'Impossible' EmDrive Space Thruster May Really Be Impossible | Hacker News Reader

'Impossible' EmDrive Space Thruster May Really Be Impossible

space.com

45 pointsjackfoxy86 comments

Disappointing, but not really surprising.

Yes the actual laws of physics can't be broken. We don't know those. We know an approximation of them based on centuries of careful observation and analysis. That approximation is incredibly good, too. It lets us make microchips with 7nm features, synthesize high atomic weight elements, and build spacecraft that have traversed the solar system. Even so, there are things we know we don't know and there are bound to be things we don't know we don't know.

It would be nice if the EmDrive turned out to be a ground breaking demonstration of a hole in our model. We could probably squeeze a century of amazing science out of it. Not to mention the potential practical applications.

But groundbreaking new discoveries that shake the foundations of a developed scientific theory are incredibly rare. So, as awesome as it could be if it were right, the odds were always against it.

It is also much more likely for such a discovery to appear at the edge of our knowledge where our understanding of the rules is at least a bit fuzzy (like near black holes and whatnot). Something like EmDrive, which is in a regime that has been exhaustively tested, would be like discovering that Bin Laden was hiding in the Oval Office all along.

> It is also much more likely for such a discovery to appear at the edge of our knowledge

This instantly reminds me of the Casimir Effect. Yes, it's way more likely for problems to appear at the edges of our knowledge. But discoveries within the noise of experiments people run every day are not unheard of.

I would also expect any such discovery coming from public science made by somebody wondering why their instruments are off, not by for-profit research looking for patents. But that's not unheard of either.

> discoveries within the noise of experiments people run every day are not unheard of.

That's also at the edge of the parameter space we're familiar with.

This one is on the edge of low noise measurement of torque. In fact, the entire exercise pushed the edge way lower than it was.

Which is really hard to do, because there aren't any corners in it.

You need pull a Comey and hide in the drapes.

EDIT: Apologies for the incorrect wording, I want to point out that you should stand AMONG the drapes, not IN the drapes.

And then you found out about Operation Cyclone and your allusion gets some cracks!

And the reason you didn't notice him was you needed to stand on your head and close one eye while chanting "abracadabra" before he could be seen.

Yes the actual laws of physics can't be broken

Conservation laws should be considered more as powerful analysis tools than as foundations of reality. So more Newtonian gravity than Relativity. In fact, you can regard Newtonian orbital mechanics as elaborations on conservation of energy and momentum. Conservation laws themselves have an underlying principle:

https://www.youtube.com/watch?v=04ERSb06dOg

Another analogy: Conservation laws are like Euclidian geometry. On the scale of building a cabinet or an office building, Euclidian geometry rules! On intergalactic scales, it simply fails. So too, with the conservation of energy. If someone tries to sell you a perpetual motion or over unity machine, it is pretty rock solid. But if you're going to try to make some kind of cosmological argument using conservation of energy, that's just plain facepalm territory there. (Some HNer actually reported me for using <facepalm> in response to conservation of energy in a cosmological argument.)

> Conservation laws themselves have an underlying principle:

(before checking the link - "I bet it's Noether's theorem")

(checks link - yep)

I've always thought Noether's formulation and the conservation laws are more like equivalent shapes of the same features of reality.

Except that any particular conservation law might fly out the window, but so long as there was some sort of symmetry, Noether's theorem would still be there.

I've been curiously watching this story although, like many other folks, I realized that the chances were greater than 99% that it wouldn't work out.

What saddened me is the glee that some scientists and students took with dismantling all of these home experimenters. It was shameful to watch. What could have been a great opportunity for many to learn more about science and experimentation turned into a bunch of jackasses mocking a bunch of interested laymen. Not exactly the public image you want for your field of study.

Agreed. Science is hard, especially without funding...but amateurs do contribute to science. This is most noted in astronomy.

When I was 17, I took independent math study at my high school. (I had ran out of math classes to take by doubling up on them) The math department lead mentored me.

One day I asked "Why is 1/0 undefined? You guys said you couldn't take the square root of a negative number, but when push came to shove, you just invented new stuff. Why not do that with 1/0?"

He did not laugh at me or mock me. He looked earnestly at me and said "Why don't you try doing that. See what happens."

I spent the next week or two trying to set up another number system. When I ran into trouble, I did a bunch of research at the library.

When I came back the following week or two? I was able to give him a mini lecture on why it didn't work.

Curiosity, humility, ignorance, and passion is the beating heart of science. The more we make it into some modern priesthood, pro football team, or rock band, the more the general public doesn't want anything to do with it.

https://en.wikipedia.org/wiki/Riemann_sphere

Yep. That's where I ended up. What a great way to get an introduction to a really neat concept.

That's roughly similar to how one of my maths teachers introduced the idea that 0.999... = 1. Introduced x = 0.333..., 10x = 3.333..., 10x-x = 9x = 3, x = 1/3, and then waited. Eventually one of us made the connection what happens if you set x = 0.999... (or equivalently if you times x by 3 - you get x = 0.999... and x = 1).

Of course it is. Nearly every physicist knew this was quackery from day one. It should of course be tested properly, but nobody should be surprised.

Such a drive would necessarily be a perpetual motion machine. Perpetual motion machines are a dime a dozen and we don't test all of them rigorously because approximately none of them will ever work.

https://arxiv.org/abs/1506.00494

No, it wouldn't.

I thought the arxiv paper was fairly convincing. What's wrong with it?

Per the arxiv paper "...Thus, at some point, the kinetic energy of the device-driven mass exceeds the energy input, and if this energy is collected via decelerating the mass (via regenerative electromagnetic braking, for example), then there would be a net gain in energy."

So, first of all, the author of the paper certainly acknowledges the simple priniciples that make EM drive propulsion function. The author is simply arguing how feasible the technology would be for achieving any significant acceleration.

He doesn't claim that EM drive is in principle a perpetual motion machine, rather that 'ar some point' [during acceleration] the energy input to achieve any signiificant acceleration to make the technology feasible would require output energy greater than the possible energy input aboard an engine of a given mass (a statement which is certainly dependant on the efficiency of the 'power conversion' of the system, which can be improved with, for example, more efficient conductors/resonator cavities, etc.).

The paper is fine IMO; it's the comment that cited the paper that is inacurrate.

To quote from the abstract: "any device with a thrust-to-power ratio greater than the photon rocket would be able to operate as a perpetual motion machine of the first kind, and thus should be excluded by the First Law of Thermodynamics" seems pretty categorical to me.

The problem is, when you look at an object with a given velocity v, it has a kinetic energy proportional to v^2. To accelerate it from 0 to v, you must have put in all that kinetic energy (ie, v^2 energy). But reactionless drives give you velocity directly in proportion to energy. At some point, the v^2 and v lines will cross, and your object will have more kinetic energy then you put into it in the first place.

For every unit of energy you pump into the drive, the velocity goes up by a unit, but since the kinetic energy of the object is proportional to v^2, at some point the kinetic energy you get out will be more than the electric energy you put in. The problem goes away with photon drives because that crossing point turns out to be at the speed of light.

This is the argument the paper is making: If you can generate velocity directly from energy, you will eventually end up generating more energy than you started with.

you have to be careful, many experimentally verified observations could have been called quackery, too (an electric current has an effect on a magnet only if it's out of the plane of the magnet? QUACKERY!) If you go to biology and biochemistry, there's plenty of things that still have zero (or very tenuous) basis in underlying theory that are observationally verified (lots of sweet-tasting compounds, basically all anaesthetics, tylenol, e.g.)

Actually no you don't have to, if the claim breaks some fundamental property like conservation of energy or momentum. Tremendous skepticism is most warranted.

Atomic fission breaks the 'fundamental property' of 'conservation of energy', hence https://en.wikipedia.org/wiki/Nuclear_binding_energy#Mass_de... and E = mc^2

The energy released was put into the atom when it was created through fusion.

Of course, but acknowledging that required that the supposedly immutable "laws of physics" be rewritten.

Except it happened the other way around: E=mc^2 was 1905, people only had the means to test for that mass difference after the discovery of the neutron in 1932.

They hypothesised earlier than that, of course, but the experiments followed updated laws nearly 30 years after the new laws had become necessary because of lots of previous experiments showing Newtonian physics was wrong.

And sure, GR will be replaced with something else (it predicts but cannot fully describe black hole interiors), but…

…but expecting that upgrade, whatever it might be, to have the ability to defy known conservation laws in a measurable way in a device made of quarks and electrons, is akin to thinking one possible refinement to the shape of the Earth (curved, sphere, oblate spheroid, …) is the existence of a 29 mile long tunnel connecting Hawaii to Moscow.

Any new physics will almost certainly involve things we can only just build at this point, not stuff like the EM drive which we could’ve built by accident from WW2 onwards.

That makes no sense to me. Which laws rewritten how?

It used to be called the "Law of Conservation of Mass". Now it's called the "Law of Conservation of Mass-Energy".

The underlying principles of the EM druve are very straight forward, easy to understand, and don't violate any laws of physics. The momentum and energy of photons emitted from the antenna (or any antenna) do not exert a reverse force on the antenna or EM drive. All that's occuring is orbial ANGULAR momentum/energy are transferring from the excited atoms in the antenna into generated photons. The exited atom lose energy/momentum as they release photons, the photons ofcourse gain energy/momentum and ultimately transfer it again to a surface on the drive.

Photons are not classical and do not "launch" themselves of of an emission surface. (see more detailed response below)

No mystery, no quackery, just basic physics.

What's your definition of quackery? I have always thought of it as a claim in the absence of experimental evidence.

the em drive had experimental evidence, albeit very tenuous/questionable. Large amounts of skepticism were warranted, but IMO the real world rightly demanded further investigation because that decision should be based on "expected yield" - think "how likely do we think it's real" * "how awesome would it be if it's true".

"I have developed a device that creates a pocket universe with TREE(3) conscious beings living productive, happy lives, and we can go talk to them!" This is very unlikely, but is it so unlikely that multiplying the awesomeness (which is beyond astronomical) that the expected yield isn't also too large not to take it seriously? Or do you have to just round down to zero likelihood at some point.

Just because one cannot explain a phenomenon using known physics doesn't mean they should outright dismiss it. A good many discoveries started with, "Hmmm, that's odd...".

And, science mistakes are good lessons, such as "why didn't we think to check that?"

I think you are not doing credit to the many physicists and science communicators who thoroughly explained why this was quackery. And in a world with quacks - people who have deceived themselves or are deliberately deceiving others - I think it is defensible for most of us to ignore people who claim to have violated the conservation of momentum with scant evidence.

well, I should either trust you, and your idea that we shouldn't give a second thought to crazy ideas like this...

or I should trust NASA, who believed there was enough interesting things to learn there that they experimented with it on their own.

It's close, but NASA wins this round.

For the 100,000,000th time, the Eagleworks thing isn't a NASA operation, except in as much as NASA owns the property it's on.

Is that correct? Brief research indicates that it is not.

Don't confuse "NASA" and its agenda with small groups inside of NASA with open mandates and the ability to spend time on side projects.

If it had been thoroughly explained then there wouldn't have been anomalous experimental results. I agree that it's rather likely that we're seeing experimental artefacts rather than new unexplored areas of physical law, but we must scientifically explain the results or else be very cautious, in general, when declaring something to be impossible. As the joke goes, that is well known to be the surest way in science to eventually be proven wrong. Even if it turns out these pseudoscientific dismissals were reaching the correct conclusion; well, as they say, even a stopped clock is right twice a day.

I'm not claiming impossible. I'm claiming so unlikely that it's not worth our attention. The problem is that we live in a world of people trying to fool us. People trying to actively undermine the scientific process. There is a long history of charlatans claiming free energy; it's not feasible for us to entertain all of their claims.

It find it super funny how it was a cable related issue

just like this famous one https://en.wikipedia.org/wiki/Faster-than-light_neutrino_ano...

Re: you are not doing credit to the many physicists and science communicators who thoroughly explained why this was quackery

To my knowledge, general intent to deceive was NEVER proven. Or are you using a different definition of "quackery"?

Yes, it appeared to violate known physical laws, but it was still possible that A) physicists missed a curious aspect of current models, or B) the existing models of physics are wrong or need tuning. It happens. Not common, but it happens.

There are plenty of quacks who fully believe in their own nonsense.

But other people could reproduce the results, at least to some degree. If the scale had said 0.00000000 for the second tester, it would have died as a story.

> Just because one cannot explain a phenomenon using known physics doesn't mean they should outright dismiss it.

I don’t think most people appreciate how solid our models of well-trodden physics is.

Any new engine is going to be more exotic than microwaves in a chamber.

Why is it impossible? Is it the amount of thrust that's generated that's impossible, or the mere concept of generating thrust from electromagnetic waves?

The reason that I'm asking is because light has momentum. If I shine a beam of light out the back of my spacecraft, then I will accelerate in the opposite direction. Not a lot, just a little.

Was EmDrive impossible because it was claimed to generate way more thrust than could be explained from the energy levels involved? (Edit: It looks like an additional reason it's impossible is because the EmDrive was claimed not to emit radiation, and rather keep it bouncing around in the cavity. Thanks for the replies and explanations!)

Imagine that I attach a 1 megawatt laser in the 630 nm spectrum to my spacecraft. From some rough calculations based on (1), it looks like the laser will generate about 0.003 newtons of thrust (equal to the weight of 0.2 pounds on Earth), which is very little thrust, and impractical for accelerating a spacecraft over regular time scales.

(1) Momentum of a laser beam: http://umdberg.pbworks.com/w/page/50455623/Momentum%20of%20a...

Any reactionless drive better than a photon drive is probably a perpetual motion machine.

https://arxiv.org/abs/1506.00494

Again, this is incorrect interpretation of the cited paper.

The EmDrive claimed to violate conservation of momentum. To extend your analogy, instead of shining the light out the back of your spacecraft, you shined it inside the spacecraft at the back wall. It bounced around and came out as net positive thrust. Hence the extreme skepticism.

Ok, if you shine it on the back wall one might expect the radiation pressure balances out and there is zero net force.

What happens if you used a waveguide to turn it around 180 degrees to the front wall? Photons have no mass, so turning it in a waveguide would result in zero force on the waveguide right?

Photons have momentum and energy.

Right, but if you bend their trajectory, does the device doing the bending experience any net force?

Edit: just learned light cannot be bent by electric or magnetic fields, I must have been thinking of radiation particles.

You can absorb light and emit new photons, but the only way to bend their trajectory is bending spacetime with gravity. Needless to say, the EMDrive doesn’t contain a black hole.

It will exert a force. I don’t know how waveguides work, but momentum is conserved (and photons have it) so it must exert a force.

What you're talking about is a photon drive, which still obeys conservation of momentum. The EMDrive was (claimed to be) reactionless, and a genuine reactionless drive would shatter physics at the foundations.

In addition to the other replies, a working EMDrive would imply that the results of physics calculations would depend on your choice of coordinate system, because of https://en.wikipedia.org/wiki/Noether%27s_theorem, which would be absurd.

As I understand it yes. Momentum is always meant to be conserved. Light does have momentum, but since energy has mass the loss of energy to procuce the light allows it to work out. It's also a very small amount of thrust even relative to the EM drive.

It's the amount of supposed thrust in addition to the violation of conservation of momentum, many orders of magnitude above pure electromagnetic propulsion. Obviously, physicists know that photons carry momentum.

> Of course it is. Nearly every physicist knew this was quackery from day one.

Source of disproof?

From two years of atomic physics study at my university, I don't see any reason you can't generate thrust using EM emission from an anetenna attached to the object being propelled. No laws are violated.

First of all, before getting to conservation laws, the author's statement that an EM drive requires no fuel, right after stating the engine is powered by microwaves (generated by some stored energy source, a.k.a. fuel) is nonsensical.

So, about those conservation laws. EM drives absolutely fully conserve energy and momentum in the system. Anyone who studied physics at the university level should have a firm understanding that photons do not "launch" themselves off of the surface from which they are emitted, like a rocket from a platform. Instead, the energy and momentum change forms as they transfer from the source (e.g. an atom) to the photom (e.g. a microwave).

In atomic orbital decay (say spontaneous emission from an exciton), emitted photons recieve their energy and momentum from the atom's decaying orbital energy and angular momentum. That's it. See here (1) for a deeper breakdown of what's going on, but if you're generally trying to figure out where the energy and momentum are coming from, at the high level that's all there is to it - no quackery here.

If you need a mental image, just try to imagine throwing (imparting a force on) a photon emitted from a light bulb in your hand. You can't do it. A photon travels at the speed of light in a vacuum the instant it is created at the bulb's filament. Because you're hand and the lightbulb (both made of matter and traveling in the same direction as the photon) cannot travel at the speed of light, the photon will always be forward of your hand/light bulb and no energy can be transferred from your hand/light bulb to the photon, and vice versa. There is no "recoil" felt by your hand. That's exactly the same for the EM drive's antenna (or any antenna or emitter for that matter). However, that same photon can certainly interact with matter and transfer it's energy and 'linear' (angular) momentum to an incident piece of matter, for a kinetic reaction.

So, again, I don't see the problem.

1. http://www.eng.fsu.edu/~dommelen/quantum/style_a/consem.html

I found this paper pretty convincing. Any "reactionless" drive more efficient than a photon drive is probably a perpetual motion machine: https://arxiv.org/abs/1506.00494

The claimed thrust of the EMDrive was incompatible with photon thrust- it was much too powerful.

Again, this paper acknowledges the operating principles of an EM drive, but simply argues that the efficiency isn't sufficient to achieve a significant acceleration. "Significance" is relative to the application and efficiency can certainly be improved. For example (and I'm just spit-balling here), using expensive superconducting materials or perhaps a metamaterial could drastically improve the q-factor (1) of the engine.

1. https://en.m.wikipedia.org/wiki/Q_factor

The paper argues that any reactionless drive that produces more force per newton than a perfect photon drive can be turned into a perpetual motion machine, regardless of how it's built:

"Since applying a constant force results in a constant acceleration, the kinetic energy of a mass driven by such a device increases quadratically with time, while the energy input increases only linearly with time. Thus, at some point, the kinetic energy of the device-driven mass exceeds the energy input, and if this energy is collected via decelerating the mass (via regenerative electromagnetic braking, for example), then there would be a net gain in energy."

With a photon drive, the break-even point is at the speed of light (which you can never actually achieve), but with any drive that's more power-efficient, the break-even point will be less, and you will be able to extract limitless free energy by accelerating the drive past that point.

The last time I tried to tell the story behind the academic community's attempts to stop this in its tracks I was downvoted. Rest assured, academia tried its best.

Pity, I really hoped we hacked the rounding error of the Universe :-/

Hmm, could we use interactions with the Earth's magnetic field for thrust when in LEO? With the Sun's magnetic field? It would be slow going.

Yes, sort of, but that’s mainly useful for torque rather than thrust as thrust relies on (if I’m getting this right) field gradients rather than absolute field strength. It’s not nothing, but it might as well be.

I back-of-enveloped a launch system with a very big magnet on the ground, supplying a much steeper gradient to push against, and it was still crazy expensive.

(I wonder if it would be useful for skyhooks?)

this is suggested frequently in these threads. it is already used in the very limited application where it is useful.

https://en.wikipedia.org/wiki/Magnetorquer

"It is also impossible to control attitude in all three axes even if the full three coils are used, because the torque can be generated only perpendicular to the Earth's magnetic field vector."

the non-earth orbiting mission i worked on didn't have any; it used little thrusters to desaturate its reaction wheels.

Yes there are various forms of electromagnetic drives that use the earths magnetic field for small tasks like rotation and (maybe) station keeping. There are a few types, some which use propellant, and some that do not. More info here: https://en.m.wikipedia.org/wiki/Electrically_powered_spacecr...

Yes, I remember reading about people using this to deorbit satellites... though I can't find the link right now.

Yes, satellites already do this.

It would have been a perpetual motion machine of the first kind (you could get an unlimited amount of energy out of it): https://arxiv.org/abs/1506.00494

The evidence for it was no better than the evidence for any other perpetual motion machine, there was nothing that made it special other than not claiming to be a perpetual motion machine (despite being one).

I wanted to believe it could be true, but that's the beauty of science.

Test. Test and re-test. If the conclusion isn't want you want it to be, refine the accuracy of your test and increase sample size. If you test is already at maximum resolution, then tough tacos buddy.

Re: "I wanted to believe it could be true"

Indeed. Star Trek dreams shatnered, I mean shattered. Sorry, no visiting green Orion space-babes. Live long and cuss.

And, wouldn't somebody typically check electrical interaction with Earth's magnetic field? It's a factor already known about.

The paper discussed on HN:

https://news.ycombinator.com/item?id=17096175

The measured thrust was evidently interactions of electrical cabling with the Earth's magnetic field as per the new study.

Wasn’t that just a theory - has it been conclusively proofed it was the cabling?

Unsurprising, but still unfortunate.

Yes, the laws of physics are rarely to be messed with. Doesn't mean we should stop trying, as long as we stay grounded.

We still can't bridge classical mechanics and quantum mechanics so obviously we are missing at least one huge piece of the puzzle.

We can approach classical behavior from the quantum mechanical formulism via the application of the Ehrenfest theorem [1] in limiting situations (e.g. as per the article, cases where "the wave function is highly concentrated around a point").

I suspect you are instead referring to unification of general relativity with quantum mechanics, which has indeed not yet been accomplished. (However, it's worth noting that quantum mechanics has been found to be consistent with special relativity, and the union of the two yields much of modern particle physics).

[1] https://en.wikipedia.org/wiki/Ehrenfest_theorem

What problems with bridging them do you have in mind? I thought this was entirely solved.

Einstein theory of gravity is still considered classic (since it's not quantified).

You mean general relativity?

Sometimes the laws of physics change.

Reply on news.ycombinator.com