Posts by Fabian Munoz

    There are few things you can do but before lets talk about this A colleague told me that, if I use "J". I think he mint save the point with joint representation and not cartesian.

    I use this method when I want to "untangle" the robot. The point is define with angles and the robot will go always on the same (know) direction. Usually I do this in between the pick (at any angle) and the place. The problem is that the offsets that you use are in cartesian so it does not apply to your angles. An offset of, lets say, 3 on axis 1 will rotate the robot 3 degrees instead of moving forward 3 mm.

    As far as your problem if you are picking parts at 0 or 180

    I would approach the point with the gripper half way of each direction, then the rotation is only 90 deg

    I would use numbers like 179.99 instead of 180 or 0.01 instead of 0


    La foto te muestra donde encontrar ACC.

    He programado cientos de robots y nunca he usado ACC. Para que la necesitas ? Porque no aumentas la velocidad ?

    EL problema que tengo conn ACC es que al usarlo te desvias de las acelerariones propias del robot. Estarias forzando al robot a ir mas despacio or lento de lo que ha sido construido

    I dont understand what you need. The description that you wrote indicates to me that you know how to work with PR and R

    When you mention the 356mm that indicates to me that you know the concept of PR (i,J) as a command.

    Can you rephrase the question OR show us the program part that you are talking about ?

    What I see is

    When you drop the first two boxes, the drop of point is higher than the other boxes , so you actually "drop" the boxes (I like that) and then when you retreat the robot is perfectible normal

    When you place the second set, you do the same after the boxes dropped and the tool now is empty, the robot goes down and compresses the boxes, then it goes up

    I wonder if that weird motion on the robot is just the tool "going back to normal" after being under pressure


    The standard Robot control software limits the TCP speed and the wrist flange center speed so it does not exceed 250 mm/sec in T1 Mode. In addition, DCS will shut down motor power if the TCP speed or the

    wrist flange canter speed exceeds 250 mm/sec in T1 Mode. The T1 Mode Speed Check function allows

    the user to further restrict the TCP Speed and wrist flange center speed to a speed lower than 250 mm/sec while in T1 mode.

    - T1 Mode Speed Check can be enabled / disabled for each motion group. (Default is disable)

    - The speed limit of the T1 Mode Speed Check can be changed, but it cannot exceed 250mm/sec.

    - In T1 Mode, when the DCS TCP speed exceeds the specified speed limit, the alarm "SRVO-340

    DCS T1 TCP speed" occurs and motor power is shut down immediately.

    - In T1 Mode, when the DCS flange center speed exceeds the specified speed limit, the alarm

    "SRVO-341 DCS T1 flange speed" occurs and motor power is shut down immediately.

    - The target TCP is the DCS Tool Frame that is selected by the "Tool select" item in the DCS Tool

    frame menu. When "Tool select" in the DCS Tool Frame menu is set to -1, the DCS Tool Change

    function is enabled and the target tool frame number is changed when the ordinary TCP is changed.

    it will be something like

    L PR1 Move to the point

    PR1,1 = PR 1,1 + 5 Add 5 mm on the x

    L PR1 Move 5 mm on the x from the point

    PR1,3 = PR 1,1 + 7 Add 7 mm on the z

    L PR1 Move 7 mm on the z from the point

    If you want to , you can put the 5 and 7 on R and do the math with the Rs

    If you want some help you are going to have to answer the questions

    I don't believe you have a colour camera. The XC-56 is the basic B and W camera Fanuc offers

    Do you know how to work with iRVision ?

    Do you know how to program the robot and integrate it with iRVision?

    If you have a black and white camera, I hope you can see the orange as a greyish colour and the blue one a bit darker. The you could use a tool called HISTOGRAM

    Also you cold define on the same vision program two MODEL IDs

    What software from Fanuc are you using ? Are you picking the parts while they are moving ?


    You know, I don't think you can do it. I think I tried that few years ago and it was not possible

    Think about the DCS design on the robot from factory. Each individual axis has a bubble around it and they are actually intersecting, but they work as a one and they don't monitor each other

    I believe if you put a box on the arm, this box will be part of this manipulator itself and it will never hit itself. It will be like having an extra axis. This box by design is already "touching" the robot arm. Let say that you put it around axis 6 and 5 and you want to avoid collision with axis 1 or 2, well, as I mentioned before, there is no axis 1 or 2 or 3..., its all one DCS

    Just had a crazy idea while typing. Maybe you can eliminate the original Fanuc bubbles, Create you own, lets say around axis 1 and 2 and you box on axis 5 and 6. Maybe now that you broke the link among then you can check for collision. I don't even know if you can do that but it's an idea .

    I will say 4 years ago, we have the same issue. We were troubleshooting the vision at the end of the day and I guess we forgot we left the Log All on. The next day we found "your problem" I think we lost the whole morning, I end up calling Fanuc and they told me to check the Log All, and there it was . No issues after

    I tell you something, if it goes good for a couple of days, turn it on again and you see the problem coming back


    Usually what you described is a UF issue.

    I know it was working but:

    Can you check the grid size ? Can you teach the UF5 again ? or make sure that matches what you did using the right tool ?

    Teach PR4 again

    Your program logic looks good