Messages

Exactly, with flat panels etc. the height makes a big difference!
Best,
Johannes

In general you want to avoid coinciding axes, it can also happen between e.g. A1 and A6, but A4 and A6 are usually the most common ones. I'm sure that you have noticed that before, when you move through a position and the A4 starts flipping - basically the tooltip slows down, whiler there is much movement happening "behind" the tool.
Here's a nice video: https://www.youtube.com/watch?v=zlGCurgsqg8

Best,
Johannes

Some milling spindles actually give the tool a 45 degree inclination, that helps you prevent kinematic singularities (e.g. when the axes of A04 and A06 coincide) as the arm will more often be tilted (compare e.g. to the Elte spindle in the library). However, from our experience this may again cause reachability problems when processing elements vertically.
As often with the robot there is no "best" solution, but a trade-off!

Best,
Johannes

For a final confirmation it may be a good idea to call the KUKA hotline. However, chances are good that they will also only quote the manual ;)

Phew, not really my area of expertise, but I'm sure that the KUKA values already contain all kind of safety coefficients.
After all, they assume a maximum force of 4600N for a 200kg machine.

Best,
Johannes

Hello,

Here are the relevant values for the KR6-2, which should be very similar to your KR6 arc. That should make simulation and dimensioning easier!
Fv = Vertical force Fvmax = 4,600 N
Fh = Horizontal force Fhmax = 5,000 N
Mk = Tilting moment Mkmax = 5,200 Nm
Mr = Turning moment about axis 1 Mrmax = 4,200 Nm

Best,
Johannes

Hello,

The most common crash is when the Settings window is loaded and some libraries are missing (but this would then happen all the time, not just sometimes).

If you run into any more crashes please let me know and I'll try to fix it. I can't think of any occasion when changing a geometry caused Rhino to crash - if anything is corrupted it may turn one of the other component red, but usually nothing worse. But of course this doesn't mean that it's not possible, so keep me updated. I can then also get in touch with McNeal if it's not directly PRC-related.

Best,
Johannes

Haha, have to prepare for next week's workshop and presentation at eCAADe ;)
I skipped the part with the closed curve, here's a quick fix attached. It's a basic problem of the conversation from CAD to CAM (but a good idea to integrate a function for that in the divide curve component...).
The reason why the component is working the way it is working lies in our early research into flank milling, where we used one curve to define the tooltip and one for the toolaxis, see e.g. here http://www.robotsinarchitecture.org/wp-content/uploads/2011/09/ecaade_ddp.pdf
I'll look into integrating a switch to flip the tool axes.

Ah, and nice video! In the new PRC, the robot should be orange/black, though... Did you intentionally change something or is this how it looks for you?

Best,
Johannes

Hello,

Here is how I would approach that problem (attached, using the member version of PRC as you are also using).
You can define the tool direction as a vector, and the use the vector to create a plane with the Plane Normal component. As the milling tool is symmetrical, you can rotate it around the tool axis to improve reachability. The Orient Plane component rotates planes around their Z-axis so that the are facing to the given point.

Hope that helps!
Best,
Johannes

Great, definitely do keep us updated in regards to your projects - and don't hesitate to get in touch if you run into any issues!

Best,
Johannes

Hello,

Sorry for the late reply, I somehow missed your reply!
And yes, there are the index holes, but quite frankly I never use them with smaller robots. I guess they are good to have if you need to mount a tool accurately, i.e. not having to calibrate it on the robot again. However, seeing as the Kress spindle doesn't even have a  (e.g. HSK) tool holder, you have to manually calibrate it anyway.

Anyway, I don't believer that you have to worry about anything with a 6kg payload robot!

All the best,
Johannes

Hello,

The KR3 is indeed a bit old, but you can find a 3D model of it on the regular KUKA downloads page (Downloads, CAD, Archive). Here's a direct link that should hopefully work: http://www.kuka-robotics.com/res/sps/8ba89f50-3ccd-4361-9b9a-60bdd0662ac6_kr3_STEP.zip

We do integrate new robots for our members for free, so if you decide to become a member it's no problem integrating it into PRC as a native component!
Best,

Johannes

Hello,

Definitely make the aluminum plate shorter if you only use it for the Kress spindle, there's no advantage in having a longer aluminum plate and I would rather scratch the robot with a plastic part then a sharp aluminum edge ;)
That being said, we never had any serious problems with the plate, so it somehow worked.
What do you mean by "so low"? If you refer to the distance between the flange and the tooltip you've got two factors that act against each other: The closer the distance is, the more it is likely that the robot's A5 collides with the ground when you do five-axis milling - the further away it is, the less stable the setup becomes. So you just have to find some compromise between those two.

I quickly threw together a definition where you can dynamically adjust the length of the plate and test it in PRC - that should make it easier to decide!
Note: Only works with the most recent member version of PRC!

Best,
Johannes

Hello Bob,

I just sent you an eMail, but maybe this is helpful for others as well - instead of using a helix you can also slice it into circles (in the case of a tube) and then flip the direction of every second circle so that the robot automatically unwinds itself.
The attached example uses the new PRC for the robot components, but most pathplanning is native Grasshopper anyway!

Best,
Johannes

Hello Alex,

Well, proper multi-robot support is hopefully incoming - I'm working with the RMIT guys and FabUnion on finding a nice solution.
I just checked again and robot|tool collisions seem to be working (see attached image). For a multirobot setup you would currently have to take the GEO output of the core component and then look for intersections/collisions, e.g. using the Collision One/Many component or the Mesh Intersection.

Hope that helps,
Johannes