velocity positron annihilators.
I'm reminded of a long weekend, ~1967, ten of us from IBM Kingston (NY), spent at Stanford University trying to "exorcize" an IBM host (S/360-67) that had been getting excessive "machine checks". It was a little red light but somewhat like the "Blue Screen of Death". While we found things that could be tweaked and replaced, at the end of the weekend, we were defeated. It was back home and to the drawing boards.
The problem was a combination of things. Stanford had the finest glossy raised floor money could buy, fantastic electrical insulation. The high speed printer was used without metallic discharge braids, so the continuous paper was quite charged up. The machine operator would of course unload the printer paper periodically. The operator's chair had fiberglass fabric. That's good for building up a charge as well. The computer console was getting zapped with 7-8 KVolts on a regular basis. Sometimes, it was too much and the machine would fail.
Until a proper solution was found, and it was, Stanford had to use a conductive mat in front of the CPU, an old wood and leather chair for the operator, and of course metallic discharge braid put in the output path of printer.
The fix (IBM called them ECs, engineering changes) was put into production and propagated to all similar machines in the field.