The only ones I can think of are...
Back in the day when RAM came on chips, instead of on DIMMs, my brother was installing an upgrade to his, I want to say 386 16MHz (it may have been an early 486, I don't recall exactly). I think it was from 256 to 512KB of RAM. The conversation...
"Hmmm, I wonder if this have a "right way" to go in. I mean, they fit both directions..." 10 seconds later and the power button is flipped. The cloud of smoke that immediately started emminating out of the case told us that, yes, they had an insertion polarity that needed to be followed (after the fact, we noticed the tiny white dot on each, helpfully on the UNDERSIDE of the RAM chips). Amazingly, after much coughing, choking and yanking out of power cords...it still worked once we swapped the chips around.
Next one, more smoke. Around abouts 2000 helping a friend upgrade an old machine (Pentium 166, or actually might have been a K6-166) with a new SCSI HDD in it. Again, polarity is queen. No keying on the cable, inserted it wrong, powered it up...all was fine for a few seconds, and then cloud of smoke coming out of the case. Yanked the power, one of the wires melted clean through the sheathing on the SCSI cable. Reversed it, continued to work (despite the wire being bare. We just wrapped the whole thing in electrical tape so that the bare wire couldn't contact any metal in the case).
Final hijinx. A friend was working on an old CRT monitor that had ceased working. It had just been plugged in. I warned him that before monkeying around in the thing, he might want to wait a half an hour or so for any capacitors to discharge. He thought I was being silly. His screw driver crossed a capacitor. The resulting current discharged caused his muscles to contract so hard he lept backwards (involuntarily) 5ft in to a concrete wall and then lay on the ground for about a minute before I could get anything coherent out of him.
I'll admit about 45 of those 60 seconds I spent doubled up with laughter. Suprisingly we are still friends.
*Edit* A couple of post-posting corrections supplied by my brother who both did the surgery and remembers it better. Apparently it was the L2 cache (now THERE are memories, off-chip replaceable L2) and a 486sx25/33. Also apparently it was depressions on the chips, not white marks. Otherwise, same lead in and result.