The main issue of spin-gravity space habitats is that at some point the centrifugal force will overwhelm the tensile strength of the construction materials, even if perfect graphene were available. Thus, even the biggest spin habitats can't deliver more than a continents worth of living area per drum. One can of cause trade terrain features and gravity for more livable area, but even this has its limits.

Knowing that dynamically supporting a structure is usually the answer to get megastructures much bigger (think of atlas towers and dynamically supported orbital rings, which use mass stream technology), I was wondering if active support could be applied to Banks Orbitals and Ringworlds, which are usually considered impossible without unobtainium. Obviously, "materials" with high compressive strengths won't really help me here.

However the basic job of the spinning ring is to resist the centrifugal force. So what if we could compress the ring somehow to supply a counterforce. The main issue that arises from this is of cause that the friction between the ring, which will spin at hundreds of kilometers per second and basically anything, will have spectacular results. We would need a frictionless surface. For now, let's assume that the engineers got some kind of absolutely frictionless unobtainium coating. (maybe superconducting magnets could provide a frictionless interface?)

The basic design of such a structure would be a spinning ring around a central source of gravity and an encasing, slowly counter-rotating support mass. The support mass is well below its orbital velocity, thus it would fall down towards the gravity source. However, the spinning rings centrifugal force would balance out the forces.

I'm aware that the counter masses own gravity might become a problem at some point. At some point we might very well be dealing with a circumstellar donut planet, though I think the ring world would be the better bet from a construction materials point of view. I case the question comes up, where one would get the materials to build such a thing, this is pretty much per definition a K2+ project. So starlifting is on the table. And most of the support mass would most likely be metallic hydrogen stored in graphene and metal (what else to do with several planets worth of metals) containers.

The central object could be any of a number of options: a planet, gas-giant, star or black hole. It might very well be the power source of the whole structure, either a fuel depot, a Dyson Sphere or a Penrose Sphere.

I such a structure fundamentally possible? Did I mess up the physics somewhere? Could magnets handle the pressures of the interface?

  • 2
    $\begingroup$ It is generally very difficult to answer a hard-science question involving future-tech of any sort. Consider removing the tag and refining the question to asking for realistic answers only. $\endgroup$
    – GOATNine
    Aug 31, 2020 at 12:08
  • 1
    $\begingroup$ @GOATNine replaced it with science-based $\endgroup$ Aug 31, 2020 at 12:15
  • $\begingroup$ Could it not be a solid ring but segmented and joined by cables, like the Dyson swarm idea but kept in the ring shape and kept closer together? $\endgroup$
    – user78658
    Aug 31, 2020 at 14:18
  • 1
    $\begingroup$ You could trust author Larry Niven and his Ringworld novels. When he published the first one, a bunch of students did the math and proved a passive ringworld was unstable. He added Bussard Ramjets in Ringworld Engineers to solve the problem. Apparently no one has complained about instability since then. $\endgroup$
    – JBH
    Aug 31, 2020 at 16:43
  • 1
    $\begingroup$ @JBH NIven also needed Scrith a fictional material far FAR stronger than any known form of matter to keep his ring from being torn apart. Scrith has a tensile strength similar to the strong nuclear force. and blocked as much radiation a lightyear thick piece of lead. $\endgroup$
    – John
    May 28, 2021 at 5:08

9 Answers 9


Use an outer bearing ring that does not spin and is solid metal (or glass fibre - whatever's handy) with a superconducting inner surface (making stuff cold is cheap in space - possibly nitrogen filled heat-pipes from the inner surface to the shaded outer surface of the bearing.)

Mag-lev the spinning ring off the stationary ring.

The stationary ring can reduce the load on the the spinning ring in two ways.

By its own tensile strength which does not need to compensate for a spin it that does not have, and by its weight in the stellar gravity.

So in essence the spinning ring is a mag-lev train running inside an iron hoop. It could be thought of as an inverted mass-stream the spinning ring being the mass stream that keeps the stationary ring in tension.

As pointed out in Ringworld Engineers a method of station-keepeing is still needed, perhaps solar sails would be enough.

  • 1
    $\begingroup$ This was my first thought when I was reading Ringworld, but a lot of the story wouldn’t happen without scrith as the handwavium solution, and Niven didn’t seem to understand superconductors so maybe a maglev didn’t occur to him at all. $\endgroup$
    – StephenS
    Sep 1, 2020 at 15:48
  • $\begingroup$ @StephenS Niven's Ringworld can't use solar sail stabilization because the whole ring is moving. The sail will be experiencing 1g--no sail big enough to be of any use can take this force. $\endgroup$ Sep 8, 2020 at 1:10
  • $\begingroup$ @LorenPechtel What solar sails? Niven’s Ringworld had ramjets for stabilization, which would work equally well on the rotor or stator rings proposed here. $\endgroup$
    – StephenS
    Sep 8, 2020 at 1:25
  • $\begingroup$ @StephenS Oh, I misunderstood--I was thinking you were talking about sails instead of ramjets. Note that there's one big problem with the maglev route--you need IIRC 12x the mass of Niven's Ringworld and you can't have his sheer walls, they have to be gradual--no spill mountains. $\endgroup$ Sep 8, 2020 at 3:13
  • $\begingroup$ might need a roof instead of just walls. $\endgroup$
    – Jasen
    Sep 8, 2020 at 8:58

I don't know if this will work; I haven't done ALL the maths, but my gut, mspaint, and the John Hopkins Uni Truss Simulator all suggest it'll work. So I'm presenting it; here goes:

enter image description here

Underneath the buildings, is a large tube kept at vacuum (shown in grey), which has metal pellets (shown in red) running through it, with large coilguns (brown) applying force from the structure into the pellets. They can apply both positive and negative forces, which result in both compression and tension respectively.

Centrifugal force (acting as a sheer force on the "ground" members) goes through the truss, which is redirected to alternating compression (C) or tension (T) on the wall with the tube in it, of which the compression and tensions is absorbed into the pellet stream by either accelerating it or decelerating it.

This would need to be carefully monitored and dynamic, changing mass distribution within the ring could change the force distribution, which would need to be countered. Your morning commute could require different forces to correct than your evening commute. Humid conditions on one side may make one side heavier, which needs to be correct, etc. etc.

This would also be used to counter wobbles, vibrations, and help stabilise the spin speed to an exact day length.

  • $\begingroup$ Isn't the tube itself spinning with the structure? How do you prevent it from flying apart due to the centrifugal force? $\endgroup$ Aug 31, 2020 at 16:07
  • $\begingroup$ The entire thing is essentially static, except the pellets. Coils are connected to the tube. The weight of the tube would contribute some tension to the equation, which factors into the maths of how much force to apply at each coil. Also the tube need not be load bearing - it just needs to keep the pellet stream in vacuum. So long as the members connect to the coils somehow the design should work. $\endgroup$
    – Ash
    Aug 31, 2020 at 16:21
  • $\begingroup$ The coils can only accelerate the pellets along the tube, this would apply a torque on the whole ring around the central axis. I don't see how that helps. $\endgroup$
    – mart
    Sep 1, 2020 at 9:13
  • 1
    $\begingroup$ the pellets are sidereal stationary and the ring is rotating around them, the weight of the pellets (in stellar gravity) is compressing the ring. $\endgroup$
    – Jasen
    Sep 1, 2020 at 12:29

Helium balloon.

Your scheme should work. The active support structure is spun the way you say, so that gravity and centrifugal force cancel out, with standard Ringworld attitude jets. We fill it with liquid helium-4 (the common isotope) at 2K, i.e. a superfluid with no viscosity. Inside that, we erect a ring-shaped environment with zero buoyancy. No matter how fast it is spun, its weight is the same as that of the helium (barring relativistic mass considerations, but we have a ballast system to keep things adjusted precisely).

This should be entirely feasible with current technology and no exotic materials other than what are needed to keep the Ringworld cylinder from bending/breaking under a 1 g load. (Try not to overheat the helium, because then there would be ... a mass flow)


A ringworld doesn't need a central object; in fact, a central object is a problem, because a ringworld is not stable w.r.t. a central object.

On the other hand, a ringworld with a 24-hour period (and no central object) could be built from existing materials and placed in a suitable L₄ or L₅ point (or some other stable orbit). Depending on the orientation, seasons would be odd.


As @GOATNine said, there is no good answer as we could find some cool stuff in the future.

But the same argument has been made with FTL that "maybe in the future it is possible". And as with FTL, I would say structures of this scale are impossible.

It's kind of insane to think that something like this, even if you could build it, would survive even a day of operation. Not to mention that you need to make stuff up just to make it sort of maybe possible.

The main problem I see is just the amount of force that rests on the ring. It's quite simple and we don't even need fancy equations, you can't build something this long without any supports.

Of course, this is where your active support comes in, but it doesn't really solve the problem. The best I can see MAYBE working is a metric f*ckton of other rings that each rotate a bit slower until either gravity or the amount of rings deal with the problem. So in a sense, you would have a few 100 rings or so resting on each other. Although the otter most rings would probably be more massive than the star itself.

But even this layered approach won't work too in any realistic scenario. I mean it probably could work if the outermost rings are 10.000km across and more massive than a black hole.

So in the end, as with everything, there might be a way to do it, but it's one of those "you really can't do that" way.

On a different node, if you write a story and really want a ring world, just make some material up that can support it. If you try to make it realistic people will get pissed. Nobody is going to blink an eye when you say "di-calinium is a meta material with the property that its tensile strength increases when an electric current flows through it". Does it make sense? No, but it's better than to try and force something that can work.

  • 1
    $\begingroup$ Do you have any supporting evidence? This looks like a lesser science-based answer. I understood it that active support structures are more or less achievable with known physics and material science. Consider a long copper mass being spun about in a magnetic containment structure at much greater than orbital velocity. The inertia of the container would be great, letting you rest a good deal of mass on it. With clever manipulation of the orbiting mass, you could perhaps rotate the magnetic container up to speed, simulating gravity. youtu.be/LMbI6sk-62E $\endgroup$
    – BMF
    Aug 31, 2020 at 13:34
  • $\begingroup$ Why exactly doesn't my active support solve the problem? As I see it the active support mechanism is a pretty straightforward application of orbital mechanis and newton's third law. Would you care to explain why you don't think this could work? $\endgroup$ Aug 31, 2020 at 14:19
  • $\begingroup$ Its kind of pointless arguing since you cleary want to use the idea. What evidence do you want ? There is no Material that strong, there is no way you can even build something like this and there is no point to it to. I mean i was writing a big long Answer with fancy Math but really, what would it change ? $\endgroup$
    – Erik Hall
    Aug 31, 2020 at 15:00
  • $\begingroup$ @ErikHall the entire concept essentially hinges on how well magnetically you can control the super-orbiting, er, filament. Small changes, e.g., a spacecraft docking/disembarking the superstructure, are actively countered by these magnetic stabilizers, hence "active" support. We more or less have the material science for such a thing, just not the space industry or economy. No new science needed. $\endgroup$
    – BMF
    Aug 31, 2020 at 23:09
  • $\begingroup$ I hardly think it was you but someone downvoted a lot of my stuff yesterday xD Like ok, that suddenly makes me wrong but ok. Back to you. Again, when i say it cant be done, that dosnt mean it is theoretical impossible. Just that no one will even try to do as it is way to unstable. $\endgroup$
    – Erik Hall
    Sep 1, 2020 at 10:51

Yes, but it would be fail deadly

Active support rings not in orbit can be suspended above black holes as put for by a crap ton of exotic-hard science fiction.

So yes

That being said, such a thing wouldn't be especially safe. Considering the scale of a bank's orbital such magnetically contained pellets and dust would be moving at a non negligible fraction of causality, and so carry a crap tone of energy. If the magnetic containment broke, you end up with a rail gun inside your habitat on top of a lack of support. Yikes. Not to mention the power required meaning you'd probably be better of with something passive, like solar panels in orbit around a star.

Or in other terms, yes, but such a structure would be fail deadly, even with some redundancy.


Here's one option that (spoilers) doesn't work:

  1. Build a superconducting toroid.
  2. Slowly add current and start spinning it.

The best we can do is a ~10T magnetic field. Let's see how far that gets us.

The pressure works out to B^2/2μ0, or ~40MPa. How does that compare with mundane solutions?

Answer: terribly. The tensile strength of carbon nanotubes is in the 10-100 GPa range.


Here's an option that outperforms mundane materials, but still isn't strong enough.

Consider a stacked series of thin plates of alternating electric charge. Now consider taking such a stack and bending it into a large ring.

The force between the plates of a capacitor is $$k\epsilon_0 A \left(\frac{V}{d}\right)^2$$. So the pressure is then $$k\epsilon_0 \left(\frac{V}{d}\right)^2$$.

Now, consider CCTO ceramic. It has been manufactured with a dielectric constant of up to 10^5, and a breakdown voltage of 2.38GV/m. (Admittedly, not both at once yet.)

This works out to a corresponding max pressure of ~5TPa. (I'd take this number with a grain of salt - this is large enough that these approximations will come back to bite us.)

How does that compare with mundane solutions?

Answer: quite well. The tensile strength of carbon nanotubes is in the 10-100 GPa range.

CCTO ceramic has a density of ~4.8g/cm^3. Stress on a thin rotating hoop producing $g$ acceleration is $T = \rho g R$, or $R \leq \frac{T}{\rho g}$. For 1g, this works out to a radius of ~10^8m. Still nowhere near enough for a ringworld, but better than maundane materials at least.

(That being said, you'd end up requiring a fair bit of energy to ensure that said capacitors remain topped up, and a single short could cause the entire megastructure to collapse...)


It wouldn't work.

What would work is a rotating ring inside of a very large non rotating ring. Without some sort of scifi telekinesis gravity control stuff, this is the only active support that will work.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .