Slashdot Mirror
SpaceX Rocket Engine Explodes During Test (space.com)
According to The Washington Post, a SpaceX rocket engine exploded Sunday (Nov. 5) at the company's test facility in McGregor, Texas. The explosion reportedly occurred during a "qualification test" of a Merlin engine, the type that powers SpaceX's workhorse Falcon 9 rocket. Space.com reports: SpaceX has suspended engine testing while it investigates what caused the incident, which didn't injure anyone, the Post added. In a statement provided to the Post, SpaceX representatives said they didn't expect the explosion to affect the company's launch schedule. That schedule has been pretty packed this year. SpaceX has already launched 16 missions, all of them successful, in 2017 -- twice as many as its previous high in a calendar year. And all but three of these missions also involved landings of the Falcon 9 first stage, for eventual refurbishment and reuse.
The incident in question did not occur during an engine firing. Rather they were performing a "LOX drop" test which basically involves pumping LOX through the engine and checking for leaks. Something went wrong in this process, causing the damage. Until the investigation is completed, there's no way to know whether it was an issue with the engine, the test rig, or the setup. It might be that a tech just dind't tighten something adequately, or a filler hose leaked or whatever. SpaceX won't know until they complete their investigation, and we may never know.
To quote Adam Savage of Mythbusters fame "LOX makes anything flammable. LOX makes something flammable into a high explosive." So even if they just had a sufficiently large leak, and the LOX leaked onto/into asphalt or similar, all it takes is a spark to cause that asphalt to detonate like a bunch of dynamite.
...si hoc legere nimium eruditionis habes...
LOL what?
I'm no Musk fan, but what's why you test?
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
That’s what happens when you cut corners and half ass things to make things cheaper.
Launch costs need to be cheaper. The trick is to figure out which corners can be cut, and which can not. Engineers learn by trying and failing, and I am sure SpaceX learned some valuable lessons today.
"Ever tried. Ever failed. No matter. Try Again. Fail again. Fail better." - Samuel Beckett
Obviously failed the test.
That's a 1650-cubic inch V-12.
Exactly!
âoeWhy do we fall sir? So that we can learn to pick ourselves up.â -Alfred Pennyworth
At least this happened with the new Merlin Series 5 redesign, scheduled for flight next year.
The current Series 4 engines have been pretty reliable so far...
Procrastination; I'll think of a sig tomorrow.
Launch costs need to be cheaper.
No they don’t.
Most famous British engine in the war. Bit rude of SpaceX to reuse the name.
the nature of test is to find faults
if everything was perfect by design the test people would be flipping burgers or work as perfect-design engineers
4wdloop
it is an EXTERNAL combustion engine is it not?
4wdloop
At least this happened with the new Merlin Series 5 redesign, scheduled for flight next year.
Exactly. That's important-- this is the next generation engine, not the one currently flying.
Some alternate sources, some with more information:
https://www.space.com/38712-spacex-rocket-engine-test-explosion.html
https://www.geekwire.com/2017/next-generation-spacex-rocket-engine-goes-flames-texas-test/
https://arstechnica.com/science/2017/11/an-experimental-spacex-rocket-engine-has-exploded-in-texas/
https://www.theverge.com/2017/...
No, then they wouldn't test at all.
My reaction to this is more like:
Yeah so, this is why they test.
Considering they LOX drop and test fire each engine, eventually with this many engines you were bound to have one with a flaw. All I can say is, good job.. better on a test stand than on a Falcon 9.
Yes Francis, the world has gone crazy.
No, not quite; more like "this is why we have tests."
The engine, not so much.
The Rolls Royce "merlin", like other Rolls Royce aero engines of the time, was named after a bird of prey. The bird they named the engine after is a type of falcon which is called a "pigeon hawk" in North America and "merlin" in Europe.
So, did SpaceX name their engine after the bird (and Rolls Royce's engine) or the mythical wizard?
Doing a quick search and there's no clear answer.
Mimetics Inc. Twitter
You're partly right. If you're being smart economically, or advancing technology, you'll test some things that don't work. If everything you test works, you're a) doing the same boring shit that's been done and b) over-engineering, making things much more expensive than they should be.
Testing is how you find out what works and what doesn't, and how much you need to spend to make things work reliably.
All the fanbois that refer to Musk as a real life Tony Stark seem to be missing in action.
If only Elon could step in a wave his hands and fix it like Tony Stark does.
There is a marvelous history of the development of rocket fuel called, "Ignition!", written by John D. Clark, one of the field's insiders who has an ascerbic wit. The foreword was written by Isaac Asimov, which contains the following fantastic quote:
Now it is clear that anyone working with rocket fuels is outstandingly mad. I don't mean garden-variety crazy or a mere raving lunatic. I mean a record-shattering exponent of far-out insanity.
There are, after all, some chemicals that explode shatteringly, some that flame ravenously, some that corrode hellishly, some that poison sneakily, and some that stink stenchily. As far as I know, though, only liquid rocket fuels have all these delightful properties combined into one delectable whole.
Explosions are par for the course. Rocket science is hard.
Put my fist through my alarm clock with its ding-dong death inside my ear. - The Blackjacks.
Ok, then, can we get ULA on a level playing field with cutting corners? Right now ULA is required to do the full engineering work up for every launch, v.s. spaceX not having ever done the engineering and pencil-whipping the paperwork.
Ok, then, can we get ULA on a level playing field with cutting corners?
You are being silly. This was R&D, not a production launch.
Right now ULA is required to do the full engineering work up for every launch, v.s. spaceX not
1. This was not a "launch"
2. It should be up to the market.
ULA provides expensive reliability. SpaceX provides discount access to orbit. If you are launching a 5 billion dollar GSO comsat, you will go with ULA. If you want to dump a van load of cubesats designed by high school science clubs into LEO, you go with SpaceX.
SpaceX will get more reliable much faster than ULA will get cheaper. In ten years, ULA will be out of business.
SpaceX provides discount access to orbit. If you are launching a 5 billion dollar GSO comsat, you will go with ULA. If you want to dump a van load of cubesats designed by high school science clubs into LEO, you go with SpaceX.
DoD is launching with SpaceX now, so they have definitely jumped up in the rankings compared to ULA, and the various state-owned launchers. Cubesats and science projects are becoming the domain of start-ups that NASA is funding
FWIW the accident involved a new block-5 merlin engine that was undergoing lox load testing for leaks and 'something' caught fire, damaging the test facility, and presumably the engine, severely. It has not been determined if the engine, which was not firing, was at fault.
They were testing an engine, and this particular test failed... That's the whole point of testing, try new things and see if they work.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Elon immediately held a press conference to say the exploding rocket was actually a fireworks show to commemorate daylight savings time.
I suppose those expensive GSO sats (TV & comm) that SES, largest commercial sat operator, is having launched by SpaceX, even on refurbished SpaceX launchers (SES being first commercial customer), must be cubesats... NOT!
SES being a for-profit corporation, they run the numbers... and they don't seem to do much business with ULA (as compared to SpaceX, Arianespace, Proton...).
Elon Musk"s rocket explodes, reports Jeff Bezos.
Can they meet ISRO's cost?
Working with LOX is just hard. Everyone that work with high purity oxygen eventually burns up something unexpected.
NASA Oxygen Gleaning Team website
Lox makes almost everything flammable, even the pipes we use to convey it.
LOX Safety Video
Don't do this, common materials become DANGEROUS when exposed to LOX
LOX as a fire starter
Guess - some FOD was inside the engine, perhaps some lint or hair. Proving the cause is going to take work.
If you aren't blowing things up now and then, you aren't on the frontier of exploration. You cannot know where the line designating the frontier is unless you occasionally step over it.
You're so very right. Software developers in particular sometimes test software only with valid, expected inputs. Unexpected inputs then result in a security failure.
Back when software ran locally, we we used to say "garbage in, garbage out". That's no longer acceptable for internet-connected software. With Heartbleed, the garbage that came out was random memory contents, which could include the server's private key.
>I suppose those expensive GSO sats (TV & comm) that SES, largest commercial sat operator, is having launched by SpaceX, even on refurbished SpaceX launchers (SES being first commercial customer), must be cubesats... NOT!
You win today's internet for not getting the point.
The shepherds did so well protecting the flock that the sheep no longer believed that wolves existed.
1. This was not a "launch"
This was not even an engine firing. It was during a LOX Drop test, where the engine is tested for leaks. And since pure Oxygen outside of where it is supposed to be is nasty, it could be just about anything in the area.
The post you were replying to was pure did not RTFA classic Slashdot.
You never ever ever take shortcuts with rocket engines. They are channeled explosions or deflagrations if you wish. They might be made cheaper, but shortcuts are instantly punished.
https://arstechnica.com/scienc... here's a nice example of changing an engine. The F1's were a testament to balls to the wall, but hand assembled with so much hand welding. We'd never want to reproduce that today. The F-1B is a nice modern update to a sound concept. No shortcuts allowed in the world of LOX and RP-1. That doesn't have all that much to do with this particular story, but its good reading. Maybe OP will give it a read and learn something.
The shepherds did so well protecting the flock that the sheep no longer believed that wolves existed.
At least they have an idea now that there's a problem somewhere that needs fixing, and it's not hidden for an actual launch.
Good test. Try and break it.
"Hmm. I am to metaphor cheese as metaphor cheese is to transitive verb crackers!"
Oh, yes. A failure during a test is nothing unexpected. It is where failures are supposed to happen. Anybody that does not understand that does not know the first thing about engineering. And a "qualification test" in particular serves to find the occasional manufacturing fault still present before it does real damage.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
Wasn't it Alan Shepard that remarked that his Mercury capsule was all built by the lowest bidders?
I was thinking the same thing.
Mishaps on the test stand is what the thing is built for.
Even a massive explosion with all equipment lost is a success because it thus did not happen on a launch pad where in addition to the lost equipment you very well may/will lose:
* The Payload.
* The Launchpad Facility.
* The actual Launchpad.
* Lives.
* Delay to future launches of unknown duration because of aforementioned damage.
Sure it's a suboptimal success, but it is still not a failure.
whois gawk date unzip strip find touch finger mount join nice man top fsck grep eject more yes exit umount sleep dump
I am reasonably new at my current (very small) company. I am the second (and still not fully dedicated) QA person.
I came from 17 years at a multinational Corp with huge QA.
To say I encountered culture shock is an understatement.
I have started implementing things like automated regression tests, and Fuzzers. My Fuzzer based tests break the shit out of things and the devs look at me with the "why would you do that?" look. They still have to go fix the issue though.
whois gawk date unzip strip find touch finger mount join nice man top fsck grep eject more yes exit umount sleep dump
What is this "full engineering work up" and what forces ULA to do it and allows SpaceX to avoid it? If it's an extraneous requirement for ULA that is unnecessary then it should be removed. However given their history I would imagine it is something created BY them to create a barrier of entry into the launch market for competitors (like the "launch readiness" contract that gave them an extra Billion per year)
Right now ULA is required to do the full engineering work up for every launch
Not really; since they're not making the engines, they're clearly not "required to do the full engineering work up for every launch".
Ezekiel 23:20
Grandparent's arguing with ULA is stupider for another reason, too: SpaceX is the only company in the recent US history to have manufactured 1000 rocket engines worth 700 MN of total thrust in a timeframe of several years. Nobody else is doing anything like this in the US anymore, so clearly nobody can possibly have any failures in engine manufacturing in the US. The closest "competitor" is Aerojet currently manufacturing no more than 15 small RL-10 engines per year for ULA on their old manual production line.
In fact, SpaceX may be the most productive large rocket engine manufacturer in history of launch vehicles (or poised to become one very soon), period - even at the height of the Cold War, there were at best 47 Soyuz launches per year, which translates to 235 first-stage rocket engines per year, and SpaceX has manufactured most of their engines in the last few years. Now tell me how many Russian test failures you're aware of, either in the Soyuz line or recently in the RD-17x/18x/19x line. It's easy to blame SpaceX for a manufacturing or testing failure when their major competitor offshores the failures to places the common American hasn't even heard about and only imports finished products.
Ezekiel 23:20
Found the Arianespace CEO.
Ezekiel 23:20
The question is whether ISRO can ever meet SpaceX's costs. Even at $25M per launch, ISRO's PSLV it can only carry one fifth of the load that Falcon 9 can (despite having 60% of Falcon 9's lift-off weight!). That makes it about twice as expensive per kg of payload, or even more. It's even more striking if you take Indian 1:8 to 1:10 labor price advantage in the aerospace sector into consideration.
Ezekiel 23:20
"The devs look at me with the "why would you do that?" look. "
It's not just small companies which suffer that.
Airbus have suffered a number of "why would a pilot do something that stupid?" issues where pilots DO these kinds of things when testing to trying out the aircraft to see how it will react under worst case conditions.
Certain switch/router manufacturer R&D departments have come back with the same question when I've asked them to check certain conditions. They may not ever happen under normal operation, but when someone's trying to break in all bets are off.
Back in High school days my tutors used to criticise me for running tests on every input as a waste of time/memory, but it's a habit which really should be ingrained in coders along with testing every condition which might occur, no matter how unlikely. Assumption is the mother of all fuck ups.
Indeed. It is one form of redundancy employed to make the final product as reliable as it reasonably can be. Definitely not a failure.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
What is the point? Companies are launching important payloads with SpaceX.
I completely concur that tests on all inputs (external to the system) are mandatory.
By external I mean any input from outside the code.
Last big project I worked on there were two "rings" in ring 0 code, one that faced userland and one that only faced ring 0 code. The former validated *everything* the latter, not so much. Embedded system where memory and compute cycles were at a premium.
whois gawk date unzip strip find touch finger mount join nice man top fsck grep eject more yes exit umount sleep dump