-
#1780 by Mulletron on 16 Jan, 2016 17:58
-
It's not looking good for any sort of generation of gravitational waves via the Gertsenshtein effect.
http://www.redorbit.com/news/space/1613925/gravitational_wave_invesitigation_nonsense/
https://fas.org/irp/agency/dod/jason/gravwaves.pdf (page 11 is about cavities)
Same thing but with the quanta called the graviton:
http://publications.ias.edu/sites/default/files/poincare2012.pdf
Science fiction discussed/debunked in the first two links:
http://www.drrobertbaker.com/docs/Aerospace%20HFGW%20Applications.pdf
Other side of the fence:
http://m.rspa.royalsocietypublishing.org//content/462/2071/1987
http://arc.aiaa.org/doi/abs/10.2514/6.2001-3913
http://arxiv.org/ftp/physics/papers/0410/0410022.pdf (Gasers
)
My take on the current situation with gravitational waves, the only rock solid bits of information one can really rely on is the well known indirect evidence the 1993 Nobel Prize in Physics was awarded for related to the Hulse–Taylor binary system PSR B1913+16. Anything else is purely theoretical. Despite rumors across the internet, we have not officially had our "Heinrich Hertz moment" for gravitational waves. Even more speculative, anything related to HFGW simply isn't reliable without direct evidence. They could be generated with floobie dust for all I know. Looks like we need about 20 orders of magnitude more of the stuff. -
#1781 by SeeShells on 16 Jan, 2016 18:00
-
Nice Shell...remember, 90 degrees out of phase is OK, 180 would be "problematic"
Sure Glen it's not perfect and I welcome thoughts on how I could do a straighter shot into the cavity.
This design addresses the shortcomings of the prior test in phasing the two incoming RF signals.
Not being an RF type, but trolling the RF posts, would it be reasonable to ask for the dimensions and see what MEEP would predict for the injection port characteristics and any other effects that would happen in the waveguide?
Having worked a few microwave backhauls in my time, there could be a few oh-by-the-ways buried in the design.
In my world it's usually a straight haul from the transmitter port to the antenna. Something like this would floor the RF guys I work with, not that it won't work, I can just picture the reaction I'd get if I said, "we're going to use this on the next link."
I started this basic layout about three months ago and as I learned more about waveguides and what I needed I ran across John F. Gerling's VP of Gerling Applied Engineering, Inc very nice paper on the basics about a month ago. He had almost the same design I'd done but with a few additions I liked. I'd used only one phase shifter (I saw going into the cavity off phase with one side being out of phase wasn't going to be right). Also I speced flexible waveguides for the bends and my tuner was different.
I even have a design using one waveguide to the frustum and splitting it there but I'm worried about the extra weight.
Shell
This is a good way to "stir the EM pot" at the frustum, but it might interfere with mode stability. Good thing about it is being adjustable is you can find the optimum phase angle. This will be VERY interesting to see. If a wormhole vortex opens up over Colorado, we're blaming you!
hmmmm. Could have said 360 out of phase.
Stand corrected.
From John F. Gerling's paper I just posted.
Gerling Applied Engineering
<Quote>
The phase shifters operate by rotating a thin dielectric slab inside the waveguide with its rotational axis in the plane of the electric field. When the slab, having low dielectric loss and high dielectric constant characteristics, rotates between positions perpendicular to and parallel with the waveguide centerline the phase shift alternates sinusoidally from near zero to maximum. Adjusting the slab geometry varies the phase shift amplitude. Rotating both phase shifters synchronously and exactly 90 degrees out of phase with each other will then cause a sinusoidal reciprocation of the standing wave pattern inside the applicator at constant amplitude. -
#1782 by SeeShells on 16 Jan, 2016 18:12
-
Hi X_Ray,Before I set upon the what I'll call the burning probe matchstick test (about a month ago) that ended with me toasting my waveguide antennas into the frustum and the direct designed coupling from the magnetron into the coax I was hoping to have it last long enough to get some results. Sadly it didn't, but I had other plans in the works that I have been actively perusing since then to stabilize the dual phase reversed injection into the drive. I feel this is the best way to have a stable high power injection into the drive.
