-
#1260 by aero on 01 Jan, 2017 18:28
-
I guess the few MHz difference is caused by different mesh conditions as well as by the fact that i used a point source instead of a loop.Monomorphic, can you please run a FEKO sim and post the cross-section image of this particular frustum:
mode: TE013
Freq: 2.450 GHz
SD: 149.30 mm
BD: 275 mm
LN: 275 mm
cone half-angle: 12.9°
Df: 0.832
Here you go. Same antenna size and location. Peak resonance at 2.44825Ghz These E-fields are weaker.
JMO - The few MHz difference is caused by the different mesh conditions. The Weaker E-fields are caused by the point source instead of the loop. That is, the different mesh conditions introduce a difference in the numerical discretization error, not a real effect.
The point source vs. loop ... in meep it's easy to tell because a loop requires more geometry to describe, leaving the source spread across a larger area. It is a trick to insure that the drive power of the simulation is equal when using a loop vs. a point source. But I get the impression that the loop couples to the cavity better than does a point source. A point source is theoretically the same as a small loop but it seems for that theory to hold, it must be a very, very small loop.
Again, JMO -
#1261 by X_RaY on 01 Jan, 2017 18:37
-
Due to the assumption of the impact on Q regarding small end below cutoff I did a few short sims using FEKO.
What I found is surprising but thats the result.
It implies, that Q is higher in the case with the undersized small plate.
The only differences are the cavity size itself and an adjustment of the loop antenna to get an reflection coefficient less than 0.1 (linear magnitude).
-
#1262 by Rodal on 01 Jan, 2017 18:46
-
Due to the assumption of the impact on Q regarding small end below cutoff I did a few short sims using FEKO.
In other words, the conjecture that the Shawyer cut-off rule may have been related to a lower quality factor of resonance Q when the the small end is below (what would be the) cut-off (for a constant cross section open waveguide), is proven false. The truncated cone that is below cut-off has actually higher Q than the truncated cone above cut-off.
What I found is surprising but thats the result. It implies, that Q is higher in the case with the undersized small plate.
The only differences are the cavity size itself and an adjustment of the loop antenna to get an reflection coefficient (linear magnitude) less than 0.1
However I expect that the heating of the small end plate is greater for the truncated cone that is above cut-off, as the highest intensity fields are closer to the small end plate and better able to heat (by magnetic induction) the small end plate when the truncated cone is terminated before what would be cut-off.
http://forum.nasaspaceflight.com/index.php?topic=41732.msg1624710#msg1624710 -
#1263 by X_RaY on 01 Jan, 2017 18:55
-
At the moment i am careful to state this is proven to be false.Due to the assumption of the impact on Q regarding small end below cutoff I did a few short sims using FEKO.
In other words, the conjecture that the Shawyer cut-off rule may have been related to a lower quality factor of resonance Q when the the small end is below (what would be the) cut-off (for a constant cross section open waveguide), is proven false. The truncated cone that is below cut-off has actually higher Q than the truncated cone above cut-off.
What I found is surprising but thats the result. It implies, that Q is higher in the case with the undersized small plate.
The only differences are the cavity size itself and an adjustment of the loop antenna to get an reflection coefficient (linear magnitude) less than 0.1
However I expect that the heating of the small end plate is greater for the truncated cone that is above cut-off, as the highest intensity fields are closer to the small end plate and better able to heat (by magnetic induction) the small end plate when the truncated cone is terminated before what would be cut-off.
http://forum.nasaspaceflight.com/index.php?topic=41732.msg1624710#msg1624710
This calculations are based on a relative coarse mesh and somewhat more than 20 single runs to get almost impedance match. This took a whole day. It must be repeated with a finer mesh and this will take lot of days. But yes at the moment it seems to be as it looks like. -
#1264 by Rodal on 01 Jan, 2017 18:59
-
It is confirmed that there is no such thing as cut-off in a truncated cone closed cavity.
At the moment i am careful to state this is proven to be false.Due to the assumption of the impact on Q regarding small end below cutoff I did a few short sims using FEKO.
In other words, the conjecture that the Shawyer cut-off rule may have been related to a lower quality factor of resonance Q when the the small end is below (what would be the) cut-off (for a constant cross section open waveguide), is proven false. The truncated cone that is below cut-off has actually higher Q than the truncated cone above cut-off.
What I found is surprising but thats the result. It implies, that Q is higher in the case with the undersized small plate.
The only differences are the cavity size itself and an adjustment of the loop antenna to get an reflection coefficient (linear magnitude) less than 0.1
However I expect that the heating of the small end plate is greater for the truncated cone that is above cut-off, as the highest intensity fields are closer to the small end plate and better able to heat (by magnetic induction) the small end plate when the truncated cone is terminated before what would be cut-off.
http://forum.nasaspaceflight.com/index.php?topic=41732.msg1624710#msg1624710
This calculations are based on a relative coarse mesh and somewhat more than 20 single runs to get almost impedance match. This taked a whole day. It must be repeated with a finer mesh and this will take lot of days. But yes at the moment it seems to be as it looks like.
Electromagnetic resonance continues, and the Q if anything appears to be a little higher. -
#1265 by Rodal on 01 Jan, 2017 19:24
-
These results are very important as they indicate that Shawyer's strange "cut-off" rule cannot be defended arguing lower Q for a more pointed truncated cone.
Two explanations for Shawyer's strange "cut-off" rule remain (can readers think of other explanations ?)
1) EM Drive as an experimental artifact due to asymmetric heating of the truncated cone: when the truncated cone is extended beyond what Shawyer calls "cut-off" there is a dead area near the small end plate and hence the small end plate cannot be as efficiently heated by induction heating.
2) The EM Drive as a real drive able to propel a spacecraft, as explained for example by WarpTech (Todd)'s theory due to asymmetric dissipation on the metal. When the truncated cone is extended beyond what Shawyer calls "cut-off" there is a dead area near the small end plate and hence the small end plate does not have as much dissipation.
What does McCulloch's theory, Minotti's theory and other theories have to say about extending the EM Drive beyond what Shawyer calls "cut-off"? -
#1266 by Monomorphic on 01 Jan, 2017 19:52
-
Due to the assumption of the impact on Q regarding small end below cutoff I did a few short sims using FEKO.
What antenna position and dimensions yielded the best impedance match? I'm curious how different it is for flat vs spherical endplates.
What I found is surprising but thats the result. It implies, that Q is higher in the case with the undersized small plate.
The only differences are the cavity size itself and an adjustment of the loop antenna to get an reflection coefficient less than 0.1 (linear magnitude). -
#1267 by Rodal on 01 Jan, 2017 20:04
-
Here is a quick explanation as to why the slightly higher Q with the truncated cone extended beyond what Shawyer calls "cut-off":Due to the assumption of the impact on Q regarding small end below cutoff I did a few short sims using FEKO.
What antenna position and dimensions yielded the best impedance match? I'm curious how different it is for flat vs spherical endplates.
What I found is surprising but thats the result. It implies, that Q is higher in the case with the undersized small plate.
The only differences are the cavity size itself and an adjustment of the loop antenna to get an reflection coefficient less than 0.1 (linear magnitude).
the quality of resonance Q is proportional to the (frequency times the) ratio of the electromagnetic energy inside the volume of the closed cavity to the electromagnetic dissipated on the surface of the cavity.
Beyond what Shawyer calls "cut-off", the electromagnetic energy inside the volume of the closed cavity stays practically constant, and so does the electromagnetic energy dissipated at the big end and at the portion of the side walls above cut-off. The main thing that changes is the amount of energy dissipated at the small end, which becomes smaller as one extends the truncated cone beyond cut-off.
Since the denominator of the Q expression becomes slightly smaller (as there is less energy dissipated at the small end), the Q becomes slightly larger, as one extends the truncated cone beyond what Shawyer calls cut-off.
See: http://forum.nasaspaceflight.com/index.php?topic=41732.msg1624968#msg1624968 -
#1268 by X_RaY on 01 Jan, 2017 20:14
-
Due to the assumption of the impact on Q regarding small end below cutoff I did a few short sims using FEKO.
What antenna position and dimensions yielded the best impedance match? I'm curious how different it is for flat vs spherical endplates.
What I found is surprising but thats the result. It implies, that Q is higher in the case with the undersized small plate.
The only differences are the cavity size itself and an adjustment of the loop antenna to get an reflection coefficient less than 0.1 (linear magnitude).
The mesh was coarse as stated above.
I guess, due to past experience, the loop must be moved into the direction of the big endplate to exclude overcoupling when a finer mesh is used. The center resonant frequency also shifts a bit. Since I only have the limited students version and I was interested in getting an overview first, I choose to try it with less precise conditions first. Now we can try to improve the results.
I know it is a lot of work and massive time consumption(to get Impedance matching), but it would be nice if you are able to confirm or reject the general result (Q difference) using a finer mesh.
Wire radius 1mm
Material:copper
I am already on the way with improved mesh, but it will take a while.
-
#1269 by Monomorphic on 01 Jan, 2017 22:50
-
I know it is a lot of work and massive time consumption(to get Impedance matching), but it would be nice if you are able to confirm or reject the general result (Q difference) using a finer mesh.
I used a finer mesh and got different results. Solutions converged after analyzing 38 frequencies. The frequency was lower with the finer mesh. Same antenna now shows over coupled and low reflection coefficient as well. Interesting results.
What were your mesh settings?
-
#1270 by demofsky on 01 Jan, 2017 22:58
-
These results are very important as they indicate that Shawyer's strange "cut-off" rule cannot be defended arguing lower Q for a more pointed truncated cone.
There has been some excellent work on this in the last day or so. Once again this forum simply amazes me!
Two explanations for Shawyer's strange "cut-off" rule remain (can readers think of other explanations ?)
1) EM Drive as an experimental artifact due to asymmetric heating of the truncated cone: when the truncated cone is extended beyond what Shawyer calls "cut-off" there is a dead area near the small end plate and hence the small end plate cannot be as efficiently heated by induction heating.
2) The EM Drive as a real drive able to propel a spacecraft, as explained for example by WarpTech (Todd)'s theory due to asymmetric dissipation on the metal. When the truncated cone is extended beyond what Shawyer calls "cut-off" there is a dead area near the small end plate and hence the small end plate does not have as much dissipation.
What does McCulloch's theory, Minotti's theory and other theories have to say about extending the EM Drive beyond what Shawyer calls "cut-off"?
The potentially very nice thing about this is that the two slightly different geometries may provide a relatively easy way to filter out the various theories out there based on whether they could predict performance differences between the two geometries. -
#1271 by X_RaY on 01 Jan, 2017 23:00
-
Exactly as expected https://forum.nasaspaceflight.com/index.php?topic=41732.msg1624985#msg1624985. I found the same issue using finer mesh. The antenna need to be shifted into the direction of the big end or make it smaller to decrease the coupling.I know it is a lot of work and massive time consumption(to get Impedance matching), but it would be nice if you are able to confirm or reject the general result (Q difference) using a finer mesh.
I used a finer mesh and got different results. Solutions converged after analyzing 38 frequencies. The frequency was lower with the finer mesh. Same antenna now shows over coupled and low reflection coefficient as well. Interesting results.
What were your mesh settings?
Lets hope the final Q result holds.
EDIT
I am calculating the long version at the moment (with HOBF and standard mesh, antenna z=7mm ) -
#1272 by Monomorphic on 01 Jan, 2017 23:14
-
Not to change the subject, but I was working on visualizing how a parabolic major end-plate might focus the RF towards the small end using different focal points. This was one of the first runs that I wanted to share. This is a TE013 spherical end-plate frustum with sidewalls removed.
-
#1273 by Rodal on 01 Jan, 2017 23:56
-
Not to change the subject, but I was working on visualizing how a parabolic major end-plate might focus the RF towards the small end using different focal points. This was one of the first runs that I wanted to share. This is a TE013 spherical end-plate frustum with sidewalls removed.
Excellent proof that there is radiation pressure on the side walls of an EM Drive at microwave frequencies, and that an EM Drive cavity is not at all like an optical cavity.
Something already known from their exact solutions, as an optical cavity has different mode shapes than the one in a microwave cavity.
Proof that it would be very dangerous for somebody to remove the side walls of an EM Drive, even one having curved ends. -
#1274 by Monomorphic on 02 Jan, 2017 01:11
-
I have also been working on refining the so-called "parabolic geometry." I think it has great potential.
-
#1275 by sanman on 02 Jan, 2017 03:35
-
I have also been working on refining the so-called "parabolic geometry." I think it has great potential.
Gee, so instead of discarding that payload faring when leaving the atmosphere, just re-use it to propel your departure stage.
-
#1276 by OnlyMe on 02 Jan, 2017 04:15
-
BTW the recent discussion re: the KISS low watt thruster is really no more than a curiosity. Even should it function, it would not be functioning at a power level that would make it useful for any of the kinds of tests that would or will be needed to attempt to figure out how any thrust is being generated. In a way it is like a flea circus, if it moves great but there is not enough going on to rule out the many issues that have been raised and re raised or even really show how or why it moves. The design is not robust enough that as a KISS experiment that the mechanism behind any positive results could be positively identified.
2mN using 8W Rf is a curiosity? NASA used 80W to generate 100uN.
So we are going from 1.2mN/kW to 250mN/kW. Best Ion Thruster that I know is around 60mN/kW. Best that SPR had reported is 326mN/kW. My experiments suggest 0.6N/kW is doable with non superconduction thrusters.
The engineering is done. It can scale as this KISS build will show. SPR engaged a team of academic and industry experts to examine, test and develop a underlying theory, which has been independently tested. It may not have been peer reviewed but the theory is sound and supports the engineering.
There is no curiosity here. What is here is a propulsion revolution.
TT it is misleading to compare your current KISS design and the EW frustum, even small changes in geometry seem to have significant impact on resonance and thus one would expect thrust. The frustums are not only of differing dimensions the equipment generating and controlling the MW introduced into the frustum are not equivalent.
Setting that aside if, you hit 2mN on the nose it would be miraculous. Getting close would even be an achievement. The problem is that by moving essentially backwards with the bookshelf design the reliability of any thrust measurements becomes unreliable. The rig becomes subject to many of the issues that have been raised in the past.
But my real point in that last paragraph quoted above was that even if it works.., rotates and thrust is measured anywhere in the neighborhood of the predicted 2mN, you haven't proven anything about how the device will function at Kw power levels. To prove that you would need to build and test a device operation a Kw levels.
However, the reason I used the terms curiosity and flea circus, is specifically because the project seems more focused on low cost mass production than producing a design that could be used to experimentally confirm the how and/or why of any observed thrust/rotation. The battery powered 8w design itself is not a bad idea. A mass produced low quality test platform for the world, is.
Redesign the build into a fully self contained device. And have it experimentally tested under the best of conditions and with the best test equipment available. Forget about a design for the masses, which makes it a flea circus curiosity, where unless you intend to sell mass procured kits, every builder will be building a frustum just that much different in one dimension or another that the results could be all over the place.
You believe you know why/how the device works. Others have different theories. No one including yourself or even Shawyer in a public manner has proven any theory of operation to be the one... Theories can accurately describe, even predict without accurately describing the how or why.., the mechanism...
General relativity (GR) is a good example. It describes and even predicts the tidal effects of gravitation, without proving the how and why. The observable in the case of GR is the tidal effects of gravitation, not the geometric description. For the EmDrive the observable is an annomolous thrust. No theory has been proven to describe the functional mechanism, resulting in thrust.
-
#1277 by TheTraveller on 02 Jan, 2017 11:21
-
TT it is misleading to compare your current KISS design and the EW frustum, even small changes in geometry seem to have significant impact on resonance and thus one would expect thrust. The frustums are not only of differing dimensions the equipment generating and controlling the MW introduced into the frustum are not equivalent.
Yup. The process is called EmDrive Engineering 101.
Get it wrong and there is little thrust generation but get it right and there is. -
#1278 by rfmwguy on 02 Jan, 2017 11:27
-
Not to change the subject, but I was working on visualizing how a parabolic major end-plate might focus the RF towards the small end using different focal points. This was one of the first runs that I wanted to share. This is a TE013 spherical end-plate frustum with sidewalls removed.
Interesting. Can't help but see the e field visual and compare it to a speculative visual on pilot waves.
-
#1279 by TheTraveller on 02 Jan, 2017 11:28
-
Sure seems to be a lack of side wall eddy current and E field in the dark narrow end.
Wonder where the photons are that normally produce those effects?
Wonder how the photons will be reflected from the small end, time and time again to build a real world thrust generating Q?
Guys sometimes believing in what software generates is a really bad idea. This is where engineer's gut feeling comes in and looking at the lack of photon activity in the small end confirms the software, as being run, is not telling the engineering truth. Your mileage may vary.
Old engineering rule. Always do a reality check when the software is showing you what looks like a non real result.
It implies, that Q is higher in the case with the undersized small plate.
