NASASpaceFlight.com Forum
SpaceX Vehicles and Missions => SpaceX Reusability => Topic started by: ThePonjaX on 03/30/2017 12:28 pm
-
In the reddit the user DSBromeister give some very interesting details about the process:
[–]DSBromeister *
I've been waiting so long for this! I interned at LC-39A while the refurb was going on and boy did B1021 give us trouble! I'm so happy to finally see my baby fly!
Edit: since people are asking for more info, I'll give a couple fun problems we ran into.
Trying to upgrade parts from block 2 to block 3, failing to install them three times, then giving up and trying (and succeeding with) a method from block 1
Trying to remove parts that weren't originally intended to be removable
Discovering parts on the booster that theoretically didn't exist before it launched
https://www.reddit.com/r/spacex/comments/62aqi7/rspacex_ses10_official_launch_discussion_updates/dfl9xge/ (https://www.reddit.com/r/spacex/comments/62aqi7/rspacex_ses10_official_launch_discussion_updates/dfl9xge/)
and
[–]FredFS456
Parts that theoretically didn't exist? What do you mean?
permalinksaveparentreportgive goldreply
[–]DSBromeister
This part was (is?) made of a stock material on assembly rather than fabricated, but wasn't officially given a part number until after the launch of CRS-8. It must have been created during B1021's original assembly and installed, but with no way to officially record its installation since a part number didn't exist. Fast forward to refurb and it calls for the removal of a part that was never officially installed, so I had to dig up some other paperwork detailing what occurred.
https://www.reddit.com/r/spacex/comments/62aqi7/rspacex_ses10_official_launch_discussion_updates/dflcu5g/
This give us a good idea the some of the changes happening on Block-5
-
Thanks for posting this, although not sure SpaceX will appreciate this being made public on Reddit!
Assuming the posts are accurate, interesting that SpaceX's CM at the time allowed the booster to fly with a part that couldn't be officially recorded ...
-
And some idea on "hey, just refuel and refly"
-
Thanks for posting this, although not sure SpaceX will appreciate this being made public on Reddit!
That's why I copied the text because I'm afraid it's going be removed in no time 8)
-
Trying to upgrade parts from block 2 to block 3, failing to install them three times, then giving up and trying (and succeeding with) a method from block 1
Why was a v1.2 vehicle flying with parts from Block 2? Unless our understanding that v1.1 = Block 2 is incorrect, that's very odd.
-
Trying to upgrade parts from block 2 to block 3, failing to install them three times, then giving up and trying (and succeeding with) a method from block 1
Why was a v1.2 vehicle flying with parts from Block 2? Unless our understanding that v1.1 = Block 2 is incorrect, that's very odd.
It has been suggested before that SpaceX is continuously iterating such that every vehicle is a little different. At some point, there is a threshold (or some other large major change) that constitutes the creation of a new "block".
-
It has been suggested before that SpaceX is continuously iterating such that every vehicle is a little different. At some point, there is a threshold (or some other large major change) that constitutes the creation of a new "block".
Yes appears to be Elon's MO. Tesla works the same. When a change is ready it goes on cars from that point onwards; no such thing as this year's model.
-
It has been suggested before that SpaceX is continuously iterating such that every vehicle is a little different. At some point, there is a threshold (or some other large major change) that constitutes the creation of a new "block".
Well SX have certainly used most launches for testing various new operational ideas. If you have a vision of where you want your vehicle to be and the current version is not there why would you not keep changing it till it got the performance you wanted (provided you're OK with the risk of a payload losing mishap if you get things seriously wrong)? After all if they are not coming back....
Obviously that fairly freewheeling approach has to be slow down a bit once you start getting stages back and you want to relaunch them.
Note that with modern ERP systems continuing to make every LV a bit different is possible (certainly in the US) but that also means you have to maintain mfg capability for each seperate BOM for each separate 1st stage you wish to maintain.
And if there's anything we've learned about SX it's they really don't like to run multiple parallel supply chains of anything.
On this basis I suspect the operating lives of this group of first stages is going to be quite brief and the bar for how difficult something has to be before it's a case "No, too difficult. Send the stage to Testing for tear down and destructive tests" is going to be fairly low.
The first stage being built from now on are likely to be longer lived, but honestly, does anyone know what "long lived" means in this context?
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
-
Trying to upgrade parts from block 2 to block 3, failing to install them three times, then giving up and trying (and succeeding with) a method from block 1
Why was a v1.2 vehicle flying with parts from Block 2? Unless our understanding that v1.1 = Block 2 is incorrect, that's very odd.
It has been suggested before that SpaceX is continuously iterating such that every vehicle is a little different. At some point, there is a threshold (or some other large major change) that constitutes the creation of a new "block".
I believe it was Jim who stated, quite some time ago, that all launch vehicles have iterative changes. Which does make sense, if you encounter a minor problem, you fix it.
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
I wonder if this was part of the systems engineering concerns NASA raised with SpaceX following the CRS-7 failure?
Of course the booster was built over a year ago now, so things may have changed.
-
Trying to upgrade parts from block 2 to block 3, failing to install them three times, then giving up and trying (and succeeding with) a method from block 1
Why was a v1.2 vehicle flying with parts from Block 2? Unless our understanding that v1.1 = Block 2 is incorrect, that's very odd.
It has been suggested before that SpaceX is continuously iterating such that every vehicle is a little different. At some point, there is a threshold (or some other large major change) that constitutes the creation of a new "block".
I believe it was Jim who stated, quite some time ago, that all launch vehicles have iterative changes. Which does make sense, if you encounter a minor problem, you fix it.
You see the same thing with aircraft.
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
Is this "haphazard"---or is a necessary consequence of trying to document every piece of electrical tape and every sheet of cork that sometimes minor things fall through the cracks. The OP never said they weren't able to determine how the part got there, just that they needed to consult backup documentation. Which is what that backup documentation (and close-out pictures, etc) is for.
The difficulty mostly seemed to be trying to refit a block 2 core as block 3 when that had never been done before, and wasn't designed to be do-able.
Sure, bugs were discovered, but the overall process looks very documentation-full to me. I'm not sure it deserves to be tarred with "haphazard".
EDIT: the OP says, "this wasn't really a mistake by anyone in particular, and it didn'treally affect anyone except the part inventory people, so maybe I dramatized the issue a little too much haha."
-
Reading both of his posts, it seems he misspoke on his ideas of blocks. He withdraws it later.
-
What appears "haphazard" to a low level worker may be appear much more rational to the higher ups. Also, Elon wasn't just joking around when he said SpaceX was like the dog that caught the bus. Yes they were planning all along to reuse stages, but I think they started getting them back in useable shape faster than they thought they would
Matthew
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
On the other hand, what other vendor has had a different crew try to modify a fully completed booster based on the supposed configuration? I'd not be surprised to find a lot of configuration errors at the bottom of the ocean. Configurations are a lot like code - everyone swears theirs is completely under control, but when you actually check, you often find it's not so perfect.
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
If only NASA and the USAF were deeply involved with SpaceX as part of the certification processes being run by both instituations...
-
I believe it was Jim who stated, quite some time ago, that all launch vehicles have iterative changes. Which does make sense, if you encounter a minor problem, you fix it.
As Leonard McCoy was famous for saying on Star Trek:
"I know engineers, they LOVE to change things."
And there is a saying we using manufacturing when introducing a new product:
"Shoot the engineers and ship the product!"
Constant iteration is not necessarily a bad thing, and in fact it can be a very good thing. For instance:
- For low rate production it sometimes takes a long period of time to dial in your production processes, since not everything engineering designs is easily manufacturable.
- Especially in the electronics world component availability and supplier changes to components is a constant challenge, so engineering has to keep up with the world as it is, not as they would like it.
- Over time engineering may discover better ways of designing parts and assemblies, and it makes sense to incorporate those changes as soon as practical.
The approach that SpaceX is using tends to be on the active side of the scale with regards to configuration stability, and it's understandable that some may not be comfortable with that. It's a philosophy that is always trying to balance product iteration speed with product success - all while running a business that requires success.
Reusable stages are going to help them to iron out issues that would have taken time and money with expendable stages, so we'll see if it works out in the end. Just as an outside observer, I think they are doing a good job so far...
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
If only NASA and the USAF were deeply involved with SpaceX as part of the certification processes being run by both instituations...
As noted above, this probably explains some of NASA's concerns following the CRS-7 situation.
Disagree all you like, but configuration control issues have led to more than one fatal aviation incident (both civil and military). And these are well-established fields with incredibly meticulous certification procedures. Those procedures exist for a reason: because when people don't document what they did and things go bad, you can never be certain you actually figured out why they went bad - unless, as had happened more than once - you discover parts in the debris that weren't supposed to be there, or evidence of prior repairs or construction that weren't supposed to be have been done that way.
So, yeah, if what this intern said initially was accurate (*), it's a mark of poor configuration control.(**) And let's be honest - configuration control is HARD. It's arguably among the hardest parts of the any serial production business, even when the product itself isn't changed from unit to unit. Suppliers change specifications, processes by *their* suppliers get changed without notice, et cetera. We never heard the final root cause for why those struts all failed in the post-CRS 7 batch testing, but again, configuration control failed.
Worse than being hard, it's expensive and it's not sexy. Which, I guess to some in this crowd, seems to be the most important element.
(*) The kid may have been overstating for the purposes of making an "exciting" post on Reddit; yet another reason I stay away from that cesspool of misinformation and horrible user interface.
(**) If you read the L2 ISS Daily Status Reports, configuration control issues is not a SpaceX-specific problem. There have been a number of recent minor issues that have been found to include pieces installed incorrectly, connectors reversed, and so on, despite all the best efforts by NASA and its contractors to maintain rigorous control.
-
The irony is that configuration control has been a strength of that organization.
It has to be if you practice agile development in aerospace, otherwise the situation is not diagnosable in the slightest. As it is already a nightmare with traditional aerospace, you're pushing it.
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
That's an understandable PoV but now that the first stages are starting to be reused I think the rate of change will have to be lowered considerably to limit the number of supply chain variants their ERP system has to track.
Making and tracking N different variants of a part because N different stages were all built slightly differently is a real PITA, in cost, mfg scheduling, machine resources etc. I expect the first stage design will continue to evolve but in more block orientated fashion.
I also think the early recovered stages will have quite short re-use lives to purge the supply chain of too many variants. My instinct is less than 10 flights each at most, but that's a guess.
I think it's very interesting they are talking about "fairing reuse" but not 2nd stage recovery. [EDIT this implies 2nd stage changes can continue at whatever rate they are currently running at and continue to provide hard data for future plans. However it's hard to say how useful this can be without an end goal of fully reusable upper stage, which implies a whole new F9 sized vehicle on LOX/Methane. ]
-
I don't quite get why everyone in the naysayer camp seems to harp on second stage recovery. They said it's not happening and they're moving on.
Maybe they'll do a Raptor upper stage, maybe they won't...
-
Making and tracking N different variants of a part because N different stages were all built slightly differently is a real PITA, in cost, mfg scheduling, machine resources etc. I expect the first stage design will continue to evolve but in more block orientated fashion.
I also think the early recovered stages will have quite short re-use lives to purge the supply chain of too many variants. My instinct is less than 10 flights each at most, but that's a guess.
I'd guess even less, maybe at most one reflight for most of them, and none for the oldest ones. When they need to pick a core to refurbish, they will pick the one that's in the best shape and needs the least work. After that the next best, and so on. The newly built cores will naturally be at the top of this list, since they have the most things already fixed, and are closest to the current configuration. The oldest cores will never be used, as enough newer cores are available.
-
I'm not shocked by any of this, they are in block transition (multiple blocks) to block 5.
A view this flight as an initial proof of concept rather than a view of how every refurbishment will be in the future.
-
I'm not shocked by any of this, they are in block transition (multiple blocks) to block 5.
A view this flight as an initial proof of concept rather than a view of how every refurbishment will be in the future.
Given this is the first time a first stage has been reflown ever this must be viewed as an extreme outlier.
According to the story 1021 first flew April 8th 2016, close to a year ago.
I don't think anyone believes it has taken SX close to whole year to refurbish it. I suspect most of that time will have been spent working out and documenting the refurbishment procedure (including deciding which processes are unique to this stage, and which applicable to it and all future ones).
How much the actual elapsed time (and staff) needed to do a full refurb will remain anyone's guess.
I don't quite get why everyone in the naysayer camp seems to harp on second stage recovery. They said it's not happening and they're moving on.
Maybe they'll do a Raptor upper stage, maybe they won't...
Only die hard amazing people have been talking about second stage recovery.
The rest of us took Musk at his word and moved on long ago.
-
Only die hard amazing people have been talking about second stage recovery.
The rest of us took Musk at his word and moved on long ago.
Elon just mentioned possibly trying for second stage recovery in his press conference...
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
If only NASA and the USAF were deeply involved with SpaceX as part of the certification processes being run by both instituations...
As noted above, this probably explains some of NASA's concerns following the CRS-7 situation.
Disagree all you like, but configuration control issues have led to more than one fatal aviation incident (both civil and military). And these are well-established fields with incredibly meticulous certification procedures. Those procedures exist for a reason: because when people don't document what they did and things go bad, you can never be certain you actually figured out why they went bad - unless, as had happened more than once - you discover parts in the debris that weren't supposed to be there, or evidence of prior repairs or construction that weren't supposed to be have been done that way.
However, adding more and more checks comes with a cost, which is snowballing ever-so-slightly. And if you do not pay attention to this snow-balling, you may inadvertently end up with a situation when replacement of a single bolt requires one person to do the work and ten people documenting it, cross-verifying it, then documenting and cross-verifying the process of documenting.
Not only it makes the process much slower and more expensive, it also makes people working on the hardware averse to the changes (since they need to do tons and tons of paperwork for each).
And then you find your rocket engine designer, totally frakked off at the fact his "rocket design" work entails only doing endless paperwork, quit you company to work for some ridiculous startup.
And then this startup's rocket whooshes past you, while you stare at the glowing letters on the wall: "Prepare to Become Obsolete".
-
That kind of haphazard configuration control does not bode well for assuaging worries of government customers. :(
If only NASA and the USAF were deeply involved with SpaceX as part of the certification processes being run by both instituations...
As noted above, this probably explains some of NASA's concerns following the CRS-7 situation.
Disagree all you like, but configuration control issues have led to more than one fatal aviation incident (both civil and military). And these are well-established fields with incredibly meticulous certification procedures. Those procedures exist for a reason: because when people don't document what they did and things go bad, you can never be certain you actually figured out why they went bad - unless, as had happened more than once - you discover parts in the debris that weren't supposed to be there, or evidence of prior repairs or construction that weren't supposed to be have been done that way.
However, adding more and more checks comes with a cost, which is snowballing ever-so-slightly. And if you do not pay attention to this snow-balling, you may inadvertently end up with a situation when replacement of a single bolt requires one person to do the work and ten people documenting it, cross-verifying it, then documenting and cross-verifying the process of documenting.
Not only it makes the process much slower and more expensive, it also makes people working on the hardware averse to the changes (since they need to do tons and tons of paperwork for each).
And then you find your rocket engine designer, totally frakked off at the fact his "rocket design" work entails only doing endless paperwork, quit you company to work for some ridiculous startup.
And then this startup's rocket whooshes past you, while you stare at the glowing letters on the wall: "Prepare to Become Obsolete".
Sure. But none of that invalidates anything I actually wrote. That's entirely not the point.
-
I'm not contradicting you. I'm pointing out there is a drawback in having way too much configuration control.
-
I'm not contradicting you. I'm pointing out there is a drawback in having way too much configuration control.
No, actually you're not. It's quite possible to have excellent configuration control and do it efficiently. The issue I have (*) here is that it wasn't done CONSISTENTLY. Consistency, repeatability, and frequent repetition is the key to efficiency.
(*) Assuming arguendo the kid's story is, in fact, accurate. I'm not entire sure he wasn't exaggerating for effect.
-
I'm not contradicting you. I'm pointing out there is a drawback in having way too much configuration control.
No, actually you're not. It's quite possible to have excellent configuration control and do it efficiently. The issue I have (*) here is that it wasn't done CONSISTENTLY. Consistency, repeatability, and frequent repetition is the key to efficiency.
What evidence is there for this interpretation?
(*) Assuming arguendo the kid's story is, in fact, accurate. I'm not entire sure he wasn't exaggerating for effect.
So... based of on an iffy source, you make the worst assumption?
-
Only die hard amazing people have been talking about second stage recovery.
The rest of us took Musk at his word and moved on long ago.
Elon just mentioned possibly trying for second stage recovery in his press conference...
I've just skimmed the SES10 webcast. Musk has a short segment. [EDIT re-checked it. He talks about booster reuse and how not throwing the whole thing away every launch saves money. There is no mention of upper stage reuse. It runs about 3 mins round 40+ mins into the cast ]
[EDIT in the interests of completeness Shotwell said the last planned spin of the F9 design is to give 10 reuses, suggesting all the 1st stages till then (sometime later this year) will have less than 10 flight design lives. TBH 10 flights seems quite low, and a long way from "aircraft like." I would have guessed a generation after that with a longer life, but in truth only SX know (because they've had real hardware to take apart) what cumulative damage is done during the whole launch/sep/land/refurb cycle ]
Has he given another longer press briefing?
-
What evidence is there for this interpretation?
The initial quoted Reddit post by the alleged intern that indicated parts were found during refurbishment that apparently didn't even exist at the time of assembly; that means configuration control was not applied consistently during the flow from assembly to flight prep to launch, because - by definition of the term - the use of said part would not have come as a surprise during refurbishment of the stage.
So... based of on an iffy source, you make the worst assumption?
Go back and read my initial quote - in context - without the defensiveness for all things SpaceX. It's quite possible to be a vehement fan and still worry about flaws in the processes used to achieve their goals. If what the kid said is accurate, their processes were flawed. People are flawed; the processes are supposed to work around those human flaws.
-
I've just skimmed the SES10 webcast. Musk has a short segment. [EDIT re-checked it. He talks about booster reuse and how not throwing the whole thing away every launch saves money. There is no mention of upper stage reuse. It runs about 3 mins round 40+ mins into the cast ]
I'm just getting out of work and haven't been able to watch the Q&A session. According to Jeff Foust, Elon said might be "fun to try a Hail Mary (https://twitter.com/jeff_foust/status/847599210513944577)" and recover the second stage. Chris Gebhardt relayed the same message here on NSF:
Musk: Upper stage reuse is next.
EDIT:
Found it. Around 14:20 in this video (https://youtu.be/jC3LQFpuzqs?t=860) of the press conference Elon mentions potentially trying for reuse of the second stage.
-
https://youtu.be/jC3LQFpuzqs
But then the only thing left is the upper stage, which we didn't originally intend for Falcon 9 to have a reusable upper stage, but it might be fun to try like a hail mary, and you know. What's the worst thing that can happen? It blows up. You know, it blows up anyway. [Martin Halliway chimes in humorously. "We need to discuss this."]
(Fairing recovery talk started at 12:05 when a guy walked in and showed Elon a photo of the floating fairing, but that's another thread.)
-
Key info from Elon at the press conference on booster 21 refurbishment for the SES-10 flight:
With this being the first reflight we were incredibly paranoid about everything. So we sort of, the core airframe remained the same, the engines remained the same but any auxiliary components we thought might be slightly questionable we changed out.
Now our aspiration would be zero hardware changes, reflight in 24 hours the only thing that changes would be we reload propellant.
-
I'm just getting out of work and haven't been able to watch the Q&A session. According to Jeff Foust, Elon said might be "fun to try a Hail Mary (https://twitter.com/jeff_foust/status/847599210513944577)" and recover the second stage. Chris Gebhardt relayed the same message here on NSF:
Musk: Upper stage reuse is next.
EDIT:
Found it. Around 14:20 in this video (https://www.youtube.com/watch?v=jC3LQFpuzqs?t=860) of the press conference Elon mentions potentially trying for reuse of the second stage.
I heard that too. Yes Musk is remains interested in full reuse. If you've just spent 15 years and about $1Bn on reusability you're not going to let go easily.
I'm guessing if some of his team come up with a (not too expensive) plan to give F9 US reuse a go with at least a 1 in 10 chance of success he'd give it a go, if they can find a customer willing to take the risk.
Otherwise F9 US reuse is dead as a dodo. :(
Having started to watch the actual press conference, rather than the launch coverage Musk mentions they are looking at a life of 10 first stage flights with no refurb but 100+ with work being done.
That suggests 10 flights is how long the TPS is good for without any repair or replacement. Beyond that it's certain it will burn through somewhere.
What I'd really like to know is what (if any) is the cumulative fatigue to the main structure caused by the hoverslam landing.
I keep thinking of the old saying that boxers and football players go in the knees first. :(
-
https://twitter.com/elonmusk/status/847882289581359104
Considering trying to bring upper stage back on Falcon Heavy demo flight for full reusability. Odds of success low, but maybe worth a shot.
-
I'm guessing if some of his team come up with a (not too expensive) plan to give F9 US reuse a go with at least a 1 in 10 chance of success he'd give it a go,
Otherwise F9 US reuse is dead as a dodo. :(
Reuse is not the only reason to try recovering a second stage given the Mars goal. Just learning how to get one back down in a small integer number of pieces, compared to burning up, would be valuable information
-
Working my way through the press conference video (had some sound problems)
Around 32mins in says they are switching to forged Titanium alloy for the grid fins from TPS coated Aluminum and they are going to be the worlds biggest Ti forgings.
This is an unusual choice given that there are very few really big presses EG 50 000 tons. They are used for things like the rotor hubs of really big helicopters, or nuclear reactor pressure vessels. They're very specialized and it's an odd choice given getting a mfg schedule on them is likely to be quite tricky.
It's even more so given that Ti alloys are quite attractive for diffusion bonding (the oxide layer is soluble in Titanium), so making up a grid structure in a vacuum furnace (even a big one) would be quite viable.
[EDIT They payoff seems to be that with these new fins will give the F9 a L/D of 1, presumably in the hypersonic regime. Given that this is a bottom heavy (very) blunt body that's potentially pretty high]
-
Cross-posting, suggest follow-ups go in other cost thread:
Shotwell: cost of refurbishing F9 first stage was “substantially less” than half of a new stage; will be even less in the future. #33SS
https://twitter.com/jeff_foust/status/849679544923697152 (https://twitter.com/jeff_foust/status/849679544923697152)
Edit to add: important extra point
Shotwell: Falcon booster refurbishment cost substantially less than 1/2 cost of new build; more done for SES-10 than future flights. #33SS
https://twitter.com/flatoday_jdean/status/849679956988674048 (https://twitter.com/flatoday_jdean/status/849679956988674048)
-
So many threads this could be posted in ...
Here's a video montage of the life of booster 1021 as presented by Gwynne Shotwell at the space symposium on Wednesday:
https://youtu.be/mQNI4bipS0g (https://youtu.be/mQNI4bipS0g)
-
So many threads this could be posted in ...
Here's a video montage of the life of booster 1021 as presented by Gwynne Shotwell at the space symposium on Wednesday:
https://www.youtube.com/watch?v=mQNI4bipS0g (https://www.youtube.com/watch?v=mQNI4bipS0g)
I might be mistaken, but is 0:41 the first time a closeup of the (hard) Thaicom-8 booster landing, from the barge's deck, was made public?
-
Karim Michel Sabbagh and Gywnne Shotwell (#SpaceX) present @EtienneSchneide with a piece of the flight proven rocket that flew SES-10 into GTO.
https://twitter.com/ses_satellites/status/931083202705461248