Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1880887 times)

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
I've been going over the figures here again
http://emdrive.wiki/Experimental_Results
Yang is ahead of the pack by a very decent margin on both thrust and specific thrust (N/W).

Consider:
- Yang is at a Chinese establishment with ties to the military.
- Yang does not respond to questions about her experiment.
- It does not seem unreasonable to imagine that disinformation is at work here.
- A perfect conspiratorial excuse for any such data is "military secrecy".

So I'll just state flat out that I don't give Yang's results any credence.
I also know that nobody here can prove me wrong unless they reproduce her results.

So that means that you give Shawyer the most credence is that right? as you say that he has the coolest experimental setup? ;)
« Last Edit: 08/06/2015 07:44 PM by Rodal »

Offline jmossman

  • Member
  • Posts: 73
  • San Jose, CA
  • Liked: 58
  • Likes Given: 180
@Rodal: Well, because the specific heat of copper isn't zero, it takes a finite time to incur a given temperature rise for a given power input. And when there's an impedance mismatch, the dissipated power will fall and so the corresponding temperature changes will be less than in the matched condition.

On the other hand, measurable mismatch changes are sensed at the ports with response times on order  roughly the cavity dimensions divided by the speed of light. This is far faster than measurable temperature changes.

However, if the EmDrive is actually generating thrust, the statement about the equivalence of high- and low-Q dissipation may not be true. That's really the reason I'm pointing this out.

There seems to be some assumption of reasonably monotonic behavior such that thermal latency will not be a problem...

While the assumption of reasonable behavior applies to other known systems, the EM Drive output (non-repeatable, different by orders of magnitude between researchers) seems to not abide by that assumption

For example, the issue of vibrations of unknown magnitude and frequency being required to "engage" the drive, etc...

As a real-world example, fine-grained CPU power management (i.e. fast decisions) has moved away from using thermal diodes since their response time is too slow.  (fast decisions need microsecond responses vs 10's or 100's of milliseconds from thermal diodes)  Since the EM drive cavity charge/discharge times are on the order of microseconds, I'm rather doubtful that the thermal data would be useful in a true control loop feedback for any EM drive experiment.

However, perhaps the temperature might be useful to post-process and evaluate how much power was actually received/dissipated by the copper in the frustum during an experimental run.  (i.e.  temperature = integration of ohmic losses over time)   Unfortunately, the possible effects of heating moisture within the frustum along with any number of other variables will introduce noise into the temperature reading (cooling from turbulent air around exterior of frustum, humidity at moment of test, etc). 

Still worth a look to see how well frustum temperature correlates with impedance matching...  but as an input into a real-time EM drive feedback loop, I'd put my money on VSWR measurement as being far more useful.

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I've been going over the figures here again
http://emdrive.wiki/Experimental_Results
Yang is ahead of the pack by a very decent margin on both thrust and specific thrust (N/W).

Consider:
- Yang is at a Chinese establishment with ties to the military.
- Yang does not respond to questions about her experiment.
- It does not seem unreasonable to imagine that disinformation is at work here.
- A perfect conspiratorial excuse for any such data is "military secrecy".

So I'll just state flat out that I don't give Yang's results any credence.
I also know that nobody here can prove me wrong unless they reproduce her results.

So that means that you give Shawyer the most credence is that right? as you say that he has the coolest experimental setup? ;)
Your turn  :D

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1314
  • Do it!
  • Vista, CA
  • Liked: 1351
  • Likes Given: 1820
...
I played with the numbers yesterday for F=2P/Vg -> F=2*P*Sqrt(1-(c/(2*(a+x)*f))^2) and found ...
...
Disclaiming again, I probably broke the math.

The equation would be F = (Vg/c^2)*P = P/Vp. IF you use standard waveguide physics that is. In the case of the EM Drive, this may be, F = (Vg*K/c^2)*P, where K is a function TBD by experiment.
Todd

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2335
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2982
  • Likes Given: 2601
About cavity energy.

When an optimal impedance match is achieved between the RF source and its load (the cavity), all the forward power gets dissipated in the cavity walls as real ohmic heat. Thus maximum heating corresponds to maximum energy within the cavity, the condition that is being sought. Notionally then it would be possible to keep the impedance match optimal over temperature changes by simply monitoring the cavity temperature (I'm assuming that the mode does not switch). The latency of thermal feedback is slower than a 1-port or 2-port VSWR measurement, it's true, but it should nevertheless work well. And it can be made contactless.

Because there's no free lunch, a perfectly matched high-Q cavity and a perfectly matched low-Q cavity will dissipate exactly the same amount of power in ohmic heating for the same input power. At least, that's the theory.
I would agree with you if we had a perfect resonator.

What we have is a condition internally in the cavity whenever it's met, with the Q rising to a point to overcome a tipping point, is a decay and the modes shifting into a collapse then to be regenerated again. I've see this in every simulation that's been done, some are faster than others (like the one we just did) and some are slower.

I know there has been talk of the Q being the largest denominator but it's not the whole equation. It would seem that a good buildup of a mode and the collapse into evanescent decay is one of the keys. That supports some theories. Shawyer's layout.

If you believe that a high Q is the denominator than the cavity I just did with Q's in the 10's of millions would be a key and support other theories. Todd and Rodal's (little of me kicking too) decaying into evanescence.

We are at the testing phase of those two thoughts of operation and I think require three different cavities to test. That's where I'm at.

Also you have several other theories that can be tested notsosureofit's and Unrah radiation.

Just remember in meep with perfect walls that don't shift and chage Q we have seen mode shifting and decay.

Tired and time for a nappie.

bbl

shell

Take a look at RFMWGUY's current test.

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
Indeed, one can actually build a self-accelerating system. But it has limited utility because of its limited lifetime. It works like this.

Take a railroad car and inside place another smaller car or "puck". The puck starts out full of sand, but continually leaks sand straight out the bottom via a slit, contactlessly. The puck bounces between the two walls of the cart, elastically. You give the system an initial push and you'll see it self-accelerate until the sand is all gone.

That's because it always hits the forward wall with more momentum than it will hit the back wall

Here you have three brothers that were granted an actual US patent for <<process produces vertical motion and relies only on the aether.>>  Ha ha.  Proof that the US Patent Office continues to grant patents for concepts that violate Physics:


http://www.google.com/patents/US7900874#v=onepage&q&f=false

Device to move an object back and forth
US 7900874 B2
Publication number   US7900874 B2
Publication type   Grant
Application number   US 12/009,852
Publication date   Mar 8, 2011
Filing date   Jan 22, 2008
Priority date   Jan 22, 2008
ABSTRACT
Disclosed herein are two separate processes that do not require a propellant and do not produce an equal and opposite reaction against any external form of matter in the Local Inertial Reference Frame and do not violate Newton's Laws in the Universal Reference Frame. The first process produces horizontal motion, relies on the earth's gravitational field as an external force, and has been successfully tested. The second process produces vertical motion and relies only on the aether. It has been successfully tested considering the effect of the earth's gravity. Due to the law of conservation of angular momentum, the first process is normally considered not possible, but with the proper use of an external field (for example, gravity) and the phenomenon of precession, it becomes possible. A clear distinction is made between a simple rotor and a gyroscope which is a far more complex device.

Of course you also have the patents by Dean:

System for converting rotary motion into unidirectional motion
US 2886976 A
http://www.google.com/patents/US2886976?dq=2886976

Variable oscillator system
US 3182517 A
http://www.google.com/patents/US3182517?dq=N.L.+Dean
« Last Edit: 08/06/2015 09:57 PM by Rodal »

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I'm considering coming out of retirement and applying for a job at the new patent office they've just opened up here in San Jose. The only other one is the main one in Washington DC. Quite frankly, it seems they could do with some guidance.
« Last Edit: 08/06/2015 10:03 PM by deltaMass »

Offline ElizabethGreene

  • Member
  • Posts: 69
  • Nashville, Tennessee
  • Liked: 138
  • Likes Given: 3
...
I played with the numbers yesterday for F=2P/Vg -> F=2*P*Sqrt(1-(c/(2*(a+x)*f))^2) and found ...
...
Disclaiming again, I probably broke the math.

The equation would be F = (Vg/c^2)*P = P/Vp. IF you use standard waveguide physics that is. In the case of the EM Drive, this may be, F = (Vg*K/c^2)*P, where K is a function TBD by experiment.
Todd

If F=P/Vp is correct for traditional physics, then Mr. Shawyer's theory falls apart.  Specifically, as the size of the waveguide decreases Vp increases.  An increase in Vp decreases F.  That would make the force on the little end --smaller-- than the force on the big end.

Is that right?

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
...
I played with the numbers yesterday for F=2P/Vg -> F=2*P*Sqrt(1-(c/(2*(a+x)*f))^2) and found ...
...
Disclaiming again, I probably broke the math.

The equation would be F = (Vg/c^2)*P = P/Vp. IF you use standard waveguide physics that is. In the case of the EM Drive, this may be, F = (Vg*K/c^2)*P, where K is a function TBD by experiment.
Todd

If F=P/Vp is correct for traditional physics, then Mr. Shawyer's theory falls apart.  Specifically, as the size of the waveguide decreases Vp increases.  An increase in Vp decreases F.  That would make the force on the little end --smaller-- than the force on the big end.

Is that right?

Your thinking is correct.  Recall that according to Shawyer, the force is larger on the Big End (he calls the net force on the Big End, the thrust although according to him there is nothing coming out, hence nothing thrusting, -which is a contradiction-), so he would probably answer right on, that what you just explained supports his theory.

[http://3.bp.blogspot.com/-L98u8XFEOiU/VUJzuM81C3I/AAAAAAAA9cw/iIMo2-6wgMo/s1600/emdrive2014.png
As to how you can reconcile that the force being larger on the Big End results in motion towards the Small End (or to a larger force being measured at the Small End as measured by NASA), then you get into more of this Shawyer's "theory".  Here:  http://www.emdrive.com/EmDriveForceMeasurement.pdf  he goes into a mumbo jumbo of action-reaction and Newton's third law that nobody in academia can follow.

So, you are better off taking Shawyer as a cook giving you a recipe.  What matters is the recipe, the chef doesn't necessarily know the Physics or the Chemistry that make the recipe taste well (although if you press him for an explanation he may give you one that is probably not scientifically correct), but, he may know (from experience) what recipes are the ones that make the best dessert.  :)
« Last Edit: 08/07/2015 03:31 PM by Chris Bergin »

Offline CraigPichach

  • Member
  • Posts: 59
  • Calgary Alberta
  • Liked: 42
  • Likes Given: 4
You think if you achieve cavity resonance (charge/discharge) in microseconds (say less than 20microseconds) that you would see the EM-Drive/Q-thruster thrust phenomena in microseconds?


@Rodal: Well, because the specific heat of copper isn't zero, it takes a finite time to incur a given temperature rise for a given power input. And when there's an impedance mismatch, the dissipated power will fall and so the corresponding temperature changes will be less than in the matched condition.

On the other hand, measurable mismatch changes are sensed at the ports with response times on order  roughly the cavity dimensions divided by the speed of light. This is far faster than measurable temperature changes.

However, if the EmDrive is actually generating thrust, the statement about the equivalence of high- and low-Q dissipation may not be true. That's really the reason I'm pointing this out.

There seems to be some assumption of reasonably monotonic behavior such that thermal latency will not be a problem...

While the assumption of reasonable behavior applies to other known systems, the EM Drive output (non-repeatable, different by orders of magnitude between researchers) seems to not abide by that assumption

For example, the issue of vibrations of unknown magnitude and frequency being required to "engage" the drive, etc...

As a real-world example, fine-grained CPU power management (i.e. fast decisions) has moved away from using thermal diodes since their response time is too slow.  (fast decisions need microsecond responses vs 10's or 100's of milliseconds from thermal diodes)  Since the EM drive cavity charge/discharge times are on the order of microseconds, I'm rather doubtful that the thermal data would be useful in a true control loop feedback for any EM drive experiment.

However, perhaps the temperature might be useful to post-process and evaluate how much power was actually received/dissipated by the copper in the frustum during an experimental run.  (i.e.  temperature = integration of ohmic losses over time)   Unfortunately, the possible effects of heating moisture within the frustum along with any number of other variables will introduce noise into the temperature reading (cooling from turbulent air around exterior of frustum, humidity at moment of test, etc). 

Still worth a look to see how well frustum temperature correlates with impedance matching...  but as an input into a real-time EM drive feedback loop, I'd put my money on VSWR measurement as being far more useful.
« Last Edit: 08/06/2015 10:40 PM by CraigPichach »

Offline flux_capacitor

  • Full Member
  • ****
  • Posts: 563
  • France
  • Liked: 679
  • Likes Given: 929
As to how you can reconcile that the force being larger on the Big End results in motion towards the Small End

That's why I prefer McCulloch's theory: according to MiHsC, as they travel towards the big end, photons gain momentum. When they bounce back towards the small end, they loose momentum. So to obey CoM, the cavity has to move from the big end towards the small end.

Offline flux_capacitor

  • Full Member
  • ****
  • Posts: 563
  • France
  • Liked: 679
  • Likes Given: 929
I tried to model TheTraveller's EmDrive Mark 2 in 3D, which was a bit difficult since the only dimensions TT provided were end diameters Ds and Db and the length (Rb - Rs)*

Ds = 159 mm
Db = 400 mm
Rb-Rs = 240.7 mm


Moreover the dimensions in his original drawing are not proportionally on scale, maybe due to a phenomenon known as the Yang Effect ;)
But with trial & error I finally found his "magical proportions": Mr. T inscribed his spherical cone (cone + spherical base) within a cube! Hence:

Rs = Ds = 159 mm
Rb = Db = 400 mm
half-cone angle = 30░


@TheTraveller: Very clever! The more I look at your design, the more I like it.

@Rodal: can you please measure mode and resonant frequency with COMSOL for such a cavity? This is important because this cavity will be monolithically fixed (no tunable small end).

@SeeShells: It should please you in your quest for those kinds of mathematical relations, following your love with the golden ratio :)


* Big and small radius from the cone apex
« Last Edit: 08/06/2015 11:14 PM by flux_capacitor »

Offline CraigPichach

  • Member
  • Posts: 59
  • Calgary Alberta
  • Liked: 42
  • Likes Given: 4
I like that explaination too however if this is photon momentum related (or even relativistic mass conservation) than why all the talk of issues of non-results without vibrations of unknown magnitude and frequency being required to "engage" the drive? Shouldn't the EM-Drive/Q-Thruster thrust effect than be essentially instantaneous then given it would occur at the speed of light (or are the vibrations required to fluidize the photons?)? Or do you think it just takes time for true resonance to be established? Or is that just a function of the low power magnetrons and slow circuits being used?

As to how you can reconcile that the force being larger on the Big End results in motion towards the Small End

That's why I prefer McCulloch's theory: according to MiHsC, as they travel towards the big end, photons gain momentum. When they bounce back towards the small end, they loose momentum. So to obey CoM, the cavity has to move from the big end towards the small end.
« Last Edit: 08/06/2015 11:21 PM by CraigPichach »

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
No promises, but tomorrow night, I'm going to test out my live cam setup. If I do, I'll make it a live chat session with some of you who might be around. I'll give you the link and time if I get things together.

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
I tried to model TheTraveller's EmDrive Mark 2 in 3D, which was a bit difficult since the only dimensions TT provided were end diameters Ds and Db and the length (Rb - Rs)*

Ds = 159 mm
Db = 400 mm
Rb-Rs = 240.7 mm


Moreover the dimensions in his original drawing are not proportionally on scale, maybe due to a phenomenon known as the Yang Effect ;)
But with trial & error I finally found his "magical proportions": Mr. T inscribed his spherical cone (cone + spherical base) within a cube! Hence:

Rs = Ds = 159 mm
Rb = Db = 400 mm
half-cone angle = 30░


@TheTraveller: Very clever! The more I look at your design, the more I like it.

@Rodal: can you please measure mode and resonant frequency with COMSOL for such a cavity? This is important because this cavity will be monolithically fixed (no tunable small end).

