-
#1940 by rfmwguy on 19 Jan, 2016 12:23
-
Doc,
Definitely worthy of emdrive.wiki update as null. It might be wise to add a column characterizing rf source as cw or pulsed, narrowband or wideband. Rfplumbers setup reminds me of our aachen friends, scaled down to 2.4 ghz. Both tests appear to be narrowband CW although I've seen no spec an pics or videos.
p.s. I updated the wiki page with my stuff when you were on break...don't think anyone objects to your caretaking there. -
#1941 by TheTraveller on 19 Jan, 2016 12:31
-
...
This shows precisely what I was talking about: rampant CONFUSION about what mode is being excited.
BTW the FEKO solver shows the single end plate and 2 x sidewall current loops I believed should be there for a TE012 excitation of Shell's frustum. I seems to be a very capable solver for frustum simulations.
IslandPlaya excellent work using FEKO shows mode shape TM112, a completely different mode shape than TE012, for SeeShell experiment, even when the rectangular waveguide enters the frustum from the Big Base, as shown iin the pictures below.
The mode shapes, as we learnt at school, are determined by the shape of the electric and magnetic fields, and not by the current loops !!!
The mode shapes are IslandPlaya's work and they are correct for TE012 mode excitation of a 2.45 GHz TE012 resonant frustum.
https://drive.google.com/folderview?id=0B6juR48k_XoTb1hTaDJuTElld3c&usp=sharing&tid=0B6juR48k_XoTdWF0Z2FFZWk5MFU
And yes you are 100% correct, the mode shapes for TE013 and TM112 are very different, internally inside the frustum and the current flows they induce in the end plates and side walls.
I guess you have never considered what the current flows in the sidewalls and the end plates that different excitation modes would create? Well with the FEKO solver you can now plot the E field shapes, the H field shapes and the current flow shapes the H field induces in the end plates and the sidewalls of the frustum. Those current shapes, in the end plates and the sidewalls, should cause localised heating and be visible with a thermal camera.
Do plan to test for and show the heating bands that should occur with TE013 excitation of my frustum. With 100W of power, should be reasonable easy to make they appear and to confirm there is a high quality frustum wide TE013 excitation occurring. -
#1942 by TheTraveller on 19 Jan, 2016 12:38
-
The mode shapes, as we learnt at school, are determined by the shape of the electric and magnetic fields, and not by the current loops !!!
As any EE learns at school, the skin depth current loops are driven by / caused by the H fields entering the metal and generating current flow.
In the attached, the relationship between H field intensity and surface eddy current flow in the end plate of a resonant cavity is made very clear.
What this says is the surface eddy current loops and related heating reveal the H and E field mode shape inside the frustum. No where have I ever claimed the end plate and sidewall surface eddy current flows drive the mode shape inside the frustum.
-
#1943 by TheTraveller on 19 Jan, 2016 12:43
-
Mode shapes are defined by the Electric and Magnetic field distributions !
For sure.
And the distribution patterns / effects of those E&M fields on the end plates and sidewalls of a frustum will cause identifiable current flows in those surfaces, which will cause thermally identifiable unique signatures for each and every resonant mode.
So to confirm a good mode excitation, maybe all that is needed is a thermal image of the end plates and the sidewalls? -
#1944 by rfmwguy on 19 Jan, 2016 12:52
-
This is a good position to take doc, just be aware of the stated position the poster of the feko analysis on another forum has taken..."there is no emdrive effect" paraphrasing the poster whose qualifications and lack of experimentation are unknown. My advice is to get feko into the hands of more objective/respectable/non-anonymous people here before the waters are muddied too much with potentially falsified data analysis. Your position is quite valid but must apply equally...just my early morning thoughts anyway...fwiw....This is my pathway. I have no doubt it will be successful.
You have no doubt, ahead of time, that your test will be successful ?
The scientific method of testing is that one doesn't know ahead of time what the outcome of a future test will be. -
#1945 by TheTraveller on 19 Jan, 2016 12:57
-
1) You are attaching a picture about an open waveguide, for an open waveguide mode TE01,
while what we are discussing is the mode shape in a closed cavity. Either you don't understand the difference between the mode shapes in an open waveguide or are you trying to confuse the issue?
Please read the lower line in the image.
"Fig 7. End plate currents in TE01 mode."
It is a plot of a end plate H field and induced current in a resonant cavity. The number of 1/2 waves resonant makes no difference to the end plate H field and induced surface current so it is not mentioned.
Where do you read "Open Waveguide"?
I did post a link to the paper which contained the above image, which it seems you may have never opened:
End Plate and Side Wall Currents in Circular Cylinder Cavity Resonator
http://www.ganino.com/BSTJ/images/Vol26/bstj26-1-31.pdf -
#1946 by Rodal on 19 Jan, 2016 12:59
-
1) The FEKO analysis discussion was introduced into the thread by somebody else. I answered the confusion that has been produced by incorrect identification of mode shapes of those FEKO results (and Meep done here?) for SeeShell's frustum with waveguide feeding.
This is a good position to take doc, just be aware of the stated position the poster of the feko analysis on another forum has taken..."there is no emdrive effect" paraphrasing the poster whose qualifications and lack of experimentation are unknown. My advice is to get feko into the hands of more objective/respectable/non-anonymous people here before the waters are muddied too much with potentially falsified data analysis. Your position is quite valid but must apply equally...just my early morning thoughts anyway...fwiw....This is my pathway. I have no doubt it will be successful.
You have no doubt, ahead of time, that your test will be successful ?
The scientific method of testing is that one doesn't know ahead of time what the outcome of a future test will be.
2) The discussion with TT is about interpretation of mode shapes. TT says that they should be identified by the currents, I say that goes against all standards. The standard interpretation of mode shapes is that they are identified by the Electric and Magnetic field distributions. That is why mode shapes are known as TM (transverse magnetic field) and TE (transverse electric fields). They are not identified as TC (transverse currents)
3)While TheTraveller initiated the discussion by posting FEKO results from someone else (https://forum.nasaspaceflight.com/index.php?topic=39004.msg1479711#msg1479711), I posted my Wolfram Mathematica calculations (https://forum.nasaspaceflight.com/index.php?topic=39004.msg1479718#msg1479718).
TheTraveller has not posted a single image for a mode shape he has calculated on his own and instead he is the one that posts images calculated by somebody else in another forum, adding his own subjective interpretation. Are you going to ask TheTraveller to post his own images for the case in hand instead of posting somebody else's that you object?
-
#1947 by TheTraveller on 19 Jan, 2016 13:07
-
2) The discussion with TT is about interpretation of mode shapes. TT says that they should be identified by the currents, I say that goes against all standards. The standard interpretation of mode shapes is that they are identified by the Electric and Magnetic field distributions
I said they can also be identified by the H field induced currents and as those currents cause localised thermal heating, it may be possible, in higher powered EmDrive, to thermally identify the excited mode shape using a low cost thermal camera attachment to a phone.
What I also said was each excited mode has it's own unique end plate H field pattern, which will causes a unique current pattern on the end plate. As FEKO can plot E field, H field and surface current flow, it offers another way to identify the excited mode by just looking at the end plate induced current patterns. -
#1948 by TheTraveller on 19 Jan, 2016 13:19
-
This is a good position to take doc, just be aware of the stated position the poster of the feko analysis on another forum has taken..."there is no emdrive effect" paraphrasing the poster whose qualifications and lack of experimentation are unknown. My advice is to get feko into the hands of more objective/respectable/non-anonymous people here before the waters are muddied too much with potentially falsified data analysis. Your position is quite valid but must apply equally...just my early morning thoughts anyway...fwiw....This is my pathway. I have no doubt it will be successful.
You have no doubt, ahead of time, that your test will be successful ?
The scientific method of testing is that one doesn't know ahead of time what the outcome of a future test will be.
Dave,
FEKO is openly available and I expect will be used by a lot of EmDrive builders.
The TE012 images I posted are not incorrect. They show the correct E & H fields on the end plate for a frustum at TE012 resonance. That they also show the H field induced surface currents is to me a bonus. Others may have their own opinions and that is their right.
What is happening here is detailed frustum analysis that was once restricted to a small few, but will now be widely available to any who wish to learn FEKO, as I and Shell are doing.
Would suggest in a few weeks there will be FEKO frustum models being openly exchanged and members here and on other forums sharing FEKO tips and techniques.
I believe FEKO will leapfrog many forum members knowledge of how to build a frustum and to tell why it is not working as expected. It is a really great tool and very easy to use. Can even generate S parameters (S11 curves) and Smith charts to compare against real world VNA data. Wonderfull software.
When I get the full version download, will model your wire mesh frustum with the maggie antenna in the centre of the big end. Should be real interesting to see the data it produces. that is of it can model copper mesh. -
#1949 by rfmwguy on 19 Jan, 2016 13:47
-
Its for everyone, but your statement was what I thought was relevant to the general discussion. I've personally been on both sides of the issue and can spot predispositions fairly well as I'm sure you can. The offsite poster has been creating feko models and I believe that person has disqualified themselves as a fair and objective person numerous time. That person also admitted they were banned twice on nsf (before my time). That is my only point. I trust feko analysis done by people here, not elsewhere. You are correct is wanting to keep things objective. Now more java...
1) The FEKO analysis discussion was introduced into the thread by somebody else, as you know. I answered the confusion that has been produced by INCORRECT identification of mode shapes by incorrect interpretation of FEKO results (and Meep ?) for SeeShell's frustum with waveguide feeding.
This is a good position to take doc, just be aware of the stated position the poster of the feko analysis on another forum has taken..."there is no emdrive effect" paraphrasing the poster whose qualifications and lack of experimentation are unknown. My advice is to get feko into the hands of more objective/respectable/non-anonymous people here before the waters are muddied too much with potentially falsified data analysis. Your position is quite valid but must apply equally...just my early morning thoughts anyway...fwiw....This is my pathway. I have no doubt it will be successful.
You have no doubt, ahead of time, that your test will be successful ?
The scientific method of testing is that one doesn't know ahead of time what the outcome of a future test will be.
2) The discussion with TT is about interpretation of mode shapes. TT says that they should be identified by the currents, I say that goes against all standards. The standard interpretation of mode shapes is that they are identified by the Electric and Magnetic field distributions
3) On both occasions that FEKO analysis was introduced here, it was introduced by 2 other posters. This last time it was introduced by TheTraveller. Why do you address your post to me rather than TheTraveller, or to both of us?
4)TheTraveller has not posted a single image for a mode shape he has calculated on his own and instead he is the one that posts images calculated by somebody else in another forum? Are you going to ask TheTraveller to post his own images for the case in hand instead of posting somebody else's?
While TheTraveller initiated the discussion by posting FEKO results from someone else (https://forum.nasaspaceflight.com/index.php?topic=39004.msg1479711#msg1479711), I posted my Wolfram Mathematica calculations (https://forum.nasaspaceflight.com/index.php?topic=39004.msg1479718#msg1479718).
Instead TheTraveller posts images from FEKO done by somebody else, and to compound the problem TheTraveller has his own interpretation that goes against standards (the standard is to use the electric and magnetic fields to identify a mode shape). -
#1950 by rfmwguy on 19 Jan, 2016 13:59
-
Thanks phil...its really the offsite work/modeling that has me concerned. Think we here can objectively hammer it out. BUT, I'm speaking for myself not in a moderator role. My formal moderating duties are very few thanks to this community...so typically I post like anyone else. Hope everybody understands that. I'm just a NSF staff liason when I need to be...and that is thankfully not very often. I wouldn't have assumed the mod duties if I believed this community was full of skallywags and madhatters
-
#1951 by Rodal on 19 Jan, 2016 14:33
-
Take a look at the surface current picture from FEKO (attached below) for SeeShell experiment being fed from a waveguide (which is what was being discussed in this NSF thread and it is the relevant NSF user experiment):2) The discussion with TT is about interpretation of mode shapes. TT says that they should be identified by the currents, I say that goes against all standards. The standard interpretation of mode shapes is that they are identified by the Electric and Magnetic field distributions
I said they can also be identified by the H field induced currents and as those currents cause localised thermal heating, it may be possible, in higher powered EmDrive, to thermally identify the excited mode shape using a low cost thermal camera attachment to a phone.
What I also said was each excited mode has it's own unique end plate H field pattern, which will causes a unique current pattern on the end plate. As FEKO can plot E field, H field and surface current flow, it offers another way to identify the excited mode by just looking at the end plate induced current patterns.
https://drive.google.com/folderview?id=0B6juR48k_XoTOWJMMHRHdm1fUmc&usp=sharing
it does NOT look like the surface current distribution for mode shape TE01, it looks like the surface current distribution for TM112
in complete agreement with the Magnetic field distribution from FEKO and in agreement with my Wolfram Mathematica calculation for TM112 (a transverse magnetic mode !!!! )
This is no surprise, since I'm tired of repeating, the surface currents are eddy currents produced by the magnetic field, not by the electric field
Previously (in a discussion started by SeeShells) we showed the images for SeeShell's fustrum being fed with a waveguide from the side conical wall (instead of being fed from a waveguide on the Big Base), which also showed mode shape TM112.
-
#1952 by TheTraveller on 19 Jan, 2016 14:42
-
Take a look at the surface current picture from FEKO (attached below) for SeeShell experiment being fed from a waveguide:2) The discussion with TT is about interpretation of mode shapes. TT says that they should be identified by the currents, I say that goes against all standards. The standard interpretation of mode shapes is that they are identified by the Electric and Magnetic field distributions
I said they can also be identified by the H field induced currents and as those currents cause localised thermal heating, it may be possible, in higher powered EmDrive, to thermally identify the excited mode shape using a low cost thermal camera attachment to a phone.
What I also said was each excited mode has it's own unique end plate H field pattern, which will causes a unique current pattern on the end plate. As FEKO can plot E field, H field and surface current flow, it offers another way to identify the excited mode by just looking at the end plate induced current patterns.
https://drive.google.com/folderview?id=0B6juR48k_XoTOWJMMHRHdm1fUmc&usp=sharing
it does NOT look like the surface current distribution for mode shape TE01, it looks like the surface current distribution for TM112
in complete agreement with the Magnetic field distribution from FEKO and in agreement with my Wolfram Mathematica calculation for TM112 (a transverse magnetic mode !!!! )
We agree.
The images are showing a TM11x excitation
Attached is the Bell Systems end plate current image from the paper posted earlier.
You do note the frustum is being fed differently than in the TE01x example and that is why the TM11x mode is excited instead of the other side of the coin TE01x mode. These 2 mode are degenerative of each other and the frustum can be excited into either mode depending on how it is excited.
-
#1953 by Rodal on 19 Jan, 2016 14:45
-
Take a look at the surface current picture from FEKO (attached below) for SeeShell experiment being fed from a waveguide:2) The discussion with TT is about interpretation of mode shapes. TT says that they should be identified by the currents, I say that goes against all standards. The standard interpretation of mode shapes is that they are identified by the Electric and Magnetic field distributions
I said they can also be identified by the H field induced currents and as those currents cause localised thermal heating, it may be possible, in higher powered EmDrive, to thermally identify the excited mode shape using a low cost thermal camera attachment to a phone.
What I also said was each excited mode has it's own unique end plate H field pattern, which will causes a unique current pattern on the end plate. As FEKO can plot E field, H field and surface current flow, it offers another way to identify the excited mode by just looking at the end plate induced current patterns.
https://drive.google.com/folderview?id=0B6juR48k_XoTOWJMMHRHdm1fUmc&usp=sharing
it does NOT look like the surface current distribution for mode shape TE01, it looks like the surface current distribution for TM112
in complete agreement with the Magnetic field distribution from FEKO and in agreement with my Wolfram Mathematica calculation for TM112 (a transverse magnetic mode !!!! )
We agree.
The images are showing a TM11x excitation
Attached is the Bell Systems end plate current image from the paper posted earlier
We also agree on the advantages of FEKO !!!
It resolves all the BS, because technical people will always agree on technical facts !!
Just think about this:
1) FEKO model being discussed uses the Boundary Element Method (*) while Meep uses Finite Difference method. It is well known that Boundary Element method for a resonant cavity is much more accurate than the finite difference method (and I'm an expert on these methods).
2) FEKO shows you the Electric fields, Magnetic fields and Currents in the 3D truncated cone as a vector resultant with a number that anyone can understand
3) Compare that with the horrible mess and confusion resulting by showing 6 different components (3 for E and 3 for H) in Cartesian coordinate components without any number field by Meep solutions presented here (**), so that people cannot even tell apart what is practically zero from what is significant !!!! And people cannot do vector addition in their heads
________
(*) The German author of FEKO refers to it as "Method of Moments" , a terminology only used in Electromagnetics, while the Boundary Element Method is a numerical method of general applicability known generally as the Boundary Element Method
(**) The advantage of Meep is that it is an open code where one can write your own constitutive equations, and has nonlinear capabilities (FEKO Boundary Element Method is best for linear solutions), but none of these capabilities have been used here. At the minimum, when showing Meep results, they should be shown with a number field, and the fields vector components should be post-processed to show the norm vector resultant instead of Cartesian components. And even then programs like COMSOL and FEKO and ANSYS and ADINA will win hands down, since they have post-processing capabilities to show the fields in 3D while Meep does not. -
#1954 by TheTraveller on 19 Jan, 2016 14:52
-
Just think about this:
1) FEKO model being discussed uses the Boundary Element Method while Meep uses Finite Difference method. It is well known that Boundary Element method for a resonant cavity is much more accurate than the finite difference method (and I'm an expert on these methods).
2) FEKO shows you the Electric fields, Magnetic fields and Currents in the 3D truncated cone as a vector resultant with a number that anyone can understand
3) Compare that with the horrible mess and confusion resulting by showing 6 different components (3 for E and 3 for H) in Cartesian coordinate components without any number field, so that people cannot even tell apart what is practically zero from what is significant !!!!
FEKO is good and it is easy to use. Only issue is the LITE version is difficult to get good meshing on a frustum.
Maybe your skills can work out a way to make the LITE version deliver what EmDriver DIYers need? -
#1955 by rfmwguy on 19 Jan, 2016 14:58
-
I've got a major concern...if people using meep are nicknamed meepers, what about feko users?Just think about this:
1) FEKO model being discussed uses the Boundary Element Method while Meep uses Finite Difference method. It is well known that Boundary Element method for a resonant cavity is much more accurate than the finite difference method (and I'm an expert on these methods).
2) FEKO shows you the Electric fields, Magnetic fields and Currents in the 3D truncated cone as a vector resultant with a number that anyone can understand
3) Compare that with the horrible mess and confusion resulting by showing 6 different components (3 for E and 3 for H) in Cartesian coordinate components without any number field, so that people cannot even tell apart what is practically zero from what is significant !!!!
FEKO is good and it is easy to use. Only issue is the LITE version is difficult to get good meshing on a frustum.
Maybe your skills can work out a way to make the LITE version deliver what EmDriver DIYers need?
don't answer that....
-
#1956 by Rodal on 19 Jan, 2016 15:02
-
I've got a major concern...if people using meep are nicknamed meepers, what about feko users?Just think about this:
1) FEKO model being discussed uses the Boundary Element Method while Meep uses Finite Difference method. It is well known that Boundary Element method for a resonant cavity is much more accurate than the finite difference method (and I'm an expert on these methods).
2) FEKO shows you the Electric fields, Magnetic fields and Currents in the 3D truncated cone as a vector resultant with a number that anyone can understand
3) Compare that with the horrible mess and confusion resulting by showing 6 different components (3 for E and 3 for H) in Cartesian coordinate components without any number field, so that people cannot even tell apart what is practically zero from what is significant !!!!
FEKO is good and it is easy to use. Only issue is the LITE version is difficult to get good meshing on a frustum.
Maybe your skills can work out a way to make the LITE version deliver what EmDriver DIYers need? don't answer that....
-
#1957 by SeeShells on 19 Jan, 2016 15:08
-
Its for everyone, but your statement was what I thought was relevant to the general discussion. I've personally been on both sides of the issue and can spot predispositions fairly well as I'm sure you can. The offsite poster has been creating feko models and I believe that person has disqualified themselves as a fair and objective person numerous time. That person also admitted they were banned twice on nsf (before my time). That is my only point. I trust feko analysis done by people here, not elsewhere. You are correct is wanting to keep things objective. Now more java...
1) The FEKO analysis discussion was introduced into the thread by somebody else, as you know. I answered the confusion that has been produced by INCORRECT identification of mode shapes by incorrect interpretation of FEKO results (and Meep ?) for SeeShell's frustum with waveguide feeding.
This is a good position to take doc, just be aware of the stated position the poster of the feko analysis on another forum has taken..."there is no emdrive effect" paraphrasing the poster whose qualifications and lack of experimentation are unknown. My advice is to get feko into the hands of more objective/respectable/non-anonymous people here before the waters are muddied too much with potentially falsified data analysis. Your position is quite valid but must apply equally...just my early morning thoughts anyway...fwiw....This is my pathway. I have no doubt it will be successful.
You have no doubt, ahead of time, that your test will be successful ?
The scientific method of testing is that one doesn't know ahead of time what the outcome of a future test will be.
2) The discussion with TT is about interpretation of mode shapes. TT says that they should be identified by the currents, I say that goes against all standards. The standard interpretation of mode shapes is that they are identified by the Electric and Magnetic field distributions
3) On both occasions that FEKO analysis was introduced here, it was introduced by 2 other posters. This last time it was introduced by TheTraveller. Why do you address your post to me rather than TheTraveller, or to both of us?
4)TheTraveller has not posted a single image for a mode shape he has calculated on his own and instead he is the one that posts images calculated by somebody else in another forum? Are you going to ask TheTraveller to post his own images for the case in hand instead of posting somebody else's?
While TheTraveller initiated the discussion by posting FEKO results from someone else (https://forum.nasaspaceflight.com/index.php?topic=39004.msg1479711#msg1479711), I posted my Wolfram Mathematica calculations (https://forum.nasaspaceflight.com/index.php?topic=39004.msg1479718#msg1479718).
Instead TheTraveller posts images from FEKO done by somebody else, and to compound the problem TheTraveller has his own interpretation that goes against standards (the standard is to use the electric and magnetic fields to identify a mode shape).
This mode confusion has been going on for months. Add to it the fact not even sure if one mode over another or dielectric inserts or waveguides or antennas will give positive or null results. I feel better in we're getting somewhere.
In 6 months look what has happened.
I was up most of last night reading and going over the hundreds of pages of postings and we all have to realize we have come far to narrowing this abnormality. We should be proud of it.
EagleWorks is in peer review, I'll be ready to run again soon, rfmyguy is redoing his tests and RFPlummer (just did his tests) and TT says he will be soon as well. (missed someone I'm sorry)
Aero has spent countless hours working trying to get meep to squeeze out clean reliable data, and I commend him for that. Dr. Rodal and others have offered their services opinions and ideas as well. I looked at meep and what was needed to run it and the user interface is not a good user interface and very hard to use.
To clarify what was done recently. IslandPlaya has 2 of my models he is running and he has modeled them both correctly.
NOTE:
I've requested a verson of FEKO to run other than the lite that you can use and I'll reconfirm other findings.
************
Congragulations RFPlummer on your recent run, I guess I owe you a PowerBall ticket as your thrusts were a null or in the noise.
The one thing that was a complete surprise to me on your setup was when you ran it for the first time with ~30 watts you caused your endplates to deform. That small of a wattage into a large cavity warping copper plates (I wonder if something could be gleened from Dr. Rodal's great paper).
A thought on your line to the rafters supporting your DUT you could split it and then isolate it with a section of sorbothane. Or even use a sole insert for your shoe (many use sorbathane) to dampen. That line will transmit vibrations. (remember the two cans and a string phone you made as a kid?)
http://www.ebay.com/sch/i.html?_from=R40&_nkw=sorbothane&_oac=1 -
#1958 by Notsosureofit on 19 Jan, 2016 15:09
-
@ RFPlumber
FWIW: The hypothesis gives 6.56mN for your numbers.
EDIT: That should be micro Newtons !
-
#1959 by TheTraveller on 19 Jan, 2016 15:26
-
This mode confusion has been going on for months. Add to it the fact not even sure if one mode over another or dielectric inserts or waveguides or antennas will give positive or null results. I feel better in we're getting somewhere.
I'm excited to be able, with enough power, to ident modes from their end plate and in your case side wall thermal patterns. Yes it may take some time for the thermal image to form and it may not be as high resolution as the FEKO current images but each mode will be different and over time, using FEKO, we can build up a library of modes that can be identified by their thermal signatures.
So to put the question of which mode is excited to bed and to get on with generating good thrust levels.
Phil
don't answer that....