Blue Origin's rocket reuse achievement marred by upper stage failure | Hacker News Reader

Blue Origin's rocket reuse achievement marred by upper stage failure

arstechnica.com

78 pointsrbanffy45 comments

Once Elon showed how to do it, and how cost-efficient it was, a rocket company that doesn't do it is not viable.

Spacex first landed an orbital booster just over 10 years ago and have now landed 600 times.

The entire rest of the world combined has done it twice.

For a long time people would scoff when it was said they had a 10 year lead, and that others would catch up quickly. Proof meets pudding.

FTA: "SpaceX suffered upper stage failures on three test flights of the massive Starship rocket last year. "

SpaceX has also had numerous failures with the larger generation of second stages and currently doesn't have a lead there. Nobody does.

It's a hard problem, and both SpaceX and Blue Origin will probably have failures in the future too, I am encouraged that they both see failure as a way to do better and looking forward to both of them eventually succeeding. It's a good time to be a space nerd.

There's a saying in the racing business. If you're not walking back to the pit now and then carrying the steering wheel, you're not trying hard enough. If you're walking back to the pit too often, you're incompetent.

If you always fail, you aren’t trying.

If you never fail, you aren’t trying.

If you always fail, you aren’t learning

Isn't that better?

True, but then you have to differentiate trying and failing vs not doing anything and failing by default.

There's another aspect. If you're launching men in rockets, you cannot tolerate failures, so the development cost is way, way higher. The cost effective method is to launch unmanned ones, tolerating a lot of failures, and when the bugs are worked out then launch men.

Nobody else has anything remotely like Starship. If they pull it off, and it's looking like they will, they will extend their dominance for another decade if not more.

Yes, Starship development has been slow and occasionally explodey, but they've successfully demonstrated all the fundamentals and it's "just" iteration from here. (They haven't gone into full orbit, but that's by choice, not lack of capability.)

> If they pull it off, and it's looking like they will,

I really wonder about this psychological effect where non technical people champion people like Musk so hard without any basis for doing so. Is is some sort of wanting to belong to some ideology that makes you just make shit up in your head about how Starship is a success, despite many indicators of it clearly being a stupid idea born from Musks ketomine episodes?

For the record, Starships engines are the equivalent of taking a Toyota Corolla and making it run on nitrous continously on the verge of self destructing. You may be able to do technology demonstrations here and there, but making it work reliably for actual missions is much much harder.

Bringing together the money and people to make this stuff happen is the basis. That’s the most impressive part. Debatably the only truly impressive part.

There’s no ideology. You can watch a really big rocket take off every month or two and watch a smaller rocket take off every couple days. I’m sure there are better designs out there… on drawing boards.

Its not a video game where you put enough resources into "science" and stuff just works.

There are fundamentals at play that Musk certainly doesn't understand, and its ridiculous to think that he would be smart enough to account for them.

> There are fundamentals at play that Musk certainly doesn't understand

Examples?

Musk thinks that if you crash things enough until it works first time, the problem is solved. Which is fine for something like Falcon. As loads get bigger and heavier, and you start running into margins of performance (for example, raptor engines are required to generate the thrust to lift it). And then, just cause it works the first time, that doesn't mean there is enough margin on the system to not fail due to an external unforseen event for which the narrow margins can't account for.

Musk's technology ventures have been incredibly successful. I wouldn't bet against him. In fact, I've bet on him.

Pre 2018 Musk before he destroyed his mind with ketamine, sure. Apart from a few very autistic events, I would have also bet on him.

But now, you wanna tell me the person that thought the Cybertruck was a good product is going to solve a very complex problem of reusable heavy launch vehicle? Much less go to Mars? Ok.

"Starships engines", also known as Raptors, are among the most extensively tested rocket engines in the world.

More than 600 of them (across all variants) have been produced and tested on McGregor test stands and in flight, with relatively few explosions, and for some of that test stand explosions, we do not know if those were deliberate overload tests.

Raptor is about the least problematic of all the crucial Starship components. They had a lot more problems with ullage and the gorilla in the room is the heat shield. It must be very reliable and at the same time quick to check and fix. The time to fix the heat shield will be the critical component of the total turnaround time.

Personally, I don't believe in 1 hour turnaround. 1-2 days just might be plausible.

Is that why you see several engines not functioning during flight tests?

The issue is that the controllers have to maintain a crazy unstable balance of the mixture to keep these things running due to the dual preburn cycle and cryogenic storage, that any unforeseen circumstance can lead to failure.

In the days of Falcon, where Space X was attracting people willing to disrupt the industry, I would have been in the "its possible" camp. Nowdays, with everything going on, I would place a large chunk of money on the fact that they will never get it to be reliable enough. They haven't even gotten it to orbit yet, despite the massive experience they have.

The rockets were designed to be tolerant of a couple of the engines not functioning.

This is actually good engineering, as perfect reliability is very expensive.

>This is actually good engineering,

Lol.

First its "engines are reliable", then its "well actually they are not, but the starship can function with a few out"

The Elon simps never cease to amaze me.

An engine should not go out by itself. Some catastrophic event can take out an engine, and then if you design your rocket to fly with a few out, thats fine. But if the engines just are so unstable that they fizzle out, thats a huge risk because that means any non planned event can cause more engines to fizzle out, leading to loss of vehicle.

By far the most common reason for engines to go out is when they don't have enough fuel, and that is mostly caused by faults in ullage, which is what I mentioned in my GP comment to be a significant problem. Once you spent most of your fuel, relighting the engines is not easy.

Ullage is a plumbing matter, though. External to the engine itself and its intrinsic (un)reliability.

On the ascent burn, where fuel is plentiful, Raptor flameouts have become way less frequent over time. The # of engines failing on the Super Heavy during the ascent burn is 0/33, 1/33 and 0/33 during the last three IFTs. Not bad for a test vehicle.

The Saturn V's had a big problem with "pogoing" which they never solved. The Soviet moon rocket apparently was abandoned because of unstable coupling among the grid of engines that powered it.

SpaceX must have a handle on these problems, I haven't heard anything about it.

The full flow cycle of the raptors means they do 2x preburn for oxidizer and fuel to spin the turbines to generate the pressure required for the pressure ratio. If a small part of this fails for whatever reason, or is out of sync, either engine doesn't start or goes kaboom.

Ascent isn't the problem its the relighting of them on the way down.

I agree, relighting is tricky, not just because of the limitations that you describe, but because there is a lot less cumulative experience with relighting of engines than simply lighting them at the launch. We (as the entire humanity) have seen tens of thousands of rocket motors lit on ascent, orders of magnitude than those relit.

Raptors don't seem to be prone to kabooms during flight, probably because the command unit switches them off in case of any suspicion.

> An engine should not go out by itself.

Airliners have twin engines, and are designed to be able to fly one one. Why? Because engines fail now and then. Every flight critical part on an airliner has a redundant workaround, sometimes two.

Redundancy is a cost-effective way to successfully deal with unreliable parts.

When I go dirt bike riding in the wilderness, I only ride with a buddy. I wear protective gear. I carry a phone, water, and a minimal survival kit. I also text a photo of the trail map to a friend before hitting the trail.

Small things, but potential life savers.

Yeah and if you look at all the jet engines produced, the amount of failures are extremely tiny compared to total flight hours.

Not the same for raptor engines.

The ratio was different in the late 50s when jet engines were new. Experience counts.

It's the 4-minute mile except it's taking everyone else too long to copy it. Really shows how far ahead Musk is.

I know insurance for a launch is typical, but seems really tough to do that for this still “rather experimental” launch. I got to imagine it has costs something like 50% on a project like this.

The failure of the upper stage is a bummer. If it triggers a months-long review, that will almost certainly bump back the schedule for the prototype Blue Moon lander launch.

Video of the booster landing: https://xcancel.com/JeffBezos/status/2045874068763632017

Stupid question I know, but are there people on that boat?

It's a drone boat, so no.

IIRC people standby on a boat at a safe distance, then come onboard to secure the booster when it's safe enough

What I was not aware of is how many satellites Amazon already has in LEO for it's own Internet service.

They've been flying under the radar there it would seen.

I will be good to have competition for space Internet. It’s unclear though if the market will really support two players. Satellite radio and data quickly ended up consolidating down to one.

Amazon is trying to become more vertically integrated but they seem at a structural disadvantage here competing against SpaceX.

You might be counting out the value of government and military contracts that might not want to do business with a wild card.

SpaceX is killing it because the US government gives them a bunch of contracts, but if stability is slightly more important than cost or speed, amazon has a contender.

They "only" have about 250 but they're authorized for 3000. They just bought a satellite company this week though that might boost the numbers a bit.

As late as 2010 there were "only" around 1000 satellites in orbit.

Apparently they lose authorization if they don't get to 1500 hence the ""

It came to my attention recently how many TOTAL objects currently exist in LEO. And that a study said that due to light deflection of these objects, that the earth’s night sky is an average of 10% brighter than it was in 1980s… although I generally am excited by technological advancement, that fact (if true) made me feel somewhat melancholy.

I wonder how a company would be able to catch up with SpaceX, and make this no longer a monopoly.

I think we will see many soon. India and China alone have something like ten promising space launch startups.

Space is hard.

Losing payloads hurts though, especially for a new platform.

Reply on news.ycombinator.com