Hacker News Reader

Comment by guardiangod | Hacker News Reader

https://twitter.com/elonmusk/status/1336808486022258688

Successful ascent, switchover to header tanks & precise flap control to landing point!

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!!

It seems that the Raptor engines didn't failed due to defects at the end of the landing. The fuel pressure from the header tank wasn't high enough and the engines became fuel starved. It is obvious in the video as the engine exhaust became bright green at the end, a sign that the engine was restarting by TEB.

Edit:

https://twitter.com/elonmusk/status/1336810077555019779

Mars, here we come!!

Indeed

View on HN

> It is obvious in the video as the engine exhaust became bright green at the end, a sign that the engine was restarting by TEB.

They wouldn't carry enough TEA/TEB to make that much flame - it's likely the engine itself burning. Also I don't think Raptor uses TEA/TEB as that would put a cap on the number of restarts - not something you want for an interplanetary mission!

It uses electric spark igniters.

If I remember correctly, green from Raptors has always been the result of “engine-rich combustion” of copper in the combustion chamber/nozzle.

Just amazing to watch.

The flip down to gliding just looked so bizarre and futuristic.

It's the big WTF I got from first seeing Grasshopper stop and hover in mid air and land again. That's not what rockets are meant to do!

Except of course they are meant to do it!

I've been following SpaceX for years, and this is exactly my reaction today! Watching the belly flop and the (relatively) slow descent, it felt completely alien, surreal. I'm in utter awe. Congrats to the teams involved, can't wait for SN9 and the rest of Starship!

the apollo lunar lander did pretty much the same thing.

The Apollo lunar lander objectively did not belly flop through atmosphere at 1G while having to deal with aerodynamic control surfaces.

Are you seriously comparing the Lunar Lander to this?

The Lunar Lander was piloted since the computers couldn't do everything, but at the same time a human was good enough at piloting to be able to take over for what the computer couldn't do.

This is the result of modern day CFD that that would take humans millions of years to do by hand

>This is the result of modern day CFD that that would take humans millions of years to do by hand

the package as a whole, yes.

belly reorientation for flight has been a thing since the early American arms' race. It's a useful missile trait.

Gimbaled vector rockets have existed since the 1940s.[0]

Yes, restart is very important -- it's not a requirement for atmospheric belly flight, though.

[0]: https://en.wikipedia.org/wiki/RTV-A-2_Hiroc

Nailing the transition to and from belly flight is the hard part, especially so when it's so close to the ground with little margin for error.

Military have been doing the belly flop for years

https://www.youtube.com/watch?v=o9vljzfywJI

Alpha Draco was testing body lift in the fifties https://en.wikipedia.org/wiki/Alpha_Draco

Robert Brulle has a fascinating memoir https://www.airuniversity.af.edu/Portals/10/AUPress/Books/B_...

Aside: the book Digital Apollo[0] gave me the impression that the computer might well have been able to land, but we'll never know since the astronauts didn't trust it enough to find out.

[0] https://mitpress.mit.edu/books/digital-apollo

And they had enough confidence that it might work that they attempted to land the first attempt on solid ground instead of in the water.

Yeah I thought they might do the same kind of 'pretend to land' hover over the water the first time, but no!

Maybe because they want to be able to keep it for posterity if it makes it - or launch it again!

It is the flight back that they need data for, not landing.

Not much different from early work on Falcon 9 landings -- the Grasshopper test rig and F9R-dev-1 prototype both did a bunch of land-to-land hops (the last being the crash that destroyed F9R-dev-1 after an engine failure) before they transitioned to attempted water landings from orbital flights

> (the last being the crash that destroyed F9R-dev-1 after an engine failure)

From memory I think it blew itself up in the air when something went wrong, rather than crashing.

The F9 water landing attempts were just because it was so far downrange though right? Not because a water landing would have been somehow better for the vehicle?

iirc when spacex asked permission to land a falcon9 on the ground it was a firm "are you crazy? no." So lots of water practice.

Hmm yeah you may be right. I definitely remember some failures that stopped above the water then tipped over but I don’t remember the circumstances.

If they'd decided for a water "landing" with SN8, I'd have assumed it was because they wanted any failures to happen further away from their test infrastructure, not necessarily because it was easier.

this sort of thing existed before spacex

https://www.youtube.com/watch?v=o2sHf-udJI8

It really didn't though. I'm not aware of any vehicle that does the belly flop flip to vertical landing maneuver, or has even attempted it.

Why I love me some DC-X footage, and bemoan what we lost when NASA killed the project for their own boondoggle that failed, talk about missing the point.

• 45,000 feet is a lot farther that 450 feet. • Falling and using movable aerodynamic surfaces to descend in reentry position, then rotating into landing position is kind of hard.

• Running rocket engines for over four minutes straight is hard, BO has barely run their engines more than 4 minutes total in years of ground tests, without the realism of actual flight and its associated vibrations and accelerations.

• Restarting rocket engines is hard.

Lastly DC-X cracked up its last landing too.

Retractable leg failed to extend and it tipped over when the engines were throttled down.

The real issue is why did it have retractable legs in the first place, esp on a test vehicle. Sounds like a great to start writing a post mortem is in the design phase.

It was a super cheap program as well. The post mortem should have been let’s build a bigger, better version with a similar budget, but instead NASA killed it to go all in on the 10X+ more expensive X-33 and killed that well before first flight.

RUD = Rapid Unscheduled Disassemble

For the people not familiarized

> With liquid propellants (but not gaseous), failure to ignite within milliseconds usually causes too much liquid propellant to be inside the chamber, and if/when ignition occurs the amount of hot gas created can exceed the maximum design pressure of the chamber, causing a catastrophic failure of the pressure vessel. This is sometimes called a hard start or a rapid unscheduled disassembly (RUD).

https://en.wikipedia.org/wiki/Rocket_engine#Ignition

Raptor doesnt use TEB. It uses spark ignition

It must be the engine burning its internal alloys then. Probably the turbopumps' due to the fuel running rich?

Probably oxygen, since the fuel tank had low pressure

That's my suspicion as well. If it's methane rich then the exhaust would have a lot of soot due to incomplete combustion. If it's oxygen rich then the burn temperature would be much higher than designed, and every internal metal surface would get oxidize in a hurry.

This line of thought is very reasonable for normal engines, but I wonder if it’s not quite right for the raptor.

I’m not 100% on this, but I think the raptor actually runs really oxygen rich. According to Elon’s tweet and Wikipedia, it runs 3.5-1 O2-CH4. It runs so lean that the unburnt o2 cools things.

If it were running closer to 1-1 (low ch4) then it would run much hotter and burn things like we saw.

Edit: reading more I’m not positive they run O2 rich anymore

Minor correction: 4.0 lox/methane is stoichiometric. 3.8 is rich. 3.55 is very rich. It runs fuel rich to keep cool and possibly increase specific impulse.

You were probably thinking for the turbo pump? There is a lox rich condition somewhere to get full flow. Just not the overall ratio

To clarify this point. The Raptor uses two turbopumps. One turbopump hangs off the side and runs fuel rich, and this fuel-rich exhaust is the fuel input for the combustion chamber.

The second turbopump, to which tdy721 is referring, sits right above the injectors and runs oxygen rich. This oxygen rich exhaust is the oxidizer input for the combustion chamber.

Why would oxygen rich combustion be hotter than stoichiometric combustion?

I know very little about rocket engines, but I would imagine that the fuel pre-burner, which appears to be intended to run extremely fuel rich, might have some serious problems if it didn’t have enough fuel coming in. Aside from just generally not working, it could internally burn much less fuel rich and therefore much hotter than intended, resulting in any number of failure modes.

The adiabatic temperature peak is stochiometric. However dynamics can shift the peak temperature lean or rich.

However, SpaceX is also tuning for engine temperature control and specific impulse. Raptor runs ~3.55 lox/methane (very rich), 4.0 is stoich.

https://www.researchgate.net/publication/255632584_On_the_Of...

Running fuel-rich is cool since the fuel won’t find extra oxidiser to burn.

Excess hot oxidiser on the other hand can often fuel. Like the combustion chamber wall.

This is the answer that the GP was looking for, but it's phrased funny.

sitharus means that the excess oxygen will oxidize (burn) things other than methane if there is enough oxygen present at the temperatures and pressures found in a rocket engine combustion chamber. It will oxidize (burn) things that you wouldn't normally think of as "fuel", such as the copper in the metal alloys.

Once you hit a propellant bubble, the combustion temperatures are the least of your worries. The turbines are designed to ramp up gently, work fully loaded and transfer the energy from the hot side to the liquid propellants; the power density is enormous, 100.000 horsepower in the space of a truck engine 100 less powerful. Without an energy sink this immense power is directed at the turbine itself, which spins out of control and will dismantle or shatter when hitting liquid again.

Your engine is now full of solid projectiles that take out other parts, fuel lines and injectors, damage film cooling nozzles and coolant canals, and in general transform your precision machine into a smoldering pile of gunk.

heh describing the ramp up of a turbo pump on a rocket engine as "gentle" seems funny. zero to 100k HP in about 500ms or so right?

also, from the pictures I've seen, the turbopumps are about the size of a propane tank, even smaller than a truck engine.

Well, half a second is quite gentle compared to ramming a fluid with a speed in the tens of m/s into a surging turbopump, especially one that is designed to minimize mass and exactly handle normal operating load and not more.

The exhaust was engine rich :)

It's probably the engine bells, since they are known to contain copper, and copper burns green.

Source: https://mobile.twitter.com/elonmusk/status/11773871411160023...

copper gets used because it conducts heat very well and so assists in cooling. In one of the first test stand videos of a raptor run you can see a little bit of green and I think Musk even tweeted it was running "a little engine rich". the first time I watched a night flight of a falcon9 there was a green glow right at ignition which caused me to hold my breath but that was the hypergolic ignition fluids which also burn green. Raptor uses a torch ignition so any green from a Raptor is very bad news.

> It is obvious in the video as the engine exhaust became bright green at the end, a sign that the engine was restarting by TEB.

The Raptor doesn't use TEB. It uses spark ignition.

The green is most likely 'engine rich' exhaust. Meaning the cooper cladding of the engine bell started to burn away.

Wow, thanks for that link, clarifies what happened. I never thought I'd see a "flash gordon" shaped space ship actually fly, and no here we are.

I think once the aerodynamics are tried a little bit more, it can be refined to a lifting body. Strakes and deltaish wing, something a bit like the space shuttle or X-37.

Having completely cylindrical body and two sets wings now for the prototype phase makes control authority good and iterating changes very easy (you can change either wing shape easily), even if it's not optimal mass wise.

What would cause the fuel pressure to be too low?

Centrifugal forces of doing the belly flop maneuver. Ever try to flip a spinning object sepaofrom the axis it’s spinning on? Imagine a fuel pump or turbine full of liquid oxygen. The engineering to get this to work is astounding.

Can you explain this more? I’m intrigued to learn about it

Not sure, but they might be talking about this: fill a bucket with water and swing it upside down over your head. The water stays in the bucket because centrifugal forces apparently push the water outwards.

The same may have happened for the fuel, I guess lowering the pressure? I know nothing about rockets.

Aside from what a sibling comment mentioned about fuel sloshing from the flip maneuver, Raptor/Starship use autogenous pressurization (i.e. some of the gas produced by the engines is tapped off to pressurize the tanks), so there could have been an issue with that, too.

"Fail fast, fail often"

Seems to be working for SpaceX. Makes for impressive videos too.

Congratulations to all involved.

Raptors don’t use teb or hypergols. This is some other fuel mix reaction

Were all the engine outages planned?

I believe some were. The plan was to switch over to two engines on ascent, and use two engines during the rapid deceleration. The full shut down was intentional. It looks like they might have briefly went to one engine during ascent but it wasn't entirely clear. The flame out of the first shut off was probably not intentional as well. It's really hard to tell.

None of that information was official though so take it with a grain of salt, it's all come from personal knowledge and analysis by some people watching.

According to Musk, yes: https://twitter.com/elonmusk/status/1336818987389181952

The orientation change just prior to shutoff is also a clue: https://twitter.com/joebarnard/status/1336820348071849986

This test of lossless convexification showed good results with the stochastic control problem, but the fluid dynamics model could use some work before:

> Mars, here we come!!

https://www.youtube.com/watch?v=RXJKdh1KZ0w

A nice try to make an explosion sound like a success - reminds me of the Soviet times when the Party was never wrong!

It's a test flight. Musk gave only a 1/3 chance of mission success before the flight.

Yes, but a 1/3 success chance doesn't not mean a 2/3 one of an explosion.

After engine ignition, isn't it pretty much that success or massive explosion are the only outcomes?

Nah engines are able to turn off after ignition. Thats what the Raptor safety is. Usually the blow or moment is when it lifts off the tower with one notable exception that I know of.

https://en.wikipedia.org/wiki/Mercury-Redstone_1

Technically the Shuttle go or blow moment is when the SRBs are turned on.

But yes, otherwise in the spirit, it will either blow up or be generally sucessful.

So you’re saying that if the engines had been off the starship wouldn’t have exploded? Try again bud. That’s still an object full of compressed methane and liquid oxygen falling at terminal velocity.

This is such a straw man that I don't even know where to begin to answer.

So you are admitting that what Musk meant was they had a 2/3s chance of an explosion.

Was it a complete failure?

Well, nothing is a complete failure ever from your point then - you always learn from mistakes.

Yes, but if it explodes during launch you learn only a little(you don't have any data for flight and landing) Here it went almost perfect because they have the data for everything to be able to fix issues

We have launch, controlled flight through the desired test regime, and a botched landing.

And as others have said, it's either complete success-or-explosion that's expected in this area.

What point was that? I just asked you a question.

No, I'd say it was an almost-complete success. The point of this was to test several things which they've never tested before and gather data to drive design improvements. They successfully tested: ascent under three engines, flap-controlled aerodynamic descent to the landing site, the landing flip maneuver - including drawing propellant from the header tanks - and final landing burn.

They didn't stick the landing, but it looks like they got test data from all the planned test activities. The explosion will prevent gathering some data from post-flight inspection, but other than that I doubt it has much affect on future development.

To look back on SN5/SN6 testing, they stuck the landings with those prototypes, but the prototypes themselves are now obsoleted by the new prototypes in production (up to SN16?!)

It sounds like they've got SN9 just about ready to go, too, so they should be able to put the data that they gathered to good use quite soon.

Seems to be a great benefit of this method that SpaceX is following with Starship: if you build hardware fast and cheap enough, you can afford some partial failures since you just put the lessons learned into the next flight in a few weeks.

They had to get a lot of things right on the first try for a successful landing. They got a lot of them right, but had trouble with fuel pressure at the end and landed too hard because there wasn't enough thrust.

As problems go, that's a relatively simple one that they ought to be able to address. The ascent appears to have been successful (though for a few moments it looked like something caught on fire and I'm not sure if that was planned or expected). The glide worked. The rocket was able to navigate to the landing pad.

Next time they'll be working with a lot more information. This seems like overall it was a success, even though they're down one rocket and three engines. A failure would have been the rocket blowing up on the pad during launch or other similar mishap from which they wouldn't have learned much aside from "make sure that doesn't happen again".

The fire after first engine cutoff was just extra gases from the cutoff engine getting trapped and burning off.

Since you're are such impressively unimpressed, maybe you could tell us what the objectives of this test were and how SpaceX failed to achieve them?

No, I'm commenting on Elon Musk's comment. I would have taken a humble: "Sorry, we blew up this (literally), but we're gonna get better after learning from our mistake." You blow things up, you fail. No excuses, sorry! Don't start a test knowing it most certainly will fail! Start a test when you're a lot more confident!

But it didn’t fail, landing wasn’t a test objective.

Landing may not have been one of the primary objectives, but it clearly was an objective, since they didn't just let it crash at will after having completed the flight maneuvers.

> Don't start a test knowing it most certainly will fail! Start a test when you're a lot more confident!

Good thing you aren't the one designing these things.

The purpose of a test is to gather the data you need to improve. Scientists and engineers constantly start tests knowing it will most likely "fail". Actual failure would be a test that didn't produce any data.

No, an explosion and destruction of equipment is a failure. Why do you keep dancing around this truth? It is a failure everywhere else in the world! You should've listened to the commentary of foreign anchors! They don't look at it as a success!

They are currently building like 10 other Starship prototypes.

Even if this model had worked perfectly, it's unlikely they ever would've flown it again. Remember, SpaceX's goal is to mass produce these.

Much better to fail hard now and in many different ways, as the more problems you can find with your expendable prototypes the better (before your manufacturing process starts to ossify).

Things are successfully tested to destruction every day. That is the truth. Some of the most successful tests involve explosions and loss of equipment; it's the only way to be sure where the limits are.

News anchors? They're often pretty clueless about technical things, so their mistakes don't mean much.

I'm very confused by your take here... SpaceX made it clear beforehand that the primary objectives of this test were the ascent and descent, not the landing. The "belly flop" horizontal descent is entirely new, and something they have only modeled so far.

The primary objectives of the test were completed without any issues. If that doesn't meet your definition of success, I don't know what to tell you...

You're right that the destruction of equipment and loss of material resources in general is a failure that could have been prevented with a more cautious approach, but that's not the entire picture. Another resource, one that Mr. Musk cares more about, is time. He both affords to and is willing to throw material resources at the technology development problem, even at a loss (i.e. the stated "1/3 chance of success"), just so he could "move fast and break things" (from time to time, if it comes to that). In this context, even with that prodigious blow at the end, having at least some of the testing goals met can be a genuine success in Musk's own eyes, not just a sell for the rest of us. The political actors, on the other hand, had to play the perception image game because that's the world they reside in. The politicians' primary objective in that space program has always been the prestige that comes attached from bearing it to fruits. Sure, in a race they went on in a similar fashion and threw copious amounts of material resources, but for them the cost of technical failures were not so much about loosing material as were about loosing face, hence the propaganda that had to handle/mitigate that. Same rocket science development game, different rules for the actors involved.

Reply on news.ycombinator.com