-
#220 by rfmwguy on 10 Aug, 2015 19:18
-
Thanks, Doc...you cheered me up enough that I feel like thisOff topic alert - Today, I turned the big six-zero; a milestone birthday. Which should be used to convince others that one can still think, dream and build well into their senior era. So many companies and institutions look to replace older folks once they turn 50...I know from experience. I also know that its a waste of talent to do this.
So, here's to the older generation still plugging away at science. How I wish something like this project were around when I was much younger. THAT should be a message to the young folks out there to get involved in science early. Push the envelope, think outside the box, do something different...who knows, what you may come up with could change the world.
Onward and upward - Dave (rfmwguy)
Just think of it, Dave: Donald Trump is over 69 years old, so you have more than 9 years left for you to run for President of the US
https://urbanfragment.files.wordpress.com/2012/09/albert-einstein-ad.jpg -
#221 by X_RaY on 10 Aug, 2015 19:22
-
Happy birthday rfmwguy
Danke, herr x_ray!
Is it possible to record your next live run?
I have little problems with the time difference between our countries, last time i was sleeping, dont see your live show
Did you see this file?
There is a description for possible Q measurement with only one port.
The problem with a magnetron is that is only a source, there is no IQ mixer (or six port) inside the magnetron... So it may be the simplest way is to use a probe port in that case.
http://uspas.fnal.gov/materials/09UNM/ResonantCavities.pdf (big file be patient)
Thanks Mulletron for post the link
BTW: I hope you got a great party today!
-
#222 by rfmwguy on 10 Aug, 2015 19:34
-
I will have a live session the day before I run the actual live test, basically walking throught the setup. This should be 2 weeks or less and hope to announce it next week sometime.Happy birthday rfmwguy
Danke, herr x_ray!
Is it possible to record your next live run?
I have little problems with the time difference between our countries, last time i was sleeping, dont see your live show
Did you see this file?
There is a description for possible Q measurement with only one port.
The problem with a magnetron is that is only a source, there is no IQ mixer (or six port) inside the magnetron... So it may be the simplest way is to use a probe port in that case.
http://uspas.fnal.gov/materials/09UNM/ResonantCavities.pdf (big file be patient)
Thanks Mulletron for post the link
BTW: I hope you got a great party today!
Last thing I need to to get the Galinstan for the liquid metal power connectors. -
#223 by deltaMass on 10 Aug, 2015 19:39
-
@rfmwguy: Happy Birthday!
About that spendy Galinstan: did you at least consider a fully-self-contained rig using batteries? - and what put you off it? -
#224 by rfmwguy on 10 Aug, 2015 19:45
-
@rfmwguy: Happy Birthday!
Thks dm...I was going to use battery until the junk exciter spoiled my day. It was only going to put out about 8 watts after amplification...far too low power for me to be able to measure on the Floobie Stick balance beam...so, the magnetron was chosen.
About that spendy Galinstan: did you at least consider a fully-self-contained rig using batteries? - and what put you off it?
I agree, batteries would have been best, but 4kV bias?...uhhh maybe not so much. -
#225 by deltaMass on 10 Aug, 2015 19:56
-
Yeah, the inverter is the tricky bit. But...
http://powersupply33.com/converter-9v-to-13-5kv.html -
#226 by X_RaY on 10 Aug, 2015 20:06
-
I was thinking about the air ballooning effect... again
And yes i am sure that was already discussed several times.
It is cause by the higher kinetic energy of the particles, they need more space and generate a pressure at the surrounded walls...
I think that effect is only true for a cavity that allow to escape the gas while heating. If it is airtight the number of gas molecules per volume stays constant.
But than there is more energy inside the cavity and based on the restmass+velocity of the gas particles the system will be a little bit heavier the higher the temperature is in contrast to the surrounding.
The gas around the cavity also heats up, that causes a upstream of air.
The last effect is already true for rfmwguy and SheShells
This can be eliminated by testing the cone in the "up" and "down" direction.
In the horizontal direction using perforated copper a force based on this effect would always working forward the small diameter(inner sidewall) same would act at the lower outside wall but forward the big end, so it cancels each other...
Is there some argument i miss in my memory?
(only for my brain protocol...)
https://en.wikipedia.org/wiki/Schlieren_imaging -
#227 by ElizabethGreene on 10 Aug, 2015 20:56
-
For Shawyer to try and patent this before there was even a dream of a viable commercial product is ludicrous.
I vigorously disagree. Mr. Shawyer has, hypothetically, discovered something that we thought was impossible. As such, in addition to his Nobel prize, he is fully entitled to own that intellectual property for the full duration of the patent and take full financial advantage of it. His right to make a billion dollars on this is just as real as your right to make a billion dollars on any improvements you make to it.
A patent is a reward and incentive for the inventor's investment of effort, risk, failures, money, time, and life in creating a device or idea. If they choose to open source it or freely license it then that is their choice. They are under no ethical or moral obligation to do so.QuoteHe should have opened it up to research at the beginning.
He did. He's been writing about it publicly for a decade to anyone who would listen. For his trouble he's been called a fraud and charlatan. He's been mocked and ridiculed. To imply that he's hoarded this to himself is a particularly bitter insult.QuoteBased on what I know, the current iteration is obviously not viable.
Your knowledge is incorrect. He discovered the Shawyer effect and then spent years inventing, experimenting, and refining devices that generated thrust from RF energy. The device accomplished the "impossible". That certainly seems "viable" to me.QuoteIf members of the public are able to own a personal stake in something as evolutionary as this, some will gladly invest enormous amounts of effort, even with no guarantee of any financial reward or any success; like us.
Mr. Shawyer built this technology "with no guarantee of any financial reward or any success." How is his contribution somehow inferior to what we would create?QuoteSo can it be done? Should it be done? If so, we need help...
We just have to make it roll. The rest is history.
Patent law is not simple at all. Even worse, it varies wildly by country and it changes over time. In the US, there is a tiny window for building a copy of a device to determine the accuracy of the specification, for the sole purpose of gratifying a philosophical taste, or curiosity, or for mere amusement. I currently work within those exceptions. Any use beyond that should be negotiated with Mr. Shawyer.
-eag
P.s. ... don't forget that Cannae has patents on this kit too. Even though their drive !did not work! they still own part of the IP. -
#228 by deltaMass on 10 Aug, 2015 21:00
-
I call it "ballooning" when the cavity is perfectly airtight and the walls bulge slightly - the thinner the walls, the greater this effect. Due to the increase in external volume at constant contained mass, the cavity density decreases and thus experiences increased buoyancy in air. There is however a counter-effect present; because the air around the cavity is also heated, it enjoys a lower density, which in turn reduces this buoyancy. Which of these two competing effects wins out depends on all the details.
When the cavity is not airtight, there are two effects to be expected:
a) a loss of all-up cavity mass due to the mass of the expelled air with increased heating
b) the possibility of a rocket effect due to convection of cavity gas in a preferred direction, which in turn depends on where it's leaking. Note that, for certain leakage patterns, one would see the artifact of reversed thrust when the cavity is flipped, since now the "jet" direction is also flipped.
Note that, in vacuum, none of the above occur if care is taken to evacuate the cavity. -
#229 by X_RaY on 10 Aug, 2015 21:06
-
I call it "ballooning" when the cavity is perfectly airtight and the walls bulge slightly - the thinner the walls, the greater this effect. Due to the increase in external volume at constant contained mass, the cavity density decreases and thus experiences increased buoyancy in air. There is however a counter-effect present; because the air around the cavity is also heated, it enjoys a lower density, which in turn reduces this buoyancy. Which of these two competing effects wins out depends on all the details.
This thought was based on inelastic walls who are thick enough..
When the cavity is not airtight, there are two effects to be expected:
a) a loss of all-up cavity mass due to the mass of the expelled air with increased heating
b) the possibility of a rocket effect due to convection of cavity gas in a preferred direction, which in turn depends on where it's leaking. Note that, for certain leakage patterns, one would see the artifact of reversed thrust when the cavity is flipped, since now the "jet" direction is also flipped.
Note that, in vacuum, none of the above occur if care is taken to evacuate the cavity. -
#230 by WarpTech on 10 Aug, 2015 21:15
-
@rfmwguy: Happy Birthday!
About that spendy Galinstan: did you at least consider a fully-self-contained rig using batteries? - and what put you off it?
You can buy a cheap UPS Pure-sine inverter relatively cheap, and use that to power the magnetron transformer.
Todd
-
#231 by SeeShells on 10 Aug, 2015 21:21
-
https://drive.google.com/folderview?id=0B1XizxEfB23tfmVCZlhnNUVtTGRoTTlZSGJiT3ZfSHBvYnNELUc1WlN0Sk15ZDZud3pSWmM&usp=sharing
Since the mode shapes are shown on different scales, without showing the numbers, a mode shape that has magnitude that is dozens or hundreds of times smaller amplitude (and thus insignficant) looks significant when compared without numbers. Hence the images are very difficult to understand .
I took both Center x and Center Y pages and opened them into separate folders, sized them so one row was 14 images across and the vertical columns were synced 00, 01, 02, 03, 04... same time slice. Then switched back and forth between the pages comparing a image and mode for X to Y and at the same position.
In some cases the modes match X and Y but several they don't even come close. Just another flag of an interesting anomaly of how you could see in a relatively symmetrical cavity mode variation just between a X slice through the middle and a Y slice through the middle in the same time stamp.
Or has everyone seen this weirdness and you're going to go... phhhsst, silly old woman...
Shell
Remember how we progressed:
1) At the beginning of MEEP runs, the same fields were shown on different scales from time step to time step. It looked so crazy that even the field outside the EM Drive had flashing colors. And inside the EM Drive we could see fractals that were numerical artifacts.
This was fixed by showing all the time steps at the same final scale. This fixed the fractals and the crazy colors outside the EM Drive.
2) We still have the problem that different fields in different directions: Ex, Ey, Ez, Hx, Hy, and Hz, are shown with different scales. Thus one is not able to distinguish, say if E in the x direction is 1000 times less than E in the y direction. Thus we cannot distinguish "noise" from signal.
ALL electric fields should be shown on the same scale and all magnetic fields should be shown to the same scale. Otherwise there is no comprehension.
3) I cannot tell what the mode shape being excited is, from looking at these images: is it TE ? is it TM? nobody knows. Why is it that nobody knows? Because the fields are not shown to the same scale, thus nobody can tell whether it is TE or TM. We don't know what is noise and what is signal.
Is the field in the longitudinal z direction an electric field (and hence a TM mode) or is it a magnetic field (and thus a TE mode). We cannot tell, both mode shapes are being shown and we don't know which is large and which is negligible.
If they would be shown to the same scale, and one mode shape happens to be 100 times smaller magnitude than another one, it would look like zero, and we would not look at it, thus we would better understand because we would be looking ONLY at the mode shapes that are high in magnitude. The way it is now, we don't know what is high magnitude and what is low magnitude.
Imagine what life would be like if your senses would be such that we would not be able to distinguish between what is near and what is far, between what is high and what is low, we would be lost.
Sigh. Lost? No, just regroup.
Thanks for the recap, I did't realize we handicapped the visual data as much as we did in truncating data simply making it unusable as a visual aid.
I hope the numbers haven't been truncated in the CSV files as well. Back to lurk and dig.
Shell
-
#232 by Rodal on 10 Aug, 2015 21:40
-
...
The numbers are in the csv files, but it is much more time consuming to post-process the csv files than to just click and look at the images. All my 3 computers are otherwise occupied at the moment. I had a small window of time in one computer which I used to calculate something that Todd will be interested in
Sigh. Lost? No, just regroup.
Thanks for the recap, I did't realize we handicapped the visual data as much as we did in truncating data simply making it unusable as a visual aid.
I hope the numbers haven't been truncated in the CSV files as well. Back to lurk and dig.
Shell Hope to get to this soon.
-
#233 by Rodal on 10 Aug, 2015 21:52
-
...
Agreed, they misuse the word "constant" when pertaining to the variables, alpha and beta. It is a terminology issue though, not math that is incorrect.
It is not magic. Equation 12 is the derivative of equation 2, per equations 9 and 10, for TE modes. Equations 13 and 14 are the derivatives of equation 5, for TM modes. Since there is no component of Er in equation 2, there is no radial component of attenuation for the E field in the r direction. But there is for the TM mode, in equation 5 to give equation 14.
Todd
It may be interesting to plot the γ function, the logarithmic gradient of the electric field:
γ = - dLog[E]/dr = - (1/E)*dE/dr
defined by Zeng and Fan, and apply it to the case of standing waves in a closed resonant cavity: Yang/Shell for TE011 and TE012, to see what it looks like.
In this case γ = γθ = γφ
γθ = - (1/Eθ)*dEθ/dr = γφ = - (1/Eφ)*dEφ/dr
γ is real, these are standing waves hence there is no imaginary component of γ.
γ grows without bounds, to Infinity, at each end because the transverse electric fields are zero at the big base and at the small base in order to satisfy the boundary condition that electric fields parallel to a metal boundary must be zero. Since γ is defined as the ratio of the gradient of the electric field with respect to r, divided by the electric field, when the field is zero at the bases, while the numerator is maximum, γ is infinite at the boundaries.
For TE012, γ also grows without bounds at the middle node of the two half-wave patterns because at that point the transverse electric field is zero while its gradient with respect to r is maximum.
Notice that γ is negative at the small base (small r) and positive at the big base (large r)....
Yes, basically γ = γ (r) hence it does not make sense that γ is treated as a constant to define
If one has a function with exponential format F=exp(k.x) and k is a constant then one can write k=(dF/dx)/F.
But if one has a function F=exp(k(x).x) then (dF/dx)/F=k(x) + xdk(x)/dx
If F has no exponential format is worse.
The wave solutions in spherical coodinates are the form exp(+ik.r)/r and exp(-ik.r)/r only when r goes to infinity.
E = A e - γ r
γ = - (1/E) dE/dr = α + j β
because this is only true for γ = constant
So, really
- (1/E) dE/dr = γ + r dγ/dr
So Zeng and Fan's expression is exact when dγ/dr = 0, that is when γ is constant.
and approximate for dγ/dr ~ 0 (γ nearly constant)
so in the images shown in http://forum.nasaspaceflight.com/index.php?topic=38203.msg1414705#msg1414705 above and in Zeng and Fan's figures, the γ expression is more accurate as a measure of attenuation where γ is flat (where the gradient dγ/dr ~ 0 ) and nearly constant, which is more nearly the case for values such that:
γ = 0 (NO ATTENUATION)
Thanks to our "NSF Peer Review Committee" with members Ricvl, Todd "WarpTech" and yours truly we have peer-reviewed Zeng and Fan's paper better than was done by the original editiorial review by Optics Express Journal:
Electromagnetic fields and transmission properties in tapered hollow metallic waveguides
Xiahui Zeng and Dianyuan Fan
https://www.osapublishing.org/oe/fulltext.cfm?uri=oe-17-1-34&id=175583
Ricvl uncovered the important error that Zeng and Fan failed to properly derive the γ exponent, as they improperly considered it to be a constant in the initial part of their derivation and then a posteriori they inconsistently considered it to be a function of r.
I have properly solved the first order differential equation for γ(r):
r dγ/dr + γ = - (1/Eθ)*dEθ/dr
that arises from solving for the exponent γ in the following expression:
Eθ = A e - γ r
(observe that Zeng and Fan did not solve this equation, instead they considered r dγ/dr << γ and hence assumed
γ = - (1/Eθ)*dEθ/dr).
When properly solving the differential equation (numerically, since it does not have a closed form solution, and it is quite difficult to solve numerically because it contains highly oscillating functions and it diverges at the boundaries)
I get the enclosed attenuation curve for γ for Yang/Shell in TE011 mode, with the RF feed off, standing waves, which does make sense (the solution based on Zeng & Fan's equation γ = - (1/Eθ)*dEθ/dr does not make sense because it contains a large geometrical magnification at the small end).
The attenuation is positive over the whole EM Drive, being close to zero over most of it but diverging to high attenuation at both ends because (1/Eθ)*dEθ/dr) diverges at the ends as previously explained due to the fact that the electric field Eθ must be zero at the bases. Therefore the tangent electric field Eθ gets 100% attenuated at the small base and at the big base. The attenuation curve for the other tangent electric field Eφ is identical.
Notice the asymmetry: there is more attenuation going on towards the small base.
Thus we have properly solved this problem for the case in which Eθ satisfies the bondary conditions for a resonant cavity.
Recall
1) Yang/Shell geometry:
bigDiameter = 0.201(*meter*);
smallDiameter = 0.1492(*meter*);
axialLength = 0.24(*meter*);
equivalently:
r1 = 0.693281 (*meter*); (*spherical small radius *)
r2 = 0.933978 (*meter*); (*spherical large radius *)
θw = 6.15933 (*degrees*); (*half-cone angle*)
2) the geometry we consider is as shown in this image:
-
#234 by SeeShells on 10 Aug, 2015 22:00
-
...
The numbers are in the csv files, but it is much more time consuming to post-process the csv files than to just click and look at the images. All my 3 computers are otherwise occupied at the moment. I had a small window of time in one computer which I used to calculate something that Todd will be interested in
Sigh. Lost? No, just regroup.
Thanks for the recap, I did't realize we handicapped the visual data as much as we did in truncating data simply making it unusable as a visual aid.
I hope the numbers haven't been truncated in the CSV files as well. Back to lurk and dig.
Shell Hope to get to this soon.
That's ok Jose, I'm going to take some time and do it. I need to.
Thanks,
Shell -
#235 by Possibles on 10 Aug, 2015 22:02
-
Hi everyone...
Been in lurking mode for a while now, and have been wringing my head for answers. No joy. I do have an idea, but I need to see the results of more vacuum tests to be reasonably confident.
I would like to take this opportunity to say this to the DIY guys. I would dearly love to build myself, but simply have no time or shop. I can only think in the rare moments I have. You really are pulling together and doing a great job.
But I need more null tests! I need someone to put a heating element in and test for deformation etc...
From different shaped frustums.
I need to see possible radiation leakage from the endplate.
I need a ramped test in a vacuum chamber bringing the power up to 100% and I need to see what happens when the magnetron pops. And in atmosphere too!
I need this because if on the off chance that this thing works, it will help me to understand the underlying process.
And I'm saying this because I believe that we are exciting a very particular set of circumstances. I honestly don't believe for a moment that the shape of the frustum is THAT important. After all, we have seen potential thrust from varying designs. I mean- look at the Cannae drive - Totally different. I can understand Temjar. He realized this too. Hence a hole you could stick your head through and peek around in.
MEEP will prove powerful in the end, I'm sure of it. But just not for the moment. Don't kill me Aero, I know you have done such fine work there...It can help guide us. But should not be considered so pertinent until we have figured the basics.
I think its time to balance the tests, and not get overloaded with results. There is also a high chance I've missed out on stuff here. It's been a busy month. And @Shell- how goes the measurement setup? Sorry to hear about your loss.
Bedtime now. -
#236 by Rodal on 10 Aug, 2015 23:42
-
It is thanks to aero running Meep that we are able to calculate and show:
1) The huge difference between having RF feed ON with travelling waves vs. the case examined by Greg Egan (RF feed OFF with standing waves)
2) The fact that there is Poynting vector directionally oriented with the RF feed ON that keeps growing with time exponentially, with the RF Feed on.
3) The fact that the stresses and hence the forces at the small and big bases are greatly influenced by the antenna with the RF feed ON
4) The difference between placing the antenna at the small end vs the big end
5) The difference between placing the antenna at the axis of axi-symmetry vs being offset
6) The difficulties of exciting TE modes with dipole antennas
I say no, nobody here is overloaded, rfmwguy and SeeShells can carry much more on their shoulders, so aero, keep it going with
7) Loop antennas to excite TE modes
8 ) Showing what happens when your turn the RF feed OFF
9) Modeling a waveguide entering the cavity to excite a TE mode in the cavity
10) Showing how steady-state is approached
etc.
-
#237 by Ricvil on 11 Aug, 2015 01:40
-
Dr Rodal, I think there is a problem again.
Standing waves.
In the cavity, at least, the fields are superpositions of two counter propagating waves, then one has (to simplify)
E=[a.exp(gamma.r)+b.exp(-gamma.r)].F(tetha,phi)
Repair, the expression above is just a aproximate model to fit the fields by exponential functions, where the correct is to use fractional spherical hankel functions (for spherical ends cavitys), where a,b and gamma will fataly be functions of r. At r>>1 the spherical hankel function decay with 1/r.
Then how to differentiate a destructive interference at node fields from a infinity attenuation?
