The world of mechanical wonders is fascinating.
My dad made his living as scientist in Institute of engineering research; they had this case of turboprop airplane blades breaking - always the third one. The reason was they were hand sanded and the toolmaker would take new sand paper for every third blade.
I've got another one for ya...
In 1995 I was working on building a fast machine that loaded catheters onto syringes (for inserting IVs into people in hospitals). This machine had to snatch 4 catheters at once, whip around 90 degrees and place them onto some (moving) needles on a conveyor, and on the return trip grab the bar the needles were in and place it 180 degrees opposite the loading site, onto an exiting conveyor. This all had to happen in 2.8 seconds. So, we picked a fast servomotor system from Allen-Bradley, the "King of PLCs" (i.e., industrial computers) in those days.
I programmed day and night for 6 weeks, trying to make the servo move. It would not. Numerous calls to Tech Support told me to do what I was already doing, leaving them baffled. Finally, angry, I called their "hit men" that they call the Global Tech Support Team, then $500 per day (1995 dollars!) to have them come from Allen-Bradley to fix your problems. I told them that if they could fix it, I'd pay it, but if not, I'd ram it up their A$$. They laughed at me on the phone, then sent one guy out.
Three days later, they had 2 guys out.
Next day, there were 3 guys, and they wouldn't even talk to me.
On the 5th day, they asked to use my phone to call Allen-Bradley, so I said 'Sure". After 3 hours of calls, the first guy who came out called me over and said, "You won't believe me if I tell you, so talk to this guy at Allen-Bradley yourself for the explanation." He handed me the phone: at the other end was a kid (he was 23 years old, a Senior Design Engineer - ?). He explained how he had just come from college to Allen-Bradley (when he was 21, graduated early, apparently a 'genius') and his first assignment was to build the servo module I was using. The specs called for a set of 24 instructions to be built into the firmware, including the sharp acceleration interrupt commands I was needing to make this fast machine work in time. He then explained how, after 5 months, the management came to his desk and told him he was done, despite the fact that he had only so far made 4 of the 24 instructions work up to that point. They put the module into Production, and sold it with a manual stating that ALL of the instructions were in there, "Just do [thus and so] and the motor will respond".
Not.
I didn't pay them, and I REALLY enjoyed ramming those servos and modules up their... I even got 20% EXTRA back on the refund. The next Monday, I ordered a Compumotor brand servo, installed it that Wednesday, and shipped the machine Thursday. (Yes, I do this for a living, don't attempt this at home...)
In much of industry since the mid-1990s, there are astounding numbers of these events occurring. People like me, who build machines with robots, computers, PLCs, blur-fast movements, and usually load explosives, noxious liquids and gases, dangerous materials and [occasionally] women's cosmetics, have come under amazing pressures the last 20 years to make this crap work. Today (2014) I would estimate that nearly 20% of it does not even function as advertised, which brings on things like the massive GM car recalls and food recalls that we all hear about. It really makes me wonder what will happen when the aging group of guys like me retire?
