sevenperforce

It's hard to tell for sure, exactly. My guess is still that a good bit of unburned liquid methane was released at engine shutdown, spraying the inside of the skirt. Part of it went up in flames immediately but flamed out due to either fuel exhaustion or lack of oxygen. However, the skirt is still an extraordinarily volatile environment and so the remaining residues on the right-hand side ignited spontaneously a few seconds later. I feel like there's really nothing flammable in that skirt. (so much innuendo) Speaking of which, I wonder exactly how long you would live if you tried to ride inside the skirt. Like, obviously you would die pretty fast, but...how fast? Really remarkable. As amazing as it was to watch the camera view live, it must have been an entirely different experience to watch it in person. From the wide-angle view, it REALLY looks like it's never going to actually flip -- that it's going to drop straight down and pancake. I'm sure that was a nail-biter. My most educated guess is that pressurizing the CH4 header tank is going to be more of a challenge than they thought. The CH4 header tank is inside the CH4 main tank, while the LOX header tank is independent of the LOX main tank. The CH4 flow path goes through the header at all times. Presumably, then, they close valves between the CH4 header and main at engine shutdown to ensure that it remains full of propellant at engine restart. The simplest solution at restart would be to reopen those valves immediately after ignition, allowing the autogenous pressurant in the main tank to provide pressure to the header: In the diagrams above (which are obviously not to scale but aren't THAT bad), you can see the thin pressurant lines running from the engines to the tanks. The methane pressurant line runs straight to the main fuel tank, while the GOX line splits to pressurize both the main LOX tank and the header LOX tank (not shown). Yesterday was the first time they ever did an in-air restart. Elon said outright "fuel header tank pressure was low during landing burn" so my guess is that the pressure in the main tank dropped faster than it was able to be replenished by the pressurant line. The easiest way to remedy this would be to split the methane pressurant line, just like the GOX pressurant line, and keep those valves closed between the fuel header tank and the fuel main tank. I referenced this yesterday just before launch, LOL. The test altitude was 12.5 km and it certainly followed the planned ascent properly so I'm guessing 12.5 km. Also I did a thing for the public who doesn't understand this:

It’s catastrophic weld failure. Remember that the tanks are pressed to a pretty significant extent with warm methane and warm oxygen. In fact, they are at a higher pressure than some pressure-fed rockets. So when the spaceship hits the ground, instant deformation of the steel skin results in that internal pressure busting straight through the skin like aluminum foil. The fireball was mostly the gases mixing and igniting, not the liquid. There was very little liquid propellant in the tanks at this point. WHY DON’T I HAVE ANY MORE REACTIONS TODAY

It didn’t catch fire; that was just hydrocarbon deposition undergoing combustion in a low-pressure atmosphere. My guess is that the legs have an experimentally-determined rating for impact velocity and the simplest way to trigger deployment is at some predetermined multiple of that velocity, like 3x. So the legs deploy when the ship reaches a velocity which is less than 3x the maximum survivable impact velocity.

Yep. Oxygen-rich combustion became engine-rich combustion.

That was just a glut of unburned propellant igniting in air. The skirt can handle the heat.

Yes, correct. I was curious about what their flight profile would be, and how they would get a "failure over water" if something went wrong. What they did was actually quite inventive. They flew out over the water, pitching over upside down like the Shuttle. They maintained heading on two engines. Then, they reduced to a single engine which allowed them to fly at a high AoA, basically hovering sideways. They used this engine for a tiny kick at shutdown to help tip back over in the opposite direction and then used the skydive to fly back. Roll and pitch control with those flaps was awe-inspiring. If you play at reduced speed you see that they never did extend the legs. Clearly the leg-extension was supposed to happen a few moments later.

Rewatching for the fourth time at 25% speed. Raptor SN42 is never ignited; just the other two. They provide the entire kick-flip torque. At T+6:32, a few green flashes appear on the left-hand Raptor relative to SN42. By T+6:37, the exhaust from the left-hand Raptor is bright green: It appears to correct a moment later and becomes the familiar pink-purple again: One second later the green has returned and it's almost violent in its hue: The green disappears, and the other engine, on the right, is shut down, apparently because they were planning on a 2-1 landing burn: The green comes back, this time to stay, and becomes so bright that it fills up the entire skirt section. The length of the plume is noticeably shortened as copper residue fills the exhaust, which in turn is saturated with soot: So it looks like they planned a 2-1 landing burn, but the engine they chose to land with had an anomaly on restart that led to it eating itself from the inside out over a period of about 4 seconds. EDIT: I hadn't seen Elon's tweet when I posted this, yet. But that green flame and sooty exhaust is clearly off-nominal. If the fuel tank header pressure was low, then could that have caused insufficient flow in the preburners, resulting in an oxygen-rich burn-through?

Merlin uses TEA-TEB but Raptor does not -- it uses electric gas-gas torch igniters.

Ridiculously, wildly impressed. Everything worked. Methalox produces just water and CO2, nothing else. The only thing cleaner is hydrolox, which produces only water. Could have been that one engine failed and that was the thrust that they needed to land. Yes, after rewatching a few times that's my best guess as well. It can do a two-engine landing burn but I'm guessing they timed it for a three-engine landing burn.

If anyone wants to replay the flight only, here's a link to the terminal count:

That was the most exciting 6 minutes and 42 seconds of my life.

I don't even know that it lost an engine. I think they may have used a two-engine landing on purpose, to induce more torque.

Exactly. The kick-flip worked flawlessly. Agreed. Right up until the last second

And we have a successful lithobrake

KABLOOEY

HERE WE GO RESTART

OH MY GOD THE RUSH THE FLAPS WORK Even if she craters, we still have a flight test showing controlled descent using flaps, which was always the hardest part of this whole thing

LOOK AT THOSE FINS GO IT IS WORKING IT IS BLOODY WORKING

FINAL RAPTOR MECO AND SKYDIVE

Flying at extreme AOA now...you can tell from the way the vent gas is trailing.

I'm guessing it's a shutdown to reduce gee loading?? IT IS SO HIGH Down to one engine. Looks like the Shuttle on ascent. So beautiful.

ONE ENGINE SHUT DOWN

THE GROUND DROPPING AWAY THE ROAR THE FINS

MOTHER OF GOD

LIFTOFF!!!!!!!!! GO STARSHIP!!!!!!!!!!