kuka PRC for parametric Laser Cladding

Started by sialam, October 30, 2020, 06:05:42 PM

Previous topic - Next topic

Johannes @ Robots in Architecture

Hello Syed,

I've tried to propose some changes, but please remember that I don't know the exact process parameters or tool geometry.

Some things that I noticed/changed...
...When cutting the mesh, in Rhino use Mesh / Mesh Edit Tool / Mesh Split so that you can get just the surface. In your original file, there was some extra geometry from - I guess - a Boolean operation.
...I've added some components so that you can get a zig-zag movement.
...I've made some changes to the calculation of the external axis angles, which was reverse before.
...Finally, the change between the geometry of the mesh geometry that you provided at the turntable and the one you used to program was due to the base values you used. The turntable simply places the geometry on the table, so if you put it at Rhino's origin, it stays at the origin. In contrast, your toolpaths were transformed according to the base values that you provided. So I've just added an inverse transformation. Note that this only affects the simulation, but not the code generation.
...To make the geometry more lightweight, I've also reduced the complexity again. The split geometry is at the initial resolution, though.
...Before you bring it to the robot, change the base number to the base number you used to calibrate the base. At the moment it's 0, which cannot work for a synchronized rotary axis.

Hope that helps!
Best,
Johannes

sialam

Hi Johannes.

I believe we made a bit of a mess in our calibration process and that is why we had a few issues, that by now we are trying to fix. By this moment I am writing to you, we already seem to fix a few things and now we are writing any tiny change in our logbook so this doesn't happen again.

This was in our side in the robot, not in the software so when we finish our calibration verification , we will just verify our software values. Thanks for pointing us this as well.

We might be able to try this today hopefully we have it working soon!

We'll study the changes you mention, we want to get to understand it as much as we can.
Again thanks we'll keep you posted!


Johannes @ Robots in Architecture

Great, fingers crossed!
Regarding calibration consider just mounting a simple tip on your tool and just go to a few selected points (i.e. not just the three points you used to measure the base) to check the accuracy. If those points look fine, you can have confidence in the points in-between as well.

Best,
Johannes

sialam

Hi Johannes its been quite some time since we had a word. This time I am here with Daniel writing to catch up with you on what we have been doing for the past weeks.

First, Daniel went out to Toronto to have a KUKA retraining on programing 1 & 2 as his experience was with ABB robots. Everything went well and he seems to feel more comfortable and knowledgeable of tiny but important details.

We installed the chuck, did the calibrations that were not as accurate and mounted the scanned pipe in place in order to do a dry run. Here I am sharing a tiny bit of the first dry run where the error of the e2 axis appeared (Maximum Following Error Exceeded) , I believe it is regarding the bigger load on the E2 axis.https://www.youtube.com/watch?v=BFUUn94hGnQ Nothing fancy but we try to record as much as possible.

We are characterizing our laser beam with a beam profiler in order for us to compute the catchment efficiency and the correct cladding parameters.

I was wondering if there is a possibility to add another angle configuration with the existing code or would we have to add some code in the script that you made. We want to add another degree of configuration to our beam for more catchment efficiency. We might start to dig into this possibility inside the code but we where wondering if you could help us with a 30 min or so session in order for us to understand better the workflow and the script.

I hope this is not to much to ask, but we would also like to learn a bit more in depth.

Also find attached a few images of the setup that we will run just for reference.

We hope things are getting back to normal and you are safe and sound. Sorry for the very very late response.
Syed & Daniel

Johannes @ Robots in Architecture

Hello Syed,

I replied on your DKP400 post a few days back! Regarding the meeting, please get in touch via eMail!

Best,
Johannes