Author Topic: EM Drive Developments - related to space flight applications - Thread 2  (Read 2168210 times)

Offline Mulletron

  • Full Member
  • ****
  • Posts: 1116
  • Liked: 779
  • Likes Given: 1030
Would seem from Shawyers latest presentation, recently shared by Mulletron, that flat end plates are out and convex / concave end plates are the new standard to reduce cavity losses and boost cavity Q / thrust generation per kW of cavity stored microwave energy.

While Shawyer did share this cavity variation applied to a superconducting cavity, why it was used and what the curves are based on is new information. Would seem to make very good sense to incorporate it into non superconducting cavities.

NOT NEW information at all.

Discussed already much earlier in the thread.  I pointed out the reason.  Please take a look at the ends of the truncated cone in my exact solutions and also in Greg Egan's.

The reason for the spherical ends is because the standing waves in a truncated cone cavity are spherical waves (this is known since the 1930's as per Shelkunoff's analysis)
Is it not new information from Shawyer?

I plan to build and test a Shawyer EM Drive with his new convex / concave end plates. Will take my design lead 100% from Shawyer at 1st as he is the only source that has built multiple EM Drives, measured generated significant force / thrust and shared enough real / practical data to allow replication of his many years of blood, sweat and tears.

The theory talk here is good background but not focused enough to be something an engineer can use to built a successful working EM Drive, at least not in the 1st instance. I believe in not reinventing the wheel, until I have several working wheels to do further development from.

I also plan to duplicate Shawyers vertical force measurement system as used in his Flight Thruster qualification. To me that is a very KISS solution.

Would welcome any suggestions as I have read others here have started down this pathway.

No, the information you posted from Shawyer was not new information on this thread.

The information you posted on Shawyer's including Shawyer's presentation (which dates 6 months ago) has been discussed multiple times in this thread.

One of the discussions was with an European Architect/Designer.  We also discussed paraboloid cavities, for example.
We are talking about the same presentation that Mullerton just received from Shawyer and posted a day ago? The attached information was known 6 months ago?

Never discussed before:
Quote
The cavity comprises a small convex end plate, a
truncated conical side wall section, and a large
convex end plate. The end plate Radii R1 and R2
are selected such that R2-R1 = L1 where L1 is the
length of the side wall.
This geometry ensures that
the EM wavefront propagates between the end plates
with every point on the wavefront travelling along a
radius line of length L1, centred at point O. This
constant path length over the wavefront ensures that
phase distortion over the very large number of
reflections within a high Q cavity, is minimised, and
the value of Q that is achieved in practice
approaches the theoretical maximum. Note that this
configuration ensures that there is no orthogonal
component of the guide velocity reflected from the
side wall, thus ensuring a zero side wall force
component in the axial plane.

Source IAC-14 paper:
http://tinyurl.com/ofl4527
« Last Edit: 05/04/2015 07:35 PM by Chris Bergin »
Challenge your preconceptions, or they will challenge you. - Velik

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
Would seem from Shawyers latest presentation, recently shared by Mulletron, that flat end plates are out and convex / concave end plates are the new standard to reduce cavity losses and boost cavity Q / thrust generation per kW of cavity stored microwave energy.

While Shawyer did share this cavity variation applied to a superconducting cavity, why it was used and what the curves are based on is new information. Would seem to make very good sense to incorporate it into non superconducting cavities.

NOT NEW information at all.

Discussed already much earlier in the thread.  I pointed out the reason.  Please take a look at the ends of the truncated cone in my exact solutions and also in Greg Egan's.

The reason for the spherical ends is because the standing waves in a truncated cone cavity are spherical waves (this is known since the 1930's as per Shelkunoff's analysis)
Is it not new information from Shawyer?

I plan to build and test a Shawyer EM Drive with his new convex / concave end plates. Will take my design lead 100% from Shawyer at 1st as he is the only source that has built multiple EM Drives, measured generated significant force / thrust and shared enough real / practical data to allow replication of his many years of blood, sweat and tears.

The theory talk here is good background but not focused enough to be something an engineer can use to built a successful working EM Drive, at least not in the 1st instance. I believe in not reinventing the wheel, until I have several working wheels to do further development from.

I also plan to duplicate Shawyers vertical force measurement system as used in his Flight Thruster qualification. To me that is a very KISS solution.

Would welcome any suggestions as I have read others here have started down this pathway.

No, the information you posted from Shawyer was not new information on this thread.

The information you posted on Shawyer's including Shawyer's presentation (which dates 6 months ago) has been discussed multiple times in this thread.

One of the discussions was with an European Architect/Designer.  We also discussed paraboloid cavities, for example.
We are talking about the same presentation that Mullerton just received from Shawyer and posted a day ago? The attached information was known 6 months ago?

Never discussed before:
Quote
The cavity comprises a small convex end plate, a
truncated conical side wall section, and a large
convex end plate. The end plate Radii R1 and R2
are selected such that R2-R1 = L1 where L1 is the
length of the side wall.
This geometry ensures that
the EM wavefront propagates between the end plates
with every point on the wavefront travelling along a
radius line of length L1, centred at point O. This
constant path length over the wavefront ensures that
phase distortion over the very large number of
reflections within a high Q cavity, is minimised, and
the value of Q that is achieved in practice
approaches the theoretical maximum. Note that this
configuration ensures that there is no orthogonal
component of the guide velocity reflected from the
side wall, thus ensuring a zero side wall force
component in the axial plane.

Source IAC-14 paper:
http://tinyurl.com/ofl4527


The statement that R2-R1=L1 is not new information, it is a definition, a necessary tautology.

The length L1 must be equal to R2-R1, so no new information of any kind is given.



The dimensionless ratios for Shawyer's superconducting drive were estimated, for example:

http://forum.nasaspaceflight.com/index.php?topic=36313.msg1335023#msg1335023

Including the most important parameter: the cone angle

It would be appreciated if Roger Shawyer could provide the actual dimensions for R1, R2 and L1 for the superconducting design, since while the dimensionless ratios can be accurately estimated from a diagram, the actual dimensions cannot (as known from painstaking efforts by aero,  Lasoi, Flyby, Fornaro and other people in this thread).

@aero reports in previous posts that he asked Shawyer for the dimensions of Shawyer's EM Drives used in his reported experiments and received the answer that the small base was chosen on the basis of the cutoff frequency (and no further ado  :( ).  Therefore dimensions for the Flight Thruster, Experimental, Demonstrator and other Shawyer's experiments would be helpful to @notsosureofit, @aero, McCulloch and others in order to compare with the predicting formulas.

_____________

PS: Thanks once again to Paul March for patiently and willingly providing all geometrical data he was asked for regarding the NASA experiments.  We are lacking the actual geometrical dimensions for Shawyer's experiments and for the Chinese experiments.
« Last Edit: 05/05/2015 08:22 PM by Rodal »

Offline zellerium

  • Full Member
  • *
  • Posts: 171
  • Pittsburgh, PA
  • Liked: 279
  • Likes Given: 400
I have not yet spent enough time working with the proposed models, thus I'll stay out of the theory debate.

Instead, I am focusing on the difference in parameters of each experiment--what does the thrust depend on? EW saw a clear dependence on a dielectric, but Shawyer, NWPU and Fetta observed high efficiencies without one. But why does Shawyer mention the "dielectric-filled section" in one of his papers and not reference it in later experiments? http://www.emdrive.com/theorypaper9-4.pdf

Has NWPU tried their magnetron cavity using a dielectric? I believe I read a comment referencing a 2014 paper by Yang but I have not been able to find it.
Did EW test without a dielectric in a vacuum?

We don't have the money to build a sensitive torsion pendulum, thus my team is building a teeter-totter. It seems any spurious effects present with a teeter-totter would also be present on a Cavendish balance, but correct me if I'm wrong. I admit connection to an external power supply is going to be an issue, but most likely a consistent issue. High power batteries are not in our budget at the moment.

Instead of eliminating all extraneous possibilities, we plan to change the parameters and observe the change in thrust. We start by changing the dielectric parameters: 4 inches thick to .5 inch thick discs. Next we change dielectric location, and maybe even shape while we're at it. We can also test upward and downward orientations to observe a change due to the gravity vector. I would be very surprised if we saw identical thrust from all of these setups, hopefully we can single out the best and then figure out why it is the best.

We have designed a cylindrical cavity to simplify the problem. Solutions to resonant freq are much easier and more intuitive in a cylinder, and hitting resonance (when we adjust the dielectric) with a constant frequency (from a microwave oven magnetron) will be possible by adjusting the height of the cylinder. We will know this week whether we will receive funding for a variable frequency/power setup which will make things much easier as far as resonance goes. But I am worried that small powers (50 W) won't produce enough thrust to be seen on the teeter-totter, and a high power amp is too expensive.

   
I think this entire discussion should be moved to reddit.com/r/emdrive. The comment structure will allow the most informed individuals in this thread to quickly rise to the top. Replies to certain comments will be separated and various topics can be explored in separate threads. Just a more efficient system in my opinion...



Offline zen-in

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


Chinese did report their cavity bandwidth data:
http://www.emdrive.com/yang-juan-paper-2012.pdf

Interesting.   Earlier in the first thread the discussion centered around cavity Q and the need for a high Q.   This cavity only has a Q = 1531.
« Last Edit: 05/04/2015 07:25 PM by zen-in »

Offline zellerium

  • Full Member
  • *
  • Posts: 171
  • Pittsburgh, PA
  • Liked: 279
  • Likes Given: 400


Chinese did report their cavity bandwidth data:
http://www.emdrive.com/yang-juan-paper-2012.pdf

Interesting.   Earlier in the first thread the discussion centered around cavity Q and the need for a high Q.   This cavity only has a Q = 1531.

The 2013 paper from NWPU: "It was found that the thruster
cavity made by copper and resonating on the equivalent TE011
mode has a quality factor 320400 and generates total net EM
thrust 411 mN for 1000 W 2.45 GHz incident microwave." 
http://iopscience.iop.org/1674-1056/22/5/050301

Offline Rodal

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


Chinese did report their cavity bandwidth data:
http://www.emdrive.com/yang-juan-paper-2012.pdf

Interesting.   Earlier in the first thread the discussion centered around cavity Q and the need for a high Q.   This cavity only has a Q = 1531.

