October 19, 2004
Upside Down... Let's take five with Moira Gunn. This is "Five Minutes."
As soon as it happened, people started asking me what I thought. The incident in question was the crash of NASA's Genesis probe right smack into the Utah desert. I had one uniform response: I need to know more before I can begin to comment. And now, a little over a month later, NASA actually has some answers.
It turns out that four switches, whose job it was to invoke the parachute sequence, were installed upside down. To be clear, these switches weren't installed wrong — the technicians installed them exactly the way the specifications indicated. The error was in the design documents.
I know for a fact that many smart, educated and reasonable people just don't believe this should happen. They say — usually with a touch of indignance — "Can't you just look at it and see that it's upside down?"
And the answer is quite simply "No."
--
There's a reason why the classic image of the engineer is one in which he is staring at his hand — his thumb, index finger and middle finger all aimed in different directions. You see, he's trying to imagine how — what he's thinking about — will play out in three dimensions, in the reality of the physical world. This is an everyday challenge for working engineers.
You strive to get it all worked out, but until you move into the real world, a drawing of a switch is not a switch. It's a concept. An idea. A representation.
And once everyone reviews and signs off, the rest of process often doesn't help to uncover the problem. A Quality Assurance person — and probably at least two back-ups — will all look at the drawing and inspect how the switches are installed, guaranteeing that the endproduct is built exactly to spec. Only the spec starts out wrong, and steadfastly stays that way.
Having moved from engineering concept to reality any number of times, I've come to realize this ability is hard to come by.
--
I was once asked to build an assembly line robot for a large international corporation, and since it had a burning timeline, they asked me to begin programming before the assembly line was actually built. Everything — except the robot — was pretend. I pretended there were trays of product trundling into my robot's reach, I pretended I had a laser to measure the surface of pallets moving by, and I pretended there was a master assembly line computer signaling there was a jam. I pretended everything I could think of, when the manager informed me that I was never going to see the assembly line. To reduce costs, they were installing it directly into the plant some 3,000 miles away. The following week, we would ship the robot, turn it on and everything would work smoothly from there.
When I balked at his rosy take on how this would all come together, I was soundly chastised. To make a long story short, reality set in when the assembly line didn't exactly work the way we imagined it would.
Even with simulations on Earth and testing galore, sometimes you don't know until you get there. I can easily see how the problem with the Genesis spacecraft happened, and I can also believe there's no simple solution.
When it comes to things we build for space, it's astonishing to think how successful we've already been.
I'm Moira Gunn. This is Five Minutes.
Back to Five Minutes List
