[colinb83]

When i tried to calibrate my TCP using RDK its giving me a strange result that is approx. 24mm out from the TCP figures i get when calibrating the TCP on the robot, and by using the Delcam long spike short spike method, this is on a KR240 2500 Prime running KRC4 8.6.6



these are the Figures i got from the RDK TCP calibration, and i have also attached the RDK station

-4.7270064,-71.8829651,108.2701720,-7.9387870,-36.6675034,4.8695169
5.4036212,-67.1676636,110.9625930,7.7603900,-44.0661240,-55.7711601
-16.5013866,-63.2351494,108.1887740,-22.7271709,-47.2785454,75.6447906
2.6902754,-68.8284000,113.9103090,74.3806686,-69.5186539,13.6651077
-5.7562475,-71.4935608,105.4217990,-106.4113920,-50.1194954,16.0552654
-10.4045801,-53.6684000,70.7701340,-121.6261440,-59.9532280,120.1800230

[Jeremy]

Hi Colin,

You didn't attach your .rdk station to your post.
I'll take a look as soon as you do.
Could you also provide the values calibrated with the robot? (Btw, in the meantime, you can simply enter these values directly in RDk and skip RDK's tool calibration all together.)

Jeremy

[colinb83]

apologies, I forgot to click add attachment,

Also I sent an email to you and Albert regarding Calibration license purchase about a week ago and still have not had a reply

Thanks

Colin

[Jeremy]

Hi Collin,

What are the values you got from calibrating using the Kuka controller?

For the calibration license, did you CCed me in the email?
Could you forward that, I'll take a look right away.

Jeremy

[colinb83]

Hi Collin,

What are the values you got from calibrating using the Kuka controller?

For the calibration license, did you CCed me in the email?
Could you forward that, I'll take a look right away.

Jeremy

Please see screen shot of the Delcam calibration, you essentially calibrate 3 times with a short and 3 with a long and it works out the errors and calculated the centerpoint of the nose of the spindle

although it does seem that Delcam works with the Z facing the opposite direction in its output to RDK, but you can see the 3 short spike tests are just 3 calibrations ran on the robot itself

[Albert]

Hi Colin,

In RoboDK you can calibrate each TCP individually as explained here:
https://robodk.com/doc/en/General.html#CalibrateTCP
So you can calibrate a short tool and a long tool.

You can then use one of these scripts in C:/RoboDK/Library/Macros:

1. TCP_2_axis_projection.py Project a TCP to an axis defined by 2 calibrated tools given the names CalibTool 1 and CalibTool 2. You can also align the Z axis of the TCP to be along this axis.
   
2. TCP_Set_Relative.py Define a TCP based on a distance offset with respect to a reference TCP.
   

To run a script, simply load it in RoboDK and double click it to execute it.

The first script is also explained here:
https://robodk.com/doc/en/PythonAPI/exam...cp-to-axis

I attached a sample project that goes a step further and recalculates all your tools based on tool lengths. This script is based on the TCP_Set_Relative_All.py example and it takes the tool length/offsets from the name.

We can help you customize it or create a simple UI so you can recalculate your tools.

Albert

[colinb83]

Thanks, that looks great, it always amazes me how powerful the RDK software is.

did you manage to work out why it was giving strange results when I tried to calibrate the tcp using RDK and the kuka driver to retrieve points vs the tcp calibration on the control itself?

as im also assuming whatever is causing that will cause my TCP setting on my rotary table to be out of line too

Also please look at your emails regarding the calibration licence, many thanks

Col

[Jeremy]

Hi Colin, 

I'll agree with the title of your post, it is strange. 

I did take all the joint positions available in the "Define TCP" window of RoboDK and created targets with them. 
They all line up perfectly with the TCP calculated by RoboDK. 
The semi-transparent TCP you see is the one calculated in Delcam. 

And according to the TCP errors graph, everything lines up well. (see picture)
   

If you use the Delcam values in the controller, does the jog rotation around a fixed cartesian position works?

Jeremy

[colinb83]

Yes the figures from the robot controller, and from Delcam both line up on the real robot to move around a fixed cartesian, Its the RDK define TCP that is giving the strange values, i used the driver to pull the joint positions so it wasn't finger error on my part I shouldn't think

[Jeremy]

That's very weird.

One quick test you can run would be the following.
If you bring the robot to a certain set of joint values, do the same in RDK, does the cartesian position of the flange (TOOL[0]) match in both environments.

As the difference seems very small, I'm looking to see if the kinematic of the robot you are using in RDK and the one in the controller are the same.

Jeremy