• Jan. 9, 2021, 5:05 a.m.

    Hello, Everyone.

    I've been kicking around an idea for a high-thrust spacecraft engine to make transport to orbit much safer and cheaper than chemical or nuclear rockets, and I'd like your input on the physics and practical engineering of it.

    I imagine most of you are familiar with the basic physics behind a rocket engine, namely conservation of momentum: "speedy stuff goes down; speedy thing goes up." And there are (essentially) four factors in that equation: rocket mass, rocket speed, propellant mass, and propellant speed; If you want to get your rocket to a certain speed, like orbital velocity, you can only make it so light, and you can only carry so much propellant before you get diminishing returns, so if you want to make the best rocket possible, you've got to get the propellant moving as fast as possible. I figured that the fastest you could ever get your propellant is light-speed, or at least close to it, so why not design an engine to do that, instead of bothering with chemical reactions and gasses heated by nuclear fission that, at best, can only approach the speed of sound?

    We kind of already have the technology to do that, in the form of particle accelerators, which I suppose an Ion Drive already is, specifically a one stage linear accelerator. However ion drives have pretty low mass flow rates, and their propellant atoms only get to around 0.6-1.6% light-speed; I think we can do a lot better.

    So imagine an engine that works more like a cyclotron than an linac, where the propellant atoms are accelerated tangentially about a cylindrical vacuum chamber by an electric field, and centrifugally by a magnetic field, until they reach a speed at which they cannot be contained by the magnetic field anymore, and exit the vacuum chamber tangent to its edge. I did some calculations and found that, for a 1 meter diameter engine with a magnetic field strength between 1 and 1.4 Teslas (which Google told me was average for rare-earth magnets but I'm not convinced) you could accelerate oxygen and nitrogen molecules to 5-8% light-speed; hydrogen atoms would give the best possible performance at 15-21% light-speed. This should result in fuel-to-mass ratios for launch vehicles of 0.22-0.15% and 0.0058-0.0042% respectively, though an oxygen/nitrogen engine could be air-breathing and not need to carry virtually any fuel at all; compare this to the the fuel-to-mass ratio of the Space Shuttle, which is about 82%. To put that in perspective, a 2000 kg spacecraft would only require 3-4.5 kg and 0.08-0.12 kg of total fuel respectively.

    That sounds like a game-changer to me, so why hasn't someone made one yet? While I've managed to dig up a few scholarly papers discussing whole spacecraft that travel at relativistic velocities, I haven't heard anyone even suggest getting the propellant up to those speeds.

    Is there a problem with my designs that the experts know but I don't? I've considered things like energy requirements, thermal velocity, cyclotron radiation, and maintaining vacuum inside an open cavity, but so far I've come up with solutions that seem to make those problems manageable. Or maybe it's just that strong permanent magnets haven't been around long enough for this to cross anyone's mind yet.

    Here's a link to my calculations if anyone wants to dig into the math I've been using: NCIE Calculations Spreadsheet.
    I'll try to keep it up to date; sorry that the diagrams didn't translate over too well to Google Drive.

    I'd appreciate hearing everyone's thoughts on this, particularly on technical problems that either need solving or are actually unsolvable, so I can stop wasting my time. And, hey, if it turns out to be feasible after all, maybe we'll all be taxiing to and from orbit in ships like the Millennium Falcon before too long. : )

  • Jan. 22, 2021, 2:20 p.m.

    Your diagrams are clear. I wish I understood this topic better. From the Machine Thread I've been reading more about particle accelerators lately.

    I think one edge cyclotrons have over linear acceleration is that they can apply force on the same propellant many times with the same hardware. Imagine if you are a rocket and your propellant is a bowling ball. First you push the ball out the stern and get some thrust, but a second later you can reach out with a long pole and push against the ball again and get some more thrust, and a second later again, and so on, until nearly all of your momentum comes from the ball's speed and not its mass. To match this a linear accelerator would have to be extremely long, with many redundant rings.

    So we're sort of talking about an iterative ion drive 'recycling rocket'? It could be a whole species of propulsion with a lot of different approaches.

    Again I can't answer your questions, but if you are looking for flaws, maybe one of them would be that something of any significant mass spinning at 20% of C could tear your engine apart. Don't know if that is a deal breaker though. The mass is held together by a magnetic field. So if the propellant is on both sides of the loop it may pull against itself as much as against the magnet.

    So perhaps the next math problem is to figure out the centripetal 'force' a regular ion drive would have to endure, if the ions were moving in a circle at relativistic speeds?

  • Jan. 22, 2021, 9:56 p.m.

    Wouldn't there be an energy cost associated with breaking from the permanent magnets field?

  • Jan. 23, 2021, 7:41 a.m.

    Yes and no.

    Magnetic fields acting on charges are not like gravitational fields acting on mass; charges do not gain energy by falling into them. Energy is conserved in a static magnetic field; it will change the direction of a charge's motion, namely bending it into a circle, but never it's speed, AKA its kinetic energy. A changing magnetic field can add energy to a system, which is how induction motors and generators work, but not a static one.

    "Yes" you need to add kinetic energy to the ions to get them to "orbit" at the outer radius of the cyclotron, but 1) that's energy we wanted to pump into the ions anyway, and 2) once they reach the edge of the magnetic field, that field starts decreasing in strength, which increases the radius of the ion's orbit, which takes it further from the magnetic field, which decreases the magnetic field strength again, etc, until it completely exits the field, specifically tangent to a cylindrical one. Of course, we want the ions to come out along one specific tangent line in a cyclotron, so we can either add a discontinuity to the magnetic field at the exit point, by either changing the shape of the magnets, adding a smaller permanent magnet, or by using an electromagnet to weaken the field at that point on command; the latter will take some energy but not a lot.

    So the short answer is "no"; it doesn't take energy for a charged particle to enter or leave a static magnetic field.

    The centrifugal force calculations are in my spreadsheet under "B-Field Requirements".

    Because the ions are being contained by a magnetic field and not the engine chamber, the engine itself is not actually going to be under any mechanical stress; in fact you could get away with not having an engine casing at all in the vacuum of space; on earth it's just to keep the air out. At worst, the ions in the vacuum chamber may act like the coil of an electromagnet and pull on the permanent magnets outside the chamber, but that's easy to fix by using some kind of support struts under compression.

    Magnetic fields are kind of weird, and the only reason that the Lorentz force works at all is because charges will do almost anything to avoid making a changing magnetic field; it costs energy, so they'd rather move out of the way (it's much more complicated than that, but that's the simple explanation). So although I've heard of plasmas generating their own self-interacting magnetic fields, I don't think ions can do that; the Coulomb force should push them away from each other rather than allowing them to orbit each other. You may be thinking of electron spin pairs in atoms, which is a completely different, though equally fascinating phenomenon, related to chemical bonding and superconductivity.

    I think you've hit the nail on the head though with your "recirculating Ion drive" analogy, @MultiTool; that's exactly how this engine is supposed to work. However, that's also what bothers me most about it:

    Your basic ion drive is a couple of screens with different charges on them; if any of the ions hit that screen, they are going to 1) erode it, and 2) neutralize some of their charge, requiring you to pump energy into it to maintain a particular accelerating voltage. Any ions that make it through the grates will exit faster than they entered, and that's my big problem: where is the energy coming from? They pick up speed from the electric field between the grates, but if they don't touch the grates, they don't change the voltage on them, so no energy drain on the field; they could recirculate almost indefinitely and gain more energy over repeated passes than was put into the field to begin with! They don't get energy from the magnetic field, because magnetic fields conserve energy. It takes some energy to separate the electrons from their nuclei in the first place, but nowhere near as much as the ions gain by the time they leave the engine (I did the math to check). You could argue that momentum is conserved when the engine and propellant accelerate in different directions, but you could say the same of any rocket, and the rest of them still require some kind of energy to accelerate their propellant in the first place. So where does the energy come from?

    I really don't like that this engine appears to violate the Law of Conservation of Energy, but at the same time I'm having a really hard time finding a physical mechanism, or particle/field interaction, to explain how it wouldn't. Technically an ion drive has the same conceptual problem, so it's not like I can say it's only possible on paper! I've actually been trying to build a small prototype cyclotron over the last couple of years to test the theory, but I'd also like your insights on this problem; perhaps there's a simple explanation that I'm too stupid or too close to the project to see.