Hacker News Reader

Comment by bryanlarsen | Hacker News Reader

Saving NASA $3B by doing so. It would have cost $2B to use the SLS rocket that Congress wanted them to use, and it would have cost another billion to rebuild Clipper to withstand the intense vibrations from SLS's solid rocket boosters.

View on HN

Realistically, what stage of testing/maturity starship need to pass so that SLS is officially cancelled?

It's irrelevant. For Congress, SLS is primarily a jobs program. They want it to continue as long as they can make up any reason to do that, regardless of the practical use.

The jobs program is the practical use - national security requires maintaining an industrial base, or at least that's the theory (don't know how valid it is).

If only there was such a project for advanced fabrication for chips maybe we wouldn’t find ourselves in the current situation.

We do, EUV LLC. was subsidized by the US govt (via DARPA) and a US industry consortium in the 90s. Lots of the tech went to European allies in I think a sale/collaboration at some point but we maintained power due to that initial seed and have used it for things like trade restrictions on the technology.

(1998) https://www.semiconductoronline.com/doc/euv-llc-enters-devel...

I will never understand why jobs programs can't also be useful. Didn't we build a lot of infrastructure that way in the 30s?

Senator Richard Shelby's retirement, for one.

If Starship works, it's going to be almost a state-change moment in the space industry. They aren't just building a single test vehicle for a demo mission, they're building factories churning out engines, rockets, GSE, building launch pads, etc. They might take a few tries to nail a first orbital launch, once that happens (and assuming no undue regulatory hurdles) their launch cadence may start picking up dramatically within 3-6 months of a success.

If eighteen months from now Starship has launched a ~dozen or more times and SLS once, I'd imagine they might wrap up SLS with Artemis 2, which should be mostly complete by then. At a total program cost of ~$24 billion for 2 flights it would be quite ... something, but it's hard to see them continuing much beyond that. They'd just figure out a way to use Crew Dragon/Starliner to get to Starship for lunar missions (assuming they're not going to want to launch/land humans on Earth on Starship for a while).

When. When Starship works

I share your optimism but I think it it is unrealistic to unquestionably assume that SpaceX will continue to succeed. The very nature of the problems that they are trying to solve is such that failure is always a possibility.

I don’t say this as a slight to SpaceX… on the contrary the risk is what makes them so exciting!

I mean, the way they succeed with Starship is failing over and over again until they don't anymore. In that sense it is (now, not before SpaceX was scaled up) solely an issue of funding: will it work before the company runs out of cash?

At this point, I imagine most every investor in Earth would be tripping over their own legs to buy SpaceX stock if they could. If they ever run low, their ability to raise more money is as sure a thing as anything in this world these days. Additionally, Starlink is going to be an absolute revenue monster very soon, even if Starship for some reason ends up taking another decade to get working.

The only way Starship fails is if SpaceX goes broke. SpaceX is not going to go broke. Therefore, Starship will eventually work, after some variable and as yet indeterminate number of prototype failures.

> If eighteen months from now Starship has launched a ~dozen or more times and SLS once, I'd imagine they might wrap up SLS with Artemis 2, which should be mostly complete by then.

Starship needs to be crew-rated by NASA before it can replace SLS+Orion. I'm sure that will happen at some point, but that point may well be after Artemis 3 has happened.

Some people propose hybrid architectures – putting Orion+ESM inside Starship, and using F9+Dragon to ferry the crew to LEO. However, I think NASA will view those hybrid architectures as even more complex, and hence riskier, than the current SLS+Orion(+Gateway)+StarshipHLS architecture. Plus, pro-SLS forces in Congress will oppose NASA doing that. Bridenstine proposed launching Orion on Falcon Heavy, and Senator Shelby's response was to ask Bridenstine for his resignation. Shelby may be gone from Congress but there are other members of Congress with the same attitude. It is harder for Congress to oppose a NASA project to crew-rate Starship because it would be less obviously directly competitive with SLS+Orion, even if the death of SLS+Orion would be its ultimate consequence.

Out of curiosity, do we know how much SpaceX has spent in total on the Starship program?

dumb question - why are we using Saturn V derived rockets? afaik none of these new rockets can match its max payload

I assume you typoed and meant to ask “why aren’t” instead.

We aren’t because the Saturn V moon rocket (and related projects) were consuming over 4% of the US government’s budget, way more than even the SLS and definitely more than the Falcon Heavy.

Starship exceeds Saturn V’s lift capacity.

Not true.

Saturn V's lift to LEO was 120-140 tonnes (https://en.wikipedia.org/wiki/Saturn_V) while SpaceX claims Starship's is "over" 100 tonnes (https://www.spacex.com/media/starship_users_guide_v1.pdf).

It exceeds Saturn V on the second launch that day. ;-)

Well, if you have a 140t payload it doesn't matter how many 100t payloads you can lift per day, it just isn't going to lift it.

On the other hand, obviously, payloads are planned for the rocket they are going to be lifted by, so most likely somebody will set constraints on their project to either have smaller payload or split it into parts.

Right; I think Starship, if it pans out, is going to entirely change how we plan, build, and carry out space missions, both manned and unmanned.

It's also possible whoever has such a large payload might want their own second stage & a fairing, instead of Starship. A single-use stage (no reserve fuel) with no re-entry needs could be a lot lighter than Starship, though vastly more expensive. That might be worth it for some payloads. Super Heavy is planned to be a ridiculously powerful booster.

I doubt that will happen though. It's far more likely in-orbit assembly and orbital refueling will progress dramatically as launch costs drop, and anyone with such a payload would find a way to design it for such assembly.

Don’t final inspections with payload attached by itself take more than a day? Or is that going to be a relic of the past?

I doubt anyone knows for sure at this point, but I'd imagine the goal is something more akin to loading cargo into the belly of an airliner; consistently sized modules with standardized attachments for power etc.

Even given a few weeks, the relaunch of an Saturn V will be a bit rough.

Starship brings any 100 tonnes to orbit, while of Saturn V's 120+ tons a half constitutes the 3rd stage with unspent fuel. So it's less of payload per ce for Saturn V compared to Starship.

On the other hand, Starship didn't fly to orbit yet. We'll see better how it goes when it does. And we can't fly Saturn V now anyway. And Starship - even if supposedly with lower payload - promises to be much cheaper. And...

In Skylab mode it was pure payload.

Which weighted less than 80 tons - much less than what other superheavy rockets are bringing to LEO - Energiya, SLS, BFR...

That’s looking at a single launch and neglecting the refuel ability.

And will be able to refuel in orbit

We don't have the infrastructure to build them anymore, and a lot of the design wasn't captured in the blueprints because things were tweaked by hand to make them work. We could build something with Saturn V-like performance, but it would be a new rocket, not just resuming production of the older line.

That was seriously considered. It was rejected because schematics consist only a small portion of the knowledge necessary to build a rocket, most of the rest had been lost to time.

A massive amount of redesign was required to build the shuttle derived SLS, deriving from Saturn would have been worse.

But the real reason was that deriving from Shuttle kept those employees employed. SLS has never been a rocket program, it's always been a pork program.

SpaceX's Starship has twice the thrust of Saturn, so could be tweaked to send more payload by sacrificing reusability. But they have an even better plan -- refueling allows massive payloads.

In particular, all Saturn V engines were completely hand-built and the process was not documented. A lot of very fine craftsmanship went into that that we can't replicate today.

Seems like that situation where the first engineer builds an inscrutably clever codebase. Lots of fine craftsmanship indeed.

Nitpick: we can replicate it, it's not like we've technologically regressed. We're just not willing to spend the money - at it's peak NASA's budget was ~4.5% of federal spending when the Saturn V was designed (it's now only 0.5%).

> Nitpick: we can replicate it, it's not like we've technologically regressed.

Saturn's materials and computer technology are more than 50 years old. Not only would we have to want to recreate this historic technology, we would have to recreate the knowledge, skills and competencies of the experienced craftsmen that used it, across the massive supply chain, as well as the tooling required.

Actually, you’d be surprised how many techniques have been lost. They’re just not relevant, for example the actual practice of riveting battleship hulls is gone, but that changes suddenly when you want to replicate the past.

That’s because federal spending has increased across the board. NASA’s funding isn’t that much different now than then.

NASA's spending was near $5 billion in the mid 1960s. [1]

That's equal to $42 billion today per the BLS. NASA's fiscal 2021 budget is $23.2 billion. So while you're correct that Federal spending has increased as a share of the economy (declining NASA's share of spending even faster), NASA's funding is realistically down by at least half now vs then when you adjust for inflation.