The 2013 paper from NWPU: "It was found that the thruster
cavity made by copper and resonating on the equivalent TE011
mode has a quality factor 320400 and generates total net EM
thrust 411 mN for 1000 W 2.45 GHz incident microwave." 
http://iopscience.iop.org/1674-1056/22/5/050301
The Q's reported in the tables 1 and 2 of http://www.emdrive.com/NWPU2010paper.pdf  and in the above quote are much higher than the calculated Q based on the bandwidth. 

Zen-In's calculated Q from the bandwidth is correct: Q = f/bandwidthf (by definition)

Can anybody make sense of the discrepancy?
« Last Edit: 05/04/2015 07:49 PM by Rodal »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
I repost here this post which may be of interest to people here as well:

So this made me think about the time I took a 3 port circulator apart. For those that are not familiar with circulators, the work like this:
Put power in port 1 and it comes out port 2, put it in 2 and it comes out 3, put it in 3 and it comes out 1. All with out much loss. But if you try to go backwards, say 3 to 2,  you loose 99% of the power.

Cool little device. So when I take it apart all it is is a flat triangle of copper, 2 triangle shaped pieces of ferrite, and a magnet.

If you don't know the math behind it, is looks at first blush as "silly" as the emdrive. No way could it do that. But it does. This thing may well work, we just don't know the math.

I'm an Engineer and I've studied the Polarizable Vacuum Model of General Relativity. What it would say is the following;

As a waveguide, the group velocity is something like;

v_g = c x sqrt(1 - (c/2d*f)^2)

Where, c is the usual speed of light, d is the diameter of the cylinder, and f is the frequency of the microwave excitation.

c/2d = fc,  is the Low cut-off frequency of the waveguide.

The refractive index depends on the Low cut-off frequency as a function of the diameter,

K = 1/sqrt( 1 - (fc/f)^2)

For f >> fc, K~1. But for frequencies in the band fc1 < f <~ fc2, K is much larger.

There is a strong gradient in the refractive index from one end of the cone to the other. This "mimics" gravity, as interpreted in the PV Model.

Therefore, we can assume there is a "gravitational" gradient in the microwave band refractive index, along the length of the cone. At one end they have diameter d1, and at the other end they have diameter d2, and d1 > d2. Below fc1, the mode frequencies exponentially decay to zero. Just like the Casimir effect.

Here is how it conserves momentum;

In the PV Model, momentum transforms as,

p => p*sqrt(K)

In a resonant cavity, p is the SUM of all the photons “in phase", minus the losses of the cavity. 

However, as photons “fall” from the large end toward the small end, they gain momentum, which is passed on to the cone when they are reflected from the small end. The photon then loses momentum as it travels back to the large end, where it imparts “less” momentum to the large end. The result is a NET propulsion in the direction of the small end. In other words, the photons are blue-shifted falling forward, and red-shifted going backwards, due to the gradient in the refractive index. It is literally gravitational red & blue shift, according to the PV Model.

The interesting thing is, the refractive index in the waveguide does not depend on the power of the microwaves, or the energy density. It is simply a matter of the geometry and frequency band relative to the cut-off. What matters more, is having enough resonant momentum stored to make the effect noticeable.

That’s IMHO as an engineer of course. Any comments?

See PV Model: https://www.researchgate.net/publication/223130116_Advanced_Space
_Propulsion_Based_on_Vacuum_%28Spacetime_Metric%29_Engineering

Todd D.

In what sense does this conserve momentum?

Treat the device as a black box. I don't know or care what is happening inside it. At time T0 it has no momentum. Turn it on and let it accelerate so that it has some velocity and so momentum at time T1. Unless you can point to something with the same amount of momentum going in the other direction then by definition you have violated conservation of momentum. What happens inside the box simple does not affect that fact.

When you drop an object and it falls to the ground. Relative to you, it gained momentum from the gravitational field. It did not expel any propellant to fall. The gravitational field is simply a gradient in the refractive index of the vacuum surrounding the Earth. If you can explain conservation of momentum for falling objects in a gravitational field, then you have your answer.

(Edit) In other words, if the cavity were not tapered, then you have equal momentum inside traveling left and right. At T0, it will go nowhere. However, because it is tapered such that you have a gradient in the refractive index, then "just like gravity", photons will be blue shifted moving into higher K, and red shifted moving into lower K, because momentum,

p => p*sqrt(K)

THIS is a violation of conservation of momentum. Therefore, the cavity must move to conserve momentum, as it tries to establish equilibrium with it's own internal stress.


Regards,
Todd D.

What would be the equation for the acceleration or the force, given the geometrical dimensions of the truncated cone, the Q, the input power, the frequency, and any other variables? Do you have a closed-form solution that could be compared to actual experimental results and also compared with the equation of Shawyer, and also to the equation of McCulloch ?

Regards,

JR

Not yet... I'm just now coming to grips with this myself. My light-bulb went off when I realized if the frequency of the microwaves is very close to the cut-off frequencies, then the speed of light will have a very large gradient inside the Frustum. Relative to the "traveling" waves (photons) attempting to move at the speed of light from end to end. When they approach the small end, their wavelength is squeezed by the reduced group velocity. Momentum depends on wavelength;

p = h/lambda

wavelength depends on velocity, and v_g is a variable inside the frustum.

That is where the momentum is coming from. Inside the Frustum, relative to the traveling waves you have an accelerated reference frame, into which you are injecting photons that are affected by this manufactured "gravitational" field, that must be compensated for by moving the Frustum.

I'll see what I can come up with for a formal equation, but I've got a day job. As for @ppnl, you will never get a Newtonian-type COM equation out of this. The two frames are the Frustum, and the frame of the moving photons inside it. The acceleration is caused by the geometry of the waveguide or a variable refractive index, i.e. the GR or PV Interpretation lead to the same result.

Todd D.
« Last Edit: 05/04/2015 08:21 PM by Rodal »

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Although most of what's being discussed here concerns the static force produced by an EmDrive, I would like to look ahead to the dynamics in free space. In this regard, and needing no recourse to relativity for low velocities, all types of propellantless propulsion device are created equal; i.e. we model them with constant input power which results in a constant thrust, which in turn produces a constant acceleration. A little high school physics and a dash of algebra (see below) should suffice to convince you that, for any and all such devices, there exists a break-even velocity, above which more energy has been produced than has been consumed. As the device continues to further accelerate, a continuous source of free power is available. Thus we have not only perpetual motion, but free energy to boot.

It is worthwhile to quantify the value of this break-even velocity, which turns out to be a very simple expression. The relevant equations we use (symbols having their usual meaning) are:
k := F/P in Newton/Watt
v = a t = (F/m) t
Ein = P t
Eout = 0.5 m v2

From the first 2 equations we get
t = (m v) / (k P)

At breakeven, Ein= Eout and v := v0
or
2 P m v0 / (k P) = mv02
so
v0 = 2/k

Armed with this expression, the performance of any propellantless propulsion device may be examined, if its 'k' value is known.
« Last Edit: 05/04/2015 08:51 PM by deltaMass »

Offline UneducatedNitwit

  • Member
  • Posts: 5
  • Liked: 4
  • Likes Given: 0
Would @Chris Bergin be willing to set up a private thread just for talking with Mr. Shawyer, undisturbed by outside agitators?

We could create a standalone Q&A thread, like this one?

http://forum.nasaspaceflight.com/index.php?topic=37295.0

(Notice the format is different to a normal thread).

Please do this.  Although I haven't seen the moderated posts, I realize from what @Rodal has said recently that perhaps not everyone who came here from forums and link aggregators has been polite or even spent very long reading your discussion.  That's incredibly disappointing, but I guess from what I have seen on Reddit.com/r/physics it should be expected every time there is a new development.  There are very many people who are angered by the idea of taking this experiment, your discussion, and the credentials of the people involved here seriously.  They don't realize that this discussion has already covered all the bases.

It took me nearly a week to skim through the entirety of both threads.  You can't expect most people coming from other websites to dedicate their time that way before posting.  However, I think we should be given the opportunity to read what is written without participating unless we have something of value to say.

« Last Edit: 05/04/2015 08:20 PM by UneducatedNitwit »

Offline Flyby

  • Full Member
  • ***
  • Posts: 385
  • Belgium
  • Liked: 447
  • Likes Given: 48
Did anyone notice the appearance of piezoelectric elements in the updated 2nd generation drawing?
Apparently, a computerized system to adjust the walls of the supercooled frustum, in order to compensate for the Doppler effect...

I thought that is new, compared to previous drawings of EMdrive2.0 ?

When reading that new text i do start to wonder if this is a description of an already build item, or a description of an item to build?

Offline MazonDel

  • Member
  • Posts: 37
  • Worcester, MA
  • Liked: 41
  • Likes Given: 36
Speaking of KM, is there anyone out there that has what it takes and is willing to volunteer and set up a wiki or something?
http://en.wikipedia.org/wiki/MediaWiki

I've set up a MediaWiki server before and it wasn't too bad. That is a good platform. I'm simply stretched too thin right now to try it again. My previous attempt at KM (just a simple Google Doc, which was pretty bad didn't catch on so I deleted it:
https://drive.google.com/folderview?id=0B4PCfHCM1KYoTXhSUTd5ZDN2WnM&usp=sharing

@Mulletron et al., I am happy to assist in this.  I agree, this forum is excellent for reviewing the latest contributions by all the excellent minds here, but newer folks are finding it difficult to get a handle on how we got here over the past hundred pages.  As a result some of the same questions are being re-asked repeatedly.  Ideally it might become the FAQ section that was discussed previously (which @Rodal noted some valid concerns regarding).  Worth a try, at least.

To this end, I have set up a MediaWiki server at http://emdrive.echothis.com and will start by the adding the relevant links in today.  However, I'm just a (non-practicing) mechanical engineer, not a physicist so my role in this would be limited to setting up the organization and linking to the relevant content already posted by people far more knowledgeable than I.  Anyone who has been tracking the forum and would like to join in this endeavor, welcome!

I would be most happy to join this effort!

Quite amusingly, I only ever look at updates on this thread during the week (worktime simulations are...boring) and this weekend I had decided that what seemed like the best thing I could do to help out the efforts here (since I'm not particularly knowledgeable with this sort of physics) would be to set up the wiki. When I showed up this morning I saw that I had a rather astounding 13+ pages of forum thread to catch up on! (Usually it seems to be about 4-5).

So when I read Mulletron's suggestion my thoughts were "Ah ha! I'm on the case! But I should finish catching up first." and then I saw you had started one up.

My provisional plan had been to slowly work through from page 1 again creating topics for things as they came up and beginning to organize them as the groupings became obvious.

