I was working for DEC when the VAX first came out. Most of the system software development took place in Building #3 of the old mill in Maynard, Massachusetts. This building was there during the American Civil War and they made blankets for the Union Army. Everything except the outer walls was made out of wood, including the massive beams and all the flooring. The floor sagged in between the beams from the weight of the old wool-processing equipment, now long since removed. There were Maximum Floor Loading signs posted in various places. I remember quite a few said "25 pounds per square foot". I guess we were not supposed to stand with our feet right together. If you watch the movie "Pajama Game" you can see a building that looked a lot like ours, except inside we did not have flat studio floors, but wavy wooden ones. An advantage of a wooden building is that it is easy to run new conduits and there were actually factory "carpenters" who did this sort of work. Need a new staircase? Sure no problem.
Anyway, the computer lab was on the 3rd floor and in the early days it was the only air-conditioned space in the building. The prototype and first-generation VAX 780 processors were very heavy so the building engineer did some calculations and directed that these computers had to be placed directly over the floor beams. This resulted in a comfortable amount of space between the racks, but no raised flooring so the entire room was kept quite cold. The very first 780 prototype had hand-built wire-wrapped connections. The next phase were manufacturing prototypes with soldered motherboards. These motherboards were huge - at least 2x2 feet each - and each machine held two of them. This is the hardware that VAX VMS 1.0 was developed on.
So there was a trade show and it was decided to show off one of these. One was carefully disconnected, wheeled over to the cargo elevator (the only kind of elevator in those buildings actually) and taken down to a truck for the trip to the trade show. Everything was fine so far. But on the trip back from the tradeshow (as I recall it - I was not actualy present when this happened) either when loading the machine into the truck or taking it oiut, the whole thing was dropped. Not real far, but enough that the connections between the parts of the motherboards were cracked. All of them. II think it was a write-off.
So we were down one development machine until the next one came out of manufacturing.