Schunk_canopen_driver launch file crash

asked 2017-05-11 09:28:05 -0500

matkok
1 ●1 ●2 ●3

updated 2017-05-13 14:27:41 -0500

Hi,

I am using Schunk LWA4P robot arm, and I recently downloaded schunk_canopen_driver package. After catkin_make_isolated, I launch standalone_interpolated.launch or some other launch file, and it crashes after trying to do "Commutation search for node 3". I was using arm previously with ipa_canopen and it worked good. Besides that, my controller_manager/spawner is waiting for service, and never spawns a controller. (ros_control and ros_controllers are installed).

I'll give you my output of my cmd.

Update: I found that it crashes when SchunkPowerBallNode::CommutationCalibrated calls m_sdo.upload(false, 0x2050, 0, commutation_status);. Then in SDO.cpp in SDO::upload on argument if( m_data_buffer[0] == SDO_SEG_ABORT_TRANSFER ). That is true and sends break exception.

Update2: In SchunkPowerBallNode.cpp in void SchunkPowerBallNode::commutationSearch(), when I comment the line bool calib_ok = CommutationCalibrated(); and set bool calib_ok = true;, in other words, when I don't go to CommutationCalibrated(), everything works, and all nodes initialize. BUT, I don't know why is that, and I don't want to do it that way. I have a feeling that the id 0x2050 in bool SchunkPowerBallNode::CommutationCalibrated() doesn't pass. Any opinions?

Thanks in advance, Matko

SUMMARY
========

PARAMETERS
 * /hardware_control_loop/loop_hz: 125
 * /hardware_interface/joints: ['arm_1_joint', '...
 * /joint_limits/arm_1_joint/has_acceleration_limits: True
 * /joint_limits/arm_1_joint/has_velocity_limits: True
 * /joint_limits/arm_1_joint/max_acceleration: 6.0
 * /joint_limits/arm_1_joint/max_velocity: 1.25
 * /joint_limits/arm_2_joint/has_acceleration_limits: True
 * /joint_limits/arm_2_joint/has_velocity_limits: True
 * /joint_limits/arm_2_joint/max_acceleration: 6.0
 * /joint_limits/arm_2_joint/max_velocity: 1.25
 * /joint_limits/arm_3_joint/has_acceleration_limits: True
 * /joint_limits/arm_3_joint/has_velocity_limits: True
 * /joint_limits/arm_3_joint/max_acceleration: 6.0
 * /joint_limits/arm_3_joint/max_velocity: 1.25
 * /joint_limits/arm_4_joint/has_acceleration_limits: True
 * /joint_limits/arm_4_joint/has_velocity_limits: True
 * /joint_limits/arm_4_joint/max_acceleration: 6.0
 * /joint_limits/arm_4_joint/max_velocity: 1.25
 * /joint_limits/arm_5_joint/has_acceleration_limits: True
 * /joint_limits/arm_5_joint/has_velocity_limits: True
 * /joint_limits/arm_5_joint/max_acceleration: 6.0
 * /joint_limits/arm_5_joint/max_velocity: 1.25
 * /joint_limits/arm_6_joint/has_acceleration_limits: True
 * /joint_limits/arm_6_joint/has_velocity_limits: True
 * /joint_limits/arm_6_joint/max_acceleration: 6.0
 * /joint_limits/arm_6_joint/max_velocity: 1.25
 * /joint_state_controller/publish_rate: 50
 * /joint_state_controller/type: joint_state_contr...
 * /pos_based_pos_traj_controller_arm/joints: ['arm_1_joint', '...
 * /pos_based_pos_traj_controller_arm/publish_rate: 50
 * /pos_based_pos_traj_controller_arm/type: position_controll...
 * /robot_description: <?xml version="1....
 * /rosdistro: kinetic
 * /rosversion: 1.12.7
 * /schunk_canopen_node/autostart: True
 * /schunk_canopen_node/can_device_name: auto
 * /schunk_canopen_node/chain_arm: [3, 4, 5, 6, 7, 8]
 * /schunk_canopen_node/chain_names: ['arm']
 * /schunk_canopen_node/frequency: 30
 * /schunk_canopen_node/node_mapping_3: arm_1_joint
 * /schunk_canopen_node/node_mapping_4: arm_2_joint
 * /schunk_canopen_node/node_mapping_5: arm_3_joint
 * /schunk_canopen_node/node_mapping_6: arm_4_joint
 * /schunk_canopen_node/node_mapping_7: arm_5_joint
 * /schunk_canopen_node/node_mapping_8: arm_6_joint
 * /schunk_canopen_node/ppm_profile_acceleration: 0.3
 * /schunk_canopen_node/ppm_profile_velocity: 0.3
 * /schunk_canopen_node/ppm_use_relative_targets: False
 * /schunk_canopen_node/traj_controller_name: pos_based_pos_tra...
 * /schunk_canopen_node/use_ros_control: True

NODES
  /
    robot_state_publisher (robot_state_publisher/robot_state_publisher)
    ros_control_controller_spawner (controller_manager/spawner)
    schunk_canopen_node (schunk_canopen_driver/schunk_canopen_driver_node)

auto-starting new master
process[master]: started with pid [11738]
ROS_MASTER_URI=http://localhost:11311

setting /run_id to fce92254-366f-11e7-8af9-fc4dd456d934
process[rosout-1]: started with pid [11751]
started core service [/rosout]
process[schunk_canopen_node-2]: started with pid [11769]
process[ros_control_controller_spawner-3]: started with pid [11770]
process[robot_state_publisher-4]: started with pid [11771]
flags:2050
<2017-05-11 19:33:48.039> CanOpen(Info)::init: CAN Device was set to auto. 
<2017-05-11 19:33:48.039> CanOpen(Info)::init: Trying CAN device: /dev/pcanusb0... 
[INFO] [1494524028.524145]: Controller Spawner: Waiting for service controller_manager/load_controller
Invalid CanDescriptor!
flags:2050
<2017-05-11 19:33:49.039> CAN(Error) tCanDevice(PEAK-CAN)::tCanDeviceT: Error open CAN-device '/dev/pcanusb0' (errno=Bad file ...

(more)

edit retag flag offensive close merge delete

Comments

And if someone can tell my how to put my Terminal output so it is easyer to read. Thanks :)

matkok ( 2017-05-11 11:57:52 -0500 )edit

Edit your question, select the terminal output, press the Preformatted Text button (the one with 101010 on it). Save.

gvdhoorn ( 2017-05-11 12:09:56 -0500 )edit

Does anyone know how to start tests, like test_sdo.cpp, that are in the package?

matkok ( 2017-05-12 07:01:36 -0500 )edit

add a comment

answered 2017-05-24 09:18:40 -0500

heppner
55 ●2 ●4 ●5

Hi matkok,

we looked into the problem, thanks for your detailed writeup which helped tremendously.

The error is not as assumed an SDO error which would produce a much more verbose output but (at least that is the theory) a PDO timing error which is resulting from the way we had implemented the commutation search.

The following lines in SchunkPowerballNode.cpp (135 following) are responsible for this:

  size_t counter = 50;

   // Try commutation search for counter times
  while (!calib_ok && counter--)
  {
    usleep(100000); // sleep for 100 ms
    downloadPDOs();
    sendSync(m_can_dev);
    uploadPDOs();
    calib_ok = CommutationCalibrated();
}

The routine waits for the commutation calibration by doing a fixed time wait. This is not the best way to do it as no SYNC signals are sent anymore during this time. If the sync signal is missed for a predefined number of steps a PDO exeception (resulting in a ProtocollException) is thrown and the nodes will enter the fault state.

Please try the following patch:

  size_t counter = 500;

   // Try commutation search for counter times
   while (!calib_ok && counter--)
   {
     sendSync(m_can_dev);
     downloadPDOs();
     uploadPDOs();
     calib_ok = CommutationCalibrated();
   }

The explicit sleep is actually not relevant as the sendSync will already take care of the wait without skipping the SYNC signal.

Please let me know if this works or you encounter further problems.

edit flag offensive delete link

Comments

Hi, I am sorry to inform you that is not the answer. I get exactly the same error, and output on my cmd. If you have any other idea, please let me know.

matkok ( 2017-05-25 11:11:45 -0500 )edit

add a comment

Schunk_canopen_driver launch file crash

Comments

1 Answer

Comments

Question Tools

Stats

Related questions

Schunk_canopen_driver launch file crash edit

Comments

1 Answer

Comments

Question Tools

Stats

Related questions

Schunk_canopen_driver launch file crash