A few groupings I had identified earlier:
Possible Error Sources of Thrust: Things like thermal buckling, atmospheric effects, etc would get pages of their own and this grouping would just allow for easy perusing of these topics. It might be wise to organize them into some sort of 'Busted', 'Under Investigation', 'Other' groupings. Of course new information can elevate one of the topics out of the busted area or descend it as necessary.

Current Theories: Relatively simple, listing out Shawyer's, White's, the Chinese Teams, Mulletron's, etc theories on what is going on.

Thrust Modes and Results: This one probably just ends up being a table of its own, but is probably particularly specific to the Eaglework's team.

Status of DIY Efforts: Lets Mulletron, Notsosureofit, DIYFan, and others that are replicating this provide easy to reference status updates and other things (future plans, etc).

What I was envisioning was to attempt to use the wiki-source system to provide pointers from information to the forum posts concerning them.

Thoughts?
« Last Edit: 05/04/2015 08:26 PM by MazonDel »

Offline Notsosureofit

  • Full Member
  • ****
  • Posts: 656
  • Liked: 704
  • Likes Given: 1367
I repost here this post which may be of interest to people here as well:

So this made me think about the time I took a 3 port circulator apart. For those that are not familiar with circulators, the work like this:
Put power in port 1 and it comes out port 2, put it in 2 and it comes out 3, put it in 3 and it comes out 1. All with out much loss. But if you try to go backwards, say 3 to 2,  you loose 99% of the power.

Cool little device. So when I take it apart all it is is a flat triangle of copper, 2 triangle shaped pieces of ferrite, and a magnet.

If you don't know the math behind it, is looks at first blush as "silly" as the emdrive. No way could it do that. But it does. This thing may well work, we just don't know the math.

I'm an Engineer and I've studied the Polarizable Vacuum Model of General Relativity. What it would say is the following;

As a waveguide, the group velocity is something like;

v_g = c x sqrt(1 - (c/2d*f)^2)

Where, c is the usual speed of light, d is the diameter of the cylinder, and f is the frequency of the microwave excitation.

c/2d = fc,  is the Low cut-off frequency of the waveguide.

The refractive index depends on the Low cut-off frequency as a function of the diameter,

K = 1/sqrt( 1 - (fc/f)^2)

For f >> fc, K~1. But for frequencies in the band fc1 < f <~ fc2, K is much larger.

There is a strong gradient in the refractive index from one end of the cone to the other. This "mimics" gravity, as interpreted in the PV Model.

Therefore, we can assume there is a "gravitational" gradient in the microwave band refractive index, along the length of the cone. At one end they have diameter d1, and at the other end they have diameter d2, and d1 > d2. Below fc1, the mode frequencies exponentially decay to zero. Just like the Casimir effect.

Here is how it conserves momentum;

In the PV Model, momentum transforms as,

p => p*sqrt(K)

In a resonant cavity, p is the SUM of all the photons “in phase", minus the losses of the cavity. 

However, as photons “fall” from the large end toward the small end, they gain momentum, which is passed on to the cone when they are reflected from the small end. The photon then loses momentum as it travels back to the large end, where it imparts “less” momentum to the large end. The result is a NET propulsion in the direction of the small end. In other words, the photons are blue-shifted falling forward, and red-shifted going backwards, due to the gradient in the refractive index. It is literally gravitational red & blue shift, according to the PV Model.

The interesting thing is, the refractive index in the waveguide does not depend on the power of the microwaves, or the energy density. It is simply a matter of the geometry and frequency band relative to the cut-off. What matters more, is having enough resonant momentum stored to make the effect noticeable.

That’s IMHO as an engineer of course. Any comments?

See PV Model: https://www.researchgate.net/publication/223130116_Advanced_Space
_Propulsion_Based_on_Vacuum_%28Spacetime_Metric%29_Engineering

Todd D.

In what sense does this conserve momentum?

Treat the device as a black box. I don't know or care what is happening inside it. At time T0 it has no momentum. Turn it on and let it accelerate so that it has some velocity and so momentum at time T1. Unless you can point to something with the same amount of momentum going in the other direction then by definition you have violated conservation of momentum. What happens inside the box simple does not affect that fact.

When you drop an object and it falls to the ground. Relative to you, it gained momentum from the gravitational field. It did not expel any propellant to fall. The gravitational field is simply a gradient in the refractive index of the vacuum surrounding the Earth. If you can explain conservation of momentum for falling objects in a gravitational field, then you have your answer.

(Edit) In other words, if the cavity were not tapered, then you have equal momentum inside traveling left and right. At T0, it will go nowhere. However, because it is tapered such that you have a gradient in the refractive index, then "just like gravity", photons will be blue shifted moving into higher K, and red shifted moving into lower K, because momentum,

p => p*sqrt(K)

THIS is a violation of conservation of momentum. Therefore, the cavity must move to conserve momentum, as it tries to establish equilibrium with it's own internal stress.


Regards,
Todd D.

What would be the equation for the acceleration or the force, given the geometrical dimensions of the truncated cone, the Q, the input power, the frequency, and any other variables? Do you have a closed-form solution that could be compared to actual experimental results and also compared with the equation of Shawyer, and also to the equation of McCulloch ?

Regards,

JR

Not yet... I'm just now coming to grips with this myself. My light-bulb went off when I realized if the frequency of the microwaves is very close to the cut-off frequencies, then the speed of light will have a very large gradient inside the Frustum. Relative to the "traveling" waves (photons) attempting to move at the speed of light from end to end. When they approach the small end, their wavelength is squeezed by the reduced group velocity. Momentum depends on wavelength;

p = h/lambda

wavelength depends on velocity, and v_g is a variable inside the frustum.

That is where the momentum is coming from. Inside the Frustum, relative to the traveling waves you have an accelerated reference frame, into which you are injecting photons that are affected by this manufactured "gravitational" field, that must be compensated for by moving the Frustum.

I'll see what I can come up with for a formal equation, but I've got a day job. As for @ppnl, you will never get a Newtonian-type COM equation out of this. The two frames are the Frustum, and the frame of the moving photons inside it. The acceleration is caused by the geometry of the waveguide or a variable refractive index, i.e. the GR or PV Interpretation lead to the same result.

Todd D.

Yep, right on target.  Accelerated frames:  GR likes to use time dilation, PV likes variable speed of light (ie refractive index)  They both should be able to get the formula.  The key point is that force needs accelerated frames.  Inertial frames won't do the job.  Also the force will diminish w/ the actual acceleration and get to a balance depending on the total mass of the system compared to the "weight" of the photons.

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
....
We have designed a cylindrical cavity to simplify the problem. Solutions to resonant freq are much easier and more intuitive in a cylinder, and hitting resonance (when we adjust the dielectric) with a constant frequency (from a microwave oven magnetron) will be possible by adjusting the height of the cylinder. ....
But a cylindrical cavity displays no attenuation and no focusing in the axial direction (unlike the truncated cone).  The experimenters in the US, UK and China are using truncated cones. 

If you obtain no net thrust for the cylindrical cavity, will that have been worth the effort you went through (instead of testing a truncated cone, preferably with the same geometry dimensions as NASA's or one of Shaywer's ? )

Concerning obtaining the frequency and mode shape, does your University give you access to a Finite Element analysis package like COMSOL, or ANSYS Multiphysics, etc. so that you could then obtain a numerical solution for any arbitrary geometry and inserted dielectric ?
« Last Edit: 05/04/2015 08:56 PM by Rodal »

Offline zellerium

  • Full Member
  • *
  • Posts: 171
  • Pittsburgh, PA
  • Liked: 279
  • Likes Given: 400
....
We have designed a cylindrical cavity to simplify the problem. Solutions to resonant freq are much easier and more intuitive in a cylinder, and hitting resonance (when we adjust the dielectric) with a constant frequency (from a microwave oven magnetron) will be possible by adjusting the height of the cylinder. ....
But a cylindrical cavity displays no attenuation and no focusing in the axial direction (unlike the truncated cone).  The experimenters in the US, UK and China are using truncated cones. 

If you obtain no net thrust for the cylindrical cavity, will that have been worth the effort you went through (instead of testing a truncated cone, preferably with the same geometry dimensions as NASA's or one of Shaywer's ? )

Concerning obtaining the frequency and mode shape, does your University give you access to a Finite Element analysis package like COMSOL, or ANSYS Multiphysics, etc. so that you could then obtain a numerical solution for any arbitrary geometry and inserted dielectric ?

Dr. Rodal:

I was under the impression that a symmetric shape should produce a net thrust, based on the null Cannae test article at EW. The only asymmetry there was the dielectric slug in the end, correct?

I don't believe we have access to COMSOL, but I will look into another method.

We have yet to purchase materials and may be able to test a truncated shape as well as a symmetric shape with the same copper tube and end plates.

Offline MazonDel

  • Member
  • Posts: 37
  • Worcester, MA
  • Liked: 41
  • Likes Given: 36
Zellerium, I am uncertain about if your University allows for this sort of thing or not, but for the purposes of the nascent Wiki under development, would you be willing to provide (later once you have them) some Bills of Materials as well as instructions documenting how you assembled and set up your rig?

I imagine that if you were to provide almost an Instructables-level documentation on how your team does what it does, that it would also better allow for people here to make suggestions on possible tests to look into or to help overcome problems you may run into.

Speaking of KM, is there anyone out there that has what it takes and is willing to volunteer and set up a wiki or something?
http://en.wikipedia.org/wiki/MediaWiki

I've set up a MediaWiki server before and it wasn't too bad. That is a good platform. I'm simply stretched too thin right now to try it again. My previous attempt at KM (just a simple Google Doc, which was pretty bad didn't catch on so I deleted it:
https://drive.google.com/folderview?id=0B4PCfHCM1KYoTXhSUTd5ZDN2WnM&usp=sharing

@Mulletron et al., I am happy to assist in this.  I agree, this forum is excellent for reviewing the latest contributions by all the excellent minds here, but newer folks are finding it difficult to get a handle on how we got here over the past hundred pages.  As a result some of the same questions are being re-asked repeatedly.  Ideally it might become the FAQ section that was discussed previously (which @Rodal noted some valid concerns regarding).  Worth a try, at least.

To this end, I have set up a MediaWiki server at http://emdrive.echothis.com and will start by the adding the relevant links in today.  However, I'm just a (non-practicing) mechanical engineer, not a physicist so my role in this would be limited to setting up the organization and linking to the relevant content already posted by people far more knowledgeable than I.  Anyone who has been tracking the forum and would like to join in this endeavor, welcome!

I would be most happy to join this effort!

Quite amusingly, I only ever look at updates on this thread during the week (worktime simulations are...boring) and this weekend I had decided that what seemed like the best thing I could do to help out the efforts here (since I'm not particularly knowledgeable with this sort of physics) would be to set up the wiki. When I showed up this morning I saw that I had a rather astounding 13+ pages of forum thread to catch up on! (Usually it seems to be about 4-5).

So when I read Mulletron's suggestion my thoughts were "Ah ha! I'm on the case! But I should finish catching up first." and then I saw you had started one up.

My provisional plan had been to slowly work through from page 1 again creating topics for things as they came up and beginning to organize them as the groupings became obvious.

A few groupings I had identified earlier:
Possible Error Sources of Thrust: Things like thermal buckling, atmospheric effects, etc would get pages of their own and this grouping would just allow for easy perusing of these topics. It might be wise to organize them into some sort of 'Busted', 'Under Investigation', 'Other' groupings. Of course new information can elevate one of the topics out of the busted area or descend it as necessary.

Current Theories: Relatively simple, listing out Shawyer's, White's, the Chinese Teams, Mulletron's, etc theories on what is going on.

Thrust Modes and Results: This one probably just ends up being a table of its own, but is probably particularly specific to the Eaglework's team.

Status of DIY Efforts: Lets Mulletron, Notsosureofit, DIYFan, and others that are replicating this provide easy to reference status updates and other things (future plans, etc).

What I was envisioning was to attempt to use the wiki-source system to provide pointers from information to the forum posts concerning them.

Thoughts?

I just visited the page.  Great work saucyjack and MazonDel.  Will there be a section for Null tests of the drives as well?  That would be a great addition IMO- maybe to the "Building" section.  Here's the only null test I'm aware of:
http://enu.kz/repository/2009/AIAA-2009-5070.pdf
though Dr. Rodal and other posters might be able to provide more information as far as that goes.

I also think a link at the top or bottom of the most current page of this forum would be very helpful.  People coming to this forum for the first time will continue to broach covered topics if they aren't aware of the wiki link.
« Last Edit: 05/04/2015 10:48 PM by lasoi »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
....
We have designed a cylindrical cavity to simplify the problem. Solutions to resonant freq are much easier and more intuitive in a cylinder, and hitting resonance (when we adjust the dielectric) with a constant frequency (from a microwave oven magnetron) will be possible by adjusting the height of the cylinder. ....
But a cylindrical cavity displays no attenuation and no focusing in the axial direction (unlike the truncated cone).  The experimenters in the US, UK and China are using truncated cones. 

If you obtain no net thrust for the cylindrical cavity, will that have been worth the effort you went through (instead of testing a truncated cone, preferably with the same geometry dimensions as NASA's or one of Shaywer's ? )

Concerning obtaining the frequency and mode shape, does your University give you access to a Finite Element analysis package like COMSOL, or ANSYS Multiphysics, etc. so that you could then obtain a numerical solution for any arbitrary geometry and inserted dielectric ?

Dr. Rodal:

I was under the impression that a symmetric shape should produce a net thrust, based on the null Cannae test article at EW. The only asymmetry there was the dielectric slug in the end, correct?

I don't believe we have access to COMSOL, but I will look into another method.

We have yet to purchase materials and may be able to test a truncated shape as well as a symmetric shape with the same copper tube and end plates.

1) At the moment we have a situation where there is only one research location worldwide that has provided research data to analyze (force vs. time curves, geometrical dimensions, testing parameters, testing equipment details, etc. etc.).  This is the only location that has also tested a truncated cone in a vacuum.  This location is NASA Eagleworks, of course.  NASA did not do any testing with the Cannae device in vacuum, to my knowledge.

2) There is (some of them are Ph.D. thesis) research data on this type of measurements going back for more than 50 years, that indicates the problem with thermal convection currents, but there is a documented way around it that apparently neither NASA, nor Shawyer, nor Yang in China have utilized.  More on this lately if you don't have a vacuum chamber and are planning to test in ambient pressure.  Ditto for Mulletron and everybody else trying to do experiments in ambient pressure: there is a better way that has not yet been discussed in this thread.

3) If you don't have access to COMSOL, can you inquire whether you have access to ANSYS-Multiphysics [which is just as capable for these purposes, really] or another FEA package for multiphysics ?
« Last Edit: 05/04/2015 10:17 PM by Rodal »

Offline TheTraveller

Instead, I am focusing on the difference in parameters of each experiment--what does the thrust depend on? EW saw a clear dependence on a dielectric, but Shawyer, NWPU and Fetta observed high efficiencies without one. But why does Shawyer mention the "dielectric-filled section" in one of his papers and not reference it in later experiments? http://www.emdrive.com/theorypaper9-4.pdf

Has NWPU tried their magnetron cavity using a dielectric? I believe I read a comment referencing a 2014 paper by Yang but I have not been able to find it.
Did EW test without a dielectric in a vacuum?
Early on Shawyer did use a dielectric but gave it up saying he found it reduced Q and increased losses. Mullerton shared this email he recently received from Shawyer where the non use of a dielectric is stated.
« Last Edit: 05/04/2015 10:25 PM 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 zellerium

  • Full Member
  • *
  • Posts: 171
  • Pittsburgh, PA
  • Liked: 279
  • Likes Given: 400
Zellerium, I am uncertain about if your University allows for this sort of thing or not, but for the purposes of the nascent Wiki under development, would you be willing to provide (later once you have them) some Bills of Materials as well as instructions documenting how you assembled and set up your rig?

MazonDel:
I am also unfamiliar with our school's policy on the matter, but I will check to make sure it isn't an issue.
I'll finish the schematics for the cavity and counterbalanced teeter-totter soon and post them here.

Dr. Rodal:
Do you think the thrust signiture observed from the null Cannae drive could be due to air currents?
I assume the frustum yielded similar thrusts while in a vacuum and in air, which shows how much the air is effecting thrust.  Unless the shape of the Cannae drive led to a significant air current, it seems a symmetric cavity should produce thrust.

Although we may be able to find the resonant frequencies for the frustum using software, changing parameters like the dielectric thickness will change the frequency required. And if we have to use a microwave magnetron then we have to change the frustum's size which is more challenging, but still possible.
However, if we recieve enough money for the variable frequency test a tapered cavity would be a much better option.

I just downloaded EM Pro recently, but I only have experience with 3-D design software (Creo) and Matlab. We may have licenses available for Multiphysics, I'll look into it.

TheTraveler:
Thank you, I didn't see that email.
Perhaps we are overestimating the effect of the dielectric and our experiment is headed in the wrong direction... On the other hand, could cavity geometry and signal manipulation be the only factors determining efficiency?
« Last Edit: 05/04/2015 11:28 PM by zellerium »

Offline TheTraveller

if what  you say is true there is no way Boeing or anyone else involved with any form of propulsion but especially aviation or space flight would lose interest. any reduced mass is astoundingly useful. Even if the thing only reduced mass by 1 percent it would revolutionize everything. Nothing would ever be the same again. Automobiles, trains, planes, rockets; everything.
Is not what I say. Is what Shawyer explains in his how to measure the forces pdf and in his explanation of how his space plane would do a 0.05g vertical lift.

As another example of EM Drive ratchet mode operation, assume we had a EM Drive with a motor & generator mode maximum force generation of 9.8 Newtons and the device had a mass of 0.5kg.

Now support it 1 mtr off the ground via say a small pedestal table, oriented such that the force of gravity would put the EM Drive into Generator / force resistance mode.

Next switch it on.

Now remove the pedestal support table.

Observe it is hovering as the downward 0.5kg weight is opposed by generator mode, which has a max ability to resist 9.8 Newtons of force or 1kg of mass at the Earth's surface.


So if I read you well and equivalence principle holds, having a rocket in deep space accelerating, by conventional mean, at 1g, a floor that is orthogonal to this acceleration, a pedestal resting on this floor, a powered EM drive resting on this pedestal in same configuration, remove the pedestal and one will observe the EM drive hovering above that floor (that is still accelerating at 1g). Meaning we now have a powered EM drive not needing to be "pushed" to accelerate at 1g (there is no longer any interaction between rocket and device).

Stop the conventional thrust of the rocket : the rocket will stop accelerating and proceed as an inertial mass at constant velocity (relative to whatever inertial frame). Let the device escape from an open front bay : it will continue to accelerate at 1g since it was no longer interacting with the rocket when the change in acceleration of the rocket occurred (no interaction => whatever change in rocket trajectory ignored). We now have a "conventionally" accelerating (thrusting) EM drive needing no added force.

So why bother with a Hall thruster ? Just put behind a big dumb powder booster that makes your EM drive accelerate at 1g for a fraction of a second the time it takes for the EM drive to "record" that acceleration as a "starting point". BTW, same argument above could be made if acceleration was 0.5g instead of 1g : this means we now have to add a new intrinsic variable to a moving object. What physical mechanism explains this memory effect of "initial acceleration" into an ongoing acceleration of given magnitude ?

Quote
...
I trust this shows how the EM Drive is unlike anything humanity has experienced before. So please do not move forward thinking it is like a propellantless Hall thruster or rocket motor that generates thrust when the cavity is filled with microwave energy.

Don't underestimate the aptitude of intelligent people to integrate counter intuitive formal systems when they show internal consistency.
Anyway after some sleep and morning coffee, I can now answer your question.

As I see your thought experiment, the hovering EM Drive in the 1g accelerating ship is actually accelerating at 1g, just not in reference to the ship accelerating at 1g. So if it can hover in the 1g accelerating reference frame of the ship, when the ships stops accelerating, the EM Drive will of course continue to accelerate at 1g.

Which BTW it really can't do as energy demand to increasing device kinetic energy conversion is going up with the square of the velocity. As the EM Drive has a max microwave energy capacity in the cavity to draw upon for conversion into device kinetic energy and same for the microwave source and same for the primary energy supply, once the accelerating kinetic demand exceeds that of the cavity / microwave source / primary energy source, the EM Drive's acceleration will stop as there is not enough energy generation capacity to meet the increasing velocity / kinetic energy needs.

Shawyer has covered this.

Can't get blood from a stone or more kinetic energy conversion than the EM Drive's system components will deliver.
« Last Edit: 05/04/2015 11:28 PM 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

Tags: