-
#3700
by
qraal
on 07 May, 2017 05:43
-
I think most of us agree that Shawyer's explanations are wrong - or, more charitably, incomplete - yet somehow that led him to an interesting suite of phenomena. I do wonder just how he came up with his EM-Drive.
-
#3701
by
RotoSequence
on 07 May, 2017 05:53
-
I think most of us agree that Shawyer's explanations are wrong - or, more charitably, incomplete - yet somehow that led him to an interesting suite of phenomena. I do wonder just how he came up with his EM-Drive.
We still need to eliminate the possibility that it's a combination of systematic error and wishful thinking.
-
#3702
by
qraal
on 07 May, 2017 20:29
-
Of course. The best way to do that is via experiment rather than theory we think we know.
-
#3703
by
mwvp
on 08 May, 2017 00:05
-
...
All this results in F= (h/c) * df/dt.
Interesting observation. Bradshaw in his paper on dispersion notes that dispersion scales energy in time and space. He's concerned with applications like spectroscopy where dispersion can be used to miniaturize the optical path, increasing the spread of spectral components.
In our case, it functions similar to a nozzle; energy flowing out exerts a force on the nozzle.
...
Back to force, F = (h/c) * df/dt and recall this applies to each photon individually, so
F-cavity = N# * F, which equals 1.39025001889807E-14 * df/dt
There are some problems here because, first, I need that number to be on the order of 1.2 E-6, and second, it assumes that all of the Doppler shiftings happens at one end only.
One, meep tells me that the difference in frequency between the two ends of the cavity is only like from 1 to 1000 Hz. I’m assuming that that is the value of df/dt.
And two, how is it possible to justify the frequency shift occurring at only the big end of the thruster? The equation for F-cavity needs a multiplication factor of frequency in there somewhere because multiplying 1.39 E-14 times 2.45 E+9 gives 3.4E-05 which is just right to fudge the quality factor and to let most but not all of the Doppler shift to happen at one end.
Does anyone see where a frequency multiplier can be inserted?
Or does anyone care to do the dimensional analysis to see if a factor of per second is missing?
Some further thoughts. I can’t arbitrarily insert a frequency multiplier into the system … where else is there a term of 10^8 to 10^10 in the cavity … It could be that I’ve misinterpreted the meep result. I noted that the frequency differs end to end in the frustum over a range of 1 Hz to 1000 Hz, but for most cases, the difference was near the 1 Hz end of that range. In the above calculations, I treated that frequency as the doppler over one second, but meep gives instantaneous numbers. Also, a shift in frequency of 1 to 1000 Hz over a period of a second isn’t enough to shift the frequency out of resonance but we know that operating conditions (heat expansion?) will shift the resonant peak significantly over a relatively short time interval. Even 1000 Hz/second shift will not move the resonant peak away from the drive frequency in any reasonable time, notwithstanding that the older photons will decay anyway with the quality factor of the cavities being what they are. So maybe the frequency difference end to end within the frustum is instantaneous from bounce to bounce. That is, they differ in time only by the light transit time from one end to the other of the cavity. If that is the case …
Assume the cavity length is a quarter meter, so the light transit time becomes 8.34E-10 seconds which gives 1.20E+09 end reflections per second for each photon in the cavity. That number is the right order of magnitude.
Using end reflections (end-r) as the doppler multiplier gives
F-cavity = N# * end-r * h/c * df/dt . Use df/dt = 1 Hz / reflection, so
F-cavity = 1.667E-05
That does give the order of magnitude that we think we’re looking for, 16.667 micro-Newtons/Watt, or 16 mN/kW.
It also allows for the doppler to be only slightly different at the small end and large end. Say on the order of 40%-60% of the shift occurring at the big end - small end. That still leaves the number in the range that our best controlled experimental results give.
Where is a Doppler shift coming from? Are you taking the case of an accelerating cavity?
Are you using Meep to give you an "instantaneous" frequency? When you excite a multi-mode cavity, energy may slosh around. You'll have a spread of frequency components as energy sloshes around. And there's losses from the cavity walls. I'd expect lower frequencies to persist at the large base, with higher losses. IIRC, losses shift phase in such a way to increase high frequency components.
And photons aren't bouncing end-to-end like little elastic balls. The integral of the path would be a bank-shot off the walls at an angle.
...
In case anyone is wondering where this fits into the EM Drive discussions, I would consider it a correction to Shawyer's theory. F=ma is not applicable within the frustum as TheTraveler has pointed out, but F = dp/dt is certainly applicable. Doppler shift within the operating resonant cavity produces an unbalanced internal momentum which is balanced by the mass times acceleration of the EM Drive hardware and the device (spacecraft) to which it is attached.
Energy and momentum are conserved.
Yep
-
#3704
by
aero
on 08 May, 2017 00:44
-
@mwvp
You hit on it. Meep gives resonant frequencies, quality factors, and field energies as I have it set up. So no, I can't justify calling the end to end frequency difference a Doppler effect. It is the difference in the resonant frequency from one end to the other of the cavity. More precisely, the difference in the frequency that meep curve fitted at a point 2 mm off-set from the inside surface of the two flat end plates of the frustum. I had nine more detectors coded equally spaced between those two end detectors but those frequencies were usually no different from the frequency at one of the ends. I tried to be very careful to make sure the end detectors were the same distance from the inside surface of the end plates, and not touching them. Touching would have voided any result. Similarly, the detectors were all set at one-half the local cavity radius, with the x and y components equal.
Oh, and I made these runs using Monomorphic's Integer frustum with dimensions 30, 18, and 24 cm.
So I think it would be the instantaneous resonant frequencies. As for photons bouncing off the walls, I don't have a good handle on how one might determine the trace of the momentum vector of the radiation. So I assumed shortest path. Would it be the same as the x, y, z components of the energy? The Poynting vector?
-
#3705
by
spupeng7
on 08 May, 2017 02:29
-
I think most of us agree that Shawyer's explanations are wrong - or, more charitably, incomplete - yet somehow that led him to an interesting suite of phenomena. I do wonder just how he came up with his EM-Drive.
Shawyer was able to pull this rabbit out of his hat due to a combination of open mind, focused determination to achieve these specific ends, and competent co-operative peers.
-
#3706
by
InterestedEngineer
on 08 May, 2017 03:14
-
I'm mapping out how this first round of low powered tests will be performed. I want to explore this TE013 mode by testing specific frequencies along the return loss trace to see if this has any effect on anomalous thrust. I have my doubts that holding at maximum return loss is best, so these tests may allow me to home in on any type of thrust signal.
I'm also getting very good Return Loss numbers if I spend time impedance matching. I'm not sure if -55dB will be reliably repeatable so I may pick a number like -40dB and try and perform all tests at least at that level.
Why can't you sweep the frequencies instead of picking specific frequencies?
-
#3707
by
qraal
on 08 May, 2017 11:42
-
That reminds me somewhat of Blieder Drive's genesis in Blieder's obsession with levitating a penny.
But I'll let the audience discover what that means for themselves...
The Great ExplosionI think most of us agree that Shawyer's explanations are wrong - or, more charitably, incomplete - yet somehow that led him to an interesting suite of phenomena. I do wonder just how he came up with his EM-Drive.
Shawyer was able to pull this rabbit out of his hat due to a combination of open mind, focused determination to achieve these specific ends, and competent co-operative peers.
-
#3708
by
Monomorphic
on 08 May, 2017 12:45
-
Why can't you sweep the frequencies instead of picking specific frequencies?
I can sweep the frequencies and may do that eventually. However, if different points along the return loss trace have different anomalous thrust signals, then those signals may be difficult to distinguish from one another during a constant sweep. I think a more methodical approach is required in this case.
-
#3709
by
mwvp
on 09 May, 2017 04:38
-
@mwvp
...I can't justify calling the end to end frequency difference a Doppler effect. It is the difference in the resonant frequency from one end to the other of the cavity. More precisely, the difference in the frequency that meep curve fitted at a point 2 mm off-set from the inside surface of the two flat end plates of the frustum.
Doppler shifts change frequency. A dispersive waveguide doesn't change the frequency of the propagating energy, but the wavelength in the waveguide (guide-length) does change with the velocity/complex impedance/propagation factor/inductance/capacitance/refractive index. It takes a nonlinear material, or a material changing its properties with time to affect the frequency.
As for photons bouncing off the walls, I don't have a good handle on how one might determine the trace of the momentum vector of the radiation. So I assumed shortest path. Would it be the same as the x, y, z components of the energy? The Poynting vector?
For radiating (far-field) traveling-wave energy, there is the Poynting vector. E X B, with zero phase shift. I don't think its the shortest physical distance, but the path of least-action for the EM wave energy. Perhaps Rodal will set me straight? He's posted a lot of good stuff on radiation pressure, the Poynting vector and its space/time divergence.
E & H in a cavity, in a standing wave, is at a 90 degree phase angle. One could consider the instantaneous, or the short time average at group velocity and long-time average of oscillating/sloshing group velocity. Then considering the near field, phase-shift from losses, and cavity acceleration Doppler shifts would make the problem interesting.
When I see some of the simulations with rotating and spinning modes, I would assume there is angular momentum and even orbital angular momentum in higher-order cavities.
I look forward to improving GPUs and simulators using them. And superconductors.
-
#3710
by
aero
on 09 May, 2017 19:55
-
@mwvp
Let me look in more depth here. Do you agree that p = h*f/c is valid within the frustum? If not, why not, and if so, consider. If the frequency is constant, but wavelength changes in the frustum, then, look at the definition of wavelength, λ = V/f. The conclusion must be that velocity V, or c in the equation of momentum above, must change. Meep doesn't go there.
By design, meep uses natural units where c, the speed of light is defined equal to 1. But meep solves Maxwell's equations, and dozens of runs of multiple frustum simulations consistently give different frequencies at the two ends of the cavity. My response to your information that wavelength changes is that since meep does not allow a change in the value of the speed of light, then the numerics must force that change, hence wavelength change into the one free variable of the calculation. That is frequency.
So the change in frequency is not valid but p = h*f/V is still valid where c is replaced by propagation velocity. Now look at the resulting force difference, F = dp/dt at each end of the frustum. Force now becomes F = h*f *d(1/V)/dt. My point is that it really doesn't matter whether the parameter that changes is frequency, wavelength or velocity of the radiation, something changes and that change results in a force difference between the ends of the frustum. I further assert that as calculated by meep via the solution of Maxwell's equations, that change is functionally equivalent to the difference in calculated frequency. That is because, in my use of the information, only the discrete difference at two points are considered, so the path taken is not yet a great concern.
The concern is the recognition that the force caused by radiation pressure can be and is different at the two internal ends of the frustum. And perhaps that (h/c) * Δf = h*f *d(1/V)/dt in magnitude.
Yes, there are many other factors that affect the radiation within the frustum. Once a phenomenon is recognized as the root source of the force, then those phenomena can be usefully analyzed.
-
#3711
by
tchernik
on 09 May, 2017 20:21
-
That reminds me somewhat of Blieder Drive's genesis in Blieder's obsession with levitating a penny.
But I'll let the audience discover what that means for themselves...
The Great Explosion
Amazing little gem of a story. Life imitating art?
There's something deeply amusing (and/or unsettling) in the potentiality that someone that would be deemed unworthy ends up making such a discovery. But the universe has no sense of humor. Or maybe it has the best one there is?
Thanks for the link.
-
#3712
by
SeeShells
on 09 May, 2017 22:50
-
@mwvp
Let me look in more depth here. Do you agree that p = h*f/c is valid within the frustum? If not, why not, and if so, consider. If the frequency is constant, but wavelength changes in the frustum, then, look at the definition of wavelength, λ = V/f. The conclusion must be that velocity V, or c in the equation of momentum above, must change. Meep doesn't go there.
By design, meep uses natural units where c, the speed of light is defined equal to 1. But meep solves Maxwell's equations, and dozens of runs of multiple frustum simulations consistently give different frequencies at the two ends of the cavity. My response to your information that wavelength changes is that since meep does not allow a change in the value of the speed of light, then the numerics must force that change, hence wavelength change into the one free variable of the calculation. That is frequency.
So the change in frequency is not valid but p = h*f/V is still valid where c is replaced by propagation velocity. Now look at the resulting force difference, F = dp/dt at each end of the frustum. Force now becomes F = h*f *d(1/V)/dt. My point is that it really doesn't matter whether the parameter that changes is frequency, wavelength or velocity of the radiation, something changes and that change results in a force difference between the ends of the frustum. I further assert that as calculated by meep via the solution of Maxwell's equations, that change is functionally equivalent to the difference in calculated frequency. That is because, in my use of the information, only the discrete difference at two points are considered, so the path taken is not yet a great concern.
The concern is the recognition that the force caused by radiation pressure can be and is different at the two internal ends of the frustum. And perhaps that (h/c) * Δf = h*f *d(1/V)/dt in magnitude.
Yes, there are many other factors that affect the radiation within the frustum. Once a phenomenon is recognized as the root source of the force, then those phenomena can be usefully analyzed.
aero,
Just a simple question. Are you measuring the pressures on the surface areas of both ends accordingly, or is your detector a small square the same size for both ends, or do you accommodate the difference in surface areas between the two?
My Best,
Shell
-
#3713
by
aero
on 10 May, 2017 01:57
-
@Shell,
I'm not using the square force detector that you are referring to. Instead, I am using Harminv which detects frequency, quality factor and field component energies. To determine force I am using Force = h*f/c, and should probably use F = 2* h*f/c because mostly the energy reflects from the end plates. I'm only using two detector points, at each point detecting the information for all six field components so 12 lines of detector output in total. The detector points are located one meep voxel width from the inside surface of the small and big end plates, respectively.
Using Monomorphic's Integer frustum which is 24 cm high makes it possible to locate the detector points precisely 2 mm from the inside surface of the end plates. That is because by coding the model geometry in SI Units, not scaled, and using meep resolution equal 500 I have that the voxel width is 1/500, or 2 mm and the height of 24 cm or 240 mm divides evenly by 2 mm. The detectors then are placed +/- 11.8 cm from the frustum center and offset in the x, y-direction.
Steve
-
#3714
by
Bob Woods
on 10 May, 2017 02:50
-
@Shell,
I'm not using the square force detector that you are referring to. Instead, I am using Harminv which detects frequency, quality factor and field component energies. To determine force I am using Force = h*f/c, and should probably use F = 2* h*f/c because mostly the energy reflects from the end plates. I'm only using two detector points, at each point detecting the information for all six field components so 12 lines of detector output in total. The detector points are located one meep voxel width from the inside surface of the small and big end plates, respectively.
Using Monomorphic's Integer frustum which is 24 cm high makes it possible to locate the detector points precisely 2 mm from the inside surface of the end plates. That is because by coding the model geometry in SI Units, not scaled, and using meep resolution equal 500 I have that the voxel width is 1/500, or 2 mm and the height of 24 cm or 240 mm divides evenly by 2 mm. The detectors then are placed +/- 11.8 cm from the frustum center and offset in the x, y-direction.
Steve
Thanks, Aero. You've done an amazing amount to help others.
-
#3715
by
mwvp
on 10 May, 2017 05:28
-
@mwvp
Let me look in more depth here. Do you agree that p = h*f/c is valid within the frustum? If not, why not...
Well, uh, yes-but... for what? I mean, in what context? I hate to tear open a sharp, steel frustrum of fetid (
https://en.wikipedia.org/wiki/Abraham%E2%80%93Minkowski_controversy ) purifying worms with my tender, raw flesh.
It seems from
https://en.wikipedia.org/wiki/Poynting_vector only the real component counts when it comes to radiation pressure.
...
If the frequency is constant, but wavelength changes in the frustum, then, look at the definition of wavelength, λ = V/f. The conclusion must be that velocity V, or c in the equation of momentum above, must change. Meep doesn't go there.
...
dozens of runs of multiple frustum simulations consistently give different frequencies at the two ends of the cavity. My response to your information that wavelength changes is that since meep does not allow a change in the value of the speed of light, then the numerics must force that change, hence wavelength change into the one free variable of the calculation. That is frequency.
I see now I've messed up and contradicted myself. Your fiendish, diabolical and insidious rhetorical trap has caught me.
