LIVE: SpaceX Falcon 9 (Flight 2) - COTS-1 - Launch Updates - December 8, 2010

[sdsds]

Excellent comments, blazotron!  Lots of insight there.  Maybe you can answer a few questions about this.  Forgive me if they seem confrontational; the intent really is to get at the deeper cause of the vent design failure, to see if there is something systematic about it, or whether it was single, unfortunate design oversight.

Use of diffusers is fairly common practice in mundane HVAC applications.  Is it uncommon in aerospace settings?  More pointedly, do other modern space launch systems use diffusers on purge gas vent lines?  Then to the core of the matter:   had SpaceX used a design process that required studying best practices in the industry, would this vent line have been designed initially with a diffuser?

[Hanol]

Assuming the image is cropped (rather than scaled), an 800mm lens on a Canon 5D Mark II (as listed in the EXIF) subtends 1.653 arcseconds per pixel
http://www.howardedin.com/articles/fov.html

The Dragon in the image is about 16 pixels across, so 26 arcseconds for the 3.6m wide capsule.

S=O/H,  H=O/S, distance = 3.6m/sin(26/3600) = 29km

Intuitively that seems too far to me. Please could someone check my maths?

Two small errors. Rounding 16*1.653 = 26.448 to 26 and using sine instead of tan. I get 3.6m/tan(26.448/3600) = 28.1 km.

Also, from the press conference I thought I heard the spacecraft land within "a 100 metres" instead of "800 metres" of the target point.

The whole point is wrong.  Yes, image was scaled by the web server upon rendering.  May be 10 times, may be more.  You never know.  So all the math doesn't make any sense.

I downloaded the full image and worked from that, not the inline server-scaled version. It could also have been scaled, but that generally strips the EXIF data so I'm guessing it wasn't.

Steven: Thanks for the confirmation. Given the vagueness of pixel counting, I'm comfortable with the rounding! Precision is likely +/- 2km.

All pictures on the web are scaled, unless it's a separate download not rendered by the web server to your browser.  Resolution of the Canon 5D mark II is 5616x3744 pix.  Was this what you've been working with?  And scaling done by the server doesn't modify EXIF.  Also if picture was smaller then this, it was either scaled, or was shot using low resolution mode.  So all this arithmetic is just a guess.

[butters]

I think that the original "process failure" comment by Jim and backed up by others was prompted by the initial reports of "weld porosity", which steered people toward the conclusion that good welds don't just fail and that SpaceX should have caught the bad weld during the manufacturing QC process.

Now I'm not sure if welding or porosity had anything to do with the failure. If the initial reports had implicated the GN2 purge vent as the root cause, then I don't think Jim makes that "process failure" comment. That sounds more like a design flaw that wasn't shaken out on the first flight.

It's a bit funny, really, that "design flaw" is considered a much lesser sin than "process failure", which is thrown around as a really serious indictment of an entire organization. Whatever it is, they appear to have gotten to the bottom of it, and while they repaired the nozzle extension using a "process" that many people found unbelievably cavalier, SpaceX didn't see it that way at all -- and they were right.

They have a different process. They trade exhaustive analysis for more flights to shake out the bugs, but they are well-informed by a considerable depth of analysis. Given the relatively benign bugs they've shaken out so far while still managing to rack up a 2/2 flight history with F9, it's a process that seems to be working.

A couple more flights down the line with these initial teething bugs ironed out and a reliable launch vehicle in service with highly competitive pricing, they're going to look really smart for not analyzing everything to death on the ground. Maybe luck had something to do with it, but they've got a winner in Falcon 9, and from what we can tell from the first flight, Dragon is no joke of a spacecraft, either.

[kevinof]

Well it strikes me that Spacex must be doing something right. They launched after quick fix (successfully) , separated the dragon (successfully), orbited (successfully) , reentered (successfully) and splashed down (successfully).

Not a bad day for people that criticize them for having limited or poor processes.

[Pedantic Twit]

Well it strikes me that Spacex must be doing something right. They launched after quick fix (successfully) , separated the dragon (successfully), orbited (successfully) , reentered (successfully) and splashed down (successfully).

Not a bad day for people that criticize them for having limited or poor processes.

That's all pretty awesome, but are the gripper arms meant to fall off a toasted strongback?

[docmordrid]

If replacing hoses, adding diffusers & having to toughen up strongback arms are their major issues after this flight...

[Jim]

If replacing hoses, adding diffusers & having to toughen up strongback arms are their major issues after this flight...

Or maybe the strongback is in the wrong place to survive repeated launches. 

All these add up to costs creep and soon they will no different than ULA.

[MP99]

I think that the original "process failure" comment by Jim and backed up by others was prompted by the initial reports of "weld porosity", which steered people toward the conclusion that good welds don't just fail and that SpaceX should have caught the bad weld during the manufacturing QC process.

Now I'm not sure if welding or porosity had anything to do with the failure.

I was going to post something very similar to this.

They have a different process. They trade exhaustive analysis for more flights to shake out the bugs, but they are well-informed by a considerable depth of analysis. Given the relatively benign bugs they've shaken out so far while still managing to rack up a 2/2 flight history with F9, it's a process that seems to be working.

A couple more flights down the line with these initial teething bugs ironed out and a reliable launch vehicle in service with highly competitive pricing, they're going to look really smart for not analyzing everything to death on the ground. Maybe luck had something to do with it, but they've got a winner in Falcon 9, and from what we can tell from the first flight, Dragon is no joke of a spacecraft, either.

My concern is that we are used to launchers with this depth of analysis behind them. I'm not convinced you can just assume that a couple of flights will have drawn all the bugs out of the design, and that the underlying launcher would then have the same inherent reliability as demonstrated by ULA (for instance).

cheers, Martin

[swervin]

Has there been any word on the 'black box' recovery in terms of preliminary findings/lessons learned?

i would love to SpaceX to be able to at least recover a (albeit, likely damaged) first stage on their next flight. i've read they are working hard on this, and it is the end-state goal.

i've read they have cork-like TPS on it (i think?), where are the parachutes on the first stage and what is the velocity they are trying to hit the water with?

words?

thanks!

[billh]

All these add up to costs creep and soon they will no different than ULA.

Jim, is there another thread on the forum where you expand more fully on your belief that SpaceX cannot possibly improve on the cost performance of existing launch vehicle providers? I'd like to understand your thinking more fully. It seems axiomatic that new competitors entering existing markets sometimes manage to best the existing competitors by lower prices or superior products. Is there something about rockets that makes this impossible? No doubt it is difficult, but you seem convinced that it can't happen. Or is it particular shortcomings in SpaceX that convince you they can't do it?

[mikes]

Assuming the image is cropped (rather than scaled), an 800mm lens on a Canon 5D Mark II (as listed in the EXIF) subtends 1.653 arcseconds per pixel
http://www.howardedin.com/articles/fov.html

The Dragon in the image is about 16 pixels across, so 26 arcseconds for the 3.6m wide capsule.

S=O/H,  H=O/S, distance = 3.6m/sin(26/3600) = 29km

Intuitively that seems too far to me. Please could someone check my maths?

Two small errors. Rounding 16*1.653 = 26.448 to 26 and using sine instead of tan. I get 3.6m/tan(26.448/3600) = 28.1 km.

Also, from the press conference I thought I heard the spacecraft land within "a 100 metres" instead of "800 metres" of the target point.

The whole point is wrong.  Yes, image was scaled by the web server upon rendering.  May be 10 times, may be more.  You never know.  So all the math doesn't make any sense.

I downloaded the full image and worked from that, not the inline server-scaled version. It could also have been scaled, but that generally strips the EXIF data so I'm guessing it wasn't.

Steven: Thanks for the confirmation. Given the vagueness of pixel counting, I'm comfortable with the rounding! Precision is likely +/- 2km.

All pictures on the web are scaled, unless it's a separate download not rendered by the web server to your browser.  Resolution of the Canon 5D mark II is 5616x3744 pix.  Was this what you've been working with?  And scaling done by the server doesn't modify EXIF.  Also if picture was smaller then this, it was either scaled, or was shot using low resolution mode.  So all this arithmetic is just a guess.

What do you mean "all pictures on the web are scaled"? A normal web server (eg. Apache) just serves up the files it's holding, without modifying them. Some particular applications, such as the forum software on NSF, will provide a scaled thumbnail for inline use, but any scaling of the full image will be done manually by the person putting it online. Webservers do not generally "render" - that's a function of the browser.

Having said that, I take your point that the image must have been at least cropped without stripping EXIF, so could equally have been scaled. The image I worked from was 1024x683. If we assume it was scaled without cropping (which given the matching aspect ratio seems plausible) then that makes the distance camera to capsule 5.1km, which also seems intuitively more likely.

Can we be hopeful of either figure? Well, no: the 3:2 aspect ratio is a standard one, so it could have been both cropped and scaled. All we can reasonably say is that the distance is 5-30km. Which isn't particularly useful

On the SpaceX file server
https://send.spacex.com/bds/Login.do?id=A043517252&p1=naj20dpsbfegcidgdlgffcj20
there's another image
20101208_F9-002_V_Dragon_IMG_9043.jpg
which from the pixel noise looks to me to be unscaled (Hanol: you probably have more experience in judging that than I have - what do you reckon?)
The capsule is about 110 pixels across in that, which gives a distance of about 4.1km
The photo was taken 21 seconds before IMG_9070, so they should be comparable (10m/s = 210m higher, +/- lateral drift?)

[Dave G]

Or maybe the strongback is in the wrong place to survive repeated launches. 

All these add up to costs creep and soon they will no different than ULA.

You can't draw conclusions on normal system durability/longevity when there is a failure.

[edkyle99]

Or maybe the strongback is in the wrong place to survive repeated launches. 

I would expect SpaceX to be driven toward the Zenit solution, which uses a strongback that pulls back shortly before launch, with a separate umbilical mast that pulls back at T-0.

 - Ed Kyle

[go4mars]

“we’re confident we could get a fully operational vehicle to the pad for $2.5 billion—and not only that, I will personally guarantee it,” Musk says. In addition, the final product would be a fully accounted cost per flight of $300 million, he asserts. “I’ll also guarantee that,” he adds, though he cautions this does not include a potential upper-stage upgrade.[/i]

No.  This part is talking about the plans for the current heavy lift program.  "Based on a roughly evenly split $10 billion budget for heavy lift, with half for the boost stage and half for the upper stage,"

The rest is talking about SpaceX doing it for a lot less.  "we’re confident we could get a fully operational vehicle to the pad for $2.5 billion—and not only that, I will personally guarantee it,” Musk says. In addition, the final product would be a fully accounted cost per flight of $300 million, he asserts. “I’ll also guarantee that,” he adds, though he cautions this does not include a potential upper-stage upgrade.

There's some other recent article out there where he talks about taking 20% of the current budget for heavy lift (2.5 billion) and doing the entire system, while the other 80% would go toward the currently planned heavy lift.  That way NASA would have redundancy.

[go4mars]

I think he very well thinks he can build a 125 mT HLV from where he is right now with only $2.5 billion. That's his point.

He may or may not be able to do that, but that's what he claims.

150 mT.

[luke strawwalker]

This willingness to look at the data, take a calculated risk, make the decision, do the fix, and suck it up and light the candle is VERY VERY REFRESHING-- reminds me of the "old days"...  Hopefully some of that will "rub off" on NASA before NASA just completely becomes the "Amtrack of space", surpassed by everything around it...

One could argue that such is what got both Shuttle crews killed, and that that is why a couple of cracks in the stringers of an external tank are being looked at with such care.

Not really... more like "ignoring the recommendations of the engineers most familiar with the affected systems and ignoring previous warning signs on prior missions". 

In both the Challenger and Columbia tragedies there were plenty of warnings given, both from engineering staff and from damage to the hardware itself that indicated a problem, but didn't cause a LOC/LOV/LOM that it was conveniently "ignored" by management more concerned with other factors than they were with safety. 

What I was referring to here was a more refreshing approach than getting mired down in "analysis paralysis" and burdensome procedures and red tape which contribute NOTHING to the ultimate success or failure of the risk taken...  Make a good decision based on good information and you'll get a good result.  Make a bad decision based on bad or incomplete information, or worse yet IGNORE CONFLICTING "INCONVENIENT" INFORMATION, and you'll get a bad result, unless you happen to be extraordinarily lucky... and luck doesn't last forever... " 

Later!  OL JR  

[luke strawwalker]

Kudos to Elon Musk!  This is patriotism, pride, and passion in action!

Isn't Musk from South Africa? He moved to the USA when he was 17, according to wikipedia. So perhaps it's just passion on his part. He certainly does love space.

edit to correct - that was when he moved out of South Africa, but initially he went to Canada, not the US.

What difference does it make where he's from??  Werner Von Braun was from Germany originally, and actually worked during wartime for an enemy power...  yet he came to the U.S., worked for the US military, and became a US citizen, and contributed quite substantially to the US achieving it's goal of landing a man on the moon. 

I thought we were a "world community" nowdays, and that a person's "national origin" didn't matter anymore...

Course I guess it depends on who's ox is getting gored...

later!  OL JR

[luke strawwalker]

This willingness to look at the data, take a calculated risk, make the decision, do the fix, and suck it up and light the candle is VERY VERY REFRESHING

Have to disagree with this. If you increase your risk, eventually it will cause an unnecessary fail.

At the press conference, the NASA guy said that every time they came up with another question, SpaceX could show that they'd already considered it & it was OK. I took that to mean that NASA were happy that they could do this without increasing risk.

cheers, Martin

Martin is correct. Luke, what you're admiring was not taking a "calculated risk," it was having the independence and ingenuity to make a quick repair on the pad. But the reason they were able to make the decision and proceed quickly with the repair was that, as Elon said, they know their hardware extremely well. They knew the stresses in that area were low, and that they could cut off that portion of the nozzle without risk, but they re-did all the analysis anyway, to be extra careful and to satisfy NASA as well. If, in their mind, there had been anything "risky" about cutting away a portion of the nozzle, they would have taken a lower-risk option, like replacing the entire nozzle extension.

Well, maybe I didn't choose my words properly...

The basic point I was trying to make was, that they had the information, and made the decision to make the repair, even though it was a novel and 'untested' approach that created somewhat of a stir amongst the "professional rocket community" and JUST DID IT, without a long, drawn out, bureaucratic process of infinite delays and "ball-tossing" type hand wringing where nobody makes the decision so everybody makes the decision, so if something happens nobody's at fault because everybody's at fault...

That was my point... OL JR

[luke strawwalker]

What was truly incredible was Elon Musk saying that a future Dragon is being planned to do a full powered landing and land on a helipad similar to landing a helicopter. You would launch on the Falcon 9 achieve orbit and have a powered reentry to a pinpoint landing with parachutes as a backup. That's truly incredible and it totally negates the need for flyback plane type vehicles using runways. Sorry Sierra Nevada, you have just been outsmarted and outclassed. Why have runways when you can land on a pad? By the time Dreamchaser is developed it will have already been ruled unnessessary.   

Well, let's just say, "I'll believe it when I see it..." 

Remember years ago we were all promised jetpacks, flying cars, home nuclear power plants the size of a water heater you buried in the backyard to power your home, and all that...  Heck in the late 40's early 50's the helicopter was promised to eventually replace the automobile... 

First and second stage recovery, and especially this "propulsive landing on a helipad without parachutes" idea seems a pretty far off bridge to me...

For now, though, I'll settle for the fact that a privately-owned company working in a gov't sponsored program has achieved the first private orbital spacecraft safely returned to Earth... that's no small achievement in itself and is QUITE historic! 

Good luck to them on the rest, but I'll believe it when I see it...  

Later!  OL JR

[marsavian]

No, what he is saying is that he could build it for $12.5bn which is still very good. The $2.5bn only refers to his fixed-price rocket development/integration costs on top of the $10bn cost-plus for development of the constituent parts but it is disingenuous to just quote it by itself. It may also not include Raptor.

http://web02.aviationweek.com/aw/mstory.do?id=news/awst/2010/11/29/AW_11_29_2010_p28-271784.xml&channel=space&headline=NASA%20Studies%20Scaled-Up%20Falcon,%20Merlin

Based on a roughly evenly split $10 billion budget for heavy lift, with half for the boost stage and half for the upper stage, “we’re confident we could get a fully operational vehicle to the pad for $2.5 billion—and not only that, I will personally guarantee it,” Musk says. In addition, the final product would be a fully accounted cost per flight of $300 million, he asserts. “I’ll also guarantee that,” he adds, though he cautions this does not include a potential upper-stage upgrade.

No.  This part is talking about the plans for the current heavy lift program.  "Based on a roughly evenly split $10 billion budget for heavy lift, with half for the boost stage and half for the upper stage,"

The rest is talking about SpaceX doing it for a lot less.  "we’re confident we could get a fully operational vehicle to the pad for $2.5 billion—and not only that, I will personally guarantee it,” Musk says. In addition, the final product would be a fully accounted cost per flight of $300 million, he asserts. “I’ll also guarantee that,” he adds, though he cautions this does not include a potential upper-stage upgrade.

There's some other recent article out there where he talks about taking 20% of the current budget for heavy lift (2.5 billion) and doing the entire system, while the other 80% would go toward the currently planned heavy lift.  That way NASA would have redundancy.

No, it is what I said it was.

http://cosmiclog.msnbc.msn.com/_news/2010/12/06/5600599-spacex-gets-set-for-next-giant-leap

Musk said his $2.5 billion figure for a super-heavy-lift rocket was based in part on the concept that 80 percent of the money Congress is expected to devote to heavy-lift development would go toward the standard cost-plus method for funding spacecraft development, with 20 percent going to the kind of fixed-price, milestone-based approach that is being used for the NASA program that's funding SpaceX's effort. "I find myself in this bizarre position where people are saying, 'You couldn't possibly do it for such a low amount as $2.5 billion,'" he said. "And actually, I have trouble trying to figure out how we'd spend so much money. In order to get to $2.5 billion, I'd have to assume that a whole bunch of things go horribly wrong during the development process."