Re: 17 years
I totally can echo that sentiment (the car parts) - however considering the economy involved we are out of luck. People[*] want cheap and buy cheap, and you get what you pay for.
Great that scientific instruments are not built that way!
--> I toast the engineers involved in building the equipment!
[*] most, but admittedly not all, and some of us actually think long and hard about a purchase, but in the end there is an intersection between the lines of "costs this much to manufacture for that reliability" and "I'm willing to pay this much over the projected lifetime - or in total". That sweet spot is different for all of us, but I am pretty certain that the probability distribution can be modelled as bivariate normal (cost, reliability), as a decent approximation - it is so far from the origin that the needed cut-off at zero cost can be safely ignored[+]
[+] this shows that I am no mathematician
 Edited to add: and Sperical Cow (who judging from the handle might be a physicist?) is totally right about the ease of swapping out parts. You are willing to spend more and overengineer more when you cannot just swap out parts, because it is inconvenient or impossible.