Several years ago, we designed and built a system that contained two motor controllers that were programmed from a PC to coordinate their movement with other controls in the system and data acquisition. Our fabrication group built the controllers in their own separate chassis to keep noise out of the other controllers and away from the data acquisition circuits. As usual, their efforts were flawless and the system worked as expected.
The programming was completed and all of the system components were integrated and tested. The testing went smoothly and everything continued working for several days of testing. The system was then considered “ready for primetime,” and everything was buttoned up and delivered to the customer’s lab. The next day we got a call from the researcher who said that one of the motors quit operating properly after a few hours of use.
We brought just the motor controller chassis back to the shop and proceeded to use our PC-based test program to troubleshoot it. As Murphy would have it, everything worked fine for several hours. Since we could not get it to malfunction, we buttoned it up and returned it to the user. Again, we got a call that the same motor stopped operating after a few hours. So, it came back to the shop for more troubleshooting, and again we were not able to duplicate the problem. This time we even ran it overnight. Back to the user it went, and again we received a call that it quit after a few hours.
This time we switched the motors to the other controllers to see if the problem followed the controller. It didn’t. In this instance, it was the other motor that stopped functioning; the problem was following the controller. Our next step was to replace the controller. The system worked fine in the shop, but when we took it back to the lab, it didn’t work.
After some head scratching, we realized that when we were testing it, we had the cover off the chassis, and when we delivered it, the cover was on. Aha! Heat buildup. Something we should have realized from the beginning, so we drilled a few vent holes and tested it with the cover on. Success. Apparently, the vent fan was drawing cooling air from places that did not include around the one controller. Lesson learned: test under actual user conditions.
Thomas R. Clem, Sr. is retired from the National Institutes of Health. He worked as a consultant and contractor for 12 years after retirement. Thomas graduated from the University of Cincinnati in 1963, and took some graduate courses at Purdue University.
Tell us your experience in solving a knotty engineering problem. Send stories to Lauren Muskett for Sherlock Ohms.