US watchdog grounds SpaceX Starship after that explosion America's Federal Aviation Administration has grounded SpaceX's Starship to conduct a safety investigation after the heavy-lift rocket destroyed a chunk of the launch pad, malfunctioned, and had to be blown up mid-test. SpaceX was unable to get Starship, said to be the world's most powerful rocket, into orbit during that …
... but it's suggested that SpaceX desperately needs to get Starlink V2 up into space to become operationally solvent, and that's not going to happen until they have Starship working - which in turn depends on a functioning launch site and reliable engines.
They've destroyed their launch site, and demonstrated that the engines are not reliable at this late stage of development.
That makes it fairly unlikely we'll see Starlink V2 going up any time soon (not this year), which leaves SpaceX surviving purely off government subsidies. How long they can do that is an interesting question.
In the mean time, the promise of lower mass to orbit costs do not seem to have been realised.
They destroyed the launch pad, not the launch site. Other than having some of ts support pillars exposed the launch tower survived relatively unscathed. I am very curious to see if the giant steel water cooler that was not ready for this test will work as expected. It, along with other changes should end the derbies problem. With 31 rocket engines I fully expected some engine failures on the first try at running all of them at close to full power on a test launch vehicle. There were more non-working engines than I would have liked to see but please note that most of them did work. I also wonder if any of them were damaged by the copious amounts of derbies generated when they destroyed the launch pad. I expect this problem to be fixed although it may take a few more tries along with a major upgrade to the launch pad. I don't believe that there was any expectation to launch StarLink V2 satellites using the Starship this year. As explained many times by Elon and SpaceX this will be a work in progress for at least the rest of this year and probably some of next year too.
The promise of lower mass to orbit obviously depends on a fully operational Starship and as most know that is still more than a few test launches away. Given SpaceX's record so far I hope/expect that it will become fully operational sometime (late?) next year.
"the copious amounts of derbies generated when they destroyed the launch pad"
That's a good trick if they can repeat it ... and if they can manage shoes (oxfords, not brogues of course) and suits, it could put some Savile Row businesses in jeopardy.
(@Garry Stewart - as always, when I make an appalling joke, I upvoted you)
This is very dubious stuff..
If you have steel reinforced concrete damaged to the point that the reinforcement is exposed, it is structurally compromised. Not something you can just put a patch over.
From the video of the car being smashed by debris, most of the launch site has been pelted with concrete and other projectiles. Every single building and facility will have to be inspected and repaired.
As the launch demonstrated, failures amongst the Raptors were enough to ultimately doom the flight. Statistically, the number of failed engines seems consistent with their reliability reports, and that number of failures is too high to trust the platform as a reliable launch vehicle. This is a serious problem for any craft, even more so for a 'reusable' vehicle. It seems madness to test engine reliability en-masse when they already knew that individually the failure rate was too high. Yet, despite a decade of engine development they've not got to the required level. It would be lovely to think that they'll just tweak the design and it'll magically work, but there's no evidence that this is going to happen in the one month Elon has promised.
And the point of all this is that SpaceX are developing on Elon's timescales, without infinite financial resources. They have some really smart people working for them, but one of the laws of physics is that money can run out. After this test flight, which may have set back their plans by a significant time, the risk of financial rather than technical failure is much higher.
> failures amongst the Raptors were enough to ultimately doom the flight
Citation needed.
The flight was terminated because the stage separation didn't happen. I've heard no suggestion that stage separation was in any way related to first stage engine failures.
The bigger problem seemed to be that the booster proceeded with its flip and boost back manoeuvre, despite the separation not working. If it had waited for confirmation, the outcome may have been different.
the rocket was purposefully terminated, when stage separation was supposed to happen.
I think if you watch the video back, you'll see the stack performs more than a double backflip before abort (it's quite impressive that it didn't break up of its own accord). If they'd seen the mission was doomed and said "scrub when we go for separation, they wouldn't have pulled a double backflip first. Or they would have gone through stage separation (to validate the mechanism) and then aborted each component separately (but that also makes no sense, because they could still have done the simulated landing of the first stage, even if Starship couldn't make orbit). The flight went well past the planned separation before they hit the big red button.
Now, it's entirely possible that the failure of engines and the course correction caused buckling or torquing that locked up the separation mechanism. But there's no specific evidence at this point that the engine failures were the root cause.
Unless the remote kill system failed in multiple ways it wouldn't have any discernible delay before it broke the stack up. Not a thing the powers that be have any sense of humor about either, some of the other smaller upstarts have been slapped down and grounded for issues in those essential systems.
The other question is that since SpaceX is operating it's own range in this case, are they doing anything different than the range safety operator at one of the (generally military controlled) government launch complexes. The live stream didn't seem contain a clear call to abort, so maybe they were still considered to be in acceptable safety parameters. By that point it was over water and still at considerable altitude, so they have chosen to wait and collect as much data as possible.
Worst case is that they tried to kill it and didn't succeed, but I also suspect that the amateur rocket watchers would have picked up the radio blip of the command and telemetry data. They have done the same for other launches in the past, and I'd expect screaming by now if there was a smoking gun there.
No need to see the telemetry / command data. Scott Manley has pointed out in one of his shorts on the firing that there was 50 second delay between the termination system firing (evidenced by new leaks opening up in exactly the right places - clearly visible in the videos) and the thing eventually breaking up. That there was a substantial delay between firing and effect is not in question.
If it's going to stay intact for 50 seconds, engines still firing, it can go an awfully long way in that time. Put into the context of if having veered off course having left the pad and heading towards Port Isabel, you could expect it to arrive intact, probably under power, and still full of fuel. It'd be like delivering a small nuclear bomb into the town. Arguably, they've had a lucky escape. Certainly, if I lived in that town, and the FAA said "yeah, 50 seconds is just fine", I'd be telling them it wasn't.
The idea of flight termination is that the vehicle is disintegrated ASAP, so restrict the distance it can travel following termination. Previously, that's not been a problem; rockets are traditionally built from the flimsiest possible aluminium by the cheapest contractor. Punch a hole anywhere, and it'll fall to pieces anyway. But, if you then go and build a rocket out of 4mm thick stainless steel, it evidently does not fall to pieces immediately (it didn't even manage to set the leaking fuel on fire...).
This could be really, really bad news for SpaceX; if they have to replace the point charges they did use for the FTS with linear cutting charges wrapped around it like a fishing net to be certain of cutting it up on demand, this makes reentry very difficult. You'd not want to reenter the atmosphere with explosive cutting charges still on the outside. They'd have to be got rid of, somehow, which sounds messy. If it turns out that an effective FTS is not going to play nicely with a heavy stainless steel rocket doing atmospheric reentry, then that could invalidate the entire design ethos, which could leave them with no program.
"You'd not want to reenter the atmosphere with explosive cutting charges still on the outside."
That's a very good observation. I never had to work with explosive RST gear and the vast majority of rockets aren't expected to come back. A first stage isn't going to orbit, but once it's been landed, safety will be a big issue.
I don't know how fast the first stage goes, or whether atmospheric friction heating ever gets high enough for external explosives to be "dodgy". Certainly things like the external fuel tank for the Shuttle got fast and high enough that it would disintegrate on falling back to earth. I don't think the StarShip first stage gets that high or that fast, so maybe it could safely carry external explosives for the whole trip. But, to my knowledge, getting explosives "too hot" is not a good idea, stability wise.
When we look at other programmes, the bit that comes back is generally not the bit that's dangerous at launch. Gemini, Apollo, Soyuz style rockets; it's just the small mostly fuel-less capsule on the top that comes back. The Shuttle had its fuel tank on the outside, and never took it into orbit. All previous manned flight has had the range safety package on a component that did not need to go into space, or did not need to survive re-entry.
It'll be interesting to see if the FAA decides that range safety for StarShip does need to be beefed up. On the face of it, it looks like it is. Solving it could be a tricky problem. When one thinks about the problem of re-usable space in the context of being able to blow it up if necessary, it starts to make the Shuttle's architecture look fairly good.
So Near to Success, So Near to Ruining it All
StarShip feels like it's a good idea, but only if it's become reliable enough that (like a Jumbo Jet or any other airliner) one has high confidence in how well it can be operated. Get it to be as reliable as, say, a 747 and it'd not need a flight termination system.
Safe designs can be dangerously operated. Dangerous designs can be safely operated. All mechanised transport is of the latter. SpaceX have shown with this test that they've got a dangerous design (nothing inherently wrong in that - all things that move fast can be dangerous), but they're not very good at operating it (which is why the FAA has grounded them for now). Fixing that could be a whole lot more painful than any technical design changes.
For example, a lot of commentards (I include myself in that description and set) are saying "flame trench + water deluge" is the way to go. Everyone knows that they'd have a much greater chance of getting cleanly into they sky, with that launch pad set up. So, why not do the thing mostly likely to be successful in 1) being permitted, 2) actually working? Why throw yet another novel, untested, unverifiable idea (the water cooled steel plate) in the pathway to success, proposing to try it out for the very first time underneath the worlds biggest rocket? That feels strongly like trying to persist with a way of operating which the FAA clearly isn't happy with. SpaceX clearly didn't understand their current launch pad's resiliency, what makes them think they'll know any better with a steel plate? Why should the FAA believe them?
They've come this far; they could ruin it all. On the whole, it looks like if they can get it cleanly launched, it would at least reach orbit. If they can reach orbit, they can use it, dispose of it safely, fit it with the mother of all FTS's to make sure that it really does terminate if required. Once it's reliably safe with more relaxed FTS requirements, they can take some time to nail re-usability (remember, re-usability was an add-on to Falcon 9, added once the rocket itself was more or less working).
They don't need to innovate in launch pad design. No one buys launch pads.
Soil Mechanics
When they light up this rocket, they're creating a very intense, very localised earthquake. The way soft sandy waterlogged soil behaves in an earthquake is that it liquefies, flows, etc, starts oozing out of the ground all over the place. So, what they've probably got in the way of soil underneath the launch pad at Boca Chica at launch time is soft jelly moving all over the place. No wonder their concrete failed - it likely had nothing of value underneath it.
So, if they're assuming that their concrete pad, steel plate is going to get any support from underneath, that's likely not the case. I think to do what they're attempting, they'd need a very large stiff plate of something to spread the load out over a large area, and (more importantly) to have some way of isolating the soil underneath from the vibrations. That can be done with, say a thick bed of gravel (as is done with nuclear bunkers). Trouble is, the gravel bed would compact each launch, and the whole thing may be too heavy and sink into the sand anyway
What you likely don't want is deep-sunk stiff foundation piles conducting the vibrations downwards and acting like tuning forks stirring up the soft, sandy soil. The way such piles are sunk in the first place is using shock and/or vibration. That ought to be a clue as to how suited to extremely strong vibration environments such foundations are...
If they're going to persist with Boca Chica and no flame trench, the best thing they can do right now is hire an experienced Japanese construction engineer. They know about structures, earthquakes, and rubbish soil. If one of those folk says, "Can't be done", they'd best give up and go elsewhere where they can have a flame trench and water deluge.
Of course, the beauty of the water deluge is that, not only does it protect the rocket, it protects the structures underneath it. In absorbing shock and vibration, the foundations are protected. Better to hammer away at disposable fluid, than have something that you'd like to remain solid fluidised.
"The live stream didn't seem contain a clear call to abort, so maybe they were still considered to be in acceptable safety parameters."
There is only a certain amount of course correction that can be done before it's uncorrectable. The small landers I worked with were limited to about 15deg of command authority (gimbal movement) so if they tipped past that it would be like that point where you are trying to balance a pencil on your finger and you know you can't get it back upright. This flight was way past that and it's usually right around that point where any automated range safety system is set to detonate. The only exception I see would be a short hold to make sure the debris won't be raining down on something such as an oil rig. I hope it won't take years to see a report from the FAA or other agency with details and mandated corrections.
"I'm not entirely convinced that the RTS wasn't fired substantially before the eventual explosion - punching a couple of holes in the tanks, but the tanks simply held together far better than might have been anticipated."
Some better independent video appears to show that the FTS did make holes in the tanks, but it took an explosion at the aft end to break the booster up and a second explosion destroyed Starship just afterwards. The upshot is that the FTS wasn't the cause of the explosion.
recent reports said that it was the onboard(automatic) fight termination system that killed it. So it isn't at all clear if a remote abort had been tried based on the information I have seen.
I doubt that the other posters concerns that the termination system failed to break up the stack after "punching holes in it" hold much water, but the steel assembly certainly rode out those twisties better than I expected, and that's maybe not great for range safety. A failed stack may need to be "force quit" to prevent a ground level hazard as it can apparently survive for quite a while operating way past intended lateral loads. That really means that the circular error probable from a failed launch includes a couple decent sized towns.
I doubt that my renters insurance includes coverage for Starship damage.
The initiation of the abort was visible in the video - Scott Manley highlights it in one of his. So, the time between initiation is very publicly visible - no need to rummage through telemetry. What I don't know is if there is a second abort system on board; SM didn't seem to say so. None the less the idea of one of them having no useful immediate effect destroys the safety case.
It seems clear that there's far too many unexpected outcomes from this flight, if the FAA has called a halt. Having done so, there can only be a full up re-evaluation of the performance of all the safety-related things. For example, having lost engines on the pad, it now must be considered possible for this thing to struggle off the pad, tilt northwards, fly 5 miles and fall to earth; bye bye, Port Isabel.
Another bad thing about the flight; clearly, a lot of damage was caused on the launch pad. That likely invalidates any telemetry they have on things failing. If you're going to smash the underside of a rocket with large lumps of concrete, of course things are going to break. We know they lost engines at start, and soon after, but who can say what damage was inflicted. One could argue that with that much debris flying around and it still getting off the pad is some sort of triumph, but they'd far prefer to know the nature of the failures.
Initial broadcast marked this out as a failure of the second stage to separate after a "flip" maneuver, so not due to the lost engines(which appear to have cut our early in the launch) and it had maintained control and stability at launch and long enough to get to the initial stage separation.
That part, the failed second stage separation, seems to fit the facts and the timeline of the launch. The more current reports were that the onboard automatic flight termination system blew it after it did three loops and was totally out of control, which still fits the basic timeline. We don't know what downed the engines, so we don't know if was a fuel supply problem(engine works but no gasses=no fire) a manufacturing or designed in problem in the engine itself, or if they we simply the victims of debris from the launch.
Clearly the launch pad needs to be redesigned, with major safety upgrades. Exactly what the right list of "fixes" are will only play out after real engineers do real failure analysis. That won't stop the rest of us from floating a million fan-theories, as we are all rocket experts. (I only claim to come from a family of rocket men, grandad worked on Apollo, as did MANY other peoples grandparents.). But clearly it shouldn't be shedding chunks of concrete and launching them miles away.
Another interesting question is if the crumbling debris damaged the engines that failed leading to an excess of improperly combusted fuel immediately after ignition. That fuel isn't going to behave the same if it isn't being burned in a rocket cone. It seems like the stack took a long time to actually get moving, but I'm not sure what the math of the new stack has to say about slow this take off was, or how fast it should have been. The peanut gallery will beat me to loading it into Kerbal Space Program, so I leave that as an exercise for the reader. A chain reaction where a damaged launch platform lead to damaged engines that than led to more damage to the platform and losing more engines seems plausible at least.
But the team certainly identified quite a few new problems to address. The government will have it's own opinions on which of those need to be addressed before a subsequent launch attempt.
The one I am most interested in is still unanswered. The ground crew should have probably hit the remote termination button from the control room when the thing started it's second loop/tumble. Unless that was a jackass stunt and they planned to have it fly in circles then blow up, somebody needs to say why the dude in the control room with exactly one job and one button to hit didn't successfully terminate the flight before the onboard termination system did. There may be perfectly good reasons for that, that we deserve to know in the interests of transparency.
If they had a failure in that system, the range needs to be shut down until they can demonstrate they have regained remote flight control and termination that is as good as Vandenburg or Cape Canaveral. People live across the bay from there, and there wasn't a rocket complex when those people moved in.
Regardless, a two month launch window seems optimistic, and NASA and the FAA will be doing them a favor by tapping the brakes and giving them time to implement engineering fixes for more of these problems.
"And do remember that SpaceX have experience with running 27 engines at a time already..."
Different engines in a different configuration. Some of those Merlins do go out on launches. Elon is really pushing the Raptors for the most power ever in an engine their size. He may be backed into a corner where they know (or at least some of the fired managers on the Raptor program know) that there is no time to do any major redesign or start again. I learned that lesson when I had a manufacturing company. I had to explain it to an investor when he squawked. I could either try to make a product hit the required specs that were, in theory, possible, but on the edge or I could make a slightly larger version that had a much better chance. The larger version had a full set of drawings in two days and parts were being prototyped immediately. Maybe I could have made the smaller one work, but it could have taken a month and been a real problem in production. I knew to cut my losses. It's a hard thing to do, but when you find yourself up a dead end alley, turn around rather than trying to scale the buildings to get to the other side. I have to wonder if SpaceX has any other engine programs in process.
Not long ago I watched a show about the original space race and this was how the USSR did things back in the day.
They would engineer to a point test, note failures and correct the US at the time would put far more effort into the engineering component,
Not sure which way is better, but they did get to space first...
The full up, whole problem engineer everything approach is better. It doesn't even have to be slow, if you've got an experienced team. The trouble is that such teams don't exist very often, and get put together only now and then, and you have to keep them busy to keep the team together.
For example, the original Skunk Works - wartime priorities resulted in the best of the best engineers in Lockheed being lumped into a single team that persisted up to the 1980s and built such iconic aircraft as U2, A12/SR71, and the F117. U2 was done fairly quickly, the A12 wasn't too long especially considering how much stuff they had to work out. The F117 - kinda the last hurrah of the original team - took very little time and cost comparatively little money; this was most definitely not a try-it-and-see project.
NASA built a team and got man on the moon inside the decade, by knowing what it'd take to get to a point of launching a Saturn V and have it work first time.
Another - Airbus, created out of the Concorde program between Britain and France. Ok, Concorde itself wasn't a big commercial success, but the "we're working together" aspect was; that lead to the formation of Airbus, putting the best engineers from across all of Europe together. The A300 is a distant memory for a lot of people, but the successors have been wildly successful. Airbus got pretty good at maintaining this team, they (in fairly rapid order) did A320, A330/340, A380 (some development screw ups), and then A350 (perfect development) and now the A320 / 330neos (also have gone well).
SLS has had a difficult, politically influenced birth, but it has to be said that they successfully didn't blow it up several times before finally getting a near-perfect first flight of it. That team, if kept together and left to get on with it, could probably now do anything.
Unthinking iteration is all very well, but if it keeps going wrong people get worried and bored. Right now there's a bunch of SpaceX engineers who must surely know that the program has taken a distinct turn for the worse, and wondering if it's worth hanging around. The will, but it won't take too many failures (especially for "predictable" reasons like too flimsy a pad idea) for people to start looking elsewhere. Engineers are very good at knowing when projects are being managed to the detriment of achieving success..
"Engineers are very good at knowing when projects are being managed to the detriment of achieving success.."
There can also be the appearance of poor management and the meddling boss with no credentials that is also a huge problem. I was working with a software person across the country on a brushless DC motor project and he was afraid of 24vdc and wanted me to build him a test rig to run on 12vdc. Since he was the main investor in the company, I couldn't come straight out and tell him to grow a pair. I wanted to move the avionics to include a 48v bus to make it easier to control bigger actuators as our vehicles grew, but he shot me down. Everybody else on the engineering team thought it was long past time since the lower voltage was adding mass to the landers.
"the US at the time would put far more effort into the engineering component"
There is a thing called "Analysis paralysis" where too much time is spent doing the engineering and its become worse with all of the modern simulation software. It takes experienced engineering management to break off the cubicle work and get the drawings to the machine shop at the optimum time. Once you are well inside the error bars considering normal material tolerances and operation conditions, there's no point in running any more sims.
Yes, it is a problem. Generally I've seen it most commonly exhibited by what I call the "MatLabists", who think they can model anything.
The hilarious part is that such people often seem to think they can turn their model into a working system, with direct synthesis to an FPGA, or similar. I've never, ever seen that succeed in anything but extremely trivial ways. The worst part of it is that MatLab is a very bad way of implementing an entire system, so the MatLabists don't ever get to explore the entire system, and therefore tend to have a skewed view of how much of the system is composed of their part (generally, not much) and how much of it is other stuff (databases, networks, people, etc). Given that model to implementation synthesis only ever works on a very limited range of hardware, and that there's many, many other factors that go into hardware selection than "does it play nicely with MatLab?", they often totally fail to deliver anything at all useful. And they really, really do not like being told that their part is perhaps only 5% of the entire thing.
I generally budget for starting again from scratch.
Though I will say that I have seen some very good results using kit from Annapolis Microsystems, largely because they build pretty good hardware and their own tooling. They make it easy for their kit to be part of a functioning system.
The problem these days lies in organisations that don't have experienced engineers leading them. Such management gets all sorts of messages, and the Matlabists generally promise generally offer cheap deep insights, certainty, guarantees, where as the more realistic engineers are (in effect) saying "we have to spend some money". If you're not an engineer and know no better, and one "engineer" is saying "you don't need to spend money yet" whilst another is asking for a budget, who are you going to go with?
The best advice I can give to management is to look at track record very, very carefully. If someone has a track record of delivery and can point to successful in-service systems with their name on it, listen to them, give them the cheque book. If someone is putting up a lot of pretty pictures but can't actually show you some physical hardware (or software) that's been in use for a lengthy period of time, ignore them (or fire them).
> Agile strikes again
The frustrating thing is that I feel this is giving all engineers a bad name (Musk isn't what I'd call an engineer, not by a long shot).
We're supposed to be people who can design things, get it mostly right, have a reasonably good idea as to whether something will work or not, and have a good handle on the risks being run in a project. Ok, so there's a lot that is "unprecedented" about Starship, but what most definitely is not unprecedented is that flame trenches / water deluge systems work. They've launched this thing, and are now suffering consequences they themselves predicted (Musk's tweet from 3 years back). This is not "smart" in anyway whatsoever.
This is a problem for engineers outside of SpaceX too. Say SpaceX had a really bad accident - e.g. people in Port Isabel got hurt. That would be scandalous, and there'd be a lot of "told you so" going on. That would dampen down any other similar endeavour, as people (quite rightly) would be questioning the abilities of engineers and the government regulator to assure people are not hurt. Arguably, we're lucky that launch didn't seem to hurt anybody, even if it did make a mess, but one feels it could have been very much worse.
The vertical pillars look undamaged. The construction industry has plenty of experience patching concrete. Even the buried horizontal section that was stripped to the melted and bent rebars is repairable. The salt water incursion will be a pest when they fill the crater but even that is not a major problem.
You are quite right about the holy buildings on and around the launch platform. That probably goes double for their contents. I am expecting damage to the winch motor that lifts the chopsticks. Perhaps SpaceX will tell us about that damage and perhaps not. The (small) good news is that some of the covers that were blown away would have had to have been removed anyway to install the deluge.
You are quite right about engine failures dooming the flight, but the causes of the failures matter. The outer ring of engines require high pressure gas from the launch table to spin up so two of the failures could be related to the ground support equipment. Strangely enough, when you use rocket engines to dig a crater some of the debris goes straight up even against the exhaust. Some engine failures could be from FOD. The booster still had plenty of methane when it ran out of LOX. That would require a massive LOX leak into the engine section. That too could cause multiple failures - such as the explosion that took out the hydraulics for engine gimbling. Putting in a flame diverter will fix multiple issues.
You are right that improving engine reliability is not going to happen in the next month. It has already happened. These were obsolete engines. The next generation - already mounted on the next booster - use electric gimbling. The hydraulics have been completely removed.
Simulations showed that the improved concrete would survive one launch. This proved to be spectacularly wrong but go back to the planning stage and re-plan the mission with this false assumption. You could scrap 24/7 (no space in the rocket garden) and wait a month or two for the new deluge and flame diverter or you could launch 24/7. Perhaps some of the engines would not have lit on the launch table but you could abort if too many failed. Demonstrably the rocket cleared the tower so if the concrete had held there would not have been significant damage to the GSE. Perhaps the rocket would have got to the required speed and altitude for stage separation and given useful data for that too. It would have been madness not to launch. As it is, SpaceX have discovered that their model for rocket blasting concrete (or Luna regolith?) is wrong and their model for debris from flight termination is wrong. Both are valuable discoveries that come with real data for fixing the models.
The point here is that hardware rich development is massively cheaper than the alternatives. Look at the costs for SLS R&D, GSE and operations. Following that example would be insane enough for a career in politics. The other big example is Blue Origin. That hardware scarce project is still plodding on only because of a giant cash fire hose from Jeff Bezos. His pockets are not as deep as the US tax payer but few other people on Earth can afford to build rockets at his pace.
We may well get another Elon tweet demanding engineers work through thanksgiving to prevent SpaceX going bankrupt. I would file it next to "funding secured".
I had something different in mind. (Explanation for those whose beards have not yet gone completely grey.)
LOX is pretty terrifying stuff, especially when there is a large amount of it around. If you are correct that would have all sorts of possible ways to make the damage to the launch stand worse, and would have helped make the exhaust plume act like a cutting torch where it contacted concrete and steel.
We may have just seen the world record for the largest thermal lance ever (unintentionally built).
>The point here is that hardware rich development is massively cheaper than the alternatives.
It is, provided that one has not made a serious error of judgement. The poor performance of the flight termination system (50 sec between trigger and break-up) alone could write-off the whole concept of a heavy stainless steel built-like-a-tank StarShip.
Superheavy, which doesn't leave the atmosphere, can probably carry linear cutting charges on the outside to ensure instant and good fragmentation without those explosives being exposed to high temperatures that'd cause them to go off unintentionally. Same can't be said for StarShip. It's going to have to re-enter the atmosphere.
If they need to be more sure of the FTS cutting it to pieces when commanded, I'm not sure how they then lose those charges before reentry so that it doesn't get cut to pieces by mistake. It looks like they've tried to get away with two boxes on the outside of the rocket, which I'm sure are intended to be disposed off just before reaching orbit. Easy to drop off a box. Less easy to drop off linear charges attached all around your rocket's body.
Nothing ever launched before has had a need to bring a FTS safely back to ground. The Shuttle managed it by dropping off its FTS in the upper atmosphere before reaching orbit; it was in the external fuel tank and/or SRBs.
A giant water cooled steel patch. Last I heard steel is much more resistant to heat damage than concrete. Even concrete especially designed to resist heat damage. And i seriously doubt even if it did fail that it would not generate anywhere near the derbies the concrete did. There are other serious problems with the amount of dust generated and widely dispersed throughout the environmentally sensitive area. If most of that was concrete then the giant steel water cooled "patch" should solve most of that problem and kill two (hopefully not endangered) birds with one stone. If a large amount of the dust was from the surrounding area other fixes will be needed. Perhaps a plume tunnel that Elon for some reason does not want to use.
Every single building will have to be inspected and some of them will have to be repaired.
This was the first fully operational (well mostly) TEST vehicle launch! JFC, it was expected to have problems, some serious and some not so serious..The Falcon Nine, a much "simpler" rocket took several tries to get it right and now it is one of the most reliable launch vehicles available. And to top it off the boosters are reusable. At least two boosters have been reused 15 times with several more in double digits. So SpaceX does know how to do reliability. Sometimes it took tweaks, and sometimes it took serious re-engineering but they managed to get it right eventually. As long as the money holds out, and I have no idea if that is really a problem or not although I am sure that infinite does not accurately describe the amount needed, I fully expect them to be able to do the same with the Starship. Given the complexity, it will certainly take longer to do this time.
As for the rockets flips and flops described below, as explained the rocket was supposed to do a single flip to position it for a landing back burn while at the same time imparting the momentum needed to separate the second stage from the booster. A complicated maneuver that I have doubts about, we will see. Since this separation did not occur it is not surprising that things got dodgy after that. I suspect that the added mass and momentum of the second stage caused the loss of control when the flight computers could not properly compensate for it. At that point it was time to destroy the rocket which the on board computer or ground flight control did.
That also failed to reach the stretch goal of hitting orbit. Which it probably wasn't going to do with that many failed engines (just from an altitude/velocity standpoint) even if it had managed separation.
This isn't a great failure mode for commercial heavy lift launches, as a failed orbital insertion will still screw most payloads completely. It also hasn't ever been demonstrated that Starship can successfully land with a payload attached, but it may be possible in some cases, which would potentially hedge some of the risk of partial launch failure in the first stage without sacrificing the space in the payload for a separate bailout system.
Even if it did land the payload safely, the whole thing would probably have to be recertified before it could be relaunched.
So I expect that they will need to get a bunch more nines back before anyone else decides to risk their payloads on the new stack.
"They destroyed the launch pad, not the launch site. Other than having some of ts support pillars exposed the launch tower survived relatively unscathed. "
The horizontal braces that are just rebar now used to be below ground level. Without hard data, there's no way an outsider can claim that the support stand and tower are unscathed. One must also consider that in regular use, the rocket is supposed to come back and be caught by the tower, be lowered to the stand, refueled and flown again up to 5x per day. With plenty of tanks and other GSE damaged, there may be much more repair work than SpaceX is going to own up to.
"That makes it fairly unlikely we'll see Starlink V2 going up any time soon"
They are launching some on the Falcon, but they are bigger and heavier so not as many can be put on each rocket which adds a bunch to the cost of the system. I suspect it's more about volume or they'd use the Falcon Heavy but perhaps it's just as possible to use 2 Falcon 9's rather than the F9H. I've seen some analysis that makes it appear that Starlink isn't viable whether the satellites are launched on an F9 or Starship.
Starship still needs to demonstrate that it can open and securely close any doors that satellites are dispensed from so it can return back to a landing site. And, that's after they can show that they can deploy the satellites from Starship since that's something that haven't been testing.
I know that there are people in the sticks with few internet options (Hughesnet and Viasat in the US), but are there enough to make another service a viable business. Going further, are there enough people that can pay enough who need the only advantage Starlink has, low latency.
The town I live in is pretty small and we now have cable, fiber and 5G home service. I could even get on a local businesses wireless internet system, but that's not really a public option and only available since I know him. There's no point for me to pay a premium for Starlink and invest a bunch of money in their hardware that comes with no warranty.
Does Starlink work, yes. Are there enough people in the boonies that want/need it? Probably not. The vast number of people in the world can't afford it or a computer to access the internet.
Yes, you thought correctly. Computer modelling predicted that the upgraded concrete could survive one launch, so it was believed at the time that the launch pad was sufficiently complete for a test launch.
The limiting factor for Starship development is the 5 launches / year license at Boca Chica. SpaceX can already build new complete rockets at twice that pace. The choices were launch it or scrap it because there is a limited amount of space in the rocket garden. Using all this year's launch slots while waiting for the flame diverter would have been close. With hindsight, using all the launch slots will now be difficult because of the 2 (probably x2) months delay for repairs to the ground support equipment.
Many people really do not get the differences between SLS and Starship development. SLS costs $4B per launch (excluding R&D and ground support equipment capital costs) and can launch ½ per year. It is made from shuttle parts that were well understood (to be difficult to work with) and modified sufficiently to get R&D to over $27B (plus a similar amount for GSE). If Artemis I had failed, that sweet cost plus contract would have died. The cost (not price) of a fully expendable Starship launch is about $100M. The price (not cost plus) of two Lunar Starship missions is $3B. By the time SLS launches Artemis II with people on board, Starship will have failed many times, with each failure generating a huge amount of data showing what not to do next time.
If you really want to see how ignorant people on the internet can be, imagine how hard you could laugh at them if they proposed a flame trench. Cannot dig one because the launch site is at sea level. Cannot build one up because the area they can build on at the launch site is tiny and surrounded by nature reserve. What is required is a tall launch table (already built and easily repaired) to get the engines far from the ground. They will also need a catch pond for all the fresh water used by a flame diverter and deluge (already dug).
It is almost as if decisions were being made by a bunch of experienced rocket scientists instead of random voices from the internet.
even with what he has done to Twitter?
{shakes head in amazement}
I recently heard him described as 'Trump minor'. Lots of 'I know more about this than anyone ever' but little substance behind the rhetoric. Then trying to look cool on stage with that hat and shades.
While he ran just Tesla he was sorta ok. Since his empire has expanded and to some extent got out of control, he seems to be lacking what it takes to sort any of the major problems out.
Whatever happened to that tunneling thingy that was supposed to revolutionise travel?
Whatever happened to the Tesla Roadster 2?
Tesla might be annoyed they just blew up the only Roadster 2 concept but as you can see the chutes didn't deploy. But accidents do happen, and think of the increase in the value of the only one in space currently. (clearly the most collectable car ever).
I suspect Elon may try to pick the old gal up when she spins by next.
"and think of the increase in the value of the only one in space currently. (clearly the most collectable car ever)."
Considering Muskolinis love of crypto, I'm surprised he's not auctioned it off as an NFT yet :-)
Musk has risked money. Hardware rich development causes RUDs and Musk has not flinched when they happened. One of the popular complaints that about SpaceX is that is it tax payer funded. The truth is more complicated. There was and still is NASA and DoD funding for small launch start-ups (SpaceX Falcon 1, RocketLab Electron, Firefly Alpha, ...). SpaceX's big break was Commercial Resupply Services for NASA. NASA picked two suppliers: Rocketplane Kistler and SpaceX. Kistler failed from inability to get investment and was replaced by Orbital ATK / Cygnus. (Sierra Nevada Corporation missed out on the big contracts but have still made impressive progress on Dream Chaser with their limited funds). Cygnus (now operated by Northrop Grumman Space Systems) costs more than Cargo Dragon, delivers less and cannot return cargo to Earth intact. The Cygnus spacecraft is still an impressive achievement with price issues related to its succession of launch vehicles forced by relying on Russian engines.
SpaceX abandoned Falcon 1 and went straight to Falcon 9 to get something big enough to launch Dragon. One happy accident of that was that Falcon 9 launched with a large (for the time) number of small (at that time) engines which made vertical landing a possibility later. It also gave them a rocket to compete for other DoD and NASA contracts. Back then, the US tax payers were being royally screwed by the monopoly supplier ULA, which operated Atlas V ($110M-$150M) and Delta IV ($450M). Compare that to Falcon 9 ($60M) and Falcon Heavy ($90M) respectively. Perhaps one day people will be able to explain to me why they are so angry about SpaceX getting government funding when the alternatives were so much more expensive. (ULA also got about $1B/year to maintain ground support equipment on top of the launch prices).
The next really big SpaceX contract was Commercial Crew. SpaceX bid less than Boeing. Congress helped by delaying CC funding and reassigning some of it to SLS. Because of funding issues, NASA's work on CC got delayed which cost SpaceX and Boeing time and money. SpaceX took the hit on CC1 and increased the cost on their bid for CC2 to make up much of the deficit. Boeing whined and got extra money for CC1. Crew Dragon is still cheaper and has already flown 6 full NASA missions (and is getting additional income from space tourists). Boeing's first crewed test mission might be flying real soon now (to be fair some of the final delays come from lack of an available docking port on the ISS). Without SpaceX, NASA would still by buying rides to the ISS from Roscosmos.
The transition from Atlas V (Russian engines) to Vulcan (Blue Origin engines) is not going as fast as was hoped. The DoD is keeping options open by ensuring they can switch to SpaceX if required. Some of their really expensive satellites require vertical integration. SpaceX plans to pass the cost of that service 100% to the DoD (no commercial customers require it). Evidently this is still cheaper than redesigning those DoD satellites to work with horizontal integration.
SpaceX is charging a fixed $3B for an uncrewed test and a crewed landing on the Moon. Compare that to Blue Origin's fixed cost bid: $6B. Starship cargo to the Moon is a difficult guess because of the frequent design improvements. 100,000kg minus crew habitat mass was a sensible early guess but it is probably more now. Blue Moon: 4500kg. The up mass cargo for Starship is probably 50,000kg but is currently irrelevant because Orion cargo return mass is about 0kg (about the same of Blue Moon). The RFP for HLS did not include the uncrewed test mission returning from the Moon. This is because only SpaceX would have been able to bid. Blue Moon must be partially dismantled by astronauts before the ascent vehicle can return. Blue Moon is a single use vehicle. Starship HLS could become a reusable system. (Blue Origin may be able to make a better bid if they make good progress on New Glenn.)
SpaceX has achieved a substantial lead in the launch market because they got tax payer funding (less than Orbital ATK and Boeing) AND Musk funding. They made more constructive use of their funding than their competitors. SpaceX would not be where it is today without Musk's money (from Thiel running and selling Paypal). With apologies to people who found Twitter useful, I am very glad that has kept him busy for months.
Danish? In any case, while it's technically possible, that doesn't mean it's easy or the most cost effective solution (especially as the nature reserve might not take too kindly to changing the subterranean water flow).
I don't get why people are so keen to make engineering challenges and justifiable experiments seem like unnecessary and unforced mistakes.
>I don't get why people are so keen to make engineering challenges and justifiable experiments seem like unnecessary and unforced mistakes.
Most of business doesn't have the budget for more than one try -- everything's pared to the bone so its definitely "one and done" (unless someone's really good at hoodwinking upper management). This corrosive mindset has permeated US industry resulting in a lot of 'by the book' unimaginative designs and mediocre products (and a tendency to ship prototypes as finished products, leaving customers to help sort out the inevitable bugs).
They made an engineering decision to try to do without - partially because of the challenges of the local environment, partly because of the eventual design goals for SH/SS.
In the same manner they decided to leave the grid fins extended during ascent - the potential issues resulting from the technically unstable centre of lift being outweighed by the benefits of reduced mass and complexity.
There are quite a few decisions they make which might look odd, but without being in their meetings it's very hard to say that they're mistakes.
They thought the pad would survive one launch - they were wrong - but the rest of stage zero has held up very well indeed from what we can see. There are some dents in the outer walls of the tank farm (remember there is a serious amount of insulation between that dented wall and the internal wall), and there is going to need to be some more concrete work (I hope that the base of the pillars can be tied back together without having to dismantle the OLM entirely), and then the actively cooled steel base/diverter will be added.
Given the amount of debris which was flying around, I was somewhat surprised that so *few* engines had issues - and some amateur analysis I've seen suggests that one of the failures was in the hydraulic unit (which is already an obsolete component)
> They thought the pad would survive one launch - they were wrong
Be interested to know which “they” decided to go ahead…
Clearly, someone expected the pad to fail and hence commissioned the “ water-cooled steel plate”.
Musk suggests it was a management decision to go ahead rather than wait until the plate was installed. Which would indicate SpaceX have a mindset similar to that NASA had when the O rings failed….
This comes up time and again. There are bugger all available launch sites. You (almost) need United States to get DoD and NASA funding and because of ITAR. (RocketLab can launch from New Zealand because the New Zealand government put some real effort into jumping through the required hoops. Even so, RL plan to launch Neutron from the Mid-Atlantic Regional Spaceport.) You need excellent transport infrastructure to get rocket parts, payloads and rocket scientists delivered (some of the small launch startups try Kodiak or Kwajalein because existing sites are too congested to get a license for an experimental launch. They do not stay once they prove they can get to orbit). Polar orbits launch north or south, equatorial go east (except Israel because launching east from Israel would drop spent boosters onto people who would shoot back). You launch over the ocean to avoid dropping spent boosters onto people (Russians and Chinese don't count). You launch from nature reserves or you get complaints from deafened NIMBYs because you smashed all their windows.
When you check for sites against requirements the results are:
1) Merritt Island National Wildlife Refuge (AKA Kennedy Space Center)
2) Vandenberg State Marine Reserve (AKA Vandenberg Space Force Base)
3) Eastern Shore of Virginia National Wildlife Refuge (AKA Mid-Atlantic Regional Spaceport)
4) Boca Chica Lower Rio Grande Valley National Wildlife Refuge (AKA Starbase)
On the plus side, property developers have been kept out of nature reserves because of rocket launch complexes.
The cost (not price) of a fully expendable Starship launch is about $100M.
Bullshit. There are 33 engines which at a minimum cost $3m each so just the engines are over $100m so the full stack (there are more engines in the second stage) is going to be at least $250m, probably a lot more.
The true cost of the engines once everything is considered is probably nearer to $10m pushing the stack cost to probably over $500m.
$3m each - you got a citation on that cost?
The target price (and I am fully aware that we aren't there yet) is ~$250k.
In 2019 the marginal cost of each engine (i.e. the cost to build one, ignoring the design costs etc) was stated to be "approaching $1m".
I suspect that R2 is rather cheaper than R1, but they are still (almost) prototypes, so $1m is probably too high, but certainly not by a factor of 4.
The rest of the stack is mostly steel pressure vessel, so is probably on the order of $200k for materials.
I'd be expecting ~$50m as a first order approximation for these prototypes (the ~$900k of fuel is less than a rounding error)
A fully expendable, fully developed, ship would have ~40 engines ($10m planned) so probably an all up cost of $15-20m
Once you factor in all the ancillary equipment and amortize the failed units the cost starts to (for want of a better phrase) rocket.
Also Musk always underestimates future costs, you can generally add a zero to any sum he mentions.
The rest of the stack is mostly steel pressure vessel
I hope not, have you any idea what cryogenic hydrogen does to steel? (obviously not). The containment is probably aluminium with steel on the outside for structural integrity and such vessels are far from simple or cheap. Then you have plumbing, turbo-pumps and a shitload of other odds & sods which all cost over your $200,000 each. Rocket stacks are expensive (if you want them to work).
A modern FAB costs several billions, and more for all the support to provide raw materials and consumables. Chips have huge NRE costs. There is no way you could ever get 32GB of flash storage for £6 retail... Oh wait, that was the first price I just found with a quick web search.
SpaceX currently makes about 1 raptor engine per day and are aiming for 500/year. Take a look at the simplifcations from going from (late) Raptor 1 to (dated) Raptor 2. Ignoring NRE for RS-25 (2/year) requires the intelligence of a congressman. Failing to divide NRE by cadence and life time is nearly as bad.
So in that case all the buildings must be fully allocated to the first engine, and all the other engines get a free ride?
When do you stop amortising, or rather when do you plan to stop amortising.
If you intend to produce a million of something then you don't say after 100 that 1% of the tooling costs are attributed to each thing.
The rest of the stack *is* mostly stainless steel - perfectly capable of holding both oxygen and methane at cryogenic temperatures (just one reason they aren't using hydrogen).
Yes there are ancillaries, but the turbo pumps are part of the engine. There are a handful of pressure tanks, and plumbing - but the vast majority of the stack is two massive propellant tanks - You might have missed that I allocated ~$10m to the "rest" of the rocket, not alot of that is taken up by steel costs.
Yes - Musk is a natural optimist, but by aiming for ambitious goals, even if he falls short he's still ended up with more progress than others.
However since the prototypes of the raptor 1 were under $1m four years ago... I don't think the final price per engine (remember the simplified design of the R2, and the focus on design for manufacture) will be $2.5m - It might end up at $300k rather than 250... but that's still substantially less the $100m for the RS25e
"NASA announced May 1 it had awarded a contract to Aerojet valued at $1.79 billion to produce 18 RS-25 engines."
When do you stop amortising, or rather when do you plan to stop amortising.
You amortize over the projected manufacturing run. So if you are going to produce 1 million units each unit will have 1/1000000th of the setup and design costs added to the production costs and profit. If you go over the projection you can either drop the sale price or (more likely) increase the profit margin, if you go under the projection you fire your forecasters.
Rocket engines are never going to be a mass consumer item, even using 33 at a time (and they claim they'll be able to reuse which reduces demand) they'll never need significantly more than 1000 working units which raises another cost, poor quality of the engines, the cost of each engine that fails QC must be wrapped into the cost of the ones that do work if they intend to stay in business.
They *are* designing and building these as a mass production item.
To launch enough people and materials (not in that order) to colonise mars is going to take a huge number of launches - They are looking at sending a *thousand* starships to Mars in a transfer window... that's a *huge* number, and each ship will need ~8 launches of a superheavy (seven with fuel tankers) - even with significant reuse, it's still going to be a lot of engines that need to get made.
The R2 is designed for manufacture in a way that other engines haven't been, last year NASA stated that spaceX were producing seven Raptor 2 engines each week... Compare that with the manufacturing rate for the BE4, or the RS25e.
They are deliberately choosing to run a "hardware rich" development cycle - which results in very visible test results, rather than hiding the calculations and simulations and running those exhaustively and then tentatively (because you're afraid of breaking it) firing a prototype. SpaceX already have the next hardware ready to test, so it doesn't matter if you blow this one up (assuming you get good test data whilst you do it)
They are looking at sending a *thousand* starships to Mars in a transfer window
Wow! - You actually believe a word of that?
I will bet any sum you care to mention that will never happen. I'd be quite surprised if a single StarShip makes a successful landing on Mars.
With current technology colonising Mars is a death sentence for anyone daft enough to go, maybe a research base with a small rotating staff of a dozen or so but a mass colonisation is a fantasy for at least 50 years at the most optimistic, gradual colonisation over a few centuries is more likely to succeed and be both practical and affordable.
Just who is going to pay for ~10,000 StarShip launches? At the moment SpaceX is existing on government handouts, when that ends they are screwed.
"At the moment SpaceX is existing on government handouts, when that ends they are screwed."
SpaceX raises hundreds of millions a year in new funding rounds. Apparently there are still a few born every minute. This might be why they dropped to their knees on launch prices. If your competitor is charging 100, you don't and shouldn't start your pricing at 50. Why aren't the other guys price matching? They know that funding R&D for future vehicles and having cash on hand for contingencies is very important. Elon is counting on there being a never ending source of outside investment to keep the company going. Not his money, of course. Imagine if SpaceX had Elon on tap for funds as needed.
The costs you are arguing over are already paid, and the costs build and launch the test flights don't need have the ground stuff added in here. They aren't billing customers for this, they don't need to mark it up.
Those costs very much WILL matter if they can make it work, but you will probably never see an honest answer for either number, because sales.
Last thing the bean-counters at any company want is potential customers knowing what stuff actually costs. The price usually starts at 20% more than the customer actually has, and is negotiated up from there. So when the competition is coming in a ~$4 Billion, why give them a reason to gripe over that piddly 200 million extra you tack on.
(this also applies to the SLS and the per launch costs will probably be at least 50% higher)
There is a slight difference between something that is made by the million every day for decades by a myriad of producers to published designs for a fraction of a penny per unit and something highly complex with extremely limited demand which had to be designed from the ground up for which there is no available off-the-shelf tooling to build it with.
"There are going to be alot of these engines made... they are designing them for mass production for good reason."
Well yeah, they tend to go off bang, so mass production is important. It's not a good conclusion to say they'll be making this version of the Raptor engine a lot. They may finally figure out that they need to go back to a clean white board again and start something new that will work consistently and reliably enough to put on a manned rocket.
"Well yeah, they tend to go off bang, so mass production is important. It's not a good conclusion to say they'll be making this version of the Raptor engine a lot. They may finally figure out that they need to go back to a clean white board again and start something new that will work consistently and reliably enough to put on a manned rocket."
Really?
How many have gone off with bang when you exclude tests to destruction (where the aim is to make it go bang).
The raptor 2 isn't an unreliable engine... Can't quite call it merlin levels of reliability yet, because we haven't seen any non-prototype flights.
If, as you say "we haven't seen any non-prototype flights" how do you know that "The raptor 2 isn't an unreliable engine" - Are you taking SpaceX or Musk's word on that because there can be no other source and they might be just a tiny bit biased.
What we DO know is the fail rate as seen in test flights and stand QC testing is far too high for it to be a viable launch system as the engine is now, that should improve in time.
It's not looking bad in terms of reliability - despite having large chunks of concrete thrown at them the vast majority of the engines were still generating alot of thrust.
When you're in the process of designing something, do you consider the final product to be unreliable?
Here we have a company with more than a little experience in building rocket engines, building a rocket engine - I therefore need more than "oh there are some issues with prototypes" to declare it unreliable.
Even the flight tests we have seen have had good success:
- Star hopper (July 2019)
- Star hopper (August 2019)
- SN5
- SN6
- SN8 (failure from a loss of methane pressure in header tank)
- SN9 Finally an engine issue. One engine didn't relight - (now changed process so that they light three and shut one down)
- SN10 (helium ingestion from the fuel tank caused a hard landing)
- SN11 Methane leak in an R1 caused a hard start - (now changed to the R2 which has substantially less opportunity for leaks)
- SN15
21/23 engines operated correctly, with both failure cases severely mitigated.
When we get some decent news out about S24/B7 I expect we'll find that the booster was missing substantial pieces of the failed engines - and there were still no major (i.e. cascading) failures.
Is it human rated - no, is it viable as a launch system... I'd say yes.
It wasn't an engine failure which caused the RUD
"It wasn't an engine failure which caused the RUD"
Some independent high res video is suggesting just that. Venting can be seen where the flight termination explosives were mounted and the first of two big explosions is on the aft end at the engine bay followed by the explosion of Starship (the second stage). All the FTS seems designed to do is vent the fuel and oxidizer tanks rather than break the rocket into bits.
The FTS should have done more - it clearly needs an upgrade. 40 seconds is *far* too long between termination and disintegration.
The RUD was caused by the RTS, which was triggered because they had lost gimbal control due to hydraulic power unit failures.
They had had some engine issues (three were stopped rather than ramping to full power before the launch, others lost communications later), but they don't yet have evidence that there was any damage from the concrete which is pretty wild.
"What we DO know is the fail rate as seen in test flights and stand QC testing is far too high for it to be a viable launch system as the engine is now, that should improve in time."
The engine design (or re-re-design) might turn out to be too sketchy to ever be stable. I suspect if they had a clear handle on why they eat themselves on the test stand, they'd be fixed by now. Elon is on the second or third team managing that project and they still don't have a working engine. These engines are supposed to run up to five times a day with no more than a cursory inspection upon landing. The cadence is important right away, not in several more years. The lunar lander is going to require several flights of tankers which also will need SpaceX to have a working design for in-orbit fuel transfer that has a NASA investment but no word on where that project is.
