SpaceX (general discussion)

Falcon 9's fairing half falling onto the deck of the recovery ship through the catching net:

LastCatch-1024x578.png


 
Space X will no longer attempt to catch the fairing halves instead in future they will just fish them out of the water for reuse.


 
Space X will no longer attempt to catch the fairing halves instead in future they will just fish them out of the water for reuse.



They probably handle water better than assumed.
 
I think they have over a dozen used boosters in inventory, but I assume there are rare clients who still want new builds and there are some missions where the booster gets sacrificed. Also I think the core of the heavy has never been successfully landed to date. But I can't remember the last launch that was a first use. Probably just slipped my notice.
One Heavy core did land successfully. Unfortunately, the Octagrabber had not yet been modified to work with a Heavy core and so it couldn’t be used to secure the booster, and rough seas later caused the booster to tip over and break. But it did at least land successfully.
 
They are churning them out just to be blown up like heretics thrown in the pyre during the Middle Ages...

Did they lost faith in sane engineering?
You mean like the kind that takes a decade to build a "new" rocket out of existing parts and not fly? That kind of "sane" engineering?

I don't know what the hell the SLS team's excuse is but it better be a doozy given the way that SpaceX is advancing in leaps and bounds. Very difficult not to be cynical.
SLS is being developed according to the Zeroth Law of aerospace development (Thou Shalt Get It Right The First Time) and its corollary (Thou Shalt Not Lose The Vehicle). It becomes a feedback loop—you have to get it right the first time, so you design extremely conservatively, study it to death, and test the snot out of everything at the component level. That drives cost through the roof. When the cost is then stupendously high, you really can’t afford to lose the vehicle (Congress, shareholders, the Board of Directors, and The Public see that as a waste of money), so you get even more conservative, study things more, and do even more testing, and the cost goes still higher.


SpaceX is taking a completely different approach. They’re betting that a “hardware-rich” development program, even after a couple dozen failures, will in the end cost less (even including the cost of lost hardware) and go faster than the traditional approach described above.

They’re able to keep the build cost down because they’ve eschewed another traditional aerospace mindset that says “use whatever large-scale equipment and cutting-edge technology it takes to build a super-optimized maximum-performance structure”. We see this all over, from small aircraft to new rockets, where all kinds of large, expensive, dedicated tools (autoclaves, jigs and fixtures, massive milling operations) and advanced techniques (friction-stir welding, automatic filament-winding, large CNC machines, etc) are used to make highly-optimized structures, and then companies try to squeeze cost out later.

SpaceX started at the other end, with little more than equipment you could find at any large construction site or well-equipped hobby garage, and tried to start building rockets. They’ve added more equipment only as they’ve determined what’s actually needed to get a good product. And even now on their prototypes, they’re pulling off-the-shelf components (Tesla motors and battery packs, for example) instead of outsourcing for brand-new dedicated items intended to be production-representative. The whole process reminds me of what any DIYer has probably experienced on a new project—you start with the tools you have, and buy more as you go along when you find out you need them.

Another factor in the “don’t lose the vehicle” mindset is that loss of the vehicle probably meant the loss of most of your data and possibly a life or three. SpaceX isn’t as worried about this—Starship prototypes are unmanned and they instrument the heck out of everything and collect all the data in real time. I’d be willing to bet a beer that a single Starship hop generates at least as much (non-video) data as the entire Saturn V flight program.

Finally, as others have pointed out, much of the data they need to get simply can’t be replicated subscale or simulated—and even if it could, the full-scale testing is probably still cheaper—plus, it’s getting them experience with manufacturing and ground handling on a vehicle this large.
 
@gtg947h : great post!

But let's not forget that the technology their are inspired from (large infrastructure projects and pressure tanks don't have much experience (and the willingness!*) in small scale experiments or how to go from the latter to full scale without an infinite trial and errors stages.

That, IMOHO, limits Musk choices and my idea was that they could go from now, temporarily, forward through a blended approach with a small scale demonstrator solving the belly flop manoeuvre.

*When you have dozens and dozens of contractors on site, the last thing that will emerge from your team is radical changes...
 
SLS is being developed according to the Zeroth Law of aerospace development (Thou Shalt Get It Right The First Time) and its corollary (Thou Shalt Not Lose The Vehicle). It becomes a feedback loop—you have to get it right the first time, so you design extremely conservatively, study it to death, and test the snot out of everything at the component level. That drives cost through the roof. When the cost is then stupendously high, you really can’t afford to lose the vehicle (Congress, shareholders, the Board of Directors, and The Public see that as a waste of money), so you get even more conservative, study things more, and do even more testing, and the cost goes still higher.


SpaceX is taking a completely different approach. They’re betting that a “hardware-rich” development program, even after a couple dozen failures, will in the end cost less (even including the cost of lost hardware) and go faster than the traditional approach described above.

<Snipped for brevity>

@gtg947h Good description of the legacy Western space launch players. The last two and a half years in the AF were in the space launch program office, so I got to see the risk adverse behaviors first hand in the endless pedigree reviews and what not. The only place I see things a bit different is with SLS where they kept the shuttle tech alive and well to support the NASA infrastructure and USA, a bit of a self licking ice cream cone...

Good description of the current state of BFR/ICT/MCT/SH&SS, there's a bit more history behind it and they didn't start hardware rich in the beginning. Round about 2002 a Silicon Valley billionaire with a battery powered Lotus wanted to some day get to Mars...five people and a Mariachi band later Falcon 1 development begins and looks pretty much like all of the small sat start ups look like, sorta replicate what the big guys are doing in miniature on a shoe string budget and pray that it all works. Three failures and two successes later with small rocket Falcon 1 they gained enough credibility with NASA to get just under a billion to develop Falcon 9.

Now, Falcon 9, at least in it's 2010 incarnation was still a get it right on the first try rocket. Nothing wrong with that, given the timeframe and prevailing industry culture. There's a lot to be said for over delivery/unexpected delivery opening the the path for innovation here, but I digress. What Elon did different here was introduce a design for manufacture mindset for Falcon 9, think Toyota Production System for space launch. Version 1.0 gave up a lot in performance to the Atlas V, but it had tons of room to grow and continued to build a track record of success which buys some patience for secondary objectives like reuse...speaking of which does anyone remember the parachute recovery into salt water method of reuse?

Later versions of Falcon 9 increased payload to orbit and proved out the rational (and business case) for reuse. The virtuous cycle of reuse and design for manufacture helped Falcon 9 corner the commercial launch market and provide spare cores to launch Starlink sats to raise funds for Mars. Speaking of which how many folks remember the carbon fiber versions of ICT/MCT and the building of some of the largest composite structures ever before moving to nice shiny 30X stainless?

All of this is to say, today's hardware rich approach never started that way. It was earned through nearly two decades of hard earned success when very few thought that success could be achieved. First principles won't get billions of capital unless there's some record of success to calm the VC's down when they see pictures of their capital blowing up several times a year...
 
On a tangent:
 
Space X will no longer attempt to catch the fairing halves instead in future they will just fish them out of the water for reuse.



They probably handle water better than assumed.
Or it's simply cheaper to refurbish them than try to recover them during launch.
 
i had some thought about the cost of SN
They are according Musk tweets mass around 150 short tons or 136 metric tons
mostly 304L steel that cost $1880/short tons
if we count 15 short tons for equipment, engine hydraulics and flight electronic
remain 121 short tons tank structure wings

cost estimation
121 short tons steel 304L = $253800 o_O
3 Raptors each $25 million ? = $75000000 (estimation)
1 Tesla Battery pact = $3500 dollar
x COPV tanks = $1000000 (estimation)
Flight electronic = $1000000 (estimation)

A SN Could cost around $77 million
 
They are churning them out just to be blown up like heretics thrown in the pyre during the Middle Ages...

Did they lost faith in sane engineering?
You mean like the kind that takes a decade to build a "new" rocket out of existing parts and not fly? That kind of "sane" engineering?

I don't know what the hell the SLS team's excuse is but it better be a doozy given the way that SpaceX is advancing in leaps and bounds. Very difficult not to be cynical.
SLS is being developed according to the Zeroth Law of aerospace development (Thou Shalt Get It Right The First Time) and its corollary (Thou Shalt Not Lose The Vehicle). It becomes a feedback loop—you have to get it right the first time, so you design extremely conservatively, study it to death, and test the snot out of everything at the component level. That drives cost through the roof. When the cost is then stupendously high, you really can’t afford to lose the vehicle (Congress, shareholders, the Board of Directors, and The Public see that as a waste of money), so you get even more conservative, study things more, and do even more testing, and the cost goes still higher.


SpaceX is taking a completely different approach. They’re betting that a “hardware-rich” development program, even after a couple dozen failures, will in the end cost less (even including the cost of lost hardware) and go faster than the traditional approach described above.

They’re able to keep the build cost down because they’ve eschewed another traditional aerospace mindset that says “use whatever large-scale equipment and cutting-edge technology it takes to build a super-optimized maximum-performance structure”. We see this all over, from small aircraft to new rockets, where all kinds of large, expensive, dedicated tools (autoclaves, jigs and fixtures, massive milling operations) and advanced techniques (friction-stir welding, automatic filament-winding, large CNC machines, etc) are used to make highly-optimized structures, and then companies try to squeeze cost out later.

SpaceX started at the other end, with little more than equipment you could find at any large construction site or well-equipped hobby garage, and tried to start building rockets. They’ve added more equipment only as they’ve determined what’s actually needed to get a good product. And even now on their prototypes, they’re pulling off-the-shelf components (Tesla motors and battery packs, for example) instead of outsourcing for brand-new dedicated items intended to be production-representative. The whole process reminds me of what any DIYer has probably experienced on a new project—you start with the tools you have, and buy more as you go along when you find out you need them.

Another factor in the “don’t lose the vehicle” mindset is that loss of the vehicle probably meant the loss of most of your data and possibly a life or three. SpaceX isn’t as worried about this—Starship prototypes are unmanned and they instrument the heck out of everything and collect all the data in real time. I’d be willing to bet a beer that a single Starship hop generates at least as much (non-video) data as the entire Saturn V flight program.

Finally, as others have pointed out, much of the data they need to get simply can’t be replicated subscale or simulated—and even if it could, the full-scale testing is probably still cheaper—plus, it’s getting them experience with manufacturing and ground handling on a vehicle this large.
That's about the best description for what the SLS team is doing that I've heard.

I'd agree with your comment re the SpaceX approach too except maybe for their engine development process.
 
i had some thought about the cost of SN
They are according Musk tweets mass around 150 short tons or 136 metric tons
mostly 304L steel that cost $1880/short tons
if we count 15 short tons for equipment, engine hydraulics and flight electronic
remain 121 short tons tank structure wings

cost estimation
121 short tons steel 304L = $253800 o_O
3 Raptors each $25 million ? = $75000000 (estimation)
1 Tesla Battery pact = $3500 dollar
x COPV tanks = $1000000 (estimation)
Flight electronic = $1000000 (estimation)

A SN Could cost around $77 million

How much to replace THE SLS if they'd blown it up on the test stand? How many YEARS to replace?
 
i had some thought about the cost of SN
They are according Musk tweets mass around 150 short tons or 136 metric tons
mostly 304L steel that cost $1880/short tons
if we count 15 short tons for equipment, engine hydraulics and flight electronic
remain 121 short tons tank structure wings

cost estimation
121 short tons steel 304L = $253800 o_O
3 Raptors each $25 million ? = $75000000 (estimation)
1 Tesla Battery pact = $3500 dollar
x COPV tanks = $1000000 (estimation)
Flight electronic = $1000000 (estimation)

A SN Could cost around $77 million

No. Almost all of that price is for the Raptor, and we know that they are already cheaper than $2M. SpaceX has never paid anywhere near $25M for any engine they have ever flown.
 
No. Almost all of that price is for the Raptor, and we know that they are already cheaper than $2M. SpaceX has never paid anywhere near $25M for any engine they have ever flown.
i had no info on price of Raptor, they really cost only $2 million ?
any source i can check ?
 
I thought at one point they wanted to get them at around a quarter of a million with a large production run. That doesn't seem remotely feasible if it's currently $25 million per. So I'd expect it's much less.
 
cost estimation adapted
121 short tons steel 304L = $253800
3 Raptors each $2 million ? = $6000000 (estimation)
1 Tesla Battery pact = $3500 dollar
x COPV tanks = $1000000 (estimation)
Flight electronic = $1000000 (estimation)

A SN Could cost around $8 million
 
No. Almost all of that price is for the Raptor, and we know that they are already cheaper than $2M. SpaceX has never paid anywhere near $25M for any engine they have ever flown.
i had no info on price of Raptor, they really cost only $2 million ?
any source i can check ?

*Below* $2M.

There are no proper public sources about SpaceX engine costs, the closest we do have is basically Musk's twitter. This is the latest I found:


Do note that this value is almost certainly "marginal cost of making a single extra engine", and is therefore not directly comparable to actual prices of other engines. That is, SpaceX pays for R&D as a separate item, while any commercially available engine would have to distribute the R&D costs into the prices of the engine to recover them from their customers. Then again, that marginal cost is the right one to refer to when considering what blowing up engines costs them.

The philosophy of SpaceX engine manufacturing is that if a <ton lump of copper costs more than a million dollars to make, there is something very, very wrong with how you are making it. They chase cost reduction as the primary goal.
 
They chase cost reduction as the primary goal.
Though that can end up being a major fallacy in itself.
Well it does allow for preservation of test material to be a very secondary goal, which seems to be working to date just by the measuring stick of Phalcon 9. I honestly thought Elon would spectacularly fail in the Falcon 1 days, and even when he made Falcon 9 cheaper than most comparable boosters as a throw away, I expected failure in recovering boosters. But I guess having a mass production point of view and a lot of money is better than a Persian cat and Agent #2 with a piranha pool...
 
Those on here concerned about the Raptors used on the Starships so far these are old in all senses of the word. See L2 on NSF for more info. If you look at the new Raptors now you can see they are significantly different looking.
 
View: https://twitter.com/austinbarnard45/status/1382350413559590912


SpaceX crews are hard at work following two cryo tests, meaning Raptors could be installed within days! Progress on the orbital pad is in overdrive, with the launch tower taking shape and the first GSE tank in place. Meanwhile BN1 is being scrapped for BN2, which will fly!

View: https://twitter.com/elonmusk/status/1382582177943015431


Aiming to launch next week
 
Last edited:
 
A sole source award will be politically hugely unpopular in some quarters I am guessing. NASA maybe looking for more budget so they can bring a competitor onboard.

 
It looks like Space X cut their cloth to meet NASA’s budget so to speak to win the award.
 

Similar threads

Back
Top Bottom