SpaceX Dragon chokes at the last second The Falcon 9 rocket from private space company SpaceX, intended to launch this morning and send a Dragon capsule loaded with supplies to the International Space Station, has failed to take off. The rocket's computer aborted the launch automatically at almost the final possible moment, when its engines had already ignited but the …
I'm a little disappointed that they didn't get to launch today (was watching the live stream), I really hope Elon Musk and his team figure out what happened to Engine #5 and are good to go for another launch attempt on Tuesday. There is nothing about this venture that I don't want to succeed. From where I'm sitting, the COTS program (whether intentionally or not) has helped to galvanise a new interest in space tech and travel, with this new batch of 'firsts' waiting to be had.
Good luck to the Space X team for their next effort, keep up the good work!
While not a successful launch, it has proved that the safety measures in the software and hardware work. Not only has this prevented a potential disaster for this launch (and possibly the project), but it also helps confirm its safety credentials toward a future manned launch.
A delay is better than a disaster.
The challenges and difficulties in designing and implementing such systems are a perennial problem. Such development processes (since one wishes to err on the side of caution) are always bedevilled by false alarms and are eliminated (we hope) by a "trial and error" process where the "error" is that an operation may be unnecessarily suspended. However, that is the whole point, it is a deliberate choice to go the route of getting "false alarms" from time to time rather than catastrophic failure. Which would you rather have, hmm?
Half silly, half serious - we've been building rockets for over 50 years, for most of that time using hugely simple computers. Our materials science and computer technology is so far beyond what was used to reach the moon, and they can't get a simple rocket to take off?
Anyone who knows any specifics and can rebut this based on solid reasons/fac - I genuinely welcome being corrected.
People have been building cars for much longer, but still every new model has its teething problems. At the moment the parameters on the Merlin engine are set very tight, one of the possibilities for the postponed launch is to widen the acceptable limits so the reading seen this morning wouldn't be a cause for abort. Don't forget that only 23 Merlin engines have flown so far, and even with the static test firings that's not a lot of data.
The Russian Soyuz and Proton launchers are the only ones that reliably go first time, but there have been nearly 400 Proton launches and around two thousand Soyuz family launches.
I can't help thinking the Russian rockets had less engineers, therefore it was a more single minded and simplified (comparatively, I'm not rocket scientist!) effort. Rather than teams of people fixing connected problems separately, like it just seemed to be at NASA.
'The Russian Soyuz and Proton launchers are the only ones that reliably go first time, but there have been nearly 400 Proton launches and around two thousand Soyuz family launches.'
And as the Russians have found out over the last couple of years, even these rockets still throw up unexplained problems.
Does Elon Musk have the best job in the world - designing rockets by day, electric cars by night?
Building something with an 80% chance of working, getting to orbit and not blowing up unintentionally is easy. Trivial, even, if you have a couple million lying around. Building something with a 95% chance is a little harder, and 99% harder still. Shutdowns like this are part of the 99% "not blowing up" work, putting in a little gizmo that can say "Hmm, this might blow up and we can still stop it. Let's do so."
There was a great anecdote in the Murray&Cox "Apollo" book about how von Braun settled on five-9 reliability (99.999%) for the Saturn booster. He asked his five German rocket colleagues "Any reason why this won't work?" and the replies: "Nein", Nein", Nein", Nein", Nein". "There you are: five-9s."
@RosslynDad, that made me grin way more than it should. Well done.
@JDX
I Have done some FEM analysis work on just the "reaction chamber, throat and expansion" geometry of a simple rocket engine design. I think I can reasonably state that even with our modern computer calculating powers, designing a rocket motor is not a trivial matter.
The forces and pressures are incredible for the amount of weight allowed for the construction. (560 KiloNewtons of thrust at sea-level. 6.77 MPa chamber pressure. Thats 982 psi for you weird unit users) We are talking about several hundreds of tonnes pushing out on the chamber walls, trying to make the thing blow up. On a construction that can't weight more than a few hundred kilograms.
We are talking about dozens of liters of fuel and oxidizer entering the reaction chamber from above, violently mix and ignite, accelerate to over mach 1 in the throat and then accelerate even further in the expansion nozzle. All the heat generated doesn't just flow out with the exhaust either. A lot of it gets absorbed by the chamber walls, throat and expansion nozzle. All that heat has to be extracted or it melts the rocket, causing an explosion. Just the heat flow problem alone takes hundreds of FEM calculation hours to get even close to a detailed enough model to be able to make any reasonable assumptions on validity and design functionality. And then hundreds of hours more to calculate the next iteration. And then they still find local hotspots caused by imperfections in the manufacture or design that didn't show up on the models. Or pogo oscillations caused by the interactions between chamber pressures and turbo-pump behaviour. So they go back to the designs some more.
I'm very often baffled by how the boffins of days gone by even managed some of the designs they did with the means they had. One of the reasons the russian Soyuz range is now so successful is that they've had years and years and hundreds of launches to test and improve the designs. Even successful launches give a lot of useful data They had quite a few incidents in the past. In those days losses were not really tolerated but were still "part of the program". With the eyes of the world now on commercial companys like spaceX a single catastrophic failure could mean the end of the company. So they now have to use all that extra computer power to make their rockets 99,99999999% safe. Because you can be damn sure that if a cargo or even a crew is lost due to a design error, SpaceX can kiss its contracts goodbye. A lot of the worlds media (and probably some NASA boffins) will simply start screaming: "you SEE. We told you, you can't leave this sort of stuff to commercial penny-pinching companies!"
A good question.
There is a continual drive to improve efficiency and reduce costs - payloads are much higher these days.
Also, because materials are pushed near their limits there are limits to what can be simulated. Trivial issues such as vapour bubbles in fuel pumps, resonances in structural members can get out of control very quickly which is why they monitor as much as possible.
This monitoring was what allowed the safe shutdown today before there was a blow up on the pad, or worse in the air.
This kind of things is why when someone has a reliable launch vehicle it keeps getting used until there is another with a significant advantage adn this is what SpaceX offers.
Re: "payloads are much higher these days".
At the risk of sounding like a broken record, and leaving aside the question of average versus largest payloads, no. Back in the day, the largest payloads were in the 120 tons to LEO category, something we seem to have difficulty with today, and at a time when we knew less about rocketry and materials engineering than we do now.
Maybe that's the problem. Today, we know too much about what we can't do, and we seem to have forgotten some of what we *can* do.
Well over a Billion cars have been built with an immeasurable number of journeys.
Same for planes.
Whereas orbital rockets number in the low thousands.
Additionally as the financial investments required vs returns are relatively low compared to automotive, there is minimal financial investment.
There are a number of projects around the world to reduce costs of space travel, but these will take time and investment in a time of austerity.
The technology hasn't advanced dramatically since the 60's, there is plenty of stuff in concept or development, but governments and investors see space travel as high risk with low returns.
This is why that billionaire club for mining asteroids was setup, it's an attempt of making commercial space operations financially viable.
Unlike automobiles, you don't get to crank a rocket engine very many times and take it for a test drive. Boeing does around 2000 test flights on an airplane before it goes to the first customer. The Space Shuttle only flew 135 times ever. By airplane standards, the Shuttle was still in early test when they retired it, and this particular Falcon 9 is attempting it's one and only flight, ever.
So the difficulty isn't that rocketry is harder from an engineering standpoint, it's that you get fewer chances to find design and production defects, and any defects are often catastrophic.
Okay, so I'm risking a reply to what may be an obvious troll, but... The Soviets were thought to have enormous apparent advantage in rocket design 50 years ago, and well, look up N1.
Soviets couldn't build a "simple rocket" to reach the Moon, it makes the American Saturn V that much more remarkable. (Some try to say it was a lucky break, but the perfect safety record says otherwise).
I didn't say it wasn't complicated and tremendously difficult. I claimed that it's half-century old tremendously complex technology and that with modern science's leaps, a well-funded organisation should be able to get a rocket to the launchpad fairly sure it will work.
Similar to nuclear reactors - they are still very difficult to make BUT compared to 50 years ago, it's not an unknown anymore.
Also why all the car comparisons?
You quoted Top Gear, it's a car program.
It's not that things are known/unknown, it's that growth and innovation come from commercial prospects. (Most of us live in a Capitalist society)
The technology in rockets hasn't evolved much since the 60's, we've had maybe 4-5 generations of rockets, for 2 reasons:
1. Cost
vs
2. Returns
Cars are a good analogy, because of the problems with rockets, while more complicated than a car. Cars also had problems to begin with, but Cars generally have a generation every 4 years or so, given they've been around for a 100 year, so around 25 generations.
It's evolution, Cars also have a commercial market with an established demand from customers.
Rockets are in lower demand, relatively and due to their high development and operating costs, the frequency and volume is low, reducing investment stimulation.
The commercial opportunities (currently) are also low, due to these high operational costs.
Cars require much lower investment and operating costs, have a higher commercial return and experience generational upgrades on a higher frequency.
Until that generational leap is achieved in propulsion that allows orbit to be achieved for a much lower cost, it will remain a venture for Governments and the Super rich.
I do hope Reaction engines get their pre-cooler to work, it will be a real game changer in the industry.
The SpaceX guys seem to be hyper-sensitive to the prospective issues that would suddenly appear in their faces if they launched a rocket which didn't make it to orbit. Governments can get away with it on occasion but it'd be pretty bad news for the private sector to have their gear blowing up overhead, if only because it seems like a lot of folk at Congress are itching for an excuse to shoot SpaceX down and a nice spectacular failure would be as good an excuse as they're likely to get.
It doesn't surprise me that their automatic safeties are quite severe sounding in those circumstances, and I'm happy to wait and see a great launch.
One of the Falcon's nice features is that the vehicle is clamped to the ground until the computers decide that all the engines are running reliably. So it's possible to start all the main engines, throttle them up, *then* realise something's wrong and cleanly shut everything down again, resulting in an intact, reusable rocket standing on the launch pad.
In most other designs, once you start up the engines, your main failure mode is 'fireball'...
I think you will find that this is pretty standard.
I certainly cannot think of one liquid fueled rocket system that isnt clamped down until all engines are running within specs. For one thing you really do not want to allow a launch to occur when one side is at a higher thrust level than the other. In the case of the shuttle it was held down, and the SRB not started, until the engines were running and the whole structure had swung back after the "twang" (shock) of the main engines starting.
It is of course one of the nice things about liquid fuel systems. Once a Solid Rocket is lit it cannot be turned off.
I can see how blowing the top off a solid rocket booster would cut off the thrust but only by having equal amounts of thrust coming out of both ends, possibly not a good idea if you are sat on the top.
You cannot stop the burning, more of an explosion really, once started as the fuel and oxidiser are intimately mixed. The only way to stop it is to let it burn out or self destruct the booster.
The shuttle SRB uses as a propellant the same chemical composition that display fireworks use to explode mortar rounds in the sky and that provide that real thump you feel in your chest. But display fireworks only use a few grams the shuttle SRBs use 500 tonnes each, good luck with turning that off...
(The miracle is that they can sort of control it in the first place.)
Historically the early part of the has used heavily pre-computed on lookup tables. Each entry was subjected to a trajectory simulation to confirm the settings (for gimal angles and sometime for engine thrust) would keep the whole stack going in the right direction.
It's slow, time consuming and very inflexible if things go wrong that you have not already (expensively and extensively) simulated. But all that *recurring* work (for every flight) is expensive.
AFAIK SpaceX are looking at a full on board GNC. You give it a destination and it figures out what those settings need to be.
One of the *main* reasons cited for the delays has been to verify the full software load in a way that NASA is comfortable with. This means the whole tool chain and all the procedures that went into its writing. While Dragon/F9 should be mechanically simpler than Shuttle its on board self testing is pretty extensive.
The original benchmark for the Carnegie Mellon CMM model was the team who wrote the on board software for the Shuttle.
Hopefully Spacex will prove as good at identifying the fault and fixing it as the have proved in the past.
Pity Lewis didn't include a link to the video of the event. I saw it here:
http://www.washingtonpost.com/lifestyle/travel/raw-video-spacex-rocket-launch-aborted-in-fla/2012/05/19/gIQAnDuUaU_video.html
The disappointment and confusion in the controller's voice as he tries to unsay "Liftoff" makes me hurt for him.
That wasn't the controller, that was the NASA PR commentator. The same old guy that emcees the pre/post flight meetings.
Anyway, Must tweeted that it was a bad turbopump valve. They're replacing it and making sure it isn't a common problem to worry about for the other engines.
Oh, wait...
(Somebody had to say it!)
And on a more serious note - I too would far rather have the control system decide to shut things down safely on the ground and leave me with a usable-in-future rocket than be picking up spare parts two hundred miles down-range. Particularly if I'm (hah, fat chance) going to be sitting on top of it.
SpaceX's safety record is and continues to be nothing short of incredible, kudos and keep up the excellent work.
This is easily the most exciting project going on in North America right now, maybe even the world (LHC might give it a run for the money). Every piece of news from and about the company makes me giddy. I'm very happy that they are avoiding the trap of "go fever" and doing everything right.
It would save on fuel, yes. But the trade-off is having to use a MASSIVE balloon and a LOT of very expensive lifting gas to get any meaningful load to altitude. A more viable option is simply air-dropping the rocket from a plane at as high an altitude as possible (There's a commercial company trying to get that method "off the ground" for actual LEO missions)
on 12 September 1962 : «We choose to go to the moon in this decade and do the other things, not because they are easy, but because they are hard, ...» Half a century later, activities in space remain difficult - the gravity well still exists and the basic propulsion techniques used to get something into low Earth orbit remain much the same as they were then. For this reason, cooperation - between all countries willing to make the effort - and between public and private interests is necessary. SpaceX's caution is the way to go - sooner or later they'll get their machine into orbit. Slow and steady wins the race....
Henri
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
