Hi, Allyssa, interesting project! Would like to know more about it and its progress.
I went that way too - used genserkins in 6 axis industrial robot arm retrofit project. 1. Drifting or error from perfect line. While possible, but I doubt there are errors in genserkins math. And perfect visualisers confirm that. More likely reality differs from your theoretical parameters. Are you sure scale of all axes is correct? Are homing or zero positions in place? Are your entered DH parameters accurate enough? Usually arm drawing specifies the dimensions, but in reality they might differ slightly. I saw people using precision 3D digitisers to get those parameters right. Joint position might be not exact due to loose PID. Mechanical gear unit, including toothed belts and shafts might be worn out, thus having backlash and looseness. Robot programming with original robot controller is done by teaching method usually. This eliminates or decreases similar errors, which exist even with a new robot and a brand controller. 2. Impossible joint angles. This happens at near-singularity postures. Have to dig deeper in math to understand this. For start, I would just recommend not to approach such postures and use joint mode to go-by or re-arrange the joints. For more or better solutions, I would recommend searching for information or people. There are more people who went similar ways too. Maybe they could add something more... Good luck! Marius 08/03/2016 09:37 PM, Alyssa Wells rašė: > We are currently working on a project to retrofit a 6 DOF industrial robot > with LinuxCNC controls. We have communications working over EtherCAT and can > use Axis to jog all six axes of the robot. We have tried our hand at setting > the D-H parameters for genserkins, with mixed results. The kinematics appear > to be at least a little bit off, despite looking correct in numerous > kinematics visualizer softwares. > For the most part, the robot tends to jog in a straight line, but drift in > sort of a curve in at least one direction. Sometimes the kinematics solve to > joint angles that are impossible, and the robot faults for "over speed" from > trying to go to positions > 10000 degrees. > > Thank you for taking the time to read this! If you have any ideas or tips on > how we can fix our kinematic model, it would be greatly appreciated. > > Thanks, > Alyssa > ------------------------------------------------------------------------------ > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
