Maybe they’ve been playing KSP? : )
That type of macro-modular design seems like it could have a lot of potential but also pitfalls. I guess that’s what all the expensive testing is for.
Maybe. One of the most valuable aspects of the SpaceX approach is their building in of margin. Margin comes in many forms–structural, failure resistance, propellant, etc. SpaceX has taken the approach of not designing right to the edge of their specifications: building a rocket that is more massive than it really needs to be but can then spend that extra margin on fun things like landings.
Some of this margin, I expect, is spent on covering for core mismatches. Does one core have slightly reduced thrust due to wear? No problem; there’s thrust to spare (after all, they can also have a successful mission with an entire engine failure). What about a core that’s a little heavier than the others, because it was older and didn’t come with the same mass reductions? Also no big deal. The engines have plenty of gimbal range and there’s also extra propellant.
The fact that they’re landing at all also gives them margin–they can always choose to sacrifice one or more cores to ensure the payload makes it to orbit. They can ensure their primary mission succeeds even with significant failures along the way.
All of this means they can have more of a backyard mechanic vibe as opposed to building pristine lab specimens. Cars have a ton of margin, and you can bolt together just about any set of aftermarket parts and get it to work. Aerospace will never be quite like that (at least not without a revolution in engine design, like fusion rockets), but clearly SpaceX has taken a small step away from the usual design style, and this has given them significant flexibility.
I expect Starship/Superheavy to have even more margin. It has better propellant, more engines, generally milder engine conditions, “easier” materials, and so on. It’s so big that it’ll be very mass efficient, and the fact that both stages are to be reused means they get a ton of margin back if they choose to sacrifice a landing.
The aerospace industry’s insistence on things like the payload-to-total-weight ratio has struck me as myopic: It ends up spending $90 in R&D to save $10 in materials and propellant. It’s armies of highly skilled workers which are in short supply and cost dearly, not rocket fuel. If you over-engineer by 10% and materials + propellant cost you at 10% of total cost, then that only increases your costs by 1%. If that allows you to bypass a lot of lengthy and costly R&D, that can be worthwhile. They’re focused on the wrong type of efficiency.
That’s the word.
It’s not the only area like that. Up and thru WWII, people thought there would be breakthrough tanks (slow, heavily armored and heavily armed to breach enemy lines), exploitation tanks (light and fast) and tank destroyers. Then people realized that instead of having highly specialized vehicles that are just so, like pristine lab specimens, it made more sense to take a looser, more flexible approach with MBTs (main battle tanks) even if it wasn’t 100% effective in the narrow case it was intended for. Because it was more important to make room for the unknown, both good and bad. The ability to respond to non-predicted problems or opportunities. Full-power battle rifles and low-power submachineguns were largely replaced by assault rifles which can handle everything in a way that’s not the most excellent at any one task but can perform pretty well across the board; Like a CPU. Same thing for machine guns; LMGs (light machine guns) and HMGs (heavy machine guns) largely gave way to GPMGs (general purpose machine guns).
“General purpose”, does that ring a bell to you, to take a piece of technology which was conceived as being useful for 1 thing and put “GP” in front of it?
Now that I think about it, the US restaurant business has figured out that when labor is expensive and materials cheap, you pile on the materials (food, propellant). That’s why US restaurants and cinemas have such big portions. The main cost isn’t the popcorn or potatoes, it’s the rent/facilities/employees/contractors: Land & people.
Modular design: One of the advantages of modular design whether it’s in software, rocket science or anything else is that you don’t have to spend much time figuring out how to make parts fit together; If they really are modular, they should fit together as easily as Lego blocks. It saves a lot of expensive labor when you don’t have to extensively test how thousands or millions of parts being pushed to their limit interact together.
When I was running a movie theatre, the most expensive thing was the popcorn container, followed by the staff making it. The contents amounted to a rounding error.
IIRC, propellant cost is about 2% of launch costs. It’s ridiculous to worry about using propellant efficiently. It’s like worrying about gas efficiency in 1950s America. The walls of the rocket can’t be all that expensive either. Get rid of the complicated pump system by making it solid propellant and you can keep most of the lift really simple and cheap, only using liquid-propellant rockets for the parts of the flight that require the most precision. Even then, you could make either the oxidizer or the fuel solid while the other is liquid. My gut says it should be the fuel that’s solid and the oxidizer that’s liquid but guts lose to brains most of the time and I don’t know enough to figure it out.
It’s not just the cost of propellant, it’s the tyranny of the rocket equation. Every pound you add to the rocket is a pound of payload you can’t take into space. But more importantly, if the rocket gets too heavy you can’t even get into space. That’s why the holy grail of Single-Stage-To-Orbit has been so hard to achieve. If the mass of the ship gets too big, either you can only get a tiny payload into orbit, or you can’t get into orbit at all. It’s why we stage rockets.
What SpaceX showed was that there’s enough margin in a staged rocket to allow at least the first stage to propulsively land, and that the cost in lost payload and capability was far less than the savings from reusing the stage. But there’s a limit to how far you can take that, as can be shown by the fact that SpaceX abandoned plans to re-land the second stage.
Now, the new Starship is much larger, which decreases the mass of the physical rocket compared to the mass of the fuel. That gives them the margin to also propulsively land the upper stage, which is what will make that rocket so cheap to fly. 100% reusability is the holy grail of rocketry, and Starship and its booster may achieve that.
It’s not the cost of the propellant; it’s about whether you can even get to orbit. If your Isp and dry mass fraction both suck because you don’t care about efficiency, it doesn’t matter how much you scale up you’re not going to get enough delta v.
I was thinking not just in terms of making individual rockets bigger but also of increasing the number of stages. Has someone tried 4 or 5-stage rockets? The rocket equation may prevent the scaling up of any one stage but that doesn’t mean it prevents the further scaling up of staging. As long as the first stage to ignite can get off the ground a couple kilometers and you direct that first spent stage to a landing/impact ground, you can take it from there.
You run into another problem: More stages means more engines, more interstages, more complexity, more weight that the first stage has to lift. And then you have to lift all the fuel needed for that weight as well. Assembling it all is also difficult. And the bang for the buck goes down as the first stage is by far the largest.
Depending on how you count stages, the Saturn V/Apollo craft had a few. The Saturn V itself had three stages, Then the Apollo service module had its own engine for lunar orbit and Trans-Earth Injection. Then there’s the descent engine in the LEM and the ascent engine. But all that stuff would generally be considered payload of the rocket, so the Saturn V had only three stages.
The most number of stages I could find on any rocket is five. But it seems like the ‘sweet spot’ for rocket staging is two or three.
When entertaining possibilities in aerospace, keep in mind that one of the reasons Musk is so successful is because he doesn’t let himself be held back by people telling him “everybody knows you can’t do that!”. He’s not interested in common sense and consensus. He’s interested in principles, just like Marcus Aurelius. If he were concerned with what everyone knows you can’t do, SpaceX would be like every other aerospace company.
If you have the time, here is a quite interesting 20 minute report about a marine tech inventor who, because he was an outsider who didn’t know that “Of course, you can’t do that!” went ahead and did it: Sucessful Inventor Tells You How He Did It Yet AGAIN. - YouTube
You may also want to give the movie Moneyball a chance by just watching its first 20 minutes and then deciding if you want to watch more.
ETA: This post isn’t a reply to the one immediately above it but an addendum to mine. I’ll come back to your post when/if I have a good reply to it.
Sam Stone and Gorsnak are right about efficiency–propellant is cheap, but your rocket won’t make if your efficiency is too low. So it’s wise to make your rocket big, but you can’t make too many efficiency sacrifices.
IMO, solid fuels are a bad idea. They are difficult to produce and handle, they don’t allow for ground testing, and just generally have a host of problems.
However, your thoughts about the pumps are sound. Behold the Sea Dragon. It’s a design for an absurdly large rocket–490 feet tall and 75 feet diameter. It had a single gigantic first-stage engine. And most relevant to your point: it’s pressure fed and made from common materials (the kind of steel used for submarines). It could have been manufactured in a drydock, which was extra convenient because it was also to be launched floating in water.
So it should have been very cost effective; probably somewhere between Starship and Falcon Heavy costs. Unfortunately, the design never went anywhere, though the water launch part was tested successfully.
I’m not sure how to have a good reply to, ‘All we need to do is think outside the box’. That’s not really a statement you can refute, because it makes no claim that can be tested. I could say that about anything technological, but it doesn’t advance the debate.
Musk didn’t think totally outside the box anyway. What he did was discover new possibilities for powered landing created by advancements in materials and computing power, and realized that the old paradigm no longer obtained. He didn’t so much break through a technological barrier than through the resistance to change from a hidebound bureaucracy and a contracting system that did not incentivize innovation and cost reduction.
NASA could have done it, but its aging management and the political demands of maintaining current assembly lines and facilities in many states meant they locked themselves into aging Shuttle derivative hardware. Because they weren’t interested in profit and used taxpayer money, there was no price to pay for this. And because their contractors use cost-plus accounting, anything they innovated to cut costs would also cut their profit. So we got stuck with basically the same infrastructure we had in the 1970’s. THAT is what Musk broke - not physics or huge technological barriers.
That’s very different than just saying, "Hey, maybe someone will come up with a nine-stage rocket’. They can, and they might even make it fly, but you can’t cheat the rocket equation, and diminishing returns with each stage makes a large number of stages useless. Also, recoverability requires size. The smaller the stage, the bigger the penalty for landing legs, extra fuel, grid fins, and the structure to hold it all together. Also, every stage is going faster and would be harder to land - which is why Falcon 9 only recovers the first stage. So a many-stage rocket would need throwaway stages, and that’s exactly what we’re trying to get away from.
As long as the entire spaceship has a T/W of at least 1.1 on take off, increasing as the propellant is expelled and with every stage, what is “too low”?
Which is why they’re the most common type of propellant for modern NATO ICBMs?
Note that I’m not saying only solid prop should be used: The upper stages require the gradual control over thrust that liquid prop provides.
Can you give me an example of a kind of claim I could make which would have the potential to satisfy you and not require me to have access to proprietary data, several engineering degrees or go through reams of spreadsheets? I’ll do the best I can with what little I have.
I did some math on what it would take to place a 1 kg payload into orbit using nothing but Estes B-class motors. It requires approximately 14.4 sextillion motors, with a total mass of 275 quintillion kilograms. That’s about 20% of the mass of Earth’s oceans.
Tsiolkovsky will make you his bitch with that exponent in the rocket equation.
They want storeability and shock resistance above almost all else. Cost isn’t a big consideration, nor is payload capacity, really (warheads are pretty lightweight these days).
Solid propellants are used in ICBMs because they cannot take the time to be fueled, and you can not leave liquid fuels loaded in a rocket indefinitely. It’s a special use-case.
Solid fuel rockets cannot be throttled or shut off if something goes wrong. When solid fuel rockets explode, they spray flammable, burning solids everywhere. One of the problems they had with Ares 1 when they tried to use a solid-fuel man-rated booster was that in an explosive emergency there was no way for the recovery system to get away from the cloud of burning propellant unless it was sent a LONG way from the rocket, which would have made it too heavy.
Solid fuels are also very hard to pack perfectly, which leads to power pulses and vibration - so much so that it was determined that the vibration in Ares could have been fatal if a crew were aboard.
That’s why solid fuels are now relegated to boosters and not primary stages.
A specific claim works. Like, “we will be able to do X soon, because research into X shows that once we solve problem Y it will be feasible.” Then we can have a discussion of X and Y and see if that’s true.
But if you say, “Hey, people throught X couldn’t be done, but a smart guy did it. So Y can also be done - we just need another smart guy to think outside of the box” then you are essentially engaged in hand-waving. There are lots of problems out there that we have never solved, and there will be plenty more. Just because someone comes up with a clever solution to a problem does not mean that all problems are solvable.
Then it would probably be a bad idea to use a model rocketry SRB with 0.9 second burn time and 4.3 Newton-seconds. Could you do the same calculation with the lowest stage of an LGM-118 Peacekeeper or UGM-133 Trident II?
I was responding to your argument that “They are difficult to produce and handle, they don’t allow for ground testing, and just generally have a host of problems.” I have difficulty understanding how they can be difficult to handle but have high storeability, shock resistance and be launched from trucks, submarines and aircraft.
You asked how low is “too low”, so I figured I’d try to give a lower bound :).
Solid-fueled 3- and 4-stage rockets make it to orbit just fine. ICBMs get regularly repurposed into orbital launchers already. But no one would build a launcher that way to start with; it’s only practical because military leftovers are cheap (they might even pay you to haul them away).
Once assembled into a rocket, solid fuel is very stable and doesn’t require any special conditions, like cryogenic storage. I suspect you could hold a torch to it and not set it off–if not, at least a match.
But here’s what can happen at a solid propellant factory. Of course, liquid oxygen and hydrocarbons can both be dangerous, but the overall level of handling expertise is greater if for no other reason than ubiquity. LOX and kerosene trucks drive all over with no problems.
A solid fuel rocket can’t be tested before flight. So it needs to be manufactured to the highest precision possible and then inspected, and reinspected, and reinspected again. And then you still cross your fingers and hope for the best. Whereas with a liquid fueled engine you build it and then do a solo ground test (static fire), and then another static fire when integrated with the other engines and the rest of the rocket. You have very high confidence that the system will work at that point.
So¿id fuels definitely have their uses. As you said, they’re great for smaller rockets that need to remain fueled and stored. They just have major problems for large rockets, and definitely for man-rated rockets. The issues with Ares 1, a shuttle booster derived rocket, lead to its cancellation.
The first ICBM’s like the Atlas and Titan were actually liquid fueled. The Titan II used a hypergolic liquid fuel that led to an accident in 1980 where a Titan II was damaged and began leaking hypergolic fuel into the silo, blowing the hell out of it. The blast blew off the 740 ton silo door and launched the second stage and the 9 megaton warhead about 100 meters from the perimeter of the facility. No radiation was leaked, but it shows how dangerous liquid rocket fuel in an enclosed silo can be.
There are still liquid fueled ICBMs around - that’s what North Korea uses, and Russia’s newest ICBM design uses liquid fuel. China also has liquid fueled ICBMs.
The other issue with Hypergolic fuels and oxidizers is that they are toxic as hell and corrosive. Keeping missiles fueled for a long time can cause problems when the tanks start to dissolve.