Falcon doom found: a whisper of thrust

By Phil Plait | August 7, 2008 10:30 am

Elon Musk of Space X has released the news of what happened to the Falcon 1 rocket that failed to reach orbit last week.

The problem arose due to the longer thrust decay transient of our new Merlin 1C regeneratively cooled engine, as compared to the prior flight that used our old Merlin 1A ablatively cooled engine. Unlike the ablative engine, the regen engine had unburned fuel in the cooling channels and manifold that combined with a small amount of residual oxygen to produce a small thrust that was just enough to overcome the stage separation pusher impulse.

We were aware of and had allowed for a thrust transient, but did not expect it to last that long. As it turned out, a very small increase in the time between commanding main engine shutdown and stage separation would have been enough to save the mission.

In other words, the stage separation didn’t happen because the mechanism that splits the two stages was still feeling thrust from a little bit of leftover fuel the pneumatic pusher that physically pushes the two stages apart is rather weak. It works, but the still-accelerating first stage overtook the too-slowly moving second stage, so the separation didn’t work. (My thanks to a mole at Space X for correcting my misinterpretation; and if it’s still wrong it’s my fault not his/hers). That mechanism needs the rocket to be coasting, not accelerating, so it didn’t work. Without stage separation, the rocket couldn’t go up anymore, and it fell back to Earth.

The third Falcon 1 launch attempt in Auguust 2008

Musk goes on to say they are confident they understand the problem, and can solve it by simply waiting a little longer after first stage engine cutoff before separating the stages. That should allow the extra propellant to burn off.

The fourth Falcon launch is already prepping, and may happen as soon as September.

CATEGORIZED UNDER: Space

Comments (43)

  1. IBY

    Yaeh!! Go Falcon! I really hope they succeed next time. The stage separation thing was unfortunate. Good thing it was only a minor mistake.

  2. wade

    It’s amazing to think of all the work, time, energy, thought put into this and then this one little glitch ruins the whole endeavor.

  3. madge

    I’m SO glad they figured the problem out. Science in action. If at first you don’t succeed look at the data and make the necessary fine adjustments. Better luck next time SpaceX. We are all (even us Europeans) rooting for you. :)

  4. Ryan

    Congrats to them on such a quick analysis. Hope to see the next one soar!

  5. madge

    @ wade
    Not quite “the whole endeavour” ruined. Elon Musk has affirmed his commitment to keep going until SpaceX succeeds. Onwards and Upwards :)

  6. drow

    heh, oops. the next one should be brilliant.

  7. almurray1958

    Video is posted behind photo in update section, or full url is here http://mfile.akamai.com/22165/wmv/spacex.download.akamai.com/22165/F1-003.asx

  8. Thomas Siefert

    Musk goes on to say they are confident they understand the problem, and can solve it by simply waiting a little longer after first stage engine cutoff before separating the stages. That should allow the extra propellant to burn off.

    That is a classic blind running program, I want to know exactly what has happened in the real world before I continue to the next stage, not what should have happened.
    Install an acceleration meter for FSM’s sake.

  9. Phy

    SpaceX posted a link to a video of the launch on their page, here it is if anyone missed it: http://mfile.akamai.com/22165/wmv/spacex.download.akamai.com/22165/F1-003.asx

    (will require something that can stream Windows Media files)

    At the end of the video: BONK

  10. Togan

    I sure hope they name one of their rockets “Millennium” … ;)
    But please not just a testing thing that’ll never reach orbit, name a real one after a real cool spaceship!

  11. Gary Ansorge

    AH, the old “Oops, sorry ’bout that,,,”.

    Gee, they must be scientists over there. Always so ready to admit to a mistook,,,I guess absolute certainty is restricted to the “faithful”.

    Now that they know the problem, WILL they be adding an accelerometer to the mix. It can’t be THAT expensive in terms of fuel, can it?

    GAry 7

  12. Well, that’s good news coming out of bad. I’m amazed that they didn’t take that long to sort it out.

  13. Bob Brashear

    Timing is so damned important and can be a stone-cold b*%ch to figure out. My sympathy and hope that the next one gets to orbit.

  14. stober

    “Both halves of the rocket then fell into the Pacific Ocean well east of the U.S. Marshall Islands and were destroyed along with its payload of two small NASA satellites and the Trailblazer demonstration satellite for the Pentagon. A container containing the cremated remains of people, including those of astronaut Gordon Cooper and actor James Doohan of television’s “Star Trek”, who had paid to have their ashes launched into space was also lost, according to the space memorial firm Celestis, Inc.”

    Guess you really can’t change the Laws of Physics…

  15. amphiox

    Any time I see or hear about one of these attempts failing, I think back to all the footage of NASA and army rockets blowing up on the launch pad or some other spectactular fubar from the 50’s and 60’s.

    If it were easy to get into space a whole lot of us would already be there today.

  16. amphiox

    I also am reminded of the Chinese astrologer in the Ming Dynasty who strapped 40 bamboo rockets to a chair, sat down, lit them, and according to chroniclers “was taken to another world in a flash of light.”

  17. Cheyenne

    He should have used 40 helium weather balloons!

  18. Nic

    Nice video. The Falcon does seem slightly ‘twitchy’ in roll, it seems to oscillate very slightly between roll limits.. I haven’t the faintest idea if that’s important though.
    It’s nice that SpaceX have already nailed down the cause of the problem – but it’s such a shame that they seem to be plagued by some really bad luck.

  19. NoAstronomer

    A superb example of why we refer to Rocket Science in one of best known English aphorisms.

  20. Michelle

    It fell back to earth… did they recover anything?

  21. ioresult

    They should use the accelerometer feed to detect coasting and separate only then (with a timeout value of course to prevent false positive values of accelerations). Trusting the whole mission to a blind timeout is unwise in my profane opinion.

  22. Torbjörn Larsson, OM

    Classic engineering mistake (or rather, what engineers wants to avoid but management always pushes for) – changing a working component to something “better” or “cheaper” without enough understanding of the component or its relation in the system.

    Space-X may have a low risk concept, but as Thomas Siefert’s comment implies, they may have a high risk project approach. Let’s see if they can get two clean stage burns next time, and keep the fingers crossed for satellite delivery. (I presume they will test that even if the 4th launch is claimed to be a dummy load.)

  23. Torbjörn Larsson, OM

    Trusting the whole mission to a blind timeout is unwise in my profane opinion.

    Sensors and electronics are cheap today (at least the non-space variants) so it seems like a logical choice. OTOH basic failure theory tells you that you want to cut down on the number of components and functions, their complexity, et cetera. Timing may ultimately be more reliable, so it may be wise to test that first.

    But that doesn’t contradict that they should pepper their launcher with harmless diagnostic sensors, if they haven’t already.

  24. Bigfoot

    I have high hopes for them but we have to guard our expectations. After all, some of the engineers are probably suffering from stage fright, and the remainder are likely afflicted with separation anxiety …

  25. Philip

    first I must say that I admire the openess with which Elon diskusses the troubles encountered. A very educative process.

    I wonder wether it is cheaper for SpaceX to risk a vehicle for maturing design rather than invest in another thousands of engineering hours (firing rockest in vacuum chambers a.s.o.) to investigate every single possible failure mode.

    At this stage of operations, having vehicle 4 near to the pad, I’d rather understand the residual thrust problem and set timing accordingly than introducing a new staging initialising system with accelerometers. Maturing that will be a headache in itself.
    K.I.S.S.

    and
    All the best for flight 4.

  26. Buzz Parsec

    Somewhere (spaceflightnow.com?) said that that the residual overpressure in the thrust chamber was about 10 PSI, or 2/3 sealevel atmospheric pressure, which is very hard to measure accurately during a static test. They knew there was residual thrust, but they didn’t know how long it lasted before it dropped low enough to not be a problem. They guessed based on their test data, but they guessed wrong.

    I don’t know if it is possible to test this on the ground. To get an accurate reading, you would need to run the engine for at least 5-10 seconds to get it up to working temperature, turbopumps running at speed, etc., before shutting it down, which would require an enormous vacuum chamber or the else the chamber would fill up with exhaust and become a pressure chamber instead. Remember, the engine burns hundreds or thousands of pounds of propellant per second, and that all has to go somewhere!

    The easiest way to test it would be to somehow get it to about 40 miles altitude and shut it down there. I.E., to launch it. :-) As long as they’re going to launch the 1st stage, they might as well put a second stage on top of it and (hopefully) test that too, if the 1st stage works. And if they’re going to launch two stages, they might as well put a test satellite on top, since it’s going to orbit if it all works.
    Which is what they did :-)

    When NASA did this with Apollo 4, there was a lot of controversy. Some people wanted to launch a couple of test flights with just the 1st stage live, before testing the 1st and 2nd stages and then all 3 stages, and then with a live Apollo capsule. That is the procedure they followed with the Saturn I and Saturn IB. But the reasoning was that it the 1st stage worked, you would miss an opportunity to test the upper stages, and if it broke, the delay to fix it would be independent of whether there were live or dummy upper stages. They’d lose some engines, but a big tank full of water (what they usually use for ballast in dummy upper stages) would cost almost as much to build as a real tank full of LH2 and LOX. They would still need a guidance system (atop the 3rd stage) and a dummy Apollo with a live escape system for accurate aerodynamics, and the 3rd stage and Apollo had already been tested on Saturn IB flights. If worst came to worst, it would be an extra opportunity to test the LES.

  27. gopher65

    *points up* I agree with what Buzz Parsec said.

  28. t@ Those in favor or more diagnostic tools,

    I understand what you’re saying, but here’s where I think you may be wrong:

    You don’t know what’s going to go wrong until it does.

    Therefore, you have to treat every aspect of launch as being equally susceptible to failure.

    Therefore, you have to equip every aspect of the launch that can go wrong with a cheap sensor.

    Therefore you lose man hours, resources, and time installing and calibrating each and every cheapo sensor, which in the end adds up to the total cost of a very expensive rocket which you will not know for sure will succeed.

    Therefore when something does go wrong, Which you are pretty sure it will otherwise you would have put someone/something important inside, you end up with a wealth of useless sensor information which will take weeks to sift through, if sensors are your diagnostic of first resort. Never mind the fact that once you have a certain number of feedback mechanisms, you have increased the chance that one of them is going to feed you false or misleading information. If they are not your diagnostic of first resort and can only add a certain amount of extra useful information to solving a problem, then it may be reasonable to restrict feedback monitoring to only the most crucial elements.

    It’s the same when trying to debug a program, throw in too many checksum requests and sure you’ll end up with a perfect bug-free program, after about five hundred years. To quote a recent movie with lots of bad science in it:

    “Sometimes you got to run before you can walk.”

  29. Joey Joe Joe

    @madge:

    “Elon Musk has affirmed his commitment to keep going until SpaceX succeeds. Onwards and Upwards :)”

    I just hope they succeed before the money runs out and, perhaps more importantly, they give the Falcon a bad rep. After all, how many companies will want to launch on a rocket that has had 4-5 failures before its first success, especially when there are alternatives with good track records?

    I really think they are going to have to nail the next flight. I certainly hope they do, which is why I’d hate to see them launch on a hunch and get it wrong. It would be a pointless waste of what seems like a very reliable powerplant (Merlin).

  30. Al Navarro

    From the Falcon 1 flight 3 press kit (page 8 – which includes “Stage Separation Reliability”):

    “Thrust vector control is provided by electro-mechanical actuators on the engine dome for pitch and yaw.”

    This implies to me that the attitude was controlled electronically. I would think that a fly by wire system with accelerometers and rate gyros with electro-mechanical actuators would be lighter (lighter always = cheaper in this case). The standard IMU in this case would include accelerometers and rate gyros or the controller would have trouble keeping it’s trajectory.

    It seems very surprising that they wouldn’t have been able to measure:

    while(current_vehicle_acceleration > PNUEMATIC_PUSHER_FORCE)
    {} //wait for first stage to cool it’s jet

    (may be current_vehicle_acceleration_z or current_vehicle_acceleration_x ;)

    {I tried to come up with a more detailed description of the control system. I thought there was a reference on the SpaceX website, but I couldn’t find it. So, my apologies to Umption, this may all be crap.}

  31. yeliabmit

    This is the kind of thing that Scotty could have warned them about, I’m sure.

  32. Newscientist has been doing some first rate reporting on the Falcon and the overall managerial issues at Space-X: http://www.newscientist.com/blog/space/2008/08/spacex-rocket-failure-due-to-new-engine.html?DCMP=ILC-hmts&nsref=specrt12_head_Rocket%20collision

    There is also a link to a previous post regarding the failure. Frankly, I am becoming worried. Some class mistakes are being made here and there is no reason they have to be made.

  33. Why can’t the second stage fire even without separation? If failure to fire guarantees mission failure, then they don’t have anything to lose by lighting the engines and seeing what happens…

  34. quasidog

    Bummer.

    As a comparison, I wonder how the failure rate is next to other rocket tests, like that of NASA, and the early Russian rockets. I’m sure the older ones had a higher failure rate in tests, seeing as they were treading new ground. These new mission have hindsight to assist, but I wonder how they compare for failures to the older ones.

  35. Spiv

    Nic: Yeah, people get used to seeing NASA’s real-time roll control systems. It’s not a huge deal at all, but I noticed that too while watching the videos. They seem to have plenty of control over the thing, just a broader range of acceptable positions.

    quasidog: The old nasa tests were something in the range of 1 in 5 were failures. Not exactly a confidence booster for the guys who were getting ready to climb onto them for the first manned flights.

    For a bit more explanation, the spaceX engines are a regenerative system- they have fuel pumping through the engine bell structure for cooling (and pre-heating of the fuel prior to ignition). This makes for a more efficient engine, but at the expense of a slower shutdown time. Fuel continues to burn in the bell longer, even if at a very low pressure. They didn’t catch the pressure at ground tests because it was being pushed against 32psi, but at that altitude 10psi becomes a lot more thrust than expected.

    There’s a forced separation (think spring loaded), then a timeout to let the first stage clear the second stage, then second stage ignition. The first stage hit cut, separated, and then both were supposed to coast apart for a few seconds. Since the first stage still had a bit of thrust it actually came back and collided with the second stage.

    You can see all of this on this awesome video that was posted on NASAWatch: http://www.nasawatch.com/archives/2008/08/flacon_1_flight.html#more

    Go Space-X, the guys at NASA are rooting for you. Next launch sounds like it might be the one!

    So in short, this is rocket science. It’s hard stuff, and plenty of unexpected variables. That sai

  36. Bo Babbyo

    “As a comparison, I wonder how the failure rate is next to other rocket tests, like that of NASA, and the early Russian rockets.”

    I was wondering the same thing.

    I’m too lazy to look up the actual numbers, but I’m willing to bet that the USAF and NASA lost DOZENS of Atlases alone. Maybe more rockets that will be built in the lifetime of the Falcon program.

  37. Lary

    “Why can’t the second stage fire even without separation? If failure to fire guarantees mission failure, then they don’t have anything to lose by lighting the engines and seeing what happens…”

    The stages did separate. The problem was that after separation the first stage caught up to and recontacted with the second stage. The second stage did fire fire at that point but the nozzle was still inside the first first stage. It was not designed for that and bad things happened.

  38. Does this mean that, because the separation failed, the next stage failed to light, or that it lit and then failed because the stages were still attached?

  39. Torbjörn Larsson, OM

    Some people wanted to launch a couple of test flights with just the 1st stage live, before testing the 1st and 2nd stages and then all 3 stages, and then with a live Apollo capsule.

    I think it makes sense in this type of systems to launch the whole enchilada. You don’t test a car with a working front end but a mock up (well, monocoque mock up in a modern car) back end either.

    The problem with building systems is to know when to upgrade. It’s very tempting to push late improvements into demos because it is a lot easier to change before freezing the design. But this was a late demo and they payed the prize for it.

    @ The Chemist:

    here’s where I think you may be wrong:

    Your list makes a lot of sense. However, seeing the inexpensive and unobtrusive type of multisensor systems that you can use today, it is probably a lot easier to install them, record all of them and pick out the information when you need it.

    For example to understand failure – and perhaps that is exactly what SpaceX has done, all they needed where a conveniently located camera, accelerator or pressure sensor to confirm what happened. I have at times followed another startup, Armadillo Aerospace, which have done this type of sensor integration to great effect.

    Dunno about the calibration need, modern cameras and accelerators won’t need much if any.

  40. AJ

    @ Ken B.: That’s something I’m puzzled about too…

    anyone?

  41. Lary

    @Ken B & AJ
    “Does this mean that, because the separation failed, the next stage failed to light, or that it lit and then failed because the stages were still attached?”

    Let’s try again. The separtion itself did not fail. They had separated then the first stage bumped back into the “attached” postion due to the residual thrust that wasn’t accounted for in the separation timing. At that point the stages were nested together but not physical connected. Then second stage lit while it’s nozzle was inside the first stage. Resulting in damage to both the first and second stage. I don’t know the details of the damage but it was enough to result in the mission failure.

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

ADVERTISEMENT

See More

ADVERTISEMENT
Collapse bottom bar
+

Login to your Account

X
E-mail address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »