[finilias]

Hello, I'm having an issue with a robot arm I'm working on using RoboDK 5.9.1. Additionally the robot has camrob installed with custom Robotic Solutions Programming (spindle controls, tool changer, etc). 

The issue I'm having is that using the same joint angles on the kuka pendant vs RoboDK, the cartesian coordinates are different by a varying amount depending on the robot angles. At first I thought it was something to do with my TCP, but I set the TCP to the flange base on both the pendant and RoboDK and the cartesian coordinates are still different. From the values I've taken, the differences average about x40 y-10 and z-50 with a variance of about 10mm in both directions.

I don't know a terrible amount about the robot parameters/kinematics and I don't have the tools to do a robot calibration, though I suspect that may be my issue here.

I've also noticed that using the 4 point TCP measuring tool in the pendant produces a different value from the RoboDK TCP utility by a similar amount, and using the TCP made by robodk does seem to produce a better TCP result when tested.

Does anyone know why this would be happening? Is there a way I can tune the robot in RoboDK to compensate for this?

I've attached my robot cell for review.

[Albert]

Did you make sure your tool and coordinate system are the same on both systems?
Is the robot calibrated by KUKA?
Are you using any external axes?

If you can send us your RoboDK project and some sample points you are using to compare we can help you better.

[finilias]

Hello Albert, thank you for taking the time to look at my post.

Did you make sure your tool and coordinate system are the same on both systems?
I did. I was using the robot base (all zero) as the coordinate base in both RoboDK and the Pendant.

Is the robot calibrated by KUKA?
I'm not sure, I believe that was done by Robotic Solutions, but we got it in an auction and I'm having trouble getting through to them to ask them questions about it.

Are you using any external axes?
Yes I am using a turntable. If I use the same coordinates for the turntable as my base, the positioning is off in the same way.

If you can send us your RoboDK project and some sample points you are using to compare we can help you better.
Not even attempting to run a project yet really. Just having trouble getting things lined up because of this. I'm attaching my cell again but this time with a few reference points labelled. The points are numbered and correspond by number to one another. E.G. RoboDK Point 1 is the cartesian value given by RoboDK of the same angles using the same TCP and coordinate system as Pendant Point 1.

[Albert]

There are different ways you can integreate a turntable or an external axes with your KUKA robot arm. This example post processor modification specifically shows different was to define the BASE frame for a KUKA system and can help you understand some different options:
https://robodk.com/doc/en/Post-Processor...amesFilter
The BASE frame can specify a static reference frame (no external axes), or a system with external axes such as a reference frame attached to your turntable.

If this is the first time you use a KUKA robot, setting up an external axis and properly calibrating it can be challenging. I would recommend you to remove the synchronization with the turntable and run programs without moving the turntable.

In any case, if you want to setup your external axes these should still help you point in the right direction:

1. Properly find the position of the turntable (KUKA calls it the ROOT point of the turntable): https://robodk.com/doc/en/General.html#C...xTurntable
   
2. This thread can help you understand how to configure this with your controller: https://robodk.com/forum/Thread-How-to-c...ernal-axes
   
3. This thread may provide additional information: https://robodk.com/forum/Thread-Problems...-KUKA-KCR2