The SpaceX Launch Anomaly

I just got back from dinner, but apparently people have begun discussing it in comments at this relatively unrelated thread.

I don’t know much about it, so all I’ll say is that if they had a first-stage engine problem, it proves out their design capability to have mission success with engine out.

Clark Lindsey has some notes from the post-launch press conference.

[Update a few minutes later]

I would go further, and state explicitly that this increases my confidence in their vehicle. It was an inadvertent flight test of a capability that they would hope they don’t have to use, but now know that it works.

[Update a few minutes later]

Never mind my first response to Paul Breed in comments. Apparently it wasn’t just an engine shutdown — a first-stage engine exploded.

That puts a different complexion on things.

Now the question is, what is different about that Merlin than the one on the second stage, other than nozzle size? Well, there’s the environment. It occurred during Max Q, so the second stage wouldn’t have to worry about that.

If it had happened on the second stage, would it have just resulted in a lost of thrust (that is, does the second stage have a similar shrapnel protector) or would the stage itself have exploded?

If the former, it would probably be survivable (that is, payload or crew recovered) without an abort system, though a mission failure, so a concern for payload customers. If the latter, it might have happened too quickly for an abort system to be activated. It all depends on how much warning they had that the engine was going south. But obviously this would be of concern for commercial crew. Of course, as discussed in comments, this engine version will be retired after the next flight, currently scheduled for January, so if that one goes all right (and they certainly won’t fly until they understand what happened to this one), it will have retired with a 98% demonstrated reliability, and the Merlin 1D will have a clean slate on the following flight.

[Monday morning update]

Charles Lurio notes in email that one difference besides nozzle size (and environment) between upper and lower stage engines is the lack of active cooling in the nozzle of the latter. This would only be significant if the failure was caused by a problem in one of the cooling channels.

50 thoughts on “The SpaceX Launch Anomaly”

  1. Sorry for the threadjack Rand. Boys will be Boys!

    Elon has issued an official statement:


    “Falcon 9 detected an anomaly on one of the nine engines and shut it down,” Musk wrote in an email… “As designed, the flight computer then recomputed a new ascent profile in realtime to reach the target orbit, which is why the burn times were a bit longer.”

  2. That is 2.7% failure rate 1/(4 flights x 9)
    A failure of the single M vac motor (largely common) would mean mission failure. Being unable to recover the first stage makes diagnosing this hard….

    Nice to see that the shrapnel shields work…

    1 First stage failure is ok.. 2nd is not, one second stage failure is bad.

    So assuming 2nd stage is as reliable as the first stage has demoed…

    2.7% failure rate would be worse than shuttle… do they de-rate Merlin Vac?

    Is there any reason to assume its more reliable than the First Stage Merlin?

    1. How do we know the shrapnel shield works? Was it even necessary? It wouldn’t have been if it was a commanded shutdown.

      We don’t know what the tolerances were on whatever caused the first-stage shutdown. They could be different for the second stage.

        1. That being the case, they’ll have no flight experience with it. So right now the Merlin 1C demonstrated 97.5 percent reliability in flight. The next flight, assuming success, will give it a 98% reliability at retirement.

    2. Ignore above comment — just heard about the explosion.

      Shouldn’t it be one in forty, or 2.5%, counting the upper-stage engines? And actually, for one in thirty six, it’s 2.8%, rounded.

      1. And realistically the engine is still not mature technology – so it’s a pretty reasonable failure rate with room for improvement.

    3. Virtually all mature space launch systems, including the Shuttle, converge on a ~2% total failure rate per flight – and there isn’t enough data to justify arguing the second digit. There’s nothing about SpaceX or the Falcon that would render it immune to the effects that wreck one out of fifty of everyone else’s orbital launch attempts, so I’d expect that’s about where they will wind up. Their vehicle is fairly conventional, just inexpensive, and they share launch sites and range safety with the rest of the industry.

      If a SpaceX catastrophe costs half as much as a ULA catastrophe (and we narrowly averted one of those a few days ago as well), that’s a net win.

  3. Any chance it had something to do with max-Q (it was an outside engine) or was that probably a coincidence?

  4. The rock throwing has already started in a comment over at Clark’s, with someone calling “it”[F9] “too cheap” thus low quality.

    What do you think the established launch companies will be whispering into customer ears now – or perhaps shouting? This sort of crap was anticipated in the sept 17 Av Week issue in which they took pot shots at the F9 v1.1, and the ‘unprecedented’ commercial flight reservations on it despite not having flown.

    All that being said, of course we all want to know what went wrong and be assured that it’s understood before the final flight with the Merlin 1C’s. And there will certainly be more anxiety over the multiple changes in the v1.1.

    If this were a rational world, people would recognize that getting _any_ non-reusable system to work the first and every time is well-neigh miraculous.

    Instead we have the self interested in the established industry and the porkers on the Hill who will be out for blood.

    (No doubt Mike Griffin will be again whispering that his Ares would’ve worked perfectly because he knows all.)

    1. You could retort with videos of all the rockets that failed after an engine explosion (which would pretty much be a compendium of all prior engine explosions).

      For one, I’d be much more willing to ride on a Falcon-9 after this event, because in the back of everybody’s head is the thought “What if an engine explodes?!” Well, in this case it slightly improves the mass ratio. They should document it and call it a feature.

      But a seocnd-stage engine failure is definitely more cause for concern, as Rand pointed out.

      Following the bizarrely ingenius logic use to armor British bombers in WW-II (armor them randomly and see which ones come back, then move the armor to cover the spots that weren’t covered on the lost bombers), I’d check the engine orientation vs. airstream exposure on the lost engine and rotate that one 90 or 180 degrees for the January flight. SpaceX has a sweet automated pipe-bender, so bend a bit of plumbing and double check the 1D designs.

      1. re: British Bombers

        The version of that I’d heard was that they were adding armor to a newer model, when someone realized at the last minute they were doing it all wrong – the RAF had kept meticulous records of where all the planes were hit, and they had planned to add armor to the most frequently hit areas. In the end, they added armor to the areas that were almost never hit – because they realized that they only had a record of damage on the planes that made it back.

      2. While the engine could have exploded, it could be what we see on the video is the engine bell breaking off after the engine shuts down. That kind of condition would not be unexpected while they are around Max Q, since the bell would be ‘depressurized’ once the engine has stopped.

        We’ll have to wait and see what SpaceX says.

      3. Thanks for the details on the bomber story. I knew it was an interesting insight into sampling errors (the vital spots can’t be in the observed sample).

        Depressuriing the engine bell at max-Q is an interesting thought. All the compressive loads between the bell and the throat become tension loads carried by the fasteners, which are also put in shear by the asymetrical aerodynamic loads from the side. If the flight data shows the anomally started in the turbopump and the RUD occured after automatic shutdown, the failure isn’t nearly so worrisome and is an after effect of a hot engine shutdown at max Q.

        1. That’s an interesting scenario and quite possible. We should know soon enough once SpaceX has had time to analyze the telemetry and video evidence.

    2. They don’t have a spotless record, and very few flights. What else does a customer really need to know ?

      1. That the Falcon 9 delivers the payload to its appointed orbit through rain, sleet, wind, and dark of night. That neither fire nor explosion will divert it from its task. A Falcon rocket can’t be stopped, can’t be reasoned with. It takes horrendous damage without flinching and just recomputes the optimal strategy to carry out its mission. It isn’t like other rockets that disintegrate when they get hit with styrofoam or blow up when their rubber O-rings get the shivers. It never falters. It never stands down. It doesn’t feel pain. It doesn’t know fear, and it doesn’t feel remorse as it drives its competitors into bankruptcy.

        Once SpaceX figures out how to disguise 157 foot Falcon 9’s as the Stay Puft Marshmallow man and send them walking through our cities, it’s game over.

          1. Yes, based on their number of flights they shouldn’t have a single successful launch yet, much less any contracts, much less NASA contracts, much less being swamped with private contracts.

            I’ll admit my prose is poor, and add:

            To launch, or not to launch–that is the question:
            whether ’tis profitable that we suffer
            What traditional launch providers charge us,
            Or to stake all upon a fresh new startup
            and by supporting end them. To fly, to leap–
            once more–and by a leap to say we cut
            expenses and the thousand tacked on costs
            the bastards charge us. ‘Tis an innovation
            accountants do so wish. To fly, to leap–
            to leap–perchance to dream: ay, there’s the rub,
            For in those steep launch costs what dreams may come
            whem we’ve hae squandered all our worldy toil
            on a contract, There’s an aspect
            that makes a calamity of such spending.

        1. Ha, wonderful!

          In thunder and in earthquake she comes, like a Jove.
          That, if her engines fail, ascent she will still compel;
          And bids the heavens, in the bowels of the Lord,
          Deliver up orbit, and receive mercy.

          Her metals are alloyed with equal parts Chuck Norris and sang froid. Damn the engine explosions, full speed ahead!

        2. ” A Falcon rocket can’t be stopped, can’t be reasoned with. It takes horrendous damage without flinching and just recomputes the optimal strategy to carry out its mission. It isn’t like other rockets that disintegrate when they get hit with styrofoam or blow up when their rubber O-rings get the shivers. It never falters. It never stands down. It doesn’t feel pain. It doesn’t know fear, and it doesn’t feel remorse….”

          They should rename it to the Spartan.

          1. Once SpaceX figures out how to disguise 157 foot Falcon 9′s as the Stay Puft Marshmallow man and send them walking through our cities, it’s game over.

            Little did the SpaceX engineers suspect quite what they had created, until the day of the fateful computer “malfunction”. The Grasshopper stretched its insectile legs, and ambled off the testing pad, pursuing and incinerating the hapless townsfolk of Van Horn with its LOX-Kero “heat ray”…

        3. “That the Falcon 9 delivers the payload to its appointed orbit through rain, sleet, wind, and dark of night.”

          what about the orbcomm it didn’t deliver to the correct orbit? If this was a commercial satellite insertion mission only, it would be pure failure. Instead they had 2 engines not do their job, of 10 total. Shutting down an engine by explosion is not a success. It is a failure of redline and control systems. This was not a test flight, but announced as the first operational flight. I’d like to see an astronaut pool poll of ready and willing to sign up

      2. …that each flight they have had with the Falcon 9 has been a success? Isn’t that the ultimate measurement that a customer is looking for?

        Sure, flight anomalies provide feedback on a design, and customers can decide for themselves if they want to switch to “Plan B” – ask the customers who use the Proton rocket about that.

          1. Regarding Orbcomm, you have to remember there are two different customers on this flight.

            The primary mission is the Dragon flight to the ISS, and that was unaffected by the engine failure.

            The secondary mission was the Orbcomm satellite release, and from what has already been stated publicly the 2nd stage malfunctioned and the satellite ended up in the wrong orbit. However that was unrelated to the 1st stage engine failure too.

            Bottom line is that the 1st stage engine failure did not affect either payloads or missions.

          2. The chatter over at NASASpaceFlight.com on their mission forum is that the 2nd stage ran out of fuel and then ejected the Orbcomm satellite (a new poster with no validation said that he had hardware on the satellite and has inside knowledge).

            Another poster thought he recalled that the TOC’s for secondary payloads as “best effort” orbits, and that this particular Orbcomm satellite is a prototype and might be able to be raised somewhat in orbit to lengthen it’s functional lifetime.

            Guess it just shows we need to wait for all the principles to provide us with the real status.

          3. However that was unrelated to the 1st stage engine failure too.

            You have no basis for that statement.

    3. The rock throwing has already started in a comment over at Clark’s, with someone calling “it”[F9] “too cheap” thus low quality.

      That was the inevitable result of putting all your “new space” eggs in the Elon/COTS/ISS basket. Every time a fanboy said, “If Elon can’t do it, no one can,” they were drawing a big target.

  5. If %2.5 is the odds of any one engine suffering an RUD wouldn’t the odds of one specific engine out of 10 (the upper stage) suffering an RUD be %0.25? The odds of any three engines suffering an RUD at the same time such that the first stage fails would be %0.0016. That looks pretty healthy to me (barring analysis of my rusty math).

  6. Mr. Simburg: Please refrain from using “went south” as an alternate expression for disaster. It sounds puerile. Thanks.
    SamW

      1. Do you remember when we were directed to remove SEX from engineering ?
        Male and female connecters became sockets and plugs . How about the phrase “Packing Heat” for being prepared for trouble ? Turned brown is another.
        By the way, I apologize for mis-spelling Mr. Simberg’s name. Excuse me.
        SamW

    1. Sam, the only puerile thing I’ve read here is your comment. Quit being a self-important little twit.

    2. Britain, please stop using the phrase “go pear shaped” for something going wrong, I quite enjoy pears myself, they are delicious.

      1. Some women I know are rather pear shaped. It sounds as if Britain could be in for a sexual harassment suit.

  7. Wow. What a great thing. It worked as designed and will only get better in the future.

    But the important thing is, my soda bet is safe.

    “SpaceX, we make our own luck.”

  8. What’s with the facebook schlock in the URL? Anyway…

    It is nice to know the rocket can go up with an engine out as designed but it would be better if the thing would not fail at all. Without knowing the cause for failure it is hard to know if this has relevance to the second stage or not. Plus there is the Merlin-1D with the new nozzle and turbopump coming out as other people said so the problems may not apply there.

  9. This is actually great news. It’s never a good thing when an engine fails, but it’s a very good thing when an engine fails and the passenger or payload still arrives safely at its destination. I have the impression that the engines on NASA launches are either successes or failures, on or off, ones or zeros. This launch is like getting on to a multiengine aircraft, knowing that if an engine fails, there are others that will allow a landing (under most circumstances). Like Rand, this incident can only improve confidence in the design – and the designers – of these rockets. It may also illustrate the difference between government and commercial enterprises. The government can afford to keep sending up rockets that fail until they are so failure-proof that they fly. Private companies don’t have unlimited funds so they have to come up with designs that will work even if there is a partial failure.

  10. Elon Musk has stated that future versions of the Falcon 9 will change the engine geometry to put the thrust loads more in line with the tank structure. Does anyone know when this will take place? The reason I ask is that could the new geometry have already been implemented (or is that for V1.1?) and possibly changing the aerodynamic loads on the engine bells?

  11. Turns out that SpaceX wasn’t the only one to have rocket engine problems recently. The Delta IV launch last week had an under-performing RL-10 in the upper stage. They don’t know why it generated less thrust than planned but it achieved the desired orbit by burning longer.

  12. Here’s a statement from SpaceX as mentioned on ARS Technica:

    SpaceX has released the following statement:

    Approximately one minute and 19 seconds into last night’s launch, the Falcon 9 rocket detected an anomaly on one first stage engine. Initial data suggests that one of the rocket’s nine Merlin engines, Engine 1, lost pressure suddenly and an engine shutdown command was issued immediately. We know the engine did not explode, because we continued to receive data from it. Our review indicates that the fairing that protects the engine from aerodynamic loads ruptured due to the engine pressure release, and that none of Falcon 9’s other eight engines were impacted by this event.

    As designed, the flight computer then recomputed a new ascent profile in real time to ensure Dragon’s entry into orbit for subsequent rendezvous and berthing with the ISS. This was achieved, and there was no effect on Dragon or the cargo resupply mission.

    Falcon 9 did exactly what it was designed to do. Like the Saturn V, which experienced engine loss on two flights, Falcon 9 is designed to handle an engine out situation and still complete its mission.

Comments are closed.