14 October 2019 at 23:03 #8741
We recently had a problem where we could not ssh from a PC to the e.Do. I re-imaged the SD card for e.Do, placed the SD card back in e.Do’s controller, and went through the firmware update process. Now we can ssh from the PC to the e.Do, but we can not calibrate e.Do. After running “roslaunch edo_control calibrate.launch”, we get warnings in the terminal that say “wait for service(/algo_control_switch_srv): failed to contact, will keep trying”. That warning keeps repeating until we kill the terminal. Can anyone help with this problem?21 October 2019 at 14:01 #18411
What is your objective? Do you want to calibrate the robot through a command line?
Are you talking about the real e.DO or the simulated one?
However I try to answer the question for the real robot:
The best way to properly calibrate the robot is the calibration page of the app on the tablet.
It is mandatory for the correct functioning that the calibration notches of the joints are aligned.
Best regards. 😉21 October 2019 at 15:46 #18414
I want to calibrate the real e.Do through the command line on the PC connected to e.Do.21 October 2019 at 22:33 #18417
When trying to run the calibration procedure, if I comment some lines of code regarding the algo_control_switch_srv, then I also received error messages stating the joints are not publishing states. Can you tell me how I can diagnose whether or not the joints are publishing states correctly using the command line on the PC connected to e.Do?21 October 2019 at 22:47 #18420
Additionally, I see that line 100 of AlgorithmManager.cpp in eDO_core/src references the algo_control_switch_srv, which seems to be involved in my problem. When I use the command line on the PC to see what services are running on e.Do, should I see a service for the algorithm control switch?22 October 2019 at 9:35 #18423
If you want to calibrate all the joints through a command line, remembering that should be already aligned, you could use the topic /bridge_jnt_calib publishing the message JointCalibration with mask 127 if you have the gripper, otherwise 63 for a 6 joints robot.
I want to underline that it is mandatory for the correct functioning that the calibration notches of the joints must be aligned, otherwise the encoder positions are disaligned and the robot think to be in another position with respect to the real one, the consequence will be a false collision and the disengagement procedure could cause damage to the robot’s brakes.
Best regards 🙂24 October 2019 at 17:02 #18585
I fixed the problem with not being able to calibrate e.Do. The root cause of that problem was that the ministarter file in the home directory of e.Do had the IP address for use with the internal wifi. I changed that to the IP address of the ethernet adapter and then the robot was working again.
However, the root cause of all of the problems was that the ssh service on e.Do had stopped working. That led to re-imaging the e.Do SD card and returning the ministarter file to factory defaults. Is there any explanation for why the ssh service on e.Do stopped working?28 October 2019 at 9:22 #18598
what is your strategy to calibrate all the joints from command line now?
Did you use the topic /bridge_jnt_calib?
The problem related to ssh connection could be referred to a corruption of the SD content inside the Raspberry, It could be due to different causes depending on the use of the sd, to better analyze the situation it could be necessary to know all the boundary conditions.30 October 2019 at 19:40 #18616
I am using the edo control and calibration package the ymollard posted on GitHub. I think he has also posted about it on the e.Do forum.
- You must be logged in to reply to this topic.