SpaceX FH : Falcon Heavy Demo : Feb 6, 2018 : Discussion Thread 2

[cscott]

https://twitter.com/elonmusk/status/963107229523038211

Elon Musk
‏Compte certifié @elonmusk
16 minil y a 16 minutes
En réponse à @kerrbones @nextspaceflight

Not enough ignition fluid to light the outer two engines after several three engine relights. Fix is pretty obvious.

The centre core landing failure was caused by the fact that the need for ignition fluids (TEA/TEB) increases after several engine relights. They are adding more storage to fix the issue.

So, basically, the side cores had as much of a chance to result in 2 high speed lawn darts as the center core?

I think Elon was simplifying.

I believe that the actual amount of TEA/TEB used depends on atmospheric and engine conditions at time of relight.  "Flow until a certain engine pressure is reached" has been suggested here on NSF, for example.

So it's not so much the 1-3-1 burns that did in the center core, so much as some of those burns used more TEA/TEB than expected.  That's why center succeeded and side boosters didn't, even though they had the same burn pattern.

It's also possible that center and side have differently-sized TEA/TEB reservoirs, since the two side cores were refurbished block 3/4 and changes were definitely made to the center core thrust structure.

Either way, the "fix is obvious", as Elon sez.

[Space Ghost 1962]

https://twitter.com/elonmusk/status/963107229523038211

Elon Musk
‏Compte certifié @elonmusk
16 minil y a 16 minutes
En réponse à @kerrbones @nextspaceflight

Not enough ignition fluid to light the outer two engines after several three engine relights. Fix is pretty obvious.

The centre core landing failure was caused by the fact that the need for ignition fluids (TEA/TEB) increases after several engine relights. They are adding more storage to fix the issue.

More critical about this.

One should always ask, if the fix was so simple, why wasn' t the issue determined before flight (i.e. simulation)?

People give SX too many "mulligans", too readily, at any time, and for any reason.

Yes it was good the demo launch succeeded.

[rsdavis9]

https://twitter.com/elonmusk/status/963107229523038211

Elon Musk
‏Compte certifié @elonmusk
16 minil y a 16 minutes
En réponse à @kerrbones @nextspaceflight

Not enough ignition fluid to light the outer two engines after several three engine relights. Fix is pretty obvious.

The centre core landing failure was caused by the fact that the need for ignition fluids (TEA/TEB) increases after several engine relights. They are adding more storage to fix the issue.

More critical about this.

One should always ask, if the fix was so simple, why wasn' t the issue determined before flight (i.e. simulation)?

People give SX too many "mulligans", too readily, at any time, and for any reason.

Yes it was good the demo launch succeeded.

except that recovery of boosters is experimental and spacex can experiment as close to the edge as they want. Because nobody else recovers anything.

[1]

That's why center succeeded and side boosters didn't,

Unless I misunderstood what you're saying, I think you have this backwards.

[Elmar Moelzer]

The performance numbers in this database are not accurate. In process of being fixed. Even if they were, a fully expendable Falcon Heavy, which far exceeds the performance of a Delta IV Heavy, is $150M, compared to over $400M for Delta IV Heavy.

https://twitter.com/elonmusk/status/963076231921938432

So we now have an actual price for an expendable Falcon Heavy.

And we have a price for the partially expendable FH (and an estimate for the payload penalty):
https://twitter.com/elonmusk/status/963094533830426624

Side boosters landing on droneships & center expended is only ~10% performance penalty vs fully expended. Cost is only slightly higher than an expended F9, so around $95M.

[Mongo62]

Side boosters landing on droneships & center expended is only ~10% performance penalty vs fully expended. Cost is only slightly higher than an expended F9, so around $95M.

So price to recover two side boosters and expend the center core is only $5 million over the price to expend a single-stick F9?

That does not sound right! Maybe Elon was speaking off the cuff again? Otherwise a FH with all three cores recovered should surely have a price under $10 million.

[JH]

People give SX too many "mulligans", too readily, at any time, and for any reason.

Because people like results and SpaceX provides them.

[jpo234]

That's why center succeeded and side boosters didn't,

Unless I misunderstood what you're saying, I think you have this backwards.

I think he meant engines instead of boosters.
The center engine started but the outer engines failed.

[LouScheffer]

https://twitter.com/elonmusk/status/963107229523038211

Elon Musk
‏Compte certifié @elonmusk
16 minil y a 16 minutes
En réponse à @kerrbones @nextspaceflight

Not enough ignition fluid to light the outer two engines after several three engine relights. Fix is pretty obvious.

The centre core landing failure was caused by the fact that the need for ignition fluids (TEA/TEB) increases after several engine relights. They are adding more storage to fix the issue.

More critical about this.

One should always ask, if the fix was so simple, why wasn' t the issue determined before flight (i.e. simulation)?

People give SX too many "mulligans", too readily, at any time, and for any reason.

I read this the opposite.   Starting rocket engines while facing into a hypersonic wind is not remotely a solved problem.   Probably no wind tunnel in the world can do this at scale, and with the rocket bells facing into the wind, it's going to be enormously turbulent, which combined wit combustion physics is going to make this impossible to simulate accurately.  It needs experiment, and SpaceX has all the experimental data that exists on this problem.

Suggest that SpaceX had an amount they believed necessary, with margin.  But an untried corner of the envelope (they have never started boostback at this speed, for example) led to a situation where prior starts of the main core took more TEA/TEB than expected.  So they ran out.

If anyone benefits by "mulligans", it's the engineering departments of other rocket companies.  Everyone here assumes, without evidence, that their plans would work flawlessly if only they could get management to approve them.  I'd bet each of them have at least one "oops" moment in them, but without experiment we'll never know.  At least SpaceX we know.

[LouScheffer]

Side boosters landing on droneships & center expended is only ~10% performance penalty vs fully expended. Cost is only slightly higher than an expended F9, so around $95M.

So price to recover two side boosters and expend the center core is only $5 million over the price to expend a single-stick F9?

That does not sound right! Maybe Elon was speaking off the cuff again? Otherwise a FH with all three cores recovered should surely have a price under $10 million.

No, this makes sense.  Ib both cases SpaceX expends one core and one second stage.    The only difference is the two side cores, which are launched, recovered, and refurbished.   So if they can launch and refurbish the side cores for less than $2.5 million each, that explains the difference.   It's an aggressive goal, but given that they want fast turnaround (24 hours) it's consistent with their thinking.   Of course it remains to be seen if they can really use and re-use a core that cheaply, but it's not inherently crazy.

[AllenB]

More critical about this.

One should always ask, if the fix was so simple, why wasn' t the issue determined before flight (i.e. simulation)?

People give SX too many "mulligans", too readily, at any time, and for any reason.

Yes it was good the demo launch succeeded.

Given infinite time for study, simulation, and other preparation on the ground, it's possible to ensure a flight which never happens goes perfectly.

IMHO this is a great example of Musk/SpaceX following the "Silicon Valley philosophy". Try something, throw it out there, and see if it works. You'll quickly know the answer, with a generous side-helping of useful data for the next try. Obviously not a useful approach with crew on top but hardly an issue if the occasional landing goes amiss.

Granted, nobody gets it right all the time... destroying Amos-6 during testing was unwise, but they've clearly learned from it.

People are excited about SpaceX because they've already delivered progress in ways that nobody else has. The status quo hasn't been exciting for decades now. That will buy a lot of slack from a lot of people.

[Elmar Moelzer]

Side boosters landing on droneships & center expended is only ~10% performance penalty vs fully expended. Cost is only slightly higher than an expended F9, so around $95M.

So price to recover two side boosters and expend the center core is only $5 million over the price to expend a single-stick F9?

That does not sound right! Maybe Elon was speaking off the cuff again? Otherwise a FH with all three cores recovered should surely have a price under $10 million.

No, this makes sense.  Ib both cases SpaceX expends one core and one second stage.    The only difference is the two side cores, which are launched, recovered, and refurbished.   So if they can launch and refurbish the side cores for less than $2.5 million each, that explains the difference.   It's an aggressive goal, but given that they want fast turnaround (24 hours) it's consistent with their thinking.   Of course it remains to be seen if they can really use and re-use a core that cheaply, but it's not inherently crazy.

Exactly!
Also worth mentioning is that the price point for F9 is going to be lowered as well, as time goes by and Block 5 has had time to establish itself. So there will probably still be a slightly higher price difference between expended F9 and partially expended FH.

[1]

Because nobody else recovers anything.

I think this is more what SpaceGhost is talking about.

This launch was a great success; he acknowledges such himself. And perhaps the fix really is as simple as adding more blinker fluid. However at some point in the not-so-distant future, probably once block 5 is regularly flying, a very good argument can be made that a failed droneship landing should count as a partial mission failure. Sure, the customer is happy, but if reusability is part of SpaceXs business model, they need to execute on that. As such, what any other entity is doing, or not doing, is irrelevant.

I think most of us are ok with mulligans on tests, which this flight very much was, and I agree with the various folks who don't see a big issue in regards to this particular flight. But I also agree with Ghost that, in general, more "why"s should be asked.

[CameronD]

I think most of us are ok with mulligans on tests, which this flight very much was, and I agree with the various folks who don't see a big issue in regards to this particular flight. But I also agree with Ghost that, in general, more "why"s should be asked.

As you say, this was a test.. Elon is even on record as saying so much could go wrong he would have been happy if it actually cleared the tower!  In many ways I think RocketLab's idea of naming their flights "Still Testing", etc. has much merit since it is a constant reminder to journos (in particular) that if the test doesn't work, that is what is expected - it is, after all, a test.

In some ways, from an engineering perspective, this FH launch was a failure: so many things that could have gone wrong, didn't - and it would certainly be a shame if the only "lesson learned" was that they miscalculated the amount of starter fluid needed.  Is the rocket over-engineered?  What else is lurking in the background (refer the F9 COPV issue) waiting to bite next launch??  No one really knows.. and that's not necessarily a good thing.  I'm sure they will be a lot of FH engineers going back over their calculations between now and next launch!
 

[1]

In some ways, from an engineering perspective, this FH launch was a failure: so many things that could have gone wrong, didn't - and it would certainly be a shame if the only "lesson learned" was that they miscalculated the amount of starter fluid needed.  Is the rocket over-engineered?
 

Yes, it's a shame so many things worked properly that the engineers won't know what to fix. 

But judging from Elon's comment that FH was much harder than they expected, I'd guess they learned quite a lot in the design process about coupled loads, booster separation dynamics, modal analysis, etc, etc. And apparently they learned most, or all, of those lessons well.

I'd guess that there's still plenty to fix; just nothing so broken that it resulted in a loss of mission. More fine tuning; less coarse.

[Rocket Science]

https://twitter.com/elonmusk/status/963107229523038211

Elon Musk
‏Compte certifié @elonmusk
16 minil y a 16 minutes
En réponse à @kerrbones @nextspaceflight

Not enough ignition fluid to light the outer two engines after several three engine relights. Fix is pretty obvious.

The centre core landing failure was caused by the fact that the need for ignition fluids (TEA/TEB) increases after several engine relights. They are adding more storage to fix the issue.

More critical about this.

One should always ask, if the fix was so simple, why wasn' t the issue determined before flight (i.e. simulation)?

People give SX too many "mulligans", too readily, at any time, and for any reason.

Yes it was good the demo launch succeeded.

I don't necessarily disagree with you, but perhaps some will give them a pass because they're actually developing new concepts.  They don't see a failure as necessarily a bad thing, they don't fear it... Pick-up the pieces , learn from it and fly again ASAP...

[OxCartMark]

Starting rocket engines while facing into a hypersonic wind is not remotely a solved problem. ...<snip>...  it's going to be enormously turbulent, which combined wit combustion physics is going to make this impossible to simulate accurately.  It needs experiment, and SpaceX has all the experimental data that exists on this problem.

Suggest that SpaceX had an amount they believed necessary, with margin.  But an untried corner of the envelope (they have never started boostback at this speed, for example) led to a situation where prior starts of the main core took more TEA/TEB than expected.  So they ran out.

I submit for consideration an alternate theory on why they used more fluid than they needed.  Its based on Lou's theory but mine happens on the way up rather than the way down.  Perhaps on the way up rather than just throttling down the center core (which is visible) they also switched off a few engines and restarted them.  It probably wouldn't be visible if it were the three engines in line with the side boosters. It makes sense that future  launches are going to be more efficient with fewer engines running at full thrust than with all engines on but throttled so that capability would be needed to get the full potential of FH going forward, even if the restart capability was only needed to replace engines that failed in flight.  And since this 3 stick launch had what appears to my eyes astoundingly underwhelming performance (the 3 stick launch of a ?1500kg? payload only accelerated the second stage to the same speed that some previous F9(s) have pushed very heavy payloads to) I'm lead to think that they were doing a lot of experimenting with what was going on behind that curtain of flame.  Multiple restarts?  Really deep throttling experiments?  If I'm right that engines were turned off and restarted as experiments during launch and if as usual only three engines carry sparky fluid then what Lou says comes into play - the environment they were trying to start the engines in was exceedingly chaotic and there was no previous test data (its probably been tried at McGregor but if so those tests didn't include the full vibration and G environment of a real launch) to tell how difficult and fluid draining the restart would be.  Perhaps that's where the stuff went.  I ran out of windshield washer fluid this morning and had to stop at a Shell station to get more so I completely understand.

Also, I don't discount the theory someone stated above that it may have been a simple communication failure on how many restarts were to be done and how much fluid needed to be loaded.

[atsf90east]

On the other side, Buzz Aldrin went to the pad and watched the liftoff of Falcon Heavy.

Jeez... what a way to go!  RIP, Buzz... 

He had a hot time on the way out   Seriously though, I'll bet he was either at the Saturn V Center viewing area, or on the roof of the VAB

[rsnellenberger]

https://twitter.com/elonmusk/status/963107229523038211

Elon Musk
‏Compte certifié @elonmusk
16 minil y a 16 minutes
En réponse à @kerrbones @nextspaceflight

Not enough ignition fluid to light the outer two engines after several three engine relights. Fix is pretty obvious.

The centre core landing failure was caused by the fact that the need for ignition fluids (TEA/TEB) increases after several engine relights. They are adding more storage to fix the issue.

More critical about this.

One should always ask, if the fix was so simple, why wasn' t the issue determined before flight (i.e. simulation)?

People give SX too many "mulligans", too readily, at any time, and for any reason.

I read this the opposite.   Starting rocket engines while facing into a hypersonic wind is not remotely a solved problem.   Probably no wind tunnel in the world can do this at scale, and with the rocket bells facing into the wind, it's going to be enormously turbulent, which combined wit combustion physics is going to make this impossible to simulate accurately.  It needs experiment, and SpaceX has all the experimental data that exists on this problem.

On the contrary (added: re: regarding turbulence & combustion) -- although the flow field outside the nozzle is highly turbulent, conditions within the nozzle, throat, and combustion chambers should be relatively static (behind the shock front) for both the re-entry and landing burns.  The question that that they answered in the first entry burn is whether the Merlin turbo pump & main exhausts can penetrate through the supersonic shockwave well enough to sustain stable combustion in the engine.

(edited: added clarification on the point that I was contrarying)

[IainMcClatchie]

I submit for consideration an alternate theory on why they used more fluid than they needed.  Its based on Lou's theory but mine happens on the way up rather than the way down.  Perhaps on the way up rather than just throttling down the center core (which is visible) they also switched off a few engines and restarted them.

Nice theory.

During the reentry burn 3-engine portion, you can see an "Eye of Sauron" (got that from Lars Blackmore's talk) effect, in which the longer axis of the plume cross section is orthogonal to the line of the three lit engines.  Lars claims that SpaceX doesn't really know why that happens.  I didn't see any equivalent effect on the way up.  I wouldn't expect it near sea level with the exhaust overexpanded.  But once the plume bloomed out, I thought I'd see the effect again, and didn't.  Maybe that's because there was a different pressure pattern.

And since this 3 stick launch had what appears to my eyes astoundingly underwhelming performance (the 3 stick launch of a ?1500kg? payload only accelerated the second stage to the same speed that some previous F9(s) have pushed very heavy payloads to) I'm lead to think that they were doing a lot of experimenting with what was going on behind that curtain of flame.

Agree!  Even given a bunch of experimenting, how did they manage to push so little so slowly?

And why?  It would have been cooler to have enough battery lifetime and impulse to capture the view of doing a gravity assist off the moon.