Kuka KRC4 .cnc

  • For those of You familiar with using Kuka KRC4 .cnc for milling via offline programing with cnc g and m code, what is your opinion of it ? I am familiar with g code programing for cnc machines. I am considering a krc4 kuka with an end of arm spindle for carving some wood items vs. a 5 axis cnc machine.

    1) Is the sending of the file to the robot fairly simple?

    2) Are the controls capable of very large programs ?

    3) What kind of tolerances can be held ? (How true is what Kuka claims: .0015" ?? )

    4) Are there provisions for tool setup, tool change, and tool offset ?

    4) What other issues are there to consider compared to using a 4 or five axis cnc table?

    Thank You.

  • I can't speak to KUKA.CNC specifically, but regarding the more general issues:

    What kind of tolerances can be held ? (How true is what Kuka claims: .0015" ?? )

    I could write an entire essay on robot accuracy vs repeatability. But bottom line is, robots generally can't achieve that kind of accuracy. Nobody's robots. The kinematics just don't make it possible. Consider, for a moment, the physical mass and footprint of a classical 3-axis or 5-axis CNC that has the same working volume of a 6-axis robot. Robots simply don't have the rigidity, and the "lever arm" of error at each axis just amplifies too much over the length of the reach.

    Not to mention the $ cost -- a robot is far less expensive than an equivalent CNC machine. You pay for that $ savings in other ways.

    The numbers KUKA posts aren't lies, but you have to read the fine print carefully. Those numbers are usually based on an ISO 9283 test.

    For example, a KR100-HA (a High Accuracy model) is officially listed as having a repeatability (in open air, no external forces) of 0.05mm, but an accuracy of 0.7mm. That's for arriving at a fixed point. Path accuracy along a Linear or Cirular path is listed as 0.8mm or 0.6mm, respectively, but with repeatability of 0.2mm or 0.25mm.

    Every model of arm will have different values.

    What other issues are there to consider compared to using a 4 or five axis cnc table?

    Rigidity. A robot arm is far more flexible than any CNC -- just look at the mass difference. So when doing cuts with a lot of "recoil" force, the robot will flex and bend much more than any CNC machine.

    That doesn't mean that good cut quality cannot be achieved, but it may well take multiple light passes, and repeated "spring" passes. And given how basically inaccurate robots are (compared to CNCs), you probably can't just "hit download, get part". You'll have to optimize your CAD/CAM to play to the robot's strengths and avoid its weaknesses.

    Load data accuracy is very important. Robots have to pay attention to tool mass, CG, and inertia where CNC machines mostly just ignore them by sheer brute force. Motion tuning is also a factor -- an HA KUKA has a factory tuning, but to get the best benefit from it, you really need to have the robot tuned on-site, carrying the actual production load. KUKA offers this as a service, but it's not cheap.

    I'm not trying to be discouraging here, but the move from "normal" CNC to Robot CNC is a much larger step than going from 3-axis to 5-axis. Maybe more like trying to switch from subtractive machining to additive -- you can't just throw the same CAD model into a 3D printer and expect to get identical results. Each tool set has its own strengths and weaknesses. And you need to have an idea of how deep the water is before diving in.

  • kuka.cnc is actually using g-code!

    But a robot is not a cnc machine!

    We have here hard vs. soft

    Hard tool -> soft material (machining wood, special material)

    Soft tool -> hard material e.g. (polishing)

    In the past I did not have any complains about the accuracy

    (I started about 35 years ago)

Advertising from our partners