Cannae/EmDrive

[RuBisCO]

At least three NSF users are building their own test devices and rigs currently. One of them (rmfguy) already started preliminary tests. There is also a team testing a 'baby-Emdrive' - an engine about the size of coffee mug

IF it works that sounds like a likely candidate for the first space trials. I think that if this thing really is legit it won't be until an orbital platform does a few km/s in delta-v before the world believes. Someone is going to have to get one of these things into orbit. The small the better, but aren't this things limited by the frequency, the smaller it gets the higher the frequency required right?

[PB666]

In this case, main cost is time. Just about everyone who has access to the right equipment, like vacuum chambers, and knows how to carry out the experiment precisely enough and accounting for all the variables properly, would probably consider it a wild goose chase not worth pursuing, because they have thier own research, which they actually expect positive results from.

I'll admit, though. I'm starting to get curious. If I still had access to test equipment I'd need, I'd probably be running my own experiments by now.

Hah, but still not curious enough to suggest it goes into space. It needs to be up!

[ZetaX]

Hah, but still not curious enough to suggest it goes into space. It needs to be up!

No. Just no.

There is no reason to spend tens of millions on sending something into space just to see if it produces thrust. A mission like that would not be approved by any of the larger space agencies (but sure, you can try to convince private investors to waste their money). To make this a viable mission you need a lot more than "lets see if it works". This includes tests that we cannot do on earth that are likely to produce insights that cannot be found here on earth.

And this doesn't even have to do with that thing probably not working as advertised due to violating known physics. Many other drives also went through years and years of surface testing, then vaccum chamber testing, before finally being sent to space (or canceled for bein impractical). Take almost any drive as an example (orion, solar sails, ion, ...).

[Mazon Del]

There is no reason to spend tens of millions on sending something into space just to see if it produces thrust.

Don't forget though, a cubesat sized setup won't cost more than maybe 20-50K to send up and operate. Now, you would have to get permission from NASA to test a propulsion system which they have never done before, but primarily because all of the cubesat sized propulsion systems up to this point have largely been in violation of other rules (too much pressurization, volatile chemicals, etc). They have long said that when someone get's a cubesat ion engine up they are ready to look over a process for allowing it.

[ZetaX]

Don't forget though, a cubesat sized setup won't cost more than maybe 20-50K to send up and operate.

That's very very optimistic. That's the prize for a single-cubed sat and that is probably way too small. This thing needs a chamber and a relevant power supply, so either a RTG (a no-go for cubesats, I think), a chemical battery (rendering its operation time very short) or a huge solar panel (being voluminous and heavy). I don't even expect this to fit in a 1x2x3; maybe a 2x2x3, and that's already around a million, and still not guaranteed to fit.

And that's also only the prize for LEO, but you want to get higher to exclude atmospheric effects (e.g. it propelling distant air).

Not speaking of the actual costs of designing and building it.

[11of10]

@ZetaX You don't care about propulsion by distant air if it works, you'll raise your apo that way and then try to circularise to do a final check. And even if it only works in the air, it's still a success, if it really can produce enough trust.

Just think, maybe it can replace jet engines

[ZetaX]

@ZetaX You don't care about propulsion by distant air if it works, you'll raise your apo that way and then try to circularise to do a final check. And even if it only works in the air, it's still a success, if it really can produce enough trust.

You can't even circularize that way. All the planning, execution and whatever of a rather long-term orbit-change adds into the costs as well.

And if it only works in the air, then you just wasted millions in shooting a very bad jet engine into space.

[RuBisCO]

That's very very optimistic. That's the prize for a single-cubed sat and that is probably way too small. This thing needs a chamber and a relevant power supply, so either a RTG (a no-go for cubesats, I think), a chemical battery (rendering its operation time very short) or a huge solar panel (being voluminous and heavy). I don't even expect this to fit in a 1x2x3; maybe a 2x2x3, and that's already around a million, and still not guaranteed to fit.

And that's also only the prize for LEO, but you want to get higher to exclude atmospheric effects (e.g. it propelling distant air).

Not speaking of the actual costs of designing and building it.

It all depends on how small they can get these thing, which in its self is a test of how it works. So far everyone has been testing in the 2-4 GHZ range, with "engines" too large to fit on a cube sat. One of these just a few cm wide that runs on a few watts should work according to Shawyer theory at 10-20 GHz. If the small one produces thrust and thrust to scale as predicted, then its more evidence for the extraordinary claims. Of course they will need to produce thrust, consistently for at least weeks in a vacuum chamber as well as prove linear acceleration in vacuum, before they will have enough evidence to warrant even a cubesat.

A Satellite would need at least 12 "thruster" for full attitude control, maybe 3 fly-wheels and one thruster but that would not last too long in space, they are going to need to last long enough to demonstrate km/s to have the extraordinary evidence needed for the world will to believe.

[Mazon Del]

As with what RuBisCO said, other ways of solving the power issue involve slow charging a battery for several sun cycles and a discharge over the same location in orbit. Space Command's tracking can easily see small shifts and they are generally pretty good about reporting such things if you ask. It's in their interest that you know where your space garbage is.

[11of10]

You can't even circularize that way. All the planning, execution and whatever of a rather long-term orbit-change adds into the costs as well.

And if it only works in the air, then you just wasted millions in shooting a very bad jet engine into space.

Sorry, I meant if it were made into a cubesat, and if it could hitch a ride with something else cheaply. Who knows, maybe it could even be a crowd funded test?

[ZetaX]

Sorry, I meant if it were made into a cubesat, and if it could hitch a ride with something else cheaply.

Umm, that's exactly the case I responded to.

[PB666]

No. Just no.

There is no reason to spend tens of millions on sending something into space just to see if it produces thrust. A mission like that would not be approved by any of the larger space agencies (but sure, you can try to convince private investors to waste their money). To make this a viable mission you need a lot more than "lets see if it works". This includes tests that we cannot do on earth that are likely to produce insights that cannot be found here on earth.

And this doesn't even have to do with that thing probably not working as advertised due to violating known physics. Many other drives also went through years and years of surface testing, then vaccum chamber testing, before finally being sent to space (or canceled for bein impractical). Take almost any drive as an example (orion, solar sails, ion, ...).

No, if my guess is right, the physics wont 'behave' properly until you get it into space. Any and all further testing on the ground may be useless because there is simply too much interference.

BTW 20 million dollars is nothing if there is a 1% chance that it works, it would pay of 1000 fold. And what is the primary mission of NASA (other than spying on everyone). How does the saying go, if we knew everything it wouldn't be called science. Or let me put this another way, on earth it works, but who needs to push off of whatever (unknown thing we have been handwaving about for the last 6 months) when wheels push off the ground (extremely effective) and props and jets translocate gas(very effective);

Simple. Small device on another payload heading to geosynchronous orbit. At orbit device is released and allowed to drift away. Then the device is turned on and you track it. Doesn't have to be big, all it needs to have is a radio transmitter with a GPS on it.

[ZetaX]

But the 1% is totally made up, that's part of the problem. If we put this against physics as a whole, it might easily be 0.0001% only and then it suddenly is a very ineffective way to spend the money. Instead, you could have spent it on the surface to figure out the actual way it works or at least exclude absolutely everything else. Even if you put it in space and it works, then what¿ Without a sufficient understanding of the how's and why's you still have no idea how to fully utilize it. On the other hand, detailed ground-based experiments with fine measurements may tell you how it works and then we can improve it.

[magnemoe]

No, if my guess is right, the physics wont 'behave' properly until you get it into space. Any and all further testing on the ground may be useless because there is simply too much interference.

BTW 20 million dollars is nothing if there is a 1% chance that it works, it would pay of 1000 fold. And what is the primary mission of NASA (other than spying on everyone). How does the saying go, if we knew everything it wouldn't be called science. Or let me put this another way, on earth it works, but who needs to push off of whatever (unknown thing we have been handwaving about for the last 6 months) when wheels push off the ground (extremely effective) and props and jets translocate gas(very effective);

Simple. Small device on another payload heading to geosynchronous orbit. At orbit device is released and allowed to drift away. Then the device is turned on and you track it. Doesn't have to be big, all it needs to have is a radio transmitter with a GPS on it.

This is not how it works exactly, main issue now is that they don't have vacuum rated capacitors, guess they have to be special made.

The main interest now is to ramp up power to get more trust, on an cubesat you don't get much power.

Secondary is to tweak the device to increase trust.

You can do neither in on an cubesat probe.

You want to increase effect since it will make testing easier and because it become an more interesting engine.

[Mazon Del]

Without a sufficient understanding of the how's and why's you still have no idea how to fully utilize it.

Not true at all, or rather, that isn't a good enough excuse to NOT do a cubesat test. Mankind pretty much specializes in expanding the utility of tech without knowing why it works. Now, sure, we can do it much better with the understanding, but we've gotten along quite fine without full understanding in many fields. There are still chemicals and metals that history tells us existed beyond a doubt, that we still (with our modern day understanding of such) do not know how to replicate properly. Damascus steel, greek fire, etc.

Trial and error is a real and valid methodology for advancement, ESPECIALLY if you don't have a full understanding.

And that is the great thing with the cubesat EMdrive idea, is that we don't NEED to use the 100% perfect version of the drive. That isn't the point. The sat isn't intended to go anywhere in particular or do anything useful beyond simply show that the drive works at all. So we don't care if it is maximizing it's abilities as long as we are pretty sure it will be able to prove that it works at all.

Additionally, as to why they don't have the vacuum rated caps. For the specific requirements they need vr-caps exist, but they are a bit of a specialty item. So it has worked out to be far cheaper to just replace the bad caps as necessary than to permanently fix it. Especially since this is just the version 1. The version 2+ will be designed more with this sort of thing in mind.

[SomeGuy12]

Yeah. I don't see a problem with a small probe that tries to use this drive. It will be possible to tell within a few days to a few months max if it works or not, and the signal will be unmistakable. If the thing gains tens of meters per second of dV that wasn't in the fuel tanks, there's not going to be anything the skeptics can say. If it doesn't budge at all, there's nothing the supporters can say. Understanding the theory is besides the point - if it actually works, you can set fire to most physics textbooks and start over.

I'm sorta imagining the post-event physics textbooks. "Conservation of Momentum *"

* only a suggestion, and not the law

[ZetaX]

Not true at all, or rather, that isn't a good enough excuse to NOT do a cubesat test.

It was an addendum to a much broader point: money and expected use.

Mankind pretty much specializes in expanding the utility of tech without knowing why it works. Now, sure, we can do it much better with the understanding, but we've gotten along quite fine without full understanding in many fields.

We only got there with tons of time. Which again leads to what I said: no reason to rush the sat. And a lot of fields made almost all their progress deductively, not by trying random things (the only exceptions that come to my mind are medicine and psychology, and those are about one of the most complex things we know, so simplifications become a necessity).

There are still chemicals and metals that history tells us existed beyond a doubt, that we still (with our modern day understanding of such) do not know how to replicate properly. Damascus steel, greek fire, etc.

Damascus steel is essentially over-hyped steel, worse than modern steels. In the end, it is quite easy to make something that is hard to replicate, but in most cases this also won't be the best product we could make. Try to _exactly_ replicate a computer from 1980; I bet you will fail.

And for the greek fire: this is due to us not knowing what it was. So we can only guess. It's not like we have some chemical from that time that we simply cannot produce. So to say that we cannot make it is unscientific (due to being unfalsifiable).

Trial and error is a real and valid methodology for advancement, ESPECIALLY if you don't have a full understanding.

It was up to some centuries ago. Then we realized that understanding works much better.

Also, "trial and error" is a bit ridiculous if you want to use it as an argument for sending this into space instead of testing on earth. On earth we can easily adapt and repeat experiments; good luck in space, even if we could ignore costs.

And that is the great thing with the cubesat EMdrive idea, is that we don't NEED to use the 100% perfect version of the drive. That isn't the point. The sat isn't intended to go anywhere in particular or do anything useful beyond simply show that the drive works at all. So we don't care if it is maximizing it's abilities as long as we are pretty sure it will be able to prove that it works at all.

And what would this actually accomplish¿ Knowing that it works is an almost useless information by itself in regard to making it useful.

If you understand it enough to maximize it, then chances are essentially 1 that you already know how it works, thus rendering space tests only a final stage.

[RuBisCO]

If a cube sat manages a few dozen m/s in a few weeks with EMdrive the skeptics will claim it was a hoax, that the cubesat had some kind of alternate real propulsion on it. A cubesat demo would need:

1. No alternative propulsion, not even for attitude control, this means flywheel or emdrive attitude control, flywheels will only last so long before getting saturated.

2. Needs to do enough delta-v fast enough that no conventional propulsion could achieve it.

Lets work with a 3U CubeSats that is 4 kg max weight to work with. Lets say hypotheticall if had a 3 kg rocket with ISP of 350, it would be able to do 4.8 km/s of delta-v! So we are talking about a 3U CubeSat that would need to go from geosyn loop around the moon and back a few times before the world believes that EMdrive is real beyond any and all doubt.

I estimate to fit 12-16 EM drives in a cubesat, each drive would need to to be less then 5X5X5 cm long. A 40x32.7x27.8 resonates chamber would need to operate at 9.72 GHz, that sounds reasonable but I have not seen anyone test one above 4 GHz. The two outer cube units would hold 6-8 EM drives each, the inner cube would hold flight computer and batteries. Solar array of 30x30cm would (2 side wings) produce 32 watts at 20% efficiency. So each EM drive assuming four running at once at most would top out at 7-8 watts with continuous solar power, not including other inefficiencies. Let us assume 15 W at 0.1 mN/W that is 15 mN to accelerate 4 kg, or 3.75 mm/s^2 acceleration, it would take 15 days of continuous thrusting to do 4.8 km/s of delta-V.

So say a several month mission to lunar orbit and back with a 3U cubesat would be possible IF EMdrive is real.

[Mazon Del]

And what would this actually accomplish¿ Knowing that it works is an almost useless information by itself in regard to making it useful.

Oh god no. Have you ever heard the phrase "the Chicago Pile moment"? For the longest time, proponents of nuclear power insisted it was possible to harness the atom in order to produce useful energy. But everybody was terribly skeptical and funding to investigate this was immensely difficult to come by. The largest source of it ended up being funds from the Manhattan Project. The Chicago Pile-1 was the first manmade nuclear reactor built in 1942, first chain reaction on December 2nd of that year. It demonstrated what many of the brightest minds considered impossible, that man could intentionally produce, and control to some degree or another, an artificial fissile nuclear reaction. The time in between this event and the worlds first commercial nuclear power plant? The Shippingport Atomic Power Station began construction on September 6th, 1954, producing power on May 26th, 1958. 12 years. Twelve years to go from something that was "obviously impossible and/or impractical" to a completed design and construction underway. That is unimaginably fast for a new technology back then. What changed investors minds? Some crappy little graphite box made a Geiger counter tick it's head off with no other practical application behind its existence.

So here's the advantage to having a cruddy cubesat with an EMdrive on it put its way around an orbit. Even if it doesn't convince the scientific community as a whole, what it WILL convince is the heavy risk takers that want to invest in moon-shot projects that maybe there is enough of something there to put money into. That one little $50K box could quite easily increase funding into the drive by two orders of magnitude if not more.

Alternatively, it could prove that some enthusiastic non-physicists have no idea of what they are doing and nothing happens, meanwhile science happily grinds on in its inexorable but glacial pace.

[sjwt]

Ok guys, lets play nice, keep on track and don't attack other people.

[Dodgey]

K^2 is absolutely correct. In an international forum in a thread talking about a sketchy new propulsion system that goes against known physics why should we think you mean anything less than what you say? When you said ISP you could have been referring back to the idea that was thrown around that it could be an ablating the walls of the device to produce thrust, or you could be referring to its efficiency with regards to power needed for thrust produced. Because ISP might not actually factor in to this engine (discounting nuclear fuel spent on generating electricity) we are left to guess what you mean by it.

I mean this is "The Science Labs", I do think that it is fair to have a higher standard with regards to preciseness of our terms than the other subforums.

On a different note.

The difference between the thrust of this engine and jet engines is that it doesn't change depending on how fast or high you are flying so you don't run into the same issues allowing you to fly faster and higher. It would be interesting in how this factors into its efficiency when used on an aircraft, sans power generation of course.

[PB666]

If a cube sat manages a few dozen m/s in a few weeks with EMdrive the skeptics will claim it was a hoax, that the cubesat had some kind of alternate real propulsion on it. A cubesat demo would need:

1. No alternative propulsion, not even for attitude control, this means flywheel or emdrive attitude control, flywheels will only last so long before getting saturated.

2. Needs to do enough delta-v fast enough that no conventional propulsion could achieve it.

Lets work with a 3U CubeSats that is 4 kg max weight to work with. Lets say hypotheticall if had a 3 kg rocket with ISP of 350, it would be able to do 4.8 km/s of delta-v! So we are talking about a 3U CubeSat that would need to go from geosyn loop around the moon and back a few times before the world believes that EMdrive is real beyond any and all doubt.

I estimate to fit 12-16 EM drives in a cubesat, each drive would need to to be less then 5X5X5 cm long. A 40x32.7x27.8 resonates chamber would need to operate at 9.72 GHz, that sounds reasonable but I have not seen anyone test one above 4 GHz. The two outer cube units would hold 6-8 EM drives each, the inner cube would hold flight computer and batteries. Solar array of 30x30cm would (2 side wings) produce 32 watts at 20% efficiency. So each EM drive assuming four running at once at most would top out at 7-8 watts with continuous solar power, not including other inefficiencies. Let us assume 15 W at 0.1 mN/W that is 15 mN to accelerate 4 kg, or 3.75 mm/s^2 acceleration, it would take 15 days of continuous thrusting to do 4.8 km/s of delta-V.

So say a several month mission to lunar orbit and back with a 3U cubesat would be possible IF EMdrive is real.

so you need 4 thrusters each side mounted at 0, 90, 180, 270 degrees relative to the direction of forward motion. Two of four thrusters to go forward, one thruster to turn bearing any direction, another to stop. You also need a solar panel, I suggest top mounted point forward, and a reasonable battery.

Oh god, I just realized we are trying to build a NASA mission using KSP logic (face palms).

[PB666]

But the 1% is totally made up, that's part of the problem. If we put this against physics as a whole, it might easily be 0.0001% only and then it suddenly is a very ineffective way to spend the money. Instead, you could have spent it on the surface to figure out the actual way it works or at least exclude absolutely everything else. Even if you put it in space and it works, then what¿ Without a sufficient understanding of the how's and why's you still have no idea how to fully utilize it. On the other hand, detailed ground-based experiments with fine measurements may tell you how it works and then we can improve it.

I contest this, there are alot of weaknesses in modern physics.

-The nature of dark matter and dark energy remain unknown.

-Quantum entanglement, the idea that you can alter the polarization of photon X and immediately change the polarization of its pair product elsewhere is contested idea.

and I can go on.

-Where neutrinos acquire their mass (and their momentum) and how they are able to flip states.

The black swan theory basically argues that statistical systems often assume the confidence of a much broader systems if the boundaries of those systems have never been properly challenged.

As scientist we often test the central regions of a distribution, and assume the proportions of the extreme edges. This is something that has been fortuitously tested over and over again in historical science and found often to be true. The fact that the drive is reproducibly creating a faux momentum gives us pause to ponder a new how, although it does not prove that the new how is true. Gravity, first and second laws of thermodynamics remained true to the early 20th century and god does not play dice with the Universe. . . . . . . . . . Quantum physics gets alot of leeway because you always have uncertainty at the smallest scales and via heisenberg uncertainty, BUT the full connotations of quantum physics have not been worked out, so that there is wiggle room still left for unusual things to happen.

To go back to an even more philosophical stance, in statistics there arises a point in where one has to decide whether to reject the null hypothesis or accept the hypothesis. This is known as the threshold of type I error or alpha. Alpha typically is presented that way when you fall below alpha, the hypothesis is summarily rejected. The probability that is generated below alpha is the risk of rejecting the null (or standard) hypothesis when it was in-fact true. However the other type of error is the error of accepting the false hypothesis is called the type II error. And this is where science has a particular amount of trouble. The hypothesis is often dependent on the distribution, and if you go pouring through the scientific literature (I've seen this so many many times in genetics especially prior to 2005) the distribution is never characterized and often assumed to be normal, when it is not. You cannot assume a distribution until you start looking at the margins. So the reality is that if you set on one side of a distribution and you get a certain T-test value you reject the hypothesis, but if one does a monte-carlo, sorted all possibilities analysis you accept, but on the other side the reverse is true. Bonnferronis correction and other correction modalities often assume independence between variables tested in parallel when it is known often they are not independent and grossly over correct. This is what Zar says and when you read this and think about this you realize why scientist often are strengent in their interpretations.

Whereas the probability of committing a Type I error is alpha, the specified significance level, the probability of committing a Type II error is beta a value that generally we neither specify nor know - Zar - biostatistical analysis, 2nd edition p. 46-47

Two groups are studying the same phenomena, lets just say the hypothesis is that X happens at rate Y and they are both testing the boundaries. Group I tests the boundaries and finds anamolous rate X2 but cannot reject the hypothesis. Group 2 finds the same result and publishes. Reviewer 3 combines the data from the studies of both groups and rejects the null hypothesis. Group 1 and Group 2 did not do an adequate level of power testing under the specified conditions, nor did either group specific that their conclusions were marginal (something that has changed since 2005 in some of the more powerful statistical studies). And actually this phenomena is observed quite commonly in the literature.

This is no guarantee of success of a theoretical challenge or failure, it specifically argues that before one can have confidence in ones conclusions that one must have increased the power of the argument at all points of the spectrum in which the argument might have validity.

[ZetaX]

I contest this, there are alot of weaknesses in modern physics.

-The nature of dark matter and dark energy remain unknown.

-Quantum entanglement, the idea that you can alter the polarization of photon X and immediately change the polarization of its pair product elsewhere is contested idea.

and I can go on.

-Where neutrinos acquire their mass (and their momentum) and how they are able to flip states.

Sorry to say this, but you are argueing like a creationist here (using that analogy because I know you to have some footing in biology):

The EM-drive as claimed goes against the fundamentals itself and so far has very weak evidence in that direction. To suddenly pull dark matter, which is a very recent topic and under heavy research, is a really bad argument against such basic thing such as conservation of momentum. You are essentially saying "physics as a whole is wrong cause of some unresolved problem", which is the one main reason always out forward against evolution. I hope it is clear why this is not actually an argument.

I can't say much about the neutrons, but it is new to me that this is disputed, so I would like to see some evidence.

And the entanglement one is essentially an argument from ignorance: you are definitely not changing the polarisation of a photon, and especially not to influence another photon. Instead, both photons did not have a well-defined polarisation before measurement at all. You can't send information that way. So entanglement is simply not what you claim it to be.

I wil probably comment on the statistics later, by the way. But I have probably less objections there anyway, only on the premisses.

Edited August 13, 2015 by ZetaX

[PB666]

Sorry to say this, but you are argueing like a creationist here (using that analogy because I know you to have some footing in biology):

The EM-drive as claimed goes against the fundamentals itself and so far has very weak evidence in that direction. To suddenly pull dark matter, which is a very recent topic and under heavy research, is a really bad argument against such basic thing such as conservation of momentum. You are essentially saying "physics as a whole is wrong cause of some unresolved problem", which is the one main reason always out forward against evolution. I hope it is clear why this is not actually an argument.

I can't say much about the neutrons, but it is new to me that this is disputed, so I would like to see some evidence.

And the entanglement one is essentially an argument from ignorance: you are definitely not changing the polarisation of a photon, and especially not to influence another photon. Instead, both photons did not have a well-defined polarisation before measurement at all. You can't send information that way. So entanglement is simply not what you claim it to be.

I will probably comment on the statistics later, by the way. But I have probably less objections there anyway, only on the premisses.

I pull dark matter and dark energy because if the standard model is correct neither of these should exist; second your creationist argument is a red-herring, simple as that and you should apologize for using it. Science is that which expands the boundaries of science, not playing the ostrich; we don't hide from things we can't explain; we try to come up with a more variable set of experimental conditions that would give a variation of results that might explain them.

You and K2 are the ones arguing that for the Cannae drive to function as a massless drive all of physics is wrong, I do not. I contend that either physics works the way we think or momentum transfers can occur over a range of possiblities not previously considered.

But more important than that, I see something more obviously than you see. They repeat the experiment over and over again and get the same result. Where have we seen this before, was it the slit experiment? Was it the attempt to measure light speed differences going 4 directions?

You have to begin significant alteration of the experimental conditions if you want to start creating a situation were you can plot one parameter versus a second parameter other than input voltage or amperage and output thrust.

The biggest problem that I see is that thrust is detected as the difference between the device and the surface (rotational) reference frame. There are ways of getting around this but as long as there is nearby structure and the level of near-by structure is not accounted there are means of momentum transfer nearby. The logical solution is to build a way bigger vacuum chamber, but eventually it becomes wiser to space the device than to create a very expensive chamber the size of a football stadium

The perfect earthbound for testing the device

You are at the South pole. you have neutralized the magnetic Field of the earth exactly in the center of the cube you are about to build, you have built a vacuum chamber the size of china's olympic stadium, you have a monofilament line hanging dead center from the evacuated cube. You have solar panels on the side of the devicewhich you feed with light both sides, ideally balanced with two very thin stabilizing lines coming off the side toward the sides near the ceiling, and a small battery on the device, the device is sealed hermetically with a non interacting and non-volatile material, you pulse the device at the harmonic frequency of the pendulum that is created. Estimated cost 20 billion dollars, risk to human life- high.