Do you know how many times Falcon 9 boosters exploded before they stuck the landing? I think it was four or five. They lost one in a fueling accident while learning how to use densified propellant.
In just a handful of years Starship has gone from being a paper rocket to maybe the most magnificent machine ever built. It is clearly going to work, and even if it were fully expended on each flight it would be a fraction of the cost of SLS.
And once they get on-orbit fueling worked out, which I’m sure they will but could take some time, Starship opens up a new era of space exploration.
A fully fueled Starship in LEO can put 150 tons into a transfer orbit for any planet except Mercury and Pluto. And it can do it every day.
The rate at which they take something that just barely works into a reliable, high-cadence system is really just incredible. I think I shared this animation before:
Every rocket they launched and landed up until June '23. There are multiple spots in the animation where you think “wow, they’re really cooking now”, only for them to kick it up a another notch a year later. And it doesn’t even show Dragon or Starlink as separate entities.
I don’t know when the high cadence for Starship will truly kick in, though I’d be surprised if it takes longer than a year. And it’ll be pretty amazing when it does.
Is that to me? If so, yes, I knew that. I also think they’ll blow up significantly more than 4-5 Starship boosters before they get one to land and a few after that before they can do it ultra-reliably. They’re trying to land a skyscraper with 33 engines from around 4.8k km/hr down to zero with a flip maneuver after hot staging. It’s really fucking hard. WAAAAY harder than a Falcon.
I’m trying to stay optimistic but Starship needs a bigger success than “got a little further than last time before blowing up”. If you knew nothing about the successes of Falcon 9, Falcon Heavy, Starlink or Crew Dragon, and went solely by the history of Starship and Heavy test launches, one might conclude that Space X was a tech billionaire’s fantasy, a bottomless money sink project that was attempting something an order of magnitude too hard for anyone but a major government to hope to achieve.
It sorta depends on how you count them. There was the Grasshopper test vehicle, which made 8 total flights. Then the F9R vehicles, which made another 10. Then, 5 “soft landing” attempts over the ocean. Then, two (failed) drone ship landing attempts. Finally, the first successful landing–but on land, not the drone ship. Two more (failed) drone ship attempts, and finally a successful drone ship landing.
Possibly, but if you look solely at the development of the Falcon 1, Dragon, or reusability efforts, you’d come to the same conclusion: lots of blowing things up, lots of seemingly silly mistakes, and so on. Falcon Heavy was way late (though there were good reasons). Falcon 9 1.0 was pretty smooth, though.
I suspect that in a few years we’ll have largely forgotten about all this, just as most have forgotten about the other difficult periods in SpaceX history.
SpaceX has posted a summary of Flight 2, what happened and what was accomplished:
Takeaways:
The Superheavy booster performed nominally all the way through stage separation, turnaround, and the start of the boostback burn. Then it went kablooey (technical term), for as yet unknown reasons. All 33 Raptors burned perfectly for the full duration.
Starship performed nominally for abut 8 minutes after separation, then something went wrong and they lost telemetry, causing the FTS to detonate. They verified FTS detonation by looking at the performance of the tracked vehicle or its bits.
The new pad design held up ‘as expected’, with only minor work required after the flight.
This is all very good news. The engines on both Starship and Superheavy appear to have worked fine. Starship was very near the end of its full duration burn when it went offline.
I believe SpaceX is using Starlink for communications with the rocket. I’m surprised we have no visuals. I would have thought the rocket and booster would have been festooned with cameras sending back high speed video telemetry.
I think they might trickle out new visuals over time to keep interest up. Every new set of imagery means a new set of articles in the space media. Not that Starship has customers, per-se, but it’s obviously in their interest to garner public attention. Maybe a few even hear about Starlink for the first time.
I wonder how hard SpaceX is pushing for a third flight this year. As I said, I think it’s unlikely but not impossible. SpaceX does have a significant interest in pushing for this year: their license for Boca Chica allows up to 5 flights per calendar year. Obviously, they have not used all of this year’s flights. They might actually push into that limit next year, though.
It’s hard to imagine another flight this year, considering that there’s really only a month or so before Christmas season and vacations kick in.
If we’re laying down markers, I’ll say the next test flight will be in February, with an outside chance of January. If that one is more successful, the cadence will pick up and we’ll see another flight in March/April. Just a guess. But they have a lot of analysis to do, and then they are going to have to make a bunch of engineering changes to fix what went wrong. Then they need to get another launch license. Getting the government to work fast in December is not likely.
Where I see a problem coming is in testing for on-orbit refueling. Fully fueling a Starship in orbit requires six launches in short succession, which they are not licensed for in Boca Chica. And I’m guessing learning how to do it will require even more launches before they get ir right - a minimum of two rockets at a time. How can they do that at Boca? Also, just getting to the point where they can attempt on-orbit refueling means perfecting the booster and Starship catch landings. How many flights will that take? I think there are going to be a bunch of flights before they even attempt it.
It seems to me that SpaceX is still dozens of flights away from actually landing HLS on the Moon. To do it in any reasonable amount of time will require far more than 5 flights per year. I assume they are planning to make some of these flights from the Cape, but that’s got to be years down the road since they have to manufacture the rockets there.
Maybe the plan is to fly the max, then use the data to amend the license to allow more flights per year?
The underprotected launch platform getting asploded (technical term). That went very well. At a glance, it looked like they could start stacking right away
Too few first stage engines lighting and staying lit. This time, it went of apparently without a single hitch, right up to separation.
Staging went completely nuts. This time, the hot staging system worked a treat and everything went on their merry ways for some time afterwards.
Everything that went wrong last time seemed to work right (at least by gross evidence). Now they have new faults to troubleshoot.
Engineering success is sometimes best measures by the number of errors and faults you don’t repeat.
A huge advantage of modern spaceflight over 1970s or earlier spaceflight is the telemetry bandwidth. So much more data coming so thick and fast, plus video & maybe audio.
Such that nowadays there’s usually not much guesswork about what went off-nominal and what went catastrophically wrong. Unlike Ye Olden Dayes where a lot of engineering analysis was in fact high-tech guesswork about what went wrong and how we might best improve fix it.
If Starship’s problem was purely a telemetry failure that triggered FTS on an otherwise fully functional vehicle that’s one thing. A darn shame, but probably a quick fix with few side effects.
But if there was another anomaly which cut telemetry before the data about that anomaly got down, and then FTS finished the deal that’s something very different. And something much more like Ye Olden Dayes where educated guesses play a large role in the post mortem and vNext process. Or where you’re stuck re-flying the same design with minimal tweaks and hope you get telemetry up through whatever goes wrong again.
Ideally, getting a new license should be trivial this time. There’s no brand-new launchpad or completely untested vehicle like with IFT-1. There’s no environmental or public safety issue like with IFT-2. Hopefully, FAA can get started immediately and be done in a matter of weeks.
Yeah. And it’s not just bandwidth that has improved. Solid state sensors are way better, and instead of thousands of pounds of analog wiring feeding transmitters, modern digital buses make additional sensors easy and cheap with almost no weight penalties.
Falcon 9 has three thousand telemetry channels. I’m guessing a similar number for Starship. Everything on those ships is instrumented to the max.
Your lips to God’s ears. And the FAA has been really good about licensure, I must say. They seem to be trying to help instead of impede. Still, a month to a new launch license would be a pleasant surprise for me.
Yes, I see no evidence that the FAA or anyone has been stonewalling things. But they do appear to be resource constrained–something that LSLGuy and many other reports have confirmed. So it matters a lot if this ends up as a 45-page report like we got with IFT-2, or if it’s just a few pages of “looks good.” We’ll just have to see.
Hey, SpaceX is back on YouTube. They have a 360 degree view of the launch:
There is a point where you see some small rocks/sand flying. Probably there was some degree of concrete ablation from somewhere on the pad. Nothing serious-looking, though.
I had a hard time from earlier videos figuring out how far from the Starship the booster was before exploding, and wondered if there was any chance of debris or something else hitting the Starship. But from these shots, it’s clearly quite far away.
Looked like a bunch of combustion instability in the ~30 seconds leading up to the explosion. Unsure how much of that was planned versus random. Certainly if they’re using the Raptors firing aysmmetrically to rotate the stack we’d expect to see lots of huffing & puffing.
The first stage was at ~90km altitude at staging & breakup. So ~270K feet. As fast as they were going the air drag would still have been material. The good news is if there was any venting of propellent or oxidizer from an engine that wasn’t firing, it’d rapidly be swept away by the breeze. The bad news of course is if they can’t get engines to start timely or stay lit, the situation is deteriorating real quickly at that point.
That final explosion in the vid was certainly plenty impressive enough to be used as a Hollywood special effect in some future SF movie. All the better for being real too.
It’s hard to say. There’s definitely a whole lot of strange venting going on. From the SpaceX stream, we can see that they managed to restart several of the engines, but not all of them stayed lit. I’d be somewhat surprised if they attempted to light them asymmetrically–the inner set of 13 engines can all fully gimbal (within the limits set by adjacent engines), and that’s going to have a more significant effect than lighting them asymmetrically. Though I suppose that, given that they do not relight all of them at once, they may as well light the ones that contribute most to the flip first.
I’m sure we’ll find out soon enough what really happened. I don’t think we even know yet if the booster exploded of its own accord or if the FTS activated. As you suggest, the situation was probably deteriorating rapidly and perhaps the FTS activated due to the vehicle leaving the expected flight corridor. On the other hand, maybe it was a fluid hammer effect that traveled up the downcomer and ruptured the landing tanks or the bulkhead between oxygen/methane tanks.
Oh, that’s really neat. The staged cutoff–with groups of 5 engines shutting down at a time–was visible from other footage (and very pretty). But from this footage, you can see another effect–the group that’s about to be shut off gets throttled down first. The reduction in brightness is very apparent.
They need some more launches just to clear out the rocket garden! They’re building them faster than they can launch. And they’re producing an engine a day, so those are likely to be stacking up as well (they build a new set of engines every 39 days).
