-
#480 by Peter Lauwer on 04 Jul, 2017 09:30
-
Fantastic, Jamie! I will come back to this later.I have tested it on a scale, like described in the article, and found no sign of pushing. I will further test it on my torsion balance (next month). It is quite easy to test, you just put a dummy load on the balance and see whether there is any deflection if it is powered.
Also, the mantle (cavity wall) can easily be cooled with water flow.
I was able to roughly simulate your coupling cavity/waveguide. With the coax and connectors in the cavity, which are hard to quantify since I don't have exact dimensions, it won't be exact. I'm pretty sure the second image below shows TM011. I'm not sure about the first, but it looks like the two antennas are coupling better with that mode. I would need to do more setup to run a proper S21.
Having the RF source and main power off the test rig may solve some of my noise issues. We could simplify your coupling cavity to a rectangular waveguide with E-probe. That way only a small hole is required. And that small hole is small enough that 2.45Ghz barely leaks out. This is a simplified sim of the concept that seems to check out. In reality, the waveguide and E-probe would be located at the center of the torsional pendulum, feeding RF through the bottom to a SMA cable that leads to the frustum. There wouldn't even be the need for battery operated power detectors as reflected power could be monitored off-rig by using a circulator before the waveguide.
Jamie:
I would steer clear of this isolated feed approach to testing the EMdrives due to the complaint that if any element of the RF source is mounted in the laboratory frame of reference, the argument can be made that any unbalanced forces developed by the frustum are just leveraged off the RF power supply and its mounts to the lab via its RF feed lines. The only convincing way to demonstrate these EMdrives is to treat them as "free flyers" with the controls, RF source and battery flying WITH the frustum as they would in free space. That recommendation came out of the July 2014 Eagleworks (EW) Blue Ribbon PhD panel and that was the primary reason we built the Integrated Copper Frustum Test Article (ICFTA) and Cavendish Balance test article the way we did.
Best, Paul M.
Dear Paul,
I don't really see this problem (at least, not as being a big problem).
Of course, you have to perform all kind of tests to show this way of feeding does not impose disturbing forces itself: dummy-load instead of frustum, cylindrical cavity instead of frustum, etc.
But I see everyone struggling with the 100+ W of heat generated on the measurement device when only measuring some tens of micronewtons. It is just not convincing. I think the problems are much less with this 'isolated feed approach' (and the 'convincing power' stronger).
Anyway, it is how I am going to do it. It will surely be useful if several ways are tried.
Cheers, Peter
My article, An improved method to measure microwave induced impulsive forces with a torsion balance or weighing scale, at arxiv: https://arxiv.org/abs/1706.04999 -
#481 by Peter Lauwer on 04 Jul, 2017 09:48
-
Funny Business at the ArXiv
McCulloch is not the only physicist facing this kind of omerta from arXiv anonymous administrators. I know others. Although publishing in peer-review academic, non predatory access journals, they have in common being alternate candidates to standard ΛCDM concordance cosmological model. It's a topsy-turvy world: the arXiv, which used to be a preprint server, now acts like a peer-review postprint club, at least in the field of cosmology.
Indeed. Therefore I did not include the word 'EmDrive' in the title, abstract or keywords. '-)
https://arxiv.org/abs/1706.04999 -
#482 by Chrochne on 04 Jul, 2017 10:02
-
I was just reading and noticed meberbs had figured out the 2nd order doppler effects which I thought was cool here: https://forum.nasaspaceflight.com/index.php?topic=37642.msg1413761#msg1413761 when a thought struck me.
Light is able to transfer more of its energy effectively to a lighter object such as a free electron than it is able to transfer its energy to a more massive object. Now in the tip of the frustum we have some large electric fields which could possibly ionize gas while at the large end ionization may be less so.
snip...
Could that possibly make sense?
What is curious about a plasma possibly existing in the cavity is that the plasma can more effectively absorb kinetic energy from light. I think I remember Shell mentioning the possibility of plasma in the cavity at one point in time. Possibly this is why she made her cavity see through or that screen mesh? The plasma would need to be created, so energy would start to be stored in the cavity and later plasma could form when the electric fields become strong enough to strip electrons from the gasses inside.
Snip....
Does it coinside with experimental evidence suggesting some minimal power level required for the effect to really take hold. Some minimal power required to form plasma?
Just some non-educated guess. I wonder if what you describe sounds a bit like small scale version of the ITER plasma reactor. From that picture of the corss-section of the reactor on wikipedia it do resembles a bit shape of the modified EmDrive, except of course that it is circular shape reactor in the end. Could there be some similarity to EmDrive or not at all?
https://en.wikipedia.org/wiki/ITER
Did anyone tried circular shape of the EmDrive
?
From the equation in the paper it seems there are some conditions that reduce the breakdown voltage/meter if I was reading it correctly. I suspected under certain conditions the breakdown voltage would be a fraction of the expected 3*10^6 V/m. Still that's a fairly high V/m. Read through it a few times but will have to sub some values in to see if the break down can be lowered enough. I am unsure it is reasonable to assume there is a plasma but it is interesting that there could be a plasma and at one end of the cavity.
I think what was important was that they "give a method of observing if a plasma is formed". Directly confirming or observing what is actually going on in the cavity is integral to understanding what might be occurring if anything.
Very interesting what you describe here. Do we have some specialist for plasma here? Perhaps it can be another try for the explanation of this device.
https://goo.gl/JYDaxV - Plasma classification (types of plasma) - Link edit
http://education.jlab.org/qa/plasma_02.html -
#483 by ThatOtherGuy on 04 Jul, 2017 10:32
-
Very interesting what you describe here. Do we have some specialist for plasma here? Perhaps it can be another try for the explenation of this device.
https://www.plasma-universe.com/Plasma_classification_(types_of_plasma)
http://education.jlab.org/qa/plasma_02.html
The forum mangled the URL (didn't include the closing parens)
https://www.plasma-universe.com/Plasma_classification_(types_of_plasma) -
#484 by ThatOtherGuy on 04 Jul, 2017 10:45
-
Just a crazy thought (feel free to skip this post of mine)
We have a resonant cavity with two end plates of different size, now, we inject photons into the cavity and those photons start bouncing back and forth (ok, more or less) betweeen the end plates, BUT while a large amount of those photons is able to hit the "big" plate, a somewhat reduced amount of them hits the "small" one (collisions and so on), this may cause an inbalance in the energy transmitted to the plates, where the large one gets more hits and more energy transfer while the smaller one gets less hits and, in turn, minor energy transfer.
Now, may this difference be the cause of the "anomalous thrust" being observed ?
-
#485 by Chrochne on 04 Jul, 2017 10:58
-
Just a crazy thought (feel free to skip this post of mine
)
We have a resonant cavity with two end plates of different size, now, we inject photons into the cavity and those photons start bouncing back and forth (ok, more or less) betweeen the end plates, BUT while a large amount of those photons is able to hit the "big" plate, a somewhat reduced amount of them hits the "small" one (collisions and so on), this may cause an inbalance in the energy transmitted to the plates, where the large one gets more hits and more energy transfer while the smaller one gets less hits and, in turn, minor energy transfer.
Now, may this difference be the cause of the "anomalous thrust" being observed ?
"In a star, it works something like this. Large clouds of gas in space are thought to have "condensed" into more dense blobs due to gravitational attraction within the gas. As the blobs grow more dense, the gravitational pull is stronger." quote from that link I sent earlier.
But as I said it it also only non-educated guess on my side. Perhaps Dr. Rodal or others can give better explenation on this.
Edit: I changed the link for the plasma wiki. Thank you for the notification. -
#486 by Monomorphic on 04 Jul, 2017 13:12
-
Again, thanks a lot for the sims. But the frequencies you got out of it, 3.50 and 3.35 GHz, do not really come close to the measured frequencies (3.60 and 3.25 GHz).
Does it due to, apart from the influence of the connectors etc., the fact that one of the endplates in not in contact with the wall?
Peter
Added: I guess it is best to paint the cavity and take pictures with an IR camera to determine what modes it is running.
I think the difference is likely both, but more because of the connectors and coax inside the cavity. It is a fairly small cavity with a lot of clutter inside. Unless I modeled the exact fittings and curve of the coax, there is likely to be quite a difference between the sim and measured resonance.
My workaround to this problem was to fabricate the simplest antenna that could excite the TE modes and mount it close to the end-plate so only a very small portion of the connector extended into the cavity. This yielded measurements that were very close to the simulations.
-
#487 by dustinthewind on 04 Jul, 2017 15:31
-
https://phys.org/news/2017-06-atomic-mass-photon-momentum-paradox.html
QuoteIn a recent publication, Aalto University researchers show that in a transparent medium each photon is accompanied by an atomic mass density wave. The optical force of the photon sets the medium atoms in motion and makes them carry 92% of the total momentum of light, in the case of silicon.
The novel discovery solves the centennial momentum paradox of light. In the literature, there has existed two different values for the momentum of light in the transparent medium. Typically, these values differ by a factor of ten and this discrepancy is known as the momentum paradox of light. The difference between the momentum values is caused by neglecting the momentum of atoms moving with the light pulse.
Shell
That reminds me of the article where they measure light going into water and the back reaction it had when entering the water. It would be interesting to see the model of a reflection in a medium.
The Minkowski momentum then looks to be a wave where energy gets effectively transferred to the crystal lattice and it having a larger mass is more effective at transferring momentum upon reflection. That's why when they put the mirror inside water and measured the impulse from light, that impulse appears to be greater by a factor of n (refractive index = n).
Energy lost from the photon to the lattice must reduce the photons ability to transfer energy upon impulse by a factor of n.
The wave moving outward when light enters would Doppler shift the photon possibly? Upon exiting, the wave that moves with the light pulse could possibly re-deliver that energy lost from a red shift when the photon entered. Effectively blue shifting it back to its previous frequency.
If the photon did lose frequency/energy/effective_mass as it entered the lattice I wonder if that could explain its loss of ability to transfer momentum in the Abraham part of its momentum. Thanks for sharing Shell. -
#488 by Peter Lauwer on 04 Jul, 2017 15:40
-
Again, thanks a lot for the sims. But the frequencies you got out of it, 3.50 and 3.35 GHz, do not really come close to the measured frequencies (3.60 and 3.25 GHz).
Does it due to, apart from the influence of the connectors etc., the fact that one of the endplates in not in contact with the wall?
Peter
Added: I guess it is best to paint the cavity and take pictures with an IR camera to determine what modes it is running.
I think the difference is likely both, but more because of the connectors and coax inside the cavity. It is a fairly small cavity with a lot of clutter inside. Unless I modeled the exact fittings and curve of the coax, there is likely to be quite a difference between the sim and measured resonance.
My workaround to this problem was to fabricate the simplest antenna that could excite the TE modes and mount it close to the end-plate so only a very small portion of the connector extended into the cavity. This yielded measurements that were very close to the simulations.
For the 'EMDrive frustum' that is a very good solution. But for the 'coupling cavity' it is too narrow bandwidth. In order to be usable, the coupling cavity needs to be rather broadband (at least a few MHz) since trough it, you have to feed the frustum (with a shifting resonance frequency due to temp change). -
#489 by SeeShells on 04 Jul, 2017 16:28
-
I agree with most of your assessment although if I may add a few thoughts on what I did and why, to provide a high power stable 2.45GHz narrow band frequency to the frustum.Again, thanks a lot for the sims. But the frequencies you got out of it, 3.50 and 3.35 GHz, do not really come close to the measured frequencies (3.60 and 3.25 GHz).
Does it due to, apart from the influence of the connectors etc., the fact that one of the endplates in not in contact with the wall?
Peter
Added: I guess it is best to paint the cavity and take pictures with an IR camera to determine what modes it is running.
I think the difference is likely both, but more because of the connectors and coax inside the cavity. It is a fairly small cavity with a lot of clutter inside. Unless I modeled the exact fittings and curve of the coax, there is likely to be quite a difference between the sim and measured resonance.
My workaround to this problem was to fabricate the simplest antenna that could excite the TE modes and mount it close to the end-plate so only a very small portion of the connector extended into the cavity. This yielded measurements that were very close to the simulations.
For the 'EMDrive frustum' that is a very good solution. But for the 'coupling cavity' it is too narrow bandwidth. In order to be usable, the coupling cavity needs to be rather broadband (at least a few MHz) since trough it, you have to feed the frustum (with a shifting resonance frequency due to temp change).
As most here know I did my own clean variable DC power supplies driving a thermally stabilized copper lined water jacket magnetron with an radiator heat exchanger. This drives a magnetron>antenna> cavity much like the one you did, but with a tuning endplate captured with a quartz rod through the center that allows thermal expansion in the resonate cavity. It maintains the frequency and keeps the mode locked. The magnetron has the ability to "lock" to the resonate frequency of the waveguide when driving this arrangement. This gave me a stable RF source that was stable, variable in power and some flexibility in frequency tuning.
