In another life, I was a mainframe field service engineer. You know, like in the old movies, room full of whirring tape decks and ominous boxes with twinkly lights, punchcards, 24 hour workload.
One of our babies, maybe once every couple of weeks, crashed at breakfast time. Never the same crash, analysis of many dumps just showed some random corrupt instruction in progress.
A lot of effort went into elimination and trying to identify source even before the pattern became evident. Any time between 0730 and 0830, but any random day.
Focus shifted to external causes - monitoring for mains power glitches, despite an isolated supply via motor/alternator set. The site did have serious machinery like gantry cranes, after all. Nothing.
After a year we got the companies chief designers involved, who were of course very reluctant. Not only “Stupid field engineers have missed something” but “We don’t even make that model anymore”. Corporate Sales team - who wanted to sell a million-pound replacement in future! - were able to insist.
Designers were as clueless as us field monkeys of course, but able to call on better resources. After repeating most of what we’d already done, someone borrowed a serious RF wideband monitor thingy. If I recall, a sort of didn’t really exist military secret thingy.
After a couple of days sat in the corner, this monitor caught a wideband radio scream - a nasty multifrequency blaargh - at breakfast time. Just a few hundred milliseconds. That day the mainframe did not crash, but this was at last a clue!
No-one could offer any real protection technique. The computer room was already supposedly RF screened to an extent, our supply was isolated, all the equipment cabinets properly earthed, etc. We’d been through all this already.
The design team wanted the whole building rebuilt as a huge Faraday cage.
The customer however insisted on finding the source - this was after all a government tech establishment during the cold war. It escalated to mysterious well-suited visitors who would not sign the visitor’s book because they were never really there. Whatever resources it was they deployed, the culprit was found. No Soviet spies.
About a mile away was an airfield, which did not operate at night. Each morning the tower staff switched on their ground radar. Maybe before or after coffee. The radar set had a fault. It was not supposed to connect to the dish antenna as it started up … because that would squirt out a powerful pulse sweeping across most of the RF spectrum, as some oscillator started up. Most days, the rotary dish would not be pointing in our direction when that happened…
It must have made a “click” on everyone’s radio and TV for miles around, but who would notice. Apparently the radar tech replaced some delay module costing pennies.
Which is a long story to say - @arden has done the hard part, identifying the source.
It is possible there is actually a fault in the washer - it is really not uncommon for the suppressor factory-fitted at the power inlet to burn out. It doesn’t affect operation, just leaves a smell. I have actually seen two washing machines like that.
Another maintenance issue could be worn out brushes on the motor, causing arc-sparky.
It is worth having a close look.