[Strider]

Dear members or the forum, i have a problem with the postprocessor. After creating the path, the simulation lokks fine, but after the program is generated there are strange movements tha lead to collision at the robot. I have tried to follow the tutorials and also searched the pp documentation and also this forum. Unfortunately i couldn't  find a solution within the last 2 weeks. 
 
The problem and also my attempts to find out more about it and my attempts to solve it are described in detail in this thread: https://robodk.com/forum/Thread-Problem-...4#pid19744

The problem is, that the pp, base is the "GCode_NCP", refuses to calculate the Z value of the choosen reference frame in a correct way while doing it correctly for the X and Y values of the output program. Also there is the strange rotation in the nc program from -180 to 180 and back to -180 degrees in the next liine, which doesn't makes sense at all.

Here some more information. You can find the projekt and a link to a screenvideo with the creation of the program: https://datashare.tu-dresden.de/s/XFzFSxXwY2dcQEp. There is also the program itself attached.

There is the problem with the Z value to be seen and also the strange behavior in the angle for Q1 in the following part:

"N00007 G01 X=806.000 Y=862.000 Z=520.000 Q1=-180.000 Q2=-0.000 Q3=180.000
N00008 G01 X=806.000 Y=1362.000 Z=520.000 Q1=180.000 Q2=-0.000 Q3=180.000
N00009 G01 X=306.000 Y=1362.000 Z=520.000 Q1=-180.000 Q2=-0.000 Q3=180.000"

As you can see, Q1 value jumps in N00008 from plus 180 to minuns 180 and back in the next movement. This behavior leads to the movement in the linked video: https://datashare.tu-dresden.de/s/dfgqa7GswCqiF42. At second 33 the robot trys to adapt the turning of Q1, the program has to be stopped to prevent damage.

When i change the Q1=180 to Q1=-180 in N00008, the program is running OK, only the Z value is still a problem. I can change the Z value by manipulating the temporary file "PostViereck_20kg.py" and compile it with the post processor, but this seems not a good way.

There are unfortunataley 2 main questions / problems unsolved:
1. Why is the Z value wrong while the values for X and Y are fine in the NC program?
2. Why is this jump for Q1 and how can i prevent this?

We are no professional Phyton programmers to develeop the postprocessor or his functions in detail, that's why the code sniplets we got as answer unfortunately could't hale to solve the problem finally. We just want to use RoboDK for genereating the programs and run our robot hardware. At the actual point, we can use it only for simulation.

May be there is only a problem with the settings in the project, but we have no idea in what 

Can anyone help us please?

[Sergei]

Please correct me if I am wrong, but it looks like you are using some specific controller with these Autonox robots.
The default postprocessor for these robots is the postprocessor for KEBA controllers (KEBA KAIRO).
The GCode NCP postprocessor you plan to use always outputs coordinates relative to the robot base ( the reference frame you select in CFP does not affect this). That's why you see this difference for the Z coordinates.

It seems that for your controller, the robot base should be where the NP-20 frame is located (e.g., this is the same location for Fanuc robots).
Accordingly, if you want to output coordinates for a robot base located as NP-20, it is necessary to rebuild the robot model.

I've rebuilt the robot model accordingly. Here are the changes I made:

   

   

Also, I attached your station with the modified robot.

As for the sign change for ABC angles - these are just Euler angles and usually, these values are handled and converted on the controller side.
Also, please make sure that your controller uses the same notation for angles as the postprocessor.

[Strider]

Dear Sergei,

thank you for the information and also for the modified station. I will test it today and i will give you feedback.

Again, thank's a lot for the support!

Best regards,
Frank