Can UR10e read 5-Axis G Code?

Standard 3-Axis G-code programs work very well in the Remote TCP & Toolpath UR Cap. Can UR/Polyscope interpret 5-axis G-code? What format is needed (XYZ B & C, or A & C, etc) for the robot to accurately follow a 5-axis toolpath? Thanks for any information on this!

1 Like

Can anyone respond to this? Obviously the robot can move in the same way as a typical 5-axis router; but can the UR setup interpret typical 5-axis G-Code? What format does the G-Code need to be in for this, if it is possible? I worked for many years with a Fanuc Robot that responded very well to RobotMaster, a Mastercam plugin, that was able to convert a typical 5-axis toolpath into robot-speak. Is the G-code interpreter in Polyscope able to do this, and if so, what format does the G-Code need to be in for this to happen? Thanks for any insight into this.



According to the article, it is possible to generate 5 axis Gcode on UR (Autodesk Fusion 360
Post-processor: Universal Robots (up to 5-axis)):

Unfortunately, I did not have the opportunity to test it.

Has anyone tried this?
Can anyone at UR confirm one way or the other if the UR Gcode interpreter can interpret real 5axis Gcode including A- and B-axis data?

I have tested a simple 5-axis program and it seems to work. I used a 5-axis post processor designed for Thermwood routers that outputs B and C axis g-code. On the Thermwoods, the C axis is the XY plane (rotates around Z), and B rotates around the Y axis. The UR seemed to be moving correctly, however I have not had time yet to determine exactly how pivot distance is calculated on the UR10e. I am hoping in the next month or so to run some tests and attempt to dial in a 5-axis program. I am using Mastercam; but I am currently working with a Fusion 360 team who is attempting to develop a better UR post for their software. As we are using our robot to trim parts, Fusion is quite limited at this time in creating toolpaths that follow wireframe with z-axis guides. But they seem to be actively working on this, so we’ll see how it goes.

More as this unfolds…

I did some tests with the existing F360 UR post processor; it seemed to work just fine running in the simulator. I have not tried IRL.

I also tried the urscript output option and got similar results, but it makes a script several thousand lines long for my simple example part; cant imagine what that would look like with more complex part, but it is interesting to be able to ‘generate path from cad’ without the use of the toolpath urcap.

You mentioned you are using the UR10E for trimming parts. The parts are kydex, I am assume from looking at your website? How are you teaching and or generating your toolpaths?
We are waiting on a UR10e for trimming Composite parts, with similarly complex geometries, and looking at using a Wandelbot to generate tooling paths.
Any insights would be greatly appreciated. We have a great CNC team, but are new to Arms.


I prefer using Mastercam. My generic HAAS post creates relatively clean 3-axis code. I manually go in and remove a few G and M codes that aren’t understood by the UR control. I have also generated 5-axis toolpaths (using a simple Thermwood 5-axis router post) but great attention must be paid to scaling, “pivot distance”, part orientation and type of 5-axis moves generated, so the the robot can execute without wild joint moves. I have also used Fusion 360 for 3-axis toolpaths, which does have a UR post. The code generated however seems to be bloated with unnecessary comments and superfluous code. It is not easy or possible to create detailed 5-axis code in Fusion (at least with our basic version of the program and my limited Fusion experience).

We use several remote TCPs and hold the parts with the robot. Once you understand the relationship between the RTCPs and the PCS, it is easy to create g-code that is surprisingly accurate with minimal adjustment. We are cutting 0.08" Kydex, which does not offer much resistance while cutting with a 3/16" single flute router bit; but as the cutters dull or when cutting through thicker material (side walls) the robot faults out. It does not seem to take much pressure to cause the robot to deviate from it’s expected position and throw a fault. Feel free to contact me if you have other questions. Good luck…I would be happy to hear about your experiences, especially with the composite material.


Thank you so much for your feedback. Greatly appreciated.
We are excited to get going, and I will reach out when we have traction.