Update - Fuelless microwave space drive now invalidated by 3 separate studies. Oh well.

It sounds like it would only be useful in microgravity. The power output required to generate enough thrust to overcome Earth's gravity would be insane. (terawatts? yottawatts?)

The whole thing about the EM drive, if it actually does work, is generating a tiny amount of thrust consistently, for a nearly unlimited amount of time. Constant acceleration, even if it's minuscule, can result in extremely high velocities, far higher than can be achieved with pretty much any chemical rocket.

Also, by increasing the delta-V budget by so very much, it would allow ships to just go to other planets, rather than needing to use Hohmann transfers to minimize fuel usage. So not only could they reach higher velocities, they could avoid spending years in long transfer orbits.

While the current design has a pretty low (terrestrially) thrust/power ratio, later implementations of the device can be much, much better. The Eagleworks device is copper with unknown thickness. Devices which minimize leakage of the em-field will theoretically have much higher thrust/power ratio when not accelerating. According to the designer a superconducting device would enough for terrestrial applications. As in, flying cars or trains.

The violation of conservation of energy critics typically seem to be applying math to the object that doesn't fit its theoretical mechanics. But it is true that it's not as intuitive. Take a nice liquid fuel rocket that is pressurized and whatnot so the fuel comes out at a really constant rate and is burned evenly and constantly at that rate. That rocket is producing a constant amount of thrust. If it's mounted to a stand its making that much thrust, if it's accelerating off the pad its making that much thrust, if it's already flying really fast its making that much thrust. (Yes we are completely ignoring that different atmospheric pressures effect rocket thrust, shhh).

A ship with that rocket is thus always accelerating at a pretty constant amount (I don't suppose you'll let me ignore all the fuel mass leaving too, right?) But constant acceleration means that the kinetic energy is increasing faster and faster and faster. (Because it's based on v^2) However the math all works out because that faster exhaust is losing more kinetic energy when the rocket is faster, so conservation of energy is satisfied.

Take out that exhaust, however, and you'll have problems. If your ship is gaining more and more energy out of this device, without that extra coming from the reactant mass side of things, then what is happening?

Lets take a 100 N device with an emdrive that gets 1 N of thrust and somehow can provide the 1000 kW of power to do so. That's an acceleration of .01 m/s^2.

Lets power this thing for 10 minutes.
Starting speed: 0 m/s
Final Velocity: 6 m/s

Starting Kinetic Energy: 0
Final Kinetic Energy: 1800 J
Distance traveled: 1,800 meters
Force * Distance = Energy engine applied to ship= 1 * 1800 = 1800 J
Time * Power = Energy engine used = 600 * 1000 kW = 600 MW

But wait, from Earth's point of view this device in LEO was already going 7.8 km/s!
Starting speed: 7800 m/s
Final Velocity: 7806 m/s

Starting Kinetic Energy: 3.042 GJ
Final Kinetic Energy: 3.0467 GJ
Difference: 4.7 MJ
Distance traveled: 4,681,800 meters
Energy engine applied to ship= 4.7 MJ
Energy engine used = 600 MJ

But wait, from the Sun's point of view this device was already going 40 km/s!
Starting speed: 40,000 m/s
Final Velocity: 40,006 m/s

Starting Kinetic Energy: 80 GJ
Final Kinetic Energy: 80.024 GJ
Difference: 24 MJ
Energy engine used = 600 MJ

So as you can see, the apparent efficiency of this sort of engine keeps creeping up the faster you are going, which is... weird. This specific example doesn't actually have you creating energy from nowhere until much later, but theoretically better engines would hit that time sooner. There are other things to take into account, for example this engine loses effectiveness at relativistic speeds, but still, it needs a better explanation than that, unless we're going to through general relativity out the window along with conservation of momentum. And honestly, by the time you're throwing out conservation of momentum is throwing out general relativity any more crazy?

(General relativity is the thing that says that all three of those above interpretations are equally valid and should all mathematically work identically, however if all reference frames aren't created equal, and the ones that lead to energy being created are just not valid reference frames, then that would also solve the problem. And would also be weird.)

Yonder wrote:

...for example this engine loses effectiveness at relativistic speeds...

The fact that we are talking about this seriously makes me very excited.

Great post Yonder.

I really want this thing to get to the point that we can send out a probe to test real performance. After it gets moving I think there will be a lot more believers. I love it when science gets ahead of itself to have this kind of controversy. I am really hoping this all works out. Who knows, we might visit another solar system (via probe) in my lifetime.

Hmm, most engines lose effectiveness at relativistic speeds, thanks to the Lorentz factor.

Quintin_Stone wrote:

Hmm, most engines lose effectiveness at relativistic speeds, thanks to the Lorentz factor.

That's true, but this engine theoretically has a further penalty on top of that, because it works by having photons bounce back and forth changing wavelengths and whatnot, and at high speeds the wavelength changes get lower and lower, on the bounce, lowering thrust. Something like that. However I think that explanation (from the designer) also goes against the idea that any inertial reference frame should be as good as any other.

Right now the least radical physics explanation for this is that this is coupling photons together so they can escape the device (and everything else). However I'm not sure if that explanation really holds water. If it works like that the device is basically a photon rocket/flashlight that happens to shine invisible light. The fact that the light doesn't interact with anything lets you do some interesting things from an engineering standpoint (like put parts of your spaceship behind the engine without slowing the ship down and/or being vaporized), but from a physics standpoint it's basically a flashlight.

The issue is... flashlights don't produce this much thrust. Like, nowhere near that much thrust. Even if a nice LED flashlight isn't 100% efficient it has to be, like, 30% efficient, right? But if you had a 1 kW flashlight you are NOT going to get 1.2 mN of thrust out of it. In order for that explanation to not also run into a "this device is making energy" speed bump the photons being produced would have to be at a laughably high (or low, can never remember) wavelength. If I'm remembering my "photon momentum vs photon energy" relationships right anyways.

Basically, while static testing environments can indeed show that the engine works (if everything is done right) even if it's not as intuitively persuasive as seeing the engine push a cart/spaceship around, the handful of hypotheses on varying places on the crackpot spectrum predict different behaviors while the device is under load accelerating at different rates and speeds. So until we start doing that with many different devices our ability to see which hypotheses are closer to correct is pretty limited.

I suspect it'll be at least a decade before they put one of these in orbit. Looking forward to the results, however.

Here is a pretty good reddit post of the weaknesses with this paper.

The two most valid (IMO) critiques were that their technique for distinguishing from real thrust and thermal expansion effects on their probe seemed to lack rigor. He sees a lot of risk that they stuck on the methodology for how to interpret their graph that gave them a result, not necessarily the right way. We'll see if other people with better qualifications agree with this. This is especially troubling considering that they didn't run a null test, swapping the frustrum for a cylinder and running at the same power level to get the thermal effects alone.

I seem to recall this guy running a null test like this back in the atmosphere, but still, re-running it here does seem like a no-brainer, ESPECIALLY since there is an even easier way to get a null test than this guy states. You can do it with a frustrum really easily, since the theoretical thrust comes from the field resonating perfectly inside the device. They should be able to turn on their microwave generator to the same power level, at basically any other frequency, and get their null test.

The other thing that he critiques that I also noticed/agree with is that there thrust results really don't form as nice of a trend as you would like:

He ran at three different power levels and most of the time his results at two sigmas of uncertainty don't come anywhere near his trend line, that's some pretty noisy data:
IMAGE(http://i.imgur.com/FzgH8Gv.png)

That said, this isn't completely unexpected either. I'm guessing that this device would be expected to have pretty thoroughly different field quality at different power levels, especially for such an early/simple/crude example. It's not exactly a smoking gun, but really that missing null experiment kind of is. (They do have a different type of null thrust experiment, turning on the device normally, but having it point in a direction the probe doesn't measure, but when the actual thrust action of the device is at question here, not having both kinds of null test is a big oversight(?).

missing null experiment kind of is.

Yeah, that's not good news at all. Sigh. Back to 'highly unlikely'.

Hmm, actually it looks like they did to exactly the test I'm proposing: offset the field so it didn't resonate for a null thrust. However apparently the paper reviewers had them remove it.

IMAGE(http://i.imgur.com/PrYS4iP.jpg)
IMAGE(http://i.imgur.com/oTxifDF.jpg)

But they haven't said why the reviewers removed it. It's kind of crucially vital, so if they had it removed it seems like that implies something pretty significantly wrong with the methodology. It's hard to imagine the null thrust result being removed otherwise, but if that was the case it seems like they should have fixed whatever there issue was and reran rather than publishing the paper without that. Maybe budget or scheduling precluded that, but I'm surprised the reviewers and journal accepted it without the null test.

Edit: Oh, on second look I can see exactly why the reviewers removed it, those tests were run in air, not vacuum (I thought I remembered that!). The paper has "In Vacuum" right in the title, so of course the reviewers weren't going to let them use the Atmospheric null tests. Did they just not bother/forget to run the null tests in Vacuum since they had already done that in atmosphere?

Did they just not bother/forget to run the null tests in Vacuum since they had already done that in atmosphere?

Is this tester somehow linked to the inventors? Could they benefit from helping a scam?

Moggy wrote:
Gremlin wrote:

And it won't make it easier to take off from planets (where you need a lot of delta V all at once)

It might. Imagine a circular track that allows a payload to be accelerated to escape velocity, at which point the payload is directed to a ramp and "up, up, and away". Not sure about air resistance, but if the motor is scalable, why not?

I think a railgun/hyperloop type accelerator would be quite more efficient than an EM drive when it comes to use of electricity for propulsion.

Robear wrote:

I suspect it'll be at least a decade before they put one of these in orbit. Looking forward to the results, however.

May be a lot sooner than that.

Correct me if I'm wrong, but if the EM Drive actually works wouldn't that give the power to destroy the Earth to anyone who can put a satellite in orbit? Because all you have to do is send up a satellite with an EM Drive in it, program it to fly around the solar system increasly gaining speed until it's fast enough that smashing into the planet just kills everyone?

We're currently watching every asteroid big enough to do actual damage to Earth, so I'm sure we'll have no problem tracking an EmDrive-powered vessel large enough to survive reentry through the atmosphere.

Assuming that the EmDrive obeys conservation of energy, that's not a short term concern. The EmDrive is (theoretically) far, far superior to things like ion engines for propulsion, but it's still limited by the energy you put in to it.

For example, if you run a small satellite off of a cell phone battery with an EmDrive, fly it around for max speed, and then crash into Earth, it will be going a lot faster than it would have been with an Ion Drive, but at the end of the day it will still hurt left than if your phone had just exploded with all that stored energy.

In the medium term, say 20-30 years from now, we'll start to be getting into the danger zone. That's (IMO) when big enough people will have been convinced by this tech, if it works, enough to start getting things organized to use it for the big stuff. By this I'm envisioning people bundling up Nuclear Reactors of the size we use to run Naval ships. When that happens we'll start getting spacecraft with enough power that they could cause some serious damage.

Even if your Emdrive is only 5% efficient, run a 500 MW reactor for a two week acceleration run straight for Earth and you're talking a 30 TerraJoule, or 7 Megaton explosion, ignoring the gravitational potential energy gained and what happens to all the Uranium at impact time.

Honestly I don't think that's much of a concern though. That sort of spaceship will still be really expensive, even more expensive than the nuclear subs of today. The sorts of things that you could have three layers of various fail-safes and safeguards in.

Unless there is a huge leap in fusion (and I don't mean "fusion plants exist" I mean "we can make little fusion reactors significantly easier and cheaper than we can currently make little naval nuclear reactors") then I don't think that the proliferation of dangerous spaceships will be anything to worry about.

Now, things ramming space stations, moon bases, and other stuff not protected by atmosphere, maybe martian bases too, that may be more of a concern.

Yonder wrote:

Now, things ramming space stations, moon bases, and other stuff not protected by atmosphere, maybe martian bases too, that may be more of a concern.

I'll be more worried about it if we have martian bases. Most settlement-type places on planet surfaces are going to be designed for protection against stray rocks and radiation already.

And, given that we still don't know how this thing works (if it works) that affects our ability to anticipate its characteristics. At least, I don't know enough. I'm having enough trouble picturing how the thing works, and it doesn't help that the internet in general has several competing misleading descriptions.

(It's definitely not a warp drive: I'll know that much. No FTL, unless it violates causality too.)

Here is my analogy for the most popular (and the inventors) interpretation of how the Emdrive works.

So the nozzle of your garden hose narrows down to a smaller point.

IMAGE(http://www.backyardboss.net/wp-content/uploads/2016/05/Heavy-Duty-Brass-Adjustable-Hose-Nozzle-by-Dramm-1.jpg)

It does this because compressing the water down increases the pressure and the speed that the water flows. The water is going faster at the narrow part at the end than it was going at the wide part at the beginning.

Now imagine that you immediately capped the nozzle at both ends. The water would flow down to the thin part, speed up, and hit the cap hard. It would then bounce off, flow back to the wide end, slow down as the nozzle widens, so be going slowed and bounce off the wide gap more gently. Repeat again and again. The difference in how hard the water was hitting each side would make a force towards the narrow end.

Now obviously none of that would actually work with water. It wouldn't even work with air. All of the molecules in the nozzle would very quickly bump into each other and cancel everything out so that the fluid in your capped nozzle wasn't moving at all. Even if that didn't happen the fluid would have a lot of friction with the sides of the nozzle and stop anyways.

However with this nozzle we don't have water or air molecules bouncing back in forth, we are bouncing photons back and forth. They don't interact with each other, so they don't slow each other down by bouncing back and forth past each other, and the energy the EM field loses through interaction/leaking through the cavity is small enough that the resonator attached to the frustum can overcome it and still get useful thrust. That resonator is also needed to keep putting energy into the EM field because it loses energy each time a photon bumps against the narrow part and loses some of its energy accelerating the device (at least if the device was accelerating, if it's just sitting still the bounce is theoretically perfect and the EM field doesn't get weaker, which is why this engine theoretically is much more efficient at producing thrust the less it is accelerating).

I thought photons bounced off of each other inside the center of the Sun and that's why it takes some extraordinarily long amount of time for the photons produced at the center of the sun to make it to the surface and then zip to earth in 8 minutes. Are they just bouncing off of hydrogen and helium nuclei, and not other photons?

I also thought that the interaction of photons hitting other photons is what caused the interference pattern in the double slit experiment. I thought for sure that photons interact with other photons. How else would you get stimulated emission in lasers?

Mixolyde wrote:

I also thought that the interaction of photons hitting other photons is what caused the interference pattern in the double slit experiment.

An interesting tidbit there: they see the interference pattern even when they're firing single photons. They do affect one another, but they also interfere with themselves.

jrralls wrote:

Correct me if I'm wrong, but if the EM Drive actually works wouldn't that give the power to destroy the Earth to anyone who can put a satellite in orbit? Because all you have to do is send up a satellite with an EM Drive in it, program it to fly around the solar system increasly gaining speed until it's fast enough that smashing into the planet just kills everyone?

Sitting at the top of Earth's gravity well is already a hugely advantageous place to be. If someone wants to be destructive, all they have to do is drop large rocks on our heads. That's easier with an EM drive, if they work, but quite doable with chemical rockets.

The single photons don't interact, because they have nothing to interact with, and if you fire just one, there will be on spot only on the receiver. However, if you fire multiple single photons sequentially (so they don't interfere with each other, obviously), the pattern produced by the sequential single strikes will be dictated by the waveform that corresponds to the energy of the photons.

Photons don't interfere with themselves, but they *do* follow the probability distributions dictated by their waveforms. And single waves *can* interfere with themselves when hitting the double slits, just as you'd expect if a water wave hit it. So that's why considering photons as points - quanta - breaks down at quantum levels. They aren't really quanta.

Arg.

Yonder wrote:

However with this nozzle we don't have water or air molecules bouncing back in forth, we are bouncing photons back and forth. They don't interact with each other, so they don't slow each other down by bouncing back and forth past each other, and the energy the EM field loses through interaction/leaking through the cavity is small enough that the resonator attached to the frustum can overcome it and still get useful thrust. That resonator is also needed to keep putting energy into the EM field because it loses energy each time a photon bumps against the narrow part and loses some of its energy accelerating the device (at least if the device was accelerating, if it's just sitting still the bounce is theoretically perfect and the EM field doesn't get weaker, which is why this engine theoretically is much more efficient at producing thrust the less it is accelerating).

Here's where you lost me. You explain Bernoulli's principle (which applies to fluids), then try and apply it to something that isn't fluid (photons).

What's the mechanism by which photons accelerate when moving through a narrowing aperture? What's your analog for a pressure increase? Never mind that, how the f*ck are you getting photons accelerating i.e. moving at a speed that isn't the universal constant of c?

The photons don't accelerate, they decrease (increase?) in wavelength which gives them more momentum, which means they hit harder, which is analogous to the increased velocity of the fluid. Higher energy photons certainly have more momentum than lower, that's rooted in firm scientific principles. My understanding is that the concept of condensing/whatever an em field to change it's frequency is also established science, but I don't have the background to evaluate that myself.

The expectation would be "well sure, according to those simple equations you can get your microwave resonance garden hose to make a force, but in real life these other effects make that simulation meaningless", just like that would be the response to a student that came up to you and said "Hey look this will totally work because it's an IDEAL gas, so the atoms will never interact with each other and disrupt the flow" and you have to explain "well, by never interact with each other we actually mean, 'they probably won't bump into each other for a few centimeters'" which is basically never, unless you are trying to trap a high moving air flow into a nozzle so that you can violate conservation of momentum like a wiseass".

I think what Jonman is alluding to, is that the same physics principle that makes water molecules accelerate in a nozzle is what prevents them from going backwards in the nozzle. You can't have one without the other. If particles/molecules interact, they will push each other faster, but will also block each other the other way. If they don't block each other the other way, then they won't be pushing each other faster.

No, that's not true in either the analogy of fluids or (apparently) in EM mechanics. The flow equations that correctly (and very accurately) predict subsonic fluid accelerating through a shrinking nozzle predict that both for water, but also for "ideal gases" like the atmosphere. The biggest assumption that defines a gas as an "ideal gas" is that it's molecules do not interact with each other in the fluid. It's not true, if memory serves the average collision distance for standard temperature and pressure atmosphere is a few centimeters, but it's true enough for many cases.

And it's not a perfect analogy, like I stated, a closed nozzle, filled with water or air or water, would absolutely not behave like this. It's just the closest I could come up with to a similar actual physical principle that's more intuitive.

MoonDragon wrote:
Moggy wrote:
Gremlin wrote:

And it won't make it easier to take off from planets (where you need a lot of delta V all at once)

It might. Imagine a circular track that allows a payload to be accelerated to escape velocity, at which point the payload is directed to a ramp and "up, up, and away". Not sure about air resistance, but if the motor is scalable, why not?

I think a railgun/hyperloop type accelerator would be quite more efficient than an EM drive when it comes to use of electricity for propulsion.

Robear wrote:

I suspect it'll be at least a decade before they put one of these in orbit. Looking forward to the results, however.

May be a lot sooner than that.

Who knows, maybe in this century humans will put electromagnetic rail accelerators powered by solar in orbit around (say the moon) and use EM driven payloads to super sling shot objects at a massive velocity from the solar system. One thing is clear - there's no way chemical propelled rockets will have any feasible chance of getting to a new solar system within humanity's foreseeable future.

Yonder wrote:

No, that's not true in either the analogy of fluids or (apparently) in EM mechanics. The flow equations that correctly (and very accurately) predict subsonic fluid accelerating through a shrinking nozzle predict that both for water, but also for "ideal gases" like the atmosphere. The biggest assumption that defines a gas as an "ideal gas" is that it's molecules do not interact with each other in the fluid. It's not true, if memory serves the average collision distance for standard temperature and pressure atmosphere is a few centimeters, but it's true enough for many cases.

And it's not a perfect analogy, like I stated, a closed nozzle, filled with water or air or water, would absolutely not behave like this. It's just the closest I could come up with to a similar actual physical principle that's more intuitive.

There is also the whole particle/wave dual nature of photons that complicates things. A lot of times, the decision of whether it should act like a wave or a particle can be traced back to "which way makes the math easier".

Bfgp wrote:
MoonDragon wrote:
Moggy wrote:
Gremlin wrote:

And it won't make it easier to take off from planets (where you need a lot of delta V all at once)

It might. Imagine a circular track that allows a payload to be accelerated to escape velocity, at which point the payload is directed to a ramp and "up, up, and away". Not sure about air resistance, but if the motor is scalable, why not?

I think a railgun/hyperloop type accelerator would be quite more efficient than an EM drive when it comes to use of electricity for propulsion.

Robear wrote:

I suspect it'll be at least a decade before they put one of these in orbit. Looking forward to the results, however.

May be a lot sooner than that.

Who knows, maybe in this century humans will put electromagnetic rail accelerators powered by solar in orbit around (say the moon) and use EM driven payloads to to crush your enemies, to see them driven before you, and to hear the lamentations of their women. super sling shot objects at a massive velocity from the solar system. One thing is clear - there's no way chemical propelled rockets will have any feasible chance of getting to a new solar system within humanity's foreseeable future.

Fixed that for the real way humans would use it.

No new info on this, I take it?

No, I've poked my head into the "emdrive" subreddit a few times to see if there has been any news since the very unconvincing paper was published, and never saw any progress.