Welding with CO2 lasers is common in all kinds of manufacturing environments. Despite its general reliability, changes in a weld from beginning to end as the power of the laser diminishes can create a problem.Figure 1. The beam profile at the beginning of the run had a high central peak.As a result, CO2 laser welding was not good enough for a manufacturer of metal assemblies used in an aerospace application. The company found that, if the power were optimized at the beginning of the weld, the intensity would be too low to achieve adequate welding at the end. Conversely, if the laser power were optimized for the end of the weld, the beginning would be overheated. When it came to reliability testing in the application, components were failing and were rejected. This resulted in scrap parts, excessive testing time, increased production time and higher costs.To address the problem, the manufacturer analyzed the laser's performance using a camera and beam profiler from Spiricon Inc. of Logan, Utah. The CW laser used the camera's chopper to achieve about 15 frames per second, yielding about 360 data points in a 24-second weld sequence. The data of a complete run were saved, and the measured parameters were exported to a spreadsheet for analysis. Figure 2. At the end of the run, the peak fluence of the beam was much lower, the fluence of the shoulders of the "annulus" of the beam was higher, and the beam width was smaller.The results demonstrated that the beam profile changed dramatically during the weld period. Initially, the beam had a high central peak (Figure 1). During a run, the peak nearly disappeared, but the shoulders of the "annulus" part of the beam increased in intensity (Figure 2). The beam width also decreased slightly.What effect these changes were producing was not immediately obvious, so the manufacturer analyzed the data from the spreadsheet of measured characteristics (Figure 3). A startling change in the total power was apparent, with a drop of 16 percent by the end of a run. Moreover, the beam width squared, which is proportional to the beam area, dropped by as much as 18 percent during a run. Because a focused spot size is inversely proportional to the input beam width, the focused spot would increase in area by this 18 percent. With a larger focused spot area, the intensity (in watts per square centimeter) also declined by 18 percent. Figure 3. During the welding run, the total power dropped by 16 percent, and the beam diameter squared dropped by 18 percent. Because of lower total power and a larger area after focus, the average fluence at focus dropped by 32 percent.Investing in profilingTaking into account both the drop in total power and the larger spot size, the focused intensity dropped 32 percent by the end of the run: by 16 percent in the first 5 seconds, and the remainder during the subsequent 19 seconds. With this dramatic intensity change, it was no wonder that welds were not consistent from beginning to end.Although a simple power meter could have measured and plotted the power change in this laser, the profiler offered the additional ability to capture the data needed to evaluate the change in the beam width and to discover that the actual intensity changed by 32 percent, rather than 16 percent. Also, the profiler provided the details of the instantaneous beam shape, enabling the manufacturer to evaluate whether the "hot spot" in the center of the beam and the changes in shape were relevant.The metal assembly firm was planning to purchase a new laser. After this test, it wanted a built-in beam profiler as part of the solution. With the profiler, its staff can evaluate the effect of "tuning" the laser to see if it is possible to obtain more consistent operation. Traditional beam profiling using acrylic blocks or burn spots on metal, paper or wood does not provide the instantaneous picture and temporal tracking that are needed to observe such dramatic changes in the beam. Once a user has "optimized" the beam, it still may not be sufficiently consistent to provide acceptable performance.Knowing the changes in the beam characteristics with time, a user may develop an alternate weld sequence. For example, one option for the laser described above could be to increase the run time to 29 seconds, with the first 5 seconds directed to a beam dump rather than to the part to be welded. With the 16 percent change during the first 5 seconds not included, perhaps the remaining changes could be accommodated in the welding design.Adding a beam profiler to a new or existing laser is not an insignificant investment for most industrial manufacturers. The increase in productivity, however, makes it an investment that will pay for itself in better quality parts, less scrap and waste, and reduced testing time.