The dielectric is clearly visible as a small flat douhgnut (a disk with a central hole) in these pictures of the Electric Field for their future truncated cone. The NASA researchers think that it is best located at the small flat surface.
The magnetic field in the truncated cone is in blue Observe that the magnetic field is directed along the axis of revolution of the truncated cone, while the electric field is in red and it circulates along two main cells of different rotational sign, clockwise and counterclockwise:
Quote from: Rodal on 10/09/2014 10:02 pmThe magnetic field in the truncated cone is in blue Observe that the magnetic field is directed along the axis of revolution of the truncated cone, while the electric field is in red and it circulates along two main cells of different rotational sign, clockwise and counterclockwise:Yep you got it. There's your differences in angular and linear momentum too.
Quote from: Mulletron on 10/09/2014 10:06 pmQuote from: Rodal on 10/09/2014 10:02 pmThe magnetic field in the truncated cone is in blue Observe that the magnetic field is directed along the axis of revolution of the truncated cone, while the electric field is in red and it circulates along two main cells of different rotational sign, clockwise and counterclockwise:Yep you got it. There's your differences in angular and linear momentum too.Observe that the magnetic field arrows point away from both surfaces, towards the center of the truncated cone.
Quote from: Rodal on 10/09/2014 10:11 pmQuote from: Mulletron on 10/09/2014 10:06 pmQuote from: Rodal on 10/09/2014 10:02 pmThe magnetic field in the truncated cone is in blue Observe that the magnetic field is directed along the axis of revolution of the truncated cone, while the electric field is in red and it circulates along two main cells of different rotational sign, clockwise and counterclockwise:Yep you got it. There's your differences in angular and linear momentum too.Observe that the magnetic field arrows point away from both surfaces, towards the center of the truncated cone.They are counter rotation circles in the vertical axis, meeting in the center. The magnetic flux flows like water down a drain, because the whole rf field has a rotation.
This is important:Please fact check me on this but of the three TM modes on this http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html page; the one closest resembling TM211 would be this one right? At least good enough for a rough model? I think so. Not exactly but close enough because what I'm really interested in are the position of transverse H with respect to Z. The view is of the Z axis running vertically so the blue magnetic X, and Y are just dots and crosses. I can see how highly frequency dependent and sensitive this mode is, which accounts for the big difference in Q and thrust with just a small change in frequency. Given the small size of the dielectric slug in the vertical axis at the small end, the resonant mode would very easily dip into and out of the dielectric with very small freq changes. Hence the loss of thrust. A COSMOL plot is needed for 1932.6 and 1936.7 to see this. I need to see if the magnetic field lines were in the dielectric more or less with both freqs.At 1932.6, Q was down but thrust was up. Very telling indeed. If I'm right, then eureka!
This is important:Please fact check me on this but of the three TM modes on this http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html page; the one closest resembling TM211 would be this one right? At least good enough for a rough model? I think so. Not exactly but close enough because what I'm really interested in are the position of transverse H with respect to Z. The view is of the Z axis running vertically so the blue magnetic X, and Y are just dots and crosses.
Quote from: Mulletron on 10/09/2014 11:27 pmThis is important:Please fact check me on this but of the three TM modes on this http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html page; the one closest resembling TM211 would be this one right? At least good enough for a rough model? I think so. Not exactly but close enough because what I'm really interested in are the position of transverse H with respect to Z. The view is of the Z axis running vertically so the blue magnetic X, and Y are just dots and crosses. I can see how highly frequency dependent and sensitive this mode is, which accounts for the big difference in Q and thrust with just a small change in frequency. Given the small size of the dielectric slug in the vertical axis at the small end, the resonant mode would very easily dip into and out of the dielectric with very small freq changes. Hence the loss of thrust. A COSMOL plot is needed for 1932.6 and 1936.7 to see this. I need to see if the magnetic field lines were in the dielectric more or less with both freqs.At 1932.6, Q was down but thrust was up. Very telling indeed. If I'm right, then eureka!Answer: none of the modes calculated in the closed-form solution by Greg Egan here: http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html agree with the modes calculated by COMSOL Finite Element analysis as presented by Brady et.al. (shown with red arrows in the attachment below).The first Transverse Electric mode shown by Egan has a single cell along the cone's axis of revolution.The mode calculated by COMSOL Brady et.al. has two cells along the cone's axis of revolution.The second Transverse Electric mode shown by Egan has two cells, but they are distributed in a completely different fashion: the smaller cell is closest to the large transverse surface while the mode calculated by COMSOL shows the smaller cell closest to the smaller transverse surface.It is not clear why this is so. It could be because of the boundary conditions. I have a strong suspicion that the reason is due to the cylinder inside the truncated cone shown in pink red here: When I have time I may model it with Mathematica and see what's going on but I don't have free time to do that during the next couple of weeks.
Yeah Egan's modes are being used as generic representations. I only care about the last number here. Do you see how increasing and decreasing the frequency fills the void of the cavity more or less? That's what I'm getting at. The cells get smaller with higher frequency and change shape. You can't see it exactly without a plot.
Quote from: Rodal on 10/09/2014 11:43 pmQuote from: Mulletron on 10/09/2014 11:27 pmThis is important:Please fact check me on this but of the three TM modes on this http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html page; the one closest resembling TM211 would be this one right? At least good enough for a rough model? I think so. Not exactly but close enough because what I'm really interested in are the position of transverse H with respect to Z. The view is of the Z axis running vertically so the blue magnetic X, and Y are just dots and crosses. I can see how highly frequency dependent and sensitive this mode is, which accounts for the big difference in Q and thrust with just a small change in frequency. Given the small size of the dielectric slug in the vertical axis at the small end, the resonant mode would very easily dip into and out of the dielectric with very small freq changes. Hence the loss of thrust. A COSMOL plot is needed for 1932.6 and 1936.7 to see this. I need to see if the magnetic field lines were in the dielectric more or less with both freqs.At 1932.6, Q was down but thrust was up. Very telling indeed. If I'm right, then eureka!Answer: none of the modes calculated in the closed-form solution by Greg Egan here: http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html agree with the modes calculated by COMSOL Finite Element analysis as presented by Brady et.al. (shown with red arrows in the attachment below).The first Transverse Electric mode shown by Egan has a single cell along the cone's axis of revolution.The mode calculated by COMSOL Brady et.al. has two cells along the cone's axis of revolution.The second Transverse Electric mode shown by Egan has two cells, but they are distributed in a completely different fashion: the smaller cell is closest to the large transverse surface while the mode calculated by COMSOL shows the smaller cell closest to the smaller transverse surface.It is not clear why this is so. It could be because of the boundary conditions. I have a strong suspicion that the reason is due to the cylinder inside the truncated cone shown in pink red here: When I have time I may model it with Mathematica and see what's going on but I don't have free time to do that during the next couple of weeks. If the smll end is filled w/ dielectric that would make the difference, the optical length is greater for the same physical length
It is incorrect to think that it is a truncated cone on the inside. Everything shows that it is a cylinder on one end joined to a truncated cone on the other end. The mode shapes for such a geometric body are different than the mode shapes for a truncated cone as modeled by Egan.And I don't think that the cylinder is there by accident. Somebody thought this through.