In my setting, magical civilizations have physical aversion to metals and moving metals have nasty distruptive effect to magic, so these civilizations development went completely different way from ours.

Eventually, after centuries of primitive protomagic, they discovered process to transform magic crystals(magic fuel, non-quark based matter) to magic cores, which could be endowed with spells/concepts(that still require magic crystals to run).

One of these are concepts of repelling or attracting something. This became a massive game changer, as it allowed to build them hovering ships, ships capable of flight up to hundreds of metres above water or ground (repelling solids and liquids). These ships are built similarly to normal ship (because they consume fuel to fly, they are usually in water for docking and can be sent by waterway to go cheap on fuel).

At first, they used side sails, sails to extend from sides of the ship, but later, they came up with idea of using air attraction core to propel the ship. My question is, how should such engine work? I myself came up with a rather simple but a crude design, but I feel that this design leaves a lot to be desired. I am not an engineer, and my few experiences are not even enough for me to call myself a dabbler.

So my question is: How would an engine that would be designed with use of magic of attraction (or repulsion) work? And is this kind of engine even worth it?

To set the technical restrictions: The cores require at least 0.25 second to turn on and off, so they can't use really short cycles of intake and expulsion. However, strength of their output can be controlled.

The attractor core as I imagined it works in a way where it pulls air particles (N2, O2, CO2) inwards, preventing the particles from escaping the zone, unless their velocity is too high. This in effect would creates a zone of certain volume in which air pressure stabilizes at different level(this in principle immitates planets holding their own atmosphere). The repellant core works on precisely opposite principle.

Attractor thus creates zone where air pressure increases as we approach the core. The best the attractor core can reach is for average density within volume of its effect to about 10 times density of regular atmosphere around it, the best the repellant core can do is 0.1 density of regular atmosphere around it. We should assume that this effect can cover volume of ~0.5 cubic metre.

My own idea was to use attraction cores to attract air - increase pressure in a wooden chamber, then block intake, open the thruster, and turn the attractor core off, which pushes the air into the only possible route of escape - thruster. However, with such technical constraints, I feel that my idea might produce only laughable thrust.

  • $\begingroup$ the majoirity of the periodic table is metals, you are going to have a hard time creating technology without metal. the calcium in your bones is a metal. 2/3rds of the atoms in water are metals. $\endgroup$
    – John
    Oct 30, 2019 at 2:05
  • $\begingroup$ I am currently handwaving it with 'metal has to be in its pure form', although I am aware that is not really good enough. I plan to look a little bit more at it. I wanted to restrict it a bit to only work with transitional metals, but to be honest, it feels completely artificial and I have bad time justifying this artificial decision. That being said, hydrogen is not a metal. As for what remains even after my restriction "metal has to be pure", I plan to leave that as everpresent background interference. $\endgroup$ Oct 30, 2019 at 7:17
  • 1
    $\begingroup$ Uhh... Did John just categorically state that hydrogen was a metal... Like, I'm pretty certain it's genuinely agreed to be relatively ambiguous in terms of its 'metalness'. $\endgroup$ Nov 2, 2019 at 22:37
  • 1
    $\begingroup$ Can you clarify more how the attractor works? It seems to me like it's basically just a compressor that is inside a chamber. $\endgroup$
    – Basher
    Dec 14, 2019 at 2:57
  • 3
    $\begingroup$ Are the ships flying by repulsing the ground? If so, wouldn't it be possible to repulse ground behind the ships for propulsion? $\endgroup$
    – Sebastian
    Dec 14, 2019 at 10:02

8 Answers 8


From what you've described I'm taking these effects to be a symmetrical, spherical volume around the crystal, and the density gradient to be continuous at all points (ie the density increases smoothly from 1 atmosphere at the boundary to some value (permissibly infinite) at the centre, with no sudden jumps). Otherwise you've created a force field for air, which I don't think is what you want.

Given those things, the force produced by these crystals is conservative, which means there is no static assembly of attractors and repulsors which can produce a net change in velocity of an air molecule. You have to cheat somehow by changing the configuration of the field in either space or in time. You've said that it's cumbersome to vary the crystal's strength in time, so let's consider spatial options.

The simplest engine, which uses only attractors, functions similarly to a water wheel. Take an airtight tube of radius $r$ open at both ends, and drag an attractor with effect radius $R = 2r$ down the outside of the tube, in a (loosely) sealed area; then lift the attractor at least a distance $r$ away from the tube before drawing it back solely in the sealed area to complete the cycle. The air inside the 'mechanism' sloshes around a lot, but the net airflow inside the tube is strictly in one direction. An alternating series of attractors and repulsors moving in one direction would drag pockets of high-pressure air with them. I think maximum efficiency would be achieved with a linear tube and crystals moving around on a caterpillar track, but it might be mechanically simpler to construct as a half-circular tube with the crystals fixed to a rotating wheel.

How much thrust might such a device produce? You've said the area of effect of a crystal is $\mathrm{ 0.5\ m^3}$, which is a radius of also about $\mathrm{ R = 0.5\ m}$. If you touch such an attractor to the outside of a tube of radius $r = \mathrm{0.25\ m}$, the volume contained within the intersection of the sphere and the cylinder is given by the spectacular formula *:

$$ V = r^3 \left( \frac{2\pi}{3} + \frac{4}{9}(1 - 4\eta^2)K(2\eta) + \frac{16}{9}(2\eta^2 - 1)E(2\eta) \right)$$

$\eta = \frac{r}{R} = \frac{1}{2}$ in this case, which is fortunate because it means we don't need to worry about $K(2\eta)$ or $E(2\eta)$ (which are Legendre's complete elliptic integrals of the first and second kinds, respectively); the $K(2\eta)$ term disappears and $E(1) = 1$. We finally get:

$$ V = r^3 \left( \frac{2\pi}{3} - \frac{8}{9}E(1) \right) = \frac{6\pi-8}{9}r^3 \approx 1.2 r^3 = \mathrm{0.15\ m^3}$$

You only said that the average density inside the crystal effect area is increased tenfold, which means technically we should choose an expression for the density as a function of position and integrate over that volume, but I hope you'll forgive me if I baulk at that just take it to be a constant. The density of air is about $\mathrm{1.25\ kg\ m^{-3}}$, so each pass of the attractor crystal drags approximately $\mathrm{1.7\ kg}$ of air through the cylinder (over and above 'ambient').

This doesn't sound like a huge amount, and indeed it's not, but it would not be unreasonable for this machine (especially if set up in the wheel configuration) to move the crystals extremely quickly. From a quick google search it seems that for traditional spinning wheels rotation rates of wooden parts of tens to hundreds of revolutions per minute (ie $\mathrm{1\ -\ 10\ Hz}$) are reasonable. A 5m-diameter wheel would have space for 15 attractors around its circumference, and rotating at $\mathrm{3\ Hz}$ it would shift around $\mathrm{76\ kg}$ of air per second at a speed of $\mathrm{47\ m\ s^{-1}}$ for a not-unrespectable thrust of $\mathrm{3.6\ kN}$ (in the same ballpark as the $\mathrm{4.5\ kN}$ from the first jet engine). You couldn't achieve heavier-than-air flight with such an engine, but to move a lighter-than-air vessel kept aloft by magic? Sure.

* Case II.C (equation 9), setting $\eta = \frac{1}{2}$ and then multiplying by $R^3$ at the end.


Well, this makes maglev style trains very easy.

For propulsion, have an open tube with a number of these "repulsion engines" (REs) inside it angled toward one end. That would cause air to be pushed out the "back" end and sucked in the front end. You now have jet propulsion. If you turn on of off some of the REs or can increase of decrease the repulsion effect, you can throttle the jet.

This tube would work in air and water.

You could also make a turbo prop out of this but I wouldn't unless they've had this tech for a very long time.

Note that if the RE is powerful enough, it would make a very good machine gun too.


A jet engine has four components: the intake, the compressor, the combustion chamber, and the turbine. Your magic attraction and repulsion cores are capable of reproducing the first two stages of the engine: sucking air into the engine, and then compressing it to increase its pressure while decreasing its volume.

If they can make magical heating elements, they could reproduce the effects of the ignition chamber, assuming that they don’t want to just use mundane fuel oils and a crystal-lined combustion chamber to do the job.

Then comes the turbine, which turns a portion of the engine’s thrust output into the rotational power needed to run the engine; strictly speaking, your magic engine wouldn’t need this since it’d be entirely solid-state, without rotating components, though maybe if it’s possible to generate magical energy by rotating a wheel you might be able to produce something analogous by building some sort of wooden turbine.

  • $\begingroup$ It's not possible to get magic by rotating a turbine, but perhaps I could use some device to harness the mechanical energy for the system to use to close the intake/thruster on its own. $\endgroup$ Nov 1, 2019 at 4:10

An internal combustion engines works on the principle that the combustion gases, expanding into the cylinder, put in motion the cinematic chain transforming the alternate motion into rotation.

You can use a similar setup but, instead of using a cylinder where you burn some fuel, you just use alternated attraction/repulsion to swing the piston back and forth, and then convert that motion into a rotation.

Not having to deal with high temperatures you can avoid using metals.

  • $\begingroup$ The trouble with this is, AFAIK, there are no such engines that work with solid state components, i.e. that do not require the repulsion/attraction to be switched on and off. As the OP mentioned a fairly low switching rate, while you might be able to build some sort of motor, your RPM's are going to be limited. Maybe you can make up for it in torque, though. You'll probably also want to combine a bunch of these on a camshaft. $\endgroup$
    – Matthew
    Dec 14, 2019 at 19:59
  • $\begingroup$ @Matthew certain naval motors do operate at around 60-70 rpm. That doesn't prevent them putting out enough power. $\endgroup$
    – L.Dutch
    Dec 14, 2019 at 20:14
  • $\begingroup$ ...but how big are they? I'm not saying you can't make a useful engine, just that you can't get a direct analog for the range of ICEs that we have. Powering a huge container ship, or even a train locomotive, is probably plausible. I'm much less confident about replicating something like the engine in a (handheld) gas-powered chainsaw. That doesn't make this useless; far from it! It just means there are some important limitations to keep in mind. $\endgroup$
    – Matthew
    Dec 16, 2019 at 15:40
  • $\begingroup$ @Matthew, rather big: naval engines where an engineer fits standing up into the combustion chamber of single cylinder $\endgroup$
    – L.Dutch
    Dec 16, 2019 at 16:09


Let's look at how rockets work:

Rocket engines work by action and reaction and push rockets forward simply by expelling their exhaust in the opposite direction at high speed. (from wikipedia)

This is almost identical to how your magic system operates. It would likely also be the first type of travel they discovered, because the theory behind it is simple. Put something on wheels, and expel particles as fast as you can in the opposite direction.

For propulsion force = mass * acceleration. So to increase speeds you need to either increase the mass of what your are repulsing, or the speed you are repulsing it at. Note, the heavier you get the more force is required to move you.

Major advancements to this field would be using things like nozzles to increase the repulsion velocity (and thus thrust generated.) It is also likely while some efficiency minded engines would use air and would operate like real world hovercraft, anything built for short bursts of speed would use heavier fuels.

So in answer to your initial question, just increase the speed at which your engines repulse the condensed air until you have the desired thrust.


it works in a way where it pulls air particles (N2, O2, CO2) inwards, preventing the particles from escaping the zone, unless their velocity is too high

You thought of using moving parts to work around the fact that the effect is omnidirectional, I assume. But I think you can do even better.


Let's have a nitrogen-attracting core. it will "pull N2 inward... unless their velocity is too high". So we build a long wooden tube, with two cores, one repelling, one attracting:

====================================== /
   --->         A                    R -->
====================================== \

The repulsor core will expel air from the tube, creating a low pressure area; the attractor core will be unable to attract much air from the tube because it will soon exhaust its content.

Now, air is attracted from the left end of the tube, at the very limit of the attractor effect, and pulled inwards, acquiring speed. Due to conservation of energy, the air will be able to move as far to the right of A as it fell in from the left of A; this must place it inside the effect area of R with enough speed to reach and overshoot R (therefore, the power of R must be less of that of A).

The net effect will be a transformation of magical potential energy into air kinetic energy; wind will start blowing through the pipe.

Close the left inlet to stop the flow, or attach the pipe to a vessel to provide thrust.

I was trying to explain why this cannot work in the real world (which was intuitive enough) and why, and what would be needed to make it work. @Stephen has put his finger on a far sharper and simpler explanation than mine: the magic field must be nonconservative, but more than that, it has to break symmetry. The simplest way to achieve this is to make the field non-additive - for example, in any given point of space, the total effect is not given by the sum of the effects of magic cores within range, but by the whole effect of the locally stronger field alone, much like political frontiers work (when you're approaching a State limit, you're completely subject to the laws of the State you're in; as soon as you step through, you're again completely subject to the laws of the State you just entered).

Another way of making this work is by supposing that metal can make the force nonconservatively disappear (aka magic); then placing a metal ring between A and R, close enough to A but not enough to interfere with its powers, would allow air to escape towards R with most of its kinetic energy intact, moving towards R on inertia.

  • 2
    $\begingroup$ I don't think this works as a machine, in either the steady state case (which will be static with a pressure gradient along the tube but with no net flux) or with an initial vacuum and feeding molecules in one at a time from the left. R must indeed be positioned far enough inside the area of A that the difference in the distance travelled to the left and right of A imparts enough kinetic energy to overcome the repulsive force experienced to the left of R. But since A is symmetric, then the particle still has to cover the same distance to the right of A after passing R, which will rob... $\endgroup$
    – Stephen
    Dec 14, 2019 at 17:38
  • $\begingroup$ ... it of the remaining kinetic energy. The setup can be decomposed into two separate fields, both of which are conservative. You need to introduce a non-conservative force (like an anti-friction) to achieve a motion machine. $\endgroup$
    – Stephen
    Dec 14, 2019 at 17:40
  • $\begingroup$ @Stephen that's it! I was trying to break conservation by having the field fall faster, and it didn't seem enough to me - you've put your finger on it. The forces must not be additive and symmetrical. $\endgroup$
    – LSerni
    Dec 15, 2019 at 12:31

I personally prefer a more simple method I shall call the Anti Gravity Engine

While its pretty hand wavy, it gives you a lot of flexibility in terms of use and has little impact when you think about the traditional movement of air particles and airflow when flying and generating lift

The issue with a rocket style engine is the constant stream of high speed particles you need to blast out. When your ships lift off, they are going to be expelling a ton of air particles into what ever is below them and this can be very destructive. If you ever watch a helicopter land you will notice the huge amounts of wind and noise generated. A commercial airliner can blow away cars and buses. A rocket basically makes the surrounding area uninhabitable...

A Anti Gravity Engine doesn't have this issues. It interacts directly with gravity, which means your surrounding areas remain nice and calm. No need for earmuffs, protected landing zones and the best feature? All your ships will be able to hover in place without blowing out everyone's eardrums. (You can also have ships fly close to each other. Directly above and below, in front and behind without burning them to a crisp with your high energy exhaust fumes).

The second advantage is attraction. Not only can you repel gravity, you can attract it. This means that your airships can theoretically double up as submarines as well. This can even be applied to other ships, allowing you to form a natural train of airships and submarines with an extra bit a magical research.

The third advantage... your ships can be any shape you want. A traditional flying vehicle is limited in design, because it needs to be able to move and generate lift. The Anti Gravity Engine produces a steady supply of lift without requiring wind or speed. A steady supply of mana is all you need, which means you can actually have and use sails to move horizontally in the air. You can have monstrous floating cities and you could even run a space program out of your rickety dinghy assuming your Anti Gravity Engine was powerful enough and had enough supplies to get you that high.


There's an even better way of using these kinds of magical attractions / repelling engines. And that is the legendary Zero Mass Drive. Theoretically, using negative matter, you could have a drive which has a net mass of zero, and propel itself. Hence the name, Zero Mass Drive (technically, it's called a Reactionless Drive, whatever, Zero Mass Drive is a cooler name).

Now, you specify that the crystals are non-quark based matter, but you haven't done the same for cores, which means that they are quark-based - and thus real matter - and thus you can build cores that can attract and repel the material that cores are made of. So what you do is simple. You build a core which attracts cores, and a core which repels core. Then you line up the cores. The first core repels, the second attracts. And now is when the drive takes effect. The second is attracted to the first, and the first is repelled, sending both cores forward along the line. Then, all you do is mount the cores (when they aren't on, of course) to your ship, then simply activate them.

  • $\begingroup$ I am really sorry but I have to correct you. I might not be an engineer, but I'm a sci-fi enthusiast. Key point of zero mass drive is, it requires one of cores have negative mass. We're not talking about gravitational mass, but negative innertial mass. Object with negative innertial mass is weird - if you push it, it starts accelerating towards you, if you pull it, it starts moving away from you. Cores and crystals, while both non quark matter do have both positive innertial mass. So sadly, that's not possible. $\endgroup$ Oct 30, 2019 at 18:14
  • $\begingroup$ I'm not saying you can do a zero mass drive. I'm saying you can do a facsimile of a zero mass drive. It won't actually be zero mass, but the drive itself, as long as you have the twin cores in place, will just accelerate forward once you turn them on. $\endgroup$
    – Halfthawed
    Oct 30, 2019 at 18:25
  • $\begingroup$ I don't think this would work unless you violate Netwon's Third Law. The attraction/repulsion of each core would apply equally to each core and would cancel out. $\endgroup$
    – Matthew
    Dec 14, 2019 at 0:21

You must log in to answer this question.

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