I hope spaceflight posts are welcome for now.
On 2 July 2013, a Russian Proton rocket had a spectacular crash on takeoff (original video was here with the original watermark but in 2x slow-motion). Anyway.
There's a story that a technician deliberately hammered in a set of sensors upside down, despite a system that should have prevented installation the wrong way. John Palmé @JohnPalme tweeted to set the record straight.
Not true - I ran the US Failure investigation.
The Yaw rate sensors were a blind install with off center pin holes that were supposed to make sure that the unit was installed in correct orientation. Unfortunately the pins on the mounting bracket were only press fit...
...so when the new tech installed them blind by reaching through an access hole, he slid them onto the studs, and then torqued the nuts down on the studs. The act of torquing the studs drove the press fit pins back in the bracket...
...and allowed the unit to sit flush on the bracket. The debris of the units showed the mark on the baseplate that the alignment stud made as it was driven back into the bracket. Root cause....
They were building so many Protons, they used a secondary horizontal stand at the factory for building the stage. The primary stand only allowed 1 orientation of +Y up so the install documentation did not ask for orientation verification. The secondary stand allowed...
either +Y or -Y up, and this LV was in -Y up. The tech installed the unit per the documentation, but the stage was inverted. The quality person looked through an access hatch after install and verified the orientation arrow on the unit was pointing up per the documentation.
The tech was new so did not realize the force that was needed for the nuts was atypical, and the quality documentation did not ask for verification that alignment studs were in the holes.