EGM read joint angle always 57 times larger

[ghost]

Hi All,

We were using the EGM to do real-time control and want to move the robot 4th joint, for example, from 0 degree to 30 degree. The simulation on robot studio works just fine. However, the joint angle reading changes from 0 to 1710, instead of 0 to 30, when we run the experiments on the real robot. We tested all the joints and the angle reading almost always has to be divided by 57 to get a seeming accurate reading. Since we are doing real-time control, this reading is vital to us and we would like to have an accurate reading, if possible.

I wonder if you may know what's going on. The customer service agent was not responding to our questions since they don't know this EGM library...

Thank you.

[gavanderhoorn]

This is most likely a radians vs degrees conflict.

Which version of RobotWare do you have installed on the controller?

And which version of abb_libegm are you using?

Reason I ask is this notice in the readme:

RobotWare versions less than 6.07.01 are now incompatible with abb_libegm (due to changes in the EGM communication protocol).

[ghost]

@gavanderhoorn Thank you for your quick response. We were using the 6.06 robotware and I believe we have the most recent abb_libegm.

[gavanderhoorn]

We were using the 6.06 robotware and I believe we have the most recent abb_libegm.

That would be the problem then.

[ghost]

Oh man, thank you. Previously the abb_libegm does not work with robotware versions earlier than the 6.07. I didn't notice the update.

Thank you.

[ghost]

We will be testing the robot now to see if works.

[gavanderhoorn]

Closing this as it would seem to be a user configuration issue.

[gavanderhoorn]

@jontje: would there be some way we could add some sort of version detection to abb_libegm to prevent situations like this?

With abb_librws that would probably be possible, but with just abb_libegm?

[jontje]

@jontje: would there be some way we could add some sort of version detection to abb_libegm to prevent situations like this?

With abb_librws that would probably be possible, but with just abb_libegm?

It should be possible to detect if proto fields introduced in the new EGM protocol is missing or not, I will investigate.

[gavanderhoorn]

Tracking version checking in #46.