Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
TCP Calibration (Position & Orientation)

I am interested in how I can perform tool calibration in a RoboDK environment. We need information about the POSITION and ORIENTATION of the tool tip. As for the position not a problem, I wonder how I can get information about the orientation of the tool tip?

Thank you and best regards!

hi, I don't know which robot are you using, but if you are using abb, then the controller gives the position and orientation of the tip. in your case when there is a tool, it will give a tooltip position. I am saying in reference to the abb robot and irc5 controller. if you go to jogging, it will show the value of q1, q2, q3, q4, q5, q6. those values tell-all p[position, the orientation of the tooltip. so if you open the robodk then go to the robot option, ( look like I can not upload figure for your reference) there are places where you can write the value q1 to q6. then you will get the [x, y, z] AND [Z, Y', X'] AS A TOOL FRAME WITH RESPECT TO REFERENCE FRAME. hope this will help you.
-- Dibash Adhikari
-- Architect, urban planner
-- Researcher in POSTECH

I am interested in how we can calibrate the tool (POSITION and ORIENTATION) as shown in the pictures below (in the attachment - Cal_TCP)?

Thank you and best regards!


Attached Files Thumbnail(s)
Hi Mitja,

There's, unfortunately, no direct way of defining the orientation of the TCP within RoboDK.

But I can propose you a solution that I think would work.
I hear that ABB has a system where you do the standard TCP position calibration (as presented in your attached image). Then you add an extender to the tool along the desired Z-axis and you take one more point on your fixed reference. Using this added value, ABB is able to calculate the orientation of the Z vector.

So my plan would be to use the same kind of method with RoboDK, but instead of taking just one point with the Z extender, you do a second TCP position calibration.
With the real TCP position and the Extended one, you should have enough data to extrapolate the Z orientation.

What do you think?
Btw, it's not something I tested or anything, it's really just something I have in mind. I may test it in the near future though.


Thank you for your suggestions. We will try this method.

Thank you and best regards!


Users browsing this thread:
1 Guest(s)