Before I retired I did failure analysis – figuring out why things busted or rusted. I just looked through the Tales from the Cube blog at EDN Network and found a bunch of interesting stories. But many of them have enough engineering jargon that they are difficult for an outsider to understand without looking at Wikipedia or elsewhere. (EDN stands for Electrical Design News, a publication that began back in 1956).
One of my favorites is Orin Laney’s February 5, 2015 story about Fixing a mainframe with a lunch bag. Back when he was just eighteen he had a summer job on the overnight “graveyard shift” as a supplemental customer engineer at an IBM computer center. One of their mainframe computers, a System/360 Model 50 would not turn on properly. Mr. Laney read the manuals on the power supply circuits and started checking the control relays.
As shown above, a relay just is an electromechanical switch. Applying a current to the coil (1) attracts the armature (2) and switches the moving contact (3) between fixed contacts. The fourth or fifth one was faulty – the blackened, dirty contacts moved, but no power came from the fixed ones. He knew that the contacts needed to be cleaned, but didn’t have a burnishing tool for rubbing them off. (That’s a very mild version of a file, as shown in Figure 7 of this article on Hard to find maintenance tips for electromechanical relays). But in the customer engineering room he found a discarded brown paper lunch bag. He tore off some strips, put them between the dirty contacts on the offending relay, applied power to close them, and repeatedly pulled them through until the contacts were clean enough to conduct electricity again. That improvised repair is an example of jugaad, a Hindi word that means:
“to make existing things work, or to create new things with meager resources”
A couple other stories involve external alternating currents getting in due to insufficient shielding. One is Paul Mathews’s May 31, 2012 story Day at the races, where the current came from a radio station located a mile away. A second is John Loughmiller’s January 21, 2010 story Brown’s buzzer busts business at a broadcast studio where both the remote start control for a videotape recorder and power for a doorbell buzzer connected to a button on their loading dock wall used unshielded twisted-pair cables sitting in the same cable tray. When the United Parcel Service driver rang to deliver his packages, the recorder malfunctioned.
Stanley Pitman’s June 1, 2017 story titled False alarm involved an unexpectedly large signal to a radiation detector for a mill’s incoming train track spur located about 100 feet from the main line. It turned out that false alarms occurred every time a carload of bananas came down the main line. Bananas contain a small amount of naturally radioactive potassium, and there even is a Wikipedia page titled Banana equivalent dose. To shield that detector they had to put up a concrete block wall between the main line and their spur.
Another unusual mishap is Douglas Forst’s August 25, 2011 story Going against the grain dust, where an unintentionally insulated replacement return roller on a conveyor belt became a Van de Graaff generator for high-voltage static electricity, creating half-inch long arcs.
Sometimes the wrong materials or processes accidentally get used during manufacture. In David R. Bryce’s March 17, 2011 story titled Acid test, the wrong type of room-temperature-vulcanizing (RTV) silicone was used to seal the ends of ultraviolet lighting tubes. The page on Wikipedia about RTV silicone mentions that some types release acetic acid during the curing process. That’s fine if you are caulking around your bathtub, and just get a temporary smell of vinegar. Sealing acid vapors inside an enclosure can corrode copper wires, lead solder, etc. (Two decades ago I ran into this problem in a load cell used for weighing trucks).
Images of an IBM System/360 Model 50 computer and a schematic relay came from Wikimedia Commons. The image of bananas came from the National Cancer Institute.