Author Topic: EM Drive Developments - related to space flight applications - Thread 11  (Read 646693 times)

Offline RERT

Monomorphic - I recall you being negative about these results, ascribing the measured force to thermal effects. But the thermal signal is present and rising when the force measurement is solidly zero. The temperature is also stable when measured force is rising.

Am I mis-characterising your position, or if not can you comment on why you see thermal effects as the most likely explanation of this data.

Offline meberbs

  • Senior Member
  • *****
  • Posts: 3096
  • Liked: 3379
  • Likes Given: 777
Monomorphic - I recall you being negative about these results, ascribing the measured force to thermal effects. But the thermal signal is present and rising when the force measurement is solidly zero. The temperature is also stable when measured force is rising.

Am I mis-characterising your position, or if not can you comment on why you see thermal effects as the most likely explanation of this data.
Thermal effects can take time to travel from the thermal source to the location that causes false thrust measurements. They can also continue and increase after power is turned off as heat continues to spread out. Along with the slow rise of this measurement, that is 3 effects that can potentially be explained as thermal. None of those effects are expected measurements from a working emDrive (Except possibly the slow rise, but only if the torsion pendulum is overdamped, which at least is not the intention, and even then should have a faster start before slowing.)

Offline Monomorphic

  • Full Member
  • ****
  • Posts: 1729
  • United States
  • Liked: 4389
  • Likes Given: 1407
Monomorphic - I recall you being negative about these results, ascribing the measured force to thermal effects. But the thermal signal is present and rising when the force measurement is solidly zero. The temperature is also stable when measured force is rising.

Am I mis-characterising your position, or if not can you comment on why you see thermal effects as the most likely explanation of this data.

No, you are correct, but one of the important traces is the vertical light gray. That is main power on/off to the amplifier board. If you watch the video (linked below), you notice that power was turned on way before the RF, so the idle amplifier board was drawing ~8A and beginning to heat up. Once the RF is present, the board draws ~13A and heats up even more. The new test procedure eliminates all of this and simply turns the power and RF on at the same time.

« Last Edit: 08/06/2018 09:26 am by Monomorphic »

Offline Ricvil

  • Full Member
  • *
  • Posts: 171
  • Liked: 110
  • Likes Given: 71
Ps: The Tx3xx "mode" is suspect in this theory.

In addition to Tx3xx, perhaps TM11x could also be a candidate. Using the same naming convention, I would expect TM11x could have been labelled Tx11x.

Yep.
But in "Tx3xx" case, TM and TE visual distinction  is much more clear.
Perhaps, this clear "visual distinction" may be an artifice.
These graphs are eigensolutions of electromagnetic equations, and if there are two eigenmodes per frequency ( degenerated) then that eigensolutions may be a linear combination  of TE and TM localized modes , with arbitrary weights(or arbitrary orientation on subspace spanned by the degenerated eigenvectors).
When degenerated states arises in a eigenproblem, in general, there is a additional linear operator where the degenerated states has different eigenvalues for each eigenvector.
So, what would be the operator for differentiate TE states from TM states?
The answer may be a "duality/chirality" generator in some spinnor representation of electromagnetic fields (see  the attached  article).
« Last Edit: 08/06/2018 07:15 pm by Ricvil »

Offline RERT

My iPad mIni combined with my eyesight  isn't up to viewing your video, so I'll respond more directly when back off vacation 13/8.

In general 'it's getting hot, that must be the cause' isn't particularly convincing. I would be more convinced by the opposite of your approach. Make the board draw 13A with or without RF on. Run for a long time until the system is in thermal equilibrium. Then show that RF on/ off makes no difference - or not, as the case may be.

Offline X_RaY

  • Full Member
  • ****
  • Posts: 852
  • Germany
  • Liked: 1146
  • Likes Given: 2479
Ps: The Tx3xx "mode" is suspect in this theory.

In addition to Tx3xx, perhaps TM11x could also be a candidate. Using the same naming convention, I would expect TM11x could have been labelled Tx11x.

Yep.
But in "Tx3xx" case, TM and TE visual distinction  is much more clear.
Perhaps, this clear "visual distinction" may be an artifice.
These graphs are eigensolutions of electromagnetic equations, and if there are two eigenmodes per frequency ( degenerated) then that eigensolutions may be a linear combination  of TE and TM localized modes , with arbitrary weights(or arbitrary orientation on subspace spanned by the degenerated eigenvectors).
When degenerated states arises in a eigenproblem, in general, there is a additional linear operator where the degenerated states has different eigenvalues for each eigenvector.
So, what would be the operator for differentiate TE states from TM states?
The answer may be a "duality/chirality" generator in some spinnor representation of electromagnetic fields (see  the attached  article).
This is not a classic degenerated state of two field patterns with their own solutions at the same frequency. These are patterns that are only present in the frustum of the cone due to the topology. The pattern on the end plate of a cylindrical version is located on the side wall of the conical shape.


By the way this is one of only a few things where i can not follow Frank Davies with its mode index denotation.
Due to the pattern in the Cylindrical case and only 2 wave length into phi direction i would label it as a deformed TM210 mode rather than, kind of cryptic, Tx3xx.
« Last Edit: 08/06/2018 08:51 pm by X_RaY »

Offline Ricvil

  • Full Member
  • *
  • Posts: 171
  • Liked: 110
  • Likes Given: 71
Ps: The Tx3xx "mode" is suspect in this theory.

In addition to Tx3xx, perhaps TM11x could also be a candidate. Using the same naming convention, I would expect TM11x could have been labelled Tx11x.

Yep.
But in "Tx3xx" case, TM and TE visual distinction  is much more clear.
Perhaps, this clear "visual distinction" may be an artifice.
These graphs are eigensolutions of electromagnetic equations, and if there are two eigenmodes per frequency ( degenerated) then that eigensolutions may be a linear combination  of TE and TM localized modes , with arbitrary weights(or arbitrary orientation on subspace spanned by the degenerated eigenvectors).
When degenerated states arises in a eigenproblem, in general, there is a additional linear operator where the degenerated states has different eigenvalues for each eigenvector.
So, what would be the operator for differentiate TE states from TM states?
The answer may be a "duality/chirality" generator in some spinnor representation of electromagnetic fields (see  the attached  article).
This is not a classic degenerated state of two field patterns with their own solutions at the same frequency. These are patterns that are only present in the frustum of the cone due to the topology. The pattern on the end plate of a cylindrical version is located on the side wall of the conical shape.

Yep, this is not the classic.
My  claim is: these are two degenerated modes (same frequency) localized at two different points(each one at neighborhood of each flat endplate).
I think they are two ghost modes, one TE and other TM.

Offline X_RaY

  • Full Member
  • ****
  • Posts: 852
  • Germany
  • Liked: 1146
  • Likes Given: 2479
Ps: The Tx3xx "mode" is suspect in this theory.

In addition to Tx3xx, perhaps TM11x could also be a candidate. Using the same naming convention, I would expect TM11x could have been labelled Tx11x.

Yep.
But in "Tx3xx" case, TM and TE visual distinction  is much more clear.
Perhaps, this clear "visual distinction" may be an artifice.
These graphs are eigensolutions of electromagnetic equations, and if there are two eigenmodes per frequency ( degenerated) then that eigensolutions may be a linear combination  of TE and TM localized modes , with arbitrary weights(or arbitrary orientation on subspace spanned by the degenerated eigenvectors).
When degenerated states arises in a eigenproblem, in general, there is a additional linear operator where the degenerated states has different eigenvalues for each eigenvector.
So, what would be the operator for differentiate TE states from TM states?
The answer may be a "duality/chirality" generator in some spinnor representation of electromagnetic fields (see  the attached  article).
This is not a classic degenerated state of two field patterns with their own solutions at the same frequency. These are patterns that are only present in the frustum of the cone due to the topology. The pattern on the end plate of a cylindrical version is located on the side wall of the conical shape.

Yep, this is not the classic.
My  claim is: these are two degenerated modes (same frequency) localized at two different points(each one at neighborhood of each flat endplate).
I think they are two ghost modes, one TE and other TM.
Not at all. Maybe we're just looking at the problem from different angles, but two different modes, TM & TE, would change their eigenfrequencies differently, while reducing the small end plate. I.E. the eigenfrequencies of different modes (one TE and another TM) would shift to different values when reducing the diameter of the small end plate, even if they lay at the same frequency for a special shape.  Regarding the simulations is this not the case.
To me it seems a pure geometrical property, a deformation of the field due to the very shape (and related to the boundary conditions) of the frustum as compared to the cylindrical cavity.
« Last Edit: 08/06/2018 09:32 pm by X_RaY »

Offline Ricvil

  • Full Member
  • *
  • Posts: 171
  • Liked: 110
  • Likes Given: 71
Ps: The Tx3xx "mode" is suspect in this theory.

In addition to Tx3xx, perhaps TM11x could also be a candidate. Using the same naming convention, I would expect TM11x could have been labelled Tx11x.

Yep.
But in "Tx3xx" case, TM and TE visual distinction  is much more clear.
Perhaps, this clear "visual distinction" may be an artifice.
These graphs are eigensolutions of electromagnetic equations, and if there are two eigenmodes per frequency ( degenerated) then that eigensolutions may be a linear combination  of TE and TM localized modes , with arbitrary weights(or arbitrary orientation on subspace spanned by the degenerated eigenvectors).
When degenerated states arises in a eigenproblem, in general, there is a additional linear operator where the degenerated states has different eigenvalues for each eigenvector.
So, what would be the operator for differentiate TE states from TM states?
The answer may be a "duality/chirality" generator in some spinnor representation of electromagnetic fields (see  the attached  article).
This is not a classic degenerated state of two field patterns with their own solutions at the same frequency. These are patterns that are only present in the frustum of the cone due to the topology. The pattern on the end plate of a cylindrical version is located on the side wall of the conical shape.

Yep, this is not the classic.
My  claim is: these are two degenerated modes (same frequency) localized at two different points(each one at neighborhood of each flat endplate).
I think they are two ghost modes, one TE and other TM.
Not at all. Maybe we're just looking at the problem from different angles, but two different modes, TM & TE, would change their eigenfrequencies differently, while reducing the small end plate. I.E. the eigenfrequencies of different modes (one TE and another TM) would shift to different values when reducing the diameter of the small end plate, even if they lay at the same frequency for a special shape.  Regarding the simulations is this not the case.
To me it seems a pure geometrical property, a deformation of the field due to the very shape (and related to the boundary conditions) of the frustum as compared to the cylindrical cavity.
Ghost modes arise by "local shift" of original (or undisturbed )mode cutoff frequency, and their Q, total frequency shift, and spacial extension depends on how large is the effect of "deformation/pertubation" causing it, and for me this pertubations are just the flat endplates, or better, the difference of shape/volume between the use of spherical endplates and flat endplates.
In other words, TE and TM ghost modes may coexist at a range of frequencys(or a range of small plate diameter as showed in yours simulations).
The cylindrical cavity may be thought as a limit case of conical cavity with spherical endplates when the apex point goes to infinity, and in this case there are  no pertubation causing ghost modes, just the classical standing  wave solutions.
« Last Edit: 08/07/2018 03:50 am by Ricvil »

Offline Bob Woods

  • Full Member
  • ***
  • Posts: 391
  • Salem, Oregon USA
  • Liked: 529
  • Likes Given: 1579
The Google Doodle today:

"Mary Golda Ross (August 9, 1908 – April 29, 2008) was the first known female engineer. She was one of the 40 founding engineers of the Skunk Works, and was known for her work at Lockheed on "preliminary design concepts for interplanetary space travel, manned and unmanned earth-orbiting flights, the earliest studies of orbiting satellites for both defense and civilian purposes."

Seemed kind of right for this forum to me. Women are woefully underrepresented in STEM. Think how much brain power can be unleashed in the sciences if the research had more participants.

https://en.wikipedia.org/wiki/Mary_G._Ross

Offline Monomorphic

  • Full Member
  • ****
  • Posts: 1729
  • United States
  • Liked: 4389
  • Likes Given: 1407
I've finished the simulations on the solid copper cavity Oyzw sent me from China. I am going to use an open-ended half loop antenna initially since it excites mode TE013, is very easy to fabricate, and is simple to tune for impedance. The location is 4cm above the bottom of the frustum. Simulated Q factor was ~32,000. I ended up needing a second much thinner copper gasket on the small end as under close inspection, the two flanges are not perfectly parallel. There is a ~1.5mm - 2mm difference on the small end - like a carrot that has the fat end sliced flat, but the narrow end has a slight angle cut. The two gaskets together bring the resonant frequency to 2.416Ghz, safely within the ISM band. Any compressing of the gaskets will bring the frequency higher.

I also wanted to report that I've been extended, and have accepted, another invitation to present at the upcoming Advanced Propulsion Workshop in Estes Park, Colorado. They were very interested in some of my recent simulations and tests of acoustic devices on the torsional pendulum. So I will be presenting on the torsional pendulum, my work with the emdrive and how that lead me to the acoustic tests, those results, and my plans for the future.  I will also get to meet Jose Rodal, Jim Woodward, Heidi Fearn, Martin Tajmar and many more. Can't wait, but that also means I have lots of work to get done in ~4 weeks! The flight, room, and car are all booked, so there's no backing out this time...   ;) 

« Last Edit: 08/10/2018 01:16 am by Monomorphic »

Offline Ricvil

  • Full Member
  • *
  • Posts: 171
  • Liked: 110
  • Likes Given: 71
The Traveller,

According to the two previous pages, it seems that:
1) you still base your understanding of the propellantless propulsion effect of the EmDrive in the same origin as Shawyer's, i.e. the existence of a force resulting from a non-zero sum of all radiation pressures upon materials within the cavity.
but:
2) you however now refute Shawyer's claim that the radiation pressure is greater at the big end, saying it would be the opposite: that the radiation pressures on side walls + small end combined are greater than the radiation pressure on the wide end, resulting in the EmDrive being pushed by this forward radiation pressure, small end leading. So no more invisible  "thrust force" directed in the opposite, rear direction without matter ejected, that Shawyer yet introduced to try to mimic his system with classical Newtonian action-reaction.

You argue based on the momentum exchange with all walls and the photon incident angle varying across the tapered section.

Shawyer bases his "EmDrive theory" on Cullen's experiments and his 1952 paper, extrapolating measurement made with open cylindrical waveguides to tapered closed cavities, since he assumes that a closed tapered cavity is the same as a series of many shallow cylindrical open waveguides of decreasing diameter connected the one after the others (from the point of view of travelling waves, hence a pulsed operation).

Therefore Shawyer claims that the radiation pressure (and the group velocity) of microwaves is greater on the big end of the EmDrive than on the small end, which seems sound, but doing so he may neglect the wall component, which should add and sum up to zero (he claims this zero sum is indeed the case for a standing wave, but not for travelling waves).

Cullen showed (eq. 15 in his paper) that:
F = 2P/c ( λ / λg )

Since λ < λg (always) and the smaller the waveguide diameter, the longer the guide wavelength λg, it is easy to show that the force due to the radiation pressure of microwaves at the same input power acting on a plate in a wider waveguide is greater than the force acting on a plate in a narrow waveguide.

So do you now disagree with Cullen; or do you agree with him but saying instead that what is going on in open cylindrical waveguides cannot be extrapolated to closed tapered cavities?
You should take the energy density per area into account. According to the work of Dr. Rodal we know that the field strength in the area of the smaller end plate is much larger than at the bigger plate. However, the total amount of incident power at the small end plate plus the equivalent vector component at conical sidewall should be the same per area unit squared, -F (small end plus sidewall vector component in this direction) +F (at the large plate), ...from a pure topological point of view.

https://forum.nasaspaceflight.com/index.php?action=dlattach;topic=37642.0;attach=1030954

However I think for traveling (reflected) waves there is a time related difference related to the reflection on both ends. I guess the reflection at the smaller side has a broad band characteristic compared to the big end. I.e. the wave is partly reflected before it reaches the small plate (partially earlier times). If the big end is flat there is also a phase dependent time dependent reflection involved. But for a proper curved big plate and a small end below cutoff the time difference of the reflected signal should be located at the small end. So maybe a time-delayed reflection of the incident wave at a undersized small end combined with a spherical big end plate leads to a nice net force because of the time delayed reflection at one end only?

Just a thought..  ::)
It is hard to think about such problems while the room temperature is still way over 30°C/86°F  :-\ :-[
I think this pattern can be explained by a Fano anti-resonance caused  by a higher Q ghost mode at small flat endplate interacting with a lower Q standing wave resonance caused by the big spherical endplate.
It  was a time domain  simulation?
« Last Edit: 08/09/2018 10:58 pm by Ricvil »

Offline Ricvil

  • Full Member
  • *
  • Posts: 171
  • Liked: 110
  • Likes Given: 71
Fano resonances everywhere!
Perhaps the  Tx3xx mode is the TM211 after all.
The thrust occurs exactly at Fano resonances profiles, them two resonances are required, and one with Q much higher than other.
Very interesting!!

Offline X_RaY

  • Full Member
  • ****
  • Posts: 852
  • Germany
  • Liked: 1146
  • Likes Given: 2479

It  was a time domain  simulation?
No this is not a finite-difference time-domain (FDTD) analyses, it is a simulation based on boundary-element-method (BEM) using FEKO-software.
« Last Edit: 08/10/2018 03:45 pm by X_RaY »

Offline oyzw

  • Full Member
  • *
  • Posts: 177
  • Liked: 173
  • Likes Given: 1
开放式半环耦合不是最佳,采用短路闭环耦合更好,Q值超过50000

Offline Ricvil

  • Full Member
  • *
  • Posts: 171
  • Liked: 110
  • Likes Given: 71

It  was a time domain  simulation?
No this is not a finite-difference time-domain (FDTD) analyses, it is a simulation based on boundary-element-method (BEM) using FEKO-software.

The big plate was spherical or flat in that simulation?

That chaotic poynting vector over a full cycle  was a simulation transient, or that persist over a long range of time?


Offline Mark7777777

  • Full Member
  • *
  • Posts: 109
  • Liked: 66
  • Likes Given: 64
开放式半环耦合不是最佳,采用短路闭环耦合更好,Q值超过50000

https://www.google.co.nz/search?q=Google+translate+Chinese+to+English&rlz=1C9BKJA_enNZ596NZ596&hl=en-NZ&sourceid=chrome-mobile&ie=UTF-8&spknlang=en-NZ&inm=vs&vse=1

Open half-ring coupling is not optimal, short-circuit closed-loop coupling is better, Q value exceeds 50000

Offline Bob Woods

  • Full Member
  • ***
  • Posts: 391
  • Salem, Oregon USA
  • Liked: 529
  • Likes Given: 1579
开放式半环耦合不是最佳,采用短路闭环耦合更好,Q值超过50000
Jamie I think this is for you...

Open half-ring coupling is not optimal, short-circuit closed-loop coupling is better, Q value exceeds 50000
« Last Edit: 08/11/2018 04:01 am by Bob Woods »

Offline X_RaY

  • Full Member
  • ****
  • Posts: 852
  • Germany
  • Liked: 1146
  • Likes Given: 2479

It  was a time domain  simulation?
No this is not a finite-difference time-domain (FDTD) analyses, it is a simulation based on boundary-element-method (BEM) using FEKO-software.

The big plate was spherical or flat in that simulation?

That chaotic poynting vector over a full cycle  was a simulation transient, or that persist over a long range of time?
This was a numerical artifact / problem within the simulation.
https://forum.nasaspaceflight.com/index.php?topic=41732.msg1633578#msg1633578

EDIT:

I found results for the average poynting vector field that looks very interesting in a special situation. See pics

EDIT 2:
Using a small loop antenna instead the magnetic dipole the vector pattern looks very different.
« Last Edit: 08/12/2018 01:00 pm by X_RaY »

Offline Monomorphic

  • Full Member
  • ****
  • Posts: 1729
  • United States
  • Liked: 4389
  • Likes Given: 1407
开放式半环耦合不是最佳,采用短路闭环耦合更好,Q值超过50000

I am only usinging the half-loop at the beginning since it is so easy to make. The closed loop antenna will be next, but those are more difficult to fabricate.   What I will end up doing is 3D printing a stepped cone of various diameters and use that to wrap the wire around so I can get the right diameter for the antenna. I also need to order more SMA parts as I'm running low and only have the parts for one more antenna.  :-\

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
0