Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1805269 times)

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2590
I really don't like this, to ask. I've always forged my own path through the years with very little help and built businesses in industries where a billion spent is sometimes a drop in the bucket.

Totaling up what the costs I'm facing it became clear that with the little I get on retirement I can't do it in the time frame needed. I guess I'm asking for a little help and only if you can.

I'm so driven to see real data from these test and to add to the pool of knowledge, I have to swallow my pride and ask.

http://www.gofundme.com/yy7yz3k

Thank you all for your support, your ideas and most of all sharing a dream.

Shell

Do or do not.  There is no try. :)
Putting my money where my mouth is, at least a little.
Try I do, don't I not.

Thank you!
Shell

Offline mwvp

  • Full Member
  • **
  • Posts: 254
  • Coincidence? I think Not!
  • Liked: 169
  • Likes Given: 30
5) What is most interesting to me is the apparent growth with time of the Poynting vector fluctuations during the few time steps examined.  This may show:

a) amplitude modulation

or/and

b) parametric amplification (which is a well-known amplification mechanism and one that I suspected may be involved in amplifying the fields inside the cavity)

Why parametric amplification? I don't see any variable reactances, such as the waveguide or its elements changing dimension, variable dielectrics or semiconductors.

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5845
  • USA
  • Liked: 5927
  • Likes Given: 5270
5) What is most interesting to me is the apparent growth with time of the Poynting vector fluctuations during the few time steps examined.  This may show:

a) amplitude modulation

or/and

b) parametric amplification (which is a well-known amplification mechanism and one that I suspected may be involved in amplifying the fields inside the cavity)

Why parametric amplification? I don't see any variable reactances, such as the waveguide or its elements changing dimension, variable dielectrics or semiconductors.
I have an idea about parametric amplification, but it is not well formed at the moment, it is not related to variable dielectrics, semiconductors or reactances.

I agree with Ricvl that the amplification is the natural result of pumping power into the cavity via the RF feed, and that 6527 finite difference time steps is way too early in the response to reach equilibrium

The total time from start of the RF feed in the Meep response analysis is: 
 320 ( time slices) * 4.082199*10^(-11) seconds/timeSlice =
                                          = 0.013063 microseconds

I think that this comment from deltaMass jumped the gun:

...Fascinatingly totally unstable. This clearly does not represent reality, but instead we're looking at an artifact of numerical simulation.  I would be hard pressed to call this "useful data"

On the contrary, this is what one would expect in reality after only 0.013 microseconds of response.  It is a comment written without probably realizing the magnitudes of the horizontal and the vertical axis in the graph.
It is not a "total instability" or an "artifact of the numerical simulation".

The Meep time step meets the Courant factor condition, so as shown by Lax and Ricthmyer in the 1950, there is no numerical instability.
 
I think that Ricvl wrote this in an excellent way:

The cavity is reservatory of energy, like a LC circuit. If a mode has been excited by the source is natural the growing of amplitude of the fields because of the succesives reflections of the counter propagating traveling waves ( forming the stationary wave) will summ with waves generated by the source. This grow will happens until the loss  in the walls of the cavity becomes equal the power generated by the source.
« Last Edit: 07/09/2015 06:07 PM by Rodal »

Offline mwvp

  • Full Member
  • **
  • Posts: 254
  • Coincidence? I think Not!
  • Liked: 169
  • Likes Given: 30
The Chinese may have figured out how to make a cavity with a Q of 117,500 using flat end plates.

As per the attachment they build in a short section, at each end of the tapered wave guide, of constant diameter circular waveguide that allows the spherical waves in the tapered portion of the cavity to transition from/to a planar wave that will generate no phase distortion as it bounces off the flat end plate.

Very clever.

The drawings also show how they do impedance matching.

See attachment.

I now know my 100,000 Q goal is obtainable. Yea!

It appears to be a double-tuned circuit; two cavities coupled through a small aperture. I read the NWPU paper a month ago or so. If I understand Shawyer's theory right, the only EM momentum that matters is in the frustrum, and the Q that matters is the Q of the frustrum, not an additional filter cavity that is used to couple energy to it.

But maybe it helps slow the group velocity in the frustrum in some way.

Anyways, is it the Q of the frustrum that is so high, or is it the combination of the two filter cavities that is?

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
I really don't like this, to ask. I've always forged my own path through the years with very little help and built businesses in industries where a billion spent is sometimes a drop in the bucket.

Totaling up what the costs I'm facing it became clear that with the little I get on retirement I can't do it in the time frame needed. I guess I'm asking for a little help and only if you can.

I'm so driven to see real data from these test and to add to the pool of knowledge, I have to swallow my pride and ask.

http://www.gofundme.com/yy7yz3k

Thank you all for your support, your ideas and most of all sharing a dream.

Shell

Do or do not.  There is no try. :)
Putting my money where my mouth is, at least a little.

Good for you Shell...I shared your gfm page with my facebook community. They also know I am building a prototype, so this will not be from left field. Good luck with this. I hope all supporters and detractors drop you a couple of bucks for giving it a go. Posting is cheap...building is not.

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
@Rodal
Man O Man, it's really hard to slam something when I don't know what it is!  >:(
Yes, that snapshot is way early. Let's look at some decent time evolution first.
« Last Edit: 07/09/2015 05:56 PM by deltaMass »

Offline mwvp

  • Full Member
  • **
  • Posts: 254
  • Coincidence? I think Not!
  • Liked: 169
  • Likes Given: 30

I have an idea about parametric amplification, but it is not well formed at the moment, it is not related to variable dielectrics, semiconductors or reactances.

I agree with Ricvl that the amplification is the natural result of pumping power into the cavity via the RF feed, and that 6527 finite difference time steps is way too early in the response to reach equilibrium

I did some research in parametrically pumped Faraday waves in fluids recently. Fascinating stuff. I suppose you might consider the mechanical displacement of the frustrum creating doppler shifts which increase or damp travelling waves in the frustrum as parametric pumping, but I would call it a linear Sagnac effect.

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2590
I really don't like this, to ask. I've always forged my own path through the years with very little help and built businesses in industries where a billion spent is sometimes a drop in the bucket.

Totaling up what the costs I'm facing it became clear that with the little I get on retirement I can't do it in the time frame needed. I guess I'm asking for a little help and only if you can.

I'm so driven to see real data from these test and to add to the pool of knowledge, I have to swallow my pride and ask.

http://www.gofundme.com/yy7yz3k

Thank you all for your support, your ideas and most of all sharing a dream.

Shell

Do or do not.  There is no try. :)
Putting my money where my mouth is, at least a little.

Good for you Shell...I shared your gfm page with my facebook community. They also know I am building a prototype, so this will not be from left field. Good luck with this. I hope all supporters and detractors drop you a couple of bucks for giving it a go. Posting is cheap...building is not.
This floors me, I expected nothing just silence and the chirping of crickets. that is not to be the case. Thank you very much!

(Crazy Eddie) :D
Shell 

Offline Flyby

  • Full Member
  • ***
  • Posts: 384
  • Belgium
  • Liked: 447
  • Likes Given: 47

If you draw a horizontal axis through the middle of the waveguide in the picture you show, what does the waveguide opening into the cavity on the plane perpendicular to that axis look like?



In other words, what does the other view look like?

It would be nice if we could guesstimate what the waveguide opening aspect ratio is, compared to the lateral cavity dimensions.

_____________________________________________
Eventually, we could also analyze with Meep the cavity + waveguide

If I get you correctly, you'd like to have a top view, cut through the middle of the waveguide?
It all depends on whether that waveguide is rectangular or cylindrical...

If it is cylindrical, i can generate such drawing, if it is square or rectangular, i'll need the width of it...
In the mean time, even I believe the drawing is most likely not accurate, I assumed it was and made the measurements proportional to the full height, that being 100%....

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5845
  • USA
  • Liked: 5927
  • Likes Given: 5270

I have an idea about parametric amplification, but it is not well formed at the moment, it is not related to variable dielectrics, semiconductors or reactances.

I agree with Ricvl that the amplification is the natural result of pumping power into the cavity via the RF feed, and that 6527 finite difference time steps is way too early in the response to reach equilibrium

I did some research in parametrically pumped Faraday waves in fluids recently. Fascinating stuff. I suppose you might consider the mechanical displacement of the frustrum creating doppler shifts which increase or damp travelling waves in the frustrum as parametric pumping, but I would call it a linear Sagnac effect.

Yes I was thinking along mechanical lines, based on my background in nonlinear aeroelasticity and nonlinear dynamic effects and as a byproduct of reading the effect of mechanical displacements in in asymmetric optical cavities with whispering gallery modes.  A very malformed idea that I just wrote without dedicating much time to justify it, I admit :)

Offline mwvp

  • Full Member
  • **
  • Posts: 254
  • Coincidence? I think Not!
  • Liked: 169
  • Likes Given: 30
I agree with Ricvl that the amplification is the natural result of pumping power into the cavity via the RF feed, and that 6527 finite difference time steps is way too early in the response to reach equilibrium

The total time from start of the RF feed in the Meep response analysis is: 
 320 ( time slices) * 4.082199*10^(-11) seconds/timeSlice =
                                          = 0.013063 microseconds
...
On the contrary, this is what one would expect in reality after only 0.013 microseconds of response.  It is a comment he wrote without probably realizing the magnitudes of the horizontal and the vertical axis in the graph.
...
I think that Ricvl wrote this in an excellent way:

http://forum.nasaspaceflight.com/index.php?topic=37642.msg1402125#msg1402125

Yea, that's why I was asking a few days ago. Off the top of my head, a Q of a thousand @ 2GHz is going to take a microsecond or so to "ring-up" to a steady state or equilibrium. Before which, travelling waves and an overall non-zero Poynting vector I would expect, then a much smaller thereafter going from the cavity RF feed to where the heat is dissipated. Those of us who've used spectrum and network analyzer know we can't sweep our filters to fast or we get a distorted Bode plot/transfer function, because the network hasn't had time to absorb enough energy to measure accurately.

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5845
  • USA
  • Liked: 5927
  • Likes Given: 5270
...
If I get you correctly, you'd like to have a top view, cut through the middle of the waveguide?
It all depends on whether that waveguide is rectangular or cylindrical...

If it is cylindrical, i can generate such drawing, if it is square or rectangular, i'll need the width of it...
In the mean time, even I believe the drawing is most likely not accurate, I assumed it was and made the measurements proportional to the full height, that being 100%....
Thank you so much for dedicating your valuable time to make that drawing.

Much appreciated.

Yes I was looking for a guesstimate of what it would look like from a view at the plane at 90 degrees, assuming that the cavity is conical with a circular cross section perpendicular to the axis of axisymmetry.

Offline Ricvil

  • Member
  • Posts: 62
  • Liked: 65
  • Likes Given: 22
Further info on the antenna arrangement I'll be using on my 2.45GHz version of the Flight Thruster.

As the EM waves in a cavity with spherical end plates are spherical, the antenna also needs to be spherical as per the sidewall insertion point curve of the EM wave at that point.

I could have used a single stub 1/4 wave antenna curved to match the spherical EM wave shape but I had reservations that the end point of a single 1/4 wave stub inside the cavity may shift over time.

To stop any shifting, especially as I plan to ship finished EMDrives all over the planet, the single 1/4 wave stub antenna was replaced by 3 x 1/4 wave stub antennas that are joined at the centre and attached to the centre of 3 Rf connectors fitted into the side walls.

Trust the attached crude drawing makes it clear how I intend to excite TE013 mode and do so in a way to introduce min phase distortion into the internal resonant standing wave by the excitation antenna.
I would hope you take in consideration the phases of your antennas and the distance from each other, you wouldn't want them to cause standing wave attenuations with each other canceling out a good idea.

Shell, there is also the issue that in the drawing, the assumption is made that the wave-patterns in the longitudinal direction are sinusoidal harmonic with equidistant nodal points.  This is true for a cylindrical waveguide but it is not true for a conical waveguide (even with spherical ends) (*).  With a conical waveguide the wave-patterns in the longitudinal direction are not sinusoidal harmonic: the nodes are not equidistant.  How different from equidistant the nodes are depends on the mode(s) being excited and their participation. Hopefully by placing the antenna where it is proposed he can force that mode, but a lot depends on nearby modes and mode participation.  Since he is not using a magnetron, that may help in eliminating other modes from participating.

_______
(*) The wave patterns in the longitudinal direction for a spherical wave in a cavity are not sinusoidal harmonic, they are governed by Spherical Bessel functions, which have unequal distance between nodes.

At this point there is a very important point of attention here.
There are two cavitys been modeled here. One with plane taps, and other if spherical taps.

They are very close in format, but the results are completally diferents.

The cavity with spherical taps has simple analytics descriptions, just because its geometry permits a orthogonal coodinate system be used to match exactly the boundary conditions, and the method of separation of variables to be fully applied, having modes and cut off frequencys well determined, where TE and TM modes are decoupled.

In the case of plane taps, there is no analytical solution avaiable. The geometry in this case cannot be described by an orthogonal coordinate system , and a approximate analytic description would use a method called "impedance matching" to restore a descrition of the problem as a cilyndrical cavity(where the difference of the radius between the large and the small tap is about lambda/4). In that description ( used for analyse corrugated waveguides) there is, in general, a coupling of TE and TM modes resulting an called hybrid mode, very common on dieletric waveguides like optical fibers.
As I said before, a tapered conical cavity with plane taps, can be viewed , by image theory, as a longitudinal half section of a corrugated waveguide with a triangular profile.
I was not surprised when someone ( I don't remember who) has obtained a simultaneous TE and TM modes where his expecting just one of them.
So the real cavity has spherical or plane taps?

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
I really don't like this, to ask. I've always forged my own path through the years with very little help and built businesses in industries where a billion spent is sometimes a drop in the bucket.

Totaling up what the costs I'm facing it became clear that with the little I get on retirement I can't do it in the time frame needed. I guess I'm asking for a little help and only if you can.

I'm so driven to see real data from these test and to add to the pool of knowledge, I have to swallow my pride and ask.

http://www.gofundme.com/yy7yz3k

Thank you all for your support, your ideas and most of all sharing a dream.

Shell

Do or do not.  There is no try. :)
Putting my money where my mouth is, at least a little.

Good for you Shell...I shared your gfm page with my facebook community. They also know I am building a prototype, so this will not be from left field. Good luck with this. I hope all supporters and detractors drop you a couple of bucks for giving it a go. Posting is cheap...building is not.
This floors me, I expected nothing just silence and the chirping of crickets. that is not to be the case. Thank you very much!

(Crazy Eddie) :D
Shell
You are welcome...don't know if it was my FB followers, but you just jumped about 40% in help since I last looked at it about 15 minutes ago...You are doing well...sure have see a lot less worthy efforts get money...Go Shell, go!

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5845
  • USA
  • Liked: 5927
  • Likes Given: 5270
...
So the real cavity has spherical or plane taps?
Thanks.  Excellent analysis.  NASA COMSOL FEA found some weird modes also that were a mixture, using plane ends.  Most of the researchers have used FLAT ends on their experiments with truncated cones.

All NASA experiments have had flat ends.  Some of Shawyer''s have had flat ends, from what I recall, according to TheTraveller, Shawyer's FlightThruster has spherical ends.

Prof. Yang's have flat ends.

Iulian Berca has flat ends.

Baby EM Drive has flat ends.

The exact solution I have been using uses spherical ends,  when approximating a flat end natural frequency I calculate the geometry that has spherical ends halfway between the circumscribed and the inscribed circles (that came close to NASA's COMSOL FEA analysis of all the modes between 0.8 and 2.6 MHz for.

Meep models have been using flat ends on conical cavities.
« Last Edit: 07/09/2015 06:30 PM by Rodal »

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2590
Further info on the antenna arrangement I'll be using on my 2.45GHz version of the Flight Thruster.

I would hope you take in consideration the phases of your antennas and the distance from each other, you wouldn't want them to cause standing wave attenuations with each other canceling out a good idea.


At this point there is a very important point of attention here.
There are two cavitys been modeled here. One with plane taps, and other if spherical taps.

They are very close in format, but the results are completally diferents.

The cavity with spherical taps has simple analytics descriptions, just because its geometry permits a orthogonal coodinate system be used to match exactly the boundary conditions, and the method of separation of variables to be fully applied, having modes and cut off frequencys well determined, where TE and TM modes are decoupled.

In the case of plane taps, there is no analytical solution avaiable. The geometry in this case cannot be described by an orthogonal coordinate system , and a approximate analytic description would use a method called "impedance matching" to restore a descrition of the problem as a cilyndrical cavity(where the difference of the radius between the large and the small tap is about lambda/4). In that description ( used for analyse corrugated waveguides) there is, in general, a coupling of TE and TM modes resulting an called hybrid mode, very common on dieletric waveguides like optical fibers.
As I said before, a tapered conical cavity with plane taps, can be viewed , by image theory, as a longitudinal half section of a corrugated waveguide with a triangular profile.
I was not surprised when someone ( I don't remember who) has obtained a simultaneous TE and TM modes where his expecting just one of them.
So the real cavity has spherical or plane taps?
Nice writeup Ric...

Dr. Rodal and TT and hte key gut Aero. I wonder what a run in meep would make of this solution of three dipoles in a cavity?

Shell

Offline aero

  • Senior Member
  • *****
  • Posts: 2744
  • 92129
  • Liked: 705
  • Likes Given: 240
There was some question about growth in field strength in the resonant cavity. This shouldn't be a surprise.  This is a resonant cavity, an energy storage device, with some value of Q which may be modest or very high. It doesn't matter in this instance because we are working with only 32 cycles from the start at zero energy. 32 cycles is not enough input energy to build up the stored energy in the cavity to its potential. After 32 cycles the stored energy cannot be be close to saturation. We are expecting the stored energy to reach 1000 times that value. The energy (amplitude) within the cavity is increasing due to the constant energy input.

I guess this does underline the need to look at some much longer meep runs.
Retired, working interesting problems

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2590
...
So the real cavity has spherical or plane taps?
Thanks.  Excellent analysis.  NASA COMSOL FEA found some weird modes also that were a mixture, using plane ends.  Most of the researchers have used FLAT ends on their experiments with truncated cones.

All NASA experiments have had flat ends.  Some of Shawyer''s have had flat ends, from what I recall, according to TheTraveller, Shawyer's FlightThruster has spherical ends.

Prof. Yang's have flat ends.

Iulian Berca has flat ends.

Baby EM Drive has flat ends.

The exact solution I have been using uses spherical ends,  when approximating a flat end natural frequency I calculate the geometry that has spherical ends halfway between the circumscribed and the inscribed circles (that came close to NASA's COMSOL FEA analysis of all the modes between 0.8 and 2.6 MHz for.

Meep models have been using flat ends on conical cavities.
I keep thinking about flat versus curved. You can get a higher Q shaping the cavity to the spherical wave shape, but peaking the Q will also open the door to more sensitivity to thermal expansion effects thereby dropping the Q in unwanted waveforms and patterns that almost are uncontrolled and sure to be disruptive. At higher powers it could be a detrimental effect as the copper frustum warpage changes from varying heat signatures in the copper.

A flat plat will give a Q oh let's say 50-90k and might offer a more forgiving environment.

Shell 

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
...
The paradox with propellant driven space craft was explained too, by Dr. White. It occurs anytime "constant acceleration" is assumed, without taking into account the variables in the av(t) expression I posted. In the case of the EM drive where there is no exhaust, dm/dt is replaced by the change in potential energy from the small end, to the big end, per unit time, /c2.

And BTW, if the EM drive were to store every bit of energy put into it, with a high power source it would have to explode eventually. The energy either needs to escape, generate heat or move the thing in order to dissipate. Or else the source input will eventually become saturated until nothing more will go in.
Todd
Each time I try to stay on topic with this breakeven business, which you say I have "wrong", you throw another spanner in the works. Let's try and focus here. I want to understand what's "wrong" as you see it. So far, you're not making sense to me. If you do want to make sense to me, and convince me that I'm "wrong", you're going to have to take my derivation and show, line by line, the "right" substitution.

Over to you.

Here is how it's done, starting from your format to mine. The difference in potential energy is from the initial energy state at the small end, to the final energy state at the big end.
Todd

Thanks Todd. That gets me a lot closer to an understanding of your radically different approach. I have gone through your derivation and algebraically I have no issues with it. But I can't say I'm there yet, being ever mystified by this darned dm/dt term. Some comments:

1. Writing 'u' for the phase velocity, you get
k = u/c2 Newton/Watt = 1/c when u=c,
or in other words a pure photon rocket. But experimental evidence suggests a much higher value for k, and so if your formula is correct, it is predicting a superluminal phase velocity.
Is that your intent? Do you think that this observation is important?

2. When you derive the breakeven velocity for the condition dEin = dEout, this is tantamount to the power breakeven condition. In this case, I get vp = 1/k, and so do you. So at least on this point we are in accord.

3. Since, then, you actually derive a value for k (which I do not), you can specify the value
vp  = c2/u.
Since you're implying u >> c, it's unclear what the actual value of vp  might be, but it certainly is << c if u >> c.
e.g. if we write u = b*c (b >> 1), then
vp  = c/b << c



« Last Edit: 07/09/2015 07:24 PM by deltaMass »

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
Poynting Vectors - I've been somewhat reluctant on excepting this theory, not so much that I have experience in it, but the collateral things I have read offsite have turned me off a bit...so I will share:

http://www.americanantigravity.com and Brandenburg's GEM Theory, which was discussed here in thread 2.

While this unification theory might be of interest, his martian civ theory is, uhhh...well. Its too bad that there is a wild mixing of theories out there, for a real one could be lost in the extraneous noise. For me, I began to discount Poynting Vectors simply because of an author's association with "more colorful" theories.

Moral of the story? Be scientific, not sciencefictionous (if thats a word) ;)


Tags: