My wife was watching a show about plane spotters, and them watching the last 747 flight from the UK. I was making some flippant comment, like, come on this is such a weird hobby.
She then pointed out that I know the ID of every active SpaceX booster, including prototypes, and watch nearly every launch live.
I watched the train go overhead with my wife, son and neighbor on Friday night. While I’ve seen them before, it was exciting to share my 11 year old son’s awe.
I know there’s concern about these LEO sats, but we still enjoyed counting the train as it streamed almost directly overhead from West to East.
As much as I understand astronomers' apprehension of satellite internet constellations but the first Starlink 'train' we saw from the top of a hill was absolutely captivating. I cannot begin to describe the beauty of a line of sparkling dots, moving unerringly but silently over the sky.
> More beautiful than the stars and galaxies they're blotting out?
Very much yes!
Stars and galaxies are. They are as they always[0] were, and always will be. Blobs of fusing hydrogen very far away. As beautiful as they may be if we could see them better, from where we are, they're just points of light.
(Also, wait a few seconds, the satellites will pass, and you're back to watching the stars behind.)
A constellation of satellites in the sky is also just a bunch of points of light. But it's more than that. It means something. Humans have put it there. Those satellites serve a purpose, and their appearance on the night sky reminds us of the great discoveries and feats of engineering that were necessary to put them there. They're a testament to the ingenuity and potential of mankind, and the promise that one day, our descendants will travel among them.
That's what I see when I see a satellite passing overhead.
> it's like someone admiring the engineering of a cloverleaf exchange running over a buried river
Sure, it's exactly that. A cloverleaf exchange is interesting. It's not something that just happens, and its shape is not random. A river is boring. Also full of insects that try to feast on your blood.
I'm totally serious, by the way. Ever since a kid, I've found nature boring. This actually changed for me in my adulthood - my interest in biology sparked when I realized that life is just molecular nanotechnology that wasn't designed by humans, but could still be eventually understood and controlled.
>Stars and galaxies are. They are as they always[0] were, and always will be
>[0] - On human-relevant timescales.
This was once orthodoxy, but it was overturned ~450 years ago, at least from a Eurocentric perspective.
- If you look at millions of objects, then something that only happens to each one, once in X million years, will be happening to one of them right now. Like a nova.
- If you look at millions of objects, then you will see the one-in-a-million rare one which is doing something on a human timescale. Like a pulsar.
Fair enough. Not something you can spot with a naked eye, though. In this sense, when I'm looking at the night sky, I find satellites more interesting than stars.
(My favorite thing to spot were always Iridium flares. I miss those.)
You can see meteors all the time. A bit of trivia I read somewhere is that your spot on the earth is facing more "into the wind" the later it gets after midnight, so that's when it's best to look for them.
Usually meteors are just streaks of light, but once I went partway up a mountain for a forecasted meteor shower and I saw one just like people render it for an illustration (except much smaller because of the distance) - you could see it was a white hot fireball, shading to red, shading to blue, with a long bluish tail.
> A cloverleaf exchange is interesting. It's not something that just happens, and its shape is not random. A river is boring. Also full of insects that try to feast on your blood.
The universe of knowledge one could obtain from studying the organisms and ecological connections in a healthy wetland ecosystem or the physics of stars and galaxies and supernovas and black holes and brown dwarfs is vaster by orders of magnitude than what could be won by marveling at highways and satellite trains.
The trains are only visible within a half hour of sunrise & sunset, when those stars and galaxies aren't particularly visible due to incident sunlight. If you want to see the stars and galaxies, look up at midnight when Starlink isn't illuminated by the sun. It was the best view before & after Starlink.
The half an hour is an approximation depending on latitude and season. The point is that Starlink is not visible to the naked eye without illumination, and being so close to the Earth they are shadowed by the earth most of the night. What's your latitude?
> More beautiful than the stars and galaxies they're blotting out?
At that moment, yes, absolutely. As I said, I understand the concerns and as with everything else, too much of it is going to be bad but during this moment I talked about it was absolutely wonderful.
Is there any realistic estimate for how much cheaper SpaceX will be, compared to bringing cargo up on traditional rockets? I've seen Musk claim something ridiculous like it's 1% of the cost of legacy systems but guess that should be taken with a grain of salt.
Falcon 9 is about half the price of competitors right now, though the internal cost is under half that again; a lack of competition has prevented SpaceX from needing to lower prices further.
Starship doesn't have an unambiguous cost because it's dependent so heavily on factors like market size, whether the Mars thing happens, and reliability. A marginal cost of a few million dollars is an explicit goal, but a more short-term rough goal is about Falcon 9 prices, or $50m/launch. That's both a very reasonable number, and also still more than 10x better than the competition.
There's a good chance for a lot more than that to come their way. Congress has spent almost $50B on SLS+Orion for Artemis, which Moonship would make totally redundant.
Let’s compare how much cheaper spacex is. For a 5 ton geo bird, spacex is about 2/3 the price of its competition. This however is price not cost. The best estimate for starlink is that SpaceX’s cost per satellite is about 500 thousand. This compares to oneweb’s 2 million per sat, for says that weigh half as much.
Starship could have a cost (not a price) that is very low indeed.
IIRC he was talking about the cash outlay for fuel. That’s the marginal cost to launch not counting the amortization of the rocket itself plus launch facilities etc.
They replace broken and worn parts, so at some point this becomes the Booster of Theseus problem. How many parts can you replace before it is not the same booster anymore.
I suppose the central section with tanks cannot be replaced piecemeal, and cannot be excessively patched because it's a structural part. You can consider it an airframe.
So it must be mostly the same boosters all the way, even if engines, valves, electronics, etc have been replaced multiple times.
I don't think anybody really knows; there simply isn't any data with > 10.
Back in the days when boosters were recovered, I remember hearing somewhere that SpaceX was confident that they would be able to fly them 10x, and hoped for 100x
Elon said they'll keep reusing them and pushing their luck until they have a failure on a starlink launch... so we're almost certainly going to find out.
And when they do, they'll likely have telemetry of exactly what caused the problem. Then they'll adjust, and make the problem less likely in the future.
It is quite apparent that the NASA/Government space industry has milked the public for 50+ years. We could have had re-useable first and higher stages decades ago. Then along came Spacex and revealed this scandal.
The essential features of the F9 that made it capable of propulsive landing were an accident. After the success of the F1 SpaceX didn’t have the resources to develop a larger engine, so they built the bigger rocket by just adding a cluster of engines to the first stage. This is far from ideal because multiple small engines gives poorer power to weight than fewer big engines, but SpaceX hoped to gain cost advantages by mass producing the engines.
The original plan for recovering the F9 first stage was to use parachutes and this was tried with the first couple of launches, but the boosters broke up on re-entry. That’s when the idea of a re-entry burn came up, and from that the concept of a propulsive landing.
But propulsive landing is only possible with the F9 because it has a central engine and the engines are small enough to throttle down low enough to do a suicide burn. You have to have an engine at or very close to the centre of the vehicle for it to work. But this 9 engine cluster with a central engine was not originally designed with propulsive landing in mind. It was an accident.
If SpaceX had better funding early on, they probably would have developed a larger main engine for the booster, which wouldn’t have been able to throttle low enough to stick a landing and may not have had a cluster arrangement with a central engine anyway. This is also why the other rocket companies mostly can’t adapt their rockets for reusability easily.
> This is far from ideal because multiple small engines gives poorer power to weight than fewer big engines
Are you sure about that? The Merlin engine Falcon 9 uses has the highest thrust-to-weight ratio of any rocket engine currently being used, more than double that of the much more powerful RD-180 used on the Atlas V. You'll lose some advantage because structural support and plumbing for more engines is heavier, but I'd expect they'll still come out ahead. The Raptor has been downsized from the original plans, and Musk has said that was to improve thrust-to-weight ratio.
It’s a fair point but how many engines have been through such an aggressive optimisation and upgrade program? Arguably a bigger brother of Merlin going through the same process should be more efficient.
The volume of an object goes up with the cube of it’s diameter, while its surface area increases relative to the square of its diameter. That means if you quadruple the volume of a combustion chamber and nozzle, the area of the chamber and nozzle only doubles. A similar effect applies to the plumbing. That doesn’t mean the mass only doubles, it’s a bit more than that, but less than 4X.
Rocket engines are far too complex to simply apply the cube-square law. In particular, the thrust of a rocket engine does not scale with the volume of the combustion chamber, but with the nozzle exit area (since exhaust velocity is independent of engine size). Thus, following your reasoning, thrust goes with the square of engine diameter, and weight as well, so thrust-to-weight is independent of the engine size.
Most parts of a rocket engine are structurally pressure vessels, where this rule doesn't apply, as wall thickness also scales with size.
As I understand it, the main driver of scale is that the combustion chamber has an optimal size. Too small and you get incomplete mixing, too large and you get instabilities.
> The original plan for recovering the F9 first stage was to use parachutes and this was tried with the first couple of launches
That wrong. They tried this on the Falcon 9, not the Falcon 9.
Re-usability was their goal early on and and it was just a question of how to get there. Its not an accident. You could of course claim the exact architecture they have is an accident. But then again had they developed this larger engine they might have moved to a larger vehicle with 5 of those before going to a fully reusable Starship style vehicle.
The important point is consistently worked towards a known goal.
NASA nah, they transitioned to third party launch providers many years ago. They like making space vehicles more than rockets. The SLS, yes, that's a classic pork project.
ULA is the real problem. The government allowed all the launch providers to merge into one with no domestic competition. Think space Comcast. Other countries allowed the same. Nobody innovated for decades. Russia and US ULA are flying rockets from 50 years ago with minimal updates.
> Russia and US ULA are flying rockets from 50 years ago with minimal updates.
Old Babushka proverb, if complicated explosion related plumbing is working as intended, best not to completely rework it for sake of changelog.
Minimal updates to the shape, maybe?
But I bet the electronics and manufacturing have seen steady improvements over that time.
SLS does feel quite like a pork barrel, but I understand the funding of it along with the other commercial companies as "more options, less risk", if something grounds the Dragon capsules, having a "home" alternative instead of buying seats on a foreign rocket makes sense.
> Old Babushka proverb, if complicated explosion related plumbing is working as intended, best not to completely rework it for sake of changelog.
Aye, but everyone stuck with that proverb for too long. Then Musk appeared and he doesn't seem to care that his test rockets explode. He revels in it.
To me it's the classic startup vs giant corporation. The bigger you are, the harder it is to change course. And exploding rockets look bad to management. The same reason so many companies are still on IBM mainframes running FORTRAN.
SLS is an insurance policy against SpaceX. But it's being managed by the same old wasteful monopolies as always, plus plenty of government mandated pork.
I'd guess it's pretty unlikely to occur, even when the virus detection actually deletes infected files.
Then again, I guess a filesystem filter driver can return whatever it pleases. Not sure if Windows anti-virus support works through filter drivers anymore, though.
SpaceX did not develop their launch vehicles in a vacuum. They built on lessons from the projects that came before. NASA's job is to do things that are not commercially viable.
NASA tried to build a reusable spacecraft with the Shuttle. Valuable lessons were learned.
Let's not oversimplify or cast blame. Sure, most/all of the tech has been around for decades - although Lars Blackmore & the other controls engineers did original work that was arguably absolutely critical - but there hasn't been any milking going on. The most profitable route wouldv'e been to develop this tech & then use it, after all. [Case in point: SpaceX, of course.] There simply wasn't enough interest from the public or Congress to put enough R&D money into any number of cool programs to see them through to commercialization.
Boeing and Lockheed Martin, separately and together under the name of ULA, and Northrop Grumman (as Morton Thiokol, ATK, Orbital ATK and now NGIS) have been milking the taxpayer for decades. Their business plan always involved making everything take longer and be more expensive. That is what they have to do to make more money, because they work on "cost plus" contracts.
An engineer in any one of them who presented a way to make the rocket cheaper to his boss would have been fired for not understanding how the company makes money.
The oversight is supposed to be done by congress, but they are extremely happy to provide good paying jobs in their districts, for decades.
No one even really hides it anymore. Any official, when asked about something like SLS and costs, would answer with "But NASA is going to the moon! Don't you want America to go to the moon?!" and with "tens of thousands of high paying jobs in all 50 states". There is no attempt to say the rising year over year price of orbital launch (before SpaceX), the exorbitant price of SLS + Orion, the use of solid boosters for crewed launches, etc, make any sense except to support the existing contractors to the detriment of the tax payer and space exploration.
(Or see anything by Dr. Robert Zubrin, or see the story of how the shuttle derived design that lost to RAC-2 (a Saturn V like design) was chosen to be "SLS": https://www.youtube.com/watch?v=ZNZx208bw0g&t=1970s)
I'm familiar with both Zubrin (two of his books within arms-reach as a I type this) and generally how NG/Innovation Systems does business (having worked at a legacy ATK plant).
I don't disagree that they often make money via cost-plus contracts, that SLS is mainly a jobs program, or any of that. What I disagree with is your apparent inherent assumption that it's malicious. Nobody's "milking" anything - Hanlon's razor, remember. Never assume malice when stupidity (I'd prefer "incompetence") is an adequate explanation. The project managers at ULA, at Boeing, at Lockheed, at Northrop, aren't monsters trying to squeeze more money from the taxpayer. They, quite simply, believe that this is the way you have to build a rocket and you can't possibly do it any cheaper. They do pride themselves on being able to deliver things ahead of schedule and under budget, except they can't because everybody expects projects to be behind schedule and over budget, since that's the way it's always gone. Traditional expectations have a stranglehold on the oldspace industry. I've also worked for a newspace company - there was none of that attitude.
Don't assume the astronautics primes are evil. They're just... old.
But it's not incompetence, these companies have made exorbitant amounts of money doing this. Congress rewards being late, overbudget, and inefficiently structured. The companies extoll the benefits of manufacturing in practically every state. We know Boeing was in bed with NASA, per Doug Loverro's removal. There's even the story of Boeing killing ULA's fuel depot ambitions because it threatened SLS.
> "We had released a series of papers showing how a depot/refueling architecture would enable a human exploration program using existing (at the time) commercial rockets," Sowers tweeted on Wednesday. "Boeing became furious and tried to get me fired. Kudos to my CEO for protecting me. But we were banned from even saying the 'd' word out loud. Sad part is that ULA did a lot of pathfinding work in that area and could have owned the refueling/depot market, enriching Boeing (and Lockheed) in the process. But it was shut down because it threatened SLS."
Even if the majority of employees are just there to work on the future of rocketry, if the incentives all reward the companies for being as overbudget and underperforming as possible, they are going to respond by being overbudget and underperforming.
We could’ve landed on the moon in 18th century as well, right? Given that we landed in 20th century, it’s just a conspiracy that we didn’t do it earlier.
In 1957 the Soviets launched the first artificial satellite into space.
In 1969 America landed the first man on the moon, using the Saturn V rocket, capable of launching 140 tons to LEO.
In 2021 NASA has yet to build a single rocket better than the Saturn V, with the in-construction SLS being not realistically much cheaper, and no more powerful.
We have had zero advances in launch technology from old space in fifty years. There was an attempt in 1981, but that attempt was bad and then they just gave up.
The book “Ignition!” explains the realities of chemical rockets. The Saturn V is as powerful as it gets, given chemistry and the rocket equation. Modern construction techniques and computers can bring the costs & overheads down, but they can’t make hypergolics better.
It's true we're all working in the margins of efficiency, but that's like saying there's no difference between a Falcon 1 and Starship. Saturn V isn't close to the limits of how big you can make a rocket, nor how efficiently you can make a rocket, nor how cheaply you can fly a rocket for. There have been plenty of proposals along the way for bigger, cheaper, more modular, and more reusable spacecraft, plenty of which are totally compatible with historic methods. Even literally “keep flying Saturn V” would have been better than what we did.
I think what the GP is trying to say is that reusable rockets had been known and built by NASA 50 years ago. More precisely, the Moon lander was just such a rocket that was able to land and then to lift off, albeit on the Moon, not on Earth. But the concept was there, it was just a matter of scaling it up.
By the way, I don't personally buy this argument, but it's not completely meaningless either.
They may also be referring to the McDonnell Douglas DC-X. It flew and landed itself in 1996, but it was heavily based on the 1967 Douglas SASSTO.
Neither project was government funded or ordered. There's no particular reason to think it could have even been built in 1967, and even in 1996 it may not have been possible to land a rocket after reentry. Coming in backwards at supersonic speed is incredibly challenging. Flipping around would have been even more so. Transitioning from supersonic to subsonic would have been a total guess. The DC-X certainly could not have done any of those things.
The OP may also not understand the incredible gulf in difficulty between landing a suborbital rocket and a real first stage. A suborbital rocket doesn't weigh that much more going up than it does going down. The F9 first stage weighs less than 5% of what it has to lift on the way up. Getting a rocket to output 25% thrust is easy; use four engines and turn off 3. You can't just turn off 19 engines; the F9 first stage is huge and needs huge power swings to remain stable.
SpaceX built custom sparse voxel GPU solvers to design around combustion instabilities and supersonic flows around the rocket. Even in the 90s they would have had to make huge concessions for the sake of running 20+ engines at once. There's absolutely zero chance they could have done it in the 60s or 70s.
Which, yeah, rocket science is hard! But the above is not difficult because it's impossible to solve for, it's difficult because we never tried! Nobody, government or otherwise, thought that rapidly iterating on rocket technology would be a good idea. This above all was the ticket to SpaceX's success - when you can throw rockets away for science, you can learn a lot in a short amount of time.
Did some things need to be invented before SpaceX could use them? Obviously, although I have doubts that the computations being performed by these "GPU solvers" really require a modern custom-built GPU and couldn't have been done on FPGAs or arrays of microcontrollers from the 80s. If we understand the math being crunched, it's just a matter of throwing more computing power at the problem.
To put those phrases in context, I think they would have required more effort than the Apollo project.
> Obviously, although I have doubts that the computations being performed by these "GPU solvers" really require a modern custom-built GPU and couldn't have been done on FPGAs or arrays of microcontrollers from the 80s.
The Cray 2 could do <2 gigaflops. So a single simulation with 10,000 points per side at 1 microsecond timesteps for .1 seconds with ~50 flops per cell-step would have taken 79 years to run, assuming the Cray had 50 cycle fetch/store times. Engineers need to run lots of simulations to come up with a working design. And that's just simulating something around the scale of a single nozzle- you'll need to be doing sims for a 40 meter tall stage as it flips around.
Also, the cray had a 256 Mword memory. So you could only store a single timestep of a grid of 400x400x400 cells anyway. The supercomputers of the 80s were really just not very useful for things like this.
Going to watch the video now. Thank you for the explanation! Graphics isn't my primary field so admittedly I don't know what I'm talking about. One thing I love about SpaceX is how engineers are allowed to do talks like this, I barely hear anything from the engineering division at BO or ULA.
The moon lander wasn't really reusable, though. It was a two stage disposable rocket that happened to hang out somewhere during the interstage period. Calling it "reusable" is like calling the Apollo capsules reusable because they go up, then down.
SpaceX benefits enormously from fast computers, advances in metallurgy, CAD, etc.
SpaceX is taking advantage of technologies that weren't available until recently. Of course, now that they are available, we should see SpaceX's competition adopting them.
Something very much like an expendable F9 could have been made decades ago. Simply evolving the Saturn 1B might have done it. But no, Nixon needed pork in California to ensure his reelection.
Not really. Falcon 9 is more sophisticated than it looks at first glance. Merlin has extremely high performance when compared to earlier gas generators... it has a t/w of 200 compared to the ra-27s 100 or the f-1s 95. The alluminum lithium alloys that give the rocket such a low dry mass date from the 90s. The avionics in falcon are revolutionary, and couldn’t have been possible before the early 2000s... a lot of little things like that.
That's why I said "evolving". Making engines better, introducing FSW when it was introduced, and so on. The F9-like boosters that could have been made around 1972 would still have been much better than the shuttle turned out to be.
What one should object to in this scenario is the idea that the conventional aerospace companies, and NASA, would have done anything like this, when their internal incentive structures pushed them away from it.
We did build an f-9 like booster around the time of the shuttle, the delta ii. It doesn’t compare to falcon. Neither does Antares, built at the same time as the f-9
Delta-II was not a F9-class launcher. The first stage thrust (ignoring strapons, which one should because F9 doesn't use them) was less than 1/7th that of the F9. A closer match was the Saturn IB.
F9 1.0's payload to LEO was 11.5 tons, vs. (up to) 6.8 tons for Delta II.
Yes, D9 was designed around boosters. This means D9 is a different kind of design than F9. I'm contending that a more F9-like design, like Saturn 1B, could have evolved into something like the F9 over time. D-II could not have.
She then pointed out that I know the ID of every active SpaceX booster, including prototypes, and watch nearly every launch live.
Oh no. I'm a rocket spotter.