Frank here,

Cracks in the welds, what in the world does this have to do with robotics?

Long ago I worked in a large corporation in Silicon Valley as an Equipment Maintenance Technician.

My department was in charge of maintaining production equipment. One of the processes used Laser Welders. The machines were used to spot weld small stainless-steel parts together which eventually became the flexures on which the read/write heads for disk drives were mounted on. All production for these parts were run through just two machines.

The process had been running fine for a few years when suddenly, one day the QA Department shut down the process through the two machines and stopped an entire production line completely!

We were informed that small microscopic cracks had been found in the spot welds and now we had to determine what had changed in the process, specifically in the Laser Welders which were causing these “cracks” in the welds.

Everything had come to a screeching halt while the incoming parts were inspected and re-inspected. The flexures went through destructive testing where they were literally ripped apart to test the strength of the weld. The welds were cross sectioned and cut with a diamond blade and the welds were examined and we were shown pictures highlighting the tiny cracks in the welds.

Our team dismantled the Laser systems, rebuilt the optics, and YAG Laser cavity ellipses but could find no flaw or failure in the systems. The power level of the Lasers had not changed. The depths of the welds were the same as they had always been. Eventually someone thought to test and cross section flexures that had been assembled a year before. Surprisingly, these older parts were found to have the same microscopic cracks. Hmmm! What had changed? What was going on?!

Well, unbeknownst to us, the QA Department had acquired a new microscope which was able to look at the cross section of the welds under a much higher magnification. The cracks in the welds had always been there. They just were not seen before. The cracks were a non-issue. No parts were failing in the field. The production line had been shut down for weeks all because of a new microscope.

So, why did I bring this story up? Quality and reliability are extremely important. Sometimes, however, we over test or over think an issue that is not relevant in the real world. Test procedures should be there to catch problems before they occur or to prevent premature failures but unless you can show data which correlates to how a machine is run in normal production, you may be looking for problems that do not actually exist.

Featured image from: