Love the painting of the huge toroidal space station—with the houses and forests inside.
I had a thought experiment: if you could ride a bicycle (motorcycle?) against the direction of spin of the station you would essentially be "stationary". You would still have a velocity into the always-sloping-up wheel. What if you rode up a gentle ramp? Could you break away from the surface of the wheel then and become "weightless"?
The Bigelow stuff was very promising and showed that it could work. The larger units on extruded spokes was a viable path to a .5G space station. This would be doable with three (possibly 4) Starship launches[1].
[1] Caveat Starship has to reach its goal of transporting 100 tonnes to LEO
Ya I really like their airship to orbit concept. I asked AI about lift to drag (L/D) ratios in plasma at 10,000-17,5000 mph (5-8 km/s) and it suggested that lifting bodies achieve between about a 1:1 and 3:1 L/D ratio. If we assume the generous 3:1 L/D ratio, that would seem to make a single-stage to orbit space plane possible.
A bit off-topic, but an aerospike engine is half of a rocket nozzle, with a virtual half created by the supersonic shockwave. So we could envision a retractable nozzle half that moves through subsonic, transonic and supersonic modes to power the airship.
Also the SABRE engine uses (according to AI) 16,800 thin-walled tubes filled with liquid hydrogen to cool ambient air to -238 F (-150 C or 123 K) in 10 milliseconds so that it can be compressed up to 140 atmospheres and fed into a combined-cycle engine. That would allow it to be air-breathing up to mach 5.4 (3,600 mph or 1.6 km/s) and transition to liquid oxygen after leaving the atmosphere.
I also asked it about using something like titanium to withstand the heat of exiting the atmosphere (since the titanium SR-71 reached mach 3+) but it said that it can't withstand a high enough temperature. So an ablative coating might need to be applied between launches. Quite a bit of research was done for that through about the 1970s before NASA chose the space shuttle with its reusable tiles.
It seems like most of the hard work has already been done to achieve this. So I don't really understand why so many billions of dollars get devoted to other high-risk ventures like SpaceX. When for a comparatively smaller amount of money, a prototype spaceplane could be built. I'm guessing that the risk/reward value just wasn't proven yet. But really shouldn't VC money chase the biggest bet?
This is the kind of stuff that I went down rabbit holes for when I dreamed of winning the internet lottery. Now that AI is here, I can feel the opportunity for that slipping away. A more likely future is the democratization of problem solving, where everyone knows everything, but has little or no money and doesn't want to pay for anything. So really not much different from today. So maybe it's better to let these half-baked ideas go so that someone else can manifest them.
Aerospikes are hard because you cant control the external pressure. At altitude A you have X atmospheric pressure, but at altitude B, Y pressure, that pressure is what keeps the exhaust against the surfave and exerting force, you can only design an aerospike for a certain effecient operational altitude and outside of that its just not great.
A nozzle engine doesnt have to account for this as much because the nozzle is keeping the pressure of the exhaust
Nozzle engines absolutely have to account for the external pressure. The optimal pressure as the exhaust leaves the bell should be as close as possible to ambient for full thrust.
If the pressure at exhaust is higher than ambient, the exhaust pushes outward against the ambient pressure and you get huge exhaust plumes, and lost efficiency.
Conversely, if the pressure at exhaust is lower, the ambient pressure pushes the exhaust inward into shock diamonds[1] and you, again, lose efficiency.
Engine bells specifically yield their max efficiency at one external pressure/altitude. The reason you see shock diamonds is most often from ground-level testing (or takeoff) of engines that perform best at altitude.
Airship To Orbit is JP Aerospace, not Bigelow. It seems like an utterly bonkers and fairly implausible concept and I'm definitely not equipped to analyze its merits. But the JP team have some legitimate accomplishments in the rockoon world, and appear to be honest, hardworking people. Definitely not grifters. I've been following their work on ATO since they first announced it at a Space Access conference in ... 2003, I think? Still can't figure out whether it's real or not.
I just saw a ship that looks like some of those pictures earlier today: The Beckett-Class science vessel in the Cygnus point of interest by the current Elite Dangerous community goal station at HIP 87621.
On the topic of space games, Kerbal Space Program also had a mod to provide various inflatable space station modules. I loved them - very compact to launch and provided a fun place to float around in IVA on the way to the destination. I brought mine to Moho.
I never understand why the rotating station concepts seem to all have rigid tethers, either in the form of a central boom or a rigid circular structure. It would seem like you could get a much larger diameter, so less rotational velocity and more comfort, by attaching rigid, or inflatable in this case, structures with a tether. Compressive loads are non existent, you just need to resist tensile loads.
An array of steerable ion engines hanging below the station (ie on the edge not in the center) can provide both reboots / stationkeeping and precess the axis, eg once per year to track the Sun.
Because trig, by "mixing" both maneuvers together it uses less propellant vs doing the two maneuvers separately.
This is a really nice article that covers the history of space habitats, but it also made me realize that the future of habitable structures has questionable value outside of space tourism.
We have entered an age where humanoid robots are beginning to do many tasks that we thought were exclusively in our domain. At our current pace, I expect they will be able to outperform us in most work settings within a decade or two.
As those robots scale up in their capabilities and numbers, we will send up a fleet of them to space to do the work there. They are far more suited for the environment than humans, and the cost savings would be huge.
The vast majority of work done in space is already done by machines.
Humanoid robots are potentially useful when operating in human environments, but that doesn’t really apply if we’re never sending humans to these locations.
Agreed, for space being humanoid is optional. Legs are superfluous for a start. However the issue with current automated systems we’re sending out there is they are function specific, and not very adaptable to novel or unanticipated activities.
This is why sending humans is often advantageous, we can do lots of different and new things. The ideal multipurpose space robot may not have to be humanoid, but it would need to replicate or ideally exceed this kind of flexibility of function.
The humans will end up spending 99% of their aggregate time in locations that are large enough to bore habitat out of solid rock. Moons, planetoids, substantial asteroids.
If we're talking about objects in the asteroid belt, the internal consistency varies wildly. Most smaller objects are indeed rubble piles: boulders, pebbles, and sand-like grains, down to dust. As you'd expect, they're very weak structurally. Notably, the OSIRIS-REx probe was nearly swallowed by its loose material during the sample collection on Bennu.
Some of the other large, iron-rich asteroids like Ceres, Vesta, Pallas, and Interamnia are more like protoplanets than rubble piles.
Besides the concern for structural integrity/stability, they also have reasonable amounts of water ice, volatiles, metals, ad other resources needed to supply an outpost.
One of the problems with space stations is that we can make them longer, or we can pull off pieces and replace them with bigger ones. We don’t have a way to make them larger around.
And the way contact points work, I don’t think we have a way to even inflate a new section around an existing one.
I suspect without ISRU production of bulk orbit sheet metal, the most feasible solution is to repurpose rockets in their whole.
Building a station this large is gonna be costly even within the cargo hold of starship. But six of them, gutted of insides as welded end to end could provide the vast majority of the bulk mass.
This assume rather sophisticated orbital welding and object manipulation; but its feasible we could do it with robots.
aren't rockets like the starship almost the opposite of what you want in a space station? They want to minimize the integrity of the rocket as much as possible (without blowing up) to reduce the mass while for a station you want robustness (for pressure & impacts).
Love the painting of the huge toroidal space station—with the houses and forests inside.
I had a thought experiment: if you could ride a bicycle (motorcycle?) against the direction of spin of the station you would essentially be "stationary". You would still have a velocity into the always-sloping-up wheel. What if you rode up a gentle ramp? Could you break away from the surface of the wheel then and become "weightless"?
The Bigelow stuff was very promising and showed that it could work. The larger units on extruded spokes was a viable path to a .5G space station. This would be doable with three (possibly 4) Starship launches[1].
[1] Caveat Starship has to reach its goal of transporting 100 tonnes to LEO
Ya I really like their airship to orbit concept. I asked AI about lift to drag (L/D) ratios in plasma at 10,000-17,5000 mph (5-8 km/s) and it suggested that lifting bodies achieve between about a 1:1 and 3:1 L/D ratio. If we assume the generous 3:1 L/D ratio, that would seem to make a single-stage to orbit space plane possible.
A bit off-topic, but an aerospike engine is half of a rocket nozzle, with a virtual half created by the supersonic shockwave. So we could envision a retractable nozzle half that moves through subsonic, transonic and supersonic modes to power the airship.
Also the SABRE engine uses (according to AI) 16,800 thin-walled tubes filled with liquid hydrogen to cool ambient air to -238 F (-150 C or 123 K) in 10 milliseconds so that it can be compressed up to 140 atmospheres and fed into a combined-cycle engine. That would allow it to be air-breathing up to mach 5.4 (3,600 mph or 1.6 km/s) and transition to liquid oxygen after leaving the atmosphere.
I also asked it about using something like titanium to withstand the heat of exiting the atmosphere (since the titanium SR-71 reached mach 3+) but it said that it can't withstand a high enough temperature. So an ablative coating might need to be applied between launches. Quite a bit of research was done for that through about the 1970s before NASA chose the space shuttle with its reusable tiles.
It seems like most of the hard work has already been done to achieve this. So I don't really understand why so many billions of dollars get devoted to other high-risk ventures like SpaceX. When for a comparatively smaller amount of money, a prototype spaceplane could be built. I'm guessing that the risk/reward value just wasn't proven yet. But really shouldn't VC money chase the biggest bet?
This is the kind of stuff that I went down rabbit holes for when I dreamed of winning the internet lottery. Now that AI is here, I can feel the opportunity for that slipping away. A more likely future is the democratization of problem solving, where everyone knows everything, but has little or no money and doesn't want to pay for anything. So really not much different from today. So maybe it's better to let these half-baked ideas go so that someone else can manifest them.
Pretty sure it's Scott Manley that did an episode on the aerospike and just how hard they are to get to work right.
Needless to say, getting anything to go to space is hard.
Aerospikes are hard because you cant control the external pressure. At altitude A you have X atmospheric pressure, but at altitude B, Y pressure, that pressure is what keeps the exhaust against the surfave and exerting force, you can only design an aerospike for a certain effecient operational altitude and outside of that its just not great.
A nozzle engine doesnt have to account for this as much because the nozzle is keeping the pressure of the exhaust
Nozzle engines absolutely have to account for the external pressure. The optimal pressure as the exhaust leaves the bell should be as close as possible to ambient for full thrust.
If the pressure at exhaust is higher than ambient, the exhaust pushes outward against the ambient pressure and you get huge exhaust plumes, and lost efficiency.
Conversely, if the pressure at exhaust is lower, the ambient pressure pushes the exhaust inward into shock diamonds[1] and you, again, lose efficiency.
Engine bells specifically yield their max efficiency at one external pressure/altitude. The reason you see shock diamonds is most often from ground-level testing (or takeoff) of engines that perform best at altitude.
[1] https://en.wikipedia.org/wiki/Shock_diamond
Airship To Orbit is JP Aerospace, not Bigelow. It seems like an utterly bonkers and fairly implausible concept and I'm definitely not equipped to analyze its merits. But the JP team have some legitimate accomplishments in the rockoon world, and appear to be honest, hardworking people. Definitely not grifters. I've been following their work on ATO since they first announced it at a Space Access conference in ... 2003, I think? Still can't figure out whether it's real or not.
I just saw a ship that looks like some of those pictures earlier today: The Beckett-Class science vessel in the Cygnus point of interest by the current Elite Dangerous community goal station at HIP 87621.
On the topic of space games, Kerbal Space Program also had a mod to provide various inflatable space station modules. I loved them - very compact to launch and provided a fun place to float around in IVA on the way to the destination. I brought mine to Moho.
I never understand why the rotating station concepts seem to all have rigid tethers, either in the form of a central boom or a rigid circular structure. It would seem like you could get a much larger diameter, so less rotational velocity and more comfort, by attaching rigid, or inflatable in this case, structures with a tether. Compressive loads are non existent, you just need to resist tensile loads.
Maybe I'll go ask the AI.
See Project Hail Mary.
And Seveneves
Stationkeeping would be a problem. And difficult to stop it precessing away from the rotation axis you want.
An array of steerable ion engines hanging below the station (ie on the edge not in the center) can provide both reboots / stationkeeping and precess the axis, eg once per year to track the Sun.
Because trig, by "mixing" both maneuvers together it uses less propellant vs doing the two maneuvers separately.
This is a really nice article that covers the history of space habitats, but it also made me realize that the future of habitable structures has questionable value outside of space tourism.
We have entered an age where humanoid robots are beginning to do many tasks that we thought were exclusively in our domain. At our current pace, I expect they will be able to outperform us in most work settings within a decade or two.
As those robots scale up in their capabilities and numbers, we will send up a fleet of them to space to do the work there. They are far more suited for the environment than humans, and the cost savings would be huge.
The vast majority of work done in space is already done by machines.
Humanoid robots are potentially useful when operating in human environments, but that doesn’t really apply if we’re never sending humans to these locations.
Agreed, for space being humanoid is optional. Legs are superfluous for a start. However the issue with current automated systems we’re sending out there is they are function specific, and not very adaptable to novel or unanticipated activities.
This is why sending humans is often advantageous, we can do lots of different and new things. The ideal multipurpose space robot may not have to be humanoid, but it would need to replicate or ideally exceed this kind of flexibility of function.
The humans will end up spending 99% of their aggregate time in locations that are large enough to bore habitat out of solid rock. Moons, planetoids, substantial asteroids.
I thought most asteroids were basically just gravel piles loosely held together by internal gravity?
If we're talking about objects in the asteroid belt, the internal consistency varies wildly. Most smaller objects are indeed rubble piles: boulders, pebbles, and sand-like grains, down to dust. As you'd expect, they're very weak structurally. Notably, the OSIRIS-REx probe was nearly swallowed by its loose material during the sample collection on Bennu.
Some of the other large, iron-rich asteroids like Ceres, Vesta, Pallas, and Interamnia are more like protoplanets than rubble piles.
Besides the concern for structural integrity/stability, they also have reasonable amounts of water ice, volatiles, metals, ad other resources needed to supply an outpost.
One of the problems with space stations is that we can make them longer, or we can pull off pieces and replace them with bigger ones. We don’t have a way to make them larger around.
And the way contact points work, I don’t think we have a way to even inflate a new section around an existing one.
I suspect without ISRU production of bulk orbit sheet metal, the most feasible solution is to repurpose rockets in their whole.
Building a station this large is gonna be costly even within the cargo hold of starship. But six of them, gutted of insides as welded end to end could provide the vast majority of the bulk mass.
This assume rather sophisticated orbital welding and object manipulation; but its feasible we could do it with robots.
This was considered with the orange Space Shuttle fuel tanks; they went almost all the way to orbit anyways.
aren't rockets like the starship almost the opposite of what you want in a space station? They want to minimize the integrity of the rocket as much as possible (without blowing up) to reduce the mass while for a station you want robustness (for pressure & impacts).
Anything that can be inflated can usually be deflated.
so what? it's only one bar of pressure. Bring some gum.