That BLS figure is also the questionable official conservative take. Gold was $35 / ounce back then, it's around $1,800 now (51 fold increase). The average new car was near $3,000 in the mid 1960s, it's now $41,000. Oil was $3 / barrel in the mid 1960s, it's in the $70s now.

[1] https://history.nasa.gov/SP-4012/vol4/ch1.htm

- NASA's spending fell to $3 billion by 1974. Inflation adjusted that would be $17.5 billion today, so at least they're ahead of that.

So adjusted for inflation, NASA's budget is merely half what it was in the peak of the 1960's, whereas the statistic quoted makes it sound like it is 1/10th.

Cherry-picking a peak date is a bit dubious though. If you average over the duration of the Apollo program, you get a number much closer to today.

So most, maybe even all of the difference in budgetary size is due to increase in size of the rest of the federal budget.

To be fair, the GP comment said “at its peak”. They weren’t cherry picking, just following the initial thread. But your point about relative % of GDP may not be the best measure

No, we can't replicate it because there is simply not enough description. For example, there aren't even specs for parts.

The entire thing would have to be re-designed from scratch with modern manufacturing process in mind.

And for things like F1 engine the available manufacturing methods have a lot of influence on the design, so the result will be basically a different engine than F1.

My dad was in the airline industry. Engineers kept lots of secret techniques to prevent them from being replaced.

One had to due with adding strings to some foam mixture to get it to set correctly. Not documented anywhere. When cheap labor was brought in they couldn’t reproduce the work. So old people had to be hired back at a lot more

Rocketdyne proposed an updated F-1, the F-1B, as part of the SLS Advanced Booster Competition:

https://en.wikipedia.org/wiki/Rocketdyne_F-1#F-1B_booster

https://forum.nasaspaceflight.com/index.php?topic=50339.0

Obviously it is not exactly the same as the F-1, it is rebuilt using modern parts and manufacturing techniques. But the F-1B aimed to have at least as good performance as the F-1. It appears the proposal was not successful in convincing NASA to adopt it for an SLS liquid rocket booster, and NASA is currently planning on sticking with solids even into SLS Block 2. (And Block 2 SLS is probably never going to fly.) But Rocketdyne owns the IP to F1B and could resurrect it again in the future if there was a demand for such an engine. They were working on it within the last 10 years so unlikely that much knowledge has been lost.

This serves as a counterexample to the often repeated claim that we can't rebuild the F-1 because we've forgotten how. Rocketdyne and NASA had enough institutional knowledge to design a modern successor.

Do you have a source on that? There was an industry I was adjacent to, losing all their skilled metal workers to retirement, and there was no next generation who were skilled enough to do the machining. They were literally filming them working in the hope they could at least train someone down the road.

We can’t make them the way they used to make them, but likely could do at least as good fairly easily _if_ sufficient funds were made available.

I’m basing that on https://arstechnica.com/science/2013/04/how-nasa-brought-the..., which tells me we now know quite well how these engines (or rather: the particular one they looked at) were constructed. I also get the impression we could create those parts using other means such as 3D printing.

Nobody says we can't build an engine like F1. We just can't build F1, because we don't know how the original ones were built.

Machines like that are designed with the manufacturing process in mind.

An engine designed to be manufactured with today's manufacturing processes will look different than an engine designed to be manufactured manually by ultra-skilled metalworkers and welders.

I would like to read more about this. Is there a source or anything?

Sure. Not a source but still interesting. https://www.youtube.com/watch?v=ovD0aLdRUs0 (Why Can't we Remake the Rocketdyne F1 Engine?)

TLDR: everything was handcrafted, notes were not made or lost, people have retired or passed away and today's engineers don't know the tricks that the original engineers used. Effectively, existing blueprints are not enough to rebuild the engine.

It's sort of like how a new cook can follow a recipe from an expert chef but the final result won't taste as good. There's a lot of tacit knowledge assumed which can't necessarily be communicated in written form.

Looking at the "LEO Payload" column here:

https://en.wikipedia.org/wiki/Super_heavy-lift_launch_vehicl...

  Saturn V: 310,000 lb
  SLS: 209,000 lb planned
  Falcon Heavy: 126,000-141,000 lb 
  Starship: 220,000–330,000 lb planned

Of course, the benefit of Starship is that it can (in theory) do that several times a day.

wasn't that the promise with the space shuttle, too?

Maybe at one point, but the final design did not allow rapid turnaround. Starship could actually do it the way is being designed. They have a good shot at pulling it off.

Reply on news.ycombinator.com