Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1804176 times)

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2589
One other thought that keeps popping in and out and how the evanescent momentum and spin vectors of the wave interacting with the Quantum Vacuum virtually homogeneous ZPF is it will no longer be homogeneous due to the everanscent's wave actions on it. It will literally warp the area of space 1/3 of the wavelength (where evanescent waves form) of the microwave harmonics from the frustum and...


This corresponds to my "feeling/imagination" actually, but unfortunately I cannot form a theory either. I also feel that if nothing is spewing out then there must some kind of rotation, so I keep thinking of a helical disturbance to the truncated cone; in your design that would be like giving your pyramid a twist perhaps. :)
Like this?


And I gottasa go.

Shell

Offline Fugudaddy

  • Member
  • Posts: 31
  • Rochester NY
  • Liked: 57
  • Likes Given: 70
So it's not the length of your frustum, but rather the resonance it generates in the cavity... ;)

There's a lot of fractal in the data and in the pictures that Aero's produced. If whatever makes this happen is more 'ratchety' instead of linear, then that makes a lot of sense.

It also makes it harder to separate noise from signal, since there's no nice straight lines anywhere.

The debate between Dr. Rodal and TheTraveller (get well soon) feels like a debate on how long a coastline is; it depends on if you use a meter stick or micrometer.

Data will tell which is the better way to measure. :)

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5919
  • Likes Given: 5261

There's a lot of fractal in the data and in the pictures that Aero's produced. If whatever makes this happen is more 'ratchety' instead of linear, then that makes a lot of sense.

...

The debate between Dr. Rodal and TheTraveller (get well soon) feels like a debate on how long a coastline is; it depends on if you use a meter stick or micrometer.
...
The fractal nature is known to be an artifact (to those of us who have written and used Finite Difference computer codes) a result of a not sufficiently-fine  finite-difference mesh (due to computer time and memory constraints) coupled with the use of Cartesian coordinates to enforce the boundary conditions for a circular geometry.   

The finite-difference method enforces the boundary conditions only at the grid points of the finite-difference mesh.  That's the origin of this fractality.  It is a complete artifact of the numerical discretization, and not a physical effect. 

The interested readers can verify this by looking at the circumferential boundaries and noticing that the fractality has its origin at the boundaries, and even when the fractality mostly dissappears in the interior, it persists at the boundaries.

To give you an analogy, it would be like representing a circumference with an hexagon (6 gridpoints), then outputting the geometry, noticing that the boundaries are lines instead of circles and wrongly inferring from that there is some intrinsic hexagon geometry in the problem.

To infer from this artificial fractality artifact that this is like a coastline fractality is a sure way to get lost in the waves of numerical artifacts and missing the physical aspect of the problem.

I hope that people are not inferring the idea of a ratchet from this numerical artifact...and this fractal artifact (due to a relatively coarse FD mesh) has nothing to do with the debate with TheTraveller concerning the incorrect use of waveguide cut-off modes for cylinders to represent cavity resonance of truncated cones either.  TheTraveller is not using a Finite-Difference mesh and neither am I, in our solutions.
« Last Edit: 06/27/2015 07:50 PM by Rodal »

Offline sghill

  • Full Member
  • ****
  • Posts: 1297
  • United States
  • Liked: 1427
  • Likes Given: 1974
I don't recall whether or not I've posted this video, so here it is. I was fooling around back in April trying to create evanescent waves when I came across this very unusual pattern. I don't know that evanescent waves are involved at all but something unusual is showing up. You tell me?

Superluminal velocity? Well, strange things do happen.
Yes, in the acoustic analogue, this would be tantamount to shock waves due to exit speeds faster than the speed of sound.  The shock waves would look like that.

Wow! That's cool and relevant. One of the issues I've been having is the phase velocity > c inside the cone, so what happens when it exists the big end? I think this video just showed me what happens.

Thank you.
Todd

Forgive the awkwardness of this question, but don't I recall that phase velocity can be >c, and we don't care about any information about it, we just want the momentum that's carried by the phase shift, and that the phase shift is set up by the waves in the EMDrive as the magnetron feeds into it?
Bring the thunder Elon!

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5919
  • Likes Given: 5261
...Forgive the awkwardness of this question, but don't I recall that phase velocity can be >c,...
Todd is correct, the phase velocity is greater than c (and the group velocity is less than c) for electromagnetic fields inside a cylindrical waveguide.  It is called the dispersion relation between frequency and the wavenumber k, known as the Brillouin diagram.

References:

1. Introduction to Solid State Physics - Kittel et. al.
2. Wave Propagation in Period Structures - Brillouin

in the diagram below, the straight line represents the velocity of light, the dispersion curve shows that the phase velocity is greater than c.  The group velocity is the tangent derivative of the dispersion curve, you can see that the group velocity is very low (near zero) near the origin and then increases asymptotically approaching c, but always being lower than c.


Introductory Physics MIT course (8.03)  Fall 2004 by Prof. Walter Lewin:

« Last Edit: 06/27/2015 08:37 PM by Rodal »

Offline sghill

  • Full Member
  • ****
  • Posts: 1297
  • United States
  • Liked: 1427
  • Likes Given: 1974
...Forgive the awkwardness of this question, but don't I recall that phase velocity can be >c,...
Todd is correct, the phase velocity is greater than c for electromagnetic fields

That's what I said. I think you reversed the order of "don't I..." which would change my meaning. :)

So, with that in mind, to reiterate again for a few of the naysayers that have popped up from time to time, we're not getting ANY information out of the >c phase velocity- just momentum.  Therefore we are NOT talking about a time machine or "window into the past" as some have snorted about derisively on the thread.
Bring the thunder Elon!

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5919
  • Likes Given: 5261
...Forgive the awkwardness of this question, but don't I recall that phase velocity can be >c,...
Todd is correct, the phase velocity is greater than c for electromagnetic fields

That's what I said. I think you reversed the order of "don't I..." which would change my meaning. :)

So, with that in mind, to reiterate again for a few of the naysayers that have popped up from time to time, we're not getting ANY information out of the >c phase velocity- just momentum.  Therefore we are NOT talking about a time machine or "window into the past" as some have snorted about derisively on the thread.

No, the discussion about the group velocity > c is also correct.  The FTL effect for the group velocity is associated with the quantum tunneling effect of the evanescent waves.

Electromagnetic fields inside a hollow cylindrical waveguide:  phase velocity > c and group velocity < 0

Evanescent waves quantum tunneling:  phase velocity < c  and group velocity > 0

See the big change that occurs from inside the cavity to the outside of the cavity, this is what Todd is addressing.

As I previously discussed, there is NO way to transmit information using the FTL group velocity tunneling of evanescent waves.

We also have:

phase velocity = vp = c / n 

group velocity =  vp + k ∂vp/∂k = c / (n + ∂n/∂ω)

where n is the refractive index , ω = 2 * Pi * frequency and k is the wavenumber

As I previously discussed the associations of information and momentum with group or phase velocity always have a context. 

Sorry ladies and gents, one is not going to make sense out of this without going through equations. 

Quote from: John Baez  http://math.ucr.edu/home/baez/physics/Relativity/SpeedOfLight/FTL.html#5
Phase velocity is the velocity of waves that have well-defined wavelengths, and it often varies as a function of the wavelength.  One can combine ("superpose") waves of different wavelengths to build a wave packet, a blob of some specified extent over which the wave disturbance is not small.  This packet does not have a well-defined wavelength, and because it usually spreads out as it travels, it doesn't have a well-defined velocity either; but it does have representative velocity, and this is called its group velocity, which will usually be less than c. 

Each of the packet's constituent wave trains travels with its own individual phase velocity, which in some instances will be greater than c.  But it is only possible to send information with such a wave packet at the group velocity (the velocity of the blob), so the phase velocity is yet another example of a speed faster than light that cannot carry a message.

In some situations, we can build a fairly exotic wave packet whose group velocity is greater than c.  Does this then constitute an example of information being sent at a speed faster than light?  It turns out that for these packets, information does not travel at the group velocity; instead, it travels at the signal velocity, which has to do with the time of arrival of the initial rise of the wave front as it reaches its destination.  You might not now be surprised to learn that the signal velocity turns out always to be less than c.



Wikipedia: Frequency dispersion in groups of gravity waves on the surface of deep water. The red dot moves with the phase velocity, and the green dots propagate with the group velocity. In this deep-water case, the phase velocity is twice the group velocity. The red dot overtakes two green dots when moving from the left to the right of the figure.
New waves seem to emerge at the back of a wave group, grow in amplitude until they are at the center of the group, and vanish at the wave group front.
For surface gravity waves, the water particle velocities are much smaller than the phase velocity, in most cases.
« Last Edit: 06/27/2015 09:22 PM by Rodal »

Offline X_RaY

  • Full Member
  • ****
  • Posts: 774
  • Germany
  • Liked: 993
  • Likes Given: 2144
Momentum vs. Wavevector in QM

http://www.pa.msu.edu/~mmoore/Lect12_Wavepacket.pdf

"Phase and Group Velocities

We can see that the phase velocity is

What does the
probability density
 look like?

We see that the center of the wavepacket
moves at the velocity

We call this the
`group velocity’

We can see that the group velocity correlates
with the velocity of a classical particle having
the same momentum"

note: for a photon m0=0 but E_photon=pc=hv
        E=mc --> m=E/c with E=hbar*omega
« Last Edit: 06/27/2015 08:49 PM by X_RaY »

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2589
...Forgive the awkwardness of this question, but don't I recall that phase velocity can be >c,...
Todd is correct, the phase velocity is greater than c for electromagnetic fields

That's what I said. I think you reversed the order of "don't I..." which would change my meaning. :)

So, with that in mind, to reiterate again for a few of the naysayers that have popped up from time to time, we're not getting ANY information out of the >c phase velocity- just momentum.  Therefore we are NOT talking about a time machine or "window into the past" as some have snorted about derisively on the thread.

No, the discussion about the group velocity > c is also correct.  The FTL effect for the group velocity is associated with the quantum tunneling effect of the evanescent waves.

Electromagnetic fields inside a hollow cylindrical waveguide:  phase velocity > c and group velocity < 0

Evanescent waves quantum tunneling:  phase velocity < c  and group velocity > 0

See the big change that occurs from inside the cavity to the outside of the cavity, this is what Todd is addressing.

As I previously discussed, there is NO way to transmit information using the FTL group velocity tunneling of evanescent waves.

We also have:

phase velocity = c / n 

group velocity = c / (n + ∂n/∂ω)

where n is the refractive index and ω = 2 * Pi * frequency .

As I previously discussed the associations of information and momentum with group or phase velocity always have a context. 

Sorry ladies and gents, one is not going to make sense out of this without going through equations.
Just dropped back to get some Effervescent waves for the party and saw your post. YES! she says Jose is very right. I see the equations as pictures, moving in 3d... but not in that order.

Have fun with it Jose. I will bore the gals at the garden party when they ask... "whacha been doing"?

Shell

Offline aero

  • Senior Member
  • *****
  • Posts: 2744
  • 92129
  • Liked: 705
  • Likes Given: 239
Uploading of 18x312 views of 9 inch cavity complete. Top level folder dated June 26, link is:

https://drive.google.com/folderview?id=0B1XizxEfB23tfmcxbUxsM0lVTGVkemVTX1RaMlZJb001NHVaUDRvYUtjS0lIbjdIcUNkX0k&usp=sharing

Enjoy.
Retired, working interesting problems

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5919
  • Likes Given: 5261
Uploading of 18x312 views of 9 inch cavity complete. Top level folder dated June 26, link is:

https://drive.google.com/folderview?id=0B1XizxEfB23tfmcxbUxsM0lVTGVkemVTX1RaMlZJb001NHVaUDRvYUtjS0lIbjdIcUNkX0k&usp=sharing

Enjoy.

CONGRATULATIONS.  You are getting there.

Yes, now we can see that you clearly excited a quadrapole.  Now the only thing left to understand this is to get the Max Min numerical values for the ranges plotted.  You have excited a TX21p mode  (where X is either E or M).  It is supposed to be TM212 but I'm not sure.

Since no field is obviously zero, and they all show colors and contours obviously what is being plotted is the Max Min range in each case.

For example, some of the plots may be plotting +10,000 to -10,000 while other plots may be plotting +0.00000001 to  -0.00000001. They all look the same because there is no telling what is the numerical values.
Particularly the "crazy looking" "fractal ones" are suspected to be very small values.


But the Max Min range is different from image to image, so that one cannot distinguish what is a large magnitude and what is not.  I'll look at it again later to see whether I can tell something further.
« Last Edit: 06/27/2015 09:24 PM by Rodal »

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1229
  • Do it!
  • Vista, CA
  • Liked: 1296
  • Likes Given: 1752
Regarding resonant modes in the frustum. Would there still be resonant modes if;

1. the frustum were open on the big end only?

2. the frustum were open on both ends?

IF I understand these modes correctly, the TExx0 modes resonate with the pointing vector radial in/out-ward from the axis to the walls. It seems to me, that a cone that is open on both ends would still support the same TEnm modes, just not the p modes. Correct?

If it's closed at the small end, it should still support odd harmonics of p modes. Correct?

Thank you!
Todd

It would support TExx modes but they would be travelling, out the end of the device. It wouldn't be a energy store like normal closed cavities, but simply an antenna.

https://en.wikipedia.org/wiki/Horn_antenna

Is it possible in a cylinder (not a cone) to have a TExx mode that does not travel? It just resonates at one particular spot in the cylinder? Since this resonant mode is radial, not axial, I don't see why it would have any reason to move in either direction. c is in the radial direction. v in the axial direction should be 0.
Todd

So, what happens when you have a short Horn antenna with a really bad impedance mismatch at the open end? Waves are reflected at the opening, but there is no material there to absorb the recoil. Correct? So could we still be able to establish resonance at the TE011 mode, using the VSWR as the p-mode? In other words, the reflecting surface is the vacuum. Or am I missing something again?
Todd



Offline aero

  • Senior Member
  • *****
  • Posts: 2744
  • 92129
  • Liked: 705
  • Likes Given: 239
@Rodal

Thanks.

I have considered the issue of max/min changing from image to image with and across the views. h5topng does have a switch to use the same max/min value within a data set but that leaves two problems.

1 - The Max/Min values automatically chosen are invariably from the later cycles, because the power increases as the cavity starts to resonate strongly. That means the first 100 - 200 images within a view are nearly always featureless with perhaps the antenna showing. Not very useful for showing start-up transients.

2 - As there are 18 data sets, there would still be 18 max/min ranges, so we still couldn't compare values between data sets anyway.

Still, if we didn't need to see the complete evolution, I could run and plot only the final 1 or 2 cycles (10 - 20 images). This would reduce the upload time to a reasonable value and make the Max/Min range across the data set sufficiently uniform to justify plotting with fixed Max/Min values, per data set. That would also make the task of reducing the data with MatLab to show the RSS of the E and H field components a lot less unwieldy.

The above doesn't consider the transfer of the raw data across the Internet. That is intractable with the full set of raw data as it requires 8 hours to upload one raw data field component as it stands now. That's 48 hours to upload all of the raw data, not to mention the time required to download same. Using only one full cycle should reduce that time to below 2 hours which becomes workable. Of course using only a single time point from each field component would make it quick. The question becomes, "Which time point?" so it seems that, "Use one full cycle," is the best answer.

Question: Does the data need to be so dense as 10 images per cycle, or would 8/cycle or even 4/cycle be sufficient?
« Last Edit: 06/27/2015 10:17 PM by aero »
Retired, working interesting problems

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5919
  • Likes Given: 5261
Thanks.

I have considered the issue of max/min changing from image to image with and across the views. h5topng does have a switch to use the same max/min value within a data set but that leaves two problems.

1 - The Max/Min values automatically chosen are invariably from the later cycles, because the power increases as the cavity starts to resonate strongly. That means the first 100 - 200 images within a view are nearly always featureless with perhaps the antenna showing. Not very useful for showing start-up transients.

2 - As there are 18 data sets, there would still be 18 max/min ranges, so we still couldn't compare values between data sets anyway.

Still, if we didn't need to see the complete evolution, I could run and plot only the final 1 or 2 cycles (10 - 20 images). This would reduce the upload time to a reasonable value and make the Max/Min range across the data set sufficiently uniform to justify plotting with fixed Max/Min values, per data set. That would also make the task of reducing the data with MatLab to show the RSS of the E and H field components a lot less unwieldy.

The above doesn't consider the transfer of the raw data across the Internet. That is intractable with the full set of raw data as it requires 8 hours to upload one raw data field component as it stands now. That's 48 hours to upload all of the raw data, not to mention the time required to download same. Using only one full cycle should reduce that time to below 2 hours which becomes workable. Of course using only a single time point from each field component would make it quick. The question becomes, "Which time point?" so it seems that, "Use one full cycle," is the best answer.

I definitely (No doubt about it) prefer to use the feature that:  <<h5topng does have a switch to use the same max/min value within a data set but that leaves two problems.

1 - The Max/Min values automatically chosen are invariably from the later cycles, because the power increases as the cavity starts to resonate strongly. That means the first 100 - 200 images within a view are nearly always featureless with perhaps the antenna showing. Not very useful for showing start-up transients.>>

That's PERFECT.

A number of readers of this thread unfamilar with FD methods are getting the WRONG impression that the field is fractal and that the transient is important.

If the transient is negligible (as expected) and the fractal nature is an artifact of the FD method plotting very small values that are physically zero, so much the better.

The FD method cannot exactly satisfy the boundary conditions so you never get a perfect zero, you get a very small value.  Much better to plot it as a zero, particularly when people in this thread are getting confused.

Yes no doubt about it, from now on:  << switch to use the same max/min value within a data >>

Concerning comparing values between different data sets, MEEP allows you to output NUMERICAL values (the old fashioned way).  You could output some numerical values for the very LAST time step at one of the particular data sets, to ascertain the numerical relationship between the data sets.
« Last Edit: 06/27/2015 10:20 PM by Rodal »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5919
  • Likes Given: 5261
...
So, what happens when you have a short Horn antenna with a really bad impedance mismatch at the open end? Waves are reflected at the opening, but there is no material there to absorb the recoil. Correct? So could we still be able to establish resonance at the TE011 mode, using the VSWR as the p-mode? In other words, the reflecting surface is the vacuum. Or am I missing something again?
Todd
The classical issue of SWR is due to mismatch between a load impedance (the antenna) and the transmission line.  The standing wave takes place in the feed line and not in the antenna (as part of the forward wave sent toward the load (the antenna) is reflected back along the transmission line towards the source).   The antenna acts a a resistive load.  The impedance mismatch is between the antenna and the feed line, instead of between the antenna and the vacuum.  The standing wave is formed in the feed line instead of getting formed in the antenna.

My understanding is that there's no mismatch with the vacuum or with air because the vacuum or air are not resistive loads and therefore there is no need to match their impedance. Impedance mismatch implies that energy is reflected somehow. But if you hook your 50 Ohm source up to a 50 Ohm antenna (that is sitting in air) you will see a VSWR=1, or no reflection. The power radiates away, and there is no spot for reflection (or impedance mismatch) to occur.

JR calling Notsosureofit: ring, ring, ring. 

Notsosureofit knows much more about this than I do, so I look forward to Notsosureofit's comments regarding impedance mismatch between a microwave antenna and the vacuum and standing waves produced in the antenna as a result of this impedance mismatch.

« Last Edit: 06/27/2015 11:06 PM by Rodal »

Offline Notsosureofit

  • Full Member
  • ****
  • Posts: 656
  • Liked: 704
  • Likes Given: 1364
...
So, what happens when you have a short Horn antenna with a really bad impedance mismatch at the open end? Waves are reflected at the opening, but there is no material there to absorb the recoil. Correct? So could we still be able to establish resonance at the TE011 mode, using the VSWR as the p-mode? In other words, the reflecting surface is the vacuum. Or am I missing something again?
Todd
The classical issue of SWR is due to mismatch between a load impedance (the antenna) and the transmission line.  The standing wave takes place in the feed line and not in the antenna (as part of the forward wave sent toward the load (the antenna) is reflected back along the transmission line towards the source).   The antenna acts a a resistive load.  The impedance mismatch is between the antenna and the feed line, instead of between the antenna and the vacuum.  The standing wave is formed in the feed line instead of getting formed in the antenna.

My understanding is that there's no mismatch with the vacuum or with air because the vacuum or air are not resistive loads and therefore there is no need to match their impedance. Impedance mismatch implies that energy is reflected somehow. But if you hook your 50 Ohm source up to a 50 Ohm antenna (that is sitting in air) you will see a VSWR=1, or no reflection. The power radiates away, and there is no spot for reflection (or impedance mismatch) to occur.

JR calling Notsosureofit: ring, ring, ring. 

Notsosureofit knows much more about this than I do, so I look forward to Notsosureofit's comments regarding impedance mismatch between a microwave antenna and the vacuum and standing waves produced in the antenna as a result of this impedance mismatch.

Boy, that's going way back.  Seems we always looked at an antenna as a matching transformer between the line and the impedance of free space.  (the antennas on sounding rockets, we would try to match or measure the ionosphere)  Anyway, the reaction momentum would be to the last free carrier that acted as the radiation oscillator.  Goes back to Plank, I believe.

Offline aero

  • Senior Member
  • *****
  • Posts: 2744
  • 92129
  • Liked: 705
  • Likes Given: 239
Thanks.

I have considered the issue of max/min changing from image to image with and across the views. h5topng does have a switch to use the same max/min value within a data set but that leaves two problems.

1 - The Max/Min values automatically chosen are invariably from the later cycles, because the power increases as the cavity starts to resonate strongly. That means the first 100 - 200 images within a view are nearly always featureless with perhaps the antenna showing. Not very useful for showing start-up transients.

2 - As there are 18 data sets, there would still be 18 max/min ranges, so we still couldn't compare values between data sets anyway.

Still, if we didn't need to see the complete evolution, I could run and plot only the final 1 or 2 cycles (10 - 20 images). This would reduce the upload time to a reasonable value and make the Max/Min range across the data set sufficiently uniform to justify plotting with fixed Max/Min values, per data set. That would also make the task of reducing the data with MatLab to show the RSS of the E and H field components a lot less unwieldy.

The above doesn't consider the transfer of the raw data across the Internet. That is intractable with the full set of raw data as it requires 8 hours to upload one raw data field component as it stands now. That's 48 hours to upload all of the raw data, not to mention the time required to download same. Using only one full cycle should reduce that time to below 2 hours which becomes workable. Of course using only a single time point from each field component would make it quick. The question becomes, "Which time point?" so it seems that, "Use one full cycle," is the best answer.

I definitely (No doubt about it) prefer to use the feature that:  <<h5topng does have a switch to use the same max/min value within a data set but that leaves two problems.

1 - The Max/Min values automatically chosen are invariably from the later cycles, because the power increases as the cavity starts to resonate strongly. That means the first 100 - 200 images within a view are nearly always featureless with perhaps the antenna showing. Not very useful for showing start-up transients.>>

That's PERFECT.

A number of readers of this thread unfamilar with FD methods are getting the WRONG impression that the field is fractal and that the transient is important.

If the transient is negligible (as expected) and the fractal nature is an artifact of the FD method plotting very small values that are physically zero, so much the better.

The FD method cannot exactly satisfy the boundary conditions so you never get a perfect zero, you get a very small value.  Much better to plot it as a zero, particularly when people in this thread are getting confused.

Yes no doubt about it, from now on:  << switch to use the same max/min value within a data >>

Concerning comparing values between different data sets, MEEP allows you to output NUMERICAL values (the old fashioned way).  You could output some numerical values for the very LAST time step at one of the particular data sets, to ascertain the numerical relationship between the data sets.

Would this
Quote
Given a direction constant, and a meep::volume*, returns the flux (the integral of \Re [\mathbf{E}^* \times \mathbf{H}]) in that volume.
be a useful data point? I believe it is saying Real {E* x H}

here: http://ab-initio.mit.edu/wiki/index.php/Meep_Reference#Field_computations

Maybe you will see other output features there that you like. I can see if my version of meep is current enough to use them.
Retired, working interesting problems

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5919
  • Likes Given: 5261
...

Would this
Quote
Given a direction constant, and a meep::volume*, returns the flux (the integral of \Re [\mathbf{E}^* \times \mathbf{H}]) in that volume.
be a useful data point? I believe it is saying Real {E* x H}

here: http://ab-initio.mit.edu/wiki/index.php/Meep_Reference#Field_computations

Maybe you will see other output features there that you like. I can see if my version of meep is current enough to use them.

Oh, yes, definitely !!!

That's the Poynting vector , that 1) would get around the issues of having to deal with components; and 2) would show the energy flux, the Poynting vector, which is what everybody is interested in

---------------
NOTE: It looks to me that you excited TE212.  I base this on the strong axial magnetic component, showing p=2. The problem is that the resonant frequency of TE212 is only 1.89954 GHz.  TE213 is much closer to 2.45 GHz frequency, since TE213 has 2.38836 GHz frequency, but TE213 implies p=3 and I only see p=2 in the magnetic mode.

I recall that you said that you put the antenna to excite an electric mode, that is consistent with exciting TE212.  An idea would be to run everything the same, (L=9 inches) but with the antenna set to excite a magnetic mode, to see whether you excite TM212 at 2.45 GHz or so.
« Last Edit: 06/28/2015 12:15 AM by Rodal »

Offline aero

  • Senior Member
  • *****
  • Posts: 2744
  • 92129
  • Liked: 705
  • Likes Given: 239
...

Would this
Quote
Given a direction constant, and a meep::volume*, returns the flux (the integral of \Re [\mathbf{E}^* \times \mathbf{H}]) in that volume.
be a useful data point? I believe it is saying Real {E* x H}

here: http://ab-initio.mit.edu/wiki/index.php/Meep_Reference#Field_computations

Maybe you will see other output features there that you like. I can see if my version of meep is current enough to use them.

Oh, yes, definitely !!!

That's the Poynting vector , that 1) would get around the issues of having to deal with components; and 2) would show the energy flux, the Poynting vector, which is what everybody is interested in

---------------
NOTE: It looks to me that you excited TE212.  I base this on the strong axial magnetic component, showing p=2. The problem is that the resonant frequency of TE212 is only 1.89954 GHz.  TE213 is much closer to 2.45 GHz frequency, since TE213 has 2.38836 GHz frequency, but TE213 implies p=3 and I only see p=2 in the magnetic mode.

I recall that you said that you put the antenna to excite an electric mode, that is consistent with exciting TE212.  An idea would be to run everything the same, (L=9 inches) but with the antenna set to excite a magnetic mode, to see whether you excite TM212 at 2.45 GHz or so.

That's easy enough to do, but I'm only going to output the final 12 data sets. That will assure a full cycle of data and give me a break on the excessive upload times. I'll see about outputting the Poynting vector, but next. It always takes a lot longer than it should to learn to use a new feature in meep. (I'm getting better though)
Retired, working interesting problems

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1229
  • Do it!
  • Vista, CA
  • Liked: 1296
  • Likes Given: 1752
...
So, what happens when you have a short Horn antenna with a really bad impedance mismatch at the open end? Waves are reflected at the opening, but there is no material there to absorb the recoil. Correct? So could we still be able to establish resonance at the TE011 mode, using the VSWR as the p-mode? In other words, the reflecting surface is the vacuum. Or am I missing something again?
Todd
The classical issue of SWR is due to mismatch between a load impedance (the antenna) and the transmission line.  The standing wave takes place in the feed line and not in the antenna (as part of the forward wave sent toward the load (the antenna) is reflected back along the transmission line towards the source).   The antenna acts a a resistive load.  The impedance mismatch is between the antenna and the feed line, instead of between the antenna and the vacuum.  The standing wave is formed in the feed line instead of getting formed in the antenna.

My understanding is that there's no mismatch with the vacuum or with air because the vacuum or air are not resistive loads and therefore there is no need to match their impedance. Impedance mismatch implies that energy is reflected somehow. But if you hook your 50 Ohm source up to a 50 Ohm antenna (that is sitting in air) you will see a VSWR=1, or no reflection. The power radiates away, and there is no spot for reflection (or impedance mismatch) to occur.

JR calling Notsosureofit: ring, ring, ring. 

Notsosureofit knows much more about this than I do, so I look forward to Notsosureofit's comments regarding impedance mismatch between a microwave antenna and the vacuum and standing waves produced in the antenna as a result of this impedance mismatch.

Boy, that's going way back.  Seems we always looked at an antenna as a matching transformer between the line and the impedance of free space.  (the antennas on sounding rockets, we would try to match or measure the ionosphere)  Anyway, the reaction momentum would be to the last free carrier that acted as the radiation oscillator.  Goes back to Plank, I believe.

Sort of what I was thinking. The wave doesn't reflect off the impedance of free space (vacuum), it is the current flowing in the conductor that reflects from the end of the conductor.

Tags: