Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1799595 times)

Offline TheTraveller

Actually, I forgot that Tajmar had a screw-driven reduction in length.  The length given must be the external length, the maximum length possible.

The COMSOL FEA image from Tajmar shows an internal length which is SMALLER than the big diameter. 
If I use this aspect ratio, and doubling the diameters I can get a natural frequency for TE111 around 2.45 GHz

So, yes it makes sense that:

The diameters given are actually the radii.  The diameters are twice as big as given


Good. So we agree and our modelling software agree.

Would be nice to know the internal dimensions but I suspect what we have is what we will get. I did take 2mm off the diameters (assumed 1mm thick side walls) as I believe they are probably external dimensions, like the length is the overall external dimension.

At least we how have resonance and mode numbers that make sense.
« Last Edit: 07/27/2015 12:16 AM by TheTraveller »
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.”
Herman Melville, Moby Dick

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5915
  • Likes Given: 5253
Actually, I forgot that Tajmar had a screw-driven reduction in length.  The length given must be the external length, the maximum length possible.

The COMSOL FEA image from Tajmar shows an internal length which is SMALLER than the big diameter. 
If I use this aspect ratio, and doubling the diameters I can get a natural frequency for TE111 around 2.45 GHz

So, yes it makes sense that:

The diameters given are actually the radii.  The diameters are twice as big as given


Good. So we agree and our modelling software agree.

Would be nice to know the internal dimensions but I suspect what we have is what we will get. I did take 2mm off the diameters (assumed 1mm thick side walls) as I believe they are probably external dimensions, like the length is the overall external dimension.

At least we how have resonance and mode numbers that make sense.

Taking (2*GivenDiameters - 2mm)
I have to use a length of 0.76*2*lengthGivenInArticle

(76% of the external length)

to get 2.45 GHz

so that is assuming the screws reduce the internal length by 24% , as in the COMSOL FEA picture

to get the 2.45 GHz resonance at TM010 and 73.5% of the external length to get 2.45 GHz at TE111

*****

Good job of thinking about the factor of 2 !!!!! Traveller :)
« Last Edit: 07/27/2015 07:21 PM by Rodal »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5915
  • Likes Given: 5253
And thanks and warm congratulations to X-Ray for first thinking about the factor of 2:

http://forum.nasaspaceflight.com/index.php?topic=37642.msg1409910#msg1409910

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5915
  • Likes Given: 5253
@Dr. Rodal

The new big and small end lateral antenna csv files are up and I have removed the previous version. This time they are in two folders so you won't get them mixed up. Naming convention is the same as my most recent previous data set uploaded, with only the base name changed, as you will see.

Each folder contains a Meep Data Request file to describe the run. Ask me for the information that I forgot to add.

https://drive.google.com/folderview?id=0B1XizxEfB23tfkF0Z184NHRtd0ViN28tNzRDY3JzSVc0WFBTOGZmSFZMcUpWLWJfcDRfZEU&usp=sharing

aero

I downloaded (after re-installing Google Drive, which was malfunctioning) the files for the case of the lateral antenna at the Small End of Yang/Shell, and I computed the stresses and forces.

Before posting the data, would you be so kind as to verify whether you obtained the big base and the small base at the following locations?

These are the locations where Mathematica shows the bases:


BIG BASE  :  row 15  (rows ranging from 0 to 228)

SMALL BASE : row 214 (rows ranging from 0 to 228)

Please note:

1) I adopted the same convention as yours:  rows starting at zero

2) when I plot the stress for the big base it looks very small, that's why I would like you to double-check whether the locations are correct

******************

Also, did you ever run csv files for rfmwguy lateral antenna at the Big Base?  All I recall is the antenna at the small end.
« Last Edit: 07/27/2015 01:29 AM by Rodal »

Online dustinthewind

  • Full Member
  • ****
  • Posts: 584
  • U.S. of A.
  • Liked: 226
  • Likes Given: 256
Thinking helps me to take a pause.

I have a silly question that I'm trying to wrap my head around it and need some help.
The group velocity holds the energy, momentum and information in a waveguide but the phase velocity can be superluminal in a waveguide. I see this.
But...
When the waves carrying energy, momentum and information reach cutoff and collapse in a wave guide then the superluminal phase velocity also collapses and in that instant where the phase velocities are collapsing from a superluminal speed into the group velocity wave front. What happens? Make sense?

Shell

from this link here:
video link in this post.

The phase velocity I believe is simply the light speed of the photons which are the fundamental part of the wave.  In the link above in the video there are two phase velocities.  Both appear under the speed of light.  The group velocity is the superposition of the two waves and that velocity in the video is above both phase velocities and possibly above c.  Being that information is limited by the speed of light I would suspect that what carries information is the fundamental wave and not the super position wave.  That being the phase wave or wave with phase velocity.  I was thinking this might happen in a tapered cavity as the cut off of a larger wavelength leads to a lower velocity (phase or group is the question - I think Tod said wave gets converted to evanescent wave and slows down so I would suspect phase wave but I am not sure).  The smaller wavelength then should have a higher phase velocity (if my guess is correct) and then you get those two waves moving in the same direction but at different velocities leading to a larger group velocity.  Hopefully I am not to far off in understanding this. 

Condolences on your loss.  She sounds like a great mom and a tribute to our kind. 
« Last Edit: 07/27/2015 12:56 AM by dustinthewind »

Offline TheTraveller

And thanks and warm congratulations to X-Ray for first thinking about the factor of 2:

http://forum.nasaspaceflight.com/index.php?topic=37642.msg1409910#msg1409910

At least we now know the Tajmar frustum can work at the claimed 2.45GHz.

With my model I can also gen the Df and from that and with the Q and power predict the Force generated.

It is too bad that Tajmar didn't appreciate that a return loss of 12.5dBs meant to his reflected power as he would not have quoted 700Ws as the frustum input power. Likewise any engineer that is Rf savvy knows what VSWR means and that a VSWR of 69:1 means your Rf generator is gonna get very hot, very quickly, unless your design has an isolator and reflected energy dump, which the Tajmar design doesn't have.

Which is why I say over and over again, EMDrive DIYers need a way to measure VSWR or return loss peak dBs and to understand that they mean and how to increase the return loss dBs or lower the VSWR, which are, in effect, the same thing.

It is silly, even stupid to quote data sheet magnetron output power as if it will magically all get inside a frustum and then magically turn into Force generation.

Wish it were that simple, but folks it aint.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.”
Herman Melville, Moby Dick

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5915
  • Likes Given: 5253
I hope that the Tajmar team reads these NSF pages and perhaps they can address these issues and re-run and report higher values :)

including the issue that that huge waveguide may be responsible for the extraneous forces they measured in directions in which the EM Drive is not supposed to thrust:

http://forum.nasaspaceflight.com/index.php?topic=37642.msg1409921#msg1409921
« Last Edit: 07/27/2015 01:06 AM by Rodal »

Online birchoff

  • Full Member
  • **
  • Posts: 267
  • United States
  • Liked: 122
  • Likes Given: 95
Quote from: birchoff
Whats the predicted thrust if you drop the 700W to 360W using the initial Q of 48.8? Also what do the numbers look like if you assume the model that was used was the one in equation 1 of The Development of a Microwave Engine for Spacecraft Propulsion

66uN is the Force prediction at a Q of 48.8.

Equation 1 in the 2005 Brighton paper is obsolete. My spreadsheet uses the Force equation in the current theory paper as attached.

I understand that the model in the paper I referenced is superseded. However I am assuming that since Tajmar didnt reference that paper you attached that the model used is the one from the paper I am referencing.

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
polarized slot antenna, poor rl, surprised at any results...which is a good thing for the ongoing project.


Offline TheTraveller

Quote from: birchoff
Whats the predicted thrust if you drop the 700W to 360W using the initial Q of 48.8? Also what do the numbers look like if you assume the model that was used was the one in equation 1 of The Development of a Microwave Engine for Spacecraft Propulsion

66uN is the Force prediction at a Q of 48.8.

Equation 1 in the 2005 Brighton paper is obsolete. My spreadsheet uses the Force equation in the current theory paper as attached.

I understand that the model in the paper I referenced is superseded. However I am assuming that since Tajmar didnt reference that paper you attached that the model used is the one from the paper I am referencing.

Equation 11 in the 2010 paper (referenced by Tajmar) is the current Force equation, which also details how to calc the Df.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.”
Herman Melville, Moby Dick

Offline TheTraveller

polarized slot antenna, poor rl, surprised at any results...which is a good thing for the ongoing project.

All I can add is that Tajmar is not an Rf microwave engineer.

Suspect he called in Shawyer when he got no results.

Even with Shawyers help his magnetron runs at 200C, while the small end of the cavity runs at 35C. This suggests to me the reflected power is way higher than 336W and that the real power inside the cavity is much lower than 360Ws.

This again says us EMDrive DIYers need to be able to measure our VSWR or return loss dBs and to do adjustments to get them as optimal as possible so to at least have a fighting chance to see some Force generated.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.”
Herman Melville, Moby Dick

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5915
  • Likes Given: 5253
polarized slot antenna, poor rl, surprised at any results...which is a good thing for the ongoing project.

All I can add is that Tajmar is not an Rf microwave engineer.

Suspect he called in Shawyer when he got no results.

Even with Shawyers help his magnetron runs at 200C, while the small end of the cavity runs at 35C. This suggests to me the reflected power is way higher than 336W and that the real power inside the cavity is much lower than 360Ws.

This again says us EMDrive DIYers need to be able to measure our VSWR or return loss dBs and to do adjustments to get them as optimal as possible so to at least have a fighting chance to see some Force generated.

In the acknowledgements, Tajmar et.al. thanks <<Prof. Plettemeier from TU Dresden who assisted our Q factor measurements.>>

Presumably, they have people at The Technische Universität Dresden that know about Electrical Engineering and Microwave cavities, and it looks like Prof. Plettemeier may be one of them


Prof. Dr.-Ing. Dirk Plettemeier

Head and Chair of the Department of RF and Photonics in the department of Electrical Engineering

Microwaves

Tel .: +49 351463-33941
http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_elektrotechnik_und_informationstechnik/ifn/hf/organisation

http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_elektrotechnik_und_informationstechnik/ifn/hf/forschung/hochfrequenzsysteme

http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_elektrotechnik_und_informationstechnik/ifn/hf/
forschung/mikrowellenphotonik/index_html/document_view?set_language=en


« Last Edit: 07/27/2015 02:10 AM by Rodal »

Online birchoff

  • Full Member
  • **
  • Posts: 267
  • United States
  • Liked: 122
  • Likes Given: 95
Quote from: birchoff
Whats the predicted thrust if you drop the 700W to 360W using the initial Q of 48.8? Also what do the numbers look like if you assume the model that was used was the one in equation 1 of The Development of a Microwave Engine for Spacecraft Propulsion

66uN is the Force prediction at a Q of 48.8.

Equation 1 in the 2005 Brighton paper is obsolete. My spreadsheet uses the Force equation in the current theory paper as attached.

I understand that the model in the paper I referenced is superseded. However I am assuming that since Tajmar didnt reference that paper you attached that the model used is the one from the paper I am referencing.

Equation 11 in the 2010 paper (referenced by Tajmar) is the current Force equation, which also details how to calc the Df.

thanks for the clarification.

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
polarized slot antenna, poor rl, surprised at any results...which is a good thing for the ongoing project.

All I can add is that Tajmar is not an Rf microwave engineer.

Suspect he called in Shawyer when he got no results.

Even with Shawyers help his magnetron runs at 200C, while the small end of the cavity runs at 35C. This suggests to me the reflected power is way higher than 336W and that the real power inside the cavity is much lower than 360Ws.

This again says us EMDrive DIYers need to be able to measure our VSWR or return loss dBs and to do adjustments to get them as optimal as possible so to at least have a fighting chance to see some Force generated.

In the acknowledgements, Tajmar et.al. thanks <<Prof. Plettemeier from TU Dresden who assisted our Q factor measurements.>>

Presumably, they have people at The Technische Universität Dresden that know about Electrical Engineering and Microwave cavities, and it looks like Prof. Plettemeier may be one of them


Prof. Dr.-Ing. Dirk Plettemeier

Head and Chair of the Department of RF and Photonics in the department of Electrical Engineering

Microwaves

Tel .: +49 351463-33941
http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_elektrotechnik_und_informationstechnik/ifn/hf/organisation

http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_elektrotechnik_und_informationstechnik/ifn/hf/forschung/hochfrequenzsysteme

http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_elektrotechnik_und_informationstechnik/ifn/hf/
forschung/mikrowellenphotonik/index_html/document_view?set_language=en
A Q of 50 is typical of an LC circuit and can't imagine an rf engineer using waveguides and a tuned cavity would be happy with this. Only thing I can think of is a Q of 50 roughly translates into the frequency spread of a typical magnetron...if u use the ctr freq/3db BW Q formula. Also, a normal thermal fuse for a kitchen magnetron is 160°C. Imo, mr t is right in rl generating heat. I also suspect the polarized slit reduced the effective radiation (erp) into the frustum even further. For a rough estimate, divide whatever the estimated return loss erp is by at least a factor of 2.

Online aero

  • Senior Member
  • *****
  • Posts: 2739
  • 92129
  • Liked: 704
  • Likes Given: 237
@Dr. Rodal

The new big and small end lateral antenna csv files are up and I have removed the previous version. This time they are in two folders so you won't get them mixed up. Naming convention is the same as my most recent previous data set uploaded, with only the base name changed, as you will see.

Each folder contains a Meep Data Request file to describe the run. Ask me for the information that I forgot to add.

https://drive.google.com/folderview?id=0B1XizxEfB23tfkF0Z184NHRtd0ViN28tNzRDY3JzSVc0WFBTOGZmSFZMcUpWLWJfcDRfZEU&usp=sharing

aero

I downloaded (after re-installing Google Drive, which was malfunctioning) the files for the case of the lateral antenna at the Small End of Yang/Shell, and I computed the stresses and forces.

Before posting the data, would you be so kind as to verify whether you obtained the big base and the small base at the following locations?

These are the locations where Mathematica shows the bases:


BIG BASE  :  row 15  (rows ranging from 0 to 228)

SMALL BASE : row 214 (rows ranging from 0 to 228)

Please note:

1) I adopted the same convention as yours:  rows starting at zero

2) when I plot the stress for the big base it looks very small, that's why I would like you to double-check whether the locations are correct

******************

Also, did you ever run csv files for rfmwguy lateral antenna at the Big Base?  All I recall is the antenna at the small end.
Here are a couple of lines from my bash shell file, picked at random, all the base cuts are the same.

h5totxt -t 09 -x 15 -o ./Shell-SELat-Ant-out/All-SE-Ant-csv/AxialAnt-09-hyBx.csv ./Shell-SELat-Ant-out/hy.h5
h5totxt -t 09 -x 214 -o ./Shell-SELat-Ant-out/All-SE-Ant-csv/AxialAnt-09-hySx.csv ./Shell-SELat-Ant-out/hy.h5

As you see, the hyBx.csv is at .h5 row 15 and the hySx.csv is at row 214. So yes, I cut at rows 15 and 214.

__________________________________

No, I don't recall ever having run having moved the lateral antenna to the big end for rfmwguy's NSF-1701 model. It's to late to run it tonight but I could run it tomorrow if you like.
Retired, working interesting problems

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5915
  • Likes Given: 5253
...
__________________________________

No, I don't recall ever having run having moved the lateral antenna to the big end for rfmwguy's NSF-1701 model. It's to late to run it tonight but I could run it tomorrow if you like.
Whenever you have a chance to run it, it would be very interesting if you could do it with the same mesh you used before for rfmwguy/NSF-1701 for the runs with the antenna at the small end

The reason for this is that it looks like SeeShell was right: at least for Yang/Shell it is better to have the lateral antenna at the big end than at the small end .

We need to check whether this is also the case for rfmwguy/NSF-1701
« Last Edit: 07/27/2015 02:38 AM by Rodal »

Offline zellerium

  • Full Member
  • *
  • Posts: 171
  • Pittsburgh, PA
  • Liked: 279
  • Likes Given: 400

...
Wish it were that simple, but folks it aint.

I ran through the equations and can't seem to get the numbers from that calculator, would you kindly show me what I'm doing wrong?

RL(dB) = 10*log10(Power incident/power reflected)
if RL = 12.5 dB, Power incident = 700 W, then power reflected is 39.3 W, not 339.9 W
or
RHO = 10^(-RL/20)
VSWR = (1+abs(RHO))/(1-abs(RHO)) = 1.62
MisMatch = 10*log10(1-RHO^2) = -0.251
%P_reflected = 100*RHO^2 =  5.623
which gives the same 39.3 W reflected.
Thank you

In regards to Tajmar's paper, I am curious as to why his thrust signature dropped in correspondence with the magnetron temperature. I don't believe any other experimenters observed this...
In a vacuum of 10^-6 Torr I would have a hard time believing it could be air currents. But I haven't run through the calcs yet.
Also, the difference in negative and positive horizontal thrusts tells me there must be some interaction that hasn't been accounted for. Reminds me of the difference in forward and reverse from the Cannae test in Brady et al.

Build update: we are almost finished manufacturing and aim to begin testing within a week or so on our symmetric partially loaded (with HDPE) cylindrical resonant cavity. Our laser measurement detector has been calibrated and can detect ~ 10 um in the morning (before people enter the building, AC turns on etc)  but we haven't put it on the pendulum yet which I'm sure will add considerable noise. 
I'll update again when we have some results :D

Offline demofsky

  • Full Member
  • *
  • Posts: 165
  • Liked: 96
  • Likes Given: 1662
This short length, constant diameter, cylindrical step out ("retraction" size of the cover depth) design also seems to eliminate the need for spherical end plates, which would be a really big win.

Yes but the Chinese used a noisy 1000W magnetron, which has AM and FM modulation over a wider bandwidth than your cleaner solid-state 100W RF amp. I remember Shawyer told you to use a magnetron with flat end plates, but a narrow-band emitter with spherical ends.

Correct.

But the new build Prof Yang has shared allows using flat end plates to obtain a Q of 117,500 by using short constant diameter set backs as attached. I assume the set backs convert a spherical wavefront into a planar wavefront in both directions.

When using a tapered waveguide to connect different diameter cylindrical waveguides this is what happens. Planar going in, spherical inside the tapered section and planar when emerging into the opposite side constant diameter cylindrical section.

I see this as a major advancement in EMDrive frustum design.

My 1st build has now altered to be able to experiment with this new structure as attached. Double ended tuning with sliding end plates inside constant diameter sections. Shawyer did do this for the small end in his Demonstrator EMDrive and repeated it with the Tajmar EMDrive. Both magnetron driven.

If this setup can eliminate the need for spherical end plates and still deliver an unloaded Q of 117,500, bring it on.

BTW to those "group velocity is different at each end of the frustum" doubters, wonder what the group velocity will be in those constant diameter sections? Would suggest it will be the group velocity as per the constant diameter circular waveguide equations. Which means Cullen, Shawyer & Prof Yang are correct.

This is very interesting.  However, something twigged prompting me to realize that we now have a fustrum and two cylinders!!  This must introduce some very interesting dynamics to an already complex situation.  Evanescent waves must be going crazy in there.

Seeing a meep simulation of one of our standard models with these extensions would be very, very illuminating.
« Last Edit: 07/27/2015 02:54 AM by demofsky »

Offline TheTraveller


...
Wish it were that simple, but folks it aint.

I ran through the equations and can't seem to get the numbers from that calculator, would you kindly show me what I'm doing wrong?

RL(dB) = 10*log10(Power incident/power reflected)
if RL = 12.5 dB, Power incident = 700 W, then power reflected is 39.3 W, not 339.9 W
or
RHO = 10^(-RL/20)
VSWR = (1+abs(RHO))/(1-abs(RHO)) = 1.62
MisMatch = 10*log10(1-RHO^2) = -0.251
%P_reflected = 100*RHO^2 =  5.623
which gives the same 39.3 W reflected.
Thank you

In regards to Tajmar's paper, I am curious as to why his thrust signature dropped in correspondence with the magnetron temperature. I don't believe any other experimenters observed this...
In a vacuum of 10^-6 Torr I would have a hard time believing it could be air currents. But I haven't run through the calcs yet.
Also, the difference in negative and positive horizontal thrusts tells me there must be some interaction that hasn't been accounted for. Reminds me of the difference in forward and reverse from the Cannae test in Brady et al.

Build update: we are almost finished manufacturing and aim to begin testing within a week or so on our symmetric partially loaded (with HDPE) cylindrical resonant cavity. Our laser measurement detector has been calibrated and can detect ~ 10 um in the morning (before people enter the building, AC turns on etc)  but we haven't put it on the pendulum yet which I'm sure will add considerable noise. 
I'll update again when we have some results :D

Yup my mistake. Thought the loss dB entry was return loss dBs. Turns out it is antenna loss as dBs.

VSWR is 1.622:1 and your forward and reflected calcs are correct.
« Last Edit: 07/27/2015 04:34 AM by TheTraveller »
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.”
Herman Melville, Moby Dick

Offline dumbo

  • Member
  • Posts: 15
  • I believe eleph^H^H^H^H^HEMDrives can FLY
  • Tianjin, China
  • Liked: 9
  • Likes Given: 77
Anybody know anything about SuperFish?

http://www.lanl.gov/projects/feynman-center/technologies/software/poisson-superfish.php

I know that the official download site is down at the moment. I was able to find a mirror site, so if anyone wants to play around with it, you can use the link:
http://hallaweb.jlab.org/12GeV/SoLID/download/field/PoissonSuperfish_7.18.exe

Tags: