During training I joke with students about the logical extension of a potential gear failure, "... even if your regulator explodes..." though I have never actually seen it happen. This is, typically, in the context of "Nothing other than 'I am currently not breathing' is an emergency. Everything short of that is a manageable nuisance."
Or, rather, I HAD never seen it happen.
It was not during training, I was simply guiding. The diver had one of those Omniswivel 360s, which I'm not terribly fond of. I've never actually seen one fail, but I've heard enough stories by people I know personally and trust that I'm wary of them.
Guiding two divers, I'm leading. We aren't far into the cave; perhaps six of seven hundred feet in. When I hear a BIG noise and see a briefly flashing light.
"That's probably not good," I think as I turn.
It wasn't.
One diver was lost in a cloud of bubbles as the other was closing distance fast. The cloud of bubbles cleared the affected diver as they turned off the failed tank when I see...
A second stage.
Spinning to the ground like a maple seed.
I don't quite catch it before it hits the ground. But the whole time I'm swimming to it I keep thinking, "Goddamn Omniswivels."
Turns out it wasn't the 360, though. The second stage had simply come unscrewed from the hose and blasted off when the o-ring came unseated. My next thought was, "Huh. Ain't that a bitch?"
As a testament to good training, the affected diver had turned off the tank, switched to their other sidemount reg after getting blasted in the face with all the bubbles in the world, and was now floating with perfectly fixed buoyancy and trim with a solid reference to the line and looking at me without fear, just a crooked eyebrow of surprised confusion.
"You OK?" I asked.
"Yeah," they shrugged with definite overtones of, "Well...... now what? We just leave I guess?"
I signalled to hold and to hand me the hose that had been whipping bubbles everywhere.
Looked inside... the o-ring was still there. Which I was almost disappointed about; I've always got an 010 in my wetnotes in case of this exact thing ever actually happening and I thought I'd actually get to use it.
I shook out what little debris had gotten into the regulator.
I deployed my Seadog adjustable wrench and my new Taran Tool. The diver continued watching with what I later learned was a sense of "What the fuck are you doing? You can't fix a regulator underwater!" And I did. I mean, I don't know that attaching a second stage to a hose counts as "fixing."
Valve on. No bubbles. Let the diver see me purge it a few times. Reg worked fine. I handed it back.
And asked, "So... you want to just continue the dive?"
They didn't.
But now I can say that I have seen a second stage explode.
And it wasn't a big deal.
(Again, as a well-trained diver capable of self-rescue it wasn't a big deal... it obviously could have been a VERY different day if they had not been as on their game.)
That's the biggest trick to cave diving: you should always have enough gas in case of an absolute catastrophe of a total loss of one diver's supply. So calling the dive is never the wrong answer.
But most problems are actually fixable. Here. Now. An extra moment to conserve a breathing supply in case of a known-to-be-fixable failure is a worthwhile moment.
Not wanting to risk ever repeating a similar experience was enough to convince them to take the Omniswivel off before the second dive, though.
I've had a primary regulator spin off the hose as I was suiting up. I noticed the 90 degree elbow wasn't tight. My idea of free-flow simulation in training is to spit out the regulator, switch to the back up and shutdown. The bubbles gives you an enhanced effect of danger.