I would have thought they would run every part until it died so they know roughly what it's lifespan is. It's good if the parts lifespan is twice as long as you need. However, it's even better if you could make it four times the amount you think you need, as long as it isn't horribly expensive.

I would tend to assume they've thought of that though, and stuck with this part for reasons. Maybe there werent other space-ready alternatives or something.

I don't think any of us are in a position to be too critical of the engineering done for the Voyager probes. The number of machines running continuously for longer than I've been alive with no maintenance is pretty short; the number of them out past the orbit of Pluto even smaller.

This is actually a fairly egregious oversight, but it is good to know they fixed it.

It's impossible with current technology to make a vacuum as good as outer space on Earth, which hampers testing of components in realistic conditions. See also Galileo's antenna.

When you are weight constrained the way a space probe is, every factor of safety above the bare minimum is horribly expensive. With no risk to human life and a skilled team of engineers on the ground, the cost of further risk reduction easily may not be worth it.

Maybe the expected number of times it would be expected to spin in space (or twice that to make sure it's working)

The initial Voyager program already ended, and they are now in an "extended" mission.

640k

Maybe as many times as needed to do the minimal in-solar-system observations that were planned well in advance?

On the other hand, if they had to budget for a second copy of everything in advance they may have never gotten to launch. Asking for the funding after already having good success behind them is likely a lot easier.