“Bargain Parts” from Mark

A while back I worked as a test technician on the manufacturing floor for a company that assembled and tested Printed Circuit Boards.  We had been assembling and testing one particular board for many months with very few of the boards failing their tests the first time through.  One day we started to get failures, LOTS of failures, almost ALL boards failing — and all failing the same test.  This board was tested by connecting it into a test stand that was just a spread-open product and loading test software that was provided by our customer (we only got the hex code).

We physically examined and compared bad with good boards, including under a microscope, looking for misplaced parts, solder shorts, bad solder connections, etc.  Nothing was different.  So after a few days, the customer sent one of their engineers to our plant.  Two days later, still no solution.  I was working with him and looking at the boards and happened to notice that the printing color on the processor chips was different between the good and bad boards though the brand was the same.  I mentioned it to him and he started looking at ALL the numbers on the chips.  There was a code below the part number (and above the date code) that was one string for the bad boards and a different string for the good boards.

Documentation that he had did not explain the code, so he called the manufacturer of the chip.  It was the Step Code AND it indicated that the Step was one lower on the bad chips.  He asked the manufacturer what the difference was.  Want to guess?  — It was that at the lower step, the function that was failing testing would definitely not work.  Come to find out later that our customer’s procurement department found a “good deal” for this processor and insisted that we use these “good deal” parts.  (We normally acquired all the parts through our own procurement department.)