Quote:
Originally Posted by Qbot2640
We often re-use components if a particular robot or robot system is not going to be kept. Then, when a motor or spike fails, we never seem to know how old it is. I'd like to implement a system that records the date the part was acquired - so we could start calculating mean time to failure statistics.
|
Whenever you get new components, label each one (with a good label maker, not just sharpie) with a unique serial number. Add that to the standard inventory list as Don described, and add a field for "purchase date" as well, and voila!
Calculating time-to-failure might be harder; components fail for many different reasons. It could die after hundreds of hours of use due to wear and tear, or it could die after a single day of use because of an overheat, short, physical damage, etc.
I'm no expert on statistical analysis, but I think you would need a pretty large dataset before you started seeing a trend. If I'm wrong, someone please correct me