Worse, even with a,b and gamma being constants, when one take the derivative of E above, because the two signals of gamma (outward/inward from/to apex cone) one has
dE/dr=gamma[a.exp(gamma.r) -b.exp(-gamma.r)].F(theta,phi), and we lost original E function, then imagine with a,b and gamma being r functions.
This is just my way to think. You may, of course, disagree.
-
#238 by SeeShells on 11 Aug, 2015 02:00
-
It is thanks to aero running Meep that we are able to calculate and show:
I also am very glad aero is here doing work that has effectively moved this group forward with answers garnered in meep that simply were not available to the group in any other fashion. And with your help we got some very good answers.
1) The huge difference between having RF feed ON with travelling waves vs. the case examined by Greg Egan (RF feed OFF with standing waves)
2) The fact that there is Poynting vector directionally oriented with the RF feed ON that keeps growing with time exponentially, with the RF Feed on.
3) The fact that the stresses and hence the forces at the small and big bases are greatly influenced by the antenna with the RF feed ON
4) The difference between placing the antenna at the small end vs the big end
5) The difference between placing the antenna at the axis of axi-symmetry vs being offset
6) The difficulties of exciting TE modes with dipole antennas
I say no, nobody here is overloaded, rfmwguy and SeeShells can carry much more on their shoulders, so aero, keep it going with
7) Loop antennas to excite TE modes
8 ) Showing what happens when your turn the RF feed OFF
9) Modeling a waveguide entering the cavity to excite a TE mode in the cavity
10) Showing how steady-state is approached
etc.
But a couple of questions arise in how we can accomplish this with the data being displayed as it currently is.
How can we even know that a TM mode can be excited when you declared that the displayed modes were of little value after the filtering of the data files to rid the display of the artifacts and flashing colors. The only way I can see is not to be dependent on the display of meep and post-process but use another program, Wolfram or something else?
I need to know how to forecast the future use of meep and post processing because this second build I'll be going after building a stable frustum with ceramic gold plated endplates and an active feedback system to assure mode and frequency lock. If I can't get the answers here than I'll need to take a longer path.
I was busy today tracking down and talking with old contacts that are going to try to get me the ceramics for the endplates. I'll need to tell them before too long what size I want them to be. The final shape of the ERD frustum will depend on how well meep and post processing can show what shape is optimal for it. For that we need as stable of an answer we can get. Simply for keeping a TE mode in a stabilized frustum, I think this is needed.
Shell
-
#239 by Rodal on 11 Aug, 2015 02:49
-
...
Perhaps aero (or somebody else) will find a way to:
I also am very glad aero is here doing work that has effectively moved this group forward with answers garnered in meep that simply were not available to the group in any other fashion. And with your help we got some very good answers.
But a couple of questions arise in how we can accomplish this with the data being displayed as it currently is.
How can we even know that a TM mode can be excited when you declared that the displayed modes were of little value after the filtering of the data files to rid the display of the artifacts and flashing colors. The only way I can see is not to be dependent on the display of meep and post-process but use another program, Wolfram or something else?
I need to know how to forecast the future use of meep and post processing because this second build I'll be going after building a stable frustum with ceramic gold plated endplates and an active feedback system to assure mode and frequency lock. If I can't get the answers here than I'll need to take a longer path.
I was busy today tracking down and talking with old contacts that are going to try to get me the ceramics for the endplates. I'll need to tell them before too long what size I want them to be. The final shape of the ERD frustum will depend on how well meep and post processing can show what shape is optimal for it. For that we need as stable of an answer we can get. Simply for keeping a TE mode in a stabilized frustum, I think this is needed.
Shell
1) Plot all the electric fields with the same magnitude scale, and plot all the magnetic fields with the same magnitude scales
2) Plot the contour plots for the electromagnetic fields with an attached color bar showing the numerical magnitude of the contour colors.
I am not conversant with Meep but it seems to me that this need is something that many other users of Meep also have (to plot all the fields with the same magnitude and to be able to show the numerical value of the contours) and therefore may be an already existing command or something that somebody may have already documented how to do.
I know that other codes with which I am conversant (ANSYS, etc.) this is standard. Actually the ability to show the numerical value of the contours was available in ANSYS from its first versions back in the early 1980's.