"How many have gone off with bang when you exclude tests to destruction (where the aim is to make it go bang)."
SpaceX isn't saying, but neighbors of the McGregor test site where SpaceX validates the engines are reporting plenty of failures. While it is good to test to destruction, the vast majority of testing should be production units being tested before being sent to Boca Chica and that's not what being seen.
"The marginal cost is the cost to make one engine for spaceX now. The amount already spent on design is not relevant to the *cost* (rather than price) of the launch."
As they keep blowing up on the test stands AND they're in volume production as well, the design cost continues to climb. Once they have an engine with a high degree of reliability, you can call the design costs as final.
The actual rocket scientists at SpaceX wanted the water-drenched launch platform. Musk personally vetoed it after the parts were there, wanting to launch on 4/20. Because the pad had no water, the acoustic vibration of the boosters shattered the concrete and sent it flying for miles -- water would have dampened the sound and explosive force. It was very much an error. The Starship booster is a very low cost design but dependent on the very high cost launch pad, costlier than a standard one because it needs to support the whole rocket before liftoff and ideally catch it on return. That basically blew up. It may also have been flying debris from the pad that caused the engines that didn't ignite to fail, and possibly damage the control systems that led to the whole thing's spinning out of control. His "cheap, cheap, see what blows up" approach did not work well.
SpaceX does as well as it does because Musk is distracted by Tesla and other things and Shotwell can run interference between him and the actual rocket scientists. This time he got involved personally and it caused more failure than expected.
Their models showed that the concrete could survive one launch - they were wrong.
Current theory is that the thrust from the engines likely compressed the ground under the concrete, leaving it unsupported and therefore able to crack.
Of course the fact that it ran for 5-6 seconds on the pad didn't help, that's far longer than would typically be seen.
"With hindsight, using all the launch slots will now be difficult because of the 2 (probably x2) months delay for repairs to the ground support equipment."
The construction can be sped up by throwing more workers on it, but getting the FAA to issue further launch licenses is going to take much longer. I have to wonder if Michelle Murray, the FAA signatory, was handed a decision from higher up to issue a license for this launch or if she was doing her own due diligence. I hope it was the former.
They aren't worried about the damage to the pad, but they ARE worried that debris from the pad got launched further than expected. Then there is also the remains of the vehicle falling outside the prescribed area. SpaceX is required to file a contingency plan that identifies the areas that could be impacted by the mission. SpaceX missed by a large margin so, although routine, you can expect the FAA to hold SpaceX's feet to the fire to do better next time. Note that the public is being told to report any debris they find but not touch it because they don't know what kind of contamination it might have.
Frankly, the more info that comes out this launch is starting to smell like a shit show. If it turns out the vehicle was lost because those engines were damaged by pad debris, on top of Musk refusing flame tunnels or waiting until his water cooler is complete, all so he could launch on 420, then people are going to have to start questioning his judgement at SpaceX too.
I was not aware that rocket debris landed outside.
Is there a map yet?
It does look like the flight termination system failed too.
The sand and dust is a massive issue and may well scupper them launching from this site at all. It's really difficult to control and the FAA now know that the computer modelling was simply flat wrong.
It does look like the flight termination system failed too.
In one of Scott Manley's video analysing the launch, he says the flight termination system did work: It just didn't work how people thought it would.
Scott believes the flight termination system just pierced holes in the tanks with the expectation was that this would be sufficient for the rocket to disintegrate.
But the facts will come out at their usual space, so Manley and a million other people will have plenty of time to soak up the clicks posting content based on guesswork and assumptions.
Guessing the correct answer to a question is not knowing the right answer to a question. Educated guesses have there place in the process, but at this point you can also just wait until the essential facts come out and avoid the risk of finding out what your foot tastes like.
I'm leaning to the second one, as everyone talking about this is talking bout Manley saying it. The lack of other voices from the engineering side saying "yeah, I think he's right" make me take that theory with a grain of salt for now.
Wait for the fog of reporting to clear a little more.
A lot of people thought it very suspicious that it tumbled out of control for so long before it was actually destroyed.
At the time I thought it odd, but presumed they were well within the permitted flight path and had deliberately postponed it to get as much data as possible before triggering it.
After review of the footage (including the commentary), it's pretty clear that it had in fact triggered much earlier, but for some unknown reason did not destroy the rocket.
That's what I mean by failure.
The "why" is a different, and far more important question.
I suspect that the bottom of the rocket was damaged by that flying concrete, especially given the gas/ fuel streams that seemed to he coming from the lower edge of the first stage, but above the engine exhaust, and you can see a lot of debris flying up within seconds of ignition.
Remember that it only took one dropped socket to detonate a Minuteman rocket - different technology of course here, but having tons of concrete blasted randomly up and out is not going to do precision machinery any good!
Also I wonder if this was also behind any of the other launch failures ?
"I suspect the FAA are far more interested as to why starship lost control/lost engines/lost power/never seperated rather than the damage to the launch mount"
What if one caused the other? I suspect the huge chunks of flying concrete and rock were the reason 3 engines failed immediately and more failed later, and may also have caused the failure to cut off engines and hence failure to separate. There was lots of damage to storage tanks, the launch tower, cars, trees, and much more, so it makes sense the rocket itself was also damaged.
"The pad damage is a cost for spacex to fix, and make sure it doesn't happen again"
If there is good evidence that insufficient ground support is the root cause to the failure, the pad is a problem the FAA will have to rule on. There's also plenty of video showing the shoreline being pelted with debris right at the beginning of the sea turtle nesting season. I'm sure there will be some reports on impacts to wildlife, especially the numerous endangered species in the area.
The take off was incredibly slow. How many more engines failing would it have taken for the fully loaded vehicle to crash straight down and detonate?
Well it never separated because it never got to that point - that requires no questions at all.
It lost control because they lost hydraulics, which are already obsolete - they aren't even on the next booster.
Loss of engines, not really something they care about.
They should *really* care about the FTS, and I expect we'll have some exciting tests of that at ground level to verify it, because the rocket had *far* more structural strength than had been anticipated.
They also need to be concerned about the rapid excavation and significant ballistic trajectory of the excavated material - though again, there is no intention to launch without a rather different pad surface - a (rather substantial) steel "upwards facing" shower head. That should survive much better, but a static fire sequence might be required to demonstrate the performance.
Yes, it was the NSF camera van so was intentionally placed there & they knew the risks.
There is a members only video showing the damage to it up close. One camera attached got totally destroyed with the main part of it currently awol, unknown if they will find it & it's SD card. The aluminium mount it was on was snapped cleanly at the welds to the base.
The car, however, surprisingly did start up afterwards so they could remove it to an other location.
That's a good one to add to the local lexicon.
TURD is way better than RUD.
I'd have gone with Scheduled Highspeed Automatic Rapid Termination for cases like this where the launch was in deep sh*t and it took three loops for the self destruct to kick in.
Too much methane can do that to men in their middle years(so I hear).
Why on earth would you not just build flame tunnels as point 1 in your launch pad design? Every launch pad in the world has them. Why? Because they are cheap, easy to build and effective at taking forces and debris away from your rocket.
I'd love to know what the thinking was there.
"Were building the largest, most powerful rocket in the world. Perhaps we should build some flame tunnels to deal with all that thrust?"
"Nahhh, it'll be fine."
"But..."
"It'll be fine. What's the worst that can happen?"
"Ummm..."
"Good. Now, next order of business. Who wants to work for Twitter?"
If it was fresh water then perhaps (but probably not in a small site surrounded by nature reserve). Boiling salt spray would be problematic. Flame diverters are actually proven technology from the early days of rockets and are still used on some vertical test stands. I see horizontal test stands far more often than vertical. Probably cheaper and certainly more accessible but may not be so good for an actual launch. Some of the modern use of flame trenches comes from "It was already here when we got the lease on the site. Might as well use it."
It's at sea level so tunnels would flood. They've built a raised table instead. They plan to put a water-cooled steel deflector under the raised table.
In fairness, so is Kennedy Space Centre... but building up the launch pad the way they have in Florida would be an epic civil engineering undertaking, so if this design looked cheaper and equally effective (we understand acoustics and suchlike better than they did in the 1950s) then it makes sense that they'd go with the cheaper/low-civils option.
They explained this in the SpaceX livestream.
There won't be pads or ground support equipment on the moon or on mars so they need to get the hang of launching from close to "normal" ground and coping with the dust/gravel/rocks flying around.
Obviously what happens under a super-heavy booster with 33 engines launching in earth's gravity is a bit more extreme to what it would be like under a Starship with 3 or 6 engines launching in the moon's or mars' gravity but, in typical SpaceX style they decided to have a go anyway. Turns out that probably was a step to far.
But also bear in mind that just because the space industry has always done it with flame trenches/water deluges/etc. doesn't provide strong evidence that they are actually necessary. If "the space industry has always done it this way" was used as evidence for everything then SpaceX just wouldn't exist because the space industry was 100% sure that SpaceX's plans for reusability were impossible. And now look where they are - the first test of a system where both the booster and the spaceship are fully re-useable.
Actually, water deluges are important to reduce the acoustic effects of the energy reflecting off the launch pad. It's been proven to reduce excessive vibration. Given that Falcon 9 uses nine engines, not a single large one (or three, if you are counting boosters too), the effects are likely to be less (acoustic energy interference amongst all nine gas streams).
And given Starship's booster uses 33 (but has twice the energy output compared to Saturn V), a water deluge could have been useful. But if the test was to check whether the booster could use a very bare-bones structure without too much damage, the test has provided some vital data to SpaceX (and probably the FAA and NASA).
The space industry actually hasn't always done it with all of those systems. The first Space Shuttle launch didn't use a water deluge system, and the shockwaves from the launch damaged the orbiter. In fact there was a real question about whether it would be stable on reentry due to damage to its control surfaces. So the hazards here are well understood, it's not just that NASA overbuilt everything from the start.
" If "the space industry has always done it this way" was used as evidence for everything then SpaceX just wouldn't exist because the space industry was 100% sure that SpaceX's plans for reusability were impossible."
The space industry didn't always do it that way. Diversion trenches evolved just like everything else. Reusability hasn't been thought as impossible, just not financially viable due to risk and the need to make vehicles much more capable so they can be landed and reused. NASA was landing rockets on the moon in the early 60's for the Surveyor missions. Many companies had demonstrated reusability long before Elon came along. I worked for one and before that worked as a photojournalist with stories on file of several others.
Because the Boca Chica site is small (compared to everything else in Texas) and at virtually sea level. One hurricane going the wrong way and everything floods.
Cape Canaveral is at least a bit higher and bigger, and NASA has built up enough infrastructure to make sure stuff only goes into the direction they want it to go. Saturn V went from there, so the likes of Wernher von Braun et al knew what infrastructure was required, but NASA doesn't take kindly to having its infrastructure obliterated by things going pop (like Starship going *boom* on/next to the pad). So... Boca Chica with a launch table it is.
I don't think a steel plate is quite what you'd want either... I mean, molten steel blown into the air in tiny globules... more damaging than concrete probably.
One of the reasons they didn't is that they easily could have wasted time and money making the wrong kind of trench, with no actual assurance that wouldn't also have been LESS safe.
So they blew up a launch platform they probably would have had to tear down anyway, and their existing water cooled plate may have to be scrapped and redesigned now too. It sounds like the big news maker was a vehicle that was intentionally placed inside the safety perimeter by the National Science Foundation. So while far from a nothing burger, none of this is tying up the case for having built a massive concrete structure based on guesswork vs a (literally) calculated risk. While this was far from ideal, it's not a worst case either. So now that people can more accurately assess the launch risks they can design and implement improvements.
It does make the case for them to sort out what happened, and if a failure in the LOX system turned the stack into a giant cutting torch. And it gives the regulators some basis to actually order a credible system before they allow more launches, or move the launches to one of the existing launch complexes. Better to scorch concrete in Texas than to blow out one of the main pads in Florida or California.
Also good basis for the regulators to smack Musk down if he actually pushes for another test before the site has been brought back up to acceptable standards of risk.
This whole article was written as a factually inaccurate kill piece agains SpaceX.
Stating "SpaceX was unable to get Starship, said to be the world's most powerful rocket, into orbit during that experiment." is factually wrong.
The rocket was never intended to reach orbit.
"The rocket started rotating mid-air when it failed to break free of its booster shortly after takeoff on April 27"
The whole thing is a rocket. The SECOND STAGE" did not release from the first stage. Whether there was a failure to release, or whether a release was never commanded since the required release parameters were not reached is, as far as I know, not publically known.
"The mess descended onto Port Isabel, where residents reported hearing the roar of Starship's engines and feeling the ground shake."
There is noise and the ground shakes during a rocket launch. news at 11.
It is quite clear that the launch did not achieve the set goals beyond clearing the tower. It is also quite clear that the launch site requires repairs and modifications.
The reaction of the FAA is also news.
The rest of this article is fluff of a quality way below the normal standards of el Reg.
I am a bit spoiled by outstanding coverage from specialist sites. When I read space coverage on main stream sites I have far lower expectations but it is still usually better than most of the clueless pajama pundits. The Register earned a B from me this time - not the best but several huge steps forward from average. It is up to them whether they choose to invest a large amount of time and effort into getting better insight into the space industry.
Re-read your quote. It is literally true.
We do have some clues about the lack of separation. Although the tumble started at about the right time for separation, the altitude was too low and the vehicle too slow because of engine failures. The first step for separation is to gimble the engines to one side to start a tumble. Next the engines are shut down. Finally the clamps are released and the spin causes the two stages to fly apart. On this booster, engine gimble was powered by hydraulics. One of the explosions on the way up is believed to be part of the hydraulic system failing (I have not seen this in a statement from SpaceX). Lack of hydraulics could cause the engines to gimble to one side which would start the tumble and there would be no way to recover. The engines did not shut down perhaps because the software failed badly when separation was commanded but more likely because separation was not commanded. Thrust from the engines would press the two stages together even if the clamps were released but I have not seen evidence that they were. The rocket fell until the atmosphere got thick enough to create strong aerodynamic forces. These forces cause the rocket to buckle and shortly afterwards the flight termination systems activated.
Activating the flight termination system was sufficient to cause an FAA investigation so that is not news to space enthusiasts but may well be to others. Debris from the flight termination landing outside the cleared zone is a serious issue that I expect the FAA will require SpaceX to deal with.
If you look at close up photos of the launch, you can see the bells of the failed engines are missing entirely, suggesting that they were knocked off by debris from the pad. Whether this happened because they didn't ignite is a moot point.
I'm not convinced that a big steel plate is the solution unless there's active water jets firing over the surface. Water-cooled is totally different to the traditional water suppression system, which is designed to prevent damage to equipment from the sound energy generated by the engines. If that's not addresses then the plate will just reflect the energy straight back into the engine bay.
This is not something that would be missed, so we're not seeing the whole picture.
The original concept was a flame diverter, which starts near vertical and curves or steps to near horizontal. This keeps the flame flowing fast away from the source. Going straight down onto a horizontal surface keeps the heat in one place longer - remember the flow is supersonic so it is compressible. The pressure and density increase so the flame can spend more time damaging a simple flat metal plate.
AFAIK determined SpaceX watchers have not yet photographed this new equipment so we will have to wait for it to see the light of day before we know the current design.
They won't get approval to launch there till they get this stuff sorted.
By the time they get past the blowing stuff up phase they(both SpaceX and NASA) will be able to predict if the existing launch facilities should be retrofitted or if the make SpaceX pay for new pads designed to handle the new stack be built from scratch. Either way they will have figured out what will work before then, but their customers (NASA and the DOD) will probably want them to launch sensitive payloads from their complexes, not Texas.
"I do wonder how all this infrastructure destruction will affect their ability to launch from Vandenberg. "
Do they have plans to launch Starship/SuperHeavy from Vandenberg? SLC-6 that was just awarded to SpaceX is approved for F9 and F9H. Being a military base, Elon can't just start throwing up towers and GSE that isn't approved so he won't be able to sneak it in.
Taking a break from all these heated “Musk can” “Musk can’t” debates, what we really all need to know is how soon will it be when starship can be sent up into space, to open up and swallow another spaceship, before coming back down and landing in a volcano in Japan?
Black helicopter cos that’s as close to Little Nelly as I can get!
Saturn V was ~35MN (no question about it's launch status)
N1 was ~45MN (It certainly launched, though it didn't get as far (distance or time wise) as this test)
Superheavy is rated at ~75MN
Given that the stack mass is ~5,000kg, it must have had over 50MN to lift off (it didn't scream off the pad, so I'd guess it was closer to 60M than 70MN - I've not actually looked at the acceleration yet)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
