This story took place at least ten years ago. Among the various tasks I was assigned, I had been tagged as the heat stress expert. There was a reliability problem with cellular amplifiers we were building. The design had six amplifiers in a shelf and one would often fail before the installer had all of them running.
The designers decided that we should put all amplifiers though a multi-hour heat stress in a tent to remove latent defects. The heat tent was all ready in the plant from past work so shelves to hold the product were installed in the tent with large Hewlett-Packard power supplies outside to provide the needed 24 volts.
When we started the initialize of each amp in turn after several were started they would all flash and revert to their start condition. We spent days trying to figure out why. Finally, we called in the designers from the east coast to come in and find the problem.
To make a long story short, they spent a week with us, found nothing, and went back home telling us that the configuration wasn't one used in the field. So it all fell on me.
The director in engineering would come around to check my progress; without a heat stress there was no product shipping -- no stress there. It took me a couple of days with a storage scope but I finally caught the problem. As one after another of the amps would power up there would be a large voltage spike on the 24V line, which would knock all the amps offline.
I finally figured out that the switching power supplies in the amps were getting synchronized and causing an oscillation in the 20 feet or so of cabling between the power supplies and the shelves. The quick fix was 30,000 uf caps at the self side. The long-term solution was to move the power supplies right under the shelves. This was possible because I found that I could cause the internal temperature to rise a sufficient amount by running the amps at full power without cooling fans, and then cooling them while turned off with full fan power -- all this without the heat tent.
The system ran for several years until the product was replaced.
Timothy A. Shankland worked for most of his career in the area of product test, both in company repair centers and then factory production. The story occurred in the later part of his career when he was involved in testing and heat stressing cellular base station equipment. He is now retired in the Columbus, Ohio area and working on his own projects.
Tell us your experience in solving a knotty engineering problem. Send stories to Jennifer Campbell for Sherlock Ohms.
Design engineers, New England's premier design and manufacturing event, Design & Manufacturing New England, will take place in Boston, May 6-7, 2015. A Design News event, Design & Manufacturing New England is your chance to meet qualified suppliers, get hands-on with the latest technologies, be informed, and expand your network. Learn more here.