[jcims]

I wonder how many of those engineers have woken up, wild eyed in the middle of the night, as their brain excavates some random issue or loose end that surfaced and silently trailed off years ago, never to be properly resolved before launch.

[proggy]

It’s happened. Worst one I’ve heard of is when an engineer got an idea of how things could go south just days before the final decision point, ran to the testbed, tried it out, and their worst fears were proven right. They took it to the team, and after much wailing and gnashing of teeth, everyone collectively decided they were more confident in the level-headed plans they had executed to validate the system over the past 5+ years than this last minute, panic-induced test. So...they decided to let the sequence run as planned, and loaded up on a double helping of peanuts for the big day.

Well, the big day comes and guess what? The spacecraft performs perfectly. Everyone was absolutely ecstatic.

It turns out the test bed had been misconfigured, and there was nothing wrong whatsoever with the spacecraft or any of its command sequences.

Really makes you think about how to make the best possible decisions in a crisis.

[jcims]

I've had microscopic renditions of this kind of thing in my career. There's a cognitive 'dolly zoom' effect where there's an illusion of growth in a given risk because the field of view has been narrowed by the preconditions and/or assumptions that led you to consider it to begin with.

It's like the risk of losing a winning lottery ticket. Broadly the risk is negligible, but when you zoom into the precondition of having won the lottery, it's a big fucking deal.

[bhupy]

This is fascinating. Do you have a link with more info?

[dmos62]

What's the lesson in that story? I would have thought that you'd postpone until having fixed the problem exposed by that last test.

[codeduck]

This is only really a viable option prior to launch, of if the probe is coasting and you have enough time.

If the probe is already near its destination, you have a very narrow window during which you can adjust its velocity for intercept.

If you delay too long you do not have enough deltaV to intercept, and either miss your target entirely, or end up in the wrong final orbit with no ability to correct your inclination or eccentricity to what your mission parameters call for. Then you've basically spent years to send a paperweight to Mars.

[ArnoVW]

If the issue requires physically accessing the payload (which has been prepared and packaged, and is sitting on top of a huge rocket) it's not always that easy.

The minimum-energy launch window occurs every 2 years, and lasts a couple of weeks.

[Taek]

This would make an amazing blog post.

[nuccy]

The Cassini–Huygens mission had exactly such a moment [1,2]. A year or so after it was lunched somebody pointed out that Huygens and Cassini communication may fail due to a flaw in handling of Doppler shifted radio signals on Cassini, which should relay data from Huygens. Luckly the spacecraft used a Venus gravity assist and had a close encounter with Earth again, so the issue was tested and confirmed. Engineers had to adjust the descent profile of Huygens, releasing it earlier than originally planned. This made its batteries last shorter after landing on Titan, but the whole mission was saved and data from the surface was properly relayed to Earth.

1. https://en.m.wikipedia.org/wiki/Huygens_(spacecraft)

2. https://youtube.com/watch?v=3ubr151wrHQ