Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1799424 times)

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2588

Just got this looks sweet.

Offline dustinthewind

  • Full Member
  • ****
  • Posts: 584
  • U.S. of A.
  • Liked: 226
  • Likes Given: 256


My suspicion is that one of these would provide more propulsion than the other.

I have got a question. Are these cavities of a pure dielectrically type (non metallic plates between them)? If not there may be a problem with that idea: If there are metallic plates the penetration depth is only a few m..
Was your idea to transform the SHORT at the end plate to be an OPEN a quarter wavelength away? I think there is a physically short at the end plate of the second cavity because there is a metallic plate also(boundary conditions). The currents are not at the outside of a metallic cavity resonator(for this high frequencies).

the picture of the 2 cavities is of only space and metal.  The reason I think radiation would tunnel is because of interaction of evanescent currents should attenuate current in one cavity and amplify current in the other to bring them both to some happy medium where their currents want to circulate in unison, however, the radiation fights this.  Because the currents are now no longer circulating only in response to the radiation and are out of phase with the radiation in the cavities.  The cavities can no longer reflect the radiation as effectively.  As a result some radiation should tunnel from one cavity to the next but now the radiation I think as it tunnels is also out of phase with radiation inside the new cavity.  The radiation from the 1st cavity now in the 2nd cavity traveling and still out of phase may encounter currents in the back plate of the 2nd cavity out of phase with it.  As a result a portion of its radiation may escape that cavity altogether. 

This was my line of thinking but maybe it's flawed or I am not understanding your question. 

An alternative to this was with a single cavity and two antennas where I was hoping to open up the cavity to transmit radiation (Radio frequency spectrum) that did use dielectrics to slow light but also used metal.  Attached the image. 
« Last Edit: 07/19/2015 07:58 PM by dustinthewind »

Offline X_RaY

  • Full Member
  • ****
  • Posts: 774
  • Germany
  • Liked: 993
  • Likes Given: 2144


My suspicion is that one of these would provide more propulsion than the other.

I have got a question. Are these cavities of a pure dielectrically type (non metallic plates between them)? If not there may be a problem with that idea: If there are metallic plates the penetration depth is only a few m..
Was your idea to transform the SHORT at the end plate to be an OPEN a quarter wavelength away? I think there is a physically short at the end plate of the second cavity because there is a metallic plate also(boundary conditions). The currents are not at the outside of a metallic cavity resonator(for this high frequencies).

the picture of the 2 cavities is of only space and metal.  The reason I think radiation would tunnel is because of interaction of evanescent currents should attenuate current in one cavity and amplify current in the other to bring them both to some happy medium where their currents want to circulate in unison, however, the radiation fights this.  Because the currents are now no longer circulating only in response to the radiation and are out of phase with the radiation in the cavities.  The cavities can no longer reflect the radiation as effectively.  As a result some radiation should tunnel from one cavity to the next but now the radiation I think as it tunnels is also out of phase with radiation inside the new cavity.  The radiation from the 1st cavity now in the 2nd cavity traveling and still out of phase may encounter currents in the back plate of the 2nd cavity out of phase with it.  As a result a portion of its radiation may escape that cavity altogether. 

This was my line of thinking but maybe it's flawed or I am not understanding your question. 

An alternative to this was with a single cavity and two antennas where I was hoping to open up the cavity to transmit radiation (Radio frequency spectrum) that did use dielectrics to slow light but also used metal.  Attached the image.
OK if there are 2 antennas, one inside, one outside the cavity out of phase. Let us think there may be some photons entering the metallic cavity from the outer dielectric (very thin metal film between cavity and dielectric ~1m) the resonance inside the metal cavity will be degenerate by modification of the second stimulation of some cavity eigenvalue. Its just wave-mixing. If there are in phase at a position there will be constructive interference..

I think if the metal in the real is thicker this will not be the case. There is only the effect of a photon rocket based of the antenna outside in the dielectrica...

edit:
http://arxiv.org/pdf/physics/0311061v7.pdf

There is a diagram of the group velocity at cutoff diameter... IMHO thats for examplethe why evanecent waves are in dicussion for the conical cavity :)

Sorry, the magnetic antenna you've draw doesn't work. The wave will be splitted into two parts and propagate along the wire, at half the way in the loop they will interact to form a maximum (that works like an electrically dipole). A magnetic antenna loop have to be in contact to the opposite potential at the end of the loop, whatever capacitive or galvanic...
« Last Edit: 07/19/2015 09:07 PM by X_RaY »

Offline tidux

  • Member
  • Posts: 11
  • Sol III
  • Liked: 6
  • Likes Given: 3
You guys are in luck, as NSF-1701 is easily reconfigurable. Here's what I propose. Assymetrical placement of monopole in both the small base and big base. This will not be frustum side insertion, but parallel to axis, offset from centerline for assymetry. My reccomendation is 1/4 wavelength from frustum sidewall to realize 50 ohm match. It will be more like nasa insertion but not a perpendicular coupling loop. For now, polarity is parallel to frustum length axis...the easiest way to swap ends for insertion tests. I'll start at a large diameter insertion for static temp testing of magnetron core. Now, off to get some more solder...IR thermometer arrives mid week, will video static test. Max core temp is 160C. Exceeding that typifies poor impedance match according to my research. If matched properly, power-up can exceed 5 minutes.

So for a meep run of that, would the change be from (set! antSIx (- (/ (* high .0254) 2) (/ (/ csi fsi) 4))) to (set! antSIx (* 0.25 wl_meep) or (set! antSIx 0) ?

Offline Prunesquallor

  • Full Member
  • *
  • Posts: 174
  • Currently, TeV Brane Resident
  • Liked: 157
  • Likes Given: 73
Even with very modest k (say 10-6 N/W) one can achieve excellent mission performance when lots of power is available. Let's go to Pluto (40 AU, 100 Kg). With 1 MW power it takes 1.1 years and maximum speed is 0.1%c.

Uhh, and how do you fit a 1 MW power plant into 100 kg?
Retired, yet... not

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2588
Even with very modest k (say 10-6 N/W) one can achieve excellent mission performance when lots of power is available. Let's go to Pluto (40 AU, 100 Kg). With 1 MW power it takes 1.1 years and maximum speed is 0.1%c.

Uhh, and how do you fit a 1 MW power plant into 100 kg?
Nukler Fishin?

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Even with very modest k (say 10-6 N/W) one can achieve excellent mission performance when lots of power is available. Let's go to Pluto (40 AU, 100 Kg). With 1 MW power it takes 1.1 years and maximum speed is 0.1%c.

Uhh, and how do you fit a 1 MW power plant into 100 kg?
That took long enough  ;)

Well, of course using an overunity EmDrive power generator! You can have as much power as you like  8)
« Last Edit: 07/19/2015 09:15 PM by deltaMass »

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2326
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2956
  • Likes Given: 2588
You guys are in luck, as NSF-1701 is easily reconfigurable. Here's what I propose. Assymetrical placement of monopole in both the small base and big base. This will not be frustum side insertion, but parallel to axis, offset from centerline for assymetry. My reccomendation is 1/4 wavelength from frustum sidewall to realize 50 ohm match. It will be more like nasa insertion but not a perpendicular coupling loop. For now, polarity is parallel to frustum length axis...the easiest way to swap ends for insertion tests. I'll start at a large diameter insertion for static temp testing of magnetron core. Now, off to get some more solder...IR thermometer arrives mid week, will video static test. Max core temp is 160C. Exceeding that typifies poor impedance match according to my research. If matched properly, power-up can exceed 5 minutes.
I'm getting so jazzed to see you getting close to testing! Can't wait!

Shell

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Or per this
https://en.wikipedia.org/wiki/Power-to-weight_ratio
we need > 100 MW/100 Kg or > 10,000 W/Kg.
SAFT Li-ion batteries do that.

Online Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5914
  • Likes Given: 5253
I would like to correct some statements that have appeared in another forum, concerning the Meep/Wolfram Mathematica calculations:

1) << If you get the Poynting vector, in the end you can calculate backwards force from areas and pressure. >>

I have not at all calculated the force from the Poynting vector !. 

To calculate the force from the Poynting vector would be very imprecise as it would involve time differentiation, which is always to be avoided in numerical solutions. I have calculated the force from integration of the stress tensor, as done in fluid mechanics problems, for example (instead of calculating the force based on the power flux), which is a much better numerically conditioned problem. The stress tensor has nothing to do with the Poynting vector.  The stress tensor is a 2nd order tensor, not a vector.  It is a completely different quantity familiar to anybody that has dealt with fluid or solid mechanics.



2) << there's a problem: it appears to not be stable in time >> 

The solution grows with time because what has been computed is a very early state of the transient solution  (only 0.013 microseconds !!!).  This growth has nothing to do with an inherent instability.  It is typical of transient solutions to grow with time (not just in electromagnetics, but in fluid mechanics, heat transfer and solid mechanics).  It makes eminent sense that steady-state should be preceded by a transient solution growing with time from zero initial conditions.  This has nothing to do with instability.  What would be un-physical would be to reach steady state instantaneously.  :)

(and 0.013 microseconds is a very short period of time, the last time I checked :)  )



3) << MEEP is very picky about the boundary conditions so if you change the model slightly you can get very different transient results>>

Meep is just another Finite Difference code,  practically the oldest numerical technique to solve partial differential equations (it goes back to at least a century ago). It uses the central-difference explicit operator (one of the oldest known FD operators). There is nothing particular about Meep "being picky about boundary conditions" vs. any other finite difference code.  Concerning the material model, a Meep default of perfect metal was compared to a copper model using Drude's equation, with parameters suggested by deltaMass based on careful analysis.  No significant difference was detected in the electromagnetic field results.  The statement that <<if you change the model slightly you can get very different transient results>> has no basis on reality, on the contrary, the opposite was shown to be the case.
« Last Edit: 07/19/2015 10:36 PM by Rodal »

Offline Prunesquallor

  • Full Member
  • *
  • Posts: 174
  • Currently, TeV Brane Resident
  • Liked: 157
  • Likes Given: 73
Even with very modest k (say 10-6 N/W) one can achieve excellent mission performance when lots of power is available. Let's go to Pluto (40 AU, 100 Kg). With 1 MW power it takes 1.1 years and maximum speed is 0.1%c.

Uhh, and how do you fit a 1 MW power plant into 100 kg?
That took long enough  ;)

Well, of course using an overunity EmDrive power generator! You can have as much power as you like  8)

Sorry :-) was off the grid for 1 (one!) week and was trying to catch up on 30 (thirty!) page of posts.

BTW I'm working on a write-up of the acceleration one would need from something like a Cubesat to unambiguously see orbital trajectory effects in the presence of drag. Hope to make it available in the next week or so.
Retired, yet... not

Offline dustinthewind

  • Full Member
  • ****
  • Posts: 584
  • U.S. of A.
  • Liked: 226
  • Likes Given: 256


My suspicion is that one of these would provide more propulsion than the other.

I have got a question. Are these cavities of a pure dielectrically type (non metallic plates between them)? If not there may be a problem with that idea: If there are metallic plates the penetration depth is only a few m..
Was your idea to transform the SHORT at the end plate to be an OPEN a quarter wavelength away? I think there is a physically short at the end plate of the second cavity because there is a metallic plate also(boundary conditions). The currents are not at the outside of a metallic cavity resonator(for this high frequencies).

the picture of the 2 cavities is of only space and metal.  The reason I think radiation would tunnel is because of interaction of evanescent currents should attenuate current in one cavity and amplify current in the other to bring them both to some happy medium where their currents want to circulate in unison, however, the radiation fights this.  Because the currents are now no longer circulating only in response to the radiation and are out of phase with the radiation in the cavities.  The cavities can no longer reflect the radiation as effectively.  As a result some radiation should tunnel from one cavity to the next but now the radiation I think as it tunnels is also out of phase with radiation inside the new cavity.  The radiation from the 1st cavity now in the 2nd cavity traveling and still out of phase may encounter currents in the back plate of the 2nd cavity out of phase with it.  As a result a portion of its radiation may escape that cavity altogether. 

This was my line of thinking but maybe it's flawed or I am not understanding your question. 

An alternative to this was with a single cavity and two antennas where I was hoping to open up the cavity to transmit radiation (Radio frequency spectrum) that did use dielectrics to slow light but also used metal.  Attached the image.
OK if there are 2 antennas, one inside, one outside the cavity out of phase. Let us think there may be some photons entering the metallic cavity from the outer dielectric (very thin metal film between cavity and dielectric ~1m) the resonance inside the metal cavity will be degenerate by modification of the second stimulation of some cavity eigenvalue. Its just wave-mixing. If there are in phase at a position there will be constructive interference..

I think if the metal in the real is thicker this will not be the case. There is only the effect of a photon rocket based of the antenna outside in the dielectrica...

edit:
http://arxiv.org/pdf/physics/0311061v7.pdf

There is a diagram of the group velocity at cutoff diameter... IMHO thats for examplethe why evanecent waves are in dicussion for the conical cavity :)

Sorry, the magnetic antenna you've draw doesn't work. The wave will be splitted into two parts and propagate along the wire, at half the way in the loop they will interact to form a maximum (that works like an electrically dipole). A magnetic antenna loop have to be in contact to the opposite potential at the end of the loop, whatever capacitive or galvanic...

One of the reasons stating that the near field propagates faster than light bothers me is because for a single charge the magnetic field decreases by 1/r^2 as given by the Biot-Savart law.  By induction then so does the resulting electric field.  From the electric field of an accelerating charge can be derived the electric field of light and is done so by Edward Purcell in his book "Electricity and Magnetism" in the appendix.  As a result the magnetic field of the charge is directly related to the light.  Stating that the near field propagates faster than light while the light only propagates at light speed bothers me as a result because either I am not understanding what the near field is or this understanding of light is flawed. 

http://www.amazon.com/Electricity-Magnetism-Edward-M-Purcell/dp/1107014026

I can understand group velocities being super-luminal as the waves them selves are not super-luminal but for the near-field of a single charge to propagate super-luminal is beyond me at the moment.
« Last Edit: 07/19/2015 10:30 PM by dustinthewind »

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Even with very modest k (say 10-6 N/W) one can achieve excellent mission performance when lots of power is available. Let's go to Pluto (40 AU, 100 Kg). With 1 MW power it takes 1.1 years and maximum speed is 0.1%c.

Uhh, and how do you fit a 1 MW power plant into 100 kg?
That took long enough  ;)

Well, of course using an overunity EmDrive power generator! You can have as much power as you like  8)

Sorry :-) was off the grid for 1 (one!) week and was trying to catch up on 30 (thirty!) page of posts.

BTW I'm working on a write-up of the acceleration one would need from something like a Cubesat to unambiguously see orbital trajectory effects in the presence of drag. Hope to make it available in the next week or so.
I have not seen that analysis and would very much welcome it.

There are at least two important issues to bear in mind with that:
1. Determination of position in realtime
2. Determination of perturbation and error sources.

It seems to me that much of such designs immediately simplifies out when two identical craft are flown close by one another. The trade-off for "close" is
a) not so close that e/m or gravitational perturbations become significant between them
b) not so far away that the local environment is significantly changed (gravity, solar wind, solar flux, residual atmosphere,  etc.)

The craft are outfitted with LIDAR and comms to support that data (one may contain a dummy to get back to "identical"). At predetermined times, one is switched on, and ditto off.

I think that's the best that can be done. We've factored out the local environment and are sensing only a difference between a powered and an unpowered module. The question is what sensitivity can be expected?

An alternative to LIDAR occurs to me that has a couple of advantages - some sort of instrumented high tech spring equivalent. This is cheaper and occupies less volume (a CubeSat consideration) than LIDAR, and also ensures that unforeseen drift does not occur between them due to an imperfect twin launch. A slight disadvantage accrues with this spring idea in that there will always be a finite almost-perfectly-undamped oscillation in their separation, since the spring operates in both compression and in expansion. Note that a bungee cord equivalent would not be a good idea because it is unable to provide mutual repulsion.

An even simpler differencing protocol obtains by equipping both craft with an accelerometer and comms to support its data. Probably the best bet. The flight profile would be:
1. both off
2. A on, B off
3. A off, B on (to catch up)
Lather, rinse, repeat.

The main problem is going to be rotation. The thrust vector needs to be in a defined direction - somehow.
« Last Edit: 07/19/2015 11:11 PM by deltaMass »

Online Rodal

  • Senior Member
  • *****
  • Posts: 5835
  • USA
  • Liked: 5914
  • Likes Given: 5253
Obtaining Q simply from the phase slope or phase measurements alone:

Here is a way to get an artificially high Q (pointed out to my by an anonymous friend) that if one is not careful, will give an artificially high Q of hundreds of thousands:

http://www.millertechinc.com/pdf_files/MTI%20TN114%20RLC%20Q%20Phase%20Group%20Delay.htm



The frequency in this group-delay Q-factor equation is supposed to be in radians/sec, not cycles per second or Hz (as it is pointed out later in the note).  One has to divide by 180 degrees/Pi radians = 57.296 degrees/radians conversion to get the correct Q, otherwise the calculation will give a Q about 57 times higher than what it is supposed to be, so, for example if the Q is supposed to be 5,000, this method will give a Q = 286,500 (if one doesn't divide by 180/Pi )

:)
« Last Edit: 07/19/2015 11:18 PM by Rodal »

Offline VAXHeadroom

  • Full Member
  • *
  • Posts: 194
  • Whereever you go, there you are. -- BB
  • Baltimore MD
  • Liked: 255
  • Likes Given: 151

...snip...lots...

Sorry :-) was off the grid for 1 (one!) week and was trying to catch up on 30 (thirty!) page of posts.

BTW I'm working on a write-up of the acceleration one would need from something like a Cubesat to unambiguously see orbital trajectory effects in the presence of drag. Hope to make it available in the next week or so.
I have not seen that analysis and would very much welcome it.

...snip...

There are at least two important issues to bear in mind with that:
1. Determination of position in realtime
2. Determination of perturbation and error sources.

...even more pruning...


It can be done way simpler than that.  A satellite emitting a carrier on an S-Band omnidirectional antenna.  Ground can detect a change in frequency and therefore velocity down to mm/s.  We did that on LCROSS showing that the water evaporating off the Centaur upper stage was pushing it and the shepherding spacecraft around.  I was kinda dumb-founded that they could detect the delta V to that degree...
« Last Edit: 07/19/2015 11:37 PM by VAXHeadroom »
Emory Stagmer
  Executive Producer, Public Speaker UnTied Music - www.untiedmusic.com

Offline deltaMass

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

...snip...lots...

Sorry :-) was off the grid for 1 (one!) week and was trying to catch up on 30 (thirty!) page of posts.

BTW I'm working on a write-up of the acceleration one would need from something like a Cubesat to unambiguously see orbital trajectory effects in the presence of drag. Hope to make it available in the next week or so.
I have not seen that analysis and would very much welcome it.

...snip...

There are at least two important issues to bear in mind with that:
1. Determination of position in realtime
2. Determination of perturbation and error sources.

...even more pruning...


It can be done way simpler than that.  A satellite emitting a carrier on an S-Band omnidirectional antenna.  Ground can detect a change in frequency and therefore velocity down to mm/s.  We did that on LCROSS showing that the water evaporating off the Centaur upper stage was pushing it and the shepherding spacecraft around.  I was kinda dumb-founded that they could detect the delta V to that degree...
There's still the rotation issue. 6 thrusters for stabilisation are going to be required. And then there's noise from the environment that has to be accounted for. But  a pair of such satellites - now we're talking. So long as they are "close"
« Last Edit: 07/20/2015 12:06 AM by deltaMass »

Offline aero

  • Senior Member
  • *****
  • Posts: 2739
  • 92129
  • Liked: 704
  • Likes Given: 236
You guys are in luck, as NSF-1701 is easily reconfigurable. Here's what I propose. Assymetrical placement of monopole in both the small base and big base. This will not be frustum side insertion, but parallel to axis, offset from centerline for assymetry. My reccomendation is 1/4 wavelength from frustum sidewall to realize 50 ohm match. It will be more like nasa insertion but not a perpendicular coupling loop. For now, polarity is parallel to frustum length axis...the easiest way to swap ends for insertion tests. I'll start at a large diameter insertion for static temp testing of magnetron core. Now, off to get some more solder...IR thermometer arrives mid week, will video static test. Max core temp is 160C. Exceeding that typifies poor impedance match according to my research. If matched properly, power-up can exceed 5 minutes.

So for a meep run of that, would the change be from (set! antSIx (- (/ (* high .0254) 2) (/ (/ csi fsi) 4))) to (set! antSIx (* 0.25 wl_meep) or (set! antSIx 0) ?

I'm not sure what he wants to do. I think ... but a little annotated sketch would be most helpful. I think he proposed (and it is his model) to leave the small end antenna x distance 1/4 wavelength from the small end as it is. But move it laterally in the y direction until it is 1/4 wavelength from the inside of the cavity at that radius. Need to check that the radius is greater than 1/4 wavelength at that x location. Then he wants to turn the antenna from the lateral direction as it is now, to the axial direction. Need to check that the antenna is not to long as is, to turn it. That is, using an antenna length of 58 mm,
 is 0.058/2 m < wavelength/4 I think it is Ok. In my most evolved model, there are parameters antlongx, antlongy and antlongz. At the moment, the antenna is in the lateral y direction so antlongy = 0.058 meters. Turning the antenna amounts to setting antlongx = .058 meters, antlongy = 0, and leaving antlonz = 0. If your model does not have those parameters, let me know and I will post a current copy.

Then I get the impression that he proposes a second antenna located diagonally across the cavity center and placed relative to the big end and radius at 1/4 wavelength from the cavity wall in the -y direction.  antlongx, antlongy and antlongz same as above.

Specifically @tidux, the parameter, ant SI x is the x coordinate of the antenna center in SI units. Similarly for parameters antSIy and antSIz. If you don't see those parameters, then it is probably something I have added recently. Let me know.

Here, we are in the model definition section where the parameters have not yet been completely defined, so they must be calculated based on what has been defined and set. This is the correct place to be, but there are a couple of changes/additions needed later, too.

As it is currently, antSIx = (Inside length / 2) - (wave length /4) in SI units, that is, meters from the x=0 center of the cavity. We cannot use wl_meep as it has not yet been defined and if it is defined here then the code flow would be broken, and anyway, you don't want to add or subtract meep units from SI units.  So for your second set statement, use (set! antSIx (* 0.25 (/ csi fsi))). Now you have the antenna center x coordinate at 1/4 wavelength from the center toward the small end and in meters. Your third set statement works as it because it sets the antenna center location x coordinate at the center of the cavity.

 I do not see what x location rfmwguy wants for the antenna, only the y location offset from center. If rfmwguy would be so good as to give us the antenna x, y, z center location in meters or inches from some reference surface (both antennas) modelling it in meep would be straight forward. Then what is left is to convert units to meep units, and add the second antenna source, either to the source-Gaus definition, or define source-Gaus2 using the second antenna and append it to the sources list.
Retired, working interesting problems

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I notice that the Aachen boys have gone quiet.

Offline Prunesquallor

  • Full Member
  • *
  • Posts: 174
  • Currently, TeV Brane Resident
  • Liked: 157
  • Likes Given: 73
Even with very modest k (say 10-6 N/W) one can achieve excellent mission performance when lots of power is available. Let's go to Pluto (40 AU, 100 Kg). With 1 MW power it takes 1.1 years and maximum speed is 0.1%c.

Uhh, and how do you fit a 1 MW power plant into 100 kg?
That took long enough  ;)

Well, of course using an overunity EmDrive power generator! You can have as much power as you like  8)

Sorry :-) was off the grid for 1 (one!) week and was trying to catch up on 30 (thirty!) page of posts.

BTW I'm working on a write-up of the acceleration one would need from something like a Cubesat to unambiguously see orbital trajectory effects in the presence of drag. Hope to make it available in the next week or so.
I have not seen that analysis and would very much welcome it.

There are at least two important issues to bear in mind with that:
1. Determination of position in realtime
2. Determination of perturbation and error sources.

It seems to me that much of such designs immediately simplifies out when two identical craft are flown close by one another. The trade-off for "close" is
a) not so close that e/m or gravitational perturbations become significant between them
b) not so far away that the local environment is significantly changed (gravity, solar wind, solar flux, residual atmosphere,  etc.)

The craft are outfitted with LIDAR and comms to support that data (one may contain a dummy to get back to "identical"). At predetermined times, one is switched on, and ditto off.

I think that's the best that can be done. We've factored out the local environment and are sensing only a difference between a powered and an unpowered module. The question is what sensitivity can be expected?

An alternative to LIDAR occurs to me that has a couple of advantages - some sort of instrumented high tech spring equivalent. This is cheaper and occupies less volume (a CubeSat consideration) than LIDAR, and also ensures that unforeseen drift does not occur between them due to an imperfect twin launch. A slight disadvantage accrues with this spring idea in that there will always be a finite almost-perfectly-undamped oscillation in their separation, since the spring operates in both compression and in expansion. Note that a bungee cord equivalent would not be a good idea because it is unable to provide mutual repulsion.

An even simpler differencing protocol obtains by equipping both craft with an accelerometer and comms to support its data. Probably the best bet. The flight profile would be:
1. both off
2. A on, B off
3. A off, B on (to catch up)
Lather, rinse, repeat.

The main problem is going to be rotation. The thrust vector needs to be in a defined direction - somehow.

Your comments are all excellent and are geared to a flight experiment that would WORK.

The first stage of the analysis I'm doing is to caution against flight experiments that would NOT work (the vast majority I have seen so far).  Essentially, it is trying to establish the acceleration a single CubeSat should attain up and beyond the DEceleration from orbital drag to unambiguously demonstrate trajectory modification (i.e., orbit raising).  It would assume "perfect" ground tracking of the CubeSat and perfect thrust vector control.

As you allude, there are better (and more complex/expensive) ways to do this.  This is meant to be the first step down the path to illustrate this.
Retired, yet... not

Offline aero

  • Senior Member
  • *****
  • Posts: 2739
  • 92129
  • Liked: 704
  • Likes Given: 236


My suspicion is that one of these would provide more propulsion than the other.

I have got a question. Are these cavities of a pure dielectrically type (non metallic plates between them)? If not there may be a problem with that idea: If there are metallic plates the penetration depth is only a few m..
Was your idea to transform the SHORT at the end plate to be an OPEN a quarter wavelength away? I think there is a physically short at the end plate of the second cavity because there is a metallic plate also(boundary conditions). The currents are not at the outside of a metallic cavity resonator(for this high frequencies).

the picture of the 2 cavities is of only space and metal.  The reason I think radiation would tunnel is because of interaction of evanescent currents should attenuate current in one cavity and amplify current in the other to bring them both to some happy medium where their currents want to circulate in unison, however, the radiation fights this.  Because the currents are now no longer circulating only in response to the radiation and are out of phase with the radiation in the cavities.  The cavities can no longer reflect the radiation as effectively.  As a result some radiation should tunnel from one cavity to the next but now the radiation I think as it tunnels is also out of phase with radiation inside the new cavity.  The radiation from the 1st cavity now in the 2nd cavity traveling and still out of phase may encounter currents in the back plate of the 2nd cavity out of phase with it.  As a result a portion of its radiation may escape that cavity altogether. 

This was my line of thinking but maybe it's flawed or I am not understanding your question. 

An alternative to this was with a single cavity and two antennas where I was hoping to open up the cavity to transmit radiation (Radio frequency spectrum) that did use dielectrics to slow light but also used metal.  Attached the image.
OK if there are 2 antennas, one inside, one outside the cavity out of phase. Let us think there may be some photons entering the metallic cavity from the outer dielectric (very thin metal film between cavity and dielectric ~1m) the resonance inside the metal cavity will be degenerate by modification of the second stimulation of some cavity eigenvalue. Its just wave-mixing. If there are in phase at a position there will be constructive interference..

I think if the metal in the real is thicker this will not be the case. There is only the effect of a photon rocket based of the antenna outside in the dielectrica...

edit:
http://arxiv.org/pdf/physics/0311061v7.pdf

There is a diagram of the group velocity at cutoff diameter... IMHO thats for examplethe why evanecent waves are in dicussion for the conical cavity :)

Sorry, the magnetic antenna you've draw doesn't work. The wave will be splitted into two parts and propagate along the wire, at half the way in the loop they will interact to form a maximum (that works like an electrically dipole). A magnetic antenna loop have to be in contact to the opposite potential at the end of the loop, whatever capacitive or galvanic...

One of the reasons stating that the near field propagates faster than light bothers me is because for a single charge the magnetic field decreases by 1/r^2 as given by the Biot-Savart law.  By induction then so does the resulting electric field.  From the electric field of an accelerating charge can be derived the electric field of light and is done so by Edward Purcell in his book "Electricity and Magnetism" in the appendix.  As a result the magnetic field of the charge is directly related to the light.  Stating that the near field propagates faster than light while the light only propagates at light speed bothers me as a result because either I am not understanding what the near field is or this understanding of light is flawed. 

http://www.amazon.com/Electricity-Magnetism-Edward-M-Purcell/dp/1107014026

I can understand group velocities being super-luminal as the waves them selves are not super-luminal but for the near-field of a single charge to propagate super-luminal is beyond me at the moment.

Evanescent waves have been experimentally measured as moving faster than light, there are many papers describing and discussing these experiments. How does it do so? I don't know but I have speculated that they tunnel. They are not propagating waves, so they don't propagate hence I think that releaves them of the duty to move as propagating waves do. And we do know that tunneling is faster than light.
Retired, working interesting problems

Tags: