Yep, one to save for the blooper reel. SpaceX just gained a bunch of experience the good old fashion way. A test program is not a test program without an anomaly or two…
Gained a bunch of experience the good old fashion way, AND probably cheaper than the paralysis by analysis method that takes forever in the government. Additionally, I can’t even fathom how much it cost to slap those SLS tanks on Pegasus, ship them to MSFC, and then test them in the test area.
Well, SpaceX may also be concentrating on the max-Q abort test before the end of the year. Not like everything was depending on a Starship flying before the end of the year. “You don’t get medals for on-time failures!” – Walt Williams, NACA/NASA Pioneer.
He said he thought they could do it in a couple months. He also said “it is difficult to predict these things with precision”. It was an off the cuff estimate of the best-case development time, not a promise or deadline.
Work on Mk2 in Florida will likely stop while the designs are analyzed. The last thing SpaceX needs is for the same thing to happen to Mk2. If it did, SpaceX would get a permanent reputation for having things that blow up, and then who in their right mind would want to ride in a SpaceX Starship?
Supposedly SpaceX had already decided not to fly either Mk1 or Mk 2. Musk said something along the lines of the final design has changed so much that they’re not representative of what will eventually fly. Therefore there wasn’t much point in flying them.
Testing them on the ground gives SpaceX practice at testing something somewhat close to what will fly though. So hopefully there will be many “lessons learned” so they don’t cause Mk 3 or Mk 3 to go “boom”.
They’d need to be around 270+ t for it to be a problem for the purposes of the prototype, especially a non-orbital one. It was built with constant thickness tanks for simplicity and because they didn’t need it to be lighter. 120 t is an eventual goal, not a requirement.
Obviously there’s design changes to be made, that’s the whole point of rapid iteration. The prototypes being heavy was not some kind of basic flaw, just them taking advantage of not having to carry 100+ t of payload.
That’s not a bad approach. Constant thickness tank walls is far from optimal. But it’s more than just simple and cheap to fly. If that’s enough to get off the ground and stress the system, you can do flight tests and find out where, exactly, you can trim down the structure. And by how much.
I don’t think they were expecting it to fail. If this was such a test, and some kind of failure was inevitable, why did SpaceX put the wings on the lower section?
Nope Elon Musk says they had already decided MK1 and MK2 would not fly as they had issues that could not be overcome and they needed to start MK3 and do a redesign so they deliberately tested the limits of the workings to gain data ! Remember this tank will need to be pressurized on Mars or the Moon !
That still doesn’t make sense. Workers were still hard at work — even at night — working on the upper “nose” section a couple of nights before the explosion. Now the nose section will be trashed. If they had already decided against hopping with Mk1, why were they wasting time and manpower on the nose section?
For the same reason anyone “wastes time” on non-functional prototypes and manufacturing pathfinders…it’s not a waste of time. They may well continue to make use of the pieces of Mk1 as a mock-up even when real vehicles are flying.
I think they would have mentioned something if they had planned on testing it to destruction. However, that doesn’t mean they were terribly surprised by the test ending as such.
As for the wings, they’re something that could have an impact on structural failures, or be adversely affected by shifts in the structure as the tanks pressurized, etc. They’re testing a vehicle, not some random pressure vessels to be installed in a chemical plant.
Yes, my understanding is Mark 1 would be used for the freefall from 20 km but the first orbital would be with the Mark 3 or 4. Until we know more, I do not think this is a serious setback.
Speaking about such pressure tests, some years ago I was talking with someone that was on a team which pressure test-to-failure the helium tank for the Atlas in late 1950s (or something similar). I believe this was done at the Southern Calif Santa Susana test site. Everyone in the concrete bunker became more quiet when pressure was at 80KSI, some began hiding behind cabinets when pressure reached 120KSI (I can’t recall exact figures from a conversation 20 years ago). When the tank finally exploded, they later examined for any fragments. Only found a some pieces here and there, all just a inch in size. This engineer never mentioned how loud it was.
Yes, some of the tests done back in the days and a ***lot*** of money was spent on those early rocket developments.
I do not quite understand how this is taken so lightly. Blowing up a fueled vehicle is not something you test for. Either there is a lack of understanding the margins or a gross procedural error. If you are testing to see what happens if you pressurize “until something gives” you probably should not do this with a fully fueled vehicle.
Scuttlebutt on the Internet is that it was accidentally over-pressurized during the test. Details are sketchy, and hearsay, at best. So more along the lines of a “gross procedural error” during a pressure test of the tank.
I haven’t seen a single thing from SpaceX to indicate that, and I can easily see someone technically minded coming up with it as a plausible mechanism, and getting taken by someone less technically minded as saying that’s definitely what it was.
Which is why I referred to it as “scuttlebutt”. Little more than a rumor. I’m hoping SpaceX will give some sort of official statement on the cause of the tank rupture.
A flawed weld or outright fabrication error seems entirely plausible and consistent with what’s been officially stated. Due to this thing’s nature as a manufacturing pathfinder, it had a *lot* more welds (because of not using rolls of steel), and also a lot of rework and adjustment as they figured out how to actually build the thing.
That’s not an indication that they’re doing anything wrong or that they need to fundamentally change anything, as some people are claiming…that’s the nature of such an engineering article. It would have been unusual if it had actually resulted in a flyable vehicle, and the fact that they were talking about going straight to a high altitude flight without starting with short hops was probably an indication of how likely they thought it was to survive repeated flights.
A pressure test would have been done, most likely with pressurized water. Not a fueled vehicle. I agree with you with one exception. No one is taking this lightly. They do need to “Test Like You Fly” and that can sometimes unveil problems. I have read it was a procedural error. So, fix it and move on.
There is a whole school of thought (not at all limited to SpaceX) that failed tests are good. If every test is successful, you aren’t pushing the limits or innovating. So I don’t really mind things like the joke about how you can’t be a “real rocket scientist” until you’ve blown a few things up.
That makes sense. In cryogenic systems, Liquid Nitrogen is often used instead of Liquid Oxygen because they are similar density & LN2 is colder than LOx.
water is denser and leads to its own structural problems. Fine for hydrostatically testing a tank for instance. But here, maybe they wanted to pressurise it well past design limits?
Well the question is at the time of the failure, had the pressure inside the tank exceeded the design specification? If so, then this is not a failure at all.
During LM development, all tanks were tested to their design limits but at least one of each type was tested to failure to find out where the real limits were. This is also typically done with at least one test article of new aircraft – mass is applied to the wings to find out where they will fail.
I saw a quote from Mr. Musk about this. It implies they were testing to, but not over, the design limits. So it wasn’t supposed to be testing to failure. At the same time, I think the words he used were that a failure “wasn’t totally unexpected.” That makes sense to me: If you even bother to ask, “Can we actually take it up to 100%?”, you shouldn’t be shocked when the answer is, “No.”
Yes and back in the days when tests and launches were covered live unlike Soviets (and many companies nowadays). When things go right there’s that thrill of victory. When things go wrong, there’s that agony of defeat.
Interesting you mentioned that Mercury test, engineers figured out what exactly happened (there’s a website describing that, I may add link later), however at the time many in the public and congress had doubts about our space program.
Most Soviet launches are televised now even the experimental ones. They didn’t have the internet in 1955 mate. but it is typical to assume Russia or china are hiding secrets!!
Back in the days Soviets would never disclose a launch unless it was successful. And successful launches were all crappy video or many cases bad photos. NASA would provide live television so if launch were to explode, everyone saw it live.
Fast forward to these days the new aspect are many launches are now done by private companies so there is proprietary stuff (i.e. car companies testing new cars and engines don’t post this stuff on the internet). There is also big stuff such as this SpaceX which is a bit hard to hide, and a big boom where nobody talks about will raise a lot of concern.
There were a large number of Soviet rocket and spacecraft failures. Most were concealed and people (both in the West and among the Soviet public) just heard rumors. Those failures are now well-documented based on information which came out in the 1990s (i.e. shortly after the Soviet Union stopped being soviet or united.)
Reminds me of the many explosions and implosions of the early Atlas rockets from the late 1950’s and early 60’s. That venerable Atlas rocket and Starship share a common engineering design heritage… the propellant tank(s) are also the rocket’s fuselage. Both are basically rocket powered stainless steel balloons.
The early Atlas missiles/launch vehicles (Mercury) had a skin thickness about that of a dime or maybe a bit thinner. Even with the propellant tanks empty, the tankage had to be pressurized to keep them from collapsing (as happened to one of the display missiles at IIRC Patrick AFB a few years ago). Is the Mk. 1 tankage so thin that it must be pressurized? Most of the Atlas failures that I heard about/saw happened about the time they jettisoned the boosters. Don’t think structural integrity was the problem. More likely had to do with propulsion system problems.
While it was impressive to watch, everything I’m reading says that this was a test to the limits test, on a version SpaceX had already decided was a non-flight model. As I read it, this was built mostly to see what was working and what wasn’t with the first design and what kinds of problems the construction process needed to revise. Pressurization was with liquid Nitrogen so no fire hazard and they ran the pressure to the max design limit (Probably to test the welding and assembly processes and see how good they were doing.)
Keep in mind that the construction and materials choices on this are a radical departure on past rocket design so they had a lot of theoretical, but very little practical data to work with. Testing like this is how you get it. I also feel it very much fits the Software design environment that Musk comes from where you build a rough pre-alpha with no thought that it will work first time, but enough to test the concept, then you refine it to a working alpha, then a refined beta, before releasing a version 1.0.
It’s not quite how my dad taught me to design and build (His motto was that it’s always easier to erase and redraw a line on paper than to move a wall once the design moves to the physical domain) but it is very much how some complex, ain’t never done this before, projects feel their way toward fruition. What? Haven’t any of you all ever written a draft version or two of a report before creating a final version for submission to the boss? This is the same idea, but in hardware. SpaceX is just a little more willing to test to the limits of the thing than conventional companies. Could be why SpaceX is getting there faster. Ain’t the first version that has to work right, it’s the last one that need to be bug free.
Reminds of the famous concert dialog from that Little Ole Band From Texas based just up the road from there when Billy asks: “Do you want to blow your top? …Well, I might not have a top to blow,,, Then you better blow what you got baby.”
If this was an intentional test to failure, then preparations seem very lax. Allowing the unretained/uncontrolled failure, and subsequent flight, of the upper bulkhead seems very undisciplined.
We are so used to seeing rockets from a distance; the amount of energy, and the energy density, in these critters is absolutely terrifying and easy to forget.
Especially since that wasn’t a fuel-oxidizer explosion. That was just the energy from pressure in the tanks. (Although people often don’t realize how much force and energy even one atmosphere of pressure implies.)
However, I just did a calculation which does terrify me. Could someone check my math? I just assumed the kinetic energy of the rocket exhaust was a good proxy to the energy density of the fuel-oxidizer mix. For a Raptor, the exhaust velocity is 3725 m/s. Fully fueled, a Starship would have 1.1 million kilograms of fuel and oxidizer. One ton TNT equivalent is 4.184 gigajoules (4.184e9 J). Could someone connect the same dots and see if I’ve either made a mistake or really gotten a scary number? Yes, I know the fuel and oxidizer aren’t mixed, and there’s no way an explosion could liberate even 1% of the total chemical energy in the tanks. But even that would make a big hole in the ground.
Unless more information is released it is impossible to say whether this represents a serious problem going forward. Is there any information on the pressure at the time of the failure? Why would it have been tested at anything more than a typical load factor (e.g. 1.25)? All the welding is done by hand with no X-rays. Arc welding austenitic steel (310S?) without perfect control of temperature and time can cause chromium carbide precipitation at grain boundaries and possible corrosion. The failure was apparently in the weld between the upper bulkhead and the tank. How accessible is this area? How good is the fit between the bulkhead and the tank, which are fabricated separately? There are a lot of unknowns here, at least to me.
Yeah, I have no welding expertise, but to the casual observer it sure seems most likely to be a construction issue… although not enough info to know.
Given that parts flew off the Starhopper on landing (which seemed to use similar construction techniques), it seems highly likely to me that the Mk I wouldn’t have survived its first flight intact.
Btw, some of the welds were done by a welding machine, see reply #194 2/25/19:
If this was an intentional failure test, it is surprising how quiet they’ve been about it. Maybe they think putting out a press release would only restart it’s news cycle ?
Yep, one to save for the blooper reel. SpaceX just gained a bunch of experience the good old fashion way. A test program is not a test program without an anomaly or two…
Gained a bunch of experience the good old fashion way, AND probably cheaper than the paralysis by analysis method that takes forever in the government. Additionally, I can’t even fathom how much it cost to slap those SLS tanks on Pegasus, ship them to MSFC, and then test them in the test area.
“Paralysis by analysis”. I’m going to use that now at work (NASA). So true.
Where I work we just call it analysis paralysis. I much prefer iterative development over waterfall. SpaceX is all about iterative development.
Well it looks like there will be no test flight this year.
Well, SpaceX may also be concentrating on the max-Q abort test before the end of the year. Not like everything was depending on a Starship flying before the end of the year.
“You don’t get medals for on-time failures!” – Walt Williams, NACA/NASA Pioneer.
I was referring to the 20km Starship test flight that Musk said would happen in 2019.
He said he thought they could do it in a couple months. He also said “it is difficult to predict these things with precision”. It was an off the cuff estimate of the best-case development time, not a promise or deadline.
(Sad trombone sound effect) Never mind – this is why they test. On to the next one…
Being the FL built version once they find & apply the fix.
I do hope Mr Musk and his Space X minions can recover and quickly. They are our future in space; I am more convinced of this than ever.
Work on Mk2 in Florida will likely stop while the designs are analyzed. The last thing SpaceX needs is for the same thing to happen to Mk2. If it did, SpaceX would get a permanent reputation for having things that blow up, and then who in their right mind would want to ride in a SpaceX Starship?
Supposedly SpaceX had already decided not to fly either Mk1 or Mk 2. Musk said something along the lines of the final design has changed so much that they’re not representative of what will eventually fly. Therefore there wasn’t much point in flying them.
Testing them on the ground gives SpaceX practice at testing something somewhat close to what will fly though. So hopefully there will be many “lessons learned” so they don’t cause Mk 3 or Mk 3 to go “boom”.
Too much dry mass, among other things. Mk1 was 200 metric tons dry IIRC, and they need to get it down to 120 metric tons. Some redesign is in order.
They’d need to be around 270+ t for it to be a problem for the purposes of the prototype, especially a non-orbital one. It was built with constant thickness tanks for simplicity and because they didn’t need it to be lighter. 120 t is an eventual goal, not a requirement.
Obviously there’s design changes to be made, that’s the whole point of rapid iteration. The prototypes being heavy was not some kind of basic flaw, just them taking advantage of not having to carry 100+ t of payload.
That’s not a bad approach. Constant thickness tank walls is far from optimal. But it’s more than just simple and cheap to fly. If that’s enough to get off the ground and stress the system, you can do flight tests and find out where, exactly, you can trim down the structure. And by how much.
Was this a dynamic pressure test, where SpaceX jacked up the presure until something gave? They had removed the top and put it a safe distance away..
I don’t think they were expecting it to fail. If this was such a test, and some kind of failure was inevitable, why did SpaceX put the wings on the lower section?
Nope Elon Musk says they had already decided MK1 and MK2 would not fly as they had issues that could not be overcome and they needed to start MK3 and do a redesign so they deliberately tested the limits of the workings to gain data ! Remember this tank will need to be pressurized on Mars or the Moon !
That still doesn’t make sense. Workers were still hard at work — even at night — working on the upper “nose” section a couple of nights before the explosion. Now the nose section will be trashed. If they had already decided against hopping with Mk1, why were they wasting time and manpower on the nose section?
Maybe the nose section isn’t what’s changing?
For the same reason anyone “wastes time” on non-functional prototypes and manufacturing pathfinders…it’s not a waste of time. They may well continue to make use of the pieces of Mk1 as a mock-up even when real vehicles are flying.
To simulate real loads.
That said, it sounds like they’ve iterated the design again.
I think they would have mentioned something if they had planned on testing it to destruction. However, that doesn’t mean they were terribly surprised by the test ending as such.
As for the wings, they’re something that could have an impact on structural failures, or be adversely affected by shifts in the structure as the tanks pressurized, etc. They’re testing a vehicle, not some random pressure vessels to be installed in a chemical plant.
No. I believe this particular prototype was supposed to fly eventually.
Apparently not , Elon Musk says they found a design flaw and had already shifted focus to MK3 for flight tests
Yeah, it blew up!
Seriously, all of SpaceX announcements before this test were that Mk 1 was going to fly.
Yes, my understanding is Mark 1 would be used for the freefall from 20 km but the first orbital would be with the Mark 3 or 4. Until we know more, I do not think this is a serious setback.
Speaking about such pressure tests, some years ago I was talking with someone that was on a team which pressure test-to-failure the helium tank for the Atlas in late 1950s (or something similar). I believe this was done at the Southern Calif Santa Susana test site. Everyone in the concrete bunker became more quiet when pressure was at 80KSI, some began hiding behind cabinets when pressure reached 120KSI (I can’t recall exact figures from a conversation 20 years ago). When the tank finally exploded, they later examined for any fragments. Only found a some pieces here and there, all just a inch in size. This engineer never mentioned how loud it was.
Yes, some of the tests done back in the days and a ***lot*** of money was spent on those early rocket developments.
I do not quite understand how this is taken so lightly. Blowing up a fueled vehicle is not something you test for. Either there is a lack of understanding the margins or a gross procedural error.
If you are testing to see what happens if you pressurize “until something gives” you probably should not do this with a fully fueled vehicle.
Scuttlebutt on the Internet is that it was accidentally over-pressurized during the test. Details are sketchy, and hearsay, at best. So more along the lines of a “gross procedural error” during a pressure test of the tank.
I haven’t seen a single thing from SpaceX to indicate that, and I can easily see someone technically minded coming up with it as a plausible mechanism, and getting taken by someone less technically minded as saying that’s definitely what it was.
Which is why I referred to it as “scuttlebutt”. Little more than a rumor. I’m hoping SpaceX will give some sort of official statement on the cause of the tank rupture.
A flawed weld or outright fabrication error seems entirely plausible and consistent with what’s been officially stated. Due to this thing’s nature as a manufacturing pathfinder, it had a *lot* more welds (because of not using rolls of steel), and also a lot of rework and adjustment as they figured out how to actually build the thing.
That’s not an indication that they’re doing anything wrong or that they need to fundamentally change anything, as some people are claiming…that’s the nature of such an engineering article. It would have been unusual if it had actually resulted in a flyable vehicle, and the fact that they were talking about going straight to a high altitude flight without starting with short hops was probably an indication of how likely they thought it was to survive repeated flights.
A pressure test would have been done, most likely with pressurized water. Not a fueled vehicle. I agree with you with one exception. No one is taking this lightly.
They do need to “Test Like You Fly” and that can sometimes unveil problems. I have read it was a procedural error. So, fix it and move on.
I did not mean to imply that SpaceX is taking it lightly. Just the initial comments seemed to ignore the seriousness of the event.
That’s just because such things are epic to watch.
There is a whole school of thought (not at all limited to SpaceX) that failed tests are good. If every test is successful, you aren’t pushing the limits or innovating. So I don’t really mind things like the joke about how you can’t be a “real rocket scientist” until you’ve blown a few things up.
Indeed, testing to failure is not a failed test. It gives data on the design and manufacturing processes.
PR says cryogenic fluid, not water. LN2 or liquid CO2?
The Ars Technica story said liquid nitrogen.
That makes sense. In cryogenic systems, Liquid Nitrogen is often used instead of Liquid Oxygen because they are similar density & LN2 is colder than LOx.
water is denser and leads to its own structural problems. Fine for hydrostatically testing a tank for instance. But here, maybe they wanted to pressurise it well past design limits?
They weren’t using fuel (i.e. oxygen or methane) but nitrogen.
For a full description see this:
https://arstechnica.com/sci…
Sounds like the Chernobyl method!
Well the question is at the time of the failure, had the pressure inside the tank exceeded the design specification? If so, then this is not a failure at all.
During LM development, all tanks were tested to their design limits but at least one of each type was tested to failure to find out where the real limits were. This is also typically done with at least one test article of new aircraft – mass is applied to the wings to find out where they will fail.
I saw a quote from Mr. Musk about this. It implies they were testing to, but not over, the design limits. So it wasn’t supposed to be testing to failure. At the same time, I think the words he used were that a failure “wasn’t totally unexpected.” That makes sense to me: If you even bother to ask, “Can we actually take it up to 100%?”, you shouldn’t be shocked when the answer is, “No.”
Nothing beats the Mercury test where the chute comes out the top…
Yes and back in the days when tests and launches were covered live unlike Soviets (and many companies nowadays). When things go right there’s that thrill of victory. When things go wrong, there’s that agony of defeat.
Interesting you mentioned that Mercury test, engineers figured out what exactly happened (there’s a website describing that, I may add link later), however at the time many in the public and congress had doubts about our space program.
Most Soviet launches are televised now even the experimental ones. They didn’t have the internet in 1955 mate. but it is typical to assume Russia or china are hiding secrets!!
Back in the days Soviets would never disclose a launch unless it was successful. And successful launches were all crappy video or many cases bad photos. NASA would provide live television so if launch were to explode, everyone saw it live.
Fast forward to these days the new aspect are many launches are now done by private companies so there is proprietary stuff (i.e. car companies testing new cars and engines don’t post this stuff on the internet). There is also big stuff such as this SpaceX which is a bit hard to hide, and a big boom where nobody talks about will raise a lot of concern.
There were a large number of Soviet rocket and spacecraft failures. Most were concealed and people (both in the West and among the Soviet public) just heard rumors. Those failures are now well-documented based on information which came out in the 1990s (i.e. shortly after the Soviet Union stopped being soviet or united.)
truly one for the vault …
the view from the “cherry picker” crane basket would’ve be “amazing”.
neat to see the top of the pressure vessel disappear off the top of the screen, only to fall back quite some time later !
Reminds me of the many explosions and implosions of the early Atlas rockets from the late 1950’s and early 60’s. That venerable Atlas rocket and Starship share a common engineering design heritage… the propellant tank(s) are also the rocket’s fuselage. Both are basically rocket powered stainless steel balloons.
Back to the Future !
The early Atlas missiles/launch vehicles (Mercury) had a skin thickness about that of a dime or maybe a bit thinner. Even with the propellant tanks empty, the tankage had to be pressurized to keep them from collapsing (as happened to one of the display missiles at IIRC Patrick AFB a few years ago). Is the Mk. 1 tankage so thin that it must be pressurized?
Most of the Atlas failures that I heard about/saw happened about the time they jettisoned the boosters. Don’t think structural integrity was the problem. More likely had to do with propulsion system problems.
While it was impressive to watch, everything I’m reading says that this was a test to the limits test, on a version SpaceX had already decided was a non-flight model. As I read it, this was built mostly to see what was working and what wasn’t with the first design and what kinds of problems the construction process needed to revise. Pressurization was with liquid Nitrogen so no fire hazard and they ran the pressure to the max design limit (Probably to test the welding and assembly
processes and see how good they were doing.)
Keep in mind that the construction and materials choices on this are a radical departure on past rocket design so they had a lot of theoretical, but very little practical data to work with. Testing like this is how you get it. I also feel it very much fits the Software design environment that Musk comes from where you build a rough pre-alpha with no thought that it will work first time, but enough to test the concept, then you refine it to a working alpha, then a refined beta, before releasing a version 1.0.
It’s not quite how my dad taught me to design and build (His motto was that it’s always easier to erase and redraw a line on paper than to move a wall once the design moves to the physical domain) but it is very much how some complex, ain’t never done this before, projects feel their way toward fruition. What? Haven’t any of you all ever
written a draft version or two of a report before creating a final version for submission to the boss? This is the same idea, but in hardware. SpaceX is just a little more willing to test to the limits of the thing than conventional companies. Could be why SpaceX is getting there faster. Ain’t the first version that has to work right, it’s the last one that need to be bug free.
This is EXACTLY why NASA spends $13 billion developing a rocket out of spare parts.
And they have YET to test it.
It is easier to avoid failures that way…
No flight tests of course, but they’ve obviously done engine tests, abort tests, and structural tests…
Blowups Happen – RAH
Reminds of the famous concert dialog from that Little Ole Band From Texas based just up the road from there when Billy asks: “Do you want to blow your top? …Well, I might not have a top to blow,,, Then you better blow what you got baby.”
If this was an intentional test to failure, then preparations seem very lax. Allowing the unretained/uncontrolled failure, and subsequent flight, of the upper bulkhead seems very undisciplined.
We are so used to seeing rockets from a distance; the amount of energy, and the energy density, in these critters is absolutely terrifying and easy to forget.
Especially since that wasn’t a fuel-oxidizer explosion. That was just the energy from pressure in the tanks. (Although people often don’t realize how much force and energy even one atmosphere of pressure implies.)
However, I just did a calculation which does terrify me. Could someone check my math? I just assumed the kinetic energy of the rocket exhaust was a good proxy to the energy density of the fuel-oxidizer mix. For a Raptor, the exhaust velocity is 3725 m/s. Fully fueled, a Starship would have 1.1 million kilograms of fuel and oxidizer. One ton TNT equivalent is 4.184 gigajoules (4.184e9 J). Could someone connect the same dots and see if I’ve either made a mistake or really gotten a scary number? Yes, I know the fuel and oxidizer aren’t mixed, and there’s no way an explosion could liberate even 1% of the total chemical energy in the tanks. But even that would make a big hole in the ground.
Unless more information is released it is impossible to say whether this represents a serious problem going forward. Is there any information on the pressure at the time of the failure? Why would it have been tested at anything more than a typical load factor (e.g. 1.25)? All the welding is done by hand with no X-rays. Arc welding austenitic steel (310S?) without perfect control of temperature and time can cause chromium carbide precipitation at grain boundaries and possible corrosion. The failure was apparently in the weld between the upper bulkhead and the tank. How accessible is this area? How good is the fit between the bulkhead and the tank, which are fabricated separately? There are a lot of unknowns here, at least to me.
Yeah, I have no welding expertise, but to the casual observer it sure seems most likely to be a construction issue… although not enough info to know.
Given that parts flew off the Starhopper on landing (which seemed to use similar construction techniques), it seems highly likely to me that the Mk I wouldn’t have survived its first flight intact.
Btw, some of the welds were done by a welding machine, see reply #194 2/25/19:
https://forum.nasaspaceflig…
although not sure what that implies regarding control of weld conditions.
If this was an intentional failure test, it is surprising how quiet they’ve been about it. Maybe they think putting out a press release would only restart it’s news cycle ?