SpaceX Starship SN8 Does All The Hard Stuff And Then Crashes
So let's see #SN8 took off, shut down 2 engines, made it to altitude, did the belly flop, flew, reoriented itself, descended, and then had a rough landing. pic.twitter.com/kpkJ3EScb5
— NASA Watch (@NASAWatch) December 9, 2020
When the engines came back on we were cheering. Amazing.
Watching the stream live was astonishing. I didn’t know what the plan was, and assumed after one and then two engines failed then it was about to get aborted. But no, it made it back to the pad and nearly landed. They will learn so much from that. Incredible.
Was it engine failures or staging?
It was engine staging.
The Raptors did not fail. The vehicle becomes lighter as fuel is consumed, so less thrust is needed. The engines can throttle only so low, so fewer were needed as the vehicle reached the designated altitude.
No question that to do almost everything and then fail at the last hurdle is impressive for such an early test of such a mould-shattering design.
Elon Musk has reported that the engine out during terminal descent that led to the crash was caused by low pressure in the header tank. So, that is obviously going to need to be looked at and mitigated. However, all told this was a very successful and impressive first flight that has certainly changed how we will look at space re-entry vehicles forever!
Let’s just remember that Musk actually said it was more likely than not that this test would end with a LOV of some kind.
If that was the problem, it shouldn’t be too much to increase the tank pressure. It also appeared that the vehicle wasn’t completely vertical just before touchdown. Probably due to the one engine being out. Once the smoke cleared it looked like something was sticking up, possibly one of the fins. Sounded like they were pretty happy with other aspects of the flight. Interesting to see the view of the engines from inside the vehicle. Attitude control appeared good throughout the “flop” maneuvers.
I still love Musk’s approach: build a little, test a little, analyze, build some more and test some more.
Next up: SN9.
Go, SpaceX.
That was incredible. Wow. Congrats to SpaceX.
The free flight was rock solid – in the SpaceX stream you can see the flaps actively control the flight. And, the landing, if “a tad” hard, was right in the middle of the pad. Wow.
Elon tweeted: “Successful ascent, switchover to header tanks & precise flap control to landing point!” and “Fuel header tank pressure was low during landing burn, causing touchdown velocity to be high & RUD, but we got all the data we needed! Congrats SpaceX team hell yeah!!” and “Mars, here we come!!”
Keith nails it when he says “Starship SN8 Does All The Hard Stuff And Then Crashes.”
Ascent despite the failure of 2 engines, freefall, relight, and landing control including the swing went great! It maintained control with one and then 2 engines out. Space X knows the hard landing was caused by low header pressure. SN9 ready to go after some minor improvements. This was a spectacular success!!!
It isn’t at all clear that two engines failed during ascent. I’d say they were shut down intentionally (although perhaps a bit more roughly than intended) and sequentially. If they had failed, it’s hard to see how they would have restarted during descent.
Points well taken! Either way, I am encouraged that so many new maneuvers went well.
When the first went out during descent, there was a lateral burst of flame that made something in the engine well catch fire. It burned for maybe 5-6 seconds IIRC. I was surprised that people haven’t said much about that.
It’s precisely when the side of the engine well catches fire that we start to see a white gas leaking out of the base of the rocket. Is that why the tanks were at low pressure during the landing?
In comments on the Ars article, several people mentioned that rockets tend to shut down by turning off the oxygen flow before the fuel flow, and that’s likely to produce some burning methane around the engine after a shutdown. To me, it looked like more than a little, but who knows. The venting could also be part of the shutdown process. As far as the header tanks are concerned, they’re on top, and I don’t know what the plumbing between them and the engines is like.
During a scheduled engine shutdown the now low pressure plume can get blown back into the engine well by recirculation, so not a problem. Venting near the base is also normal and has been seen on the pad.
I’m curious about the engine shutdowns during the flight. Was that to control the acceleration as they used up the fuel (dropped weight)? Or, were they having engine issues during the ascent? I know they can deeply throttle, but that still means a minimum of ~40% thrust, so I’m more inclined to the former, but I’m very curious for SpaceX to share more information. They went down to 1 and then relit one just prior to touchdown, and then something happened (loss of engine, attempt to relight 2nd or light 3rd?), so I suspect the engine shutdowns were planned but for the last one.
According the Mr. Musk’s tweets, the engines worked without problems during ascent. So the shutdowns were planned and not a result of a problem. He also tweeted, and a statement from SpaceX confirms, that the problem during landing was due to low fuel pressure from the vehicle’s header tanks. That cause low thrust and (perhaps) a engine shutdown prior to landing. So the vehicle landed at a high speed.
I learned a new term today from astrophysicist and rocket enthusiast Scott Manley. I am familiar with burning fuel-rich or oxidizer-rich, but knowing the emerald color of burning copper pretty well, I think I agree with his assessment that at the very end you had “engine-rich” combustion. I suspect running a turbopump dry could do that, but I’m sure we’ll know more soon enough.
I have never seen an aerospace company that is so completely open about their failures: when they happen, what they learned, and what they’re going to do about it. It’s quite refreshing.
That SN8 engine wasn’t running on a dry turbopump. The one with the green exhaust plume was firing all the way down, so it must have been getting _some_ methane flow. Just not enough to produce the required amount of thrust. If anything, that would be worse for the engine, and make running engine-rich a like explanation for the color. With some methane flow, the gas in the combustion chamber and nozzle would have been oxygen-rich and _hot_. Possibly hot enough for some dissociation and atomic oxygen. That’s pretty likely to burn with anything in the engine, and it seems the lining on the nozzle was copper.
It seems someone on the internet has turned this around. This may have been the first flight use of a hybrid methane-metal-oxygen tripropellent engine.
“It seems someone on the internet has turned this around. This may have been the first flight use of a hybrid methane-metal-oxygen tripropellent engine.”
Ha! You win the internet for the day!
A turbopump running dry would have led to a rather abrupt stop to the burn, one way or another. Possibly the engine that shut down had this happen. The other one kept going to the end.
The fuel is the coolant as well as taking part in the combustion, so at the same time that the ratio was shifting toward more oxygen in the mix, the combustion chamber and nozzle would have been overheating. Not ideal.
They’re not completely transparent, especially at the moment of failure. This flight was a bit of an exception to that. I’ve watched multiple launches where there was an issue that was glossed over when it happened. (Falcon Heavy demo mission’s central booster landing comes to mind.)
There was no engines out until the landing, AFAIK. The shut-downs on ascent were planned to keep acceleration and speed down.
They were planned, at least based on Mr. Musk’s tweets. But we don’t know why. Keeping down acceleration as propellent was burned is plausible. But testing guidance and control systems with one, two and three engines, as well as control during a engine shutdown, is also a plausible test goal.
To stay sub sonic and to slow down to the target altitude.
Also a good and plausible reason. Which means we have three such reasons why SpaceX might have decided to conduct the test in the way they did. But in terms of data, we can only say that they did decide to do it that way. We don’t have enough information to figure out _why_ they did it that way. It could have been for any one of several good reasons, or several of those reasons in combination.
Going slower during the climb might also help them study engine performance at different altitudes. Instead of blasting through the troposphere in sixty seconds, this was sort of a slow motion version of that, with more time spent at the different atmospheric pressures.
It’s notable that the final engine did effectively a launch-length burn as a result. On a vehicle in flight, as opposed to on a test stand.
Two engines didn’t fail on assent. Those two engines were meant to shutdown.
I count this as a success. Yes, some engine problems meant it came in fast, and blew up on the pad – but it got back to the pad after a successful launch and flight. The engine problems can be solved, and hopefully SN9 will go without a hitch.
More significantly, NASA needs to open their eyes and take note, and take a really long hard look at themselves and their approach with SLS. Musk and SpaceX are doing new things – innovating, and doing so in a manner that will be transformative for US space exploration. SLS is doing nothing and will not innovate at all.
The future lies with fully reusable space launch with regular (daily) access to LEO at low cost – not outdated expensive and fully expendable big dumb boosters that fly once a year and cost US$1bn per flight.
Low cost, reusability and frequent flights are certainly critical. But I think the more critical lesson is about how to achieve that. NASA and more conventional aerospace companies try to design and build a finished product, before doing any real test or test flights. The idea is that doing everything right the first time makes it likely that there will be no failures and the product, once delivered, will work as intended. But that approach is very expensive, time-consuming, and may fail if there are problems missed during design and development. SpaceX, with Starship, is taking the approach of building, flying and crashing preliminary designs, to learn what does and does not work. And then iterating on the design, trying again, probably losing more test articles, and repeating. That seems like a faster, more efficient and more productive way to end up with a final design which works and does what it’s supposed to.
And also build knowledge about what it does when it doesn’t do what it’s supposed to. That nose section is surprisingly intact…
One would like to think that NASA might take this to heart, but they didn’t in the 1990s with the DC-X which executed close to the same flight profile as SN8 in flight test 8, and it used closer to traditional manufacturing methods than this, so not holding my breath. Also, I think it will take thousands of flawless flip maneuvers before NASA would stick their astronauts in there. I would like to think they’ll eventually substitute the super heavy for the SLS core stage if the SH works out… we’ll see. Thomas Zurbuchen seems to be a fan, but he’s SMD.
That was sooooo much fun to watch. I think they’ll nail the entire flight with SN9. I can’t wait to see it.
When I saw it it seemed like it was all in slow motion and was really surreal. Fantastic test.
Eight years after the largely recycled technology SLS project began, SpaceX started building in a field in Texas. Two years later, they pioneered new engines, a new flight regime, new construction techniques, new aerodynamics, and flew a demonstration mission that retired probably 80% of the questions regarding their novel EDL process.
At the same time, the $billions more expensive SLS is undergoing “green run” trials and might fly in 1.5 years if we’re lucky.
I nice perspective of the flip/landing maneuver.
https://streamable.com/2y12eo
https://twitter.com/nova_ro…
I was just amazed at how slow it seemed. Especially when it was performing the belly flop maneuver. I wonder how high it got as the Starship almost went beyond Everyday Astronaut’ range camera.
I think I read 40,000 feet which is higher than most airliners.
The planned altitude was 12.5 km, which would be just a little over 40,000 feet. I’m not sure what the actual maximum altitude was. But since everything went as planned until the landing, I assume it did get to about the planned 12.5 km.
Could anyone explain the thinking on the necessity of ‘header tanks’? Why is a different set of tanks needed, in lieu perhaps of utilizing the main tanks? Is this mass distribution, or?
And kudos to our host’s headline; I’ve seen a dozen reports on the media, all focusing on the explosion. Keith got it right.
Since the landing burn relight occurs while the vehicle is freefalling sideways the nearly empty main tanks would have fuel sloshing around. By having small tanks dedicated for the short duration landing burn it’s a lot easier to pressurize them and have better control of the fuel flow.
The LOX header tank is located in the nose for better weight distribution during the belly flop.
Falcon 9 presumably gets by without header tanks because of the already smaller tank sizes which is easier to keep pressurized for the landing burn, and also because even though the booster is freefalling it is falling vertically and so atmospheric drag slowing the vehicle will help keep fuel at the bottom of the tanks, which would not be the case for Starship while it is falling sideways.
All of that, and also Starship needs to store some tens of tons of cryogenic propellant for up to several months. It makes sense to do that in smaller, more thermally isolated tanks.
How do you drop a chicken egg 3 feet from above a concrete floor without breaking it ? Drop it from 4 feet …
The same context applies to SN8. It was a very successful maiden flight. Except for that last little bit above the concrete floor…
I was extraordinarily impressed by it all… way beyond all reasonable expectations.