Engineering the SpaceX way

alfadriver said:

Funny how we discuss NASA and the locations of where they are- the ONLY place that they use that has real geographical logic in it is Canaveral-

Vandenberg (polar launches to the south) and Edwards (lake bed) are geographical too.

codrus (Forum Supporter) said:

alfadriver said:

Funny how we discuss NASA and the locations of where they are- the ONLY place that they use that has real geographical logic in it is Canaveral-

Vandenberg (polar launches to the south) and Edwards (lake bed) are geographical too.

True, but not having a massive body of water in the flight path for either makes Florida better.  Almost impossible to hit someone if there's a failure in the Atlantic.  Which happened a lot in the 50's and early 60s.  

codrus (Forum Supporter) said:

alfadriver said:

Funny how we discuss NASA and the locations of where they are- the ONLY place that they use that has real geographical logic in it is Canaveral-

Vandenberg (polar launches to the south) and Edwards (lake bed) are geographical too.

That's actually the really brilliant thing about Virgin's design.  Potentially better than Space-X, portable launch pad to avoid weather issues, and reusable everything

In reply to alfadriver :

South America is really far east.  If you want to launch to the south with a large body of water under you, Vandenberg or Edwards are where you'd want to do it from.  Florida, you'd crash into the Andes and then you'd have writers clamoring to write a disaster movie out of it.

Mr_Asa said:

codrus (Forum Supporter) said:

alfadriver said:

Funny how we discuss NASA and the locations of where they are- the ONLY place that they use that has real geographical logic in it is Canaveral-

Vandenberg (polar launches to the south) and Edwards (lake bed) are geographical too.

That's actually the really brilliant thing about Virgin's design.  Potentially better than Space-X, portable launch pad to avoid weather issues, and reusable everything

Look into plane launch. Basically, the amount of extra mass needed to survive the stress of aerial carry and launch is not worthwhile given the very small amount of delta-V it provides. It also doesn't really scale. It's been tried a bunch of times over the years and the math always works out the same. Although Virgin Orbit has actually managed to reach orbit twice, which puts it in pretty select company.

In reply to APEowner :

You're right that there's an emphasis on manufacturing - one of the things that SpaceX is trying to do is bring mass production to heavy rockets. Since this was an interview with SpaceX Musk and not Tesla Musk (I swear it's like they're different people), it's about rocket engine production. If one fails during the static fire test, you swap it out. If one fails on the way up, well, you just burn the remaining 28 a little bit longer. Failure modes in electric cars are different.

These rules were taken from a fairly long interview. There are examples given of what happens when you skip some of the steps.

Mr_Asa said:

That's actually the really brilliant thing about Virgin's design.  Potentially better than Space-X, portable launch pad to avoid weather issues, and reusable everything

Are you talking about SpaceShipOne (Virgin Galactic) or LauncherOne (Virgin Orbital)?

SpaceShipOne is reusable but suborbital only and there's no path to orbital capabilities with that vehicle, it would need about six times the delta vee to get to orbit and the re-entry heat shielding design can't scale to the levels required to return from orbit.  If an orbital launch vehicle were a car, SpaceShipOne is a golf cart.

LauncherOne is flexible, yes, but it has a max LEO payload of about a thousand pounds, roughly the same as Pegasus did (a broadly similar system).  Air launch limits how big your rocket can be, so you just can't put as much of a payload up there.  For comparison, a reusable Falcon 9 launch is 17,000 pounds to LEO.  The 747 launch aircraft is obviously reusable, but the LauncherOne rocket is not.  LauncherOne cost per pound to orbit seems to be about $12K, compared to $2500 for SpaceX.  They may have a niche for certain payloads, but there's no way in which it's "better than SpaceX" in general.

In reply to Keith Tanner :

Auto-landing rockets were also tried and failed over and over.  The envelope gets pushed more and more.

alfadriver said:

codrus (Forum Supporter) said:

alfadriver said:

Funny how we discuss NASA and the locations of where they are- the ONLY place that they use that has real geographical logic in it is Canaveral-

Vandenberg (polar launches to the south) and Edwards (lake bed) are geographical too.

True, but not having a massive body of water in the flight path for either makes Florida better.  Almost impossible to hit someone if there's a failure in the Atlantic.  Which happened a lot in the 50's and early 60s.  

Florida is good for low-to-medium inclination orbits due to the Atlantic to the east, that covers almost all manned spaceflight and most communications satellites.  Mapping or reconnaisance satellites are usually put into high inclination/polar orbits and as Pete says there is a lot of occupied land both north and south of Florida. Vandenberg has a straight shot to the south, if you look at a map you don't overfly any land until you get to Antarctica.

5. Automate.  This seems to be a manufacturing thing rather than a design thing...

 It can be; they'd call it "form factor" when I used to work with programs.    They'd design it with an eye towards needing fewer touch points or processes for a machine to perform. 

Mr_Asa said:

In reply to Keith Tanner :

Auto-landing rockets were also tried and failed over and over.  The envelope gets pushed more and more.

No, they were tried and succeeded, just not pursued further because the Shuttle was the official way to go or because there was no competitive advantage to doing so. And that's a whole other discussion! 

But it's not pushing the envelope that's limiting air launch, it's that nasty rocket equation. The "first stage" (the airplane) isn't a very good one as rocketry goes. Falcon 9 stages pretty early for a booster, and it's going mach 10 when it shuts down. You're not doing that with a 747, so now the airplane has to lift a rocket that still has to do almost the same amount of work as if it launched from the ground. It's not altitude you need for orbit, it's velocity.

SpaceShipOne (Virgin Galactic) is basically the X-15 with passengers. It's pretty much at the limits of its performance envelope and is frighteningly hands-on to fly. It's already killed one pilot, possibly due to pilot error. It's not really relevant to anything but a carnival ride.

codrus (Forum Supporter) said:

alfadriver said:

codrus (Forum Supporter) said:

alfadriver said:

Funny how we discuss NASA and the locations of where they are- the ONLY place that they use that has real geographical logic in it is Canaveral-

Vandenberg (polar launches to the south) and Edwards (lake bed) are geographical too.

True, but not having a massive body of water in the flight path for either makes Florida better.  Almost impossible to hit someone if there's a failure in the Atlantic.  Which happened a lot in the 50's and early 60s.  

Florida is good for low-to-medium inclination orbits due to the Atlantic to the east, that covers almost all manned spaceflight and most communications satellites.  Mapping or reconnaisance satellites are usually put into high inclination/polar orbits and as Pete says there is a lot of occupied land both north and south of Florida. Vandenberg has a straight shot to the south, if you look at a map you don't overfly any land until you get to Antarctica.

Right, but in the context of when NACA/NASA was building the space agency, going south wasn't really a plan- just getting stuff to space was.  Pretty much the rest of NASA's 50-60's era buildings are distributed around the country with not a lot of consideration of where- just who had clout at the time.  Except where to launch.  Since then, we have multiple places to launch as well as land, depending on the mission.

Vandenburg was not originally built for launching missiles (army base). It was used for launching (testing) missiles (the kind that went up and down and went boom) because it was an easy shot from there over lots and lots of water, it later of course had the advantage of being a straight shot to the south Pacific test ranges.  Polar orbit launces where of course a late add, but not it's original purpose (missile wise)

I am sure it's relatively close location to the (at the time) massive aerospace industry in Southern California was also a huge plus.

aircooled said:

Vandenburg was not originally built for launching missiles (army base). It was used for launching (testing) missiles (the kind that went up and down and went boom) because it was an easy shot from there over lots and lots of water, it later of course had the advantage of being a straight shot to the south Pacific test ranges.  Polar orbit launces where of course a late add, but not it's original purpose (missile wise)

https://en.wikipedia.org/wiki/Vandenberg_Space_Force_Base

Ballistic missile testing started in late 1958, first polar satellite launch was 1959 (a military recon satellite).  NASA got involved in the late 60s when planning for LandSat.

For those who find this interesting, the checklist came from this video. Elon expands on every step  and gives examples of how they haven't always been followed. It's a long interview but it's pretty interesting as he does know his stuff. The engineering process talk starts at about 13:30 and pops up again a few times.

In reply to Keith Tanner :

Thanks for posting the source for some context.  It's important to note that the list of 5 engineering principles that started this thread was not from some polished essay or white paper or business advice book, it was an off-hand remark by a tired man giving an outdoor walking tour of a factory in the Tex/Mex heat to describe how SpaceX is currently iterating its next generation of rockets.  Musk's approach seems to be working for SpaceX and Tesla, but YMMV.

PS -- I found this 3-hour, 3-part video walking tour of Starbase to be interesting and insightful.  It's rare to hear this sort of candid, in-depth conversation from any big corporate CEO.

It's also the result of some hard lessons starting a major car manufacturer from scratch and the world's only heavy lift rocket manufacturer from scratch. 

That first requirement certainly gets some more elaboration during the interview.

BTW, that video is part 1 of 3. Only the first 2 have been posted at this point.

Thanks for linking to the video, Keith.  I was going to ask for the source.  Very cool to see him in a natural environment like that.  I love the way he works through whether he even needs separate engines on the ship part during the conversation, and you can see him redesigning parts of it in his head :) 

I also enjoy watching the interplay between Tim Dodd and Elon. They are both highly intelligent but Tim is more methodical. He makes for a good throttle to Elon's stream of consciousness. 

Having participated in MANY action work-outs at various manufacturing facilities this all just looks like a simplified version of LEAN to me. Other than the way it was said, nothing about any of this is new in engineering.

infernosg said:

Having participated in Other than the way it was said, nothing about any of this is new in engineering.

A lot of times the way something is said is important. Sometimes we suck at interpreting things, and it's just a matter of hearing something put differently to grok it better.

And sometimes it's good to hear what the process at a specific company is, especially if it's managed to completely turn an industry upside down. There are other ways of engineering. I'm pretty sure this is not the Boeing way of working, for example.

I think the personal accountability for a design choice is a good one, as is the concept of every designer having a global view of the product. As noted in an earlier interview, there's no point in optimizing up to an interface if you don't know what's on the other side.

BTW, the interview was just part 1 of 3. All three parts have now been released.

There are a lot of gems in that interview / discussion.  If everyone is already doing everything like this, then why is SpaceX so far out in front?  

My experience, which is in aerospace but not on the space side, is that everyone is most definitely NOT doing anything like this.