At work, we've been battling trying to find out the causal factors for an incident that occurred on our gear during routine maintenance. What happened was that an electronically controlled winch spontaneously ran when it wasn't supposed to do. We immediately downed the gear, brought it back home (Aug. 29), and trouble-shot the incident. Until today, nothing that we did could duplicate the incident. Almost everyone on the team believed it either didn't happen or that a combination of events that was so improbable actually happened. We could not come up with that combination of events.
When testing first began, I (along with other senior engineers) were precluded from being involved in test development or testing, as we were too close to understanding the gear's operation. That kinda pissed me off, but okay...that's the rules.
When testing basically concluded yesterday without any progress, I pointed out an incident that happened immediately prior to the actual winch incident, and that idea needs testing. Basically, the software detected a voltage fault and stopped the winch, requiring a reboot after correction of the problem. When they rebooted, the winch ran unexpectedly. I hypothesized that the voltage fault and how it was handled in software may be the source of our running winch. Nah...we already tested for it, they said. But, did you test it under these conditions? Noooo....so today they did.
And guess what folks? We were able to exactly replicate the winch running incident. And who suggested testing for this? Moi! Yes, I am the Man! Four weeks of testing without looking at the most obvious thing prior to the accident.
The Navy base has ordered building contractors to widen all the doors, as my head cannot clear them as of today. :tigg
When testing first began, I (along with other senior engineers) were precluded from being involved in test development or testing, as we were too close to understanding the gear's operation. That kinda pissed me off, but okay...that's the rules.
When testing basically concluded yesterday without any progress, I pointed out an incident that happened immediately prior to the actual winch incident, and that idea needs testing. Basically, the software detected a voltage fault and stopped the winch, requiring a reboot after correction of the problem. When they rebooted, the winch ran unexpectedly. I hypothesized that the voltage fault and how it was handled in software may be the source of our running winch. Nah...we already tested for it, they said. But, did you test it under these conditions? Noooo....so today they did.
And guess what folks? We were able to exactly replicate the winch running incident. And who suggested testing for this? Moi! Yes, I am the Man! Four weeks of testing without looking at the most obvious thing prior to the accident.
The Navy base has ordered building contractors to widen all the doors, as my head cannot clear them as of today. :tigg
