-
#2460 by rfmwguy on 16 May, 2016 01:48
-
Perhaps you should actually build one and show us your capabilities. Such expertise shouldn't go to waste on a keyboard.So, I thought I would sweep the frustum BEFORE soldering closed...yeah I know, its not a good idea. So, looks like center is around 2.474 GHz, higher than the model, but within mag range. Return Loss 43dB which is outstanding without solder. If I measure Q from 3dB off best RL, it would be crazy high. Not going to go there yet, things will change after it gets soldered. Bottom line, almost ready to button up...
You folks are all re-inventing resonant cavity filters. A return loss of 43 db is OK, but not great, and the Q should be well over 100,000 for a good design, even with a basic plated aluminum cavity. Again, once more. You can design a filter for an expected signal, or you can TUNE THE SOURCE for an expected Q. You folks seems to be slowly re-learning basic microwave engineering, and I'm glad to see that there is more talk of making the source track resonance, rather than trying to micro-machine a cavity to resonate with a crap source like a magnetron.
EW references to 20 Newtons per Kilowatt give me hope, but I'm not holding my breath because I don't want to turn blue.
EDIT: The s-parameter measurements you all make are only as good as the source and receiver you have on hand. Cheap USB instruments are severly limited in performance (simple physics). Phase noise, AM/PM conversion, lack of phase information ALL contribute to lack of useful information. Again, s-parameters are impedance specific, by definition. If the system is not normalized to a particular (generally 50 ohm) impedance, the results are useless. I have seen NO attempt by ANYONE on this forum to normalize their impedance. Any reference to injected power levels, or S-Parameters, are therefore worthless.
Note: A magnetron is most definitely NOT a 50 ohm source, nor is a microwave oven a 50 ohm load. The take away is that an Em drive cavity is not a 50 ohm load, and should never be modeled as such. -
#2461 by SeeShells on 16 May, 2016 02:11
-
Shell, if you're still on...re this image of yours, I don't recall exactly; this very clean frequency/signal was the result of powering your Maggie with DC, but is it fair to assume that it is a water cooled Maggie and you experience no "signal drift" related to the active cooling ? thnx FL
The critical things you need to do for the magnetron to clean up the output. Don't forget magnetrons are used in the semiconductor industry and need to have a clean output and a stable output.
Clean DC to the magnetron. Critical as it keeps the splattering and AM and also any DC component and removes the 50% chopped duty cycles from regular microwave power supplies.
Thermal cooling other than air. Coiled copper water cooling through a heat exchanger works well. Need more cooling then submerse in chilled water.
Turn off the magnetron heater after 3-4 seconds, you don't need it after.
Shell
-
#2462 by FattyLumpkin on 16 May, 2016 02:54
-
So given the above mentioned signal/frequency "drift" is ameliorated? Thnx , FL
Looking forward to your results!!!!!
-
#2463 by Gordon Powell on 16 May, 2016 04:08
-
Use of a tungsten torsion wire eliminates magnetic effects. Also, it is helpful to provide a copper disk perpendicular to the wire axis with a stationary U magnet around the possibly moving Cu disk. Induced eddy currents dampen out the oscillation of the torsion balance. If the magnet is a copper electromagnet, the damping effect and any magnetic disturbance can be turned off.
-
#2464 by TheTraveller on 16 May, 2016 06:18
-
Some more experimental data to share and that has been confirmed by another EmDrive builder.Quote
Before my amp died some of the tests I did indicated the best thrust freq was not the same as the best rtn loss freq. I believe this is due to frustum manufacturing errors introducing phase distortion in the multiple reflected travelling waves being compensated by a slightly off freq exciton such that the say negative phase distortions introduced by the freq are compensated by positive phase distortions caused by the manufacturing defects resulting in the min overall phase distortions over the life of the travelling wave & thus generating a higher thrust than operating at best rtn loss freq.
I also found the thrust bandwidth was very much narrower than the rtn loss bandwidth.
This info is preliminary & needs to be confirmed by a lot more tests when I get my amp back.Quote...Agreed on both counts.
I’ve seen the same thing, (max force freq. not equal to min VSWR freq, and peak force output having a narrower BW than the RF resonant cavity’s -3dB bandwidth)... -
#2465 by TheTraveller on 16 May, 2016 06:22
-
Shell, if you're still on...re this image of yours, I don't recall exactly; this very clean frequency/signal was the result of powering your Maggie with DC, but is it fair to assume that it is a water cooled Maggie and you experience no "signal drift" related to the active cooling ? thnx FL
The critical things you need to do for the magnetron to clean up the output. Don't forget magnetrons are used in the semiconductor industry and need to have a clean output and a stable output.
Clean DC to the magnetron. Critical as it keeps the splattering and AM and also any DC component and removes the 50% chopped duty cycles from regular microwave power supplies.
Thermal cooling other than air. Coiled copper water cooling through a heat exchanger works well. Need more cooling then submerse in chilled water.
Turn off the magnetron heater after 3-4 seconds, you don't need it after.
Shell
All very good advise. -
#2466 by TheTraveller on 16 May, 2016 06:32
-
So, I thought I would sweep the frustum BEFORE soldering closed...yeah I know, its not a good idea. So, looks like center is around 2.474 GHz, higher than the model, but within mag range. Return Loss 43dB which is outstanding without solder. If I measure Q from 3dB off best RL, it would be crazy high. Not going to go there yet, things will change after it gets soldered. Bottom line, almost ready to button up...
You folks are all re-inventing resonant cavity filters. A return loss of 43 db is OK, but not great, and the Q should be well over 100,000 for a good design, even with a basic plated aluminum cavity. Again, once more. You can design a filter for an expected signal, or you can TUNE THE SOURCE for an expected Q. You folks seems to be slowly re-learning basic microwave engineering, and I'm glad to see that there is more talk of making the source track resonance, rather than trying to micro-machine a cavity to resonate with a crap source like a magnetron.
EW references to 20 Newtons per Kilowatt give me hope, but I'm not holding my breath because I don't want to turn blue.
EDIT: The s-parameter measurements you all make are only as good as the source and receiver you have on hand. Cheap USB instruments are severly limited in performance (simple physics). Phase noise, AM/PM conversion, lack of phase information ALL contribute to lack of useful information. Again, s-parameters are impedance specific, by definition. If the system is not normalized to a particular (generally 50 ohm) impedance, the results are useless. I have seen NO attempt by ANYONE on this forum to normalize their impedance. Any reference to injected power levels, or S-Parameters, are therefore worthless.
Note: A magnetron is most definitely NOT a 50 ohm source, nor is a microwave oven a 50 ohm load. The take away is that an Em drive cavity is not a 50 ohm load, and should never be modeled as such.
Basically I agree with this.
Your need to be able to electronically tune the excitation freq to get the best VSWR and directly measure the forward and reflected power to/from the frustum while exciting the frustum at the desired power level and then continually adjust by active feedback, to best VSWR is a good start to track frustum resonance changes.
BTW I have found as I indicated in another post, that operation at lowest VSWR freq will not necessarily generate the best thrust and that the thrust bandwidth is very much narrower than the rtn loss bandwidth. So just being close in drive freq is not good enough to ensure thrust. You need to tune for it.
Let me repeat this so it is not lost:
Thrust bandwidth is very much narrower than the rtn loss bandwidth.
A VNA should get you close to the resonance freq but you need to use the dynamics of the filled cavity and the Rf drive system to actively tune for the best VSWR at power and then use a load cell or other force or acceleration sensor to tune for the best thrust. Which means a low cost USB VNA is fine to get you to step one. But with real time forward. reflected power and freq control, a VNA (low cost or not) may not be needed. -
#2467 by X_RaY on 16 May, 2016 07:36
-
That there isn't conversation about at the moment doesn't mean it was never noticed in the forum.So, I thought I would sweep the frustum BEFORE soldering closed...yeah I know, its not a good idea. So, looks like center is around 2.474 GHz, higher than the model, but within mag range. Return Loss 43dB which is outstanding without solder. If I measure Q from 3dB off best RL, it would be crazy high. Not going to go there yet, things will change after it gets soldered. Bottom line, almost ready to button up...
You folks are all re-inventing resonant cavity filters. A return loss of 43 db is OK, but not great, and the Q should be well over 100,000 for a good design, even with a basic plated aluminum cavity. Again, once more. You can design a filter for an expected signal, or you can TUNE THE SOURCE for an expected Q. You folks seems to be slowly re-learning basic microwave engineering, and I'm glad to see that there is more talk of making the source track resonance, rather than trying to micro-machine a cavity to resonate with a crap source like a magnetron.
EW references to 20 Newtons per Kilowatt give me hope, but I'm not holding my breath because I don't want to turn blue.
EDIT: The s-parameter measurements you all make are only as good as the source and receiver you have on hand. Cheap USB instruments are severly limited in performance (simple physics). Phase noise, AM/PM conversion, lack of phase information ALL contribute to lack of useful information. Again, s-parameters are impedance specific, by definition. If the system is not normalized to a particular (generally 50 ohm) impedance, the results are useless. I have seen NO attempt by ANYONE on this forum to normalize their impedance. Any reference to injected power levels, or S-Parameters, are therefore worthless.
Note: A magnetron is most definitely NOT a 50 ohm source, nor is a microwave oven a 50 ohm load. The take away is that an Em drive cavity is not a 50 ohm load, and should never be modeled as such.
Thread 3:Quote... Use a tuner and calibrate the system before {open(r=+1_j0),load(50R),short(r=-1_j0)}
http://forum.nasaspaceflight.com/index.php?topic=37642.msg1412290#msg1412290
I know calibration kits are expensive but it works, uncalibrated your VNA is less helpful.
https://www.maurymw.com/Precision/VNA_Cal_Kits.php
...
don't know if your USB-series-VNA have a cal mode (it has to have i think
)
Thread 7:QuoteOK this was a tip for people using a noisy magnetron as source because its cheap but not tunable in frequency.
Its less of interest what is tuned, the source frequency at fixed resonator geometry or the cavity at fixed frequency of the source(movable endplate or similar). If this done well the result is similar.
But again I am with you, If a tunable source(+amp) is available (maybe PLL control) it is much easier to tune the frequency.
BTW there are automated mechanical stub tuners available at the market.
http://www.mksinst.com/docs/UR/SmartMatch-DS.pdf
A mechanical cavity tuning could be automated in the same way.
ReplyQuoteNotify...
A return loss of 43dB is not that bad for a single cavity resonator. The bigger problem is to be ensure the value holds under thermal expansion of the system using several hundred watt of input power.
Signal filtering was also already discussed in the past.
In fact there is a lot of good info in these threads.
Magnetron info: pdf-file below
-
#2468 by TheTraveller on 16 May, 2016 08:02
-
Quote
...A mechanical cavity tuning could be automated in the same way.
Yup automated physical length tuning as attached.
The Demonstrator EmDrive also had the ability to tune the magnetron frequency and used tunable stub matchers.
It is only a guess but I would suggest Roger would have used good microwave engineer practice with a very stable magnetron Dc supply and could probably reduce or eliminate magnetron filament current as well.
It would seem he used every technique I know of and maybe other unknown to put that cavity into the sweetest operation mode possible. He spend 2 years putting that system together, so EmDrive DIYers if you think building a working EmDrive is simple or quick or low cost, well you may get lucky or maybe not.
-
#2469 by Rodal on 16 May, 2016 12:37
-
So, in General Relativity the energy E and momentum p obey the relativistic energy–momentum relation
I was about to post how I assumed that the "higher inertial mass" near the large end in Dr. McCulloch's theory must be meant in a way that the total momentum in the photons is reduced when the mass increases.* This post from him confirms he understands the direction issue with Shawyer's theory, so he isn't likely to be making the same mistake, and comparisons of his McCulloch's increased mass to Shawyer's increased momentum are wrong.Interesting new comment from Dr. Mike McCulloch on his EmDrive theory:
https://twitter.com/memcculloch/status/731774299380568064?s=01
Both Mike & Roger talk about the thrust being produced by larger EmWave momentum / mass at the big end and smaller EmWave monentum / mass at the small end with the EmDrive thrust being the result of gained momentum from the lost EmWave momentum / mass differential between the big & small ends.
http://physicsfromtheedge.blogspot.com/2015/06/crit-of-shawyers-emdrive-theory.html
* I don't understand how inertial mass even makes sense when talking about photons, but McCulloch's theory has room with new physics, general relativity, and interactions between photons and other radiation that can presumably carry momentum away from the cavity. My limited understanding is that the increase in mass comes with transfer of momentum to Unruh waves, so that the momentum decreases despite the increase in mass. I could be off base, and there are other consistent interpretations I can think of.
In general relativity, gravity affects anything with momentum and energy. Photons do not have a rest-mass, but they do have momentum, and energy--and they are thus effected by gravity (photons cannot leave a black hole and their path in space is affected by the geodesics imposed by a massive object like the Sun).
...
where mo is the rest mass, which is an invariant, so since photons have zero rest mass: mo = 0, we have
which says that a photon's momentum is a function of its energy, and also its velocity is an invariant: the speed of light in vacuum, c
Therefore the "inertial mass" that McCulloch is discussing is not the relativistic concept of rest mass, which is an invariant, but McCulloch's "inertial mass" is instead a measure of its momentum.
The De Broglie expression for the photon's momentum is
One could divide this momentum by the speed of light in vacuum, c, to get a momentum that has "mass units":
p/c = h/(c λ)
which is a relativistic mass
where the Lorentz factor γ is defined as follows (and using nomenclature where the speed is v=u, sorry about using Wikipedia images for equations that use different nomenclature for the speed in different articles)
(Notice that the relativistic mass for a photon is undefined since the rest mass mo for a photon is zero while γ becomes infinite at v=u=c, since the photon speed v=u is c, and this entails multiplying zero by infinity which is undefined, so again, it is better to think of this expression as being momentum measured in "mass units")
Somebody at the "other EM Drive forum" trying to debunk McCulloch wrote:QuoteSo far I believed, that having standing waves inside the cavity of any shape implies, these waves have everywhere the very same wavelength. First of all we should demonstrate, that the wavelength of standing waves changes inside the conical resonator along its axis - if it doesn't change, then no dependence of photon mass on wavelength applies and the whole above line of reasoning is wrong with massive photons or without them (not to say, it would be also redundant with respect to the "Unruh radiation" stuff anyway).
No,on the contrary: while cavities with constant cross-section do have standing waves that have constant wavelength along the longitudinal axis, tapered cavities like the EM Drive have monotonically changing wavelength along the longitudinal axis. For example, for a truncated cone with spherical ends, the exact solution is given by spherical Bessel functions, which have non-uniform wavelength along the longitudinal axis.
See, for example: http://gregegan.customer.netspace.net.au/SCIENCE/Cavity/Cavity.html
[Small end is at left and big end is at right. The cavity cross-section increases with increasing "r" which is the spherical radius with r=0 being at the vertex of the cone]
In an EM Drive, the wavelength is smaller towards the big end and the wavelength gets bigger towards the small end:
So McCulloch's measure of momentum in "mass units"
p/c = h/(c λ)
does change along the longitudinal axis of the EM Drive:
for a cavity with constant cross-section like a cylindrical cavity, the momentum of the photons is constant along the longitudinal axis. However, for a cavity like the EM Drive, the momentum of the photons changes with their location along the longitudinal axis.
Since in an EM Drive, the wavelength is smaller towards the big end and the wavelength gets bigger towards the small end, p/c = h/(c λ) means that the "momentum in mass units" p/c = h/(c λ) gets larger towards the big end and smaller towards the small end
Hence, when McCulloch writes:
I interpret this as saying MiHsC shifts "momentum in mass units" from the microwave injection towards the big end of the tapered cavity, so as new microwave energy is put into the cavity its centre of momentum is continually being shifted by MiHsC towards the wide end.
It is also interesting that in special relativity, the center of momentum frame is necessarily unique only when the system is isolated, which is something that makes non-isolated experiments questionable. This is another reason why experiments using a power cord are questionable, since the power cord goes to a stationary power supply, hence the center of momentum in experiments with a power cord may NOT be in the moving platform, and why experiments with batteries in the moving platform (as the latest experiments by Yang that nullified her previous experimental claims, and also as the experiments by Hackaday Aachen team) are much more convincing. http://users.physics.harvard.edu/~morii/phys151/lectures/Lecture16.pdf
Experiments using power cords ignore the electromagnetic momentum being fed from a stationary power supply. -
#2470 by Monomorphic on 16 May, 2016 12:47
-
So, I thought I would sweep the frustum BEFORE soldering closed...
What are you using for a signal generator? I got official approval from the spouse to purchase the USB spectrum analyser, so now i'm looking for an economical signal generator.
Also, how do you like VNA/J? -
#2471 by rfcavity on 16 May, 2016 12:48
-
Yes, the usb vna has the open, short and load cal kit. The OP is mistaken about 43dB rl. Try to buy a 900W load or cavity and post the rl specs. Better yet, they can build an assymetric cavity themselves and prove their statement.
Conical frustrums have an infinite number of planes that divide the shape into two symmetric halves. -
#2472 by TheTraveller on 16 May, 2016 13:26
-
In an EM Drive, the wavelength is smaller towards the big end and the wavelength gets bigger towards the small end:
Maybe review the attached as Roger shows how his EmDrive cavity generates a momentum gradient that transfers lost small end momentum (big end - small end momentum) to the cavity such that big end momentum = small end momentum + cavity gained momentum to balance overall momentum.
So it seems Mike & Roger are different dogs but with the same leg action.
-
#2473 by Rodal on 16 May, 2016 13:36
-
Yes, it is true that Shawyer understood that tapered cavities have non-uniform wavelength in the longitudinal direction, and therefore Shawyer deserves credit for recognizing that the wavelength changes along the length of a tapered cavity.In an EM Drive, the wavelength is smaller towards the big end and the wavelength gets bigger towards the small end:
Maybe review the attached as Roger shows how his EmDrive cavity generates a momentum gradient that transfers lost small end momentum (big end - small end momentum) to the cavity such that big end momentum = small end momentum + cavity gained momentum to balance overall momentum.
So it seems Mike & Roger are different dogs but with the same leg action.
But the problem is that something else is needed to self-accelerate the EM Drive, beyond Newton. Maxwell and Special Relavity, because conservation of momentum and conservation of energy are principles that if violated, then just about everything else in physics needs to be re-formulated. That is why Dr. McCulloch, Dr. White, Prof. Woodward, and so many others trying to make sense of this have come out with alternative theories, because Shawyer's theory is unsound regarding these conservation principles. (His "thrust force" balance of forces is questionable, and his invoking kinetic energy (which implies a reference for speed measurement) runs contrary to his claim that he uses Special Relativity).
McCulloch, White and Woodward critics debate whether those effects are real, but they don't debate their force balance equations. -
#2474 by meberbs on 16 May, 2016 13:46
-
...
That has been done. And many issues have been pointed out. Recently an issue with the conservation of energy slide was asked very nicely to Shawyer in an e-mail exchange with rfmwguy, and Shawyer ignored the question entirely. (It was just a few pages back if you missed it)
Maybe review the attached.
As I just pointed out and then Rodal provided a link to where Mike McCulloch pointed out the same last year, the arrows on the Conservation of momentum slide indicate Shawyer's confusion with the direction of forces. Cullen's equations if applied blindly and the side walls are ignored result in a force (applied to the cavity walls) in the direction of the "thrust" arrow on that slide. Because from the definition of force, acceleration of an object is in the same direction as the force, the acceleration arrow should also be to the left.
Rather than reposting things that provide no new information, could you address any of the self-contradictions with this theory? Failing that, (since I can think of no way to resolve these contradictions other than simply saying Shawyer is wrong) you could just focus on experiment results and let other people consider theories that at least have a slim chance of explaining any anomalous forces instead of insisting people accept a theory that is equivalent to 1=0. -
#2475 by X_RaY on 16 May, 2016 13:51
-
Source file can be downloaded here: http://forum.nasaspaceflight.com/index.php?topic=39772.msg1534941#msg1534941
-
#2476 by Rodal on 16 May, 2016 14:02
-
If this works out, Shawyer may prove to have had the right stuff "inventor's intuition". What I don't understand is why Shawyer dismisses other explanations like McCulloch's, White's etc: Edison, Westinghouse and Tesla didn't debate theories with Hertz: they invented and experimented.
Its a strange beast I've never seen the likes of before. Its almost as if you take a great cavity and distort its natural resonance modes. This bifurcation is certainly something new to me. Shawyer picked up on it and as Rodal says, he might not have the correct theory, but others like White and McCullough are theorizing this new form factor produces an unexpected result. Still quite interesting that CoM might be responsible for the momentum and therefore it's prohibiting a violation. Will this classical physics law be the key to the observed effect? Don't know but this is the kind of stuff that "fascinates" me.Yes, the usb vna has the open, short and load cal kit. The OP is mistaken about 43dB rl. Try to buy a 900W load or cavity and post the rl specs. Better yet, they can build an assymetric cavity themselves and prove their statement.
Conical frustrums have an infinite number of planes that divide the shape into two symmetric halves.
Actually, I think what were doing here is creating a single frequency RF load by using a cavity filter...weird.
Even Dr. White did not discuss his theory in his 2004 AIAA paper, which concentrated on the experiments. -
#2477 by Monomorphic on 16 May, 2016 14:08
-
The VNA/J is great...no sig gen needed. Its normalized just like the old school benchtop units I used to tweek endlessly. Granted you won't get the absolute accuracy, but several hundred $ versus several thousands make it a no-brainer. Tuning cavity filters are best done with sig gens with calibrated cables and connectors, but there again, several thousands for a lab quality generator not to mention the cal kits and cables. Slow the sweep rate down and narrow the display BW on the VNA/J and you'll do just fine I believe.So, I thought I would sweep the frustum BEFORE soldering closed...
What are you using for a signal generator? I got official approval from the spouse to purchase the USB spectrum analyser, so now i'm looking for an economical signal generator.
Also, how do you like VNA/J?
Looks like you are using the miniVNA Tiny ($560) with VNA/J (open source) as the application. This would be a more economical solution than the $530 Triarchy spectrum analyser + $400 signal generator. -
#2478 by Notsosureofit on 16 May, 2016 14:16
-
(No internet). Even cyl cavity has var wavelength in grav field or accl frame of reference
-
#2479 by Rodal on 16 May, 2016 14:24
-
(No internet). Even cyl cavity has var wavelength in grav field or accl frame of reference
Yes. This effect is testable: wouldn't the anomalous force be different in different orientations?
1) EM Drive longitudinal axis in the HORIZONTAL direction (anomalous force greater than #2 and smaller than #3 ?)
2) EM Drive longitudinal axis in the VERTICAL direction, small end pointing down (anomalous force is smaller than #1 and #2 ?)
3) EM Drive longitudinal axis in the VERTICAL direction, big end pointing down (anomalous force is greater than #1 and #2 ?)
Mmmm, if this would be significant, EM Drive small end pointing up would act as "antigravity."
Wouldn't a huge (black hole-like) gravity field be needed for the gravitational field to have a significant effect on the photon momenta ?
)