@SeeShells: It should please you in your quest for those kinds of mathematical relations, following your love with the golden ratio :)


* Big and small radius from the cone apex

This is really Shawyer's prescription for his superconducting design: spherical ends, 30 degree cone half-angle, spherical radii= diameters, etc.
« Last Edit: 08/06/2015 11:34 PM by Rodal »

Offline lmbfan

  • Member
  • Posts: 55
  • Phoenix
  • Liked: 42
  • Likes Given: 42
I tried to model TheTraveller's EmDrive Mark 2 in 3D, which was a bit difficult since the only dimensions TT provided were end diameters Ds and Db and the length (Rb - Rs)*

Ds = 159 mm
Db = 400 mm
Rb-Rs = 240.7 mm


Rs = Ds = 159 mm
Rb = Db = 400 mm
half-cone angle = 30░



I get slightly different dimensions, on the order of tenths of a millimeter.  Which brings up the question, how tight are the tolerances on the manufacturing process, and exactly how tight they need to be to get good results?

Offline frobnicat

  • Full Member
  • ****
  • Posts: 518
  • Liked: 500
  • Likes Given: 151
I'll be the first to point out that the folks here have probably had quite enough of me discussing CoE and over-unity. But to Bae...

I had the same thought. The situation is indeed similar. He is asking us to believe that thrust is proportional to the number of bounces, which is directly related to the Q. So he also has a P*Q factor in his thrust equation. And the obvious question is: how can it be that a cavity which is being pumped steady-state with input power P can yield a force that depends on P*Q? How can that be sustainable for seconds and minutes? It appears that one's extracting more than goes in. 

I confess to being uncertain about this. I saw his experiment where he got thrust that was indeed 2*P*Q/c. The saving grace might well be that photon mirrors only become efficient as the mirror velocity with respect to the light source approaches c. By this logic, he is only sipping at each beam, and despite the fact that multiple reflections are simultaneously in play (which intensifies the beam of course) he is only taking a small fraction out of each individual beam.

But that's an unsatisfactory explanation too, because one could engineer around it. Probably only a mathematical argument could settle it for me.

I can't have enough of you discussing CoE deltaMass. Had only a shallow peek on answers (by meberbs and other contributor) that ensued that request, so maybe the following remark is not relevant...

After all the discussions about Q, some aspects of Q are still unclear (at least to me, so probably also for some other readers). I found this formula about Q for optical resonator :

The Q factor of a resonator depends on the optical frequency ν0 (nu0), the fractional power loss l per round trip, and the round-trip time Trt. (assuming that l << 1)

My (perhaps crude and wrong impression) was that the "boosting factor" for radiation pressure between two mirrors, compared to that F=2*P/c for one perfect mirror with single hit per incoming photon (photon sailing) was that it would be proportional to 1/l (inverse of loss, roughly "the number of bounces") but agnostic of Trt, that is the distance between mirrors. Here assuming 0 relative velocity between mirrors.

The formula can be restated as Q=4π*(d/λ)*1/l  with d distance between mirrors and λ wavelength (please check, I'm in bad mood with factors today)
Following this definition, given mirrors of fixed quality (say 99% reflection) Q can be pumped arbitrarily high simply by having mirrors at huge distance (relative to wavelength), notwithstanding difficulties of lateral losses due to diffraction at longer distances.

For a resonator consisting of two mirrors with air (or vacuum) in between, the Q factor rises as the resonator length is increased, because this decreases the energy loss per optical cycle. However, extremely high Q values (see below) are often achieved not by using very long resonators, but rather by strongly reducing the losses per round trip.

Contrary to the "often case" indicated in this excerpt, for long range beamed force as chased by BAE, we would have long resonator indeed. For instance the calculator on linked page yields Q of nearly 1012 for IR at 1Ám, round trip time of 10000ns (1.5 km, not much distance to ask in space) and 2% round trip loss.

Would that mean it is possible to have a force a trillion time that of a photon sail (at given power) ? For instance, 600 kg force between 99% reflectance mirrors 1.5 km apart, for 1W ?? Somehow the formula must be F=some_constant_factor_around_1*(1/l)*P/c, not 2*P*Q/c, at least if their definition of Q depends on distance. Intuitively, one would expect a force from the "average number of round trips per photon" times 2P/c, the amount of energy buffered in between being irrelevant (for a constant regime). Is it possible to clarify that (before tackling moving mirrors at constant velocity, then accelerating...) ?

Add : This Q=4π*(d/λ)*1/l could indicate that the apparent Q dependence in EMdrive results might in fact be a dependence on (1/l), since d/λ hasn't been varied a lot in the experiments (I know we are not talking of a 1D problem with just two mirrors with the EMdrive, but still...)

« Last Edit: 08/07/2015 12:44 AM by frobnicat »

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2335
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2982
  • Likes Given: 2601
I tried to model TheTraveller's EmDrive Mark 2 in 3D, which was a bit difficult since the only dimensions TT provided were end diameters Ds and Db and the length (Rb - Rs)*

Ds = 159 mm
Db = 400 mm
Rb-Rs = 240.7 mm


Rs = Ds = 159 mm
Rb = Db = 400 mm
half-cone angle = 30░



I get slightly different dimensions, on the order of tenths of a millimeter.  Which brings up the question, how tight are the tolerances on the manufacturing process, and exactly how tight they need to be to get good results?

Less than a thickness of a piece of paper puts it into perspective.

Sorry can't stay.. company.

Shell

Offline jmossman

  • Member
  • Posts: 73
  • San Jose, CA
  • Liked: 58
  • Likes Given: 180
You think if you achieve cavity resonance (charge/discharge) in microseconds (say less than 20microseconds) that you would see the EM-Drive/Q-thruster thrust phenomena in microseconds?

Definitely a loaded question, but I'll take a swing:

If the "phenomena" is real, and the RF energy being injected has been properly "matched" to the EM drive, then yes, I would expect the "phenomena" to manifest within microseconds. 

However, just like Phase-Lock-Loops (PLL) within CPU's need time to lock/stabilize their internal clock frequency, I would envision some delay/latency for an EM drive control loop to find the appropriate "match" to start producing a noticeable "phenomena" signature.  Designs utilizing wide-band RF magnetrons might be the exception;  time will tell.

How to best "match"/couple the RF energy to the frustum is an interesting question, and several theories have been postulated.  The concept of evanescent waves, frequency limits based upon geometry, Q factor, the excited mode, vanilla impedance matching from RF amp to antenna to frustum...  the list is only likely to grow. 

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1314
  • Do it!
  • Vista, CA
  • Liked: 1351
  • Likes Given: 1820
As to how you can reconcile that the force being larger on the Big End results in motion towards the Small End

That's why I prefer McCulloch's theory: according to MiHsC, as they travel towards the big end, photons gain momentum. When they bounce back towards the small end, they loose momentum. So to obey CoM, the cavity has to move from the big end towards the small end.

That's what I'm finding too. I've been busy lately and going through my notes for this paper, and I realized I've said some things that were confused and backwards, regarding which way the momentum of the wave changes.

As the wave moves from the small end toward the big end it expands, but it's wavelength is getting shorter because the phase velocity is getting slower. So the momentum of the wave is increasing when moving from the small end toward the big end. Therefore, to conserve momentum, the frustum must move the other way. It feels the force that moves it in direction toward the small end. This is thrust, just like a rocket.

When the wave reflects from the big end, the amount of momentum it transfers depends on the impedance match and the angle of reflection. If the impedance doesn't match exactly, or the angle of reflection is not perpendicular, then the force will vary too. Don't ask me for the equations please, I don't have them quite right yet. So this is just hand waving, based on my understanding.

When the wave is traveling toward the small end, its momentum is decreasing and therefore, to conserve momentum the frustum must again feel a force in the "forward" direction, like a sail on a boat. This is due to the gradient, which is unique to the cone geometry. According to Zeng & Fan's paper,  if the phase constant Beta goes to zero, there is a perfect impedance match and all the power and momentum is attenuated into the frustum with no reflection.

So regardless if the wave is traveling forward or backwards inside, the frustum will be pushed "forward" toward the small end. The only component that is in the other direction, opposing this thrust is the reflection from the big end. From what I've seen in the meep models, the wave is not perpendicular to the big end. The angle of reflection produces a glancing blow. Anything that reduces the momentum transferred to the big end plate, results in thrust forward.
Todd




« Last Edit: 08/07/2015 02:34 AM by WarpTech »

Tags: