Hi all,
Can I set a configuration of FANUC of the goal position in linear motion? I worked with KUKA and there we can set the configuration but KUKA doesn't look at it and moves in point with most "convenient" configuration.
Hi all,
Can I set a configuration of FANUC of the goal position in linear motion? I worked with KUKA and there we can set the configuration but KUKA doesn't look at it and moves in point with most "convenient" configuration.
Fanucs will honor the arm configuration and will fault if you try to move linearly from one configuration to another, but they do ignore turn numbers if moving to the point linearly.
You can add the wrist joint (Wjnt) instruction to the end of a linear move to move through different orientations, but the move will not be linear.
J1-J3 will move as they normally move in a linear move, but J4-J6 will do a joint move. This can be useful in tight situations where you want a predictable motion, but aren't absolutely concerned with maintaining a linear TCP move.
I'm not sure if this is a feature of newer controllers, but wrist joint does maintain a linear path with respect to the TCP. However, the orientation interpolation between the two points will not be evenly distributed (aka the arm will sometimes make some pretty wild moves occasionally).
WRJNT is intermediate motion between linear and joint.
The mayor axis interpolate linearly (but in this case is the Wrist Center Point that moves along a line)
The minor axis move like in joint (point to point interpolation)
Hence the name Wrist Joint because the wrist (axis) move in joint (point to point) interpolation
Still the motion command is Linear (or Circular) because the Wrist Center Point interpolates linearly (or in a circle))
Sometimes it is used to avoid wrist singularity (change of configuration, J5 changes from positive to negative or vice-versa)
It is imposible for the TCP to keep the line because it would eventually required to pass though the singularity (J5=0)
The only way to avoid this is to make the wrist axes interpolate point to point
The wrist center point keeps the line though using WRJNT
You can easily verify this when the robot moves between two positions with the arm extended
For the wrist center point to interpolate linearly (the shortest path) will require the arm to retract then extend again
In joint interpolation this does not happens, the arm stays extended while the base swings to reach the target position
Maybe that is how it used to function, and how it got its name, but I did a test in roboguide prior to posting my previous comment, and the TCP sphere followed the path exactly, and that was with a wrist flip in the path.
Maybe that is how it used to function, and how it got its name, but I did a test in roboguide prior to posting my previous comment, and the TCP sphere followed the path exactly, and that was with a wrist flip in the path.
In my experience it works how Nation describes. Following a straight path, but if you have singularity avoidance it can tilt the wrist to avoid it.
Sent from my SM-G930V using Tapatalk
It is difficult to see it in a real robot, but in roboguide is clear that the robot TCP follows a line with wrjnt
Thanks Nation for teach me this!
I wonder now what algorithm they use
and, how you do it in other robots? nachi, ABB, motoman, kuka?
In Nachis I don´t remember such an option and instead I define a tool with length zero (TCP at the wrist center point)
That way robot moves as I explained: mayor axis linearly, minor axis in joint