Of course, Gettysburg's true moment of infamy was the June 16th, 1995 crown sheet failure on CPR G5 class Pacific #1278. The #1278 had been restored and operated at Steamtown USA in Vermont and had made trips across PA with Ross Rowland's High Iron Co throughout the '60s and '70s, but when Steamtown made the move to the new Scranton location, they needed something with the tractive effort to climb the grades to the Delaware Water Gap. So in 1986, Steamtown traded the #1278 to Gettysburg Railroad for Canadian National S-1-b #3254, which Gettysburg Railroad had acquired in 1982 and only run for a year and a half before finding it to be too big and heavy for their tracks.
While hauling a regularly-scheduled dinner train, the top, or crown sheet, of the firebox collapsed and dumped water into the firebox, right as one of the two fireman aboard was opening the firebox door to shovel in some coal. The water instantly flashed to steam and blew out through the firebox, burning all three of the crewmembers. The fireman jumped clear of the cab, while engineer Jim Cornell dumped the brakes and then hopped out himself. I've read an account of someone who was on the train that day with his daughter, and the train all of a sudden came to a halt. They then saw the two fireman, covered in soot and with sheets of skin hanging off their arms, come tumbling out of the cab. The conductors/car hosts never said a word about what was going on, one of Gettysburg's diesels showed up, the #1278 was unhooked and dragged off and the diesel took the train the rest of the way like nothing had ever happened.
The FRA launched an investigation, with Valley Railroad's chief mechanical officer, David Conrad, being brought in as an expert. The list of mechanical issues with the #1278 were pretty staggering. The determined cause of the failure was from the crown sheet overheating, due to a lack of a sufficient water level being obtained, and the crown sheet then softened and ripped away from the crown staybolts. At the time of the incident, the lines to the water sight glasses were almost completely blocked with scale from a lack of cleaning or water treatment. The sight glass levels were not reading correctly as a result, only fluctuating about a 1/2" in the sight glass. That should have been a giveaway, since 2-4" fluctuations are normal, but none of the crew had sufficient training to know that was an issue. Even if they had been trained sufficiently or the sight glasses had been working properly, the dynamo was no longer functional, so none of the gauges were lit in the cab, making them impossible to read, especially at night. The water injector was also not functioning properly because it had been improperly repaired. The line from the feedwater heater to the boiler had sprung a leak, so when they were trying to add water to the boiler, a good portion was leaking out. The flow gauge for the injector (what adds water to the boiler) had failed, been replaced, failed again and then the owner had just decided to forgo it entirely and thus had no idea how much water the injector was moving, or if it was working at all. The injector had the incorrect brass disc installed, with too small of an opening, causing it to inefficiently add water to the boiler. A thick layer of scale was found inside the boiler, which causes the water to not cool the firebox properly because the scale acts as an insulator, due to insufficient water treatment facilities and none of the crew knowing how to blow down the boiler.
Fortunately, due to Canadian Pacific's design of the crown sheet on the G-5s, the crown sheet didn't collapse entirely. If it had, that would have possible resulted in a true boiler-off-the-frame boiler explosion. As bad as the crown sheet failure was, at least it only harmed the crew. A true boiler explosion likely would have killed many passengers. The resulting investigation caused the FRA to completely revamp the entire steam locomotive inspection process, rolling out the modern 1472 process in 2000. Gettysburg Railroad never recovered from the incident, rightfully shutting down in '96, and the #1278 was eventually sold to Age Of Steam Roundhouse, where it remains a display piece. While there are those that complain that the new process makes operating steam too expensive, the reality is, that if you can't afford to run a steam locomotive in accordance with the new rules, you shouldn't be running it anyways. This fortunately remains the only incident of this severity in the US in the preservation era.