Positioning Issues 200iD/7L with R30iB

    Hello,



    We recently purchased 200iD/7L robot with R30iB controller.


    Robot has low hours and appears to be in great shape.


    Here is the problem and I think it is a head scratcher.


    Using 250mm Tramming Stone – In WORLD align one side of the stone to X . When checking in Y direction, I can see the robot is not following the stone. So basically the robot is not moving 90deg X/Y ( not square ) .


    Second test – ( with correct Tool Frame – using a pointer that attached to J6 so only Z length entered in Tool Frame) – made User Frame using the tramming stone . Made sample program to hit all 4 corners of the stone . When moving X250 mm – the robot actually stops short about 2mm from the end of the stone ( which is 250mm away) . position screen says that robot is at X250mm. Next point X250/Y250 – robot is overshooting by a 1mm in X and .5mm in Y . So Basically the Robot is not moving square to the Frame .



    Checked Robot Mastering. All Marks line up. I also did the Fanuc enhanced mastering procedure for J2/J3/J4 and J5 . The mastering is good .



    Not sure what to check next. Any suggestions?

    I have seen issues in the past where the robot was loaded from the AP1 and the 7L configuration was not selected (they picked current configuration).


    However, I don't think that is your problem as everything is very close, not perfect but the wrong robot setup usually looks much worse.


    You didn't mention mastering J6. I've also seen this affect accuracy when the J6 was mastered with a level, or by eye. Problem arises when the robot is not level. Put the robot in the Zero position all axis. Then I take the tool off, put 2 long bolts in the lower 2 holes with the dowel hole up. put a level where the robot mounts on the base, and mark where your bubble is. Then use that level to master J6 and align to those marks, and single axis master J6. I hope this helps VS

    Doctro_C


    Thank you for your advise. I'll try that today. I was able to get the Original Master Counts from Fanuc Yesterday and remastered the Robot with them. Mastering was close but not perfect ( within 0.5degon J3 and J4 and about 2 deg on J6). I believe it is now perfect. Anyway, the problem is still there . I 'll try remastering J6 today and report back.

    Thank you

    The robot is not attached to the Table

    I tram using a granite stone by moving the robot in World X direction only, to where the indicator reads Zero so the robot is perfectly aligned with one side of the granite stone .

    I then indicate the stone using Y axis move . The robot does not follow the stone edge. Over 250mm is already off by 3mm . So the Robot is not moving square in X and Y

    Are you sure that your robot and stone parallel to each other? Do you use press-fit pins to locate the robot base on the robot's pedestal/base attachment? One more thing, do you use any other sort of measurements(caliper, measuring tape, else) to determine that robot doesn't calculate well? If robot doesn't go in line with your table edge using world frame, most likely that your robot is not installed right.


    If you use your Granite stone 250mmx250mm and you want to use it as a your golden piece for measurement, could you confirm that your table has certificate of dimensions?

    aligned granite stone to X axis of the WORLD , so the side of the granite stone is parallel to X axis travel of the robot. Robot orientation to Base is irrelevant. The robot must move in X and Y that are 90 degrees from each other and it currently does not . it moves in 90.8 degrees based on my measurements.

    I do use height gage and calipers to verify positions. If i tell the robot to move 250mm in X direction - it shows' a move of 250mm on the TP but in reality it moved 247mm.

    i' ve sent more info to Fanuc and waiting for their response.

    How is the repeatability of the robot? You should be able to measure with a test or dial indicator.


    Do you have your payload properly set? How many hours are on the mechanical unit? It may look clean because it was in a clean environment and well cared for, but that doesn't mean it didn't have a tedious past life.


    If the repeatability is spot on then you may just be running into the accuracy limits inherent to an articulated pedestal robot.

    Repeatability is good ( .02mm)


    I have hard time believing that I have reached the max capability of this robot


    I’m trying to pick parts from 260mmx 270mm Tray . There are 7 x 13 3/8” x 1.5” parts . With set User Frame at the corner of the Tray and using position registers and register to advance counts, The Arm falls shorter of the programmed distance the further is travels .


    I made the attached report to Fanuc. Waiting for their response.


    Please feel free to offer any advice as I’m lost on how to move forward.


    BTW, we have 200IC on another machine picking parts from same Trays – no problem there .

    I wonder if we are missing some sort of Accuracy Option on this Robot

    Quote from scotty

    If robot doesn't go in line with your table edge using world frame, most likely that your robot is not installed right.

    My problems with the movement of robots on X, Y and Z (Arc Mate 100iC 12L and 8L) are just manifesting in the mentioned way. Since RobotOrigin and RobotZero points are virtual points that cannot be accessed, the front surfaces (machined) on the front or right side of the robot carrier (base) should therefore be used. In relation to these surfaces, a work table, a work object should be placed, and then the UF should be determined. The UF thus determined will contain some values for W, P and R which means that the plane is rotated / tilted along the X, Y and Z axes. When you activate that UF and jog robot along the X and Y axes you have to get parallel movement with the edge of the table or object. If you said that the repeatability is good, try to write a program where only P1 is learned and you get other points through PR and offset in relation to the starting point, if the repeatability is good and you set 250 mm offset, it must be 250 mm in reality. The values for W, P and R are an indicator of robot unperfection. In my example on two robots through the calibration procedure the following values were obtained for X, Y, Z, W, P, R for Robot1 (1,080 / 0,777 / 0,873 / 0,030 / -0,327 / 0,529) and Robot2 (1,451 / -3,725 / 0,782 / 0.000 / 0.052 / -0.073), when jog robot 1 by Y axis (1000 mm) I had about 10 mm drift, on robot 2 this drift was less, about 4 mm. So this unperfection for everyone robot is his personality and it occurs between the base and the mechanics of the first axis. This you can't eliminate, you must live with this. All this applies when you work with TP, if want to work with RoboGuide and program teached on 3D model transfered to real robot, things are more complicated, very soon I will post my experience for this problem.

    Quote from VS Manufacturing

    W0

    P0

    R-90.1

    By this data You have only rotation by Z axis in CCW (-0,1 deg), as You can see in my post for two robots I have from zero up to 0,5 deg rotation almost every axis.


    Check parallelism of Yours table/working piece relative to forehead surface at robot base, determine UF, activate that UF, teach Points, must be OK. And on end but first You must precise determine TCP. If You determine E.g. UF5 and UT5 the in your program activate

    USERFRAME_NUM=5

    UTOOL_NUM=5

    Quote from fmiroslav

    If you said that the repeatability is good, try to write a program where only P1 is learned and you get other points through PR and offset in relation to the starting point, if the repeatability is good and you set 250 mm offset, it must be 250 mm in reality.

    That's exactly the problem i'm trying to solve .

    With UF set ( using three points) and using correct TCP ( pointer sticking out from the J6 , X0,Y0,Z150mm,W0,P0R0) -

    P1 set at same point as origin of UF1 ( end of 250mm long Bar stock )

    P2 X250mmY0 - programmed


    Robot moves to 247.8mm when I physically measure ( position screen of robot shows 250mm).

    You can visually see that the pointer is short of the 250mm


    - - remastered the Robot - same result

    - INIT the control to factory setting - same result

    - i'll swap the control with another 200ID/L7 we have and see if error is there.

    When I was in vision training a few weeks ago, I was informed that the witness marks are pre-molded into the robot prior to assembly. They're good for getting a rough calibration, but unfortunately they're often off by over a degree in rotation.


    I'd recommend calling your FANUC rep and requesting they perform a vision mastering. They'll attach a camera to the robot and it will remaster itself using a calibration grid.


    We did this in class and found the witness marks were about 0.5mm off.


    Note - vision mastering only masters J2~J5. J1 and J6 aren't re-mastered (though, they're not critical anyway to generating straight lines)

    schnautz


    after I read you post i went back and remastered the Robot using the Enhanced Mastering procedure from Fanuc ( J2-J5) . The Robot now positions better than before. This makes me think that precise mastering is my problem. I would think more users will have this issue , especially if relaying on the witness marks to master their robots.

    Th error now is about 1mm over 250mm distance. I 'll try to remaster again to see if I can get it even closer ( .5mm over 250mm will work for this application) , if not I'll have to call Fanuc to get them to Vision Master the robot . I think I'm on the right path now...

    Thank you all for your help.

    I'll update later on my progress.

    Quote from VS Manufacturing

    Second test – ( with correct Tool Frame – using a pointer that attached to J6 so only Z length entered in Tool Frame) – made User Frame using the tramming stone . Made sample program to hit all 4 corners of the stone . When moving X250 mm – the robot actually stops short about 2mm from the end of the stone ( which is 250mm away) . position screen says that robot is at X250mm. Next point X250/Y250 – robot is overshooting by a 1mm in X and .5mm in Y . So Basically the Robot is not moving square to the Frame .

    TCP for Your Tool isn't defined only by Z, Your Tool Frame is more like this



    Quote from VS Manufacturing

    With UF set ( using three points) and using correct TCP ( pointer sticking out from the J6 , X0,Y0,Z150mm,W0,P0R0)

    Different Tool Frame,

    TCP 1 - X0 / Y0 / Z150 / W0 / P0 / R0

    approximately

    TCP 2 - X-50 / Y-50 /Z100 / W90 / P225 / R0


    Edited 2 times, last by fmiroslav: correction 6 axis triads ().

Advertising from our partners