I saw from the graphic that not all engines were lit at launch and then a few went out during the launch only to come back on again. Anyone have an explanation?
Bad health & status (H&S) instrumentation or telemetry. The six engines are clearly out (in a couple of cases you see blowouts) and the usage bars show excessive LOx consumption, but somehow the LCH4 goes down and then jumps back up.
Stranger
Really? In the video, five to eight engines (not symmetrically located) were unlit, as of about a minute into the flight (i.e., a minute before it ended). Real failures — not just telemetry failures, surely?
ETA: I understand better now. Yes, truly out; but they never came back on (that was the weird telemetry).
All good. Carry on!
Great launch, despite the idiot headlines on CNN, etc.
At least six engines out; maybe a couple more. Although the Raptor is qualified for multiple restarts, you generally wouldn’t try to restart an engine that was cut out for violating some safety or performance limit.
Well, aside from the multiple engines out, there was a failure to separate, apparent loss of oxidizer, and at least three other potential problems that I spotted. I understand that the party line is that a ‘successful test’ would be the vehicle clearing the tower, which it certainly did (albeit with already three engines out) but most operators usually aim for something beyond that, i.e. at least getting to an orbital altitude, before declaring success. Failure is a constant in this business, of course, and you learn more from a single failure than a dozen successes but it seems like SpaceX still has quite a way to go before declaring Starship and Super Heavy as success even as an LEO delivery platform, much less their more lofty aspirations.
Stranger
Did it begin to cart wheel because of engine asymmetry? As large as the diameter of the engine cluster, it would seem to me that the rocket would turn towards the dead engines.
ISWYDT
Enlighten us! 
Using lofty both literally and figuratively.
I’m only inferring from what I see on the video and I don’t want to engage in errant speculation in advance of SpaceX doing their own failure investigation. I will say that I think the omission of a deluge system for a system that emits the level of acoustic power and amplification of 33 Raptor engines was ill-advised although how much it may have contributed to the problems that the vehicle experienced is unclear.
Stranger
Could have, although I was under the impression that part of the draw of having so many smaller engines was precisely so a lot of them could fail, and still achieve the mission.
I’d also have expected that with so many, that there would be some sort of differential thrust from the nearby engines and/or throttling back the far side ones to keep the rocket stable.
It was kind of funny; I’ve played a lot of Kerbal Space Program (a rocket-designing and flying video game), and that cartwheel is exactly what happens when the rocket thrust gets off-axis past a certain degree. I was watching the replay, and noticed it starting to yaw, and thought “Uh oh… in KSP, this is when it starts to tumble”, and sure enough, it did.
@muldoonthief and @Stranger_On_A_Train, thank you both.
In the space of an hour, I’ve gone from “oh, no!!” (initial headlines) to “oh, yay!” (SpaceX’s own video feed, and several well-informed posts here) to “let’s call it 50-50” (Stranger’s elaborations).
If you compared this launch to the second launch of the Saturn 5, aka Apollo 6, it looks pretty reasonable. That launch didn’t end so spectacularly, but presented some interesting failures, some unforeseen, some stupid, and some uncannily similar.
Nobody is riding the next Starship to the moon, but they did ride the next Saturn 5.
Overall I suspect SpaceX fell pretty pleased. It it had worked out if the box Elon would probably say they were not pushing things hard enough. Whether this is reasonable is another matter but it does underline the development philosophy.
NASA is obviously more careful before they launch anything.
True, but boy did things happen fast in the late 60s:
“Despite tension between the Apollo 7 crew and ground controllers, the mission was a complete technical success, giving NASA the confidence to send Apollo 8 into orbit around the Moon two months later.”
Hmm. I thought it was supposed to cartwheel at least partway around once, so that the second stage could separate from the first? I’m sure I read something about them wanting to try a super-simple stage separation system that didn’t rely on a mechanism pushing the stages apart?
It’s much gentler than a cartwheel, and only needs to spin a fraction of a revolution before separating. But yes, they do intend to use centrifugal force to separate the two stages.
The booster stage will need to continue rotating about 90 degrees back, to get into the right orientation for the boostback burn (so as to return to the launch site). But the second stage should continue more or less on its original trajectory.
You have to remember that this is an unfinished vehicle. It’s the Minimum Viable Rocket. SpaceX was pretty adamant that Starship would never have survived re-entry anyway.
But they managed to launch the largest rocket in history, got it through max-Q, and one consequence of the failure is that the booster and Starship didn’t even come apart when tumbling, and had to be detonated. I’d say the choice to use 301 stainless was a good one. They may be aboe to shave some weight off the rocket even.
I do worry that what we saw was a fuel pressurization/delivery issue, and they couldn’t keep all the engines lit. That might be mhy more dropped off over time, if the pressure dropped during flight. Maybe it went out of control merely because it lost too many gymballed engines to maintain stability. I’d rather it was damage from FOD - that’s easier to fix.
Still, they managed to make that 390 ft, 11 million pound skyscraper fly, almost to stage separation. And collected a ton of priceless data.
But man, there are still a lot of challenges ahead. And design/build/test cycles will get longer as the vehicle gets built out for more complex tests like orbital refueling. This program will take years before it’s fully operational. Maybe quite a few years if things get explodey or they run into hard problems.
It would be… interesting? unfortunate?.. if the root cause of all anomalies turned out to be the inadequate engineering of the launchpad, that nothing was inherently wrong with Super Heavy or Starship themselves.
Launch segment engineering can sometimes get overlooked as less important.
I have seen it opined in several places that some of those engine failures were possibly due to foreign object damage from the concrete under the pad getting demolished. IIRC there was some other damage to nearby support structures from flying concrete chunks. The pics of the damage are pretty impressive.
There is a video of a car very far away from the launch site (as in deemed a safe distance for people to watch from) that had its windows smashed from debris.
Those rockets did a number on that launch pad.