that I want to highlight Background: NASA considers configuration files to be flight software. Let me try to put this in layman terms. A ![]
When people think of flight software they often think of thousands of lines of complicated code. That code does exist, but lots of other stuff is handled by “configuration files.” Config files contain parameters & settings that you may want to change without touching “core” code.
Software Config files updates are relatively simple and don’t require a room full of coders banging out code for days straight on just coffee and vending machine snacks.
Starliner and Dragon were meant to be crewed vehicles. In crewed vehicles, NASA wants the astronauts to be in the chain of “graceful degradation” of the vehicles such that manual piloting is an option after some faults, and astronauts can help “save” themselves.
During OFT-2, Starliner had a config file(s) that allowed the ground to takeover after certain departure/return fault responses without a crew present. This current mission likely had config files updated with crew in the chain of fault responses.
I would argue that it is not a robust design to need to manually update/patch the config file to toggle between the two configurations. Seems like an inelegant design, but I understand how/why they got there. I feel like there might be somewhat of a NASA “requirements hole” here.
Just elevating all this background because extreme speculation is not helpful to anyone. NASA wanting to test the config file updates and have their operators do some SIMs with the modified fault response chain makes sense.
Probably good lessons learned here for NASA to apply more broadly. I personally would feel better knowing that crewed vehicles could quickly depart uncrewed, or with incapacitated/hurt crew.
That mostly explains why one mission could undock autonomously and the other can’t, even though they’re ostensibly the same codebase (and as someone in the software industry, this explanation does make sense–we use config files for just this kind of thing).
I get why they’d want a human-in-the-loop configuration. However, I still don’t understand why it would exclude the fully autonomous case. Suppose the astronauts were inside but totally incapacitated. Wouldn’t you still want ground to be able to take over?
Heck, in this case put the astronaut(s) in full suits, have them perform an undock & maneuver within reach of the arm, do an EVA to be “rescued” by the arm back into ISS them destructively deorbit the POS capsule.
If they REALLY need a freed docking port this’ll git 'er done. Yee haw!!!
I have some sympathy for shitty processes. It creeps up on you unless you fight it constantly.
However, at some point it should have been asked whether they could undock without an astronaut onboard, and done what it takes to fix that constraint. That they didn’t is pretty inexcusable.
If it’s “build to spec”, which I suspect it is, then NASA dropped that ball not Boeing. But I’m perfectly happy to be proven wrong on this supposition.
A certain someone has an “algorithm”, the first rule of which is: “Question every requirement”.
Cost-plus bidders would never do that, because every bad requirement is an opportunity to make more money: first, implementing the dumb thing; and again to eventually fix the dumb thing. Or better yet, doing some other dumb thing which itself must eventually be fixed.
But for fixed-price bids, it makes a lot of sense to question the requirements, because if you’re going to outbid the others, you need to eliminate as much internal waste as possible. You no longer get paid more to fix the dumb thing.
So while NASA may well be the originator of the bad requirement in this case, Boeing is on the hook for the bad outcome. It’s on them for not pushing back (if that’s indeed what happened).
The Starliner capsule “Calypso” has now been in space 64 days, its mission extended indefinitely while Boeing and NASA conduct testing in an attempt to find the root cause for why multiple of the spacecraft’s thrusters failed during docking.
SpaceX had to cancel Starlink 8-3 today due to bad weather at the drone ship site. So tomorrow we have two launches just over an hour apart. One from Canaveral Space Force Station and one from Kennedy. Damn. That would be cool to see.
That’ll be impressive if they pull it off. I wonder if we’ll ever see near-simultaneous launches (assuming they’re not headed to the same orbit). Planes do it often enough.
After their failure a couple weeks back, SpaceX launched three rockets in <28 hours across all three launch sites. Only a single non-SpaceX rocket had launched while they were figuring out the root cause.
Funny you mention that. There’s a Vandenberg launch scheduled for Sunday/Monday. Not Starlink, Artic Broadband for Space Norway (Northrop Grumman-built satellites - which has got to sting a little that they are still having to hitch rides on SpaceX launch vehicles since NG is part of ULA).
In other news, looks like Rocket Lab also has an Electron launch planned in the next couple of days. Their cadence really seems to be picking up. It’s their 2nd launch in August and 10th of the year (which will be a record for them), They have 13 other orbital launches planned for 2024. They are currently solidly in 3rd place behind SpaceX (natch) and the Long March family of rockets. It’s light-lift (150kg payload to orbit), but it’s not nothing.
Rocket Lab will be in good shape for Neutron if they can keep improving their cadence. A lot of stuff is no different between a small vs. large rocket. Any practice they can get with keeping launch services slim and efficient will pay off greatly when they start launching something bigger.
I love Rocket Lab as the little engine that could. With that said, it’s not just cadence, they have waaaay too many 2nd stage failures and need to seriously address that. They also seem to be struggling with reusability (although at least they are trying - looking at YOU United Launch Alliance).
Again, rapid, reliable, reusable. If you aren’t at least trying for all 3, get out of the game and make room for people who are. When ‘government is wasteful - why are we paying for this’ people talk about spaceflight, it’s a direct reference to the bloated, unsustainable, ridiculous practices of ULA and Boeing (I was actually wrong above, ULA is Boeing and Lockheed Martin, Northrup Grumman is not part of it - mea culpa).
Beating a dead horse, but while we need competition, we need that competition to be something other than a Congressional jobs program pouring money into dying companies who can’t figure out that these aren’t the Apollo days anymore.
I used to think the Electron’s electric pumps were a brilliant way to shortcut the development of a full-fledged turbopump engine (whether a gas generator or otherwise). I still do, to some extent–but two out of three of Electron’s operational failures (I’m not counting the first flight) were directly related to the electric pumps. And the second one was a very unusual failure where an almost-but-not-quite-total vacuum is actually fairly conductive.
Point being that they might have better luck with a more traditional engine design (as they’re doing with Archimedes). The electric pumps were clever and simple in a way, but had an entirely new set of unknown risks. Unknown unknowns instead of known unknowns, as it were.
Really only experience that reveals all the unlikely and counter-intuitive ways that things can go wrong is the only way to eventually work out what’s reliable.
]
Yee haw!!!