The state of affairs today, of course, is markedly different. Sophisticated manufacturing engineers and managers operate in a world more demanding than ever before.
Ultimately it comes down to ROI and analyzing a significant capital purchase to justify cost of ownership. The usual tool of choice is a spreadsheet comparing different manufacturers' wares. Many scrutinize maximum linear speed in X, Y, and Z-directions, in addition to angular speed and the reach of the robot arm.
While these criteria measure significant aspects of robotic performance, such information is not a reliable indicator of true productivity.
It's a common experience to talk with a welding engineer who notes that a part can be welded in just two minutes. If you were to pursue the discussion by asking, "Then can you make 240 parts per day?" the response is likely to be, "No. Not really." The answer to the parts-per-day question, in fact, usually approximates 190 - 200 parts. This discrepancy leaves the manufacturing engineer/manager with important tasks, namely, finding the real process inefficiencies--and a solution.
The good news today is that the parts-per-day question has a new answer. True artificial intelligence (AI) power supplies and robot systems--with fully integrated components designed for maximum information processing efficiency and increased CPU processing speeds--provide more finished, quality parts per day.
Measuring true productivity requires detailed knowledge of the welding process. Consider a sheetmetal application in a high-volume industry such as automotive, in which an assembly frequently specifies that many discontinuous, very short welds are to be made at high speeds, thus requiring frequent arc starts and stops.
Careful analysis shows that reliable arc starts and stops significantly increase true weld productivity in applications involving short, high-speed welds. Technologies now available employ a capacitor discharge and a closed-loop feedback of amperage from the power supply to achieve high certainty of arc starts and reductions of arc instability.
Traditional, general-purpose power supplies with analog interfaces rely on sending a signal from the robot to the power supply to initiate routine end-of-weld functions, including slowing of wire-feed speed and anti-burn-back and crater-fill procedures, typically requiring a 250-msec interval. Newer systems move this function to the power supply to save time more profitably used for productive welding. The robot is free to initiate the next programmed step after only 10 to 20 msec.
Digital technology also simplifies how floor operators learn to program and use robots. Teach pendants operating with familiar Windows-based technology enhance productivity, facilitate learning, and allow for offline programming, training, and increased robot uptime.
Additionally, process research on actual welding arc behavior in different applications has yielded extensive databases of unique and optimum arc conditions for manual/semi-automatic and fully robotic gas-metal-arc welding. Fast-acting power-source-controller systems can monitor and correct welding arc characteristics as frequently as 5000 times/sec, as necessary.
Continuing efforts are leading to new advances in information management and technology. Merging power source, controller, and robot controls into one unit, and stepping up to 64-bit CPU speeds are setting the stage for the enhanced management of process information required to achieve still better performance. Control systems now can operate 2000 to 3000 times faster than those available as recently as 2003.
This article was first published in the May 2005 edition of Manufacturing Engineering magazine.
Analysis provides opportunities for increasing robotic arc welding productivity as never before. Measurement truly has significant rewards.