On Fuses and Their Uses

Or, Zappy Zappy, Joy Joy

I believe it was having a Fluke meter that did it.
Honestly, nothing is more of a magnet for damage than an expensive piece of equipment, and multimeters are far from a counterexample to this theorem.
Back when I was in my freshman year at NJIT, I was friends with a fellow named Garrett. Somehow we'd kind of glommed together a research group of sorts, and also gotten ourselves a lab to play with.
Now it happened that, at this time, a certain 3kV power supply happened to meet with an atmospheric pressure Helium plasma chamber that we'd devised. Fun was had, and to check our threshold voltage, we used the lab tech's Fluke multimeter.
Boy, but that was a mistake.
I was manually turning down the feed volume of Helium, letting the small leak around one of the electrodes do the work of producing a smooth resistance gradient over time, when the periodic sparks from anode to cathode got less and less frequent... and then stopped.
Now, normally this would be a good thing, but we still heard the sparks discharging, and they were getting faster.
Then I look at the multimeter, and the pulsing bright light coming from behind the plastic face plate.
"Garrett!" I say, "Shut it down!"
He shuts down the power supply, and, without thinking to wait for the bleeder resistors to discharge the supply's huge capacitors, I grab the jack going into the multimeter and pull hard.
Or at least, I was going to.
A spark leaps from my left pinky knuckle to the metal chassis of the power supply, a quarter of an inch away. My other hand clutches harder on the banana plug, with its semi-exposed contacts.
The power supply had been at 3kV at 10mA. My approximation is that I got about 1.5kV at about 1-2mA.
Lesson here: the orchestra has it easy; it's the conductor who has it hard.
Oh, and as for the Fluke meter, well, expensive equipment usually has the advantage that it's well designed. We only had to replace a fuse in it for it to work again perfectly with no lasting damage. Our diagnosis was that one spark ionized the static air in the meter case, allowing for the spark to effectively create a continuous arc, shorting out the multimeter. Two bits of soldering, and it was fixed.