Posts by Doctor_C

    No, I don't think you can. I even tried the PALOAD[0] command and that says the same thing.

    But you could set to an un-used payload which would be the same.


    If you don't mind, how do you lock out the +% and -% keys?

    I have also looked at the variable $MCR.$GENOVERRIDE and compared to a R[…:speed].

    This way I can allow them to speed it up but never go slower than the register (speed is critical to the cycle time).

    Or let them slow it down with the +% and -% keys but never go faster than the register (no need for speed, very long cycle times)

    And one more use, long cycle time machining, I will move slow through picking and prepping a part for load or placing the finished part. Then have a very high speed for machine unload/load (which determines machine thru-put)

    Yes, things are different.

    When I do them, I have the robot in Roboguide. Have the page loaded into the robot

    Do a backup AOA

    Open the STM that is in My Workcells, that robot, and work with it there.

    Then you can do-copy very easily and see what it will look like in the robot.

    It's a pain, but.....

    you should also set "Abort all programs by CSTOPi" to true

    "Failed to run task" means it wasn't reset when you tried to start. You Need a delay there.

    I also noticed you have PNS setup. Do you need it? If you can decide what to do in your "Main" program you can eliminate all the PNS stuff and UI6 Start works fine.

    CSTOPi, Reset, and then Start

    Nation hit it on the head. 1 or 2 working points, why bother.

    But say 35 points around an application, I would teach a Uframe. Big reason for me using frames is we are an Integrator. So we do it at our shop, then install at customer.

    If you were applying to 35 points, with a Uframe you would only touchup 3 points (frame) onsite.

    Things are never exactly in place like on Cad. No frame, you would touchup 35 points.

    Uframes get me home quicker :D

    I've done it with DCS, but not with IIC.

    Only way I think you could do it is with different models which are tied to different UTOOL numbers.

    But if you are using points versus PR's, that won't work for you.

    I had the situation once, Fanuc says for IIC between the 2 robots to stop collision on a single rail, they have to be oriented in the same directions (ie their X directions must be the same). Mine were facing each other.

    They couldn't help me at all and while waiting for their help, I got an epiphany. I apologize now for the very quick representation of what I did.

    I created bubble zones under each robot, 10,000mm below (-Z), large enough to cover the reach of each robot.

    Link number was ZERO, Worked very well, and covers even manual jogging. Hope this can help you.

    I have used Ethernet I/P and socket messaging with Keyence cameras and Asycube vibrators.

    They had a very good setup process, and also example programs to trigger both.

    Now I am being tasked with hooking up to a Cognex 3D laser scanning vision system.

    Cognex's documentation is clear as mud.

    Any one had any experience with the Cognex and explicate messaging programs thru Ethernet I/P?

    Especially the programs (examples).

    Any help would be appreciated.

    I went back and looked at the ones I got boned with. I had nothing.

    One thought, if you could figure out the rotation, you could possibly add a J5 tool offset to your motion to try and offset your offset.

    Sorry, you just learned a hard lesson. It is the nature of a 5 axis robot.

    What you are seeing is a slight rotation of the part, building up the more you place.

    The interpolation is not the same. I got burned on this in 2010 cuz the customer bought 5 axis robots, bunch of them, cuz they were cheaper. He got burned in the end.

    J6 wants to go past your soft limits. You probably led in with a Linear move, which takes the shortest path (ie its at +150, want to go to -150---shortest path is backwards to +210)

    lead up to the motion with a Joint move, it will turn the correct way