Chuck Maggi, a regular contributor to our Sherlock Ohms blog, recounts a case of a slight misapplication of wet slug capacitors:
“A long time ago I worked for a military communications company, which was owned by a large commercial telephone company. The project that I had was to design a large “tactical” telephone switching system which would be deployed in the field in a large shelter.
Hundreds of field telephones could be “wired” to this massive switch via field telephone wires - hardly tactical in the true sense of the word. As the chief scientist from the Army once told me - “How could something that needs the power from Hoover Dam be considered tactical?” Of course my reply was it could not be deployed as conceived. But the project was funded anyway.
The several hundred line cards, which were needed to interface to the multitude of tactical telephone sets (some were digital), consumed a lot of power - several amps if my memory serves me right. Of course each line card (circuit board) needed decoupling and filtering of the power going to each of these as well as the computer and switching matrix circuit boards. All tallied up to several hundred amperes of dc power per shelter.
My main problem that I had at the time was that the chief engineer on the project insisted on using wet slug capacitors throughout the system. His reasoning was that when they failed, they would fail short and thereby we would be able to find the failed part. His argument was that if a capacitor failed open, we would not know it and the problem might surface as hum or noise in the system and we would never be able to locate the offending capacitor. I reluctantly gave in to his demand.
The prototype system was put together in large racks with the huge, fault clearing, series regulated power supplies mounted on the rear doors of every rack. These power supplies were so heavy that when you opened the rear doors, you had to have extra support to hold the door upward, otherwise the entire rack would fall on your head. These supplies had remote sensing attached to the backplanes of each card shelf. The time delay associated with the response to a changing load was a little slow due to the length of the wires and the technology employed in the supplies.
The system was designed and debugged with little consequence other than the normal ones. The completed system was finally shipped and reconfigured at the customer’s site for acceptance. The coup-de-grace came with a phone call from a fellow engineer who was deployed with the system. It sounded like he was standing inside of a popcorn machine. All I could hear was resounding loud pops followed by more pops. He had to leave the shelter to avoid being covered with tantalum material. Apparently, one capacitor decided to fail by shorting. This caused the fault clearing power supply to increase the current, which of course blew the lid off the capacitor. Once this offending short was cleared, the voltage of course shot up very high and then this caused another capacitor to decide to fail due to excessive voltage, and so on.
The next week I delivered a large box full of a few hundred damaged circuit boards to the chief engineer and told him that he was correct. We were able to find the bad capacitors without much problem at all.”
(Chuck Maggi earned his BSEE from PSU in 1967. A self-proclaimed “old and tired electrical engineer,” he has worked in oil fields, aerospace, communications, consulting, spooking, and now into asphalt and viscosity in his waning years. He is now back into the exciting world of amateur radio after a 40 year hiatus as N3CRM. )