Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1800714 times)

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I was able to confirm that it is RFI affecting the scale causing the apparent changes in force.   I used a rubber duck antenna suspended above the scale and was able to reproduce the ~30mg change with 30mw of net power, which seems like a plausible leakage value from the adjustable end which is not well sealed.  That end is closest to the scale in the "Up" orientation that produced the largest force changes.
Based on your data, this was the expected conclusion. I have made the same mistake myself with other experiments. If one insists on using an electronic balance, put it in a Faraday cage. Fo those without access to a good machine shop, this can be as simple as a lidded plastic box internally lined with copper tape that forms a closed conductor when closed. But in any case follow the mantra "calibrate, calibrate, calibrate".

Good to see that you've done that. But never rest on your laurels - when you re-invent your thrust-measuring apparatus to take care of this problem, never stop recalibrating with a null device.

As I've mentioned here before, one of the best devices for measurement is the Mettler H20, a miracle of Swiss engineering, fully mechanical, available from time to time on eBay for a not unreasonable sum, and gets you 0.1 microNewton force resolution (10 microgram-weight). A friend of mine has automated it using a laser position measuring device, and gets even better resolution plus electronic data logging capability.

The downside of the Mettler is that it takes a maximum of 200 gm. This limitation can be finessed by judicious use of a mechanical arrangement to balance out the dead weight without sacrificing resolution.


« Last Edit: 06/25/2015 06:00 AM by deltaMass »

Offline OttO

  • Member
  • Posts: 81
  • France
  • Liked: 91
  • Likes Given: 11
@aero

I would be a lot interested in a run with two additional equal sources of microwaves OUTSIDE the frustum. One toward the big side and one toward the small side. All the other parameters the same as your previous runs.

If I am correct and if MEEP is able to simulate Wood anomalies (plasmons), we could end with one OUTSIDE end surface of the frustum less reflective than the other to microwaves.

http://cpb.iphy.ac.cn/EN/article/downloadArticleFile.do?attachType=PDF&id=26571

What do you think of that?

PS: If it is not a silly idea, and if the DIY lab is near a TV transmitter or an airport radar we could have funny results :P

 http://juluribk.com/tag/free-software/
« Last Edit: 06/25/2015 08:37 AM by OttO »

Offline TheTraveller


As I've mentioned here before, one of the best devices for measurement is the Mettler H20, a miracle of Swiss engineering, fully mechanical, available from time to time on eBay for a not unreasonable sum, and gets you 0.1 microNewton force resolution (10 microgram-weight). A friend of mine has automated it using a laser position measuring device, and gets even better resolution plus electronic data logging capability.

The downside of the Mettler is that it takes a maximum of 200 gm.

This one got away:
http://m.ebay.com/itm/250916276083?_mwBanner=1

Like it:
http://vi.raptor.ebaydesc.com/ws/eBayISAPI.dll?ViewItemDescV4&item=250916276083&category=11814&pm=1&ds=0&t=1435213901184
« Last Edit: 06/25/2015 06:36 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 deuteragenie

  • Member
  • Posts: 71
  • Germany
  • Liked: 22
  • Likes Given: 0
Is the MEEP model's finite difference grid fine enough and the MEEP eigensolution HarmInv well-conditioned enough to successfully predict the frequencies measured by NASA and other experimenters, using their geometrical dimensions?

@ Rodal
Don't you mean, "The frequencies that COMSOL indicated that they should use? And the answer is that Harminv only comes close. And would you care to give an error bound on the measured (stated) dimensions of the cavities and the sensitivity of resonance to those measurements? I don't expect Harminv to reproduce the COMSOL numbers even if I do input the same numbers and precision used but we don't know what was used, do we. As for the experimental data, we have the same problems in spades. So if you could tell me what the resonance frequency sensitivities to small diameter and length are, that would be very helpful. Then we could estimate probable measurement errors and see if they are realistic. And if you can't tell me what the sensitivities are, then I can tell you, by using numerical data.

I doubt that Paul made a measurement error by as much as a tenth of an inch but unless he used a large micrometer to measure the height, he could have. And even with a micrometer, unless he was very very careful he could have introduced a slight angle to his measurement.

My  point is that you know as well as I that my computer is not up to running high resolution in 3D but Harminv does do much better in 3D than it ever did in 2D or cylindrical coordinates.

And just so you will know, over the last nearly 10 years, meep has been downloaded over 10,000 times. Some, if not most of the downloaders used meep, and many of them conducted and published peer reviewed research papers based on meep results. Meep is still widely used and does not have a reputation for frequency errors. The one thing those users may have had access to that I don't yet have is a powerful computer. Mine is a good home desk-top but at 5 years old, it is not a supercomputer. Go ahead and knock my computer all  you want but please lay off of meep.

And to answer your question as asked, "Yes, Meep absolutely does have the capability to measure resonance frequencies as well or better than other tools. I just do not have the needed tools installed. Harminv, not so much."

To install MPB and recent meep upgrades, I need to compile, link and load from C++ source code. That code is available but I am not a computer systems administrator or a professional C++ programmer and I do not want to stop producing some helpful results to produce nothing for the time it will take me to become knowledgeable enough to do that. Then take the time to learn to use the newly installed and upgraded program features. And then only to have my results flawed my my own modelling errors with many more potential sources of error. My system does what it does and if someone doesn't like it they can choose not to consider it.

And to the other 1,499,999 readers of this thread, I apologize for my rant.

Dear aero,

Thank you so much for all your efforts in this project ! I guess what is happening is that we all get very excited with the rapid "turnover" that simulation allows to achieve, as you have demonstrated time and over again. So we ask, and ask more and keep asking more... I guess it boils down to this: we need more people running simulations, as we need more people running experiments, we need the latest version of the source code compiled and packaged properly (something the Meep project should have done...), we also need a link to the Meep folks @ Meep forum, for knowledge sharing and harvesting their experience, and finally, we need an online Meep submission site with a distributed computing backend (BOINC based ?).  Am I asking too much ? Yes ! Do I have time to help ? No !

Offline KittyMoo

  • Member
  • Posts: 23
  • Liked: 8
  • Likes Given: 33
This is a good guide for installation of MEEP on Windows.
http://novelresearch.weebly.com/installing-meep-in-windows-8-via-cygwin.html

Offline OttO

  • Member
  • Posts: 81
  • France
  • Liked: 91
  • Likes Given: 11
Simulation of optical resonators using DGTD and FDTD
http://tinyurl.com/of3uh5w

It is found that FDTD suffers from phase errors and is limited by the
staircasing approximation. A further restriction stems from only second-order accuracy which
limits the geometrical problem size that can be analysed with given computational hardware.
Particularly for simulations of high-Q optical resonators, those problems prevent sufficient
convergence with reasonable grid spacing. The DGTD method, on the other hand, allows for
the approximation of curved surfaces with high accuracy using triangular elements.

 :)
« Last Edit: 06/25/2015 01:24 PM by Chris Bergin »

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I find this fascination with simulation rather curious, given the fact that all simulators seek to follow Maxwell's equations as accurately as their computational methods allow, and that Maxwell's equations predict zero thrust for the EmDrive. So what is it precisely that's so interesting about simulating the fields inside the cavity?

Offline deuteragenie

  • Member
  • Posts: 71
  • Germany
  • Liked: 22
  • Likes Given: 0
I find this fascination with simulation rather curious, given the fact that all simulators seek to follow Maxwell's equations as accurately as their computational methods allow, and that Maxwell's equations predict zero thrust for the EmDrive. So what is it precisely that's so interesting about simulating the fields inside the cavity?

The patterns and colors are nice :)

What I find puzzling in @aero's simulations is the change of colors (field strength?) outside the cavity. What can explain such change ? Or is this an artifact of the colorization algorithms ? Or am I missing something obvious ?

« Last Edit: 06/25/2015 09:25 AM by deuteragenie »

Offline OttO

  • Member
  • Posts: 81
  • France
  • Liked: 91
  • Likes Given: 11

Frankly, between the proposal that the EM Drive somehow "knows" its velocity so that it cannot become a free-energy machine and this proposal that the EM Drive has to have an unspecified level of vibration amplitude and frequency to exert a force... well I better stop here. :)


I found that (I know, I know it is not a frustum  :D):

Microwave frequency electromagnetic coupling to a thin membrane as one end of a cylindrical cavity
http://adsabs.harvard.edu/abs/2015APS..MARB35006C

This experiment shows that the TE011 mode gives rise to radiation pressure on the ends of a cylindrical cavity and demonstrates the feasibility of future work using high Q superconducting RF cavities to realize a dynamical Casimir effect (DCE) due to the membrane's motion at GHz frequencies.


EDIT And

Parametric Oscillation and Microwave Optomechanics with cm-sized SRF Cylindrical Cavities
https://escholarship.org/uc/item/0qz3w91j#page-77

And no I do not think that is what is happening here...
« Last Edit: 06/25/2015 11:52 AM by OttO »

Online Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5918
  • Likes Given: 5257
@rfmwguy - some images
https://drive.google.com/folderview?id=0B1XizxEfB23tfmcxbUxsM0lVTGVkemVTX1RaMlZJb001NHVaUDRvYUtjS0lIbjdIcUNkX0k&usp=sharing
anyone who has the link can view?

I had trouble finding resonance and basically failed. My excuse is that I ran out of daylight.
Driving at 2.45 GHz I got Q's of 145 at both 2.40189260E+009 and 2.64320588E+009 Hz.
Driving at the 2.40 GHz I got a Q of 100 and no other resonances
Driving at 2.64 GHz I got Q =  2000 at 2.40 GHz so I switched back to that number but the resonance went away.

So these images are from the cavity driven at 2.64 Ghz and so perhaps not meaningful. I did use the full 15 digits computed, not the 3 digits used here. I probably need to play some more and decrease the bandwidth of the search for resonance. Maybe I'll try that ... later.

These images are twice as dense as before. Ten images per cycle instead of five.

Thanks aero, well done. Unfortunately I am stuck at driving at 2.45 ghz and not 2.64...fortunately I have yet to cut the frustum, meaning I can tweak the small and big diameters from 6.25 and 11.01. Is it easy to plug in the slightly larger diameters for 2.45 ghz resonance?. Not wanting to load u down, but 2k Q is better than 100. 6.735 in small diameter and 11.864 in large diameter, length can stay the same. Just wanting to know if resonance occurs...no pics needed. Thanks in advance...last favor to ask as I am meepless ;)
A clarification on my previous suggestion: I had suggested to use MEEP to look at optimal antenna placement, but not to make a decision at what frequency there is resonance.  Not until the MEEP finite difference model has been verified vs. experimentally measured frequencies.

Is the MEEP model's finite difference grid fine enough and the MEEP eigensolution HarmInv well-conditioned enough to successfully predict the frequencies measured by NASA and other experimenters, using their geometrical dimensions?

Is the MEEP model's finite difference grid fine enough and the MEEP eigensolution HarmInv well-conditioned enough to successfully predict the frequencies measured by NASA and other experimenters, using their geometrical dimensions?

@ Rodal
Don't you mean, "The frequencies that COMSOL indicated that they should use? And the answer is that Harminv only comes close. And would you care to give an error bound on the measured (stated) dimensions of the cavities and the sensitivity of resonance to those measurements? I don't expect Harminv to reproduce the COMSOL numbers even if I do input the same numbers and precision used but we don't know what was used, do we. As for the experimental data, we have the same problems in spades. So if you could tell me what the resonance frequency sensitivities to small diameter and length are, that would be very helpful. Then we could estimate probable measurement errors and see if they are realistic. And if you can't tell me what the sensitivities are, then I can tell you, by using numerical data.

I doubt that Paul made a measurement error by as much as a tenth of an inch but unless he used a large micrometer to measure the height, he could have. And even with a micrometer, unless he was very very careful he could have introduced a slight angle to his measurement.

My  point is that you know as well as I that my computer is not up to running high resolution in 3D but Harminv does do much better in 3D than it ever did in 2D or cylindrical coordinates.

And just so you will know, over the last nearly 10 years, meep has been downloaded over 10,000 times. Some, if not most of the downloaders used meep, and many of them conducted and published peer reviewed research papers based on meep results. Meep is still widely used and does not have a reputation for frequency errors. The one thing those users may have had access to that I don't yet have is a powerful computer. Mine is a good home desk-top but at 5 years old, it is not a supercomputer. Go ahead and knock my computer all  you want but please lay off of meep.

And to answer your question as asked, "Yes, Meep absolutely does have the capability to measure resonance frequencies as well or better than other tools. I just do not have the needed tools installed. Harminv, not so much."

To install MPB and recent meep upgrades, I need to compile, link and load from C++ source code. That code is available but I am not a computer systems administrator or a professional C++ programmer and I do not want to stop producing some helpful results to produce nothing for the time it will take me to become knowledgeable enough to do that. Then take the time to learn to use the newly installed and upgraded program features. And then only to have my results flawed my my own modelling errors with many more potential sources of error. My system does what it does and if someone doesn't like it they can choose not to consider it.

And to the other 1,499,999 readers of this thread, I apologize for my rant.

@aero: my purpose was to warn @rfmwguy (who said was going to be cutting metal) on the accuracy of prediction of natural frequency and mode shape for truncated cone cavities.  I will respond here to your message asking for comparisons, present a spreadsheet comparison and comments based on my experience on numerical methods on what is going on.  I will abstain from making further comments about your Meep model in the future, as they seem to be unwelcome (judging from the tone of your response to my message to @rfmwguy).

______________________________________________________________

Here is a comparison  of twenty one (21) natural frequencies calculated by Frank Davis at NASA for the Brady et.al frustum, between COMSOL Finite Element Analysis solution by Frank Davis and my exact solution.  As it is evident from the spreadsheet, the Mean difference in natural frequency is only 0.177% and the Median difference is only 0.0898%.  The standard deviation (of the 21 frequencies % difference) is 1.04%. That's an acceptable difference based on my experience with numerical models.  In my estimation, Frank Davis model was a very good model, with sufficient finite element discretization to have converged close to a solution.

The mean difference in GHz is only 0.0023 GHz and the median difference in GHz is only 0.0016 GHz.

The dimensions of the cavity analyzed by Frank Davis are given in the second page of Davis' report attached below:

Height: 9.00 inch (228.6 mm)
Top diam.: 6.25 inch (0.1588 mm)
Bottom diam.: 11.01 inch (279.7 mm)
Material: 101 Copper Alloy

From what I have observed, the difference between your finite difference model and the NASA FEA model is much larger  (it is an unacceptable difference, based on my long experience with Finite Element and Finite Difference analysis).  From my education (S.B., S.M and Ph.D. degrees at MIT all using numerical analysis including FEA and FD) and experience (using and writing computer programs FEA) the problem is likely to be with your model and nothing to do with Meep (which is indeed an excellent program written by people at MIT).  Initially the problem was due to your use of a flat 2-D model instead of a 3-D model.  I don't know what the problem with your model is now (impossible to know when we lack a lot of the specifics of your particular model, starting with numerical values of the fields), but I have suggested that you should start by giving numerical values of your output to ascertain what's going on.  The first thing that I would do (with a FD or FEA analysis) is to conduct a convergence analysis to ascertain whether your model is converging to a natural frequency solution (and to find out what is its asymptotic behavior).

Another suggestion is for you to compare your Meep model (at similar discretization of the FD mesh) to a problem for which an exact closed-form solution is available: a cylinder.  This will help in assessing the accuracy of your Meep discretization to obtain natural frequencies, for people ready to cut metal to make their own models based on those predictions.


Attachments:

A) Frequency comparison table
B) NASA's Frank Davis FEA frequency and mode shape analysis


___________________________________________________
Note:

1) Brady et.al.'s frustum uses flat ends.  For my exact solution I used spherical ends that have a spherical radius equal to the mean radius, where the mean is  calculated  from 1) the spherical radii that intersect the corners between the flat ends and the cone lateral walls and 2) the spherical radii that are tangential to the center of the flat ends.  This mean results in a good approximation to flat ends, taking into account the boundary layer effect of the spherical wave trying to accommodate the flat ends.

2) Notice that the differences in the spreadsheet have different signs:

a) for the frequency difference: exact - FEA
b) for the % difference: (FEA - exact)/exact
« Last Edit: 06/25/2015 02:46 PM by Rodal »

Offline TheTraveller

Frankly, between the proposal that the EM Drive somehow "knows" its velocity so that it cannot become a free-energy machine and this proposal that the EM Drive has to have an unspecified level of vibration amplitude and frequency to exert a force... well I better stop here. :)

What Free Energy?
The Work done by the EMDrive generated Force moving a Mass, is powered by Energy from the power supply.

Electrical energy to

Microwave energy to

Mechanical energy to

Acceleration to

Kinetic energy

An EMDrive powered ship obeys A = F/M.
Accumulated ships Velocity or Kinetic Energy is not part of A = F/M.

As for getting an EMDrive to generate an external Force, there 1st needs to be an external Force that moves the EMDrive and causes an internal Doppler shift of the resonant standing waves.

With no external Force, there is no internal Doppler shift and no EMDrive generated Force.
« Last Edit: 06/25/2015 11:46 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 Vix

  • Member
  • Posts: 16
  • Liked: 4
  • Likes Given: 0
Hi!

I would like to throw some thoughts about EM drive and to suspicions whether its effects are real or not.
I think they are, despite the fact that some of our common attempts aren't showing it yet. Why? Without going into conspiracy theories, just ask yourself a couple of simple questions:
Q: 1. If you were an inventor, and invented something with a great potential, would you reveal all the details in the public?
A. No. Most probably you would reveal it to the goverment of your country, and/or the military you trust (hint: USA/Nasa)
Q2. If you were a government/military which was presented with an invention of a great potential, would you reveal all of its details to the public?
A: No. Most probably, you would say that you did a thorough research and found nothing, or at least nothing significant, and that it's not worthy pursuing further. In the meantine, you rename the project and work secretly toward its advancement. Moreover, in order to protect the invention and keep it for yourself, you may be inclined to provide fake info to the public, so a DIY model won't work, so afer an initial euphoria has subsided,  it will be effectively forgotten. (Because it would appear that it doesn't work).

So, to summarize: EM drive works, but I am affraid that we won't be able to replicate it, simply because we lack some important (undisclosed) bits of information (the devil is in the details), and we may never find them, except if someone else discovers it by accident and make public (unlikely).
In the best case we may find out that our diy EM drive provides thrust but at funny low levels, rendering it useless, while the real (undisclosed) model really works!
This doesn't mean that we should give up, quite the contrary: Go and build EM drives, but don't stick to the disclosed info. Experiment, change things, be creative. Discover.

At the end, one of my favorite Eintein quotes:
"In theory, theory and practice are the same. In practice, they are not." :)

Regards,

Vix

Online Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5918
  • Likes Given: 5257
I find this fascination with simulation rather curious, given the fact that all simulators seek to follow Maxwell's equations as accurately as their computational methods allow, and that Maxwell's equations predict zero thrust for the EmDrive. So what is it precisely that's so interesting about simulating the fields inside the cavity?
The main interest lies in precisely predicting the natural frequency and mode shape of a resonating cavity.

A secondary interest is in finding the optimal location for the RF Feed, for resonance.

Here is a comparison of natural frequencies and mode shapes for truncated cone cavities:  http://forum.nasaspaceflight.com/index.php?topic=37642.msg1393872#msg1393872

The people in this thread that are interested in making their own cavities are interested in knowing at what frequency and mode shape will their cavities resonate.  The solution of Maxwell's equations for a conical frustum are non-trivial.  If you know of other ways to solve Maxwell's equations to predict the natural frequency and modes shapes of a truncated cone cavity (other than by using numerical methods or by using exact solutions) please let us know.
« Last Edit: 06/25/2015 01:32 PM by Rodal »

Offline Chrochne

  • Full Member
  • *
  • Posts: 182
  • Liked: 104
  • Likes Given: 247
I found this article from May 2015. Roger Shawyer answers there on some interesting questions. I did not see the link to the article so far on the forum, so here it is.

It seems his cooperation with the private companies and development of the second generation EmDrive is alive and well. It fact, he is very sure about the progress.

So, make a coffee, tea or your favourite poison and read it :)

http://www.ibtimes.co.uk/nasa-validates-emdrive-roger-shawyer-says-aerospace-industry-needs-watch-out-1499141

I really can not wait to see the coming paper from him. I also noticed that media are starting to pay more attention to him.

« Last Edit: 06/25/2015 12:55 PM by Chrochne »

Offline Fugudaddy

  • Member
  • Posts: 31
  • Rochester NY
  • Liked: 57
  • Likes Given: 70
So, to summarize: EM drive works, but I am afraid that we won't be able to replicate it, simply because we lack some important (undisclosed) bits of information (the devil is in the details), and we may never find them, except if someone else discovers it by accident and make public (unlikely).

Sorry, but pfft.

Between the math-heads trying to get a wrap around this in one direction, and the builders who are actually making these things, the amount of knowledge regarding how EMDrives function has gone up tremendously in just the last few months.

The biggest problem isn't that somebody else has a 'special sauce' that the people here can't figure out the recipe to. I don't think that *anybody* yet has a solid grasp on why an EMDrive appears to do what it does.

There's nothing accidental going on here; lots of theory and serious science is. And yes, the skeptics are as much a part of that process as the DIYers, the people developing software models, etc. I wonder how many papers have been written now as a result of this exploration/experimentation ;)

In the end it could be nothing more than an interesting artifact that can be used to make thrust that is better than photons, but not as effective as something else. Or it could open up new avenues in understanding the dual nature of wave/particles and their interaction with 'regular' matter.

Either way; patience. Data is coming. :)


Online Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5918
  • Likes Given: 5257
I found this article from May 2015. Roger Shawyer answers there on some interesting questions. I did not see the link to the article so far on the forum, so here it is.

It seems his cooperation with the private companies and development of the second generation EmDrive is alive and well. It fact, he is very sure about the progress.

So, make a coffee, tea or your favourite poison and read it :)

http://www.ibtimes.co.uk/nasa-validates-emdrive-roger-shawyer-says-aerospace-industry-needs-watch-out-1499141

I really can not wait to see the coming paper from him. I also noticed that media are starting to pay more attention to him.
TheTravellerEMD posted the abstract on reddit: (bold added for emphasis)

Quote from: TheTravellerEMD
SECOND GENERATION EMDRIVE PROPULSION APPLIED TO SSTO LAUNCHER AND INTERSTELLAR PROBE
Roger Shawyer C.Eng.MIET.FRAeS
SPR Ltd, United Kingdom
sprltd@emdrive.com
ABSTRACT
In an IAC13 paper the dynamic operation of a second generation superconducting EmDrive thruster was described.
A mathematical model was developed, and in this paper, that model is used to extend the performance envelope of the technology.
Three engine designs are evaluated. One is used as a lift engine for a launch vehicle, another as an orbital engine for the launcher, and a third as the main engine for an interstellar probe.
The engines are based on YBCO superconducting cavities, and performance is predicted on the basis of the test data obtained in earlier experimental programmes.
The Q values range from 8 x 10^7 to 2 x 10^8 and provide high specific thrusts over a range of accelerations from 0.4 m/s^2 to 6 m/s^2.

The launch vehicle is an “all-electric” single stage to orbit (SSTO) spaceplane, using a 900 MHz, eight cavity, fully gimballed lift engine.
A 1.5 GHz fixed orbital engine provides the horizontal velocity component.
Both engines use total loss liquid hydrogen cooling.
Electrical power is provided by fuel cells, fed with gaseous hydrogen from the cooling system and liquid oxygen.
A 2 Tonne payload, externally mounted, can be flown to Low Earth Orbit in a time of 27 minutes.
The total launch mass is 10 Tonnes, with an airframe styled on the X37B, which allows aerobraking and a glide approach and landing.
The full potential of EmDrive propulsion for deep space missions is illustrated by the performance of the interstellar probe.
A multi-cavity, fixed 500 MHz engine is cooled by a closed cycle liquid nitrogen system.
The refrigeration is carried out in a two stage reverse Brayton Cycle. Electrical power is provided by a 200 kWe nuclear generator.
The 9 Tonne spacecraft, which includes a 1 Tonne science payload, will achieve a terminal velocity of 0.67c and cover a distance of 4 light years, over the 10 year propulsion period.
The work reported in this paper has resulted in design studies for two Demonstrator spacecraft.
The launcher will demonstrate the long-sought-for, low cost access to space, and also meet the mission requirements of the proposed DARPA XS-1 Spaceplane.
The probe will enable the dream of an interstellar mission to be achieved within the next 20 years.

It will be interesting (from a public's expectation study point of view) to hear whether this abstract meets what readers here were expecting from what was said about the paper.
« Last Edit: 06/25/2015 02:11 PM by Rodal »

Offline TheTraveller

So, to summarize: EM drive works, but I am afraid that we won't be able to replicate it, simply because we lack some important (undisclosed) bits of information (the devil is in the details), and we may never find them, except if someone else discovers it by accident and make public (unlikely).

Sorry, but pfft.

Between the math-heads trying to get a wrap around this in one direction, and the builders who are actually making these things, the amount of knowledge regarding how EMDrives function has gone up tremendously in just the last few months.

The biggest problem isn't that somebody else has a 'special sauce' that the people here can't figure out the recipe to. I don't think that *anybody* yet has a solid grasp on why an EMDrive appears to do what it does.

There's nothing accidental going on here; lots of theory and serious science is. And yes, the skeptics are as much a part of that process as the DIYers, the people developing software models, etc. I wonder how many papers have been written now as a result of this exploration/experimentation ;)

In the end it could be nothing more than an interesting artifact that can be used to make thrust that is better than photons, but not as effective as something else. Or it could open up new avenues in understanding the dual nature of wave/particles and their interaction with 'regular' matter.

Either way; patience. Data is coming. :)

There was missing info but Roger Shawyer helped me to put together my calculator, which pushes out the same data, for the same frustum dimensions as SPR's in house software does.

The resonate frequency it generates agrees with known SPR EMDrive dimensions such as the Boeing Flight Thruster.

As far as I know, no other model can generate the same resonant frequencies as this software does.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.”
Herman Melville, Moby Dick

Online SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2589
I find this fascination with simulation rather curious, given the fact that all simulators seek to follow Maxwell's equations as accurately as their computational methods allow, and that Maxwell's equations predict zero thrust for the EmDrive. So what is it precisely that's so interesting about simulating the fields inside the cavity?

Patterns! EM harmonic mode patterns and yes Maxwell's equations are written not to be violated by the software code. The simulations will not magically show us a new effect violating any fundamental laws. The actions within the cavity follow those laws, but the thrust effect I believe lays it foundation from those laws. When I was 14 a old ham that I'd babysit his kids helped me get stuck on this tar baby technology.  When I showed an interest in how his ham set worked he started showing me Maxwell's and Coulomb and ohms laws by drawing pictures and then writing the formulas. I could grasp the pictures as to what was happening and then the formulas made sense, latter I could see the images in my head by reading the formulas. Now it's impossible for me to separate the two. That is why I like them.

Einstein was good at what he did not only because of being able to write formulas but even more his visual thought experiments.

Shell

Online SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2589
I find this fascination with simulation rather curious, given the fact that all simulators seek to follow Maxwell's equations as accurately as their computational methods allow, and that Maxwell's equations predict zero thrust for the EmDrive. So what is it precisely that's so interesting about simulating the fields inside the cavity?
The main interest lies in precisely predicting the natural frequency and mode shape of a resonating cavity.

A secondary interest is in finding the optimal location for the RF Feed, for resonance.

Here is a comparison of natural frequencies and mode shapes for truncated cone cavities:  http://forum.nasaspaceflight.com/index.php?topic=37642.msg1393872#msg1393872

The people in this thread that are interested in making their own cavities are interested in knowing at what frequency and mode shape will their cavities resonate.  The solution of Maxwell's equations for a conical frustum are non-trivial.  If you know of other ways to solve Maxwell's equations to predict the natural frequency and modes shapes of a truncated cone cavity (other than by using numerical methods or by using exact solutions) please let us know.
Several hundred quotes ago you, Thetraveler and a host of others joined in to do some not so simple calculations for a truncated cone. they all kind of worked and you all agreed to disagree, and now we have meep telling us something else. I even tried pen and paper and because I didn't feel good about my numbers I decided to make the diameters of the endplates express variable width dimensions using a polygon shape and be able to adjust the endplates to a harmonic that formulas refuse to do.

And don't feel bad about spread sheets or formulas calculating the variable geometries giving you exact harmonic frequencies, look at it as a clue. Something is going on within the cavity to cause a deviation from those formulas and that to me is a large waving flag.

Shell

Online Rodal

  • Senior Member
  • *****
  • Posts: 5838
  • USA
  • Liked: 5918
  • Likes Given: 5257
...Several hundred quotes ago you, Thetraveler and a host of others joined in to do some not so simple calculations for a truncated cone. they all kind of worked and you all agreed to disagree, and now we have meep telling us something else...
And don't feel bad about spread sheets or formulas calculating the variable geometries giving you exact harmonic frequencies, look at it as a clue. ...
Concerning the predicted natural frequency and mode shape, based on my experience with Finite Difference methods at MIT (where Meep was written) the difference may be due to the discretization model and not due to a physical difference or due to a difference with Meep from other codes.  It is known that Finite Elements have much better convergence properties than Finite Difference methods (for the same number of grid points), since FE are based on variational principles and use polynomial interpolation in between.  It simply appears that the FD mesh discretization is far from being sufficient to provide convergence.  @aero writes that it is due to his computer memory and time limitations. 

This issue is well known to people familiar with FD methods, for example in fluid mechanics, where finite difference methods were prevalent (due to the fact that one has to solve nonlinear partial differential equations of Navier Stokes).  Whenever undertaking a numerical solution, the user should:


1) conduct comparisons between the meshed solution and a known closed-form solution (in this case a comparison with a cylinder's natural frequencies)

2) conduct a convergence study of finer and finer meshes to ascertain the rate of convergence, and whether there is any monotonic convergence to a solution.


Part of the problem is due to the enforcement of boundary conditions based on Cartesian coordinates for a problem that has circular axi-symmetry (this is evident from the fractal artifact on the FD solution showing the fact that the FD mesh is not fine enough).  The problem of FD methods with boundary conditions is well-known and it is one of the reasons that FE and other methods were developed.  Yet, the FD's advantage is due to the simplicity in coding Finite Difference codes (much simpler than coding Finite Element methods), and the fact that one does not have to deal with a more fully populated huge matrix that has to be inverted (as in the FE method).  The inversion of the FE matrix is very onerous for nonlinear transient problems (Meep was written for optical applications where nonlinear transient solutions are required).

« Last Edit: 06/25/2015 02:48 PM by Rodal »

Tags: