SpaceX launches Starship v3 rocket | Hacker News Reader

SpaceX launches Starship v3 rocket

space.com

441 pointsbusymom0315 comments

https://www.nbcnews.com/now/video/spacex-successfully-launch... [video]

Summary from my watch:

- Launch roughly on time, after a scrub yesterday. (Sounds like the scrub was due to ground equipment, most notably the water system.)

- Initial ascent was good, but then one engine on the booster went out.

- Relight of the booster's engines after stage separation for the boost back burn failed. Engines did light again for a landing burn, but seems to have hit the water harder than expected and was very off target.

- Starship lost one engine shortly after stage sep. Turned into an unintentional test of engine out capability. It made it to space.

- Some weird motion and lots of off-gassing after engine cut-off, with uncertainty about if it actually got a good orbital(ish) insertion. Seems to have been benign, with the motion being a weird slow flip to the orientation for payload deployment.

- Test deployment of dummy payloads was successful, including a couple with cameras to look back at Starship.

- An in space engine relight test was skipped, presumably due to the issues during launch.

- Re-entry to over the Indian Ocean seemed to go really well. Nothing obviously burning or falling off. The amazing views of the plasma during re-entry, something never seen live before starship, are now routine.

- Starship did a maneuver to simulate how they'll have to go out over the gulf and back to the landing site.

- Nailed the target, evidenced by views from drones and buoys. Soft landing before falling over and giving us a big (expected) boom.

As far as overall progress from previous test flights goes, they're at least treading water while making many large changes. I think they were hoping to try for a tower catch and actually going orbital for next flight, but I highly doubt that now. The boostback burn failing was the largest failure, with the engine failure on Starship being a close second. Good performance despite engine out seems to be an unintentional success.

Nice work by SpaceX engineering.

Good summary. The booster appeared to hit the water at 1400 km/h (a bit under 900 mph) so not really survivable :-). Engine out on ship seems to left them with just enough fuel to land but not enough to do the hover thing (simulates being caught by chopsticks). They notched it down to two engines (vs planned 3) on the landing it seems?

Basically if they can figure out the engine issues, it looks like they should be able to do a full end to end flight. That's reasonable progress. Given the IPO this was a pretty important flight and I don't think they hurt it (like blowing up on the launch pad would have). So their one step closer it seems.

Landing on two engines was the plan.

V3 Raptors are too powerful, they no longer need three engines to land. They are only going with two from here on out.

So I think it’s unlikely that they altered any aspect of the landing test due to lighting only two engines… as they was the plan anyway.

Hmm, I've seen data that landing the booster on 2 engines was the plan, but hadn't seem similar things about Starship. The difference is the chamber pressure you need in the individual engines. Lower chamber pressure has, in the past, been easier to modulate for precise control. Do you know if they've done any white papers or patents on V3's flow aeronautics?

Rocket engines can’t throttle down very much. Raptor can go down to 40% of its rated thrust, which for V3 would be 100 tons. The ship’s mass is maybe 150 tons with remaining propellant at the start of the landing burn, and probably around 100 tons at the end of the burn. Even at the lowest throttle, three engines would give it a thrust to weight ratio of 2, making hovering impossible and a suicide burn tricky. Two engines gives them redundancy, roll control, and a lower thrust to weight ratio to help with landing precision.

I’m surprised they went down to one engine at the end, because that means they lose most of their roll control. The only way to roll with one engine is to use the cold gas thrusters, which aren’t nearly as powerful.

They have a new RCS system on V3 which appeared to be hot gas thrusters. The landing was 3 engines lit to 2 to 1, using the lesson learned in very early flights to have an extra engine in case one fails unexpectedly.

Is there videos of booster crash?

I doubt it since many of the booster engines didn't seem to relight, the location of touchdown wasn't near any pre-positioned cameras (if there were any).

Not that I have been able to find, the 1400 km/h number comes from the telemetry on the video just before it contacted the water. Presumably one could estimate the return point if you had access to the telemetry and perhaps a platform in the Gulf might have eyes on it. Depends on how far east it got.

Good summary. I was pleasantly surprised that they nailed the re-entry target even after the ascent engine problems.

The re-entry itself looks amazingly smooth compared to V2. TBD whether it's good enough for re-usability (much less rapid re-usability).

But Flight 12 was definitely forward progress.

It looks like the ships software did a phenomenal job compensating by adding that much longer Second stage burn time. It was also very cool to see the sea levels vector to compensate for the thrust asymmetry. All and all that ship is look really realllllly good. Seems like they just need to add some mass to Raptor to reduce the tendency to tear itself apart..

1 out of 33 failing hardly seems like a tendency - probably some minor issue. Relight was likely an issue with new downcomer that was supposed to improve boost back engine feeds for its flip maneuver - that seems to need work.

The videos were incredible. My favorite part was watching the booster flip in such clarity. Normally we don't get full view of it, let alone 4k.

Scott Manley pointed out that it seemed to flip in the wrong direction, with one of the grid fins passing through the plume and inducing a roll. Will be interesting to hear more about that.

The videos are great!, but the rest of it is never going to work lol, just never. Even without a rethink about how to get heavy payloads to another planet this is still good entertainment.

why won't it ever work?

The hardest problem in the entire design had yet to be solved. Having a robust human rated tile system that can be rapidly turned around is a huge engineering challenge that kind of breaks the whole point of the design if it doesn't work. I wouldn't be surprised if they eventually give up and go back to a cheaper throw away second stage, or throw out the tile design completely and try for some evaporative cooling approach, again.

Bear in mind that a lot of what's happening to the tiles now is deliberate experiments to see how much weight they can shave off and how many failed tiles they can survive. Given that the vehicle is routinely surviving reentry at this point, it doesn't seem "hard" to make the tiles more robust by paying for it with added weight. The question is whether they'll have enough weight budget to pay for it? But at this point...probably? Not my area ofc.

"surviving reentry" and being reusable are two very different things, particularly if this is to become human rated.

Human rating is irrelevant to what they want to do, it is only a NASA rating - if NASA wants to ride they will have to come up with a rationale but private astronauts can fly on it.

From day one they said their vision was to replace airlines for international air travel.

I really hope to see the evaporative cooling make a comeback but it seems unnecessary when it’s returning to earth right now.

It would also be interesting to see them do shallower dips into the atmosphere then pull back out and repeat. Like a skipping stone. Lots of expansion contraction, but might work better without tiles.

They did that during the return of Ship on this test flight.

They’ve already shown they can replace all the tiles in a couple of days with removal and new install. No reason they couldn’t do even faster turn around with just re-install if that was needed.

And human rating is a NASA requirement they won’t have to worry about for a few years.

Well at least for one part of that, 90-something percent of the launches don't need to be human rated.

Even if it landed perfectly how is it going to be rapidly reusable with all those tiles breaking and needing repair? Then if that problem was magically engineered-away through some sort of materials science breakthrough, it still makes more sense to me to keep your big ships in a space staging area and your smaller ones as atmospheric gophers.

All what tiles breaking and needing repair? There was remarkably little visible damage this time around compared with previous flights.

There's no materials science breakthrough needed -- the shuttle used ceramic tiles successfully its entire service life. What's needed is engineering work, and that's what SpaceX has been doing.

You know a whole the size of a quarter can wreck the entire spacecraft and make it effectively throw away? Also, you'd want to use this many times. Making a system robust while not requiring months of refurbishment is really really hard.

The Space Shuttle had that problem because it was aluminum with a much lower melting point. It’s one of the reasons they’re using steel.

We’ve seen much larger holes than that in previous tests. Some of the control fins burned completely through.

For some of the tests, they removed a few tiles before launch, presumably to test that. Starship did fine.

coming back in one piece, and being good enough to use for 5 more missions are two very different things. For example, all existing reentry vehicles come back "fine" but they need to be completely remade to go up again.

They already demonstrated that entire tiles can be removed without wrecking the spacecraft.

The quarter thing may have been true for the space shuttle, that doesn't make it true in general.

Deliberately testing its survivability with that failure mode over different parts of the vehicle has been one of the major foci throughout the entire test campaign, and it has proven remarkably resilient. That generalisation pretty much does not hold for starship.

I doubt that. It's made of stainless steel. If it gets home safely they can patch it.

Weren't the tiles one of the worst obstacles to quick turnaround times for the shuttle? It was something like 18 months before one could be launched again, and that's if they were in a hurry.

SpaceX has been specifically engineering both the tiles themselves (e.g. manufacturing) and the way that are used on the ship to be much more rapidly repairable than the Shuttle.

By the end they could turn a shuttle around in ten weeks.

Small ships are less efficient, especially leaving the gravity well. Thats the whole point

Could you tell me more? I suppose a heavy two-stage rocket is not optimized from the point of view of the rocket equation, but I know nothing about this field.

In short, the more stages the better to discard mass once it isnt necessary, and the larger to the better to improve the ratio of (ship+payload) to fuel.

Here is a decent summary.

https://gemini.google.com/share/121466b300c1

This is only true to an extent. Yes, a larger rocket means a better mass:payload ratio, but a larger rocket also means more mass in absolute terms, and more mass means more fuel, and more fuel means more mass, and more mass means more fuel, and more fuel means more mass, and so on. This is "the tyranny of the rocket equation", and it places an upper bound on the size of rockets that need to carry their own fuel for a given gravity well. And because the larger absolute mass of a larger rocket means more fuel, which means more cost, it relies on actually being able to find enough paying customers to fill out that payload capacity every single time. This is why, for example, despite the existence of jumbo jets (which have a better mass:payload ration than smaller planes), most passenger flights are not on jumbo jets, because there's just not enough demand on most routes.

They’ve already demonstrated they can replace all the tiles in a couple of days - even if they continue to have some fall off it won’t be an issue.

The design is wrong in a fundamental level. The rocket equation says it all. By adding so much extra weight for “reusability” (which they don’t release numbers on in terms of cost and means nothing when the rockets end up in the Indian Ocean) they nerf the performance.

Hence tens of launches to make the moon and zero payload to orbit in a dozen flights.

This forum is full of fanboys but their arguments don’t hold up to reality. Saturn 5 had men on the mock. This much time and this many rockets in. Starship has launched a dozen times for $15B - over a billion per launch - and it hasn’t made orbit.

Say what you want about SLS, it works.

Musk is a conman and Starship is a failure. Like the F-35 supposed libertarians and conservatives will happily throw money at it until it achieves its targets, after multiple redesigns, and then claim success as if the neigh-sayers weren’t right all along.

While we're downing this I'd like to add I think the "I've written X in Rust" posts are also cringe :D

SpaceX does an excellent job at videography. Sad that Nasa flew its Artemis mission with potato cameras.

Hey, we have everyone watching, our funding might in part depend on interest and awe…

Not just space-potatoes… but missed the separation shot on the live feed. How in the hell!?

SpaceX probably spends a lot of money on marketing/public relations creating great media. I'm guessing NASA's on a shoestring budget for that kind of thing.

SLS was far from a shoestring budget and both NASA and Boeing/Northrop/etc have equally strong incentive to provide solid coverage for the public to keep the jobs program going.

SpaceX has much better infrastructure for video with their satellites and are just generally more competent at production

One of SpaceX’s founding reasons was to promote STEM and they continue to do that with their video productions.

I think the ship really punted the booster during stage separation. And caused the boost back failure from sloshing.

Also I think Ship now has methane thrusters on it. They were operating with a clean blue flame in short purposeful bursts.

If we look at the venting from the propellant tank (around T+16:15) it looks thick white closer to the vent, becoming more transparent and blue as it expands. That's just sunlight scattering on the particles and density fluctuations in the flow.

A good cold gas thruster produces a lower density, more expanded flow, which looks blue for the same the reason the sky looks blue.

One can compare this to the exhaust from various Falcon-9 engines and thrusters when it is illuminated by the sun on the backdrop of the night sky: https://youtu.be/JRzZl_nq6fk?t=193

From what I've read there was "unintentional mixing of fuel and oxidizer" which caused a fire in the engine section, so the engines automatically shut down. I don't thing we have official word yet, though.

I'm concerned about the cracking clearly visible on the heat shield tiles. It doesn't bode well for rapid reusability.

I thought the tiles were designed for easy replacement, so not a big concern with replacing cracked ones.

The tiles ablate. The shuttle returned from every mission with missing tiles.

Shuttle's tiles not being durable as hoped is what killed it's turnaround time.

The problem was never solved and turned what was supposed to be a few days into weeks or months. Every mission the shuttle had to go back into the assembly building and have all tiles inspected and potentially replaced.

Shuttle tiles were also unique per position and starship tiles have a few base forms that are interchangeable

I would also believe that a robot could inspect and replace tiles a lot faster than humans.

Shuttle tiles were also bonded to the body, which I don't believe is the case with most of the Starship tiles.

I thought Starship has many unique tiles, they ended up needing to vary thickness to match heat patterns to save weight and have some complicated geometry near the fins.

Shuttle had more unique perimeter shapes, but starship still has a lot of variation due to tickness. I don't know if that is easier to vary in production or still needs custom molds for the variation.

Total 6 shuttles built over 35 years. SpaceX already crashed 12 over 5 or so years.

Obviously doesn’t guarantee they’ll find solution, but fast iteration will definitely help.

The tiles are not supposed to ablate - they're supposed to be ~fully reusable. That said I think it's plausible that the much higher iteration speed and lack of a need for human-rating (at least during reentry, for now) will allow for more success than the space shuttle saw with its similar approach.

The shuttle required long expensive refurbishment after each flight.

Just made me realise, this is just like the F-35.

Its turn around time is ridiculous, it has to be maintained with specialized equipment/hangers, along with external contractor assistance.

Compared to the Gripen, as an example, which can land on a freeway and be up in the air again in a few minutes.

One was designed to be used in war, in desperate scenarios, with no ability to coddle it. The other, the F-35? Is designed around milking the taxpayer as much as possible, and employing people in as many politician's states as possible.

The shuttle was like that, I think. Which is really sad.

The F-35 is designed to be able to break into and defeat modern air defense networks.

The gripen is a much less capable non expeditonary platform designed to maximize asymmetric losses if sweden is invaded. As a small country sweden has to follow a porcupine strategy to deter invasion.

Presently the actual comparable to the F-35 is attritable drones, which is why every mid-size and major power is developing them.

The Russians have been trying to use attritable drones for years to break down the Ukrainian IADS and have not yet succeeded. Meanwhile the Israeli F35 fleet with no direct American support was able to crack open Iran’s air defenses at the start of the Twelve Day War with relative ease.

Israel used long range air launched ballistic missiles (ROCKS) and cruise missiles, mostly from F15s, to degrade Iran's ancient radar network and S300s (which allegedly weren't moved around much). That was before the invasion, so Iran didn't have much long range radar coverage when the war started.

The shahed/geran is just a low cost cruise missile. The f35 comparable is a loyal wingman that does manned/unmanned teaming and carries A/A and A/G.

Loyal wingman aircraft would have to be stealthy though to do it.

The whole point of the F-35 is that it is a stealthy platform with a very advanced set of targeting and weapons capabilities: but the issue is a similar capability drone would basically have a similar price tag since the expense of the pilot is minor in that case.

Like plausibly the best autonomous F-35 platform would just be an F-35.

Such a platform does not exist today in active service in any military (so far as we know).

>Its turn around time is ridiculous, it has to be maintained with specialized equipment/hangers, along with external contractor assistance.

>Compared to the Gripen, as an example, which can land on a freeway and be up in the air again in a few minutes.

I have no idea where people got the idea that the F-35 requires a major refit after each sortie or that it needs climate controlled hangars, but there's literally no truth to any of it.

The turnaround time for an F-35 after a mission in a wartime scenario isn't going to be much different from any other older fighter jet. Refuel, rearm, get back in the air.

One of the key requirements for the F-35 programs was to minimize extra care needed for the RAM (Radar Absorbent Material). Unlike older stealth aircraft the F-35's ram is "baked in" to the aircraft skin, rather than being a coating. The F-117 and B-2 require climate controlled hangars because their coatings are old and delicate, the F-22 doesn't, but needs regular touch-ups for its coating, the F-35 is just left sitting outside most of the time regardless of where it's operating, a desert, the arctic, a jungle, the deck of a ship, you just leave it out there. The only common maintenance done on the F-35's RAM is replacing a relatively small amount of special RAM tape which is usually used around the edges of the access panels which are opened for other types of maintenance.

I think there's also some exaggerations about the differing highway landing capabilities of various aircraft. [1] is a video showing Eurofighter, F/A-18 and F-35 all landing on a highway in an exercise. Capability with stores and fuel load is another thing but I've read material that doesn't find the contemporary aircraft drastically different in that regard. Now, maintenance hours per flight hour and general operability certainly are interesting topics and there could be large differences.

1: https://www.youtube.com/watch?v=kKbgtixpfIc

Landing is the trivial part, though the USAF traditions of "FOD walk" do seem funny to air forces where donations you found out the aircraft spent whole day flying with maintenance toolkit left in intake.

The maintenance is the real difference - US specifically USAF gear is designed for nice air conditioned hangars to do regular maintenance, Gripen, MiG-29, and to way lower effect F-18 (when compared with F-16) - the first two assume forward bases without ability to do major maintenance, and even the latter (and other carrier adapted ones) promote things like quick swap engines because that's no space for hangar queen to have deep engine maintenance just so engine vendor can claim long time between overhauls

The main reason the Mig29s have a reputation for easy maintenance is because they don't replace parts, they just throw away the whole airframe. The structural and engine service life is like 1/10th that of western fighters.

The Gripen is a light multirole aircraft like the F-16. The F-35 is a stealth strike fighter. It requires another level of special care to maintain its stealth performance. If you want mass-produced stealth aircraft, that's what's required. Stealth aircraft up to this point have been in extremely limited numbers at astronomical costs.

The f35 has been produced in sufficient numbers that its purchase price is lower than that of 4th Gen fighters. The maintenance cost is higher though.

True, its really the first mass produced stealth plane. Everything before it has been in pretty limited numbers (even the F-22 didn't break 200). We definitely could build a new non-stealth fighter that's cheap to build and maintain but that's like saying we should design a new incandescent lightbulb, the money spent on development just isn't worth it over just building more F-35s.

Gripen will not be able to fly higher than tree lines in zones with active anti-air. Russia can't really use any of it's air power in Ukraine war, for example.

F35 can actually do something in such scenarios, as detecting them in the first place is hard.

Russia can and did start the war with using airpower but stopped due to losses. Currently Russia is using its airpower to lob guided bombs which is effective due to the limited range of ukraines missiles. They have nothing comparable to the R33.

Agreed and specifically in the case of the Gripen the “test condition” was “Needs to be serviceable by a few conscripts working under the direction of one person who knows what they are doing”.

It’s an extremely different design goal, the US doesn’t mind exotic weapons that require exquisite (and expensive) methods of servicing, they have the budget and the assumption that a well equipped air field will be immaculately maintained.

Meanwhile the Mig-29 designers assumed it’d operate from damaged/poorly maintained fields, so on the ground you can shut the primary air intakes and it uses ones on top of the plane to get air, drastically reducing the FOD risk on taxi/takeoff.

I do wonder how well the F-35 would fare in an actual shooting war against near peers when all the peacetime assumptions breakdown.

That's a reason the Mig-29 is no longer in production. Point defense fights are obsolete.

The F-35 was just in a war, in Iran. It performed as expected and was able to roll back Iran's air defense network in days.

>It performed as expected and was able to roll back Iran's air defense network in days.

"Rolling back" Iran's air defense seems like very fuzzy phrasing. Certainly, Iran was not able to close its own airspace, nor prevent ongoing airstrikes on many American and Israeli targets. At the same time, my armchair observation is that a great many US and Israeli airstrikes were accomplished using stand-off weapons [1], which would not have been needed if the United States and Israel had achieved 'air supremacy'[2] as has been the case in America's conflicts in recent history.

The observed trend in USAF readiness has been downward for some time [3][4]. Air war is more than single sorties. If you have anything resembling an accurate summary of sorties flown, targets successfully hit, and number of combat-ready aircraft throughout the (currently on hold) war, and so on, please share. Absent such detailed information, all we have are various degrees of speculation.

1. https://news.meaww.com/us-used-tomahawks-himars-standoff-wea...

2. https://en.wikipedia.org/wiki/Air_supremacy

3. https://www.airandspaceforces.com/air-force-mission-capable-...

4. https://havokjournal.com/culture/military/the-hidden-erosion...

Iran innovated with the loitering optically guided missile. This was enough to dissuade a longer air campaign. Political support is low so any losses create a political problem.

    > an actual shooting war against near peers

Who could this be? I can only think of Russia or China.

Well, every mission that it returned from it had missing tiles. That is not the same thing as returning from every mission.

Starship’s tiles are not designed to ablate. They are intended to last multiple flights.

That wasn’t cracking.

I mean ... step 1 is probably fixing the part where it lands in the ocean, falls over and explodes. Once they've done that and can get their hands on the tiles I'm guessing they can continue to iterate there until they get a more easily reusable design.

That part was intentional

Dang, a random HN user solving all the world's problems yet again, what would humanity do without you random HN guy?

The final issue that led to the scrub was that a pin that held back the QD arm got bound and would not release.

It was also noticeably wayyy faster off the launch mount. These V3 raptors are pretty fierce. It took off so fast it seems destined to be stretched imminently.

V2 they were bring them up in groups. V3 they start them all up at once.

“seems to have hit the water harder than expected and was very off target.”

SpaceX’s people were saying it was on target, and it seems to have landed in about the same position relative to the camera buoy as previous flights. I don’t think there’s any evidence to call it off target. The landing and toppling looked the same as previous flights too.

You've mixed up the ship and booster.

Yep, I did. Thanks.

First item is wrong - scrub was due to hydraulic pin on launch QD arm not releasing (apparently when released it caused an arm vibration that triggered a re-lock) and they had to improve arm stabilization when pin releases.

Also Ship reached original intended orbit.

A tower catch was never in the plans for this flight.

Booster is a totaly new rocket. It did launch with acarity and by any other standard did well, but the failure to relight could be anything, but I am going for the giant fuel feeder tube bieng the failed part, based on nothing more than how tickled they were with it,and tank baffling bieng a dark art. The slosh of the fuel durring the flip is going to produce an internal tidal wave, lots of stuff gets "tested" there.

Fantastic summary. Thank you!

They've made great progress but have a bit left. It's always the last mile, isn't it?

It's so cool seeing progress in this space (sorry).

Did the landing burn light two engines as expected? It happened fast, but the graphic made it look like only one lit. If that’s true, that would be impressive as only lighting two was meant to be a test. At least according to the live stream hosts.

Thank you for your service! You nailed every single detail.

This was as good, if not better, than the livestream itself!

Thanks for the summary It was a good read and you nailed it!

And the “side effect” of polluting the ocean? SpaceX is more of a let’s only do things half way kind of company it seems.

Polluting with what? It's mostly stainless steel and other inert materials.

You do realize that every company so far has been doing that to every booster for every launch, right?

There's 2 options if you don't want to drop stages back on earth. You don't launch, you land the stage.

SpaceX is the company that pioneered propulsive landing of a booster. You can say a lot about them but not that they pollute with dropping stages in the ocean. Even in absolute sense that doesn't happen often and that's ignoring that they put over 90% of all the weight in orbit nowadays

The takeoff looked almost normal but I noticed a slight drift from vertical, likely because one of the engines was dead or dying. Overall the V3 is supposed to be an upgrade but actual progress is more or less stalling compared V2.

It is supposed to tilt away from the launch tower immediately, you can see this on previous flights. This keeps the engine plume away from the chopsticks and top of the launch tower.

Also an additional goal is to get the booster as far away from the pad as immediately possible in the event it falls back down.

The payload (100t) is at least double that of previous flights. It’s largest spacecraft ever flew. That’s some stalling

I have a hard time saying that its stalled. For one we don't really have the hard data to quantify the V3s vs V2s actual performance. On their first flight we lost one on the booster and one on the ship. I don't know that the boostback burn problems are related to the engine themselves given that they had multiple failures and a premature cutoff. That feels like thats a problem upstream of the Raptor.

There is a slight tilt normally, but I agree it was more than usual.

Flight 12 flew a different flight path than previous test flights.

>> As far as overall progress from previous test flights goes, they're at least treading water while making many large changes.

What a kind take on what is again the continuous trend of: Every flight fails for a different reason. And they still cant make the most basic use case: https://www.instagram.com/p/DWzTFAEAhSe/

> The amazing views of the plasma during re-entry, something never seen live before starship, are now routine.

The word "live" is doing a lot of work here. Astronauts used to film the plasma going past the windows of Shuttle.

I remember as a kid my science textbook had a still of it to illustrate plasma.

> The word "live" is doing a lot of work here

A latency of a few seconds for streaming video compared to several months for a still photo from the Shuttle seems an entirely valid use of 'live'.

The difference here is a matter of AV technology improvements in the last 40 years.

The difference is having high-bandwidth signal pointed up / away from the wall of plasma.

And having such a large craft that a hole in the plasma envelope exists above the craft that can reach Starlink.

Lots of engine failures. Doesn't exactly bode well for a company looking to go public immediately. One of the engine failures was not on the booster but Starship as you noted, and that is a bit unexpected. I don't think they have spoken about it being equal in capability with one engine out, right? Those engines don't move around to compensate IIRC.

Not sure how you come to that conclusion. The capabilities can overcome loss of engines. The fact it was successful with loss of engines shows it is working as designed.

No, it just means the mission happened to be salvageable because of its parameters. The booster is designed to have engines out and can compensate because it has so many engines and many of them are on gimbals. On starship, the vacuum engines aren’t on a gimbal. I’m not sure how it could compensate for one of three engines being out.

Some are on a gimbal and they specifically talked how others gimbaled out a bit to compensate. This is specifically something they designed in and not an accidental lucky save. In this flight they didn’t intend to test “one engine out” feature but it worked out that way.

See my other comment. The vacuum engines are NOT on a gimbal. None of them. The sea level engines on starship and several of the engines on the booster are on a gimbal. But not the vacuum engines for space.

EDIT: I cannot reply further in this thread, but my understanding is that the non vacuum engines are not intended to stay lit throughout the orbital flight in a typical mission. If they are, they can gimbal and compensate.

> The vacuum engines are NOT on a gimbal. None of them

I said some raptor engines are on a gimbal, not vacuum engines.

To be precise, the three central engines can gimbal up to 15 degrees. That can control the thrust vectoring when an engine fails, and that’s what happens in the last flight.

Since the flight already happened and we know it didn’t spin out of control (unless you imply their diagnostic and telemetry was completely off and the engine was actually on) something must have compensated for the failure. It wasn’t magic, it was in fact the central 3 engines that did that.

You may be confused because those are called sea level engines, but that doesn’t mean they can’t work in vacuum.

I think previous comment means "on a gimbal" as in "angled at a non completely prograde direction" (presumably angled such that each engine points through center of mass so that none of the engines impart a torque)

The person you’re replying to is trying to play rhetorical word games.

The upper stage has six engines. The outer three engines are “vacuum engines” (optimized for operation in space). The inner three engines are “non vacuum engines” (optimized for operation in the atmosphere, at sea level).

The outer three vacuum engines are not gimbaled, but the inner three sea level engines are. Thus, it is completely accurate to say that they gimbaled some of the engines to compensate for the engine failure.

They explicitly said that they have engine out capability on the ship in the stream.

That’s for the booster (the big lower part) not for starship (the upper part that continues to space). They were surprised to have a vacuum engine out. In space there’s no atmosphere so you can’t use fins or wings to change direction. And if the engines can’t move around, you only have thrust and gravity and the tiny attitude adjusters to direct your ship.

You're simply wrong. The non-vacuum-optimized engines on the upper stage are still functioning in a vacuum, and their ability to gimbal to offset the loss of one of the vacuum-optimized engines was planned for.

Watch the stream again starting with the ship burn. They explicitly said they have engine out capability on ship. The sea level engines on the ship are running and gimballing.

I am just delighted that SpaceX continues with the "good enough" pace of development here, at least at these phases. Rapid iteration of build, test, learn, and improve rather than wait for perfection.

They are willing to have "negative outcome learning experiences" to gather data quickly. and, of course, data, data, data.

I like it because I know what insane amount of red tape has built up to do anything similar in a Gov (any Gov).

Shame they're risking that ability with the IPO. We've seen how irrational and ignorant stock traders are from other publicly traded space companies. Even scrubbed launches cause the price to dip.

> We've seen how irrational and ignorant stock traders are from other publicly traded space companies.

Absolutely true, but ignorant stock traders making irrational trades only matters if company management pays attention to them. Musk will maintain complete control of SpaceX even after the IPO, so he can focus on long-term value rather than short-term ups and downs.

Of course, over time, if more shares are issued, this may change.

The argument for the IPO was to fund their space datacenters project, if Musk could afford to ignore the stock price, that wouldn't be needed.

At this point in time, Musk can afford to ignore the stock price since he will retain sufficient shares (with 10x voting rights if I am understanding correctly) to outvote any board.

However, some projections I have seen suggest even the public raise won't be enough for the entire vision (which includes LEO data centers AND trips to the Moon and Mars). That means he will either have to find the revenue or eventually sell even more shares. It will only be at that point he will care about stock price.

That's the thing, why does everyone seem convinced that the guy who has been making up pie in the sky projects and breaking his promises, won't find himself wanting more money?

Have IPOs ever been about just a one time cash infusion?

He'll be able to raise debt against the massively inflated valuation of the company. SpaceX will have access to far more capital after this than before.

As I understand it he will retain control of the company in a similar arrangement to the one Zuckerberg has at Meta. Anyone who buys SpaceX stock is just along for the ride.

They get all the money on the ipo day. They dont get more if it goes after unless they sell more stock

But by the time retail traders can affect the stock price, SpaceX will have already gotten their money.

In an article from the Financial Times [1], Elon Musk will hold "more than 80 per cent of the voting power". This is pretty similar to Mark Zuckerburg's control over Meta. I don't see any risk of SpaceX losing control of their destiny due to short term demands by stockholders.

[1] https://www.ft.com/content/00382ab9-3dfe-468c-8966-853cd787d...

Elon negotiated with Nasdaq to automatically accept it in Nasdaq 100 index after 15 days, so the stock traders don't really matter.

I asked about this in previous HN discussions. The old rule: You had to wait three months. Is there a meaningful economic difference between waiting 15 days or ~90 days? I don't see it. (For transparency: I own ETFs that track the S&P 500, which has lots of overlap with Nasdaq 100.) To be clear: OpenAI and Anthropic will sure IPO this year or next and have a similar effect -- they will be (or nearly) trillion dollar market caps upon listing.

If you short the company during the first 15 days, your short will get wiped out by the passive index buying. Nobody is going to short the stock to protect passive investors. The short will happen after the bag holders entered the market.

Can you point me to a source that actually says that Elon influenced the Nasdaq?

The whole thing stinks, and we can create stories since some incentives are obvious.

Stock traders do matter since 30% of stock is aimed at retail (very unusual % - what is the norm?)

More info here: https://www.youtube.com/watch?v=-X6YzlY_8tM (sorry, video)

> SpaceX has successfully lobbied the Nasdaq stock exchange to loosen rules governing how and when it adds companies to its Nasdaq 100 index – a group of large-cap companies that it bills as “fundamentally sound and innovative.”

https://techcrunch.com/2026/05/21/how-elon-musk-will-increas...

I’m sure the going of public of the large AI companies wasn’t on Nasdaq’s mind at all.

Doesn't Musk retain full control anyway?

Yes, sources say over 80% of voting power.

The views from Ship's engine bay looked rather ominous -- with the red glow visible in multiple places, and something venting furiously from the broken engine. It was a pleasant surprise that the ship did not explode and not only that, but it even landed exactly on target. Guidance system software engineers have done a very good job!

The booster not completing the return part of the flight was disappointing. They had a similar incident in one of the previous flights, when they tried to maneuver the booster too aggressively immediately after stage separation which caused problems with the fuel supply. If it was something similar this time, it might be solvable by changing just a few details of the maneuver. So, maybe it is not that huge of a deal.

There were many cool things in the webcast, from them showing the catamarans that are deployed at the landing site, to the views form the cameras on-board of the "satellites". The first few minutes after liftoff were just amazing visually.

Yeah, the new raptors and engine bay redesign for version 3 of ship seem a lot more robust than version 2, very impressed that the ship finished its mission so well.

Scott Manley goes into quite a bit of detail on analyzing superheavy’s failed boostback, it’s a good watch:

https://youtu.be/2kxanBYTAaY

TL;DR - seems like the hot staging kicked the booster out in the wrong direction, and the ship’s plume impacted one of the grid fins, which would’ve given it quite a big kick. The sloshing just from that could easily have caused the observed issues.

Hopefully NASA ups their game for Artemis III

Well it was supposed to be a moon landing, but SpaceX didn't get their stuff developed. Artemis II flew by the moon btw, I think that's more impressive than SpaceX crashing a bunch of suborbital flights into the ocean while behind schedule.

My comment was about the video coverage. Artemis II's coverage for the launch was especially lame. The solitary gopro on the side of the capsule had some cool footage but was just a single camera.

The actual mission was not in doubt

The best part of this flight was seeing the full reentry with no visible hot spots or burn through like we've seen on every previous reentry of Starship. Seems like they have the heat shields really nailed.

Yes, reuse of the heat shield has been the biggest question mark of the whole program and this is by far the best result of any launch so far. This is the first time it looks plausible that you could consider reusing the heat shield.

Plausible? It was always plausible. This flight does nothing to increase or decrease that - it is still very much a wide open question. The goal is full rapid reusability, and nothing we have seen yet suggests that that is possible, because no Starship has flown twice, much less re-entered twice with the same non-refurbished tiles.

It may be that the heat shield is the thing that causes them to miss the BHAG. Starship will still be cool if it needs all-new tiles every flight but let’s not pretend that that would be anything but a miss for the program.

I sure hope they can figure it out.

Keep in mind that all the prior flights had deliberately missing tiles specifically to test how bad the damage would be. This one had no missing tiles so it should have essentially no damage to the hull.

I think the worst damage on previous flights was often not related to the missing tiles.

I think it was due to voids in the hinge area causing localized plasma impingement and heating. The v3 design has a more continuous shape around the hinges.

Right. I’d have to go back and find explicit statements about the redesigns, but as I recall the v1 Starship had the flaps further forward and lots of tiles missing around the hinges. V2 moved the flaps back so that the hinge is shielded more by the curve of the body, but still had not enough tiles. V3 has almost continuous coverage using much more uniquely curved tiles. They look harder to make but not as bad as the shuttle’s.

My favorite part of this launch that others haven't already mentioned: during reentry, the dummy payload satellites were visible burning up behind the ship!

Glad you mentioned this. I was puzzled by the starry looking background during reentry

Curious! Got a timecode?

I do believe it appears at 35:00 into flight

Check around T+50:00

I wonder if the hot separation was supposed to be that hot. Going at mach 5 and doing a quick U turn while there was some weird orange color on the side of the Super Heavy, then (possibly?) losing most engines from it seemed extra chaotic

They had a similar issue with the v2 booster the first time, flipped it so hard it damaged the downcomer and they had to change it up a bit. Seems likely here, that thing thats as big as a building flipped ends pretty quick and is partially filled with liquid.

The side of Super Heavy was reflecting light from Ship’s engines.

Seeing both the Starlink mass simulators deploy and the camera view from the last simulators looking back at Starship was really cool.

The amount of data they must have at this point running so many of those raptor engines has got to be insane... at least 300+ engine launches now -- wow.

Five years ago SpaceX reported that they had 30000 seconds of test firing time on the Raptor, over 567 engine starts. Since them the program accelerated dramatically. Well over one thousand engines had been produced, and on an average day at McGregor test facility the Raptors are fired for about 600 seconds. That would give about a million seconds over five years. That's a lot for any engine development program.

I found this on Wiki: https://en.wikipedia.org/wiki/SpaceX_facilities

    > SpaceX Rocket Development and Test Facility, McGregor, Texas

    > SpaceX calls the facility the most advanced and active rocket engine test facility in the world, and said that by 2024, over 7,000 tests had been conducted at the facility since it opened, with seven engine test fires on a typical day. Despite its low-profile compared to the company's other facilities, is a critical part of SpaceX's operations, and company president and COO Gwynne Shotwell maintains her primary office in McGregor.

Consider for a moment the data requirements for the telemetry system that records those engine runs.

If there's any public info about this I'd love to read it.

Both SpaceX and NASA use LabView. NASA has a relatively detailed description of the engine test stands at Stannis:

https://scholar.google.com/scholar?q=NASA+Data+Acquisition+S...

https://scholar.google.com/scholar?q=Design+of+Electrical+Sy...

A typical test stand would have maybe a thousand channels of relatively slow data (pressures, temperatures, flow rates, valve states, etc), and maybe up to a few hundred of channels for essentially audio data from vibration sensors. This amounts to sub-gigabit per second data rate overall.

If very high speed video / multiple video cameras are used, this could generate massive data rates, but unless something interesting happens it is not clear how important this data is.

In flight, the telemetry data rate from the entire Falcon-9 used to be measured in megabits per second per stage, plus the video stream. It was not a huge amount of data. Presumably now with Starlink they send a lot more telemetry from Starship, but in flight the engines typically have far, far fewer sensors compared to the ground testing.

Very cool. Thank you!

Sort of... this was version 3 of the engine, a fairly big redesign and for version 3 this was the first flight.

Some footage: https://youtu.be/CiWX1nsvqBs?si=lE5autC2y2b8ez2X

At a minute in you can see the satellites being ejected out one by one.

This was good forward progress (V3 mostly worked, clear improvements on heat-shield, near-final Starlink deployment system).

Is this enough progress to keep a 2028 crewed landing? Don't know.

I'm curious whether they are going to try to recover a Starship before trying for in-space refueling (or the reverse). Either way, I think both have to work before they can try for an uncrewed lunar landing (presumably in 2027).

The big question is re-usability. How close are they to relaunching a Starship? They may not know for sure until they can get one back intact. If they can launch at least once a month, maybe they'll make it.

If they can re-fly a Starship this year AND demonstrate in-space refueling, then 2027 can be all about an uncrewed landing attempt. That would make me feel good about a 2028 crewed landing on the moon.

> If they can re-fly a Starship this year AND demonstrate in-space refueling

I'd bet that they'll not try in-space refueling before they demonstrated in-space relight of an engine. So they need to fly at least twice. Or even thrice because to demonstrate refueling you need two Starships in orbit.

IIRC, I think they're off the hook for a 2028 crewed landing because the spacesuits won't be ready until 2030, but if I'm wrong let me know!

It lifts off so rapidly, it’s truly incredible

“Like a rocket” is a phrase for a reason

Eh, they've left some performance on the table:

> Sprint accelerated at 100 g, reaching a speed of Mach 10 (12,000 km/h; 7,600 mph) in 5 seconds. Such a high velocity at relatively low altitudes created skin temperatures up to 6,200 °F (3,400 °C), requiring an ablative shield to dissipate the heat. The high temperature caused a plasma to form around the missile, requiring extremely powerful radio signals to reach it for guidance. The missile glowed bright white as it flew.

https://en.wikipedia.org/wiki/Sprint_(missile)

Maybe v4.

Then there's HIBEX, which hit 400g acceleration.

(Via a Wikipedia rabbit-hole thanks to your comment.)

<https://en.wikipedia.org/wiki/Hardpoint_(missile_defense)>

I’m really struggling to find a point to this comment

It's a joke. An excuse to share one of my favorite rockets. A rocket that happens to uniquely support your point.

I chuckled and appreciated the anecdote, having not read about this rocket before. Thanks.

Because you are speechless?

Such accelleration! Much wow!

Having a faultless payload deploy and a pinpoint landing after losing a whole vacuum engine (one of 3) so early was an unexpectedly amazing performance. I suppose they gimballed the inner non-vac engines to the max and burned longer, next level adaptability.

Most obvious improvement was having no re-entry heating problems, secondmost was deploying with zero issues and with a faster pace. It appears they decided to pause the "horizontal" movement of the pez dispenser before a final push away, probably to avoid vibration causing those "bonks" on the payload door, like we had once before.

Starship moved around so much there were a couple times I thought for sure it was out of control. Oh and takeoff looked very late but I’m sure I’m wrong. Heh I was watching the deluge turned on with no ignition thinking the whole things is about to hard start and explode.

Big takeaway for me is that the reentry and “landing” of Ship looked great. For the first time, it felt like they’re really on the path to achieving upper stage reuse. That was always the biggest “reach” of the entire program in my view, and today they took a major step forward.

Is it disappointing that they had a couple of engine outs, and also trouble with the booster relight? Sure. Do I have even a little doubt by now that they can fix these problems? None whatsoever.

The success of Ship 39 today was a big, big deal.

Does anyone know how important hot staging is?

It seems to give the booster a real kick - what's that do to turbo's and fuel movement?

You've got hot exhaust onto cold cryo fuel tank header?

You've got to carry more mass in terms of protection for the tank?

Is doing MECO and then push and then get 100 yards apart or something before second stage / ship engines kick on a big enough penalty to justify all the extra complexity?

All the velocity you lose in any kind of throttle back is velocity you never get back. It hurts you the entire rest of the flight.

aka "gravity losses". Every second you're not (yet) in orbit you're losing dv / cargo capability.

IIRC, it basically removes the need to do a handoff between the relatively tiny tanks that can reliably pump fuel and oxidizer in zero-g from the main tanks that struggle with that in starship, and completely removes the need for any of that on the booster.

They were originally planning to not do hot staging, aiming instead for a somewhat funny approach where they spin the ship and booster slightly so they are passively separated by centrifugal forces. A bunch of things went wrong in the first test flight, so this was never attempted, and they switched to the "simpler" hot staging in flight 2.

eliminates the need for ullage thrusters to force the blobby propellants in the tank onto pump intakes

This incremental progress, far smaller improvements than planned, has put them so far behind schedule I'm not confident this design is any good. Still haven't done orbit. This launch was not a smooth launch. SLS by contrast seems to work. Why did nasa contract SpaceX for the lander. The whole plan is bad.

There's plenty of finicky systems which go on to be good systems with a lot of work. Some things are just hard, a lot of the time you just don't see them being hard so publicly.

Sometimes a design is poor and needs a lot of modification and patches to kind of work.

That's software engineer "I'll simplify it!" logic and the score of it's successes is far too low for how common it is.

If a design with a bunch of modifications works, then it's a good design. Thinking you need to clean sheet redesign everything is how you get Second System Syndrome.

Are you serious? They just launched a completely revamped version of Starship from an entirely new pad, and still hit almost all of their planned milestones while demonstrating that the design is reliable enough to handle a missing engine.

Go back and look at the original plans and projections. Constantly redesigning is not something to be proud of. I call it vibe spaceship design.

Two successful test missions over 15 years using a dead end rocket and ship design for $50+ billion is not something to be proud of either.

Starship's design is so far beyond where the rest of the world is that even if it takes another 3 years of iteration to perfect it will still be by far the best rocket in the world for many years afterward, to the point where it would hardly make sense to choose to launch on any other existing or currently in development rocket for any reason other than political ones.

You can't perfect a flawed design. The whole thing might be too big and complex to launch reliably in the way that NASA and spacex need.

Its plainly obvious that the lander should have been launched on falcon heavy. I do believe that starship will eventually work but it'll take a long time to get 10+ launches working in a row so that a lander can work.

I am genuinely curious what causes the different colors of plasma that were visible throughout the descent. I saw at least purple, yellow, and shades of green.

This is amazing.

Cool splashdown.

https://www.youtube.com/watch?v=CiWX1nsvqBs&t=173s

So crazy that the NBC video didn’t include that! It’s still not as bad as NASA skipping the good view of the live separation.

Better to always watch from the official source, the only one that is in 4k too AFAIK

https://x.com/SpaceX/status/2057843818859209147

Thank you!! I started watching on NBC and there video quality is potato.

I don’t keep up with them. What’s different compared to v2?

Now the flaps don't melt! The tiles don't fall off!

It's a major overhaul of the design they've been working on for a long time. There was talk of v3 fixing the problems in early v2 test flights. The booster is v3 as well which presumably is why they had some problems. I believe this is also the first time they flew the v3 engines with the plumbing fully integrated in a single piece housing they 3D printed.

Quite a bit has changed. Here's the highlights: https://www.spacex.com/updates#starship-v3

Biggest things changed are the plumbing inside the booster (much beefier, now supports starting all engines at once), and a new version of the engines, raptor v3 (more robust, and saves about 1 ton of mass per engine).

Cool image comparing the three versions of the engine. In addition to making the engine itself a bit lighter, moving all that stuff to the inside of the engine let them eliminate a whole bunch of heavy shielding in the engine bay.

https://ibb.co/pr0gx688

Oh man, so glad I stayed up to watch it. Kind of a rough start (but it's the 1st flight w/ new redesign, new engines, etc), had an engine out on both booster and ship, but the views were absolutely worth it. They managed to get the last satellite to connect to starlink and download the footage of the ship in orbit. Even with an engine out, the ship managed to reach orbit, deploy all the satellites, re-enter, flip and soft splash into the ocean, near a buoy! And on top of that we got the drone views of the landing. Fucking spectacular views.

It really was an amazing sight to see.

I'm guessing / hoping that the engine outs we're planned, or that they ran the engines with slightly different parameters to test them. If it's just unreliability then it might be a hard problem to solve.

> If it's just unreliability then it might be a hard problem to solve.

It might, but it certainly helps having a ton of them around. Given that they used 42 of them today and 2 failed in some fashion, we'll call that a 1:21 failure rate. On a more typical rocket with say 10 engines (eg falcon 9), there's a good chance they wouldn't have seen the same failure till flight 3.

> Given that they used 42 of them today

20+10+3=33 on the booster, 3+3=6 on the Ship, total 39.

I remember Elon said they want to add 2 engines to the first stage, but that still would be 41. Where's the 42th supposed to be?

I messed up, for some reason I had it in my head that there were 9 on starship, so 33 + 9 = 42.

It’s something like up to 6 can fail and it keeps going, seems pretty good. I know they did some stuff like remove a heat tile to get failure feedback, wonder if engine was planned or accidental

Accidental since they didn’t make the sub-orbit they were aiming for and thus couldn’t test engine re-light.

Adjusted wald intv'l gives us for 2 observed out of 39 gives us a "true" failure rate from 1 in 200 to 1 in 6 (95% CI).

Very first flight of a brand new engine type (Raptor 3) with totally reworked heatshielding/plumbing/sensors/control systems/etc.

Which is true, but at the same time: this is Starship Flight 12.

The whole point of Starship is that it's a reusable vehicle with easy turnaround and quick maintenance. And in particular it's supposed to be different than the other reusable vehicle with easy turnaround and quick maintenance, which turned out to be sort of a boondoggle.

Yet, they've now hand-built and destroyed twelve of these things across multiple redesigns, and it still hasn't completed its design mission once. In fact basically every launch has unexpected major failures.

As poor as its safety record ultimately ended up being, the shuttle launched successfully on its very first try. And we only had to hand-build five of them. And lost two, sure, which is still a lot less than twelve.

Yes yes, I understand that iterative design has merits and that the ability to rapidly prototype and try things in the stratosphere allows for less conservative tolerances and better ultimate performance.

But does it really take 13+ tries?! At what point to we start wondering if we have another boondoggle on our hands?

But all of those 12 launches happened in just 3 years, and cost a tiny fraction of other major spaceflight development programs.

For reference, SLS has been in development for 5 times as long, and cost 15-20 times as much, as Starship, and they still haven’t landed people on the Moon, which has been one of the stated goals since the Constellation program in 2005.

I don’t see how the number of failures matters if the end result still happens faster and cheaper than anything else.

Recent SpaceX IPO filings put that 'tiny fraction' at about 1/3 or 1/2 of SLS. $15B total investment with about $4-5B of that figure from US gov. Is starship more than 1/2 or 1/3 of the way to a human rated Artemis II style mission? The main reason starship costs less to test (apart from the SLS jobs program baggage) is because of design choices which prevent it from performing such a mission without significant further tech development.

'5 times as long' is dubious too. SpaceX claims to have been working on the design since 2012 vs 2011 for SLS. Ultimately though the start date of a complex program is not well defined, as early conceptual design stages can take years without leaving the drawing board. Government needs to put a start date on such efforts for legal/budget reasons, but a private company does not.

Also relevant - SpaceX has been given a lot of tech and expertise from NASA at a tiny fraction of the cost and time it would have required them to develop it themselves. Therefore, the costs of NASA programs like space shuttle actually includes some of the development costs of SpaceX.

Both programs pale in comparison to Saturn V, which was faster, cheaper, and more technically demanding at the time.

Moreover the two lost shuttles included human lives. Better to blow stuff up with demo payloads now before sending up large contracted payloads or worse human beings!

I couldn’t believe my ears when I first heard that the second ever flight of SLS was going to be crewed.

It worked out in the end, but I can’t imagine being so confident in a new system, no matter how much money and brainpower has been spent to make it safe.

> Better to [...]

That's undeniably true. Nonetheless "Better than the shuttle, which sucked" isn't the design goal.

The question is not even just "is it better to blow up 12 Starships?", which would probably still be true. It's "Why isn't Starship working yet?" and the implied "Maybe Starship sucks too?!".

They failed far more than 12 landings before they started reliably landing boosters, and people made the same “maybe this is a dumb idea” comments back then too.

It seems premature at best to make any sort of triumphant comparison, when Starship has yet to complete an orbit of the Earth while SLS has carried humans to the moon

It's especially ironic to talk about not landing on the moon yet, when it is the landing system--aka Starship!--which is the primary bottleneck for that goal

I am in no way trying to make this some SpaceX vs NASA pissing contest, but there seems to be a tendency to overvalue Starship's promised achievements to SLS's real ones

They could have made orbit at any time on any flights that reached space. they have deliberately not done so because that is the safest thing to do when developing.

I don't see how that gives them credit. Yes of course they could have burned the engines for a little longer and reached orbit

Were they confident that they could relight the engines and get it back out of orbit? This very test seemed to answer "no"

It's Starship that's supposed to do the landing on the moon. If the next few test flights don't go well it won't be ready for Artemis III next year.

If you can afford it, I'm sure anyone developing a rocket would prefer to do it this iterative way. I don't really understand the complain.

If I was developing a rocket, I would prefer making it work first and then iterating on it.

SpaceX is iterating for the sake of a narrative. None of the development goals for Starship require reuse of the upper stages, that's a bonus that should happen after Starship started deploying Starlink satellites.

There is no point in having dedicated reentry tests. Every test flight comes with a free reentry test included.

Even if a company does iterative rocket development, it would never follow the SpaceX way of putting th most difficult but simultaneously unimportant objectives on the critical path at the expense of the essential objectives.

The development program of Starship is basically ignoring every aspect of Falcon 9 development that made it the success people take for granted these days.

The part that makes no sense to me is why they are going starship scale rather than falcon 9 scale. Had they done their prototyping on a rocket with 9 engines on the first stage and 1 on the second, they could have gotten to raptor 3 (and a falcon 9 replacement) while blowing up way fewer engines, launch complexes, etc. There's a reason Spacex started with the falcon 1 rather than the falcon 9. It's a lot cheaper to blow up fewer engines and smaller rockets while you're developing a new rocket engine.

The biggest part of this test campaign is learning to build Ships and Super Heavies quickly and efficiently. They are not just testing Starship, they are also testing and iterating on Gigafactory.

Time costs money as well. Im guessing the cost is irrelevant compared to a faster timeliness.

part of my argument is that if your test platform is 1/4th the size, you get much faster iterations. the path they took meant that every test destroys 39 engines, a huge amount of rocket fuel, and all of their construction. A falcon 9 design (9+1) with raptor engines would let them already have a falcon heavy capacity vehicle because everything about a smaller scale vehicle is way easier

They already have a falcon heavy capacity vehicle - Falcon heavy. I guess I am confused about what you think they would get out of testing with a smaller platform. It seems like most of their focus is specifically figuring out how to make a large platform work. What do you think would translate?

If they had a 1/4 scale test platform, they wouldn’t be testing what they need to learn.

They wouldn’t be testing how to transport, generate, store, load and offload massive quantities of liquid oxygen and methane quickly and reliably.

They wouldn’t be testing the simultaneous ignition of 33 engines, the massive fuel demands of that (which appears to be very difficult), and blasting the pad with 33 engines.

They wouldn’t be testing the structural loads and stresses of max Q of such a massive vehicle.

They wouldn’t be testing tank slosh and pressure issues of massive tanks.

They wouldn’t be testing the heat loads on the shield with huge mass.

All the manufacturing too would be learning how to do things on such a large scale.

In short, they simply wouldn’t be learning the lessons they need to learn to make this work.

The point was more that there is a point where (to borrow the software terminology) "iterative design" becomes "death march". Trying a few times in the early days and being willing to throw stuff out and start over is a powerful tool.

I think blowing up a handful of rockets is a fine idea. But at some point you have to ask yourself if it will ever work? Why are we on a another engine redesign? Why is this the third iteration of the second stage? How many more?

And what number is that point? Six? Nine? I'm thinking thirteen may be getting into the danger zone.

In a somewhat similar situation Sergey Korolyov stopped his colleague in front of the Party officials asking a similar question and explained: "We are exploring terra incognita, this is the process of getting knowledge". He was sort of right - even though there were many specific engineering problems, and many of those were rather solvable, especially in hindsight, overall process was stepping into the unknown.

Here we have a cutting edge rocket design - scale, sophistication of engines, design goals - and a commercial evaluation, which path would get to the intended success cheaper. NASA doesn't like public embarrassments, and, as Henry Spencer reminds us, when failure is not an option, the success could be quite costly. So NASA spends billions and many years for a fragile system. If the goal is an airline-like operations, the design should be thoroughly shaken up. It's known that no simulation, no static testing can equate the actual flights in the ability to get the data best describing what conditions the system will encounter in real use. And also, given the industrial scale of Starship production, each flight hardware costs way less than if we'd built them manually, in quantities justifying naming each unit separately.

In Soviet Union, where rocket departments were part of artillery, the testing with actual launches seemed logical. In this case the approach to run a massive test flight program seems logical too, and we can't complain about the lack of progress - first Starship had way less capabilities and performed way worse. In USA we had more than 1000 tests for injector head for F-1 engine in Apollo program, and this number was justified at that time. Starship is way bigger - but the progress is also undeniable, and it would be odd to stop test flights now, when the 3rd iteration of design looks promising.

So, while we can't pin a particular number of tests, I don't think we should worry yet. This year and the next one should be important for Starship program, given SpaceX commitments to help NASA Artemis. If we won't have orbital Starship then - we can come back to this question.

> Sergey Korolyov [...] explained: "[...]"

Not relevant to the discussion, but while it's fairly bland and feels right, I can't find a reference to that quote anywhere. Is this a hallucination? Where did you get it? The fact that you're quoting directly implies you copied it from somewhere.

My quote is inexact, of course.

https://www.nasa.gov/wp-content/uploads/2015/04/636007main_R... page 458 of 831, exact quote:

“Everything that Chertok was about to tell you will take a lot of time. An explanation of what caused all the failures while solving the soft landing problem is presented in detail on these posters, focusing on each individual launch. But there is one general cause that explains everything—this is a learning process. In our plans and schedules, we did not make provisions for the expenditure of resources and time on the learning process. That is where we made our mistake; we have paid for it, and I dare say that in the very near future the mission will be accomplished. Our learning process has taken us down a rough road, but we have gained invaluable experience. I request that the commission permit us to conduct a launch and make the final decision, if you deem it necessary, based on the results of that launch.”

So what if they blow up literally 100 rockets, if they can eventually perfect it faster and more cheaply than the traditional approach, recently typified by SLS.

SpaceX have already proven that the iterative approach works with Falcon 9, literally the most successful rocket program ever. SpaceX have also proven that this specific Super Heavy/Starship rocket design isn’t a dead end. Criticising them for failing to succeed in the future is a valid but uninteresting opinion.

The point is that by the 100th test flight, your competitor has already proven the first version of their design to the point of being retired by the second version that has a 55% higher payload capacity and had a history of dozens of moon landings and is busy manufacturing solar panels on the lunar surface using lunar regolith, thanks to the orbital infrastructure for lunar transit that they've painstakingly built over several initial launches is paying off.

Meanwhile SpaceX having proven that the iterative development cycle works by first building a successful rocket and accepting customer payloads as soon as possible and then upgrading the rocket as they flew, fell behind because they decided to abandon their origin and instead try to delay success as much as possible which is antithetical to their origin and the success of the launch vehicle. In their hubris, they doubled down on failure, excusing it at every opportunity, while pretending that this is how Falcon 9 succeeded, which is simply not true. You cannot have a radically different development process and call it the same because you're the same company.

>Why are we on a another engine redesign?

Just looking at it should tell you a lot about why:

https://www.metal-am.com/wp-content/uploads/sites/4/2024/08/...

It’s cheaper and faster to make in volume. It doesn’t require nearly as much shielding, because it’s less fragile, which saves a lot of weight. The engine itself is lighter. And on top of that, it develops more thrust, at higher fuel efficiency.

The net result is cheaper and lifts significantly more mass to space, which significantly drops the cost per kg to orbit.

It already worked, they’re making it much better, and getting it ready for a level of mass production that we’ve never seen anything close to in the space industry, even from SpaceX. They are much more ambitious than I think people who haven’t been watching them closely understand. The US grid is 1.4 TW of generation, they’re aiming to put up 1 TW of AI compute every year. Maybe they’ll stop well short of that, but their stated goal is insanely ambitious.

Reminds me of the adage - The successful have failed more times than the unsuccessful have tried.

v3 is the first version that was made with the intention of being used for actual payload delivery. The versions before were about testing and proof of concept.

It's worth remembering that, according to SpaceX's own filings, they've spent >$15 billion on the Starship program thus far with more to come. And SpaceX is burning cash still, particularly because Elon Musk bailed out his own bad decisions with Twitter and xAI with SpaceX stock, basically.

Flight 12 was a relative success. Some engines failed to light but that's an unintended good test. Rockets are typically designed such that they can have a certain number of engines fail and still achieve their mission.

At this point, the entire SpaceX project is a bet on telecommunications services, specifically direct-to-satellite handheld Internet. That's the only market that will recoup the program costs.

We don't have exact figures for the current true cost of a Falcon 9 launch factoring in reuse but many think it's somewhere betweenm $10 and $20 million. Well, SpaceX has spent 100 F9 launches on Starship so far and that's how you have to look at it. Say F9 is $20M and Starship once it starts launching Starlink is $10M that's 150-300+ launches just to break even.

You might be tempted to say there are other missions for Starship but there really aren't. Satellites aren't that bug, as evidences by there being ~1 Falcon Heavy launch per year (usually for the military and/or to geostationary orbit AFAICT). You can't economically put multiple payloads in one Starship because they all have different orbital parameters.

F9 is rated for human spaceflight. It's a long road for Starship to be certified for human spaceflight. SpaceX hasn't even begun to test in-orbit refuelling yet. Gases are weird in microgravity.

F9 is the cash cow funding all this and that too might go away if Blue Origin or one of the other wannabes ever gets a reusable launch platform to commercial operation.

There are big launches like interplanetary missions but those are few and far between.

It would be fascinating if what ends up dooming SpaceX is actually Twitter.

> At this point, the entire SpaceX project is a bet on telecommunications services, specifically direct-to-satellite handheld Internet. That's the only market that will recoup the program costs.

There's also a military angle here. I'll leave it as an exercise for the reader to look into Musk's history with Michael D. Griffin from the Reagan SDI/'Star Wars' program.

It’s an awkward comparison, but F9 can deliver a payload to orbit at a slightly lower price per kg than a Tomahawk missile can deliver it to a target. Starship would be MUCH cheaper if the economics works out the way that SpaceX would like it to.

Obviously a few hundred kg of payload in orbit are not equivalent to the same payload delivered directly to a target.

You don’t need very many kg delivered to target if it’s plutonium. The SDI program had the idea was that if you parked enough defensive weaponry in orbit then maybe mutually assured destruction wasn’t something you had to worry about. The only problem was that getting all that mass into orbit was prohibitively expensive.

Then the deputy director of the program met a young man named Elon Musk, and the rest is history.

I don’t think plutonium is the right comparison. Plutonium is expensive, and nuclear bombs are neither cheap nor particularly useful for doing things like attacking 10k different targets in some foreign country.

I’m imagining a launcher in a spacecraft that kicks out a bunch of payloads, one at a time, out the back, into orbits with perigee on or before the ground. (An LLM calculates the needed delta-V at under 200m/s, which is likely quite manageable with a small mass driver-style launcher or a very small rocket.) The payloads will lose a bunch of energy to the atmosphere, but all the remaining energy is kinetic energy delivered directly on target, assuming that you can inexpensively aim the thing at a target. Look up “Rods From God” on Wikipedia — you don’t even necessarily need any explosives.

So the question becomes: how economically can one build the guidance systems, avionics packages, and whatever heat shielding is needed to survive reentry?

(Cold War-era ICBMs with MIRV payloads are sort of in this category, but they treated launch vehicle as disposable, which means that the launch would be far more expensive but the reentry system could likely be a bit simpler as the payloads could be launched from a launch vehicle on a non-recoverable orbit. And it appears that Russia has attacked Ukraine with a MIRV-equipped missile with non-nuclear payloads, so there is precedent.)

Near real time fpv drones anywhere on the planet, free of jamming due to starlink is the real game changing capability. What would any military pay for that? Plenty I wager.

Wait a moment. How exactly is Starship supposed to usefully deliver a pile of FPV drones to a hostile area?

Starship itself is highly engineered to survive reentry and can even land anywhere that an appropriately flat surface is available. It will conveniently toast anything hanging out on the landing zone. But it is extremely far from stealthy, it seems likely to be extremely vulnerable both before and after landing (a hole in a cryogenic fuel tank = big boom), and I don’t think any of the design mission profiles involve Starship’s second stage, minus most of its fuel and minus the first stage, taking off again from Earth. And the value proposition isn’t there if you don’t get to reuse it.

A common cheap FPV drone dropped from orbit or from a suborbital ballistic trajectory is a small meteor, not a weapon.

ISTM it would be a better bet to equip an F-35 or a larger UAV to drop a pile of small drones.

Not sure how you made that leap because that would be absurd. The point was that fpv drones enabled with starlink render icbms, mervs, rods from god, etc unnecessary and uneconomical.

> At this point, the entire SpaceX project is a bet on telecommunications services, specifically direct-to-satellite handheld Internet. That's the only market that will recoup the program costs.

I seriously doubt that. Just for example, mining a single asteroid has the potential to flood the market for any number of metals. I don't pretend to know how expensive it would be to achieve that in practice; my point is that there are quite a few different ways to recoup program costs at some handwavey point in the future.

If there were infinite gold bars just sitting on the surface of the moon, it wouldn't be economical to go collect them and bring them back to Earth. No matter how expensive you think any metals are here on Earth, the cost of launching vehicles, rendezvousing with said metals and bringing them back to Earth makes it uneconomical.

An asteroid is much, much further than that but more important than distance is the delta-V required for change its orbit to reach an Earth orbit. So you not only need to get there, which, as discussed, requires in-orbit refuelling with Starship (or any vehicle), but you have to carry all the fuel you need for the orbital burn to bring it back. The rocket equation just kills this immediately.

You really hope you have to get incredibly lucky that an metallic asteroid is on a near-intercept course with Earth that is just shy or going into orbit. The odds for that are, well, astronomical.

That depends on how much a unit of delta-v costs. If you can do the whole mission for $100/kg, quite a few things become economical.

If we ever mine for gold in space, I suspect it would be more economical to leave the gold there than to bring it down to earth. Most gold ownership never results in delivery, people just own receipts for gold stored elsewhere.

I’ve always found it amusing that so much effort goes into extracting gold from the Earth when such a substantial amount of it goes right back underground into some vault.

Matt Levine has discussed similar theme for sub surface gold on earth[1]. NatGold is the company mentioned which tokenizes that.

Obviously pre-mined should trade lower than earth mined gold. Price should be something like Earth gold pricing- mining cost - cost to transport back to earth (or location based prcing, we are some time away from space based economics). Basically you will be trading rights for mining for a particular space mine.

[1] https://www.bloomberg.com/opinion/newsletters/2025-11-24/lea... (paywalled)

Are you saying "I don't know how expensive it would be to achieve that, but it won't prevent me from mentioning it as an economically viable option"?

> Say F9 is $20M and Starship once it starts launching Starlink is $10M that's 150-300+ launches just to break even.

Assuming they deliver the same payload, sure, but that’s very much not the plan.

Revenue from xai renting to anthropic this year alone will be more than starlink and launch revenue

Revenue from selling money at a discount isn't generally considered a good strategy.

xAI is burning through $1 billion a month [1]. With Anthropic as a customer, it's basically an argument that we're losing money on every transaction but we'll make it up in volume.

[1]: https://www.bloomberg.com/news/articles/2025-06-17/musk-s-xa...

> That's the only market that will recoup the program costs.

No. If it is just $15B I can think of dozens different usecases ranging from military applications(fast transportation, it is the cheapest ICBM) to asteroid deflection to moon mining to science applications to space datacenter.

Are you seriously thinking $15B is big? Artemis by comparison has spent $93B and has cost of $4B per launch.

Hexagon heat shield tiles show promise - I think.

12 crashes in 5 years to get 1/5th the capacity of a Saturn V. Good job Elon!

Starship capacity is 150 (reusable) to 250-300 (non-reusable).

Saturn V non-reusable capacity was like 140 metric tons or so.

Saturn V cost in today’s dollars roughly $1.5bn per launch. Current Falcon heavy launch from SpaceX is in the $20mm cost (to SX) range, for 64 metric tons. So that’s 1,500:40 = slightly less than 40x cost improvement over Saturn V.

Starship’s target costs would be $10mm per reusable launch or roughly another 4x cost improvement over Falcon Heavy.

Also, the SX team is huge, including some critical operators like Gwynne Shotwell, it’s not just Elon driving this 160:1 improvement in lift and cost capacity.

If Starship could meet the payload goals with their V2, they would have retired Falcon 9 already.

>Starship capacity is 150 (reusable) to 250-300 (non-reusable).

Those are aspirational numbers. V1 was supposed to have 50 tons and the actual payload was 0 tons. The confirmed V2 payload was 35 tons (optimistic). V3 has a payload closer to 50 tons than 150 tons.

many would think it was a failure lauch before going public

I am not sure results of single test launch matter much. The issues with SpaceX IPO aren't even related to rockets: it is the very low float, Musk making himself irremovable from CEO seat, and buying xAI at inflated valuation. If people are willing to buy such stock, even V3 blowing up at launch pad won't stop them.

What a time to be alive

Another flight with many explosions and a trivial payload. Trial and error, trial and error. At least these million monkeys have upgraded from typewriters to something more fun.

You don't see the progress between flights, do you?

Some of his employees have died in the meantime, but that's a price Elon is more than willing to pay

Had to check this one and you're indeed right: one death in 2014[0], and one death earlier about a week ago (May 2026)[1] (as well as 600+ workplace injuries [0]).

[0]: https://www.reuters.com/investigates/special-report/spacex-m...

[1]: https://techcrunch.com/2026/05/18/osha-probing-worker-death-...

Scaled Composites has had a few deaths, too.

Less recent that I thought: https://spacenews.com/explosion-mojave-air-and-space-port-ki...

How many deaths has Facebook or Google had in that time?

Over 100 died building the hoover dam. Over 5000 building the Panama canal, and over 30,000 if you count the failed attempt.

30,000 people trip and die in their homes per year.

70,000 Americans die per year from medication error or accidental overdose.

So when's that self-growing Moon base happening?

"potentially achieve that in less than 10 years" in Musk time which is 20 years to never in real time?

Reply on news.ycombinator.com