Author Topic: EM Drive Developments - related to space flight applications - Thread 2  (Read 2171293 times)

Offline Star-Drive

  • Member
  • Full Member
  • ****
  • Posts: 829
  • TX/USA
  • Liked: 877
  • Likes Given: 9
@Paul March,
It is kind of important to my results if you could confirm that the Teflon Rubber gaskets are installed as illustrated in the attached model. Click on the image, it will expand so you can see detail - but of course it is mostly black so use the sliders to move around to find a corner. :)

Right now, I am using a 12.5 mm coaxial dipole antenna at the inner face of the dielectric disk. I know you used a loop of some sort. How much do you think this difference matters considering that I am running a digital model?

It is also important that I correctly model the width of the Teflon Rubber gasket filled gap. You wrote that the gasket was .064." Was that after installed, or did you compress it when you tightened down the retaining ring. If so, what would you estimate the actual distance is, between the copper cone and copper base plate, as installed? I know that sounds like a nonsense question, but my simulation shows thrust force is dramatically sensitive to just a small changes in the gap width. I'd like for my model to be as close as is possible to your Copper Kettle thruster.

My final question (I hope) re. the gasket is, "Do you know what the dielectric constant is for the actual Teflon Rubber that you used?" (Did your supplier document it, perhaps.) I find values ranging from 2.1 to 2.5 and while force is not very sensitive to this value, it does have an effect.

And while I'm at it, I read that the vacuum chamber is 30 inches by 36 inches, diameter by length. Is that inside or outside dimensions?

Aero:

"It is kind of important to my results if you could confirm that the Teflon Rubber gaskets are installed as illustrated in the attached model."

I think what you are talking about is the initial pressure seal design for our aluminum frustum cavity that later went to a silicone O-ring and metal to metal compression shorting pad just inside the O-ring for both the large and small OD ends of the frustum. 

The Eagleworks copper frustum is not a gas sealed unit, so all it has for its large and small OD end-cap interfaces are copper metal to copper metal interface with #6-32 brass cap-screws, nuts and bronze internal star lock washers spaced an average of 1.0" apart on the frustum's 0.50" wide copper flanges.  As to the average air gap between these copper flanges due to their out of plane irregularities, (These copper flanges are only 0.040" thick.), my guess is that it can be no larger than 0.002" midway between the cap-screws.     

"Right now, I am using a 12.5 mm coaxial dipole antenna at the inner face of the dielectric disk. I know you used a loop of some sort. How much do you think this difference matters considering that I am running a digital model?"

I've used various OD magnetic loops made from #20 AWG copper magnet wire soldered to SMA bulkhead connector that is mounted on the copper frustum's conical side wall, 15% of the of frustum Z-axis height from the large OD end of the frustum cavity, see attached picture.  Currently we are using a 14.0mm OD loop antenna for our TM212 work at 1,937.118 MHz work.

"And while I'm at it, I read that the vacuum chamber is 30 inches by 36 inches, diameter by length. Is that inside or outside dimensions?"

The Eagleworks vacuum chamber interior dimensions are as noted except the distance front aluminum door to the rear domed portion of the 304L stainless steel spun end cap is ~38.0", see attached Kurt J. Lesker drawing.  However, Our vacuum door is hinged on the right side of the chamber as viewed from the door end.
« Last Edit: 02/15/2015 02:55 PM by Star-Drive »
Star-Drive

Offline Star-Drive

  • Member
  • Full Member
  • ****
  • Posts: 829
  • TX/USA
  • Liked: 877
  • Likes Given: 9
Folks:

In the meantime, lets ask why 60 watts of relatively harmonic free sine-wave RF power at the 1,937.118 MHz AKA the TM212 resonant frequency in this copper frustum cavity, can only generate a paltry ~60uN, whereas the Chinese claimed to have produce 160,000uN using just ~150 watts of 2,450 MHz RF signals from a magnetron?  The magnetron RF signal source that is anything but a pure sine-wave generator, that instead has a modulated FM bandwidth of at least +/-30 MHz that is also concurrently amplitude modulated (AM) with thermal electron noise. 


Taking a critical look at this question, and knowing that the spectral shape of a magnetron looks like (see below) compared to a CW spike. It seems evident that a CW spike isn't the best waveform to use if you want to maximize thrust. Dollars to donuts says the Chinese are making full use of the available bandwidth of their resonant cavity by using that noisy magnetron. Magnetrons have lots of phase noise too. You can't easily use them on phased array radars because of that for example.

...
I agree with Mulletron that the answer to Paul March's question is that it is much more effective to have a distributed power spectral density than the power concentrated at a single frequency spike.  When the natural frequency changes in an unpredictable manner, it is much more effective to have a distributed power spectral density of excitation (it is the power spectral density ( http://en.wikipedia.org/wiki/Spectral_density#Power_spectral_density ) over the spectrum of changing natural frequencies that matters).

The reason for this is that (as has been verified by Prof. Juan Yang in China by inserting thermocouples at different places in the EM Drive) the EM Drive is subjected to a very non-uniform temperature distribution, with the temperature increasing with time, that results in significant non-uniform thermal expansion of the EM Drive, and therefore the natural frequencies must shift with temperature (and therefore shift with time as the temperature changes with time) as the EM Drive expands non-uniformly with time.  Therefore, having the power concentrated at a single frequency spike (NASA) is bound to be non-efficient as the resonant frequency changes with time, the EM Drive is going to move out of resonance even if one happens to excite it at the correct frequency to start with.  The COMSOL calculations do not provide the natural frequency to enough precision within the extremely narrow bandwidth of a high Q resonance (the higher the Q, the narrower the resonant bandwidth) for NASA to know exactly the natural frequency for a given mode shape.  More importantly, the COMSOL calculations do not provide the information needed for NASA to know how to shift the frequency with time, as the EM Drive thermally expands non-uniformly to stay at peak resonance.

This is evident from the very low Q's reported by NASA (7K to 22K) compared with the Chinese, who report a Q=117K

Quote from: Juan Yang
the resonant frequency and quality factor of the independent microwave resonator system are 2.44895 GHz and 117495.08 respectively

Compare this with NASA's reported Q:

Mode   Frequency (MHz)  Quality Factor, Q   Input Power (W)  Mean Thrust (μN)   Medium      Efficiency(uN/W)
TE012     1880.4               22000                         2.6                55.4                   Air           21
TM2112  1932.6                 7320                       16.9                91.2                   Air             5
TM2112  1936.7               18100                       16.7                50.1                   Air             3
TM212    1937.115             6726                       50                   66                      Vacuum      1

NASA's reported Q for the vacuum experiment is a meager Q = 6726, which is 17 times smaller than the Chinese reported Q = 117495.

Also note that the most efficient mode reported by NASA Eagleworks is the Transverse Electric mode which gave a Mean Thrust of 55 uN with only 2.6 Watts.



The Chinese also report that they used the Transverse Electric mode



Instead, NASA Eagleworks has been running most of the experiments in the Brady report in the Transverse Magnetic mode, and the vacuum experiment also in the Transverse Magnetic mode, which NASA's own data (see above) shows to be the most inefficient mode.

Why is NASA running the vacuum experiment in the most inefficient mode (Transverse Magnetic) rather than the most efficient mode (Transverse Electric) ?  Because they report difficulties in tuning the EM Drive under the Transverse Electric mode.

Quote from: Brady et.al page 17
Prior to the TM211 evaluations, COMSOL® analysis indicated that the TE012 was an effective thrust generation mode for the tapered cavity thruster being evaluated, so this mode was explored early in the evaluation process. Figure 22 shows a test run at the TE012 mode with an operating frequency of 1880.4 MHz. The measured quality factor was ~22,000, with a COMSOL prediction of 21,817. The measured power applied to the test article was measured to be 2.6 watts, and the (net) measured thrust was 55.4 micronewtons. With an input power of 2.6 watts, correcting for the quality factor, the predicted thrust is 50 micronewtons. However, since the TE012 mode had numerous other RF modes in very close proximity, it was impractical to repeatedly operate the system in this mode, so the decision was made to evaluate the TM211 modes instead.

Why does NASA have difficulties running the EM Drive in the more efficient mode (the Transverse Electric mode) ? Because the most efficient mode results in greater shifting of its natural frequency with time.  Hence I agree with Mulletron that instead of having the power concentrated at a frequency, for a problem where we know that the natural frequency of the EM Drive changes with time in a difficult to calculate and predict (with enough precision) manner, the best solution is to have the power distributed over a wider spectrum of frequencies, as done by Prof. Juan Yang in China.



Dr. Rodal:

You seem to ask a lot of "why" questions that could be better answered by getting yourself in the lab and finding out the sought after answers for yourself.  Be that as it may, the main reason that we went with the lower-Q TM modes was because they consistently produced higher thrust levels for a given input power than the TE modes.  I will grant you though that getting the most thrust out of a particular resonant mode depended very painfully on the size, placement and rotational orientation of the loop antenna in the frustum cavity.  And it may just have been that I didn't know how best to optimize their operations at the time, since this has been a rather large learning experience for me over these last three years in the Eagleworks Lab.

BTW, thanks much for the pointer to the 2014 Chinese report.  Is there an English translation of same out in public yet?  Also, in the 2013 Chinese report that had been translated into English, see attached, you will find that their large hundreds of milli-Newton thrust results were obtained with a loaded quality factor of just ~1,530 at 2.45 GHz, see figure 13 in their 2013 report.  We think that is because that like any ac electric induction motor, this device has to load down its input energy/power source as it is generating thrusting work.  Which brings up another point.  That being all the calculated Q-Factors given in the Chinese papers, unless otherwise stated, is the very idealized unloaded Q-factors that implies that no energy is being extracted from the resonant cavity.   We must keep that fact in mind as well...

Best, Paul M.
Star-Drive

Offline Star-Drive

  • Member
  • Full Member
  • ****
  • Posts: 829
  • TX/USA
  • Liked: 877
  • Likes Given: 9
Folks:

In the meantime, lets ask why 60 watts of relatively harmonic free sine-wave RF power at the 1,937.118 MHz AKA the TM212 resonant frequency in this copper frustum cavity, can only generate a paltry ~60uN, whereas the Chinese claimed to have produce 160,000uN using just ~150 watts of 2,450 MHz RF signals from a magnetron?  The magnetron RF signal source that is anything but a pure sine-wave generator, that instead has a modulated FM bandwidth of at least +/-30 MHz that is also concurrently amplitude modulated (AM) with thermal electron noise. 


Taking a critical look at this question, and knowing that the spectral shape of a magnetron looks like (see below) compared to a CW spike. It seems evident that a CW spike isn't the best waveform to use if you want to maximize thrust. Dollars to donuts says the Chinese are making full use of the available bandwidth of their resonant cavity by using that noisy magnetron. Magnetrons have lots of phase noise too. You can't easily use them on phased array radars because of that for example.

Now to put this idea to test, Q: What is the bandwidth of the resonant cavity and what is the 90 percent power bandwidth of the signal you are driving it with? What kind of sig gen are you using? Can it do FM? Can you do any advanced waveforms like a PSK waveform? Do you have a way to produce wideband noise or a spread spectrum carrier for your testing? Can you do any waveforms like at the bottom?

Also during researching other possible theories which could explain Emdrive we found ample literature stating that molecules acquire a kinetic momentum during the switching of the magnetic field as a result of its interaction with the vacuum field. If correct, that may well be a very significant lead. So that raises the question, how does one increase the switching rate? What about phase shifting? http://en.wikipedia.org/wiki/Phase-shift_keying

Phase shifting seems important.
https://www.viasat.com/files/assets/web/datasheets/EBEM_MD-1366_043_web.pdf
One of these driving your amp would be helpful. They go up to 2ghz.







Mulletron:

When using our current 14mm loop antenna optimized for the TM212 resonance at 1,937.118 MHz in our copper frustum, there were four RF resonances spaced +/- 30 MHz around the 2.45 GHz center frequency.  And I assume that would also be the case using a higher power slot antenna in a similar location as the Chinese and Shawyer have done with their frustum resonant cavities.  So yes, a wide bandwidth RF source seems to be called for and one that can be both AM and FM modulated at the same time.  From my readings to date, that appears to be a hard nut to crack for solid state RF amplifiers at the desired kW power levels due to their limited RF power bandwidth capabilities, so we may be forced into using magnetrons and just learn how best to feed their 4-to-20 kV high voltage anode requirements while working in a hard vacuum.  However the more difficult problems are finding ways of reducing their mass and size so we can "fly" them on our torque pendulum.  Cooling the magnetrons in a hard vacuum is also another problem we need to deal with since air cooling is out of the question and liquid cooling is a giant pain to deal with as well.  About the only other way to cool these beasts in a hard vacuum is to use phase change material like paraffin wax that could give us several minutes of run times before we had to let the accumulated heat in the paraffin radiate to the vacuum chamber walls.

BTW, the paraffin wax phase change cooling was used to good advantage on the lunar moon buggy used by NASA astronauts during their lunar explorations in the late 1960s and early 1970s during the USA Apollo Moon program.

Best, Paul M.
Star-Drive

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
...
Dr. Rodal:

You seem to ask a lot of "why" questions that could be better answered by getting yourself in the lab ...
When I was in charge of R&D laboratories for private companies I always welcomed such questions, particularly questions from outsiders  that perhaps had not been asked by internal staff, because the purpose of R&D is always to find the truth as fast as possible, hence I saw it as a great benefit to our R&D efforts to get such questions.   That was also the attitude of the Professors at MIT, I was lucky to work with.
« Last Edit: 02/15/2015 04:31 PM by Rodal »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
... the main reason that we went with the lower-Q TM modes was because they consistently produced higher thrust levels for a given input power than the TE modes.  ...
The data reported by NASA Eagleworks contradicts that TM modes produce higher thrust levels than TE modes for a given input power:

Mode   Frequency (MHz)  Quality Factor, Q   Input Power (W)  Mean Thrust (μN)   Medium      Efficiency(uN/W)
TE012     1880.4               22000                         2.6                55.4                   Air           21
TM2112  1932.6                 7320                       16.9                91.2                   Air             5
TM2112  1936.7               18100                       16.7                50.1                   Air             3
TM212    1937.115             6726                       50                   66                      Vacuum      1

Apparently, the emphasis should be on the use of the word "consistently" either as that

a) the reported Brady et.al. data for TE012 was inconsistent, in which case the Mean Thrust was not 55.4 uN as reported by Brady et.al. but apparently there is other unreported data that NASA has, giving significantly lower values of the thrust (and accordingly the reported Mean by Brady et.al. was not the true Mean of the TE012 experiments with the dielectric at NASA)

or

b) my interpretation of the Brady et.al.'s statement that NASA Eagleworks had trouble staying tuned at the natural frequency near  1880.4  MHz

Quote from: Brady et.al page 17
Prior to the TM211 evaluations, COMSOL® analysis indicated that the TE012 was an effective thrust generation mode for the tapered cavity thruster being evaluated, so this mode was explored early in the evaluation process. Figure 22 shows a test run at the TE012 mode with an operating frequency of 1880.4 MHz. The measured quality factor was ~22,000, with a COMSOL prediction of 21,817. The measured power applied to the test article was measured to be 2.6 watts, and the (net) measured thrust was 55.4 micronewtons. With an input power of 2.6 watts, correcting for the quality factor, the predicted thrust is 50 micronewtons. However, since the TE012 mode had numerous other RF modes in very close proximity, it was impractical to repeatedly operate the system in this mode, so the decision was made to evaluate the TM211 modes instead.
« Last Edit: 02/15/2015 06:26 PM by Rodal »

Offline zen-in

  • Full Member
  • ****
  • Posts: 531
  • California
  • Liked: 468
  • Likes Given: 365

BTW, thanks much for the pointer to the 2014 Chinese report.  Is there an English translation of same out in public yet?  Also, in the 2013 Chinese report that had been translated into English, see attached, you will find that their large hundreds of milli-Newton thrust results were obtained with a loaded quality factor of just ~1,530 at 2.45 GHz, see figure 13 in their 2013 report.  We think that is because that like any ac electric induction motor, this device has to load down its input energy/power source as it is generating thrusting work.  Which brings up another point.  That being all the calculated Q-Factors given in the Chinese papers, unless otherwise stated, is the very idealized unloaded Q-factors that implies that no energy is being extracted from the resonant cavity.   We must keep that fact in mind as well...

Best, Paul M.

There may be a more conventional reason for the Q to decrease.  Excess RF power may just be getting dissipated inside the cavity as heat.   If the Chinese test device was in a static test fixture, like a torsion balance, and there was no movement in the direction of the measured force, then no work was done. 
« Last Edit: 02/15/2015 04:27 PM by zen-in »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
BTW, thanks much for the pointer to the 2014 Chinese report.  Is there an English translation of same out in public yet?
I have not seen an English translation of this paper posted in the Internet:

Resonance experiment on a microwave resonator system
Shi Feng Yang Juan Tang Ming-Jie Luo Li-Tao Wang Yu-Quan
(College of Astronautics, Northwestern Polytechnic University, Xi’an 710072, China)
Acta Phys. Sinica Vol. 63, No. 15 (2014)

I have only seen the Chinese original that I attached.  As to translations, my experience is that it is always best to seek a meaningful translation from people skilled in the art: in this case from people conversant in the language of the paper who are also engineers/scientists in the same field discussed by the paper.
« Last Edit: 02/15/2015 04:27 PM by Rodal »

Offline aero

  • Senior Member
  • *****
  • Posts: 2786
  • 92129
  • Liked: 724
  • Likes Given: 249
@Paul March,
It is kind of important to my results if you could confirm that
             ... snip ...
Paul - Thank you for this very helpful information

Quote
Aero:

"It is kind of important to my results if you could confirm that the Teflon Rubber gaskets are installed as illustrated in the attached model."

I think what you are talking about is the initial pressure seal design for our aluminum frustum cavity that later went to a silicone O-ring and metal to metal compression shorting pad just inside the O-ring for both the large and small OD ends of the frustum. 

The Eagleworks copper frustum is not a gas sealed unit, so all it has for its large and small OD end-cap interfaces are copper metal to copper metal interface with #6-32 brass cap-screws, nuts and bronze internal star lock washers spaced an average of 1.0" apart on the frustum's 0.50" wide copper flanges.  As to the average air gap between these copper flanges due to their out of plane irregularities, (These copper flanges are only 0.040" thick.), my guess is that it can be no larger than 0.002" midway between the cap-screws.     

Good good, very good. I have ran many cases, attached find a curve showing the EM Thruster force sensitivity to gap size. Ignore the units - this curve is only representative of force .vs. gap size for what I would consider to be a bare thruster in isolation, perhaps in orbit for example. But do note that the smallest gap corresponds to 0.011 inches, significantly larger than 0.002 inches. This was a 2D meep simulation, limited by the computer power of my home desktop machine.

Quote
"Right now, I am using a 12.5 mm coaxial dipole antenna at the inner face of the dielectric disk. I know you used a loop of some sort. How much do you think this difference matters considering that I am running a digital model?"

I've used various OD magnetic loops made from #20 AWG copper magnet wire soldered to SMA bulkhead connector that is mounted on the copper frustum's conical side wall, 15% of the of frustum Z-axis height from the large OD end of the frustum cavity, see attached picture.  Currently we are using a 14.0mm OD loop antenna for our TM212 work at 1,937.118 MHz work.

Thank you - I'll look again at my antenna placement. Unfortunately I don't know of a way to simulate the loop antenna in 2D if it is not coplanar.
 
Quote

"And while I'm at it, I read that the vacuum chamber is 30 inches by 36 inches, diameter by length. Is that inside or outside dimensions?"

The Eagleworks vacuum chamber interior dimensions are as noted except the distance front aluminum door to the rear domed portion of the 304L stainless steel spun end cap is ~38.0", see attached Kurt J. Lesker drawing.  However, Our vacuum door is hinged on the right side of the chamber as viewed from the door end.

 Hmm. So it is even larger than I am simulating. That will add CPU cycles to the simulation runs. But not so much because the model is already to large to run in 3D on my machine. I've looked at renting a compute engine from Google to overcome this problem but ...
Retired, working interesting problems

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
FYI for no good reason, here what I get w/o dielectric

Mode   Frequency (MHz)  Quality Factor, Q   Input Power (W)  Mean Thrust (μN)   Calc
TE012  1880.4                  22000                  2.6                       55.4                        16.9
TM212  1932.6                   7320                  16.9                      91.2                        60.5
TM212  1936.7                 18100                  16.7                      50.1                       146.9
TM212  1937.115               6726                  50                         66                          163.3

How much of the table do we have for the Chinese ?
The dimensions of the EM Drive are needed to estimate thrust in your formula as well as Shawyer's and McCulloch's formula.   I could not find all the dimensions needed of the Chinese EM Drive tested by Prof. Juan Yang in any of their Chinese papers.  My recollection is that @aero also tried to look for those dimensions in the Yang papers translated to English, and was unable to find the dimensions either.

We (colleagues in Thread 1 and 2 of EM Drive Developments at NASA Spaceflight.com) never worked out an estimate for the dimensions of the Chinese EM Drive as it was done for the estimates of the NASA Eagleworks truncated cone (an estimate by @aero and @Muletron that turned out to be very close to the actual dimensions recently reported by Paul March) or for Shawyer's Experimental and Demonstration truncated cones.

Dr. McCulloch has used the same dimensions for the EM Drive tested by Prof. Juan Yang as the dimensions of the Shawyer Demonstration EM Drive (see  http://physicsfromtheedge.blogspot.com/2015/02/mihsc-vs-emdrive-data-3d.html) to estimate thrust of the Chinese EM Drive, and gets a reasonable comparison.

If we assume that the drawing from the Chinese paper is to scale, that gives us the ratios of the dimensions, and as a consequence we only need to estimate one of the dimensions (for example one of the diameters). It looks from this drawing (see below) that the axial length of the Juan Yang truncated cone is significantly shorter than the small diameter, and therefore the Juan Yang truncated cone is significantly different from the one tested at NASA Eagleworks and Shawyer's Demo and Experimental truncated cones.

I estimate from this drawing:

Big Diameter     = D1
Small Diameter = D2
Axial Length      = H


Ratio of Big Diameter to Small Diameter = 1.61
Ratio of Big Diameter to axial Length =  2.38




Hence
Ratio of Small Diameter to axial Length =  1.48
(Big Diameter - Small Diameter)/axialLength = 0.91
Tangent of Cone's half angle = 0.4538
Cone's half-angle = 24 degrees



For reference. the tangent of the cone's half angle thetaw and the cone's half angle thetaw, for the following cases are:


Example (and geometry)                    { Tan[thetaw],thetaw (degrees) }

Shawyer Experimental                        {0.104019,   5.93851}
Shawyer Demo                                   {0.219054, 12.3557}
NASA Eagleworks frustum                   {0.263889, 14.7827}
Egan's example                                  {0.36397 ,  20}
Prof. Juan Yang  (2014)                      {0.4538,     24.4 }
Shawyer Superconducting 2014          {0.7002,     35}




Shawyer's latest (2014) superconducting design (see image, presented at the IAC 2014 conference in Toronto), for which there are no experimental results reported yet, appears to have a significantly larger cone angle than his previous experimental and demo geometries, and significantly larger than NASA's Brady et.al.'s.

Shawyer's (2014) superconducting EM Drive design has a cone angle  thetaw of about 35 degrees.  It is the only EM Drive designed with spherical ends, which make much more sense -from an electromagnetic wave propagation viewpoint, particularly for large cone angle geometries- than the flat ends used by NASA Eagleworks, and the other EM Drives







(*
"Shawyer EXPERIMENTAL geometry"

shawyerExpLength=0.156 meter;
shawyerExpBigDiameter=0.16 meter
shawyerExpSmallDiameter=0.127546 meter;
*)

(*
" Shawyer DEMO geometry"

shawyerDemoLength=0.345 meter;
shawyerDemoBigDiameter=0.28 meter;
shawyerDemoSmallDiameter=0.128853 meter;

*)
« Last Edit: 02/15/2015 06:50 PM by Rodal »

Offline aero

  • Senior Member
  • *****
  • Posts: 2786
  • 92129
  • Liked: 724
  • Likes Given: 249
Quote
It looks from this drawing (see below) that the axial length of the Juan Yang truncated cone is significantly shorter than the small diameter, and therefore the Juan Yang truncated cone is significantly different from the one tested at NASA Eagleworks and Shawyer's Demo and Experimental truncated cones.

Which means to me that the mode is probably like TX y,z,1. That is, the cavity is only one-half wavelength in height.
Retired, working interesting problems

Offline RotoSequence

  • Full Member
  • ****
  • Posts: 782
  • Liked: 572
  • Likes Given: 782
I wonder what kind of cash it costs to get some quality time with an additive/subtractive hybrid manufacturing 3D Printer/5 axis mills like the Lasertech series by DMG MORI, seen here fabricating a turbine housing:



These kinds of printers are capable of working with a variety of materials, including aluminum, steel alloys, inconel, titanium, etc. I suspect hybrid manufacturing of an Inconel cavity or other alloy could be the easiest, least expensive way of acquiring high precision resonance cavities with relatively consistent operational characteristics, but I have no idea how much it costs to work with these sorts of machines.
« Last Edit: 02/15/2015 05:37 PM by RotoSequence »

Offline aero

  • Senior Member
  • *****
  • Posts: 2786
  • 92129
  • Liked: 724
  • Likes Given: 249
I wonder what kind of cash it costs to get some quality time with an additive/subtractive hybrid manufacturing 3D Printer/5 axis mills like the Lasertech series by DMG MORI, seen here fabricating a turbine housing:



These kinds of printers are capable of working with a variety of materials, including aluminum, steel alloys, inconel, titanium, etc. I suspect hybrid manufacturing of an Inconel cavity or other alloy could be the easiest, least expensive way of acquiring high precision resonance cavities with relatively consistent operational characteristics, but I have no idea how much it costs to work with these sorts of machines.


As I have remarked before, to make high quality precision cavities, all you need is a spinner machine from the 20th century. Just spin the copper sheet metal over a steel mandrel, then punch-press the end caps, or use the same circuit boards that Eagleworks uses. Problem with the circuit board is that the gaps are in the horizontal direction which is less efficient at producing thrust than gaps in the vertical direction. At least that is true if meep and the differential Maxwell equations are correctly characterizing the forces.

Making cavities in this way should cost less than $5 each once the machinist has been paid for turning out the mandrel and the punch tools. Of course we could ask Mulletron for an estimate of the cost of copper sheet in bulk, it might be more.
Retired, working interesting problems

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
I wonder what kind of cash it costs to get some quality time with an additive/subtractive hybrid manufacturing 3D Printer/5 axis mills like the Lasertech series by DMG MORI, seen here fabricating a turbine housing:



These kinds of printers are capable of working with a variety of materials, including aluminum, steel alloys, inconel, titanium, etc. I suspect hybrid manufacturing of an Inconel cavity or other alloy could be the easiest, least expensive way of acquiring high precision resonance cavities with relatively consistent operational characteristics, but I have no idea how much it costs to work with these sorts of machines.

A 3D printer that would enable fabrication of an EM Drive with spherical ends and a cone half angle~35 degrees or higher would be most advantageous (Invar, as you suggest because it would greatly reduce thermal expansion, and hence allow easier tuning to a stable natural frequency).  Also one that would allow fabrication of a superconducting cavity (Invar+YBCO).
« Last Edit: 02/15/2015 06:34 PM by Rodal »

Offline TMEubanks

Thoughts on Axions. (I apologize if this has been already covered.)

The question as to whether the EM Drive could be coupling to the Axion background came up on a different forum. I am dubious, and thought it would be useful to post why.

A recent review of axions as CDM: http://www.pnas.org/.../2015/01/07/1308788112.full.pdf

Current constraints on the axion mass constrain it to be in the range ~ 1 - 1000 micro-eV, and a 2 GHz axion would correspond to 8.2 micro-eV, so that's OK, maybe the EM drive couples to the axion mass. But, check out the Axion Dark-Matter Experiment in the PNAS article. That is much much more sensitive than the EM Drive - they are looking for yoctowatts (10^-24) of RF power in the 2 - 20 micro-eV range, precisely the range of the EM Drive, by tuning the cavity's resonant frequency to the axion mass. There is simply no way that the Drive is coupling to the axion background - the ADMX would see a whopping signal. Now, maybe the ADMX is producing lots of axions, at a low enough velocity to evade the photon rocket limits. That would mean that the EM Drive can convert watts of RF -> axions, while the ADMX is NOT converting 10^-24 watts of dark matter axions -> RF power. While I guess it is possible, I just don't buy it. Nobody is accidentally lucky by 20+ orders of magnitude. (Note also that the theory is pretty clear here - if axions have these kinds of 1 - 10 micro-eV masses, they will supply most or all of the DM, so there will be a background.)

However, there is an easy test - just take an EM drive to the University of Washington, and see if you get a reaction from the ADMX. That should done at some point if the drive continues to pass experimental muster.

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
...
The question as to whether the EM Drive could be coupling to the Axion background came up on a different forum. I am dubious, and thought it would be useful to post why.

A recent review of axions as CDM: http://www.pnas.org/.../2015/01/07/1308788112.full.pdf

...

Link broken ?

I get the message "Proceedings of the National Academy of Sciences of the United States of America
Not Found The page you were trying to reach could not be found." using either Google Chrome or Mozilla Firefox
« Last Edit: 02/15/2015 06:51 PM by Rodal »

Offline SleeperService

  • Member
  • Posts: 48
  • Liked: 15
  • Likes Given: 15
Maybe the error message should have said:

Not found: The axions you are looking for do not exist.

 ;)

Offline Star-Drive

  • Member
  • Full Member
  • ****
  • Posts: 829
  • TX/USA
  • Liked: 877
  • Likes Given: 9
... the main reason that we went with the lower-Q TM modes was because they consistently produced higher thrust levels for a given input power than the TE modes.  ...
The data reported by NASA Eagleworks contradicts that TM modes produce higher thrust levels than TE modes for a given input power:

Mode   Frequency (MHz)  Quality Factor, Q   Input Power (W)  Mean Thrust (μN)   Medium      Efficiency(uN/W)
TE012     1880.4               22000                         2.6                55.4                   Air           21
TM2112  1932.6                 7320                       16.9                91.2                   Air             5
TM2112  1936.7               18100                       16.7                50.1                   Air             3
TM212    1937.115             6726                       50                   66                      Vacuum      1

Apparently, the emphasis should be on the use of the word "consistently" either as that

a) the reported Brady et.al. data for TE012 was inconsistent, in which case the Mean Thrust was not 55.4 uN as reported by Brady et.al. but apparently there is other unreported data that NASA has, giving significantly lower values of the thrust (and accordingly the reported Mean by Brady et.al. was not the true Mean of the TE012 experiments with the dielectric at NASA)

or

b) my interpretation of the Brady et.al.'s statement that NASA Eagleworks had trouble staying tuned at the natural frequency near  1880.4  MHz

Quote from: Brady et.al page 17
Prior to the TM211 evaluations, COMSOL® analysis indicated that the TE012 was an effective thrust generation mode for the tapered cavity thruster being evaluated, so this mode was explored early in the evaluation process. Figure 22 shows a test run at the TE012 mode with an operating frequency of 1880.4 MHz. The measured quality factor was ~22,000, with a COMSOL prediction of 21,817. The measured power applied to the test article was measured to be 2.6 watts, and the (net) measured thrust was 55.4 micronewtons. With an input power of 2.6 watts, correcting for the quality factor, the predicted thrust is 50 micronewtons. However, since the TE012 mode had numerous other RF modes in very close proximity, it was impractical to repeatedly operate the system in this mode, so the decision was made to evaluate the TM211 modes instead.


Dr. Rodal:

Yes, consistency of results has been a large problem for us.  And its cause is centered around consistently frequency locking the voltage controlled oscillator (VCO) frequency to the always drifting high-Q resonant frequency of the resonant cavity as it heats up while under power.  We tried building a mixer based phased lock loop (PLL) circuit and box that will work in the vacuum chamber, but found that for proper and consistent frequency lock-on, it was very sensitive to the amplitude of the sense antenna signal coming back from the resonant cavity.  Running out of budget & time, we had to give up on it for the moment when it became apparent that we also had to design, build and integrate an automatic gain control circuit for the PLL that could fit into the already physically constrained PLL box.  Thus, right now we are still using the "Paul Locked Loop" approach with me acting as the frequency control feedback loop for each test.  We've talked to our RF group here at JSC about better solutions, and they suggested using a S11 return loss based sensing circuit, which is much less sensitive to input amplitude variations, but again we don't have the budget to have one designed & built for our vacuum based application.  In other words you do what you can do and then move on...

Best, Paul M.
Star-Drive

Offline Notsosureofit

  • Full Member
  • ****
  • Posts: 656
  • Liked: 704
  • Likes Given: 1367
Just direct search

1308788112.full.pdf

Offline Notsosureofit

  • Full Member
  • ****
  • Posts: 656
  • Liked: 704
  • Likes Given: 1367
@ Paul

Back in the 60's we just used a cavity based oscillator  (that one was reentrant)

Edit:  50's,  memory must be slippin
« Last Edit: 02/15/2015 07:08 PM by Notsosureofit »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
...
The question as to whether the EM Drive could be coupling to the Axion background came up on a different forum. I am dubious, and thought it would be useful to post why.

A recent review of axions as CDM: http://www.pnas.org/.../2015/01/07/1308788112.full.pdf

...

[/quote]
Just direct search

1308788112.full.pdf

Thanks  :)

This is the link to the paper on Dark Matter search posted by Mashall:

http://www.pnas.org/content/early/2015/01/07/1308788112.full.pdf

Dark-matter QCD-axion searches
Leslie J Rosenberg

Abstract
In the late 20th century, cosmology became a precision science. Now, at the beginning of the next century, the parameters describing how our universe evolved from the Big Bang are generally known to a few percent. One key parameter is the total mass density of the universe. Normal matter constitutes only a small fraction of the total mass density. Observations suggest this additional mass, the dark matter, is cold (that is, moving nonrelativistically in the early universe) and interacts feebly if at all with normal matter and radiation. There’s no known such elementary particle, so the strong presumption is the dark matter consists of particle relics of a new kind left over from the Big Bang. One of the most important questions in science is the nature of this dark matter. One attractive particle dark-matter candidate is the axion. The axion is a hypothetical elementary particle arising in a simple and elegant extension to the standard model of particle physics that nulls otherwise observable CP-violating effects (where CP is the product of charge reversal C and parity inversion P) in quantum chromo dynamics (QCD). A light axion of mass 10−(6–3) eV (the invisible axion) would couple extraordinarily weakly to normal matter and radiation and would therefore be extremely difficult to detect in the laboratory. However, such an axion is a compelling dark-matter candidate and is therefore a target of a number of searches. Compared with other particle dark-matter candidates, the plausible range of axion dark-matter couplings and masses is narrowly constrained. This focused search range allows for definitive searches, where a nonobservation would seriously impugn the dark-matter QCD-axion hypothesis. Axion searches use a wide range of technologies, and the experiment sensitivities are now reaching likely dark-matter axion couplings and masses. This article is a selective overview of the current generation of sensitive axion searches. Not all techniques and experiments are discussed, but I hope to give a sense of the current experimental landscape of the search for dark-matter axions.
« Last Edit: 02/15/2015 07:23 PM by Rodal »

Tags: