Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1875242 times)

Offline Star One

  • Senior Member
  • *****
  • Posts: 8260
  • UK
  • Liked: 1338
  • Likes Given: 168

Has anyone from Eagleworks other than Paul March (NSF user Star-Drive) posted here?

All his posts can be seen here.  His last post was on 2015-04-30, the day after NSF's Evaluating NASAs Futuristic EM Drive feature article was published.

~Kirk

Edit: Typo

I'm not sure. Paul March was the person I had in mind the most when I was posting that. I didn't think to look at his profile for a complete list of posts. I also didn't realize that he posted so little outside of emDrive discussion. I was hoping that he had posted in the 2-3 months that I'd been away from the threads, but I suppose not.

Can check his stats here:
http://forum.nasaspaceflight.com/index.php?action=profile;area=summary;u=2074

Was last loggged on 15 June, so he does read forum posts.

I assuming any further communications will be through formal channels for EW such as at conferences etc?

Offline aero

  • Senior Member
  • *****
  • Posts: 2785
  • 92129
  • Liked: 724
  • Likes Given: 249
With all the fooling I've done with meep recently, I discovered a new trick. I can show different views of slices of a 4-D data set. Here is the closed cavity with the Gaussian drive frequency centered at 2.253 GHz though Meep/Harminv says it resonates at 2.343 GHz. These are all at timestep 1420. What mode is this illustrated.

Shown are big end, center and small end slices "x", and y and z axial slices.
What is the reason for the elliptical shape of the electromagnetic field in the circular cross-section views?



The boundary conditions should be perfectly circular for a truncated cone. 
Therefore one expects a circularly symmetric electromagnetic field instead of this surprising elliptical electromagnetic field.
There appears to be an asymmetry in the circular cross section of your model responsible for this ellipse.
What feature of the model does the major and minor axes of the ellipse align with?

They seem to align with the cut-outs your model has at 0 , 90, 180 and 270 degrees.

What is the asymmetry responsible for the major axis of the ellipse aligning itself with the cut-outs at 0 and 180 degrees instead of 90 and 270 degrees ?

Or, in other words, why should there be an ellipse instead of a rhombus-like shape with symmetry every 90 degrees (the angle between the cut-outs) instead of every 180 degrees?

The cut-outs are an artifact of the display software, they don't exist in the model.

As for the elliptical shape, probably due to the antenna being a line, like this.
Retired, working interesting problems

Offline rfcavity

  • Member
  • Posts: 37
  • Liked: 46
  • Likes Given: 0
.....
I calculated the natural frequency of the EM Drive based on the formula for a perfect cylinder ( https://en.wikipedia.org/wiki/Microwave_cavity#Cylindrical_cavity ) (using the cylindrical Bessel functions that you use in your spreadsheet), based on the Mean diameter (the average of the small and big diameters of the Baby EM Drive),  I obtained the following frequency for TE013 for perfectly flat ends:

24.41 GHz

which compares to the value I obtained using my exact solution:

24.34 GHz
......

With respect that is not now effective guide wavelength is calculated. Roger Shawyer told me how to do it and I shared this here several times. It is in my SS.

The effective guide wavelength is not based on the average of the 2 end plate diameter. It is the numerically integrated value of 10,000 diameters (well I use 10,000, could be more, could be less) including and in between the plates.

For the Baby EMD the guide wavelength for the mean/average diameter is = 0.016501. The numerically integrated effective guide wavelength is = 0.018224, which is why your resonance is too high as your effective guide wavelength is too small.

Have attached the latest version of the SS, which has a new feature.

Let's agree to disagree on this point otherwise this is going to run for ever.  You calculate the effective guide wavelength and the cut-off based on Bessel cylinder functions that do not satisfy the boundary conditions of the problem.  It is an ad-hoc solution that you like because it agrees with Shawyer's formulation .

My point was that there is no justification to include so many digits of numerical precision based on an ad-hoc formula based on cylinder functions and geometrical dimensions (runout, concentricity) of unknown precision, and not performing a spectrum solution, as the participation of other modes in the response is not taken into account (the response of any system will not be the response of a single mode, but will include other mode shapes as well).

It is not what I like or not. It is now Roger Shawyer instructed me to calculate the effective guide wavelength.

It is your method which does not agree with how Roger Shawyer and SPR do the effective guide wavelength calculation.

So if Roger said 5+5=9 you would believe him? He's not setting up the problem correctly. It's plain wrong.

Offline D_Dom

  • Global Moderator
  • Full Member
  • *****
  • Posts: 458
  • Liked: 157
  • Likes Given: 106
As for the elliptical shape, probably due to the antenna being a line, like this.

Is this antenna shape correct? When I first read the @Rodal post showing those elliptical lines of force I thought immediately of antenna.
Space is not merely a matter of life or death, it is considerably more important than that!

Offline aero

  • Senior Member
  • *****
  • Posts: 2785
  • 92129
  • Liked: 724
  • Likes Given: 249
As for the elliptical shape, probably due to the antenna being a line, like this.

Is this antenna shape correct? When I first read the @Rodal post showing those elliptical lines of force I thought immediately of antenna.

No. EW uses a loop antenna as best I know. I don't yet know how to model a loop antenna so I use a line source placed for Meep to calculate the highest resonance. Of course antenna length, location and direction make a difference.
Retired, working interesting problems

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
The 24 GHz Baby EmDrive


Tough to get the moment of inertia. Any suggestions?
Were the platform disk-shaped, it'd be easier.  Put the axis horizontal and wind a string around the disk a few times with a weight attached. Attach a shaft encoder to the axle and log the position against time as the string runs out when the weight is released and falls. Then use a little AlgebraTM and presto! you have the moment of inertia.

Has anyone else noticed that the cavity axis is not aligned exactly tangentially?
« Last Edit: 06/15/2015 04:22 PM by deltaMass »

Offline rmem

  • Member
  • Posts: 2
  • Liked: 1
  • Likes Given: 0
Hello,

I'm new to this forum and want to add to the discussion:
Thrust can be achieved by various error sources interacting with the environment as many out here already pointed out. eagleworks and many others used only small input power therefore i propose a test theory which is built on a pressure gradient caused by thermal effects.
Explanation:
The measured forces are tiny but divided by the projected area in thrust direction it becomes an even smaller pressure needed for thrust.

Example for eagleworks tests:
Area: 0.0613116 m
Force~0.05mN
=> Pressure needed=0.815mPa =0.00000008% of ambient pressure.

Note that this is such a tiny pressure change that it could still be produced in near vacuum. My theory now is that this tiny pressure difference is caused by uneven heating in near wall regions (p=RTrho). Other reasons could be vibrations and magnetic fields. This effect should fairly quickly reach a constant thrust. If this theory holds up we should see a correlation between mode shape and thrust. The node shape dictates where heating occurs. All needed to test this is therefore to integrate heat production over the surfaces and add these up with respect to the orientation of said surface since the pressure gradient should be linear to this. It would probably suffice to use the B field on the boundary as an estimate for the heat production. My prediction is that certain frequencies will produce a heating pattern that is  more uneven and hence produces more thrust. With all the reference geometries we have we might see a correlation to the data. If the math has already been done i'd like to apologize.
« Last Edit: 06/15/2015 04:51 PM by rmem »

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Hello,

I'm new to this forum and want to add to the discussion:
Thrust can be achieved by various error sources interacting with the environment as many out here already pointed out. eagleworks and many others used only small input power therefore i propose a test theory which is built on a pressure gradient caused by thermal effects.
Explanation:
The measured forces are tiny but divided by the projected area in thrust direction it becomes an even smaller pressure needed for thrust.

Example for eagleworks tests:
Area: 0.0613116 m
Force~0.05mN
=> Pressure needed=0.815mPa =0.00000008% of ambient pressure.

Note that these is such a tiny pressure change that it could still be produced in near vacuum. My theory now is that this tiny pressure difference is caused by uneven heating in near wall regions (p=RT/rho). Other reasons could be vibrations and magnetic fields. This effect should fairly quickly reach a constant thrust. If this theory holds up we should see a correlation between mode shape and thrust. The node shape dictates where heating occurs. All needed to test this is therefore to integrate heat production over the surfaces and add these up with respect to the orientation of said surface since the pressure gradient should be linear to this. It would probably suffice to use the B field on the boundary as an estimate for the heat production. My prediction is that certain frequencies will produce a heating pattern that is  more uneven and hence produces more thrust. With all the reference geometries we have we might see a correlation to the data. If the math has already been done i'd like to apologize.
Can you describe in more detail how the B-field will help to determine heat production?
Also, I believe it should be p= R T rho, because p V = n R T

Offline rmem

  • Member
  • Posts: 2
  • Liked: 1
  • Likes Given: 0
@deltaMass yes ofcourse.

Greg Egan showed this correlation near the end of the page
http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html

"For the TM modes, the rate of heat production per unit area is not proportional to any quantity weve previously considered, so we compute it directly from the magnetic field strength:"


The Nasa thermal images also showed a correlation to mode shape as one would expect. If the heating power is small enough air movements on the outside can be neglected and only a pressure change remains which will cause a thrust. I was therefore wondering if this correlates with the observed thrust magnitudes.
« Last Edit: 06/15/2015 04:52 PM by rmem »

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I suggest starting with the EagleWorks data, both in air and in vacuum. Somewhere the vacuum pressure has been recorded. Probably on the wiki page. This should be enough information to work backwards and determine what temperature would suffice to produce a given amount of thrust due to a pressure differential.

I believe that Paul March has also posted temperature plots of the cavity faces.

Offline Abyss

  • Member
  • Posts: 11
  • Houston, TX
  • Liked: 3
  • Likes Given: 3
Actually the cross-correlation looks interesting, there is a relatively clear max (magnitude) centered around 0. Interestingly this absolute max is found at a lag of 3 frames (is it 3 minutes ?).

I've been thinking more about this, we expect the em drive to produce thrust instantly, if there is a delay that means the thrust producing mechanism is most likely related to heat or another means. 

Because of this lag I'm leaning towards the thrust not being from a novel mechanism.  We'll need more data to be sure.

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
As for the elliptical shape, probably due to the antenna being a line, like this.

Is this antenna shape correct? When I first read the @Rodal post showing those elliptical lines of force I thought immediately of antenna.

No. EW uses a loop antenna as best I know. I don't yet know how to model a loop antenna so I use a line source placed for Meep to calculate the highest resonance. Of course antenna length, location and direction make a difference.

Yes, the loop EW seemed to use was a quarter-wavelength loop, horizontal polarization w/the top and bottom frustum plates. This is different from Shawyer, whom I believe used a monopole design at one time, possibly aligned with frustum side walls. Decent loop page here: http://www.robkalmeijer.nl/techniek/electronica/radiotechniek/hambladen/qst/1986/06/page33/index.html

Julian's was a monopole perpendicular to frustum sidewalls, so no one seems to have settled on one particular style. Not sure about our Aachen friends, but might be the same as Julian's.

If I seen any effects w/my test, the radiator (antenna) is the first thing I will vary. Starting out w/a Julian style monopole, not an EW loop on the small plate.

Offline PaulF

  • Member
  • Posts: 52
  • Netherlands
  • Liked: 19
  • Likes Given: 9
I just had another thought experiment but I need a couple of solid answers before I can draw conclusions. I am hoping some of you could fill in on this. No weirdo untestable stuff this time.

If I understand correctly, there is massive energy flowing through the "nodes" of the standing waves. Now let's say, we could focus those nodes to maybe the diameter of a proton. We then pump one KiloWatt of power into the cavity. The energy density at the nodes would become high.

What I don't know is, could the energy density in the nodes become so high that a singularity could form? And in this case because the mass is the result of energy flux (like a lightbulb, switch off power and instantaneously there's darkness, in case of the nodes they fade extremely quickly in energy density), would it be a an actual virtual singularity suitable as a tabletop black hole for experimenting on?

Anybody with a view on that?
« Last Edit: 06/15/2015 06:32 PM by PaulF »

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I just had another thought experiment but I need a couple of solid answers before I can draw conclusions. I am hoping some of you could fill in on this. No weirdo untestable stuff this time.

If I understand correctly, there is massive energy flowing through the "nodes" of the standing waves. Now let's say, we could focus those nodes to maybe the diameter of a proton. We then pump one KiloWatt of power into the cavity. The energy density at the nodes would become high.

What I don't know is, could the energy density in the nodes become so high that a singularity could form? And in this case because the mass is the result of energy flux (like a lightbulb, switch off power and instantaneously there's darkness, in case of the nodes they fade extremely quickly in energy density), would it be a an actual virtual singularity suitable as a tabletop black hole for experimenting on?
No.

Offline PaulF

  • Member
  • Posts: 52
  • Netherlands
  • Liked: 19
  • Likes Given: 9
I just had another thought experiment but I need a couple of solid answers before I can draw conclusions. I am hoping some of you could fill in on this. No weirdo untestable stuff this time.

If I understand correctly, there is massive energy flowing through the "nodes" of the standing waves. Now let's say, we could focus those nodes to maybe the diameter of a proton. We then pump one KiloWatt of power into the cavity. The energy density at the nodes would become high.

What I don't know is, could the energy density in the nodes become so high that a singularity could form? And in this case because the mass is the result of energy flux (like a lightbulb, switch off power and instantaneously there's darkness, in case of the nodes they fade extremely quickly in energy density), would it be a an actual virtual singularity suitable as a tabletop black hole for experimenting on?
No.
Short 'n sweet. Thank you :)

Offline aero

  • Senior Member
  • *****
  • Posts: 2785
  • 92129
  • Liked: 724
  • Likes Given: 249
As for the elliptical shape, probably due to the antenna being a line, like this.

Is this antenna shape correct? When I first read the @Rodal post showing those elliptical lines of force I thought immediately of antenna.

No. EW uses a loop antenna as best I know. I don't yet know how to model a loop antenna so I use a line source placed for Meep to calculate the highest resonance. Of course antenna length, location and direction make a difference.

Yes, the loop EW seemed to use was a quarter-wavelength loop, horizontal polarization w/the top and bottom frustum plates. This is different from Shawyer, whom I believe used a monopole design at one time, possibly aligned with frustum side walls. Decent loop page here: http://www.robkalmeijer.nl/techniek/electronica/radiotechniek/hambladen/qst/1986/06/page33/index.html

Julian's was a monopole perpendicular to frustum sidewalls, so no one seems to have settled on one particular style. Not sure about our Aachen friends, but might be the same as Julian's.

If I seen any effects w/my test, the radiator (antenna) is the first thing I will vary. Starting out w/a Julian style monopole, not an EW loop on the small plate.

Right now, I'm moving the dipole antenna around and exciting it with Ez and Hy sources. If you have a configuration you'd like to see, let me know what it is. I have not modeled your cavity but I could take the time to do that if you can wait a few days to see the field patterns.

In order to model any antenna other than a point source or a dipole, I'd need to write an antenna model in Scheme which I'm not up to. That's more of a task than I want to bite off.
Retired, working interesting problems

Online WarpTech

  • Full Member
  • ****
  • Posts: 1307
  • Do it!
  • Vista, CA
  • Liked: 1351
  • Likes Given: 1815
Actually the cross-correlation looks interesting, there is a relatively clear max (magnitude) centered around 0. Interestingly this absolute max is found at a lag of 3 frames (is it 3 minutes ?).

I've been thinking more about this, we expect the em drive to produce thrust instantly, if there is a delay that means the thrust producing mechanism is most likely related to heat or another means. 

Because of this lag I'm leaning towards the thrust not being from a novel mechanism.  We'll need more data to be sure.

Don't forget, there was a time delay in previous experiments too, at much higher power levels. If it requires a certain stored-energy threshold before any thrust can be achieved, these guys are playing with only mW where others were using Watts. It could take this Baby EM Drive 1000X longer to ramp up to speed, so to speak.
Todd


Offline deuteragenie

  • Member
  • Posts: 71
  • Germany
  • Liked: 22
  • Likes Given: 0
Actually the cross-correlation looks interesting, there is a relatively clear max (magnitude) centered around 0. Interestingly this absolute max is found at a lag of 3 frames (is it 3 minutes ?).

I've been thinking more about this, we expect the em drive to produce thrust instantly, if there is a delay that means the thrust producing mechanism is most likely related to heat or another means. 

Because of this lag I'm leaning towards the thrust not being from a novel mechanism.  We'll need more data to be sure.

Don't forget, there was a time delay in previous experiments too, at much higher power levels. If it requires a certain stored-energy threshold before any thrust can be achieved, these guys are playing with only mW where others were using Watts. It could take this Baby EM Drive 1000X longer to ramp up to speed, so to speak.
Todd

Even though there is a lot of noise in the data, it would be interesting if the Baby EM drive experimentors could generate a longer time series with much more on / off cycles, preferrably of the same duration.  This would provide a better basis for the TS analysis.  From what I have seen, there are some hints that something is going on, but it is inconclusive.
« Last Edit: 06/15/2015 09:19 PM by deuteragenie »

Offline flux_capacitor

  • Full Member
  • ****
  • Posts: 563
  • France
  • Liked: 679
  • Likes Given: 927
The Power input is the rate of change in work "W". It results in doing work on the field via the Poynting vector, and also -J*E, which is the current density flowing in the frustum conductors. So look at E and look at J and see where it can be maximized.

That Poynting vector is analogous to the Lorentz force when electric charges are able to freely flow within the electric field, right?

In more conventional (although leading edge and still in R&D) magnetohydrodynamic (MHD) thrusters using a plasma as a working fluid (magnetoplasmadynamic thrusters) the flow of charged particles q along the electric field E produced by a voltage between electrodes gives rise to an electric current density J which, combined to the induced magnetic field B, creates Lorentz forces which accelerate the plasma in one direction (the output nozzle). If a magnetic coil wraps that kind of MPD thruster, the ambient B-field is no longer mainly induced by the flow of charged particles, but is externally applied to the plasma. This is known as an "applied-field" (VS "self-field") MPD thruster. The purpose of an applied field is to increase the magnetic field, hence the Lorentz force, hence the thrust.


A self-field MPD thruster.
In dark purple: the electric field lines. In red: the induced toroidal magnetic field.
The magnetic field is stronger where J is more focused, i.e. where the E-field lines are gathered, on the axis.
Arrow: direction of the Lorentz force.


An EmDrive is also a "self-field" thruster. It induces its own E-field and B-field within, which combine together.

Following this idea, could we imagine an "applied-field EmDrive" with an external belt coil that would produce a more powerful magnetic field inside the cavity, tuned to combine with the E-field (axial B-field in the case of TE modes, and toroidal B-field for TM modes) in order to enhance the Poynting vector in a preferred direction? Or is my idea coming from magnetohydrodynamics a complete nonsense when applied to standing waves in a resonant cavity?
« Last Edit: 06/15/2015 09:28 PM by flux_capacitor »

Offline X_RaY

  • Full Member
  • ****
  • Posts: 778
  • Germany
  • Liked: 1001
  • Likes Given: 2196
.....
I calculated the natural frequency of the EM Drive based on the formula for a perfect cylinder ( https://en.wikipedia.org/wiki/Microwave_cavity#Cylindrical_cavity ) (using the cylindrical Bessel functions that you use in your spreadsheet), based on the Mean diameter (the average of the small and big diameters of the Baby EM Drive),  I obtained the following frequency for TE013 for perfectly flat ends:

24.41 GHz

which compares to the value I obtained using my exact solution:

24.34 GHz
......

With respect that is not now effective guide wavelength is calculated. Roger Shawyer told me how to do it and I shared this here several times. It is in my SS.

The effective guide wavelength is not based on the average of the 2 end plate diameter. It is the numerically integrated value of 10,000 diameters (well I use 10,000, could be more, could be less) including and in between the plates.

For the Baby EMD the guide wavelength for the mean/average diameter is = 0.016501. The numerically integrated effective guide wavelength is = 0.018224, which is why your resonance is too high as your effective guide wavelength is too small.

Have attached the latest version of the SS, which has a new feature.

What is that mean? The length between the two end plates is 0.02437m. Mode index p=3 means 3 notes of the field(2 inside the cavity and 1! for the boundary conditions at the metallic end plates). The guide wavelength is always larger inside some waveguide such as a cavity than in free space. The effective wavelength have to be larger as the ~0.0125m. Based on your relation it is :) . But is this a average over the 3 half wavelengths? Is that what you want to tell with this number(0.018224m)? The wavelength is dependent on the actual diameter and its not linear, its quadratic with the diameter over the length { (2r/l) }. So is it deceptive to talk about an effective wavelength inside the cone with p=3?
Or does it mean the relation between the freespace wavelength and that inside in cavity at whole?
An effective wavelength (m/wavelength) is only direct usefull if you have a mode with index p=1 (one half wavelength). All other is still average. ;)
« Last Edit: 06/15/2015 10:03 PM by X_RaY »

Tags: