FAILURE: SpaceX Falcon I Launch III - August 2

[KEdward5]

Is there a chance this site could get exclusive angles of the failure, the same as with the Sea Launch failure videos and images.

I'll link the non L2 content as an example:
http://forum.nasaspaceflight.com/index.php?topic=6927.0

[Chris Bergin]

Never say never, but I doubt it. Sea Launch is a much bigger company if you take into account of its parents. SpaceX is a tightly knit family. Thus I doubt someone at SpaceX is going to go to the media with new footage when that person might as well stand on his desk and shout "it was me!!"

[psloss]

Anybody in on this telecon?
http://www.spacex.com/updates.php

[hyper_snyper]

Anybody in on this telecon?
http://www.spacex.com/updates.php

Hmm...  If that's all it was I think I feel a little better about it.  On the other hand, they should have caught this.  I'm curious to see that video or a detailed report later on.  Anyway, on to flight 4.

[William Barton]

Am I correct in assuming this has to do with the "grunt" we heard at cutoff on the Merlin 1C test video, and the big billow of flame at the end of the F9 nine-engine test?

[nacnud]

So it was the shut down burp mentioned earlier. That's two down due to staging transients.

[Kolgate]

Yup, I remember Pbreed had suggested this over in the opinion thread. Here's his post:

"The real difference between this mission and the 2nd mission that came oh so close is the new main engine.

At stage seperation both the 2nd stage and the 1st stage are pretty much above the atmosphere. Both are in free fall after MECO, so any anomaly that causes the main engine to burp before the 2nd lights and flies away
will potentially cause the 1st stage to smack the 2nd.
The rumors seem to indicate that this is what happend.

A Regen motor has a lot more places to hide a "burp" or propellants than the ablative motor. Look at the ball of "Stuff" that is burning for several seconds after the Merlin 1C shuts down in the falcon 9 9 engine test.

The time line has  5 seconds from "Approaching Main engine Cut Off" to
"2nd Stage Ignition confirmed."

Only 1 sec from Approaching Main Engine cutoff to "Stage Seperation confirmed"

A burp 1.5 sec after shutdown could smack the 2nd stage pretty hard.

I would dearly love to see the un-broadcast part of the video."

[Lee Jay]

If this "burp" manifested on the test stand, why wouldn't they have been able to account for it in the flight?

[manlymissileman]

That's a very good question, a poster (or a couple) on this forum saw this as a potential problem looking at the video of the test.  Why wouldn't a bunch of professional engineers identify this as a problem.  (perhaps they did, but were given no time to solve it?)

[Herb Schaltegger]

That's a very good question, a poster (or a couple) on this forum saw this as a potential problem looking at the video of the test.  Why wouldn't a bunch of professional engineers identify this as a problem.  (perhaps they did, but were given no time to solve it?)

Most likely, they never measured the thrust of the post-shutdown "burp" and so didn't take the resultant acceleration of the near-empty stage into account.   This could easily be the result of something as simple as data measurement routines that stop recording data at the engine shutdown command, or of a control algorithm that wasn't updated to take such data into account.

Again, not to beat the dead horse further into glue, but if the failure is indeed due to little slips likes these possibilities, it goes back to the idea that these lessons have been learned by everyone else in the industry of decades of hard work.  Tossing all those procedural-type lessons aside in the interest of doing things in a new way is not always the best approach.

[Hootz]

So the plan will be to extend the staging gap if I understand it correctly? If they go with too big of a gap, how long can a rocket run under minimal thrust before there is going to be a repeat of the "slew" when the second stage finally fires?? By no means an engineer, just curious.

[jongoff]

That's a very good question, a poster (or a couple) on this forum saw this as a potential problem looking at the video of the test.  Why wouldn't a bunch of professional engineers identify this as a problem.  (perhaps they did, but were given no time to solve it?)

Most likely, they never measured the thrust of the post-shutdown "burp" and so didn't take the resultant acceleration of the near-empty stage into account.   This could easily be the result of something as simple as data measurement routines that stop recording data at the engine shutdown command, or of a control algorithm that wasn't updated to take such data into account.

Again, not to beat the dead horse further into glue, but if the failure is indeed due to little slips likes these possibilities, it goes back to the idea that these lessons have been learned by everyone else in the industry of decades of hard work.  Tossing all those procedural-type lessons aside in the interest of doing things in a new way is not always the best approach.

Actually, if you read what Elon said in the quoted article, he said that due to the fact that there's much higher back pressure on their sea level test stand than there is at staging, the data they got on the shutdown transient underestimated the effect at altitude.  Remember, the big billowing flame at the end is more due to the fact that you're firing in air, which has plenty of oxidizer to oxidize the rest of that now very fuel rich flame.  The actual thrust transient was probably much harder to get an accurate measurement of on a sea level test stand.

Also, Paul mentioned that theory after Henry Spencer mentioned it on aRocket, and after it came out that there was the rumor of the stage recontacting after separation.  This wasn't a case of someone looking at it beforehand and saying "golly I hope they handle the staging right".  Hindsight is 20/20.  It's not like the primes never make stupid mistakes like this.

Of all the "easy to solve" errors it could've been, this is one of them that I think makes SpaceX look least bad. 

Now, what I wonder is how many other "minor" "issues" came up on the flight.  Stuff like why do they keep having issues with the Helium fill taking a lot longer than expected?  Why are they always tripping abort sensors during startup?  While I think there's a good chance that had the timing been right they would've made it farther, I'm sure there was still a decent length squawk list, and I hope they make sure to be careful before next time.

~Jon

[iamlucky13]

Geeze...this still seems like something too big to miss to me. However, according to the spaceref article:

When we tested at sea level we had higher ambient pressure than what you have in a vaccuum, so this effect was effectively masked on our test stand during testing.

Ok, the pressure difference is true, but still, they saw that low-pressure billow of flame on the test stand. Surely they recognized it was different on the regen engine than on the ablative one? I guess either that didn't trigger any warning bells, or they were ignored.

I'm wondering how much thrust that is? Using a full 14.7 psi is probably way too high and I don't know the nozzle size, but guessing 0.5 m I get about 4500 pounds.

I also wonder if they're completely confident in their assessment. I guess if they've got accellerometer data that should be pretty conclusive, but they don't seem to have shared how they reached that conclusion.

As far as the question of whether increasing the staging delay will allow too much rotation to occur: the drag should be fairly consistent so that I think it may be possible to give the rocket an appropriate rotation in the opposite direction at MECO if this is expected to be a problem.

[Patchouli]

Good to hear it's something simple but I still think they should maybe had some means to give the second stage more of a push away from the first stage.
Since the recovery parachutes got cooked by the second stage engine firing after it made recontact maybe they should add a couple of stage separation motors.
 
Either put them on the first stage to pull it away from the second stage or put them on the second stage to give it a kick to get it away from the first stage before the Kestrel starts so the chutes don't get fried again.

I'd still do a dummy payload test to make sure it's fixed before carrying a satellite.
 But since nothing exploded I guess using a dummy stage to sort out the staging is no longer necessary.

It's a good thing that the F9 second stage has a full RCS so it should be a lot less likely to do what the second F1 did.

But they should also take that burp in account on F9 and maybe add some separation motors to it just to be safe.

I also wonder if the Ares Ix or later Ares Is could experience a similar failure since a RSRM is still making some thrust normally when it's discarded.

[Jim]

Good to hear it's something simple but I still think they should maybe had some means to give the second stage more of a push away from the first stage.
Since the recovery parachutes got cooked by the second stage engine firing after it made recontact maybe they should add a couple of stage separation motors.
 
Either put them on the first stage to pull it away from the second stage or put them on the second stage to give it a kick to get it away from the first stage before the Kestrel starts so the chutes don't get fried again.

Only a clueless person would suggest a significant design change,  when  simple change to a software constant will do the trick


You don't understand the design changes, the impacts to logistics (spacex doesn't have SRM storage facilities at kwaj), impacts to GSE (the need to access the vehicle at the pad), safety impacts and documentation impacts for licensing and range use, it would take to add ullage rockets or retros.   It is not just snap your finger and it happens.

[Herb Schaltegger]

Actually, if you read what Elon said in the quoted article, he said that due to the fact that there's much higher back pressure on their sea level test stand than there is at staging, the data they got on the shutdown transient underestimated the effect at altitude.  Remember, the big billowing flame at the end is more due to the fact that you're firing in air, which has plenty of oxidizer to oxidize the rest of that now very fuel rich flame.  The actual thrust transient was probably much harder to get an accurate measurement of on a sea level test stand.

Again, this is a lesson that anyone who's actually analyzed performance data for any type of jet or rocket engine should just KNOW from undergraduate studies even if not past institutional knowledge.  Missing the significance of something like this is a common "oops" tossed into an exam by sneaky professors. 

The fact that they saw the burp at all and even CONSIDERED it, yet failed to realize that the magnitude of the resultant force would change under different ambient conditions shows they don't have enough "greybeards" on their payroll.  An AE professor or consulting propulsion engineer hired to evaluate their test data and design plan for staging would have caught that in an instant, for a lot less money than the cost of a failed launch (let alone the harder-to-estimate loss of prestige and confidence from yet another failure).  This is exactly the type of probem caught and dealt with in the industry-standard PDR/RID process (you know, that expensive, wasteful old-fashioned way to fly vehicles).

[StuffOfInterest]

I know there are some pretty big vacuum chambers out there.  Is there any which could handle the test of shutting down an engine the size of the F1 (after a second or two of burn) to measure any residual thrust or reactions in a near vacuum environment?  My first thought is that even one or two seconds of firing would put enough pressure in the chamber to negate any testing.

[Jim]

1.  I'd still do a dummy payload test to make sure it's fixed before carrying a satellite.
 But since nothing exploded I guess using a dummy stage to sort out the staging is no longer necessary.

2.  It's a good thing that the F9 second stage has a full RCS so it should be a lot less likely to do what the second F1 did.

3.  But they should also take that burp in account on F9 and maybe add some separation motors to it just to be safe.

4.  I also wonder if the Ares Ix or later Ares Is could experience a similar failure since a RSRM is still making some thrust normally when it's discarded.

1.  A dummy stage would be more work and it wouldn't show the problem.  The stage would have to be exactly like the real one to be of benefit.  It is called test like you fly.   Also the stage couldn't be a dummy since the 2nd stage has the guidance system

2.  Nonsense and clueless statement .  A.  RCS has no bearing on this failure  B.  The F1 has a "full" RCS too. 

3.  It is not needed.  The D-IV doesn't have them.  "Making sure" is doing your design right and not adding unneeded hardware.  That is called rocket science. 

4.  Ares I has retro and ullage rockets

[Herb Schaltegger]

I know there are some pretty big vacuum chambers out there.  Is there any which could handle the test of shutting down an engine the size of the F1 (after a second or two of burn) to measure any residual thrust or reactions in a near vacuum environment?  My first thought is that even one or two seconds of firing would put enough pressure in the chamber to negate any testing.

You don't need to measure it in a vacuum.  As noted in my post above, ambient conditions are just part of the equations used in the design process, and those conditions change with every millisecond of flight.  Experienced designers realize this without even really thinking about it.  They know cutoff conditions for any kind of trajectory will be different from launch conditions and design accordingly.

[Jim]

I know there are some pretty big vacuum chambers out there.  Is there any which could handle the test of shutting down an engine the size of the F1 (after a second or two of burn) to measure any residual thrust or reactions in a near vacuum environment?  My first thought is that even one or two seconds of firing would put enough pressure in the chamber to negate any testing.

not needed.  other first stage engines are not tested in vacuum chambers.   Also it is not just the engine, it is the propulsion system, which includes the stage.