ros-industrial / motoman

ROS-Industrial Motoman support (http://wiki.ros.org/motoman)
146 stars 192 forks source link

Add GP110 support #476

Closed cjue closed 2 years ago

cjue commented 2 years ago

This adds a support package for the Motoman GP110 manipulator.

@EricMarcil: As far as I could gather from comparing the datasheets and technical drawings the GP110 is just about identical to the older MH110, apart from the DX200 controller for that one (and a different total weight?). Is this assumption correct?

EricMarcil commented 2 years ago

@cjue It's possible that it is very similar to the MH110 (I haven't checked to confirm). In any event, the attached file has the values for the GP110.

Robot_Specs_GP110-A0.csv

cjue commented 2 years ago

Thank you for the GP110 data, I updated the effort values in the URDF accordingly.

I found my link_6_t mesh was in the wrong place and that the collision meshes were corrupted, this should all be fixed now.

The meshes total about 4MB in size now.

cjue commented 2 years ago

I used Blender Decimate to further shrink the biggest links, meshes/ is down to 3 MB now.

@gavanderhoorn Is that size fine with you?

cjue commented 2 years ago

Is this correct? flange is supposed to have X+ 'forward' (ie: pointing out of the flange link). Should this rotation be applied to tool0 instead?

Yes, it appears there is the same mistake in the gp4 macro: https://github.com/ros-industrial/motoman/blob/kinetic-devel/motoman_gp4_support/urdf/gp4_macro.xacro#L148

The hc10dt_b10 macro does it the right way: https://github.com/ros-industrial/motoman/blob/kinetic-devel/motoman_hc10_support/urdf/hc10dt_b10_macro.xacro#L156

gavanderhoorn commented 2 years ago

Seems that was missed in #433.

Please remove 7e5caf6 from this PR.

We should not touch multiple packages in a PR introducing a new package.

cjue commented 2 years ago

Please remove https://github.com/ros-industrial/motoman/commit/7e5caf65bdaa3e74e45ee27181701c9388caccf0 from this PR.

done, also rebased to the current kinetic-devel

gavanderhoorn commented 2 years ago

I'm going to squash-merge this as well, to avoid the fixups (and non-decimated meshes) from ending up in the history.

gavanderhoorn commented 2 years ago

Thanks again @cjue