In post
http://forum.nasaspaceflight.com/index.php?topic=41732.msg1675775#msg1675775 is stated:
Are you using Meep to give you an "instantaneous" frequency? When you excite a multi-mode cavity, energy may slosh around. You'll have a spread of frequency components as energy sloshes around. And there's losses from the cavity walls. I'd expect lower frequencies to persist at the large base, with higher losses. IIRC, losses shift phase in such a way to increase high frequency components.
Following up in post:http://forum.nasaspaceflight.com/index.php?topic=41732.msg1676106#msg1676106
I stated:
Doppler shifts change frequency. A dispersive waveguide doesn't change the frequency of the propagating energy, but the wavelength in the waveguide (guide-length) does change with the velocity/complex impedance/propagation factor/inductance/capacitance/refractive index. It takes a nonlinear material, or a material changing its properties with time to affect the frequency.
That must sound confused and contradictory, but, it was
. If energy is sloshing around a multimode line, by definition, there is a frequency spread apriori.
A dispersive, lossy waveguide does change the frequency, as well as guide wavelength. A perfect monochromatic tone, a single-line in a Fourier spectrograph, is stable in amplitude from infinity past to infinity future. If it propagates as traveling or standing wave in a lossy line, as its amplitude decreases, its bandwidth spreads. The envelope of an AM signal gives you its modulation frequency and sidebands. No wonder you noticed what Meep accurately provided you.
My thinking was (poorly phrased) that it wasn't a Doppler shift. Moreover, since the mechanism of Doppler shift efficiently can convert or transducing electromagnetic, radiant energy into vehicle kinetic energy, and radiation pressure is proportional to Q, loss is the wrong way to go to produce a cavity optomechanical effect. Unless you're using an extraordinary superconductor that makes the energy half-life (Q) >> the period of the Doppler shift frequency.
Having said that, I suspect it is resistive loss that is responsible for EW & Monomorphic's thrust (providing it isn't a false). There test rigs don't appreciably accelerate for a Doppler shift to cause any imbalance in radiation pressure.
-
#3716
by
SeeShells
on 10 May, 2017 12:57
-
@mwvp
Let me look in more depth here. Do you agree that p = h*f/c is valid within the frustum? If not, why not...
Well, uh, yes-but... for what? I mean, in what context? I hate to tear open a sharp, steel frustrum of fetid ( https://en.wikipedia.org/wiki/Abraham%E2%80%93Minkowski_controversy ) purifying worms with my tender, raw flesh.
It seems from https://en.wikipedia.org/wiki/Poynting_vector only the real component counts when it comes to radiation pressure.
...
If the frequency is constant, but wavelength changes in the frustum, then, look at the definition of wavelength, λ = V/f. The conclusion must be that velocity V, or c in the equation of momentum above, must change. Meep doesn't go there.
...
dozens of runs of multiple frustum simulations consistently give different frequencies at the two ends of the cavity. My response to your information that wavelength changes is that since meep does not allow a change in the value of the speed of light, then the numerics must force that change, hence wavelength change into the one free variable of the calculation. That is frequency.
I see now I've messed up and contradicted myself. Your fiendish, diabolical and insidious rhetorical trap has caught me.
In post http://forum.nasaspaceflight.com/index.php?topic=41732.msg1675775#msg1675775 is stated:
Are you using Meep to give you an "instantaneous" frequency? When you excite a multi-mode cavity, energy may slosh around. You'll have a spread of frequency components as energy sloshes around. And there's losses from the cavity walls. I'd expect lower frequencies to persist at the large base, with higher losses. IIRC, losses shift phase in such a way to increase high frequency components.
Following up in post:http://forum.nasaspaceflight.com/index.php?topic=41732.msg1676106#msg1676106
I stated:
Doppler shifts change frequency. A dispersive waveguide doesn't change the frequency of the propagating energy, but the wavelength in the waveguide (guide-length) does change with the velocity/complex impedance/propagation factor/inductance/capacitance/refractive index. It takes a nonlinear material, or a material changing its properties with time to affect the frequency.
That must sound confused and contradictory, but, it was . If energy is sloshing around a multimode line, by definition, there is a frequency spread apriori.
A dispersive, lossy waveguide does change the frequency, as well as guide wavelength. A perfect monochromatic tone, a single-line in a Fourier spectrograph, is stable in amplitude from infinity past to infinity future. If it propagates as traveling or standing wave in a lossy line, as its amplitude decreases, its bandwidth spreads. The envelope of an AM signal gives you its modulation frequency and sidebands. No wonder you noticed what Meep accurately provided you.
My thinking was (poorly phrased) that it wasn't a Doppler shift. Moreover, since the mechanism of Doppler shift efficiently can convert or transducing electromagnetic, radiant energy into vehicle kinetic energy, and radiation pressure is proportional to Q, loss is the wrong way to go to produce a cavity optomechanical effect. Unless you're using an extraordinary superconductor that makes the energy half-life (Q) >> the period of the Doppler shift frequency.
Having said that, I suspect it is resistive loss that is responsible for EW & Monomorphic's thrust (providing it isn't a false). There test rigs don't appreciably accelerate for a Doppler shift to cause any imbalance in radiation pressure.
Taking these thoughts a step further with a the Way-Back Machine. Dr. Rodal a few thousand posts ago calculated the Electromagnetic stress tensors withing the frustum cavity which yielded what appeared to me to be unequal stresses in the cavity.
The first gif is of the poynting vector.
The next ones come from calculations of the stress vectors.
It's important to note that at this point we were looking at and just starting to understand the importance of the antenna and how it couples with the fields in a frustum, which proved to be a very critical component of the builds.
While I believe your statement on the poynting vectors are important, the energy stress tensors are even more so, more so because it represents and calculates the interaction between the internal electromagnetic forces of the frustum and consequently the mechanical momentum within the frustum.
https://en.wikipedia.org/wiki/Electromagnetic_stress%E2%80%93energy_tensorhttps://en.wikipedia.org/wiki/Maxwell_stress_tensorMy very Best,
Shell
-
#3717
by
MazonDel
on 10 May, 2017 23:39
-
It's important to note that at this point we were looking at and just starting to understand the importance of the antenna and how it couples with the fields in a frustum, which proved to be a very critical component of the builds.
If you don't mind SeeShells, would you (or anyone else), be willing to give a short explanation as to what has gone on regarding revelations with the antenna? I've been away from the forums for quite some time and am quite interested in this.
Thanks!
-
#3718
by
SeeShells
on 11 May, 2017 01:34
-
It's important to note that at this point we were looking at and just starting to understand the importance of the antenna and how it couples with the fields in a frustum, which proved to be a very critical component of the builds.
If you don't mind SeeShells, would you (or anyone else), be willing to give a short explanation as to what has gone on regarding revelations with the antenna? I've been away from the forums for quite some time and am quite interested in this.
Thanks!
You could start by reviewing the work done by monomorphic on this site using FEKO where he has posted dozens of videos on youtube with not only different geometries of frustums but different antennas. Aero also did dozens of simulations using MEEP that are posted here over the last few years although they are harder to find as the search function on the NSF site isn't the best. Oh, we also did designs of frustums driven by waveguides.
https://www.youtube.com/user/monomorph1/videos?shelf_id=1&view=0&sort=ddWhat we did find out is exciting a TE012 or TE013 required a loop to couple into the fields. First attachment.
This also can be a modified loop like the one Eugene Samsonov did in his tests in the second attachment.
http://vixra.org/pdf/1603.0153v1.pdf While he didn't see any thrusts I liked his antenna design for coupling in a TE012 in the endplate.
Eagleworks did thier report with a TM212 and a TE012 mode
http://libertariannews.org//wp-content//uploads//2014//07//AnomalousThrustProductionFromanRFTestDevice-BradyEtAl.pdf.
My Best,
Shell
-
#3719
by
aero
on 11 May, 2017 05:12
-
@Shell
Here is one of the very first simulations I posted. Remember when all we had to go on were the dimensions we could extract from the photo images in the "Anomalous" paper? This is that cavity on start-up excited by a drummed up source. It was showing the start-up of the frustum from power on, probably the first 32 cycles of a run. November 2014.
Steve
. If energy is sloshing around a multimode line, by definition, there is a frequency spread apriori.