP.

    Thanks,
    Raghu
  • Hi Raghu, 

    The error appears to occur during initialization of the RF subsystem. When the init message is being set, the mailbox tries to trigger an interrupt by writing to a register it looks like? At line 690 in mailbox.c

        ((driver->hwCfg)->baseLocalToRemote.reg)->INT_TRIG = CSL_FINSR (((driver->hwCfg)->baseLocalToRemote.reg)->INT_TRIG, 0U, 0U, 1U);
        

    Debugging through, this is where the error actually occurs. Unfortunately, this error also causes the debugger to disconnect. 

    I've attached my stack trace from just before the error occurs as well.

    Any progress on your end? Any other information I can provide that might be helpful?

    Thanks,

    David

  • David,

    The crash and error message shown in the Linux ROS message capture you sent are typically seen when the EVM does not have sufficient power. The EVM user guide specifies a 5 Volt power supply that can supply at least 2.5 Amps. Many phone/device USB chargers and computer USB ports only supply around 1 Amp and will cause the crash seen. If you are using a computer USB port or a charger/adapter that lists a current less than 2.5A (typically printed somewhere on the adapter), then you will need to change it out for a 5V power supply that can supply at least 2.5A. If your current power supply meets the requirements, you may also want to try swapping it out with a different power supply in case the one being used has an issue.

    Please mark the thread as answered if the issue is resolved or reply if more support is required.

    Regards,
    John
  • ...2 weeks of digging through code, and that fixes it immediately. Thanks John!