-
#1900 by SeeShells on 17 Sep, 2015 06:50
-
@SeeShells,
Here are some views of the CE3 model. They are very preliminary but ... well, its as far as I have gotten.
https://drive.google.com/folderview?id=0B1XizxEfB23td2ZDeDFUUFdvLUk&usp=sharing
It resonates but it's not a mode that I recognize.
Specifics-
Cavity is the SeeShell Crazy Eddie 2, upgraded by replacing the loop antenna with wave guides, hence CE3 model.
Drive frequency 2.47 GHz, Harminv calculated resonant frequency 2.47837658E+009 Hz, Q is good: ~ 1 million.
Antenna - dual magnetron wave guides, 180 degrees apart. That is, on opposite sides of the frustum using WR 340 wave guides with bottom edge flush with the big base. Phase is adjusted to place sources 3 wavelengths apart.
Gaussian noise bandwidth at 0.05 * 2.47GHz is is double the expected magnetron noise bandwidth.
This is a debug run at resolution = 200, hence it may appear more granular than the resolution = 250 runs commonly uploaded. There are no csv files, resolution of 200 would adversely affect any data comparison calculations based on the csv files.
Great to see you got the waveguide injection working aero.
It looks like you placed the wave right on the bottom to replace the antennas,
If you look at the mode generation in the attached image...
You'll see close to the middle of the frustum is where the mode wants to generate within the frustum. Did you try to do the waveguides insertion at that point? We should try.
Shell
Move the waveguides up to the point where they are not inserting into the frustum which should be halfway up. See image.
-
#1901 by rfmwguy on 17 Sep, 2015 13:38
-
I might have been part of that ionized conversation. I noticed a metallic "feel" to the air after power testing which someone proposed as copper ionization. It could be Ni, but it seemed "metallic", if I might used that term. This was noticed on all 3 Flight Tests I did in August. FWIW, I know its not exactly a scientific description, just an impression.I found a nice youtube channel, sure that's already known, nevertheless quite interesting to see the pure mechanically conditions in such a simulation
...
some more vids are available
Interesting video...
So all we have to do is pressurize a cone and heat it up, and we should get a net thrust?
Back in thread three this was discussed briefly in the context of ionized particles. Yang stated in one of her papers that the behaviour of ionized particles was important to understanding the behaviour of EM Drives. It has been pointed out numerous times that neither Yang or Shawyer published experiments ran in vacuum.
At one point I called this the "ionic wind tunnel" effect in those discussions. That said, these were general discussions and everyone felt that microwave energy was the primary source for trust...
Now we have some excellent simulations that bring this all back. Note that, say, nitrogen ions are orders of magnitude more massive than photons.
In my view, these simulations are a significant step forward in our understanding of this phenomena. Congratulations!
And once again I am in awe of the power of these forums!
-
#1902 by SeeShells on 17 Sep, 2015 14:21
-
Ni? Why Nickel?
I might have been part of that ionized conversation. I noticed a metallic "feel" to the air after power testing which someone proposed as copper ionization. It could be Ni, but it seemed "metallic", if I might used that term. This was noticed on all 3 Flight Tests I did in August. FWIW, I know its not exactly a scientific description, just an impression.I found a nice youtube channel, sure that's already known, nevertheless quite interesting to see the pure mechanically conditions in such a simulation
...
some more vids are available
Interesting video...
So all we have to do is pressurize a cone and heat it up, and we should get a net thrust?
Back in thread three this was discussed briefly in the context of ionized particles. Yang stated in one of her papers that the behaviour of ionized particles was important to understanding the behaviour of EM Drives. It has been pointed out numerous times that neither Yang or Shawyer published experiments ran in vacuum.
At one point I called this the "ionic wind tunnel" effect in those discussions. That said, these were general discussions and everyone felt that microwave energy was the primary source for trust...
Now we have some excellent simulations that bring this all back. Note that, say, nitrogen ions are orders of magnitude more massive than photons.
In my view, these simulations are a significant step forward in our understanding of this phenomena. Congratulations!
And once again I am in awe of the power of these forums!
Shell -
#1903 by rfmwguy on 17 Sep, 2015 14:23
-
Ooooops, N not Ni
Ni? Why Nickel?
I might have been part of that ionized conversation. I noticed a metallic "feel" to the air after power testing which someone proposed as copper ionization. It could be Ni, but it seemed "metallic", if I might used that term. This was noticed on all 3 Flight Tests I did in August. FWIW, I know its not exactly a scientific description, just an impression.I found a nice youtube channel, sure that's already known, nevertheless quite interesting to see the pure mechanically conditions in such a simulation
...
some more vids are available
Interesting video...
So all we have to do is pressurize a cone and heat it up, and we should get a net thrust?
Back in thread three this was discussed briefly in the context of ionized particles. Yang stated in one of her papers that the behaviour of ionized particles was important to understanding the behaviour of EM Drives. It has been pointed out numerous times that neither Yang or Shawyer published experiments ran in vacuum.
At one point I called this the "ionic wind tunnel" effect in those discussions. That said, these were general discussions and everyone felt that microwave energy was the primary source for trust...
Now we have some excellent simulations that bring this all back. Note that, say, nitrogen ions are orders of magnitude more massive than photons.
In my view, these simulations are a significant step forward in our understanding of this phenomena. Congratulations!
And once again I am in awe of the power of these forums!
Shell
-
#1904 by rfmwguy on 17 Sep, 2015 14:37
-
As a backup to the digital meter made for the LDS, I am using a basic, 10-bit A/D converter. Here it is:
http://www.dataq.com/products/di-145/
This is a USB interface and easy to use. It measures up to +/- 10 VDC and the LDS is about +/- 5 VDC. It will datalog and can provide a simple graphic display on vertical displacement at a sample rate of about 4 msec (240 Hz). The LDS has a sample rate of about 3 msec, so it matches up fairly well.
Any downward force that might briefly appear (counteracting thermal lift) should be able to be picked up in this time frame. Also, any "stickiness" or pivoting errors should also show up. We'll see in a few days.
I would encourage experimenters to datalog with this simple box as a backup to other measurement methods such as laser spots or digital displays. FWIW... -
#1905 by wallofwolfstreet on 17 Sep, 2015 16:17
-
Hmm...
I found a nice youtube channel, sure that's already known, nevertheless quite interesting to see the pure mechanically conditions in such a simulation
some more vids are available
1 - One of my first posts here (way back in thread one) dealt with this - that the EM Drive moved on account of its shape.
2 - I was told the concept violated the laws of thermodynamics. So, given that, are these simulations legit?
Ok, thought experiment. Say we use actual physical objects instead of photons. maybe old fashioned 'super balls' or something similar, though unlike photons they would slow down. A frustum in a zero G environment. Would the results match the simulation?
To answer your question:QuoteSo, given that, are these simulations legit?
No, these simulations are not legit. A frustum in a zero G environment would not have results that match the simulation.
Exactly what when wrong in the simulation, I couldn't say without seeing everything he did to the settings, etc. But the results are completely nonphysical and don't correspond to the real world.
Just intuitively, a bunch of non-zero momentum particles bouncing around in an asymetric cavity isn't a rare system. Such systems exist everywhere. Even the most energetic photons have negligible momentum if we compare them to your average gas particle at STP. So if this simulation was legitimate (which it isn't), then a coke bottle filled with air or your average household balloon ought to be rocketing around with orders of magnitude better thrust than an emdrive.
I'm not going to take the time to flesh out the solution to such a problem mathematically, but it isn't difficult. The case of infinitely many particles is equivalent to the uniform pressure case, and a person only needs some trig to be able to solve it. In fact, the intuition of the problem is simple and shows the issue with what the youtube poster believes is happening.
In the video, he says that because the walls of the cone are angled off the plane of the horizontally flat base, only some of the force they experience when struck by a particle is in the vertical direction. This is correct, but it does not lead to any net movement in the vertical direction because the angled walls also have a greater area than the horizontal base. While one bounce against the flat base creates a greater vertical force than a bounce off the angled side walls, the fact that there is greater chance to hit a side wall than there is the top base due to the greater area perfectly cancels, so there is no net force over time.
If that wasn't enough, also realize that the simulation showed the drive moving towards the direction of the "big" base, which is opposite the direction the emdrive is supposed to go.
So if anything, these simulations are a case study in how you need to have a proper understanding of first principles to tell whether or not your simulations are correct. -
#1906 by CraigPichach on 17 Sep, 2015 16:34
-
Dr. Rodal, you lost me there. I thought that the EM-Drive/Q-Thruster phenomena was linked to the concept that the frustrum had to be at the exact resonance mode (i.e. a frustrum at TE013 at 900MHz would NOT be at TE013 at 2.45GHz and thus would NOT have the same thrust??). Is not a concern in these experiments the question of how exact do you have to be to achieve resonance mode as oppose to simply resonating (to fluidize the quantum vacuum or whatever)?? Indeed with our ANSYS modelling we are finding that the cavity part itself is not so hard, but we've found it to be non-trivial in the way the rf needs to be launched into the cavity.... more complicated is the fact that we want to be able to try firing RF into the large side and be able to switch to the smaller side... though I guess the easiest choice is into the side.
On the contrary, if a cavity of given dimensions resonates at a lower frequency, it will certainly also resonate at higher frequencies, and the higher the frequency the closer to each other (on a % basis) are the natural frequencies to each other. -
#1907 by aero on 17 Sep, 2015 17:09
-
@SeeShells,
Here are some views of the CE3 model. They are very preliminary but ... well, its as far as I have gotten.
https://drive.google.com/folderview?id=0B1XizxEfB23td2ZDeDFUUFdvLUk&usp=sharing
It resonates but it's not a mode that I recognize.
Specifics-
Cavity is the SeeShell Crazy Eddie 2, upgraded by replacing the loop antenna with wave guides, hence CE3 model.
Drive frequency 2.47 GHz, Harminv calculated resonant frequency 2.47837658E+009 Hz, Q is good: ~ 1 million.
Antenna - dual magnetron wave guides, 180 degrees apart. That is, on opposite sides of the frustum using WR 340 wave guides with bottom edge flush with the big base. Phase is adjusted to place sources 3 wavelengths apart.
Gaussian noise bandwidth at 0.05 * 2.47GHz is is double the expected magnetron noise bandwidth.
This is a debug run at resolution = 200, hence it may appear more granular than the resolution = 250 runs commonly uploaded. There are no csv files, resolution of 200 would adversely affect any data comparison calculations based on the csv files.
Great to see you got the waveguide injection working aero.
It looks like you placed the wave right on the bottom to replace the antennas,
If you look at the mode generation in the attached image...
You'll see close to the middle of the frustum is where the mode wants to generate within the frustum. Did you try to do the waveguides insertion at that point? We should try.
Shell
Move the waveguides up to the point where they are not inserting into the frustum which should be halfway up. See image.
Just to be clear on this point, the wave guides do not insert into the frustum. They are placed then cut cleanly at the inside surface of the frustum. The image shows that cut in the frustum cone as projected in the 2-D image of the 3-D object. Not sure how you would actually construct them that way, but it is quite easy to model.
As currently modelled the wave guides are offset in z from the frustum exact center by -3.852 mm toward the big base. That aligns the bottom inside surface of the WR 340 with the inside surface of the big base. I will change that offset to zero from the frustum center and see what happens. -
#1908 by SeeShells on 17 Sep, 2015 18:25
-
Not quite zero aero, we are dealing in the Z direction with different issues in mode generation because of the asymmetry of the frustum. I'll need to search to find the equation Dr. Rodel used in determining the percentages. Search is hideous btw.@SeeShells,
Here are some views of the CE3 model. They are very preliminary but ... well, its as far as I have gotten.
https://drive.google.com/folderview?id=0B1XizxEfB23td2ZDeDFUUFdvLUk&usp=sharing
It resonates but it's not a mode that I recognize.
Specifics-
Cavity is the SeeShell Crazy Eddie 2, upgraded by replacing the loop antenna with wave guides, hence CE3 model.
Drive frequency 2.47 GHz, Harminv calculated resonant frequency 2.47837658E+009 Hz, Q is good: ~ 1 million.
Antenna - dual magnetron wave guides, 180 degrees apart. That is, on opposite sides of the frustum using WR 340 wave guides with bottom edge flush with the big base. Phase is adjusted to place sources 3 wavelengths apart.
Gaussian noise bandwidth at 0.05 * 2.47GHz is is double the expected magnetron noise bandwidth.
This is a debug run at resolution = 200, hence it may appear more granular than the resolution = 250 runs commonly uploaded. There are no csv files, resolution of 200 would adversely affect any data comparison calculations based on the csv files.
Great to see you got the waveguide injection working aero.
It looks like you placed the wave right on the bottom to replace the antennas,
If you look at the mode generation in the attached image...
You'll see close to the middle of the frustum is where the mode wants to generate within the frustum. Did you try to do the waveguides insertion at that point? We should try.
Shell
Move the waveguides up to the point where they are not inserting into the frustum which should be halfway up. See image.
Just to be clear on this point, the wave guides do not insert into the frustum. They are placed then cut cleanly at the inside surface of the frustum. The image shows that cut in the frustum cone as projected in the 2-D image of the 3-D object. Not sure how you would actually construct them that way, but it is quite easy to model.
As currently modelled the wave guides are offset in z from the frustum exact center by -3.852 mm toward the big base. That aligns the bottom inside surface of the WR 340 with the inside surface of the big base. I will change that offset to zero from the frustum center and see what happens. -
#1909 by Rodal on 17 Sep, 2015 18:48
-
What is the relevance of resonating at an arbitrary mode shape like TE013? (Why would one choose TE013 instead of other mode shapes) ?
What theory supports the choice of TE013? What is special about TE013?Dr. Rodal, you lost me there. I thought that the EM-Drive/Q-Thruster phenomena was linked to the concept that the frustrum had to be at the exact resonance mode (i.e. a frustrum at TE013 at 900MHz would NOT be at TE013 at 2.45GHz and thus would NOT have the same thrust??). Is not a concern in these experiments the question of how exact do you have to be to achieve resonance mode as oppose to simply resonating (to fluidize the quantum vacuum or whatever)?? Indeed with our ANSYS modelling we are finding that the cavity part itself is not so hard, but we've found it to be non-trivial in the way the rf needs to be launched into the cavity.... more complicated is the fact that we want to be able to try firing RF into the large side and be able to switch to the smaller side... though I guess the easiest choice is into the side.
On the contrary, if a cavity of given dimensions resonates at a lower frequency, it will certainly also resonate at higher frequencies, and the higher the frequency the closer to each other (on a % basis) are the natural frequencies to each other. -
#1910 by jmossman on 17 Sep, 2015 19:33
-
As a backup to the digital meter made for the LDS, I am using a basic, 10-bit A/D converter. Here it is:
http://www.dataq.com/products/di-145/
This is a USB interface and easy to use. It measures up to +/- 10 VDC and the LDS is about +/- 5 VDC. It will datalog and can provide a simple graphic display on vertical displacement at a sample rate of about 4 msec (240 Hz).
...
Great to hear about the 10-bit ADC DAQ!
I took the liberty to examine their datasheet, and I suspect the DI-145 analog inputs do not contain much low-pass filtering. I mention this as it can have a non-trivial impact on what ADC value gets generated for a given input (i.e. can get "noisy" digital output if there are frequencies greater than 120Hz in the real analog signal).
Adding a simple low-pass filter might not be a bad idea to help eliminate any aliasing noise and help guard against unknown setup/hold requirements of the ADC channel (i.e. set filter to a ~120 Hz cutoff). Also worth keeping in mind the 240Hz sampling rate is only valid for 1 analog channel; if there are additional analog channels, the 240Hz gets subdivided accordingly (i.e. 1 ch@240Hz, 2 ch@120Hz, 3 ch @ 80Hz, etc) so any pre-filter would also need to be adjusted.
As a primer for other lurkers, a key principal in DSP (Digital Signal Processing) is known as the Nyquist Sampling Theorem.
https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem
The basic summary is that when converting analog into digital, you can generally only reconstruct the original analog signal if the digital signal is sampled at a rate at least twice as fast as the highest frequency in the original analog signal. Without the 2x sampling rate (digital sample vs analog freq), the digital signal won't contain all of the information from the original analog signal; worse yet, the higher frequency information can "alias" and masquerade as a lower frequency signal in the digital output (see attached picture from Wikipedia). Since the DI-145 claims a 240Hz sampling rate, any frequency greater than 120Hz in the analog signal can result in aliasing. Pre-filtering the analog signal to eliminate the higher frequency content (i.e. >120Hz) will help avoid any alias-induced noise (although filters inevitably introduce distortion that must be accounted for; primarily a time-delay, but often also a small voltage drop due to losses in a passive filter circuit).
Thanks,
James
-
#1911 by rfmwguy on 17 Sep, 2015 19:41
-
Very nice summary! The LDS has a pre-filtered amp. The entire assembly is an Omron Z4M-W40 plus an Omron K3NX display. Here's a datasheet: http://www.limasoft.cz/omron/pdf/z4m.pdfAs a backup to the digital meter made for the LDS, I am using a basic, 10-bit A/D converter. Here it is:
http://www.dataq.com/products/di-145/
This is a USB interface and easy to use. It measures up to +/- 10 VDC and the LDS is about +/- 5 VDC. It will datalog and can provide a simple graphic display on vertical displacement at a sample rate of about 4 msec (240 Hz).
...
Great to hear about the 10-bit ADC DAQ!
I took the liberty to examine their datasheet, and I suspect the DI-145 analog inputs do not contain much low-pass filtering. I mention this as it can have a non-trivial impact on what ADC value gets generated for a given input (i.e. can get "noisy" digital output if there are frequencies greater than 120Hz in the real analog signal).
Adding a simple low-pass filter might not be a bad idea to help eliminate any aliasing noise and help guard against unknown setup/hold requirements of the ADC channel (i.e. set filter to a ~120 Hz cutoff). Also worth keeping in mind the 240Hz sampling rate is only valid for 1 analog channel; if there are additional analog channels, the 240Hz gets divided accordingly (i.e. 1 ch@240Hz, 2 ch@120Hz, 3 ch @ 80Hz, etc) so any pre-filter would also need to be adjusted.
As a quick primer for other lurkers, a key principal in DSP (Digital Signal Processing) is known as the Nyquist Sampling Theorem.
https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem
The quick summary is that when converting analog into digital, you can only faithfully reconstruct the original analog signal if the digital signal is sampled at a rate at least twice as fast as the highest frequency in the original analog signal. Without the 2x sampling rate (digital sample vs analog freq), the digital signal won't contain all of the information from the original analog signal; worse yet, the higher frequency information can "alias" and masquerade as a lower frequency signal in the digital output. Since the DI-145 claims a 240Hz sampling rate, any frequency greater than 120Hz in the analog signal can result in aliasing. Pre-filtering the analog signal to eliminate the higher frequency content (i.e. >120Hz) will help avoid any alias-induced noise (although filters inevitably introduce distortion that must be accounted for; primarily a time-delay, but often also a small voltage drop due to losses in a passive filter circuit).
Thanks,
James -
#1912 by X_RaY on 17 Sep, 2015 19:55
-
What is the relevance of resonating at an arbitrary mode shape like TE013? (Why would one choose TE013 instead of other mode shapes) ?
@Rodal
What theory supports the choice of TE013? What is special about TE013?Dr. Rodal, you lost me there. I thought that the EM-Drive/Q-Thruster phenomena was linked to the concept that the frustrum had to be at the exact resonance mode (i.e. a frustrum at TE013 at 900MHz would NOT be at TE013 at 2.45GHz and thus would NOT have the same thrust??). Is not a concern in these experiments the question of how exact do you have to be to achieve resonance mode as oppose to simply resonating (to fluidize the quantum vacuum or whatever)?? Indeed with our ANSYS modelling we are finding that the cavity part itself is not so hard, but we've found it to be non-trivial in the way the rf needs to be launched into the cavity.... more complicated is the fact that we want to be able to try firing RF into the large side and be able to switch to the smaller side... though I guess the easiest choice is into the side.
On the contrary, if a cavity of given dimensions resonates at a lower frequency, it will certainly also resonate at higher frequencies, and the higher the frequency the closer to each other (on a % basis) are the natural frequencies to each other.
I agree with you for higher performance of TE modes instead TM as you pointed out some times ago.
But what do you mean with "the higher the frequency the closer to each other"?
If you use dimensions 10 times smaller the total frequency difference between the modes is 10 times bigger
Example(random dimensions):
Dimensions bD=234mm sD=123mm Length=123mm
TE011=2,4543258071GHz
TE013=4,2396819866GHz
Dimensions bD=23.4mm sD=12.3mm Length=12.3mm
TE011= 24,5432580708GHz
TE013=42.3968198663GHz
And @All:
- The higher the mode the more sensitive it is against variation of the dimensions.
+The higher the mode the lower the BW (in particular in relation to the "p" value of J'mnp). -
#1913 by Rodal on 17 Sep, 2015 20:24
-
The statement <<and the higher the frequency the closer to each other >>, implies changing the excitation frequency under the condition of leaving everything else the same. That means,: same materials, same geometry, same boundary conditions, etc. If you change any of the parameters affecting the frequency (you are arbitrarily changing the dimensions in your example below) then not much can be said without knowing in advance the (in this case arbitrary geometrical) change you intend to make. The statement was made in the context of the experiment being discussed in the post that I was answering, in which it was implied that the geometry, materials, etc., remained the same and the only thing being changed was the excitation frequency.
...
But what do you mean with "the higher the frequency the closer to each other"?
If you use dimensions 10 times smaller the total frequency difference between the modes is 10 times bigger
Example(random dimensions):
Dimensions bD=234mm sD=123mm Length=123mm
TE011=2,4543258071GHz
TE013=4,2396819866GHz
Dimensions bD=23.4mm sD=12.3mm Length=12.3mm
TE011= 24,5432580708GHz
TE013=42.3968198663GHz
And @All:
- The higher the mode the more sensitive it is against variation of the dimensions.
+The higher the mode(p value) the lower the BW. -
#1914 by X_RaY on 17 Sep, 2015 20:31
-
The statement <<and the higher the frequency the closer to each other >>, obviously implies changing the excitation frequency and leaving everything else the same. That means,: same materials, same geometry, same boundary conditions, etc. Obviously, if you change any of the parameters affecting the frequency (you are arbitrarily changing the dimensions in your example below) then nothing can be said without knowing the arbitrary geometrical change you intend to make. The statement was made in the context of the experiment being discussed in the post that I was answering, in which it was implied that the geometry, materials, etc., remained the same and the only thing being changed was the excitation frequency.
Agreed. Also for the post before "What is special about TE013?...
But what do you mean with "the higher the frequency the closer to each other"?
If you use dimensions 10 times smaller the total frequency difference between the modes is 10 times bigger
Example(random dimensions):
Dimensions bD=234mm sD=123mm Length=123mm
TE011=2,4543258071GHz
TE013=4,2396819866GHz
Dimensions bD=23.4mm sD=12.3mm Length=12.3mm
TE011= 24,5432580708GHz
TE013=42.3968198663GHz
And @All:
- The higher the mode the more sensitive it is against variation of the dimensions.
+The higher the mode(p value) the lower the BW." 
-
#1915 by rfmwguy on 17 Sep, 2015 20:48
-
Tangent - First Orion Spacecraft Manned Flight?
2023...8 years from now.
Almost as long as it took us to start the space program and land a man on the moon.
Sometimes I wonder. -
#1916 by aero on 17 Sep, 2015 21:27
-
Not quite zero aero, we are dealing in the Z direction with different issues in mode generation because of the asymmetry of the frustum. I'll need to search to find the equation Dr. Rodel used in determining the percentages. Search is hideous btw.@SeeShells,
Here are some views of the CE3 model. They are very preliminary but ... well, its as far as I have gotten.
https://drive.google.com/folderview?id=0B1XizxEfB23td2ZDeDFUUFdvLUk&usp=sharing
It resonates but it's not a mode that I recognize.
Specifics-
Cavity is the SeeShell Crazy Eddie 2, upgraded by replacing the loop antenna with wave guides, hence CE3 model.
Drive frequency 2.47 GHz, Harminv calculated resonant frequency 2.47837658E+009 Hz, Q is good: ~ 1 million.
Antenna - dual magnetron wave guides, 180 degrees apart. That is, on opposite sides of the frustum using WR 340 wave guides with bottom edge flush with the big base. Phase is adjusted to place sources 3 wavelengths apart.
Gaussian noise bandwidth at 0.05 * 2.47GHz is is double the expected magnetron noise bandwidth.
This is a debug run at resolution = 200, hence it may appear more granular than the resolution = 250 runs commonly uploaded. There are no csv files, resolution of 200 would adversely affect any data comparison calculations based on the csv files.
Great to see you got the waveguide injection working aero.
It looks like you placed the wave right on the bottom to replace the antennas,
If you look at the mode generation in the attached image...
You'll see close to the middle of the frustum is where the mode wants to generate within the frustum. Did you try to do the waveguides insertion at that point? We should try.
Shell
Move the waveguides up to the point where they are not inserting into the frustum which should be halfway up. See image.
Just to be clear on this point, the wave guides do not insert into the frustum. They are placed then cut cleanly at the inside surface of the frustum. The image shows that cut in the frustum cone as projected in the 2-D image of the 3-D object. Not sure how you would actually construct them that way, but it is quite easy to model.
As currently modelled the wave guides are offset in z from the frustum exact center by -3.852 mm toward the big base. That aligns the bottom inside surface of the WR 340 with the inside surface of the big base. I will change that offset to zero from the frustum center and see what happens.
Well, we have 3.82 mm to work with. That is almost zero by itself, though it is nearly 1/2 the wave guide height. Frustum is 16.34 mm high, wave guide is 8.636 mm high. That means the lowest (toward the big end) the wave guide can go is 3.82 mm center to center.
I did complete the resonance run with Gaussian noise bandwidth = 0.025 * 2.47 GHz for both cases. With the wave guide aligned with the big end of the frustum, Harminv calculated a Q number in the 10's of millions. With the wave guide centered on the frustum, Harminv calculated a Q number in the 100's of thousands. (Remember, these numbers only indicate relative quality, not absolute quality) I am currently running Harminv with the wave guide moved half the distance to the Big end. That is split the difference between zero and max. We will see. -
#1917 by SeeShells on 17 Sep, 2015 21:39
-
I know that the mode TE011 and TE013 have the same EMF cross sections and I could have done TE011 or a TE013 which I seriously looked at, but the ceramic plate sizes I had available TE012 was easier for me to do. Also a TE012 mode was the goal with RS in the EMDrive he made for Boeing and also a goal for EW when they were trying to get thrust. It was an obvious choice with the materials I could get.The statement <<and the higher the frequency the closer to each other >>, obviously implies changing the excitation frequency and leaving everything else the same. That means,: same materials, same geometry, same boundary conditions, etc. Obviously, if you change any of the parameters affecting the frequency (you are arbitrarily changing the dimensions in your example below) then nothing can be said without knowing the arbitrary geometrical change you intend to make. The statement was made in the context of the experiment being discussed in the post that I was answering, in which it was implied that the geometry, materials, etc., remained the same and the only thing being changed was the excitation frequency.
Agreed. Also for the post before "What is special about TE013?...
But what do you mean with "the higher the frequency the closer to each other"?
If you use dimensions 10 times smaller the total frequency difference between the modes is 10 times bigger
Example(random dimensions):
Dimensions bD=234mm sD=123mm Length=123mm
TE011=2,4543258071GHz
TE013=4,2396819866GHz
Dimensions bD=23.4mm sD=12.3mm Length=12.3mm
TE011= 24,5432580708GHz
TE013=42.3968198663GHz
And @All:
- The higher the mode the more sensitive it is against variation of the dimensions.
+The higher the mode(p value) the lower the BW."
Shell
-
#1918 by cuddihy on 17 Sep, 2015 23:25
-
Tangent - First Orion Spacecraft Manned Flight?
No bucks, no buck rogers.
2023...8 years from now.
Almost as long as it took us to start the space program and land a man on the moon.
Sometimes I wonder. -
#1919 by rfmwguy on 17 Sep, 2015 23:55
-
Good thought, although if you look closely, there is a large budget still available. Not being an insider, it appears a lot of it is earmarked for climate observations, PR and aeronautical applications. If I recall the 60s correctly (OK, so I was a kid), it seemed like NASA went aeronautical in order to advance to space. We have done this already.Tangent - First Orion Spacecraft Manned Flight?
No bucks, no buck rogers.
2023...8 years from now.
Almost as long as it took us to start the space program and land a man on the moon.
Sometimes I wonder.
The way I read it now is LEO is the domain of the ISS and getting up there is being farmed-out. Perhaps thats the natural progression of things, but I'd like to see more focus on advanced spaceflight, manned or unmanned.
Wouldn't it be the ultimate defeat if humanity is destined to be stuck on this rock forever; I guess is my long-term concern...of course, once my time is up, I won't care ;^)
