start me up!

An anecdote about testing to failure

Once upon a time, while I was working for GreyPilgrim Inc., a disgruntled employee -- one who didn't agree with certain aspects of component design -- took it upon himself to test a component to failure. He basically tore the component open.

The EMMA robotic arm had elastic joints, made of polyurethane, connected to metal sections at either end. The polyurethane was intended to supply flexibility; the metal strength and stiffness. (Flexibility and stiffness -- mathematical inverses -- in the same mechanism. What a concept.) The joint was always problematic in design because of our lack of knowledge about how to properly fasten the two types of component together. And what we did -- long nails and epoxy in this case, though the design evolved later -- didn't set well with the disgruntled guy. He said to me, "I'll prove this thing doesn't work," and even as I said "I don't think you oughta do this," he tore the joint apart in front of my eyes. The cross-section looked a little bit like the picture above: once the tear started, stress increased around the tear, and the whole thing accelerated. It became easier for him to complete the damage once he'd started it.The problems were that (a) that joint was the only one we had at the time, and it was reserved for another test -- having nothing to do with failure -- that we could not perform until a new joint was built; (b) it took him so much more effort to tear the thing apart that even he was surprised. Even the fasteners seemed to hold, until he got serious in what amounted to an act of sabotage. So if something good came out of this incident, it was that we had much more confidence in the design than we'd had before. We were planning to test one to failure eventually, but not right then. Remember, if you're going to break stuff, you have to make sure "there are many copies."

What happened to the disgruntled guy? Well, he was fired. I didn't stop him doing what he did because he'd been above me on the company totem pole prior to the failure. But he'd become a kind of cancer to the company already and was on the edge of being let go. It got to the point afterwards that he'd threatened the company, wanting a cash payment for his shares, and said

If this offer is not accepted by Monday at 12:00 noon then I will make the destruction of GreyPilgrim my passion.

He did make this his passion, and although he failed in his aim of destroying GreyPilgrim, he made life a pain in the ass for the rest of us for quite a while.

Another lesson you must learn is to make sure anyone who becomes unhappy in your start-up is unable to become a cause of failure, whether of prototypes or anything else.

Labels: consulting

posted by Ron @ 9:30 AM