robotic welding
Chalenge with robotic welding
if a human can do it, a robot can easily do it once. the problems begin on the second try.
That is to say, interfacing a welding machine with a robot and manually teaching it to do one pass is very easy. The challenges lie in doing it on a production scale. Out of the box, an industrial robot is extremely good at one thing and one thing only : Always repeating the exact same movements.
For a weld to be good, it needs to be placed in the center of the joint. A human with 2 eyes can, fairly easily, figure out where the center of the joint is, but the robot is going at it blind. The human equivalent of a welding robot would be a one-armed blind person holding a MIG torch. Through no fault of their own, this not someone we would expect consistent quality from. We should expect even less from a standalone robot.
There is two different path we can take to aleviate the robot’s handicaps. We can either make sure the welding joints are always in the exact same positions or we can give the robot some sort of ability to sense its environment.
In reality, we usually try to do a bit of both. Below are the different options we have.
Custom jigs
a jig designed for your parts
For simpler parts, it’s possible, even common, to only have a robot and a good jig. The jig needs to be designed in such a way that the welding joints are always in the exact same positions relative to the robot. When done correctly, the cell will consitently deliver high quality parts.
The only problem with this approach is that it’s often impossible to design such a jig. For example, here is a thought experiment : Try to imagine a jig that would hold all 6 faces of a cube in place while leaving enough room for the robot to weld all the edges.
Pre-assembled parts
let the robot do the back-breaking work
For more complex parts, often, the only realistic approach is to have someone pre-assemble the parts before they go infront of the robot. The robot then gets the unenviable job of completely welding all the seams.
However, since the position of the joints now depend on human assembly, the robot will need some way to measure the actual positions of the joints.
Touch sensing
seeing with the torch
The welding torch can be used to detect the position of a surface. By combining multiple touch it is possible to accurately locate the begining and the end of a welding joint.
Only downside is the speed. It’s slow.
SeAM TRACKING
all is good when the arc is on
Seam tracking is a technology that allows a robot to follow the joint through voltage changes in the arc.
For example, in the picture to the left, the robot was programmed to weld a simple straight line, but, with Seam Tracking activated, the robot precisely followed the shape of the joint.
Laser profilers
vpr™ : our very own vision system
Feeling the need for a more versatile way of locating welding joints and seeing a gap in the market, we decided to develop our own vision system.
welding processes
it’s not about the process, it’s about the robot’s position
MIG steel, MIG aluminium, MIG copper on steel with pre-heating, etc… It’s not really about the welding process. Yes some welding processes are a lot more capricious than others, but, at the end of the day, it just means that the robot’s position needs to be even more accurate.
Side note on TIG :
– Is it possible? Yes.
– Is it recommended? No.
– Why not? Because TIG welding is a highly visual process and, for this use case, robots are blind.
– Is there any alternative? Yes. TIG is most commonly used for thin aluminium and today’s welding machine, Fronius for example, are equiped with lots of features that enable reliable thin aluminium MIG welding. Our first recommendation for automating TIG welding is always to try and see if it’s possible with MIG.