Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

After a decade, NASA’s big rocket fails its first real test

arstechnica.com

Post reply on HN
Print view View original

71–80 of 153 posts

Re: After a decade, NASA’s big rocket fails its first real test

#71

Post by dotancohen »

sandworm101 wrote: post #33
BurningFrog wrote: post #22

Earlier quoted context omitted.

> That USAF person in the back of the room with their finger on the destruct button. I'm not going up until that job is eliminated!

Then you aren't going up. Rockets are massive tanks of explosives flying through the air at supersonic speeds. That person is there to protect the people on the ground. Launch escape systems will protect the crew, but given the energies involved, the crew is less important than the potential for killing hundreds on the ground, or potentially thousands onboard a cruise ship.

Why don't airliners have AFTS, then?

Re: After a decade, NASA’s big rocket fails its first real test

#72

Post by Animats »

stretchcat wrote: post #62
Animats wrote: post #54

NASA forgot how to make rockets: “Both NASA and contractor officials explained that nearly 50 years have passed since development of the last major space flight program—the Space Shuttle—and the learning curve for new development has been steep as many experienced engineers have retired or moved to other industries.”

It seems like a textbook case of 'use it or lose it.'

The Department of Energy maintains a large nuclear weapons establishment for "stockpile stewardship", so the US doesn't forget how to make nuclear weapons.[1] That already happened once with a material for fusion bombs called FOGBANK. The original manufacturing plant was shut down in 1989. In 2000, there was a demand for more FOGBANK, to refurbish old nuclear weapons. But attempts to make more of it failed. Attempts to make an alternative material failed. It took until 2008 to get production working again.

"The new production scientists noticed that certain problems in production resembled those noted by the original team. These problems were traced to a particular impurity in the final product that was required to meet quality standards. A root cause investigation showed that input materials were subject to cleaning processes that had not existed during the original production run. This cleaning removed a substance that generated the required impurity. With the implicit role of this substance finally understood, the production scientists can control output quality better than during the original run."

That's one way an institution forgets how to make something. There are other ways.

[1] https://en.wikipedia.org/wiki/Stockpile_stewardship

[2] https://en.wikipedia.org/wiki/FOGBANK

Re: After a decade, NASA’s big rocket fails its first real test

#73

Post by tailspin2019 »

mumblemumble wrote: post #57
tailspin2019 wrote: post #53

Earlier quoted context omitted.

Waterfall got us to the moon. Agile will get us to Mars...

Lore suggests that that's a misconception, born of the Agile myth that the only thing that existed before the invention of either XP or Scrum was waterfall. The truth is perhaps more like this: agility got us to the moon, bureaucracy trapped us in low earth orbit, agility seems to be offering our best hope at getting back out of LEO again, including to Mars. For example, see https://www.computerweekly.com/news/224011…

Fair enough.

I was making a commentary more of what I see the difference between NASA vs SpaceX's approach today. I wasn't around for Apollo, and given their crazy compressed timeframe for getting to the moon back then, there must have been some agility there....

I was really agreeing with the previous commenter, which I guess is a statement about today, and not how NASA operated in the 1960s...

> SpaceX learns by doing, not by years of expensive analysis that ends up not preventing failure anyway.

My mistake for invoking the Agile vs Waterfall debate... (and potentially being wrong!)

Re: After a decade, NASA’s big rocket fails its first real test

#74

Post by beowulfey »

beerandt wrote: post #63
ryandrake wrote: post #52

Earlier quoted context omitted.

NASA probably could be as nimble as SpaceX if they didn’t have to rely on Congress, whose goal is not to build a successful rocket, but to spread the money funnel across the States. Imagine if SpaceX’s board suddenly said “the company’s goal is now not to get to Mars, but to employ as many people and contractors as possible in all 50 states!” You’d basically get NASA.

So basically... government could be as efficient as the private sector if it wasn't government?

It’s not the fact that they are government, it’s that the incentive structure is wrong.

Companies are incentivized to make money. In this case, the government is incentivized to provide lucrative contracts to contracted corporations.

In the 60s, when a different incentive existed, the government (i.e. NASA) was incredibly efficient.

Re: After a decade, NASA’s big rocket fails its first real test

#75

Post by GuB-42 »

tailspin2019 wrote: post #37

You have to fail to move forwards. A lack of failure is a sign of a lack of trying anything new. Hasn’t SpaceX proved that? The negative commentary around this seems a little in contrast to how we’d talk about a SpaceX “failure” on HN. I know NASA and SpaceX are very different, but Elon is the king of drifting deadlines too. And SpaceX do fail, but they do a good job of “owning their failures” and they keep moving fo…

SpaceX made failure part of its strategy. Rocket explodes, no big deal, we expected that, the next one is ready.

NASA is not like that, they are more conservative, spend a lot of time and money making sure things work the first time, so when they don't, then it is a problem.

This, BTW, is one thing that killed Armadillo Aerospace.

> Carmack said another mistake the company made was not to go into series production, making several versions of the STIG rockets simultaneously so that the loss of a single vehicle would not be as traumatic. “That was our critical mistake in the last few years, because we should have been able to put more of these together,” he said.

It is not that NASA doesn't have the right to fail, but in their case, it is a major setback, and it is a bit disappointing considering how much SLS went overboard.

Re: After a decade, NASA’s big rocket fails its first real test

#76

Post by mumblemumble »

darepublic wrote: post #66
mumblemumble wrote: post #57

Earlier quoted context omitted.

Lore suggests that that's a misconception, born of the Agile myth that the only thing that existed before the invention of either XP or Scrum was waterfall. The truth is perhaps more like this: agility got us to the moon, bureaucracy trapped us in low earth orbit, agility seems to be offering our best hope at getting back out of LEO again, including to Mars. For example, see https://www.computerweekly.com/news/224011…

Agility, sure, but don't confuse agility with agile

Agile agile. In tech jargon, this particular word is case-sensitive.

Re: After a decade, NASA’s big rocket fails its first real test

#77

Post by JshWright »

tailspin2019 wrote: post #37

You have to fail to move forwards. A lack of failure is a sign of a lack of trying anything new. Hasn’t SpaceX proved that? The negative commentary around this seems a little in contrast to how we’d talk about a SpaceX “failure” on HN. I know NASA and SpaceX are very different, but Elon is the king of drifting deadlines too. And SpaceX do fail, but they do a good job of “owning their failures” and they keep moving fo…

They are very different kinds of failure though... This was not an exploratory "try something new" test. This was an integration test to verify that all the "on paper" work was correct. The whole point of SLS is that its use of technology with extensive "flight heritage" is supposed to avoid having to break new ground.

This is like discovering a major issue during the shakedown cruise of a new ship. It's better to find the issue now, and that's why they test, but the point of this test wasn't to learn anything new, it was supposed to just prove that it works as expected.

SpaceX is generally very up front about tests that they think have a high likelihood of failure, and the tests that do fail tend to be doing genuinely "new things" (like belly-flopping a booster).

I'm not a SpaceX fanboy here (I think Elon in particular is pretty insufferable), but it's not really to fair to compare SpaceX's development testing failures with this integration test failure.

Re: After a decade, NASA’s big rocket fails its first real test

#78

Post by TaylorAlexander »

tailspin2019 wrote: post #37

You have to fail to move forwards. A lack of failure is a sign of a lack of trying anything new. Hasn’t SpaceX proved that? The negative commentary around this seems a little in contrast to how we’d talk about a SpaceX “failure” on HN. I know NASA and SpaceX are very different, but Elon is the king of drifting deadlines too. And SpaceX do fail, but they do a good job of “owning their failures” and they keep moving fo…

So I listened to a podcast about this... but SpaceX and Boeing/NASA are explicitly taking two different approaches to proving their rocket will work. Obviously we know SpaceX is proving their rocket through system level tests. But for SLS the plan has been to use a long accepted NASA approved method of proving their design where individual parts are so thoroughly tested and proven to work that they don’t need many flight tests of the full system. In fact I believe the original plan was one flight test without humans and then have the second flight be a manned mission! So they really aren’t supposed to fail like this.

So at the core are two different approaches to proving their system, where SpaceX totally expects some failures and SLS is supposed to be proven before it even flies. Personally I’m a fan of the SpaceX method but I can understand why the old guard space developers would choose this slower more cautious method.

Re: After a decade, NASA’s big rocket fails its first real test

#79

Post by Rebelgecko »

sandworm101 wrote: post #13
Isinlor wrote: post #11

Earlier quoted context omitted.

Great comparison of SLS vs. Starship: https://www.youtube.com/watch?v=KA69Oh3_obY Key points: https://i2.wp.com/everydayastronaut.com/wp-content/uploads/2... Mass to the Moon: - Saturn 5: 49 metric tonnes - SLS: 43 metric tonnes - Starship: 156 metric tonnes Price per launch: - Saturn 5: 1.2B dollars - SLS: 0.875B dollars - Staraship: 0.1B dollars worse case scenario (Elon aims at 0.002B$) Price per kg: - Satrun 5: 2…

>> Staraship: 0.1B dollars worse case scenario (Elon aims at 0.002B$) Lol. No. That is beyond pipe dream territory. There a innumerable costs associated with spacelaunch that have absolutely nothing to do with reusability. Range closures. Payload integration. Regulatory compliance (ie safety). Telemetry. That USAF person in the back of the room with their finger on the destruct button. That stuff doesn't happen for f…

>That USAF person in the back of the room with their finger on the destruct button.

I think (at least at the Eastern Range), they started leaving this seat empty for Falcon launches in 2017 once the automated flight termination system was qualified. This system also reduces staffing requirements at the radars in Bermuda and all over the place

Re: After a decade, NASA’s big rocket fails its first real test

#80

Post by dmix »

GuB-42 wrote: post #75
tailspin2019 wrote: post #37

You have to fail to move forwards. A lack of failure is a sign of a lack of trying anything new. Hasn’t SpaceX proved that? The negative commentary around this seems a little in contrast to how we’d talk about a SpaceX “failure” on HN. I know NASA and SpaceX are very different, but Elon is the king of drifting deadlines too. And SpaceX do fail, but they do a good job of “owning their failures” and they keep moving fo…

SpaceX made failure part of its strategy. Rocket explodes, no big deal, we expected that, the next one is ready. NASA is not like that, they are more conservative, spend a lot of time and money making sure things work the first time, so when they don't, then it is a problem. This, BTW, is one thing that killed Armadillo Aerospace. > Carmack said another mistake the company made was not to go into series production, m…

Wasn’t this simply a byproduct of Congress trying to control the process? The idea of hard deadlines and low risk.

Apparently they only have until Feb to try to test it again to make sure they can launch this year. That sounds like pure political pressure instead of organizational culture.

Congress tried to cancel this whole test to speed things up so they were apparently fortunate to have done this test at all.

Post reply on HN