-
#1800 by Mulletron on 09 Sep, 2015 23:08
-
@6:43, see the "trapnado"
Do any computational electrodynamics experts out there know how to simulate the motion of a positive or negative ion inside an Emdrive excited with TE012 for example?
-
#1801 by dustinthewind on 09 Sep, 2015 23:41
-
1...
2...
I am wondering if my post on this thread is too convoluted and no one feels they understand it, or if people feel they understand it but think that I don't know what I am talking about. It does appear everyone is very busy already and so maybe I should be posting this in a separate thread? I feel like it qualifies as a type of EM drive so maybe it should be here and just sit a bit on the back shelf? Any suggestions?
One question I have is if anyone has measured with a Hall effect magnetic field meter (http://hyperphysics.phy-astr.gsu.edu/hbase/magnetic/hall.html), an osculating signal that decreases with 1/r^3 coming from outside the base of a resonating cavity which changes in phase with distance. I do expect such a signal to be there in the frame of the meters moving charge. I am not sure a meter exists that could record data fast enough though it is possible with the right design, I think. It would at least appear as noise on a meter that was too slow.
Are there any arguments for why it won't work?
Are there any questions?
Suggestions, that I should move it to a new thread? -
#1802 by zellerium on 09 Sep, 2015 23:43
-
I don't quite get the differentiation you are describing between FDTD and FEM.
Meep uses the central difference method both in the space and time domains to solve transient problems. One can use the Finite Element method in space with Finite Difference in the time domain (FDTD) to solve transient problems. I wrote several such programs many years ago to solve very nonlinear problems. Actually most Finite Element programs use Finite Difference methods in the time domain to solve transient problems.
Perhaps when you are referring to the Finite Element method you are referring to an eigenvalue solution using the Finite Element method ? Is your FDTD solution using finite difference or finite elements in the space domain?
Yes, the FEM solver in EMPro is an Eigenvalue method, however it is different than the pure Eigenfrequency solver (another option) which does not have a source but simply finds the resonant frequency of whatever geometry you give it.
I'm not sure the distinction between finite elements and finite difference, (both of them divide the space into small cubes where electric/magnetic fields are calculated right? )
This quote from the Keysight manual will probably clarify:
"In the FDTD approach, both space and time are divided into discrete segments. Space is
segmented into box-shaped cells, which are small compared to the wavelength. The electric fields are located on the edges of the box and the magnetic fields are positioned on the faces as shown in the figure below. This orientation of the fields is known as the Yee cell, and is the basis for FDTD."
source: http://edadownload.software.keysight.com/eedl/empro/2010/pdf/emprosim.pdf
-
#1803 by zellerium on 10 Sep, 2015 00:12
-
Um can somebody explain to me why TE012 is so special? I mean this thing is looking more and more propeller like as we go on. For god stakes it even seems to be torquing the frustum similar to what a propeller does. What fluid is it acting on? I don't have the fogiest. The swimming in space article gives me some interesting ideas, but anything dealing with physics at that level tends to come down to math, not intuition.
Ok, so basically you're trying to circularly polarize the wave and TE is better for this purpose than TM. So why is one TE mode any better than any other? Are you trying to circularly polarize an evanescent wave and TE012 is somehow special? If so why the frustum? Are we somehow generating more evanescent waves by using it? More evanescent waves in one direction than another (greater area of the big base than the small base to generate them, would explain why the symmetrical test was a null)? If so why on earth would an evenescent wave do something that a normal one could not? If evenescent waves tend to be generated 1/3 wavelength from an antenna what effect does the frustum deforming the wave have? Why Q? Does the probability of generating an evenescent wave increase with each bounce? Does the effect just require a highly energetic environment?
Does this paper have any bearing on the matter at hand? (Singular evanescent wave resonances) http://arxiv.org/abs/1311.3718
This thing seems to spin up (turn on delay), spin down (turn off delay) and torques to one side. It acts like a propeller and I'm completely at a loss as to how it could be doing anything in a vacuum chambers. My only thought is that the torquing (reported here and with Tajmar) is somehow the result of the use of a waveguide making the thing look more propeller like than it actually is.
Think I'm going to read a book and not think about EMDrives for a bit.
I believe the more fundamental the mode, the the higher amplitude of resonance. Whether TE or TM will 'thrust' better is still up in the air IMO. But we do know Yang claims the TE011 mode to be the best she could simulate, and she reported the largest thrust values.
"It was found that the thruster cavity made by copper and resonating on the equivalent TE011 mode has a quality factor 320400 and generates total net EM thrust 411 mN for 1000 W 2.45 GHz incident microwave"
-Yang 2013
EW's TM212 seemed to work, but thrust to power was significantly less. And we can't forget they needed a dielectric, a mystery that still hasn't been solved. I don't know if I'd settle with Paul March's original theory that the phase modulation and amplitude modulation of the magnetron contributes to no need for a dielectric. I think it has to do with what mode was excited in EW's cavity and the need for significant field asymmetry. Perhaps the frustum shape alone can cause enough asymmetry when using fundamental modes because they span more of the tapered height. A TM 212 on the other hand does not span very much of the height so a dielectric may have been necessary to 'squish' the fields down on one side so that enough asymmetry was created for thrust.
But take all of that with a grain of salt, I haven't spent as much time thinking about why this thing works and rather focused on how to get one to work. -
#1804 by zellerium on 10 Sep, 2015 00:18
-
...
I am wondering if my post on this thread is too convoluted and no one feels they understand it, or if people feel they understand it but think that I don't know what I am talking about. It does appear everyone is very busy already and so maybe I should be posting this in a separate thread? I feel like it qualifies as a type of EM drive so maybe it should be here and just sit a bit on the back shelf? Any suggestions?
One question I have is if anyone has measured with a Hall effect magnetic field meter (http://hyperphysics.phy-astr.gsu.edu/hbase/magnetic/hall.html), an osculating signal that decreases with 1/r^3 coming from outside the base of a resonating cavity which changes in phase with distance. I do expect such a signal to be there in the frame of the meters moving charge. I am not sure a meter exists that could record data fast enough though it is possible with the right design, I think. It would at least appear as noise on a meter that was too slow.
Are there any arguments for why it won't work?
Are there any questions?
Suggestions, that I should move it to a new thread?
It seems like it would be possible, as you said, with fast enough data collection and a sensitive enough probe.
What would be the result of such a measurement?
-
#1805 by Rodal on 10 Sep, 2015 00:23
-
I don't quite get the differentiation you are describing between FDTD and FEM.
Meep uses the central difference method both in the space and time domains to solve transient problems. One can use the Finite Element method in space with Finite Difference in the time domain (FDTD) to solve transient problems. I wrote several such programs many years ago to solve very nonlinear problems. Actually most Finite Element programs use Finite Difference methods in the time domain to solve transient problems.
Perhaps when you are referring to the Finite Element method you are referring to an eigenvalue solution using the Finite Element method ? Is your FDTD solution using finite difference or finite elements in the space domain?
Yes, the FEM solver in EMPro is an Eigenvalue method, however it is different than the pure Eigenfrequency solver (another option) which does not have a source but simply finds the resonant frequency of whatever geometry you give it.
I'm not sure the distinction between finite elements and finite difference, (both of them divide the space into small cubes where electric/magnetic fields are calculated right? )
This quote from the Keysight manual will probably clarify:
"In the FDTD approach, both space and time are divided into discrete segments. Space is
segmented into box-shaped cells, which are small compared to the wavelength. The electric fields are located on the edges of the box and the magnetic fields are positioned on the faces as shown in the figure below. This orientation of the fields is known as the Yee cell, and is the basis for FDTD."
source: http://edadownload.software.keysight.com/eedl/empro/2010/pdf/emprosim.pdf
The Yee cell is a Finite DIfference in space domain cell.
Finite difference methods approximate the partial differential equation with finite differences at nodes (points in the domain). Higher order finite difference methods in space need to use several nodes.
Finite Element methods divide the space domain into polygons (they don't need to be cubes, actually tetahedrons are as common or more commonly used) with a functional interpolation (usually polynomials) inside them.
Finite Element methods can be based on variational principles (although they don't need to be, some of them are based on the Galerkin method for example) and therefore have much better convergence properties than Finite Difference methods. -
#1806 by dustinthewind on 10 Sep, 2015 00:25
-
...
...
It seems like it would be possible, as you said, with fast enough data collection and a sensitive enough probe.
What would be the result of such a measurement?
The observation would be an osculating voltage (a wave) on the magnetic field meter towards or away from the cavity base that depends on current direction in the base of the cavity and current in the Hall effect probe. It should be the static magnetic field (not -dB/dt) that is being observed to penetrate the cavity. The static magnetic field really being the relativistic dipole electric field I indicate in the image of my earlier post. (the one with the curved spokes for negative charge).2...
With a probe that is too slow it would appear as increased noise but osculating between a high and low voltage. It just wouldn't have enough data points to capture the wave behavior. A Hall effect meter would be ideal I would think. Currents should be large with a large Q of the cavity and enough stored energy.
If the signal travels at light speed and "if a wave form can be observed" then there should be a delay in phase with probe distance from the base. This delayed signal should be lacking in any static electric field (no electric field is observed in the lab frame) and should be purely magnetic if the cavity is in a low Transverse Electric mode with only circulating currents. If this is so then I suspect we can use this signal for magnetic propulsion using another cavity and the forces should be greater than radiation propulsion. (Unlike a phase antenna array which should only give radiation propulsion because of charge separation or capacitance). -
#1807 by rfmwguy on 10 Sep, 2015 01:40
-
Hey Doc, with all the recent talk about corkscrew modeling, aren't you glad I didn't wander off into an EM Vortex type analogy?
I'm very pleased with myself, holding back on my early, wild speculations about mass displacement from the center of the vortices to avoid superluminal velocities...doooh! Sorry, Doc
-
#1808 by SteveD on 10 Sep, 2015 02:40
-
Hey guys, on some reflection I think somebody knowledgeable really needs to look at this paper.
Singular evanescent wave resonances http://arxiv.org/ftp/arxiv/papers/1311/1311.3718.pdf
Yu Guo, Zubin Jacob
(Submitted on 15 Nov 2013 (v1), last revised 17 Sep 2014 (this version, v3))
Resonators fold the path of light by reflections leading to a phase balance and thus constructive addition of propagating waves. However, amplitude decrease of these waves due to incomplete reflection or material absorption leads to a finite quality factor of all resonances. Here we report on our discovery that evanescent waves can lead to a perfect phase and amplitude balance causing an ideal Fabry-Perot resonance condition in spite of material absorption and non-ideal reflectivities. This counterintuitive resonance occurs if and only if the metallic Fabry-Perot plates are in relative motion to each other separated by a critical distance. We show that the energy needed to approach the resonance arises from the conversion of the mechanical energy of motion to electromagnetic energy. The phenomenon is similar to lasing where the losses in the cavity resonance are exactly compensated by optical gain media instead of mechanical motion. Nonlinearities and non-localities in material response will inevitably curtail any singularities however we show the giant enhancement in non-equilibrium phenomena due to such resonances in moving media.
http://arxiv.org/abs/1311.3718
Sorry but I can't help but wondering, what if you can run this process backwards? I know I'm going to botch this but, get two plates at the critical distance and stick them so that they can't move. Circularly polarize the signal so that it is corkscrewing into the end plate. An evanescent wave would be generated at a 90 degree angle (am I right that with the wave corkscrewing in this would be toward the other base). If this could be run backwards, it would seem that the closer to perfect Q the setup gets the more it would want to introduce a mechanical movement between the plates. Let me go even further out and suggest that when the plates are of two different sizes one gains more energy in the direction of its movement that the other does in the opposite direction.
Does it conserve energy: Well I don't think you can get more energy out of something that you load into an object, minus heat, with this proposed effect.
Does it conserve momentum: Well maybe. The waves would give up more of their energy to the baseplate generation movement in the direction of travel than the baseplate generating "drag." That's really weird, but maybe it conserves momentum.
Does it do weird things to spacetime: um the paper proposes a "negative Poynting vector," I suspect that is going to do something odd (though wouldn't it be positive if the process were reversed).
Ok it's late and I'm showing my lack of knowledge on a paper that is already pushing some theoretical boundaries. -
#1809 by flux_capacitor on 10 Sep, 2015 09:13
-
Um can somebody explain to me why TE012 is so special? I mean this thing is looking more and more propeller like as we go on. For god stakes it even seems to be torquing the frustum similar to what a propeller does. What fluid is it acting on? I don't have the fogiest. The swimming in space article gives me some interesting ideas, but anything dealing with physics at that level tends to come down to math, not intuition.
Ok, so basically you're trying to circularly polarize the wave and TE is better for this purpose than TM. So why is one TE mode any better than any other? Are you trying to circularly polarize an evanescent wave and TE012 is somehow special? If so why the frustum? Are we somehow generating more evanescent waves by using it? More evanescent waves in one direction than another (greater area of the big base than the small base to generate them, would explain why the symmetrical test was a null)? If so why on earth would an evenescent wave do something that a normal one could not? If evenescent waves tend to be generated 1/3 wavelength from an antenna what effect does the frustum deforming the wave have? Why Q? Does the probability of generating an evenescent wave increase with each bounce? Does the effect just require a highly energetic environment?
Does this paper have any bearing on the matter at hand? (Singular evanescent wave resonances) http://arxiv.org/abs/1311.3718
This thing seems to spin up (turn on delay), spin down (turn off delay) and torques to one side. It acts like a propeller and I'm completely at a loss as to how it could be doing anything in a vacuum chambers. My only thought is that the torquing (reported here and with Tajmar) is somehow the result of the use of a waveguide making the thing look more propeller like than it actually is.
Think I'm going to read a book and not think about EMDrives for a bit.
I believe the more fundamental the mode, the the higher amplitude of resonance. Whether TE or TM will 'thrust' better is still up in the air IMO. But we do know Yang claims the TE011 mode to be the best she could simulate, and she reported the largest thrust values.
"It was found that the thruster cavity made by copper and resonating on the equivalent TE011 mode has a quality factor 320400 and generates total net EM thrust 411 mN for 1000 W 2.45 GHz incident microwave"
-Yang 2013
EW's TM212 seemed to work, but thrust to power was significantly less. And we can't forget they needed a dielectric, a mystery that still hasn't been solved. I don't know if I'd settle with Paul March's original theory that the phase modulation and amplitude modulation of the magnetron contributes to no need for a dielectric. I think it has to do with what mode was excited in EW's cavity and the need for significant field asymmetry. Perhaps the frustum shape alone can cause enough asymmetry when using fundamental modes because they span more of the tapered height. A TM 212 on the other hand does not span very much of the height so a dielectric may have been necessary to 'squish' the fields down on one side so that enough asymmetry was created for thrust.
But take all of that with a grain of salt, I haven't spent as much time thinking about why this thing works and rather focused on how to get one to work.
In the US patent #3,425,006 "Cavity resonator with mode discriminating means" by James Wolf, from 1969, already posted by Rodal in Thread 2, it is clearly stated that TE modes are better than TM modes to achieve precise tuning and high Q factor:Quote from: James WolfThe TE modes are highly desirable for tunable high-Q resonators because they do not require current crossing between the outer Wall and the circular end plates of the cavity, and because the Q factor is high. These particular modes have therefore found Wide application.
-
#1810 by SeeShells on 10 Sep, 2015 13:58
-
Hey Doc, with all the recent talk about corkscrew modeling, aren't you glad I didn't wander off into an EM Vortex type analogy?
There is nothing wrong in a helical wave injected into the frustum if done right. It also has nothing to do with superluminal wave velocities.
I'm very pleased with myself, holding back on my early, wild speculations about mass displacement from the center of the vortices to avoid superluminal velocities...doooh! Sorry, Doc
The key is to match the helical wave pattern to the one the frustum wants to generate from it's geometry. In the case of the current loop tested by meep inside the centerline of the frustum the loop generated wave patterns that were mostly subtractive to any mode generation.
Select the correct rotational pattern along with the correct cavity size it becomes a additive process instead of subtractive. Very similar to the one animation I just did, picking this mode as it's easy so see the effects of a matched frequency/helical/frustum pattern generation.
Shell
-
#1811 by SeeShells on 10 Sep, 2015 14:27
-
Can anyone here convert a .skp to a GDSII?
-
#1812 by bprager on 10 Sep, 2015 15:28
-
#1813 by SeeShells on 10 Sep, 2015 15:34
-
#1814 by SteveD on 10 Sep, 2015 16:54
-
Has anyone done a mathematical proof to make sure a stable TE012 is possible in a frustum at these frequencies? We would all feel a bit silly if we spent all this time and effort trying to do something that was impossible. (Certain ironies with the above statement and this entire project are noted).
Speaking of which, here is the dispersion measure plot from the Arecibo, Fast Radio Burst detection. I've said elsewhere that I think its either the result of the classified project Shawyer says he was working on (very likely) or possibly somebody else using an EMDrive subject to engineering restraints we are not currently aware of (highly unlikely going to extremely unlikely if that somebody is not associated with the planet Earth). Since I don't know the cutoff for Arecibo's equipment this is of limited value (if a full run starts at 1hz and goes all the way up to the high gammas it clearly stellar in nature). Still, I can't help but feel that if I could, arbitrarily, pick a frequency to start poking about around with an EMDrive, somewhere in this range might bear interesting fruit. (Paper: http://arxiv.org/abs/1404.2934)
-
#1815 by SteveD on 10 Sep, 2015 17:29
-
Follow up from that Arecibo paper, please tell me that that is not what I think it is. Alternatively, please show me that its somebodies spy sat that we aren't suppose to know about.
-
#1816 by Ricvil on 10 Sep, 2015 19:27
-
Why the field on the loop antena appears like a traveling wave along the loop? The current is not balanced?
When a loop is feeded by a transmission line there are two antisymmetric current waves, and one will see the fields on the loop arising as two couterpropagating waves and resulting a stationary wave on the loop.
What is going on? -
#1817 by aero on 10 Sep, 2015 19:46
-
I don't know what I'm doing but here is my "Magnetron/waveguide" feed.
https://drive.google.com/folderview?id=0B1XizxEfB23tUWZVY2RMc3lJbVU&usp=sharing
I've coded a constant frequency for use as a wave length ruler, and am exciting the cavity at a variable frequency parameter, both are set to 2.47 GHz, and the constant won't change.
The dimensions are x= wl/4, z = wl/2 and the length, y = 2*wl. (constants) The x and z dimensions are then increased by 30% as I read somewhere that they should be. I hope to be able to shorten the y dimension as that length will make the meep computational lattice quite large.
The structure is excited from a rectangular plate wl/4 from the big end. The plate dimensions are yp = 0, xp = 0.9*x, zp=0.9*z (all constants) excited by the Ex component. The resulting Hz component is displayed.
I can use all the help I can get. -
#1818 by aero on 10 Sep, 2015 19:48
-
Why the field on the loop antena appears like a traveling wave along the loop? The current is not balanced?
When a loop is feeded by a transmission line there are two antisymmetric current waves, and one will see the fields on the loop arising as two couterpropagating waves and resulting a stationary wave on the loop.
What is going on?
I programmed it like the receiving loop antenna on my TV set. Is that not right? -
#1819 by Ricvil on 10 Sep, 2015 19:58
-
Why the field on the loop antena appears like a traveling wave along the loop? The current is not balanced?
When a loop is feeded by a transmission line there are two antisymmetric current waves, and one will see the fields on the loop arising as two couterpropagating waves and resulting a stationary wave on the loop.
What is going on?
I programmed it like the receiving loop antenna on my TV set. Is that not right?
The point aero is how the loop antena model is feeded.
Sorry I need take a bus now. I will clarify the question later.
