totm nov 2023 SpaceX Discussion Thread

[StrandedonEarth]

I don’t think I saw this posted here; finally a cause for the last failed landing:

 

[RyanRising]

33 minutes ago, StrandedonEarth said:

I don’t think I saw this posted here; finally a cause for the last failed landing:

Sounds like the engine failed on ascent, not descent, yet they still tried to use it for the landing burn. I'd have thought you'd switch to a different set of three, but hey, maybe they thought the issue had passed. Clearly it had not.

[tater]

SpaceX link shows Starlink tomorrow:

 

[tater]

 

[sevenperforce]

19 hours ago, RyanRising said:

Sounds like the engine failed on ascent, not descent, yet they still tried to use it for the landing burn. I'd have thought you'd switch to a different set of three, but hey, maybe they thought the issue had passed. Clearly it had not.

All nine engines are lit on the pad using TEA-TEB plumbed from the ground. Only three specific engines are plumbed to the onboard reservoir of TEA-TEB needed for in-flight restart. When the first Falcon Heavy launch failed to land the core, it was because the onboard TEA-TEB reservoir ran dry just after restarting the core engine for the second time, meaning there was not enough remaining for the other two engines to relight. They fixed this in subsequent Falcon Heavy flights by making the core's reservoir bigger.

So this is the problem. If Falcon 9 had lost one of the other four six engines on ascent, it would have been okay, but since it lost one of the intended landing engines, it was stuck. It can't choose other engines to relight because they aren't plumbed with TEA-TEB. I'm guessing they went ahead and tried to force a restart of the engine that previously shut down and its burn-through got worse.

Starship, of course, won't have this problem because all of the Raptor engines use their own internal spark igniter and can relight at any time, independent of any other engine or system.

Edited March 2, 2021 by sevenperforce
because apparently I can't count

[RyanRising]

5 minutes ago, sevenperforce said:

All nine engines are lit on the pad using TEA-TEB plumbed from the ground. Only three specific engines are plumbed to the onboard reservoir of TEA-TEB needed for in-flight restart. When the first Falcon Heavy launch failed to land the core, it was because the onboard TEA-TEB reservoir ran dry just after restarting the core engine for the second time, meaning there was not enough remaining for the other two engines to relight. They fixed this in subsequent Falcon Heavy flights by making the core's reservoir bigger.

So this is the problem. If Falcon 9 had lost one of the other four engines on ascent, it would have been okay, but since it lost one of the intended landing engines, it was stuck. It can't choose other engines to relight because they aren't plumbed with TEA-TEB. I'm guessing they went ahead and tried to force a restart of the engine that previously shut down and its burn-through got worse.

Starship, of course, won't have this problem because all of the Raptor engines use their own internal spark igniter and can relight at any time, independent of any other engine or system.

Thanks! I wasn’t ware they’d only plumbed the intended landing engines for restart capability, thought it was universal and they could light up whatever they liked. Makes a lot more sense to try restarting the failed engine in that case.

[sevenperforce]

Just now, RyanRising said:

Thanks! I wasn’t ware they’d only plumbed the intended landing engines for restart capability, thought it was universal and they could light up whatever they liked. Makes a lot more sense to try restarting the failed engine in that case.

Yep. And with lower thrust due to the engine-out on ascent, the ascent burn probably took longer and cost a little more propellant, leaving it with even lower margins. Otherwise perhaps it would have attempted a single-engine entry burn.

I just realized that given the core landing failure on FH1, the tipover on FH2, the core landing failure on FH3, and the intentional expenditure of the core for the upcoming FH4, it will be NET October of this year before we even have a chance to see a recovered FH core. That one is 6.35 tonnes to GTO so it should have no trouble reusing all three cores just as it did for FH2. 

[RCgothic]

The last time they had an engine out on ascent it caused them to overshoot the droneship. Less thrust means longer to orbit means further downrange. I'm surprised they could still attempt a landing at all, the engine out must have been very late in the first stage burn.

[tater]

 

[sevenperforce]

7 hours ago, RCgothic said:

The last time they had an engine out on ascent it caused them to overshoot the droneship. Less thrust means longer to orbit means further downrange. I'm surprised they could still attempt a landing at all, the engine out must have been very late in the first stage burn.

Yep. They were probably already throttling down the first-stage engines at that point so shutting down the bad one and throttling the others back up a little was likely easy enough. If this had happened earlier in the first-stage burn, they likely wouldn't have had enough propellant to even attempt an entry burn.

With an engine-out in a nine-engine cluster in KSP, you'd probably need to shut down the opposite engine to balance thrust (even with Vectors) but I'm guessing Falcon 9 is able to keep eight engines burning since its TVC PID programming is much smarter than the SAS in KSP. 

[Lukaszenko]

20 hours ago, StrandedonEarth said:

I don’t think I saw this posted here; finally a cause for the last failed landing:

 

Good explanation, but it leaves more questions than answers. For example, how did they find out about this "hole" if they lost the booster? Did they know about it BEFORE the launch? 

[cubinator]

16 minutes ago, Lukaszenko said:

Good explanation, but it leaves more questions than answers. For example, how did they find out about this "hole" if they lost the booster? Did they know about it BEFORE the launch? 

They likely have pressure sensors inside the engines. If one part was supposed to have a high pressure, and then suddenly shifted to ambient or low pressure, that would indicate that a hole has developed. This data would be part of the telemetry transmitted during flight.

[StrandedonEarth]

Yeah, I still think that falls into the “we were dumb” category of “don’t test life-leader components on the non-redundant landing engines “

Edited March 2, 2021 by StrandedonEarth

[mikegarrison]

That whole explanation is confusing. It sounds like they lost the engine during ascent, because he's talking about "we still got to orbit". But can't they use a working engine for landing? Or are there specially located engines that must be used?

[tater]

3 minutes ago, mikegarrison said:

That whole explanation is confusing. It sounds like they lost the engine during ascent, because he's talking about "we still got to orbit". But can't they use a working engine for landing? Or are there specially located engines that must be used?

Yeah, they only use 3 for entry burn/landing. The same 3 every time.

[mikegarrison]

1 hour ago, tater said:

Yeah, they only use 3 for entry burn/landing. The same 3 every time.

So that means that even though there are 9 engines, there is no redundancy for landing?

Edited March 2, 2021 by mikegarrison

[CatastrophicFailure]

12 minutes ago, mikegarrison said:

So that means that even though there are 9 engines, there is no redundancy for landing?

Here:

21 hours ago, sevenperforce said:

All nine engines are lit on the pad using TEA-TEB plumbed from the ground. Only three specific engines are plumbed to the onboard reservoir of TEA-TEB needed for in-flight restart. When the first Falcon Heavy launch failed to land the core, it was because the onboard TEA-TEB reservoir ran dry just after restarting the core engine for the second time, meaning there was not enough remaining for the other two engines to relight. They fixed this in subsequent Falcon Heavy flights by making the core's reservoir bigger.

So this is the problem. If Falcon 9 had lost one of the other four six engines on ascent, it would have been okay, but since it lost one of the intended landing engines, it was stuck. It can't choose other engines to relight because they aren't plumbed with TEA-TEB. I'm guessing they went ahead and tried to force a restart of the engine that previously shut down and its burn-through got worse.

Starship, of course, won't have this problem because all of the Raptor engines use their own internal spark igniter and can relight at any time, independent of any other engine or system.

 

[StrandedonEarth]

31 minutes ago, mikegarrison said:

So that means that even though there are 9 engines, there is no redundancy for landing?

Nope. 

They may want to reconsider trading a bit more complexity (plumbing another pair into the ignition fluid system) for landing redundancy, or not using life-leader parts on the critical landing engines. Or accept writing off more boosters due to preventable occurrences. 

Of course, hindsight is 20/20, as usual 

[tater]

1 hour ago, mikegarrison said:

So that means that even though there are 9 engines, there is no redundancy for landing?

Only whatever is afforded by by using some fraction of those 3.

So far this the first failure they've had along those lines with Merlin, though I think.

[RealKerbal3x]

 

[CatastrophicFailure]

30 minutes ago, tater said:

Only whatever is afforded by by using some fraction of those 3.

So far this the first failure they've had along those lines with Merlin, though I think.

Added complexity (of more plumbing) means added weight, added cost, and potentially added failure modes. There’s something to be said from simply figuring out, through trial and error, how long X part can go before it need thorough inspection or just preemptive replacement. As this was a life leader part, that sound largely like what they’re doing. They don’t want to invest too much effort here, as Starship is right around the pike, and is planned to make Falcon obsolete. The question is, where does that balance point come?

[Brotoro]

I'm confused. I was under the impression that a hole in the engine boot allowed hot gasses to enter during reentry, which managed to kill the engine partway through its landing burn. 

*entry burn

[sevenperforce]

1 hour ago, tater said:

Only whatever is afforded by by using some fraction of those 3.

So far this the first failure they've had along those lines with Merlin, though I think.

Correct me if I'm wrong, but I would think that successful restart of each engine is a single failure point for Falcon 9 recovery regardless of mission profile. If they are doing a three-engine burn it's because they don't have enough margins for a single-engine burn, so a failure of either outboard engine means they don't have the thrust they need. If they are doing a single-engine burn then they only restart the center engine, and if it fails to restart they're screwed anyway.

The reason they always light the center engine first is to avoid bad juju during startup. This isn't like KSP where the engine roars to 100% the instant you stage; they have to spin up the turbopump and then start pouring in TEA-TEB and wait for a full, good ignition. There are timing variations (which, again, is why they lost the first Falcon Heavy core when the center engine took too long to start and they ran out of TEA-TEB for the other two engines).

If they tried to light all three engines at once (or, worse, tried to light only the two outboard engines), then one of the outboard engines could reach full throttle before the other engine(s) had ignited, producing a massive amount of off-center thrust. Lighting the center engine first ensures that they have gimbal control alive and operating during startup transients.

9 minutes ago, Brotoro said:

I'm confused. I was under the impression that a hole in the engine boot allowed hot gasses to enter during reentry, which managed to kill the engine partway through its landing burn. 

*entry burn

The impression I received from the explanation was that hot gas recirculation during the ascent caused a burn-through of the engine boot and triggered a commanded shutdown of that engine, resulting in the other 8 engines completing the ascent at higher throttle. When they attempted the restart for the entry burn, the damaged boot caused a thrust shortfall on that engine which was ultimately unrecoverable.

Edited March 2, 2021 by sevenperforce

[RCgothic]

6 of the 9 engines have their TEA/TEB igniters plumbed to the pad, they can't be air-started.

The other 3 engines share an onboard reservoir. I think the centre engine must ignite for a safe landing. CofG is too low to safely land on one of the outer engines with gimbal. There *may* be some redundancy in the other 2 if there's enough propellant margin (bearing in mind a failure reduces the nominal margin) and the failure is identified really enough to swap to plan B.

[sevenperforce]

5 minutes ago, RCgothic said:

6 of the 9 engines have their TEA/TEB igniters plumbed to the pad, they can't be air-started.

Aren't all nine lit together from pad plumbing, to make the onboard TEA-TEB reservoir smaller?

5 minutes ago, RCgothic said:

There *may* be some redundancy in the other 2 if there's enough propellant margin (bearing in mind a failure reduces the nominal margin) and the failure is identified really enough to swap to plan B.

They've never lit two outboard engines alone without first lighting the center engine to give them gimbal control during startup transients. I wonder if they wrote the code to attempt a two-engine relight and hoverslam if the center engine fails to restart for a single-engine landing burn. That's the only time they'd have sufficient propellant margin.