-
#3160 by Monomorphic on 25 Feb, 2016 20:07
-
Nice results!
Could you so kind to add the power/frequency graphs or even magnitude/frequency and/or the complex S parameter for your last simulations?
(I know it depends on the port/source definition and used requests )
I will see what kinds of graphs I can get out of postFEKO. I did not request S parameter before running the solver, so I do not think that is available. I can request S parameter during the 300Mhz refining sweeps.
I've also been trying to get the Time Domain Analysis going just because I think it would look cool to watch the waves bouncing around in the frustum. But I can't seem to get the time signal to match the 2 - 3 Ghz domain. Was hoping you could help with that! -
#3161 by SeeShells on 25 Feb, 2016 20:18
-
Frequency sweep finished on flat-end frustum. I'm showing TE013 resonance at 2.88889Ghz, vs the 2.73737 for the spherical end-plate frustum. This is a difference of 151.52Mhz. In addition, full power at resonance was only 1/4 that of the spherical end-plate frustum (7,500 kV/m vs 1,750 kV/m).
OK that is a difference of only 5.5 % in resonant frequency between spherical ends and flat ends.
I need to run frequency sweeps over 300Mhz centered on 2.88889 and 2.73737Ghz as stronger resonance could have been missed since there is a limited number of frequency steps. So these numbers may move around. We'll see how much of that 151Mhz I can make up, but I can't imagine it being that much.
This is what I expected (only a few % difference in resonant frequency), and it is very much in line with my previous calculations.
What is the geometrical tolerance of the frustum of a cones made so far? Early on TheTraveller was pointing out differences of ~5% as signficant in his opinion, but when it came to make his own experimental frustum of a cone he has reported that imperfections due to home-construction are such that they detract from the Q, and he is exploring manufacturing alternatives to improve the geometrical tolerance.
This has also been pointed out by the Chinese co-worker of Prof. Yang that briefly visited the NSF forum a few weeks ago: geometrical imperfections are easily going to impact the natural frequency and most importantly going to affect the Q.
What's going to impact more than the cavity shapes is the bandwidth of the signal source and the insertion RF (we figured that out a long time ago). If you look at his frequency sweep which I've seen in FEKO sims before you will see that with even across a 30MHz bandwidth you can see a significant change in the modes shape. What FEKO doesn't show with this sim is a simulated bandwidth of 20 or 30 MHz a run over just one cycle and how the standing waves of the mode react within a wider bandwidth. Is it stable? Does it tend to travel and oscillate like in a meep sim over one cycle?
Shell
This is what is important as well.
-
#3162 by Monomorphic on 25 Feb, 2016 20:18
-
OK that is a difference of only 5.5 % in resonant frequency between spherical ends and flat ends.
This is what I expected (only a few % difference in resonant frequency), and it is very much in line with my previous calculations.
That 5.5% difference is enough so there is little to no TE013 resonance overlap between the two frustums. Since I made the sweeps identical, they can be played side-by-side.
http://youtubedoubler.com/hHjy -
#3163 by Monomorphic on 25 Feb, 2016 20:22
-
Is it stable? Does it tend to travel and oscillate like in a meep sim over one cycle?
I'm trying to get the Time Domain Analysis working in FEKO, but can't seem to get the time signal to match the ~2.5Ghz range. I tried for an hour yesterday with no success. Will try again soon, but I have a dinner tonight. -
#3164 by X_RaY on 25 Feb, 2016 20:23
-
OK is this the whole frequency range of the run(2...3GHz)? I think so. Shell shows a gif generated by some cycles at almost fixed frequency.OK that is a difference of only 5.5 % in resonant frequency between spherical ends and flat ends.
This is what I expected (only a few % difference in resonant frequency), and it is very much in line with my previous calculations.
That 5.5% difference is enough so there is no TE013 resonance overlap between the two frustums. Since I made the sweeps identical, they can be played side-by-side.
http://youtubedoubler.com/hHjy -
#3165 by SeeShells on 25 Feb, 2016 20:25
-
It was 2.45GHz with a 30MHz bandwidth 10 slices over one cycle.
OK is this the whole frequency range of the run? Shell shows a gif generated by some cycles at fixed frequency.OK that is a difference of only 5.5 % in resonant frequency between spherical ends and flat ends.
This is what I expected (only a few % difference in resonant frequency), and it is very much in line with my previous calculations.
That 5.5% difference is enough so there is no TE013 resonance overlap between the two frustums. Since I made the sweeps identical, they can be played side-by-side.
http://youtubedoubler.com/hHjy
Shell -
#3166 by Monomorphic on 25 Feb, 2016 20:34
-
Simply start postfeko, go to the tab results, make a right-click on power --> Add to new-->Cartesian graph
I didn't have power scaling enabled so that chart just shows zero all across. I will run the 300Mhz sweep at 1Kw so we can see a chart. And look into the S parameter. -
#3167 by Monomorphic on 25 Feb, 2016 20:40
-
OK is this the whole frequency range of the run(2...3GHz)? I think so. Shell shows a gif generated by some cycles at almost fixed frequency.
Yes, the whole freq range 2 - 3Ghz.
This is the tutorial I had problems with matching the time signal to the 2 - 3 Ghz bandwidth range. The tutorial is in Mhz, and the math is not very intuitive. Trying to get a Time Domain Analysis of the microwaves bouncing around in the frustum. Can you see if you have better luck?
-
#3168 by rfmwguy on 25 Feb, 2016 21:04
-
Thanks Doc...I did not do a search like I should have. I contacted him via offsite contact info as he appears to not have been posting here for a while.
Usually it is not a good idea to disclose username identities, and of course I'm not going to do this. Dr. McCulloch has not been using a monicker, and it easy to find his posts using Google search, for example:
Username?I've invited Mr McCullough to join the forum and address questions...
Dr. McCullough is already a member of the NSF forum and has previously posted in previous EM Drive threads
http://forum.nasaspaceflight.com/index.php?topic=36313.msg1336056#msg1336056QuoteDear excellent NSF forum. This is just to say (for those who don't know) that I've suggested a specific new model for inertia that predicts galaxy rotation without dark matter (it is called MiHsC) and I recently compared its predictions with the 9 EmDrive results with 'some' (not perfect) success. You can see the results by looking at the Table here:
http://physicsfromtheedge.blogspot.co.uk/2015/02/mihsc-vs-emdrive-data-3d.html
I've also published a paper summarising this comparison (slightly out of date now) here
http://www.ptep-online.com/index_files/2015/PP-40-15.PDF
I'd encourage those with other interesting explanations of the EmDrive to make a similar comparison between predictions and data, so we can compare using the facts. If you have any more data points to add, or if you disagree with the numbers in my Table, please let me know.
NSF member section shows that he is still a member of NSF. -
#3169 by rfmwguy on 25 Feb, 2016 21:50
-
Here ya go doc:
@memcculloch
@rfmwguy Thank you. I'm a member & happy to answer, but not today. I'm preparing lectures for tomorrow..
-
#3170 by Rodal on 26 Feb, 2016 00:01
-
Nice results!
Could you so kind to add the power/frequency graphs or even magnitude/frequency and/or the complex S parameter for your last simulations?
(I know it depends on the port/source definition and used requests )
I will see what kinds of graphs I can get out of postFEKO. I did not request S parameter before running the solver, so I do not think that is available. I can request S parameter during the 300Mhz refining sweeps.
I've also been trying to get the Time Domain Analysis going just because I think it would look cool to watch the waves bouncing around in the frustum. But I can't seem to get the time signal to match the 2 - 3 Ghz domain. Was hoping you could help with that!
Thanks Monomorphic for reporting these FEKO runs at NSF
We are very glad that you are here ! -
#3171 by CW on 26 Feb, 2016 09:25
-
Concerning the earlier brought up idea in this thread of perfectly parallel and perfectly reflective mirrors on two spacecraft, reflecting an initial volley of e.g. E = 1kJ of photons in between them.. wouldn't this be a way to 100% convert the photons' energy into kinetic energy (of both spacecraft, flying off into opposed directions)? The photons should be red-shifted into 'oblivion' by doing so.
-
#3172 by SeeShells on 26 Feb, 2016 14:25
-
Just doing some morning reading. Thought I'd like to share.
http://boingboing.net/2014/11/24/the-quest-for-a-reactionless-s.html
(mod note - NSF has lots of additional info on this: http://forum.nasaspaceflight.com/index.php?topic=31037.0 for those interested in reading more) -
#3173 by Rodal on 26 Feb, 2016 15:30
-
Just doing some morning reading. Thought I'd like to share.
http://boingboing.net/2014/11/24/the-quest-for-a-reactionless-s.html
The first test conducted by NASA on Shawyer's EM Drive measured <<NO significant>> anomalous force, even at power inputs 12 times higher than the power input at which an anomalous force was later reported with a dielectric insert.
* 30 Watts for TE012 mode shape with NO dielectric: no significant measurable force
* 2.6 Watts for TE012 mode shape with dielectric: the highest force/InputPower measured in the NASA report !!!!!
This bears repeating: same mode shape, no dielectric results in no significant measureable force even when employing over 12 times higher input power !!!!!!
When NASA tested the Cannae device, they found the same as they did with the frustum of a cone: no significant force without a dielectric insert in the Cannae device as well. NASA only reported significant anomalous force when a dielectric insert was inserted in the Cannae drive.
Fortunately, Paul March was in the NASA Eagleworks team, since Paul had worked with Prof. Woodward, who, based on the Mach Effect and relativity, showed that if a dielectric is submitted to a varying electric power (charge or discharge), Woodward's hypothesis predicts a transient mass fluctuation, see this:
https://en.wikipedia.org/wiki/Woodward_effect#Transient_mass_fluctuation
This transient inertial mass fluctuation would enable self-acceleration of the EM Drive under such a varying electric field.
This was the reason why Paul March at NASA's Eagleworks team headed by Dr. White, came up with the idea to insert an internal dielectric when they observed no anomalous force in the EM Drive without an internal dielectric insert.
All the reported tests by NASA Eagleworks have featured dielectric inserts.
Only NASA (Dr White) and TU Dresden (Prof. Tajmar) have reported tests in vacuum.
Shawyer, through his many years in his quest for the EM Drive propulsion has failed to perform a single test in partial vacuum. This is very noteworthy.
The measured forces by NASA and TU Dresden in partial vacuum are much smaller than the forces measured in air, which shows that the large forces claimed by Shawyer and Yang may be due to the fact that they did not perform their experiments in partial vacuum. This is a major flaw of Shawyer's and Yang' experiments since it has been well known since Maxwell's time that to measure the radiation pressure one must preferably perform such experiments in a partial vacuum, (because of the effect of thermal convection). See Lebedev's paper on his experiments that finally confirmed radiation pressure measurements:
P.N. Lebedev, Ann. der Physik, 6, 433 1901
EXPERIMENTAL EXAMINATION OF LIGHT PRESSURE
http://web.ihep.su/dbserv/compas/src/lebedev01/eng.pdf
______________________________________________________________
The force/InputPower measured by NASA in a partial vacuum is significantly higher than the force/InputPower measured by Tajmar in a partial vacuum. The reason for this may be due to A) NASA has used dielectric inserts, while Tajmar (under advice from Shawyer) has not used dielectric inserts, or B) NASA reported a much higher Q. To clarify this one would have to ascertain how was the Q measured, and compare the experiments on equal ways to measure the Q, and one would have to ascertain how scalable are Tajmar's results at a higher Q(*)
(mN/kW) Q mN/(kW Q) Force (mN) Power (W) Dielectric ?
Tajmar 0.0286 20 1.43*10^-3 0.02 700 NONE
NASA 1.10 6726 1.64*10^-4 0.055 50 YES (HDPE)
(NASA's force/(InputPower*Q) is actually 9 times lower than Tajmar's)
(NASA also reports a measurement in the opposite direction of only 0.283 mN/kW, however no Q was reported for this test)
There are also issues of coupling the power in these tests. Was Tajmar's very low Q due to severe undercoupling?
Apparently Tajmar did not use an iris, apparently the test was conducted with a relatively large opening from the waveguide into the cavity. Therefore, it may not be proper to compare NASA's and Tajmar's tests based on their InputPower. Not enough information has been given by Tajmar to clarify these issues.
____________
(*) When Shawyer used dielectric inserts in the past, Shawyer used inorganic dielectric with high electric permittivity (as disclosed in Shawyer's patent), unlike NASA, who has been using organic polymer dielectrics with low electric permittivity. This is a huge difference, since Shawyer when he used dielectric inserts, used completely different materials with completely different properties than what NASA has been using. Thus Shawyer's early experience with dielectric testing in the EM Drive appears to be irrelevant to NASA's findings. -
#3174 by TheTraveller on 26 Feb, 2016 15:33
-
The first test conducted by NASA on Shawyer's EM Drive conclusively showed zero force: NO anomalous force.
This often quoted statement of your is not correct. Please check with you know who. -
#3175 by Rodal on 26 Feb, 2016 15:38
-
The first test conducted by NASA on Shawyer's EM Drive conclusively showed zero force: NO anomalous force.
This often quoted statement of your is not correct. Please check with you know who.
The fact that the first test reported by NASA was without a dielectric insert, and NASA reported no significant anomalous force is clearly reported in NASA's report:
Anomalous Thrust Production from an RF Test Device Measured on a Low-Thrust Torsion Pendulum
Brady, David (NASA Johnson Space Center, Houston, TX, United States);
White, Harold G. (NASA Johnson Space Center, Houston, TX, United States);
March, Paul (NASA Johnson Space Center, Houston, TX, United States);
Lawrence, James T. (NASA Johnson Space Center, Houston, TX, United States);
Davies, Frank J. (NASA Johnson Space Center, Houston, TX, United States)
http://ntrs.nasa.gov/search.jsp?R=20140006052Quote from: Brady et.al., page 18There appears to be a clear dependency between thrust magnitude and the presence of some sort of dielectric RF resonator in the thrust chamber. The geometry, location, and material properties of this resonator must be evaluated using numerous COMSOLŪ iterations to arrive at a viable thruster solution. We performed some very early evaluations without the dielectric resonator (TE012 mode at 2168 MHz, with power levels up to ~30 watts) and measured no significant net thrust.
(Bold added for emphasis)
this is meticulously shown, step by step, in the following post:
http://forum.nasaspaceflight.com/index.php?topic=39214.msg1469685#msg1469685
and the subsequent posts. The resonance has also been independently verified by X-Ray and by my calculations using an exact solution. Where it is proven without a shadow of a doubt that NASA's test without a dielectric was in resonance as reported by NASA.
It is also a fact that all the tests reporting measured forces in the Brady report have been with a dielectric insert, both for the Cannae and the frustum of a cone cavities.
According to the Brady report, NASA measured no significant force both in the Cannae and frustum of a cone cavities when the dielectric was removed. -
#3176 by Monomorphic on 26 Feb, 2016 17:24
-
Ok Shell, here is my first attempt at Time Domain Analysis in FEKO. It's not very intuitive, so this first sim has the wrong frequency for resonance.
In order to get it to work, this frustum is about 5 meters tall and the frequency used is around 150Mhz. Hopefully I can get better control and learn how to shrink this down to Ghz bandwidth.
Total signal duration: 100 ns Amplitude: 1 Pulse delay: 50 ns Pulse width: 50ns
-
#3177 by X_RaY on 26 Feb, 2016 17:33
-
OK is this the whole frequency range of the run(2...3GHz)? I think so. Shell shows a gif generated by some cycles at almost fixed frequency.
Yes, the whole freq range 2 - 3Ghz.
This is the tutorial I had problems with matching the time signal to the 2 - 3 Ghz bandwidth range. The tutorial is in Mhz, and the math is not very intuitive. Trying to get a Time Domain Analysis of the microwaves bouncing around in the frustum. Can you see if you have better luck?
snip
It works (click on the GIF below) but I canīt export longer animations. 
This is a first short try only. There is still a lot of optimisation potential.
-
#3178 by SeeShells on 26 Feb, 2016 17:44
-
Ok Shell, here is my first attempt at Time Domain Analysis in FEKO. It's not very intuitive, so this first sim has the wrong frequency for resonance.
Nice work! I'm sure you'll get it down. In order to get it to work, this frustum is about 5 meters tall and the frequency used is around 150Mhz. Hopefully I can get better control and learn how to shrink this down to Ghz bandwidth.
Total signal duration: 100 ns Amplitude: 1 Pulse delay: 50 ns Pulse width: 50ns
The point being is there is so much going on in this frustum when you are close to resonance. That a static slice image will not relay to you.
Shell -
#3179 by TheTraveller on 26 Feb, 2016 18:08
-
As posted to EmDriveResearch:
My PSA is increasing fairly quickly (means they still have not yet killed all the cancer). Next week starts more surgery, radiation treatment and maybe drugs. As this will put a damper on my plans, here is a short update from a bit of behind the curtain action.
THE OMG moment still fills my thoughts as what I heard was like a 2kg mass dropping several cms inside the small end down frustum (very solid THUD). Best proper resonant power on reaction force generation so far is ~8mN at ~92W forward power.
There is a new frustum in process but I have stopped it until I can properly make it happen. I feel the 8mN is just about as much as can be achieved with the current frustum. Main issue is, I suspect, a very slight non parallelism of the end plates and the end plate to frustum side wall joints are not Rf tight. Look at all the images of the Experimental, Demonstrator and Flight Thruster EmDrives to see the effort (flange thickness & number of clamping bolts) that was engaged to make this joint Rf tight. Any leaking Rf adds to the eddy current wall losses and drops cavity Q. Non parallel end plates, I suspect, results in the Rf walking off to the widest separation point and bunching up against the side wall at that point.
====================
Phil Wilson
Feb 24 (1 day ago)
To All,
Where all this came from was my own OMG moment, when early on I powered on my system, with the PA enabled by a slide switch I incorrectly left on and the scale went Over Limit. I could hear the "thump" as the downward Force hit the scale. Then all quite as all of the system powered up.
I use a master AC switch to turn everything off at night & back on in the morning. Seems I made 2 fortunate errors that night. Left the PA enable micro slide switch on & the battery powered scale switched on.
Tried like everything to make it happen again but nothing worked. So like others, move on to stuff I could make work.
Then the laser powered sail & the photonic reflection thruster got me thinking that it is more about the bounce rate per second than the Q.
All I know for sure is I witnessed and HEARD a LOT more than 8mN being generated for a few seconds by a LOT less than 90 watts that was not resonant with the frustum. I do believe, had the small end been pointing up, I would have witnessed liftoff. Really, truly do believe the Force was that large.
In his breadcrumb way, Roger just confirmed that thought.
Phil
via Note 2
On 2/23/2016 4:30 AM, Phil Wilson wrote:
Hi Guys,
Sent the lower email to Roger for his comments.
He just replied Q is used as an approximation of the number of bounces versus decaying cavity energy.
Which MAY be correct for a decaying no longer powered frustum but it is not the case for a continually powered frustum.
Guys we may have been going down a chase the high Q cavity pathway, that for sure works but just maybe it is not where the main action is.
Comments please.
Phil
via Note 2
Hi Roger,
Hate lying in bed, doing nothing but the brain is in high gear. Started thinking about your reactive Force equation F = (2 P Q Df) / c and how Q really relates to it. Came to the conclusion that Q should only be applicable as to the cavity charge and discharge time and is not involved in the Force equation of a CW Rf signal, which is driven by the standard radiation pressure bounce equation F = (2 P) / c and the number of bounces per second of the resonant RF wavefront. Which is what the photonic thruster works with as attached.
Assuming you are correct with the small end having a reduced radiation pressure versus the big end, due to longer guide wavelength at the small end versus the big end (which is your equation as attached) then applying the photonic thruster number of bounces off one end + your Df equation generates a very different force equation:
Force = ( 2 P Df Number of Bounces) / c
where number of bounce per one end plate = 2 / one way transit time as attached.
Assuming 1kW Rf energy, Df = 0.65 (typical value), TE013 mode (3 x 1/2 resonant waves) & freq = 2.45 GHz.
F = 2 x 1,000 x 0.65 x 816,666,667) / c = 3,600 N/kW
Which seems nuts but it follows the photonic thruster example.
I really don't see where Q is involved in the Force equation. Am I drug affected?
Also suggests TE011 excitation would have 3x the Force generation as the frustum transit time should be 1/3 that of TE013 excitation or 3 x more end plate bounces per sec than TE013.
Ok finally starting to get sleepy. Nite nite.
Phil
==========================
Phil
via Note 2
Could you so kind to add the power/frequency graphs or even magnitude/frequency and/or the complex S parameter for your last simulations?
In order to get it to work, this frustum is about 5 meters tall and the frequency used is around 150Mhz. Hopefully I can get better control and learn how to shrink this down to Ghz bandwidth. 
It works (click on the GIF below) but I canīt export longer animations. 