Houston: we have a door latch problem

November 20, 2019 6:51 PM

Early in the shuttle days there was a huge concern about what to do if the doors did not close and latch properly at the end of the mission. The doors had to be open during most of the orbit stay time for cooling and to allow satellite deployments or other objectives. But during re-entry and the atmospheric part of the flight, the doors had to be firmly shut and latched down. If they were to spring open and rip off, the vehicle would become uncontrollable and catastrophe would ensue.
posted by sammyo (24 comments total)

Easy: an inanimate carbon rod.

https://www.youtube.com/watch?v=kC-AXV54nMU
posted by jonathanhughes at 7:03 PM on November 20, 2019 [11 favorites]

The STS-3 test gives me the willies just thinking about it. Riding in the back would just be kinda great, though.
posted by DangerIsMyMiddleName at 7:50 PM on November 20, 2019 [2 favorites]

Wouldn't the bay doors just transition to pauldrons in gerwalk mode?
posted by GCU Sweet and Full of Grace at 8:21 PM on November 20, 2019 [10 favorites]

Cool, I had not heard that but it makes perfect sense.

For a few years in the '00s the rod cutters were part of the tools I prepped for every flight.
posted by BeeDo at 10:00 PM on November 20, 2019 [6 favorites]

Yet another case of NASA normalizing deviance. The payload bay doors have never jammed, so they will never jam, so we don't need to be able to un-jam them.

The idea of having some poor astronaut ride in the back of the payload bay is even more scary when incidents like this happen. Luca Parmitano almost drowned in his suit during an EVA. If he had been on his own in the back of the payload bay he would have died.
posted by monotreme at 10:12 PM on November 20, 2019 [6 favorites]

Sammyo, thanks for introducing me to Wayne Hale's blog. I've read several entries now, and it's great stuff.
posted by bryon at 10:53 PM on November 20, 2019 [1 favorite]

Wayne Hale is an Internet national treasure, and that entire site is worth reading and thinking carefully about if you’re at all interested in safety and complex systems.
posted by mhoye at 11:36 PM on November 20, 2019 [3 favorites]

So that is the story. Accept the risk because we think it is low; have a screwy contingency procedure ready if we’re wrong.

But that is not the way you really want to fly in space.

The story of the entire shuttle program.
posted by rockindata at 3:54 AM on November 21, 2019 [10 favorites]

Fascinating post and fascinating blog, thank you.
posted by hoyland at 4:11 AM on November 21, 2019

Every time I read an in-depth story about some aspect of the Shuttle mission profile, I go in hoping that it won't be another shit-show of normalization of deviance and lethal ineptitude, only to be disappointed.

Then I go and re-watch the SpaceX blooper reel (set to the tune of Yackety-sax) and realize that, yes, reusable rocketry is hard.
posted by cstross at 4:43 AM on November 21, 2019 [1 favorite]

Sometimes I wish they had kept the Shuttles flying, because MURICA.

This is not one those times.
posted by Brandon Blatcher at 5:31 AM on November 21, 2019 [1 favorite]

Sometimes I wish they'd kept the Saturn 5 flying, because earthicans.
posted by justsomebodythatyouusedtoknow at 7:27 AM on November 21, 2019 [4 favorites]

The Saturn V was pretty damn cool, yeah. In hindsight, we should have kept them and never built the Shuttle we built.
posted by Brandon Blatcher at 8:52 AM on November 21, 2019 [4 favorites]

Personally, I find it refreshing when organizations and the people doing the work are willing to accept what they consider a reasonable risk, at least when the risky thing isn't intended for public use. See, there are many risks that are completely unacceptable in that context that aren't actually a huge issue when the people subject to said risk are instead subject matter experts who understand the risk they are accepting and volunteered for their position.

Astronauts have yet to be anything but test pilots. Happily, it seems that will change relatively soon. In the meantime, significant risk is inherent in the endeavor because it simply hasn't been possible to build a spacecraft that has the redundancy we would expect in something like a commercial airliner without a mass penalty that renders the spacecraft useless.
posted by wierdo at 10:58 AM on November 21, 2019

Astronauts haven’t been required to be test pilots since the 70s
posted by Brandon Blatcher at 11:40 AM on November 21, 2019

Yes, they do indeed launch people whose employment history does not include "test pilot," but the level of risk and indeed the amount of understanding of the craft's systems required has not actually gone down.
posted by wierdo at 12:16 PM on November 21, 2019 [5 favorites]

So, I have an anecdotal experience involving the cargo doors of the orbiter that transferred all the way to building a plastic model.

I was given a fairly large scale model kit of the Shuttle/orbiter for Christmas once, and at that point I was pretty good at building pretty complicated model kits, including import stuff that had moving parts or were build it yourself versions of transforming toy mechas like Robotech, etc.

As fans of the space program know when the orbiter is in the rebuild and recovery bay and phase horizontally, the doors are so fragile that they have to mount a heavy duty support frame and structure to open and close them in full earth gravity.

So I'm building up this plastic model of the orbiter, and to be honest, the whole thing was a huge pain in the butt to assemble cleanly because of how unconventionally shaped the orbiter is in many places, so it had a lot of curved seam and glue joints in ways that were difficult to align.

And then I get to the cargo bay doors and there's a full on warning in the build instructions that they're very fragile because they basically have to be long, thin foils of polystyrene plastic with tiny little plastic hinge points to look right.

I'm trying to recall if the build instructions had a moveable option or not or you had to choose between gluing them closed or open, and I'm thinking it might have been the latter, that you had to glue them in place. And I may have tried to rig them so I could open or close them.

Regardless, that inherent engineering fragility was well represented and replicated in that scale model. Even before trying to apply any glue or modifications, just getting the doors to line up and fit for a fit test while closed was really fiddly. I never did get the cargo doors right either fixed so I could move them or glued shut or whatever it was I was trying to do, and I abandoned the model completely and never really finished painting it because of it.

In hindsight and knowing much more about the shuttle and aerospace in general, those doors are basically the same thing as a rocket cargo faring - just thin aero shells really that need to be fixed in place to resist nominal aerodynamic loads and stresses. If the orbiter, say, re-entered upside down in a backflop those doors would probably blossom and blow open and be the first thing to get shredded into confetti.

And in the orbiter's case they're also heat radiators and essential to on orbit operations to remain open for the duration of the mission.

It's astounding that the shuttle system flew at all, as much as it did, with so few accidents. And yet we probably couldn't have built out the ISS so fast without it and space SUV work platform overkill.
posted by loquacious at 12:19 PM on November 21, 2019 [4 favorites]

> normalizing deviance. The payload bay doors have never jammed, so they will never jam

Here is the basic issue.

Let's say you have 10 mission-critical systems, which if they fail will lead to mission failure.

And let's say they are pretty darn reliable--say 999 out of 1000 reliable.

That's pretty reliable. The shuttle bay doors were probably not that reliable.

So 10 systems at 999/1000 reliability, if you do that math, gives 99% reliability for the entire system.

Not bad, huh?

It's going to catastrophically fail just 1 time out of 100.

Of course, there were 135 space shuttle flights.

So even with that level of reliability, you are basically guaranteed to have a catastrophic failure or two over the life of the program.

100% reliability of complex systems with catastrophic failure modes operating in unpredictable environments is really, really hard.
posted by flug at 4:39 PM on November 21, 2019

Yeah, that stats like that always make me curious how many, if any, failures we would have seen of the Saturn V, if it had flown 100 times, instead of 12? 13?
posted by Brandon Blatcher at 4:47 PM on November 21, 2019 [2 favorites]

Given that I can think of three separate Apollo missions that had something off-nominal happen with the Saturn stack alone, I dare say its record might not have been so great if we had kept launching it without working the issues. (They were worked at the time, which probably explains why they never had multiple failures add up into a loss of vehicle incident)

The kind of management that led to Challenger was plainly unacceptable, but good management wouldn't have eliminated all risks, only managed them as much as possible. In that case by not launching in conditions outside the rated specifications of the system's components.
posted by wierdo at 6:17 PM on November 21, 2019

Wierdo, which three are you thinking of? The only two I can recall off-hand are Apollo 4 and 6, but those were the first two flights, not sure if they should count?
posted by Brandon Blatcher at 6:26 AM on November 22, 2019

13 lost an F1.
posted by GCU Sweet and Full of Grace at 7:02 AM on November 22, 2019

Apollo 12 was struck by lightning twice in its first minute of liftoff, though the greatest effects seemed to be on the CSM/LM.
posted by ZeusHumms at 8:14 AM on November 22, 2019

This MIT opencourseware on the systems engineering of the shuttle is indispensable for those curious about the subject.
posted by OldReliable at 10:21 AM on November 24, 2019 [1 favorite]