The Great Ongoing Space Exploration Thread

Miscellaneous and Personal Stuff I Must Share
422

Musk doesn’t put a whole lot of importance on degrees. SpaceX doesn’t hire idiots, and people who aren’t idiots are pretty likely to have degrees, but that’s correlation, not causation. Tom Mueller designed the Merlin engine, its variants, and a few other engines along the way. He has a BS and Masters in mechanical engineering, but that’s not why Musk hired him. Mueller was hired because he was building turbopump rocket engines in his garage and firing them out in the desert. His day job was designing novel engines for TRW that never flew, but lots of people in aerospace have that kind of position. To go home and build your own engines because that’s the only way you’ll see your own work fly takes a certain obsessive quality that Musk was looking for.

As an aside, I think “rocket scientist” is a bad name for the position Musk holds. “Rocket engineer” is only slightly better. I would say “rocket architect” is good, however. A normal architect designs a building at a high level, with the low-level details to be worked out by others. And yet, the architect must have a solid grasp of what is possible, given the materials, current state of the art in manufacturing and engineering, etc. The architect must also make economic decisions, and be aware of what cost tradeoffs are possible. A rocket architect is in a similar position, except that the margins for rockets are so much narrower that the architect must have an even keener grasp of the important details, even if they are still not designing any particular part.

Sergei Korolev and Werner von Braun would be historical examples of rocket architects. Musk is probably the equal of either of them at this point. He does have the advantage of having complete control over SpaceX, whereas others had to contend with political infighting.

Everyone who has worked with Musk, even people who have had a falling out or left SpaceX on less than the best of terms, has said that he is able to grasp new subjects very quickly. Robert Zubrin, for instance, said that when he first met Musk, he had essentially zero knowledge of rockets, and it showed. A short time later, they met again and Musk knew everything there was to know on the subject. That doesn’t mean he was designing rockets down to the nuts and bolts, but that ability to absorb vast detail did make him an excellent rocket architect.

The Falcon 9 is in many ways nothing special, but they made good decisions on just about every front. Choice of propellants, ease of manufacture, materials, engine commonality, horizontal integration, even things like diameter (chosen to be the maximum transportable by highway) all made for a very well-balanced rocket. There was no outside influence forcing them to use solid boosters or Russian engines or to put their factory in Mississippi. Nor did they shoot beyond what they could reasonably achieve (no spaceplane, no single-stage-to-orbit, etc.). Musk and his crew just made a nice clean-sheet rocket that didn’t do anything stupidly, and amenable to a great deal of constant refinement.

423

In many ways this is related to the “killing your friends dreams” thread. If someone comes up to you and says, “I’ve got an idea, we should make rockets reusable! You’re a rocket scientist, figure it out.” Then you show them the door. If they follow that up with, “and I’ve got billions of dollars to make it happen.” then you have a new job.

424

What happens next, however, depends on the quality and feasibility of the idea. Lots of rocket scientists had someone walk into their office and offer them millions of dollars to build an SSTO (single stage to orbit) craft, yet none have successfully done so.

425

Fair enough, but my point was more that, in that other thread, people were being offered ideas, to be done on their own time and dime, and the idea person expecting to get a cut for coming up with the idea.

It’s a whole different thing if the idea guy is also bringing the financing to the table. Maybe it’ll never work, but at least you are getting paid for it.

426

Gotcha! Yes, I agree, being the idea guy is not enough. You need to bring something else to the table too, be that expertise or money.

427

That’s not nearly enough. Jeff Bezos did exactly that, and while not a failure, Blue Origin is getting pretty much the kind of results you’d expect from an old space company.

True innovation comes from taking risks, and from allowing the people in the trenches the freedom to make choices. Organization structure matters a lot. A culture of risk-taking, rapid prototyping and incremental development matters a lot. Allowing the peoole closest to the problem to offer solutiins is critical.

In the Russia/Ukraine war thread, everyone seems to acknowledge that a key problem for the Russian military is a lack of an NCO corps. In western militaries, NCOs are the backbone of thr army. The generals make the strategic plans, but the NCOs are given wide latitude in how to achieve their unit’s objectives. They are the ones on the ground where the action is, and they are therefore the most qualified to figure out how to respond. So American military units are given high-level objectives and then given the trust of the leadership to be flexible in how to achieve it. Everyone down to the soldiers in the field understand the tactical objectives and the strategic plan so they can make smart decisions.

In the Russian military, decisions are made at the top, and then at the low level people are given specific instructions in exactly how and where to fight - by people with the least amount of information about what’s actually going on. Information other than those specific instructiins are siloed, so the soldiers don’t even know why they are doing what they are doing. So you get tank columns sitting on a road for weeks when the original plan goes south, or mass confusion when an unexpected offensive happens.

As for their big plans and strategies, as Prussian General Helmuth von Moltke famously said, No plan of operations extends with any certainty beyond the first encounter with the main enemy forces. (often stated as, “No battle plan survives first contact with the enemy”).

So when your great plans go wrong, what do you do? You rely on the people at the bottom to have agency, and allow them to improvise. Russia couldn’t do that, so it had to send its generals to the front line to figure out what was going on, and Ukraine started picking them off. So the Generals keep their distance, and the Russian military fails.

The same lessons apply to firms, governments, and society. Top-dow; planning and rigid hierarchies and silos of information and control lead to bad decisions, locked-in thinking, turf wars, and incredibly stupid decisions. Trying to ‘innovate’ by getting the big brains into a committee and having them think up ideas is a poor substitute for setting free the minds of thousands of employees who are actually working on those problems. But you have to give them the incentives to push back on bad choices and to go down different paths.

To enable this in a commercial world, you need freedom to fail. Musk, coming from a software background, understood the power of incremental development, minimum viable products, fast turnaround and the ability to survive risky choices that fail. Agile development in software is all about this: we learned a long time ago that risky projects fail when designed from the top down in giant monolithic designs, as NASA does with its rockets.

Musk brought agile development to giant rockets, gave his engineers freedom to take big swings, doesn’t punish failed choices so long as they were justified, etc. That’s why SpaceX is moving to fast and innovating so much.

Absolutely. Musk doesn’t have to be a great engineer - but he does need to understand enough engineering to be able to evaluate alternative ideas and spot areas where old thinking needs to go.

If SpaceX succumbed to modern managerialism and had hired a superstar soft drink CEO to run the company, it would have been a disaster. Too many companies think that a manager is a manager, and it doesn’t matter what they manage. Musk among many others have shown that to be nonsense.

428

The other side of this is that you have to be able to tolerate total failure. It’s entirely possible that it wouldn’t have worked out, and Musk would have been out a couple few bucks in the process. My understanding is that they were one more failed launch away from pulling the plug.

Bezos treated private space enterprise as a second mouse scenario, rather than an early bird. He expected Musk to spend a whole bunch of money and go bankrupt, giving Blue Origin a good pathway of at least avenues not to follow, and possibly even giving them the ability to buy up SpaceX’s IP and hire its staff for cheap.

I certainly wouldn’t compare Blue Origin to Russian military, that’s just a ridiculous comparison. Bezos wasn’t some rich guy that wanted a space program, Bezos became wealthy because of the work and risks he put in in creating Amazon.

I’ll agree on the NASA side of things, there is quite a bit of stagnation due to the culture, but at the same time, they are doing the things that no corporation will do, as there is little ROI to be had in most of their missions.

He also got damn lucky.

429

Pretty much. In fact, they were already almost past that point. But they realized they had just enough spare components for a fourth try. They were pretty much out of money.

Well, some bad luck too. They were hoping to launch out of Vandenberg, but were constantly delayed by conflicts with other vehicles. No real surprise that they’d be the one to get bumped, but still frustrating. They moved the launches to Omelek island, which had the advantage of fewer conflicts, but the disadvantage of being in the middle of nowhere. And it probably (indirectly) caused the first Falcon 1 failure, which had a small latent defect–an aluminum bolt where there should have been something more robust. It would have been fine had they been at Vandenberg. But the Omelek site is much more exposed to the ocean spray, and the remoteness meant the rocket was often spending its time out in the open, and roughly a month of salt exposure corroded the bolt. It failed, caused a fuel leak, and the first stage failed.

430

Perhaps the entire next century of spaceflight has pivoted on that small contingency. A bit more or less money, a bit different luck, a different bolt or two and we’d all be on a very different trajectory into the future.

431

Did he? SpaceX had lots of failures, and went down wrong paths numerous times (like trying to build the world’s largest rocket out of carbon fiber composite). The key difference is that SpaceX had the willingness to abandon failed approaches and try something else. That required a guy like Musk at the top.

Most other companies would have stuck with their carbon fiber plan. No one manager would have the ability to completely change the material of the rocket. The composites division manager would fight the change, the bean counters would engage un sunk cost fallacies, and the manager who rocked the boat that hard would be making a career-limiting move.

In a company with a culture like that, the engineers at the bottom know the limits of what they are allowed to suggest, and a huge change like that might not even be suggested. Or an engineer would suggest it and take it to his manager, and the manager would not be willing to kick the idea up the chain. Or if she was willing to, her manager might not. Eventually everyone learns that innovation at the bottom is not happening except for small things.

The change to stainless was serendipitous. It enabled the hiring of cheper and more available steel workers and welders rather than expensive composites engineers. It allowed rapid design changes to the rocket. It lowered the cost of construction and therefore the cost of failure. As a reminder, SpaceX blew up quite a few prototypes on the way to a successful soft landing of Starship. If those had been composite rockets the failures would have been much more expensive and the turnaround time between attempts much longer.

Musk has allowed many other radical changes along the way when the current path they were on was too hard or too inefficient. The composite/stainless steel swap is just the best known of them.

The old way of managing risk was to try to think of every failure mode in advance and design it out of the project at the architecture level. Musk’s way is to make sure that iterations are cheap enough that you can just try stuff and keep what works.

Eventually SpaceX will have to slown down iterations, as the ‘minimum viable rocket’ becomes more complex and expensive. We’re already seeing that with the orbital prototype. When they get to in-space refueling, the ‘minimm rocket’ will be close to a final version. That’s where I expect things to slow down and take some time.

432

Hell yeah. Not just in SpaceX either, but in some of the companies that he was able to be a part of to make a pretty nice fortune in the first place.

And the money to be able to absorb those failures without having to fold up shop.

Of course it was also Musk’s idea to go with carbon fiber in the first place. You are correct that he finally changed his mind after seeing how much more difficult and costly that would be.

Right, he was also talked out of his idea of building it all “lights out” when he found out how hard and costly that would be.

That’s because NASA was using taxpayer money, and people would complain if their tax dollars get blown up on the launch pad. Even if it cost 5x as much, it still was better optics for PR. A private company doesn’t have to worry as much about that (though they still have investors), and so some failures don’t result in congress’s phone lines being lit up by angry taxpayers.

Not really all that cheap. Musk’s way is to have lots and lots and lots of money, so that the iterations are cheap relative to his wealth.

433

This focus on the original mistakes, as opposed to the ability to recover from them, is exactly the problem with traditional aerospace (and other industries). Everyone makes mistakes. The choice to use solid rocket boosters on the SLS is far worse than any misstep that Musk has made. The same goes for selecting hydrogen propellant on a rocket meant to be commercially competitive. The trouble isn’t the mistake, but rather the unwillingness to change course when it’s discovered.

Part of the reason for stainless steel over carbon fiber is just that it is so cheap and easy to handle that you can afford to iterate. If an airframe takes months to produce and tens or hundreds of millions of dollars, you are going to treat it like a Faberge egg and do nothing to risk it. And other design errors–simple ones, like suboptimal routing of cabling, overdesigned support structures, etc.–are going to be tolerated because producing a new version is so costly and time consuming.

It’s much more that he understands the virtue of a design that is possible to iterate rapidly on in the first place.

Going back to solid boosters again, one of the many flaws they have is that they cannot be unloaded of propellant. That means they are their final weight all the time. Which means the transporter device–the “crawler-transporter”, for NASA–is huge and slow. Every little issue, if it requires driving back to the assembly building, takes the better part of a day and causes so much wear that the crawler loses a significant portion of its design lifetime.

Starship, on the other hand, is basically a tin can, and they can transport it on normal roads at relatively high speed with off-the-shelf equipment (used for other construction work). They can move it to the pad and back multiple times a day if required. There’s no need to tolerate even tiny failures (whether design, manufacturing, or otherwise) since bringing it back for a fix costs so little (both in dollars and time).

434

Musk has lots and lost of money, and I have no doubt he would probably chip in to save SpaceX from going under, but if you think Musk is currently bankrolling SpaceX I wish you provide some sort of evidence. They have no problem raising as much investment money as they want right now.

435

Either you didn’t read my post correctly, or you have chosen to imply something that I didn’t say, so could you please clarify exactly what claim you think I have made that you need evidence for, specifically?

436

I take this to mean that Musk is spending his personal money on SpaceX iterations, otherwise your statement makes no sense. So I’m simply asking for why you believe Musk is bankrolling SpaceX.

437

No, just that he had lots of money at his disposal. I don’t know how you could read the words that I actually wrote and come to whatever fallacious conclusion that you thought.

I take your posts to mean that SpaceX iterations are cheap, so I’m simply asking for why you believe that it costs $20 per iteration.

438

I don’t know if I’ve had a stroke or that you are absolutely terrible at communicating, but I’m interpreting this statement, read in context to your complete post, is stating that iterations are cheap relative to Musk’s own wealth.

Now you are claiming that Musk is not bankrolling SpaceX but that he just has a lot of money at his disposal. What is the point of comparing the two if there is no relation?

This is about as pointless as saying SpaceX’s iterations are expensive compared to my own wealth. What is the point if I’m not paying for it?

439

Do you somehow think that he has billions of dollars in cash just laying around his house, and that’s what his wealth is?

No, his wealth is in the money and resources that he controls.

Anyway, I have no idea what point you are trying to make here, other than that you thought you heard someone say something negative about Musk and you had to run to the rescue.

I won’t continue this conversation until you explain why you believe that it costs $20 to launch a Falcon 9.

440

Iterative development can easily be cheaper than large waterfall development, even if you are blowing up rockets along the way. Far from being based on ‘lots and lots of money’, SpaceX’s development costs have been a miracle of low cost.

Development cost of Falcon Heavy: ~500 million (no gov funds involved)

Development cost of Falcon 9: 390 million

In comparison, SLS to date has a development cost of $11.8 billion. Delta IV Heavy cost about $1.5 billion, and it was just an extension of the existing Delta design. Ariane 6 is at $4 billion and counting.

When SpaceX applied for COTS service provider status, NASA did an analysis of what it would cost to develop using the National Air Force Cost Model:

SpaceX did it for $390 million.

Far from winning by throwing around more money than others, SpaceX has developed its rockets for less money than any other comparable rockets in history.

That Falcon Heavy launch today saved the government about half the cost of the entire Falcon Heavy development program - in one launch. When Falcon Heavy launches Europa Clipper instead of SLS, it will save the U.S. government more money than the entire development budget of Falcon 9 and Falcon Heavy combined. Actually, about twice as much.

441

To be clear, that was the Falcon 9 1.0 (plus the Falcon 1), which wasn’t reusable and had significantly lower performance than the latest iteration. Nevertheless, it was a perfectly good expendable rocket, and would easily compete in the market today.

Including the reusability and further optimization probably puts it at $1.5B total, which is still a bargain compared to the competition.