When I was going though training to be a Planner Analyst on DSCS-III satellites back in 1987, one of our instructors told us a story about a glitch he'd discovered in the satellite design and procedures. Virtually all satellites back then used ni-cad batteries. One unfortunate property of ni-cads is that they develop a memory. If you partially discharge a ni-cad then recharge it, it may not accept a full charge. If you keep up with the partial discharge-recharge cycles, you'll soon end up with a battery that on stores a fraction of the amount of energy that it used to hold.

The DSCS-IIIs had 3 big batteries sufficient to power the satellite during eclipses (up to 70 minutes of darkness on the equinox, with lesser amounts of darkness for about 35 days on each side of the equinox [eclipse season]). Between each eclipse season, we took the batteries off line one at a time and slowly discharged it through a resistor. Once fully discharged, we'd reenable charging. The problem is that there was a resistor in the circuit rated at about 10 amps. If you turned on the battery at a full charge rate, it'd blow the resistor and permanently take that battery offline.

Our instructor had examined the power system schematics before the first launch and noticed the problem. They changed the recharge procedure to keep the current within limits and there was never a problem. Had he not caught the error, the first satellite would've been severely damaged. He got a nice bonus for that catch.

Details. You gotta pay attention to the details.