Question about accuracy

  • I have a M20-iA arm and a R30-iB controller. There seems to be an issue with the accuracy. A point saved into the program keeps its point when it is moved in the tool frames y-direction as well as its z-direction up to 1.75mm in either direction in those axes. I am not sure if this is normal for the model of robot or there is a problem with the servo motors.

    The motors seem to be a little rough when the move. I'm not sure of the proper descriptive term, but what i mean by rough is that they do not seem to move smoothly. They have some jerky motion at some points.

  • Jerky motion is not good, but could be caused by how it is programmed.


    I'm not sure what you mean when you say a point keeps its point. Would be helpful if you posted the program code and describe what's happening with each line of motion.

  • I'm using the TeachPendants coordinate system and the change is also visible. When the robot places the part into another gripper that is stationary other than rotation, it often misses and the part will fall. This also happens after the rotation during the pick from the rotary gripper. The robot gripper will miss and the part will fall.

    This is also reflected when the part is placed into a vise inside the CNC machine. The part is often off in the y-direction of the tool and unable to fit into the slot it was designed to fit in.

  • Without knowing too much about the gripper or component, I recommend setting up a dial indicator, programming a point to apply a set measurement, and then sending the robot to that point from various other positions. That would give you an idea of the robots capability.


    Is it possible the part is shifting in the gripper? 100% Fine moves could cause jarring, potentially allowing the component to move.

  • I don't know if it matters but I tried running it in STEP at 10% speed and the same errors appear.

    And by jarring do you mean that it is jarring itself? it makes no contact with anything else until the part is gripped.

  • And by jarring do you mean that it is jarring itself? i

    I suggested that the part could be jarred/slip in the gripper due to the robot's rapid acc/deceleration, but STEP mode would rule that out.


    Not sure what else to recommend other than a dial indicator or fixed/stationary object you can bring the gripper and/or part over to for determining what is varying.

  • Hi

    First of all. This line surprises me


    L P[21] 100% FINE


    I didn't know you can have percentage on a linear motion

    Anyway, what I noticed is that all your point are going at 100% to an abrupt stop (FINE)

    I don't know how long your motion are but that kind of logic could produce a vibration


    Also, how heavy and big is your tool ?

    Retired but still helping

  • I ended up changing them all to 500mm/sec FINE and they still have the same issues.

    I'm not sure the weight. I was not the one who installed the tool. I am cleaning up after a previous worker who did not finish the cell.

    The issue is that when the program is running the tool will shift further than the accuracy leniency. From what I know, the repeatability is supposed to be within 0.08mm. This is not the case when it is placing the component. It is shifting past a full mm after running only 3 iterations.


    Let me clarify, it is not just in this one spot. It happens all over at random times. It makes no sense to me that it is happening and my only thought is that it is mechanical. I was just wondering if there is any solution that could be done outside of calling Fanuc for maintenance since my company is a lot on the cheap end.

  • It turns out that a failed backup was performed and the arm had to be remastered. It was not recalibrated afterwards. That may be the issue.



    The payloads are set. the robot uses 2 grippers and they are under the max load

  • It is important to understand that by factory default is assuming that has 20 kg at the end and it with calculate all its motions based on that. You need to be sure that the tool weight and geometry matches the robot setup.

    Retired but still helping

  • The robot was properly calibrated and the payloads were correct. Sorry to not specify. We are now under the assumption that an encoder is gone bad and are trying to find ways to work around it. Service is not an option at the moment.

  • Having years of experience with Fanuc motors, I'd be shocked by a couple of things: 1) that you had a bad encoder, and 2) that the motor would keep working with a bad encoder. Those things are rock-solid and I can count on one hand the number of actual failures I've seen in 25 years. "Bad" for a Fanuc pulsecoder typically means it's not working at all, either due to same hardware failure, like the glass disc breaking, or the electronics failing. Since that encoder is used for motor commutation, a non-functional encoder means a motor that won't spin, and often results in some sort of overcurrent alarm.


    Not sure if that robot has any belts inside, but I once saw a Kuka that had been serviced by "professionals" where they'd screwed up the pulley on an idler shaft and the belt was half off. The robot positioning was doing something very similar to what you're seeing.

Advertising from our partners