Orient Positioner

[CharlesBo]

Hi all,

I've looked through the forums and followed the advice with setting up a positioner, and have been 95% successful;

- The robot is correctly oriented in the rhino scene
- The positioner is the correct distance from robot
- E2 rotations appear correct in simulations

I have two final issues I'm trying to address.

The first is the positioner needs to be rotated 90 degrees to be accurate to our setup, so the E1 tilts towards and away from the robot. This I assume is done through the Positioner Root Frame, but when I click it nothing pops up or makes a request etc. I'm not sure what's suppose to happen, but I assumed a menu would appear?

The second thing is more a discrepancy than anything else. My KUKA settings on the pendant shows the positioner has a z value of -826 which, when I type that into the gh component is way below where it is in reality. Again this might be solved with the positioner root frame, but I could be wrong?

[Johannes @ Robots in Architecture]

Hello,

Where are you changing the positioner root frame? If you right-click the positioner component you can - in your case - change the A value and it will rotate it (or more accurately the robot in relation to the positioner).
When you say -826, do you mean the height of the floor, measured from the positioner? I measured the DKP and that sounds plausible, or at least not off by a lot.

A good way to verify the position of a turntable (without using the built-in calibration tools) is to equip the robot with a well-defined tooltip and then get the global, Cartesian position of the same point on the turntable using different E1 values for rotation (e.g. 0, 120, 240). In Rhino/GH you can then create a circle from 3 points and extract the center point.

Note that KUKA|prc assumes that you are using an Offset Base - don't use those values for that, because then the coordinate system rotates along and you get the identical values, no matter the rotation. So switch into a global coordinate system instead.

Best,
Johannes

[CharlesBo]

Hi Johannes,

-826 is the z-value from the $config file our integrators have set up for precise location of the base. I would assume this gives the Z position where it meets the rail (see photo), relative to the robots base. What I was trying to ascertain before is what the positioners Z is in XYZABC in the PRC. From what you've described, Z would be the location top of the positioner, so could be found by doing -826 + height of positioner. The circle trick sounds much more sensible for this to confirm.

With the positioner root frame, sorry should have be clearer. I found the right-click options. Adjusting A does rotate the robot in relation to the positioner (in my case it would be -90.75 degrees), so it is correct in relation to the positioner, but then both the robot and the positioner are oriented incorrectly to the world XY. The positioner needs to be rotated, so it's motor sits on the X-axis, not the Y-Axis and the robot sits on the Y-Axis. From what I read online the positioner root frame (when you right-click) allows you to specify the orientation of the positioner and then you can position the robot relative to that, but I might be really off with that.

Kind Regards,

Charlie 

[Johannes @ Robots in Architecture]

Hi Charlie,

Hmm... I don't have a DKP at our lab - I will have to drop by at the local KUKA HQ sometimes, there is one in the tech center. It might be easiest for now to put together a custom DKP - see attached. It's a bit hacky, but the data inputs should be clear. Then you can set your values accordingly.

That being said, the exact value of the root point doesn't matter that much as long as it is physically/virtually at the right location in both cases. As in your case the positioner location is fixed anyways, you can match it once. Your base defining the toolpaths is relative to the moving part of the turntable, so the base values themselves don't depend on the root point.

At least that is my understanding.

Best,
Johannes

[Johannes @ Robots in Architecture]

Hi Charlie,

The DKP500 just seems to have a different standard orientation than the DKP400. I'll send you a build with DKP500 and the KR8R1440-2 arc HW in a few hours.

Best,
Johannes