Hi Shell,
This design addresses the shortcomings of the prior test in phasing the two incoming RF signals. I talked about this method very lightly about 6 months ago using two phase locked magnetrons, but I found way too many design hurdles to overcome using the direct injection method. One was the massive amount of excess heat from the magnetrons that would be strapped onto the cavity, swamping any measurements of thrust.
If the burning probe matchstick test failed because of design limitations (which it kind of did, got a large force displacement on the digital scales before it fried itself) I had plan number 2. And I'll present it here for the first time and invite comments.
Getting together pieces and parts to build, I've had to go surplus because it's not cheap. From the undocumented power on test results of the first one I know it's the way I need to go to assure a stable clean high power high Q mode control into my drive.
Shell
PS: RfPlumber this is why I hadn't answered your last question, simply I've gone beyond it and I'd welcome any inputs you or anyone may have.
Added: Speeeling corrections.
could you so kind to add the circulator direction ( or port numbers & type/datasheet)?
I have to think about the effect of your phase shifters.
If you go to the very last page on the .pdf I posted by John Gerling you will see the circulators and their directions. I haven't got the circulators yet so I'll not post the spec sheets if that's ok?
Shell
http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=39004.0;attach=1094383 -
#1783 by artefact on 16 Jan, 2016 18:15
-
@Tellmeagain: You were amused about my name and the mentioning on this other website (ECW).
I don't know if there is a language where artifact is written artefact. Actually my name is a mix of many things. Most of them are not obvious to others. In no way it is ment to ridicule.
Back to lurking for me. Good luck everyone. -
#1784 by rfmwguy on 16 Jan, 2016 18:16
-
You've probably thought about this Shell, but you might want to try using the 2 WR340 waveguides side by side, making one moment arm of the balance beam towards the frustum.
-
#1785 by SeeShells on 16 Jan, 2016 18:47
-
You've probably thought about this Shell, but you might want to try using the 2 WR340 waveguides side by side, making one moment arm of the balance beam towards the frustum.
I have thought about it and I think it's a good idea. The final configuration will depend on feedback from here and what I can get in hardware and design considerations come first of course.
Shell -
#1786 by X_RaY on 16 Jan, 2016 18:49
-
That was exactly the reason why I asked that. Take a look to the port numbers and the path (picture below). This is the way it works.
Hi X_Ray,Before I set upon the what I'll call the burning probe matchstick test (about a month ago) that ended with me toasting my waveguide antennas into the frustum and the direct designed coupling from the magnetron into the coax I was hoping to have it last long enough to get some results. Sadly it didn't, but I had other plans in the works that I have been actively perusing since then to stabilize the dual phase reversed injection into the drive. I feel this is the best way to have a stable high power injection into the drive.
Hi Shell,
This design addresses the shortcomings of the prior test in phasing the two incoming RF signals. I talked about this method very lightly about 6 months ago using two phase locked magnetrons, but I found way too many design hurdles to overcome using the direct injection method. One was the massive amount of excess heat from the magnetrons that would be strapped onto the cavity, swamping any measurements of thrust.
If the burning probe matchstick test failed because of design limitations (which it kind of did, got a large force displacement on the digital scales before it fried itself) I had plan number 2. And I'll present it here for the first time and invite comments.
Getting together pieces and parts to build, I've had to go surplus because it's not cheap. From the undocumented power on test results of the first one I know it's the way I need to go to assure a stable clean high power high Q mode control into my drive.
Shell
PS: RfPlumber this is why I hadn't answered your last question, simply I've gone beyond it and I'd welcome any inputs you or anyone may have.
Added: Speeeling corrections.
could you so kind to add the circulator direction ( or port numbers & type/datasheet)?
I have to think about the effect of your phase shifters.
If you go to the very last page on the .pdf I posted by John Gerling you will see the circulators and their directions. I haven't got the circulators yet so I'll not post the spec sheets if that's ok?
Shell
http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=39004.0;attach=1094383
I hope the insert loss is not so high for the single components.
It's a great idea!
Added:
Thanks for this Paper/Application note!
-
#1787 by SeeShells on 16 Jan, 2016 19:24
-
You're quite welcome and thanks for the drawing. I wasn't quite sure if it was going to work the way I wanted it to but John Gerling's paper and you, (who I regard very highly) has made me confident this is the right track to go.
That was exactly the reason why I asked that. Take a look to the port numbers and the path (picture below). This is the way it works.
Hi X_Ray,Before I set upon the what I'll call the burning probe matchstick test (about a month ago) that ended with me toasting my waveguide antennas into the frustum and the direct designed coupling from the magnetron into the coax I was hoping to have it last long enough to get some results. Sadly it didn't, but I had other plans in the works that I have been actively perusing since then to stabilize the dual phase reversed injection into the drive. I feel this is the best way to have a stable high power injection into the drive.
Hi Shell,
This design addresses the shortcomings of the prior test in phasing the two incoming RF signals. I talked about this method very lightly about 6 months ago using two phase locked magnetrons, but I found way too many design hurdles to overcome using the direct injection method. One was the massive amount of excess heat from the magnetrons that would be strapped onto the cavity, swamping any measurements of thrust.
If the burning probe matchstick test failed because of design limitations (which it kind of did, got a large force displacement on the digital scales before it fried itself) I had plan number 2. And I'll present it here for the first time and invite comments.
Getting together pieces and parts to build, I've had to go surplus because it's not cheap. From the undocumented power on test results of the first one I know it's the way I need to go to assure a stable clean high power high Q mode control into my drive.
Shell
PS: RfPlumber this is why I hadn't answered your last question, simply I've gone beyond it and I'd welcome any inputs you or anyone may have.
Added: Speeeling corrections.
could you so kind to add the circulator direction ( or port numbers & type/datasheet)?
I have to think about the effect of your phase shifters.
If you go to the very last page on the .pdf I posted by John Gerling you will see the circulators and their directions. I haven't got the circulators yet so I'll not post the spec sheets if that's ok?
Shell
http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=39004.0;attach=1094383
I hope the insert loss is not so high for the single components.
It's a great idea!
Added:
Thanks for this Paper/Application note!
The last test proved there might be something that it showed up at higher power although it also showed weaknesses in my basic design running at higher power. RFPlummer is right my VSWR just kicked back down the coax, found the weak points and fried it all. This will correct that and keep me from doing the burnt matchstick antenna test again.
Shell
-
#1788 by aero on 16 Jan, 2016 20:15
-
Fooling a little more with my Tajmar device model. Actually, I just added a couple or detectors for Harminv resonance detection. The locations show as spheres in the attached .gif's. I can add more if there is justification to do so.
hmm . The z-view gif doesn't seem to upload properly though it runs locally?
-
#1789 by Rodal on 16 Jan, 2016 20:17
-
Dr. Rodal and others -
This is the information I'm aware of:
Attached find my attempt, with SeeShell's guidance, to model the geometry of Tajmar's frustum. I have labelled the sections so that we can identify them for discussion.
I have made several guess: Like "What is the height of the feed and coupler sections?" I'm assuming that the width is 1/2 the height as is common for waveguides, and Shell extracted the section lengths from the image. I am still in doubt about the length of the WR 340 but Shell assures me that it is a commercial part and I'm using those dimensions. But the antenna doesn't seem to me to be the right distance from the end of the WR 340 being 0.021336 meters instead of 1/4 wavelength which is about 0.03 meters.
One other item in question is regarding a rectangular z-choke. Did Tajmar use one as Yang did, and if he did, where was it located?
In any event the fields are not forming quickly, if they ever will. The attached two gif's are 10 frames from one cycle at the end of a 64 cycle run. I haven't tried to make a resonance run because I think there is still some tweaks needed to the model.
* Between the Cavity and the waveguide Tajmar used an adapter.
Following are the dimensions in the latest corrected version of Tajmar's EM Drive paper:
http://bit.ly/1h4E0RzQuoteWe iterated our design several times by consulting with R. Shawyer to be as representative as possible. Our final tapered cavity design had an
internal top radius of 38.5 mm, a
bottom radius of 54.1 mm and a
height of 68.6 mm
as well as a side entrance for the microwaves as shown in Fig. 2. The cavity was made out of three copper pieces with a wall thickness of 3 mm where the lower and middle part as well as the side flange were hard soldered using silver and the top part was able to adapt its position in order to optimize for a high Q factor. A standard WR340 waveguide was then used to connect the magnetron to the EMDrive
The dimensions I am using, from FluxCapacitor's post give the same diameters but a greater height of 72.8 mm. Which set of dimensions do you suppose is correct?
Looking at the images in the paper from Tajmar, it looks like the cavity tested was not the same as the prototype cavity commonly illustrated and which I based the model geometry on. Assuming that the cavity imaged inside of the vacuum chamber was the one tested.
.....
aero
For a complete answer, see this post: https://forum.nasaspaceflight.com/index.php?topic=39214.msg1477307#msg1477307
As Tajmar disclosed that the tested fustrum of a cone had spherical ends, and he did not provide any drawing clarifying what is the technical meaning of the dimensions he provided, it is not possible to have a unique interpretation of the dimensions of his resonant cavity:
1) There are several possible interpretations of "height" of a frustum of a cone with spherical ends
2) The "height" dimension given by Tajmar cannot be the height of the fustrum under conventional interpretations, because that leads to incorrect natural frequencies. Hence, the dimension given must be interpreted as being 1/2 of the height.
I will show results for two likely interpretations to what Tajmar refers as "height" of the fustrum of a cone with spherical ends:
A) The lateral length of the conical walls of the fustrum of a cone, from the small end to the big end
B) The length (from the small end to the big end) measured perpendicular to the lines defining the small and big diameters of the fustrum of a cone.
All cases will assume that the small radius and big radius are the correct internal dimensions of the diameters divided by 2:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
_____________________
In this post https://forum.nasaspaceflight.com/index.php?topic=39214.msg1477307#msg1477307, I consider at length 4 different interpretations, including interpreting the height given by Tajmar being an internal dimension, and also considering it as being an external dimension. After these calculations it is apparent that the height given must be an external dimension. I conclude that the height of Tajmar's tested fustrum of a cone, given in the latest, corrected version of his AIAA paper, was the external height.
If the height was taken as the lateral length of the conical walls of the fustrum of a cone, from the small end to the big end, then the internal dimensions were:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
axialLength = 2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
= 0.1312 meter
resulting in
first natural frequency (mode shape TM010 *) = 2.47432 GHz
or
if the height was taken as the length (from the small end to the big end) measured perpendicular to the lines defining the small and big diameters of the fustrum of a cone, then the internal dimensions were:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
axialLength =2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
= 0.1312 meter
resulting in
first natural frequency (mode shape TM010 *) = 2.45701 GHz
_______________________________________________________
NOTE:
1) (*) The first mode shape in a truncated cone is NOT constant in the longitudinal direction. We label it as TM010 here following the convention in these threads of calling the mode shape closest to the one in a cylindrical cavity, but it should be understood that TM010 electromagnetic fields vary in the longitudinal direction !
2) The theoretical Q, for perfect coupling should have been a little less than 34,000. Since Tajmar's test had an awful small Q (48.8 in ambient conditions and 20 in partial vacuum), Tajmar's test had horribly bad coupling ! No doubt due to the way that they coupled the huge waveguide into the small cavity.
-
#1790 by aero on 16 Jan, 2016 20:43
-
... snip ...
The dimensions I am using, from FluxCapacitor's post give the same diameters but a greater height of 72.8 mm. Which set of dimensions do you suppose is correct?
Looking at the images in the paper from Tajmar, it looks like the cavity tested was not the same as the prototype cavity commonly illustrated and which I based the model geometry on. Assuming that the cavity imaged inside of the vacuum chamber was the one tested.
.....
aero
For a complete answer, see this post: https://forum.nasaspaceflight.com/index.php?topic=39214.msg1477307#msg1477307
As Tajmar disclosed that the tested fustrum of a cone had spherical ends, and he did not provide any drawing clarifying what is the technical meaning of the dimensions he provided, it is not possible to have a unique interpretation of the dimensions of his resonant cavity:
1) There are several possible interpretations of "height" of a frustum of a cone with spherical ends
2) The "height" dimension given by Tajmar cannot be the height of the fustrum under conventional interpretations, because that leads to incorrect natural frequencies (much high than the tested frequency). Hence, the dimension given must be interpreted as being 1/2 of the height.
I will show results for two likely interpretations to what Tajmar refers as "height" of the fustrum of a cone with spherical ends:
A) The lateral length of the conical walls of the fustrum of a cone, from the small end to the big end
B) The length (from the small end to the big end) measured perpendicular to the lines defining the small and big diameters of the fustrum of a cone.
All cases will assume that the small radius and big radius are the correct internal dimensions of the diameters divided by 2:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
_____________________
In this post https://forum.nasaspaceflight.com/index.php?topic=39214.msg1477307#msg1477307, I consider at length 4 different interpretations, including interpreting the height given by Tajmar being an internal dimension, and also considering it as being an external dimension. After these calculations it is apparent that the height given must be an external dimension. I conclude that the height of Tajmar's tested fustrum of a cone, given in the latest version, corrected version of his AIAA paper, was the external height.
If the height was taken as the lateral length of the conical walls of the fustrum of a cone, from the small end to the big end, then the internal dimensions were:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
axialLength = 2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
= 0.1312 meter
resulting in
first natural frequency (mode shape TM011) = 2.47432 GHz
or
if the height was taken as the length (from the small end to the big end) measured perpendicular to the lines defining the small and big diameters of the fustrum of a cone, then the internal dimensions were:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
axialLength =2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
= 0.1312 meter
resulting in
first natural frequency (mode shape TM011) = 2.45701 GHz
_______________________________________________________
NOTE: The theoretical Q, for perfect coupling should have been a little less than 34,000. Since Tajmar's test had an awful small Q (48.8 in ambient conditions and 20 in partial vacuum), Tajmar's test had horribly bad coupling ! No doubt due to the way that they coupled the huge waveguide into the small cavity.
Thanks - that will help -
The remaining big question has to do with the waveguide feed dimensions since there is some concern that the EM fields may have been resonating within the waveguide. If so, then the waveguide feed dimensions must be accurately modelled and I don't have anything to go on other than measurements and visuals taken from the one photograph of the prototype set-up.
Irrespective of waveguide resonance or not, accurate dimensions of the waveguide feed sections are needed in order that the Maxwell equations solution from meep can hope to match experiment with any level of fidelity at all. Its the other side of GIGO. -
#1791 by Rodal on 16 Jan, 2016 21:03
-
... snip ...
The dimensions I am using, from FluxCapacitor's post give the same diameters but a greater height of 72.8 mm. Which set of dimensions do you suppose is correct?
Looking at the images in the paper from Tajmar, it looks like the cavity tested was not the same as the prototype cavity commonly illustrated and which I based the model geometry on. Assuming that the cavity imaged inside of the vacuum chamber was the one tested.
.....
aero
For a complete answer, see this post: https://forum.nasaspaceflight.com/index.php?topic=39214.msg1477307#msg1477307
As Tajmar disclosed that the tested fustrum of a cone had spherical ends, and he did not provide any drawing clarifying what is the technical meaning of the dimensions he provided, it is not possible to have a unique interpretation of the dimensions of his resonant cavity:
1) There are several possible interpretations of "height" of a frustum of a cone with spherical ends
2) The "height" dimension given by Tajmar cannot be the height of the fustrum under conventional interpretations, because that leads to incorrect natural frequencies (much high than the tested frequency). Hence, the dimension given must be interpreted as being 1/2 of the height.
I will show results for two likely interpretations to what Tajmar refers as "height" of the fustrum of a cone with spherical ends:
A) The lateral length of the conical walls of the fustrum of a cone, from the small end to the big end
B) The length (from the small end to the big end) measured perpendicular to the lines defining the small and big diameters of the fustrum of a cone.
All cases will assume that the small radius and big radius are the correct internal dimensions of the diameters divided by 2:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
_____________________
In this post https://forum.nasaspaceflight.com/index.php?topic=39214.msg1477307#msg1477307, I consider at length 4 different interpretations, including interpreting the height given by Tajmar being an internal dimension, and also considering it as being an external dimension. After these calculations it is apparent that the height given must be an external dimension. I conclude that the height of Tajmar's tested fustrum of a cone, given in the latest version, corrected version of his AIAA paper, was the external height.
If the height was taken as the lateral length of the conical walls of the fustrum of a cone, from the small end to the big end, then the internal dimensions were:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
axialLength = 2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
= 0.1312 meter
resulting in
first natural frequency (mode shape TM011) = 2.47432 GHz
or
if the height was taken as the length (from the small end to the big end) measured perpendicular to the lines defining the small and big diameters of the fustrum of a cone, then the internal dimensions were:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
axialLength =2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
= 0.1312 meter
resulting in
first natural frequency (mode shape TM011) = 2.45701 GHz
_______________________________________________________
NOTE: The theoretical Q, for perfect coupling should have been a little less than 34,000. Since Tajmar's test had an awful small Q (48.8 in ambient conditions and 20 in partial vacuum), Tajmar's test had horribly bad coupling ! No doubt due to the way that they coupled the huge waveguide into the small cavity.
Thanks - that will help -
The remaining big question has to do with the waveguide feed dimensions since there is some concern that the EM fields may have been resonating within the waveguide. If so, then the waveguide feed dimensions must be accurately modelled and I don't have anything to go on other than measurements and visuals taken from the one photograph of the prototype set-up.
Irrespective of waveguide resonance or not, accurate dimensions of the waveguide feed sections are needed in order that the Maxwell equations solution from meep can hope to match experiment with any level of fidelity at all. Its the other side of GIGO.
Sorry to be blunt, but it is apparent to me that the Tajmar team did not do a good coupling job as their Q was horribly low (50 instead of 34,000 theoretical for copper).
It is apparent to me that they should have used an iris to do the coupling with the waveguide, as solved by a Nobel Prize winner in the 1940's. It appears to me that a lot of knowledge regarding waveguides and resonant cavities that people had at the MIT Radiation Laboratory during WWII and that several Nobel Prizes used for resonant cavities for Radar, for proton accelerators, and for the Manhattan project was just lacking in this Tajmar project (and I'm also suspect of Roger Shawyer since he is mentioned as an assistant to Tajmar's project and Roger Shawyer has written about very strange things like using cut-off formulas that are known to apply only for open waveguides and because Shawyer does not understand in his papers that a truncated cone has lateral radiation pressure on the lateral walls even when using spherical ends).
In other words: what Tajmar did concerning coupling of the waveguide into the resonant cavity could not have followed even the standard practice known in WWII, and much less today's standard practice in applications like particle accelerators etc.
So, I cannot help in this regard except to say that the way that he coupled the waveguide into the cavity ended up with a horribly low Q of 50 instead of a theoretical one of 34,000 (whatever he did, it was not very good
).
-
#1792 by aero on 16 Jan, 2016 21:13
-
Dr. Rodal - Please check your axial length numbers posted above. They are the same. I suspect you copied/pasted the same data in both places intending one set to be axial length and the other to be lateral or slant length.
-
#1793 by Rodal on 16 Jan, 2016 21:17
-
Dr. Rodal - Please check your axial length numbers posted above. They are the same. I suspect you copied/pasted the same data in both places intending one set to be axial length and the other to be lateral or slant length.
The numbers are correct.
The numbers should be the same.
The interpretation of what the numbers mean is different: one interpretation corresponds to the length of the conical walls, the lateral length corresponding to (r2 - r1) in the picture below. The other interpretation is the height measured perpendicular to imaginary lines defining the diameters.
Best thing (if you are up to it) would be to make a drawing of your interpretation of what "height" means, and I could check it as to what meaning your drawing corresponds to.
This is the geometry defining the spherical radii r1, r2 and the halfconeangle "θ"
-
#1794 by aero on 16 Jan, 2016 21:30
-
Quote
Sorry to be blunt, but it is apparent to me that the Tajmar team did not do a good coupling job as their Q was horribly low (50 instead of 34,000 theoretical for copper).
I almost feel the need to apologize for Dr. Tajmar. The documentation we have does not stand up to scrutiny at any level that I can see. However, please clarify the objective of this modelling effort.
Do you want to see a model of the set-up as it should have been? That we can do using the cavity dimensions you have provided.
Or do you want to see a model of the set-up as built/tested? That will be more challenging. Perhaps I could find a very low Q solution by first modelling the cavity itself as you have specified, then:
1- Identify the number of segments of the waveguide feed starting with the WR 340 waveguide.
2- Guess at the dimensions of the segment(s) between the WR 340 and the cavity.
3- Iterate.
I will address the spherical ends of the cavity in the mean time.
-
#1795 by Rodal on 16 Jan, 2016 21:47
-
The purpose is met by this kind of examination:QuoteSorry to be blunt, but it is apparent to me that the Tajmar team did not do a good coupling job as their Q was horribly low (50 instead of 34,000 theoretical for copper).
I almost feel the need to apologize for Dr. Tajmar. The documentation we have does not stand up to scrutiny at any level that I can see. However, please clarify the objective of this modelling effort.
Do you want to see a model of the set-up as it should have been? That we can do using the cavity dimensions you have provided.
Or do you want to see a model of the set-up as built/tested? That will be more challenging. Perhaps I could find a very low Q solution by first modelling the cavity itself as you have specified, then:
1- Identify the number of segments of the waveguide feed starting with the WR 340 waveguide.
2- Guess at the dimensions of the segment(s) between the WR 340 and the cavity.
3- Iterate.
I will address the spherical ends of the cavity in the mean time.
1) First Tajmar had the diameters off by a factor of 2
2) Upon discussions by e-mail he accepted the dimensions were wrong and corrected them
3) The "height" was still off. His student was on vacation. Eventually he measured it himself. However, the latest correction still fails to inform the reader that this is not the height but 1/2 of the height and there is no drawing showing what he means by height.
3) I spent time looking at 4 different interpretations and concluded that to be at ~2.45GHz magnetron bandwidth the height figure must be 1/2 of the external dimension. This was useful in showing that he excited TM010 which is the same mode shown on his paper as modeled by COMSOL FEA
4) Your Meep analysis shows that there is a big problem exciting a mode shape. I think that was valuable to find out.
5) In retrospect, it is not surprise since Tajmar obtained a horrible Q of 50 instead of 34,000, but it is nice to have a Meep model showing that. (Although there is no way to know whether this Meep model represents what Tajmar did regarding geometrical coupling, since he did not give the details).
6) Which means that Tajmar's coupling, including the geometrical coupling of the waveguide left a lot to be desired
7) You are welcome to modify the dimensions in your model for the fustrum itself as I give in my above post
8) Concerning what Tajmar did to screw up the coupling of the waveguide feed into the fustrum there are so many ways to screw this up, that I cannot help any further :(
-
#1796 by RFPlumber on 16 Jan, 2016 22:13
-
So, I cannot help in this regard except to say that the way that he coupled the waveguide into the cavity ended up with a horribly low Q of 50 instead of a theoretical one of 34,000 (whatever he did, it was not very good
).
IMHO, they did one thing very good, a very important thing and because of this their paper should be read by every wanna-be EmDrive fan: they showed that there exist large forces related to hot air, and that those forces are indeed / often different depending on orientation of the frustum + magnetron assembly, and all those forces completely disappear when testing in vacuum, and hence it is a really bad idea to consider any difference in those hot air-related forces to be "thrust".
-
#1797 by aero on 16 Jan, 2016 22:15
-
Dr. Rodal - Please check your axial length numbers posted above. They are the same. I suspect you copied/pasted the same data in both places intending one set to be axial length and the other to be lateral or slant length.
As I use axial height, it really is equal to the slant height times the cosine of the cavity half angle, that is:
AXhi = (r2 - r1 ) * cos (θ ) as shown in the poorly annotated image attached.
Here, (r2 - r1 ) is the slant height of the frustum.
Tell me the value of r1 and r2 and knowing the values of the radii, I can work with that. -
#1798 by aero on 16 Jan, 2016 22:17
-
oops
-
#1799 by SeeShells on 16 Jan, 2016 22:20
-
The purpose is met by this kind of examination:QuoteSorry to be blunt, but it is apparent to me that the Tajmar team did not do a good coupling job as their Q was horribly low (50 instead of 34,000 theoretical for copper).
I almost feel the need to apologize for Dr. Tajmar. The documentation we have does not stand up to scrutiny at any level that I can see. However, please clarify the objective of this modelling effort.
Do you want to see a model of the set-up as it should have been? That we can do using the cavity dimensions you have provided.
Or do you want to see a model of the set-up as built/tested? That will be more challenging. Perhaps I could find a very low Q solution by first modelling the cavity itself as you have specified, then:
1- Identify the number of segments of the waveguide feed starting with the WR 340 waveguide.
2- Guess at the dimensions of the segment(s) between the WR 340 and the cavity.
3- Iterate.
I will address the spherical ends of the cavity in the mean time.
1) First Tajmar had the diameters off by a factor of 2
2) Upon discussions by e-mail he accepted the dimensions were wrong and corrected them
3) The "height" was still off. His student was on vacation. Eventually he measured it himself. However, the latest correction still fails to inform the reader that this is not the height but 1/2 of the height and there is no drawing showing what he means by height.
3) I spent time looking at 4 different interpretations and concluded that to be at ~2.45GHz magnetron bandwidth the height figure must be 1/2 of the external dimension. This was useful in showing that he excited TM010 which is the same mode shown on his paper as modeled by COMSOL FEA
4) Your Meep analysis shows that there is a big problem exciting a mode shape. I think that was valuable to find out.
5) In retrospect, it is not surprise since Tajmar obtained a horrible Q of 50 instead of 34,000, but it is nice to have a Meep model showing that. (Although there is no way to know whether this Meep model represents what Tajmar did regarding geometrical coupling, since he did not give the details).
6) Which means that Tajmar's coupling, including the geometrical coupling of the waveguide left a lot to be desired
7) You are welcome to modify the dimensions in your model for the fustrum itself as I give in my above post Concerning what Tajmar did to screw up the coupling of the waveguide feed into the fustrum there are so many ways to screw this up, that I cannot help any further
Areo,
Please move the antenna probe 1/4 Wl from the back wall of the Magnetron>waveguide. 0.030 M @ 2.45GHz.
http://www.ainfoinc.com/en/pro_pdf/new_products/Adapter/tr_340WCAN.pdf
Might have miss-read the external size and placed the antenna probe too far away from the back wall of the waveguide.
)
In my world it's usually a straight haul from the transmitter port to the antenna. Something like this would floor the RF guys I work with, not that it won't work, I can just picture the reaction I'd get if I said, "we're going to use this on the next link."
).