I'm trying to figure out how Fanuc calculates a TCP (x,y,z only) from only 3 taught points. I've read this thread discussing calculating the TCP https://www.robot-forum.com/robotforum/rob…ulation-of-tcp/ but the 4 points on a sphere method isn't what fanuc use. I'm guessing the extra rotation information in each taught point allows for a solution to be found with only 3 points..
Does anyone know how they do it?
And what about the effects of misalignment, ie why doesn't a small error throw the solution way off since only the minimum amount of data (3 points) are used?
How does Fanuc calculate the tool frame?
-
Chojin -
December 5, 2014 at 4:52 AM -
Thread is marked as Resolved.
-
-
The reason I ask is I'm trying to create a routine to automatically set the tool frame and user frames on machines during commissioning (using input from a touch probe). Has anyone done a similar thing?
Thanks in advance for any help! -
I'm going to ASSUME that Fanuc does it like all other companies with whom I'm familiar. The required three points are (1) Frame Origin point, (2) point located ON desired new frame X-Axis, (3) point located anywhere on the positive portion of the X-Y plane of the desired frame. Some calculations request an optional fourth point (4) where to put the new frame Origin after the geometry is calculated. Generally for (4) I simply repeat (1). In this way only the {X, Y, Z} components are used to calculate the frame location & orientation. The {,,,RX, RY, RZ} components are ignored.
In structured language pseudo-code:
frmNewFrame = FRAME(pntNewFrameOrigin, pntNewFrameXAxis, pntNewFrameXYPlane, pntNewFrameOrigin)
This creates a TRANSFORM from the Robot World Frame to frmNewFrame with correct {X, Y, Z, RX, RY, RZ} components for the frmNewFrame.
Hope this helps. -
Hi
Is all that it needs! But remmeber that there is also a 6 point method! That defines not only position but also orientation of the tool
This is more a mathematical question: if you want to solve a system of linear equations you need at least the same number of values.
If only need the position of a TCP, ONLY need three points
but if you need to define tool orientation, then you need at least six points
But this applies to ALL robots not only Fanuc,
There is a nice function in Nachi robots that calibrates the robot and TCP at the same time!
calculates the axis offsets and tool distances, guess how many points you need to define? At least 12!I used it to calibrate a robot once because I didn´t have the calibration blocks and pins to master the robot and worked GREAT!
I would like to help you more if you can send me more information,
-
There is a good explanation of the three point TCP method here http://www.intechopen.com/books/advances…bot-programming
I have it working in excel and its perfect for 3 points that are ideally aligned in the simulator. But when i do the maths for imperfect points from the real robot ( the same points that have been used to calculate a TCP) my answer is different from the on-board Fanuc calculation.. They must be doing some error correction as well
-
The reason I ask is I'm trying to create a routine to automatically set the tool frame and user frames on machines during commissioning (using input from a touch probe). Has anyone done a similar thing?
Thanks in advance for any help!Have you had any luck with this? I would be intresed in using this.
-
There is a nice function in Nachi robots that calibrates the robot and TCP at the same time!
calculates the axis offsets and tool distances, guess how many points you need to define? At least 12!I used it to calibrate a robot once because I didn´t have the calibration blocks and pins to master the robot and worked GREAT!
I would like to help you more if you can send me more information,
I would like to know more about this. Do you have any reference or manual for the Nachi robot function?