We had a new 4 inch scale Foster traction engine in to test recently. It had been built some years ago, but never steamed - having recently been painted, the builder's small grandson wanted to see it go.
He delivered it to the workshop a couple of days before boiler testing, we had a couple of hours looking it over and gave it a hydraulic test in advance of the boiler inspector's arrival - everything looked fine. Come the day, he pumped up the boiler, declared himself happy and we put a fire in the thing for the first time stood just outside the workshop door. It came up to steam readily and ran nicely. Injector was fine, pumped worked nicely, all we had left was to check the setting and efficacy of the safety valves. With a deep fire and blower full on, pressure came up to 120psi, first one safety then the other opened when, with a great roar the engine disappeared completely in steam, roaring out with not inconsiderable force. I've never seen Ian move so fast!
When the steam cleared, the engine appeared to be the right way up with all important bits still attached. The grate, which three minutes before was four inches deep under a roaring fire was completely clean - there was not a crumb of coal on it, ditto the ashpan. There was some very wet coal in the spud pan, but the vast majority of the fire had shot straight through the workshop and hit the back wall beside the lathe - David, who was in helping for day, had actually seen it whizz past his boots.
The culprit was the fusible plug - it had been filled with soft solder, which is a bit marginal at the best of times but inadequate at 120psi operating temperatures. It wasn't helped by not having wetted the plug - it came out clean like a cork. However, the real problem was the builder having made the plug with a half inch hole in it - it was this produced the rush of steam which launched the fire into orbit.
We let the engine cool down while Geoff made another plug with a more sensible-sized hole, we filled it with lead and got it back in within the hour.