Is there any truth to this statement?
Representing fanuc positional data as joint versus Cartesian will increase line/travel speed.
Is there any truth to this statement?
Representing fanuc positional data as joint versus Cartesian will increase line/travel speed.
Yes. Moving linearly or in a cartesian point to point the TCP (tip of the weld wire, middle of a gripper, whatever that may be) will move in an exact straight line. When moving from point a to b using jointed programming or points it can achieve faster travel speeds because it doesn't worry about keeping the TCP in a perfect line. It will arc to a certain degree and is easier on the servo motors / joints of the robot.
Hope this helps.
What rcran2015 said is true, moving in joint is faster vs. linear movement type. But the coordinate representation can be Joint or Cartesian and still use either movement type. The coordinate representation should not affect speed, only the motion type as rcran explained.
Coordinate representation should be chosen based on if you want to have the robot position relative to a user frame or strictly defined as joint angles. I use Cartesian representation for movement relative to a "fixture" or "work surface" and use joint representation for "Home" or "Maintenance" positions that are not tied to a work surface.
I tested this concept recently with a Fanuc R-30iA robot controller and M-10iA manipulator. I ran 2 separate (and each dual) tests, with a simple pick and place program the first test, and then with a program that purposely used a large amount of extraneous wrist motions, including changes to the position CONFiguration string. In both tests, the robot manipulator was in "dry cycle" and did not actually handle a panel. Additionally, no unnecessary logic was executed in any of these tests: only servo motion, JMP LBL loops, Registers to count up, and TIMERs running.
I used system timers to measure the total cycle time, and fed those timers to R[ ] registers to remember the results of the testing. In the first case, the simple program looped 100 times. In the second case, the program with a lot of extra unnecessary wrist motions and CONF changes looped 50 times.
This was done with a modern Fanuc robot, the R-30iA controller. The only difference between the tests was the F5[REPRE] 'sentation of POSITION data.
The 2nd run of both tests was identical to the first, except that all of the J (Joint) interpolations (J moves) were re-[REPRE]sented as Joint data instead of Cartesian data. The very few L moves (Linear interpolations) in each of the programs were never modified for the test.
My results follow...: I MIGHT HAVE TO DO THIS IN MORE THAN ONE POST .....sorry
This is from the 2nd test, with lots of unnecessary wrist motions.
DSRobGuy,
I hope this answers your question.
For the newer Fanuc robots (R-30iA system or newer), there does not seem to be any advantage to re-representing P[ ] positions in any particular representation.
This result says NOTHING about the importance of choosing the correct motion interpolation (J move or L move) for the task at hand. It only speaks to how the positions are [REPRE]sented.
No advantage to insisting on a Joint repre 'sentation
Excellent test USURP_RUR!
The logic behind it is that the robot is still moving to the exact same points and using the same movement types. The positional data is just REPREsented differently.