I use the output of this waveguide to drive into a frustum that utilizes the same thermally stabilized design. A Quartz tuning rod for stabilizing the thermal expansion and contraction as the Drive cavity fills with RF. *attached image
My Very Best,
Shell
opps... 2.45GHz
-
#490 by Bob012345 on 04 Jul, 2017 17:58
-
https://phys.org/news/2017-06-atomic-mass-photon-momentum-paradox.html
QuoteIn a recent publication, Aalto University researchers show that in a transparent medium each photon is accompanied by an atomic mass density wave. The optical force of the photon sets the medium atoms in motion and makes them carry 92% of the total momentum of light, in the case of silicon.
The novel discovery solves the centennial momentum paradox of light. In the literature, there has existed two different values for the momentum of light in the transparent medium. Typically, these values differ by a factor of ten and this discrepancy is known as the momentum paradox of light. The difference between the momentum values is caused by neglecting the momentum of atoms moving with the light pulse.
Shell
That reminds me of the article where they measure light going into water and the back reaction it had when entering the water. It would be interesting to see the model of a reflection in a medium.
The Minkowski momentum then looks to be a wave where energy gets effectively transferred to the crystal lattice and it having a larger mass is more effective at transferring momentum upon reflection. That's why when they put the mirror inside water and measured the impulse from light, that impulse appears to be greater by a factor of n (refractive index = n).
Energy lost from the photon to the lattice must reduce the photons ability to transfer energy upon impulse by a factor of n.
The wave moving outward when light enters would Doppler shift the photon possibly? Upon exiting, the wave that moves with the light pulse could possibly re-deliver that energy lost from a red shift when the photon entered. Effectively blue shifting it back to its previous frequency.
If the photon did lose frequency/energy/effective_mass as it entered the lattice I wonder if that could explain its loss of ability to transfer momentum in the Abraham part of its momentum. Thanks for sharing Shell.
Why can't a ship in principle create a disposable, unattached local medium with huge index of refraction, say 1E8 and emit laser light or microwaves in it thus imparting a billion times the kick to the ship over a photon beam in free space. One has to create and release the medium but if it's a low density gas in the form of a Bose Einstein Condensate, it wouldn't be much material. Alternatively, we could seek some other method of changing the index of refraction of space using only energy. -
#491 by tchernik on 04 Jul, 2017 20:45
-
Why can't a ship in principle create a disposable, unattached local medium with huge index of refraction, say 1E8 and emit laser light or microwaves in it thus imparting a billion times the kick to the ship over a photon beam in free space. One has to create and release the medium but if it's a low density gas in the form of a Bose Einstein Condensate, it wouldn't be much material. Alternatively, we could seek some other method of changing the index of refraction of space using only energy.
If the differences of Minkowski's and Abrahams' momenta of light have anything to do with the Emdrive, the transparent medium (the probably ionized gas) inside of the cavity would be a significant part of the effect.
Which means an Emdrive on a near full vacuum tested here on Earth would be noticeably less efficient than one on an atmosphere, but probably not null, because in any "vacuum" we can make on Earth there are some traces of gas left and the cavity could sustain some out-gassing during its function.
But if we tested it in a near perfect and self-replenishing vacuum (like in deep space), there the thrust could perfectly go to zero or become negligible/undetectable. Unless we were smart and prepared some test article where the cavity is airtight and contains some inert gas. -
#492 by Slyver on 05 Jul, 2017 01:29
-
If the differences of Minkowski's and Abrahams' momenta of light have anything to do with the Emdrive, the transparent medium (the probably ionized gas) inside of the cavity would be a significant part of the effect.
Which means an Emdrive on a near full vacuum tested here on Earth would be noticeably less efficient than one on an atmosphere, but probably not null, because in any "vacuum" we can make on Earth there are some traces of gas left and the cavity could sustain some out-gassing during its function.
But if we tested it in a near perfect and self-replenishing vacuum (like in deep space), there the thrust could perfectly go to zero or become negligible/undetectable. Unless we were smart and prepared some test article where the cavity is airtight and contains some inert gas.
I have been an advocate of an EMDrive container system that is designed to have an ~uniform thermal radiation signature, and is hermetically sealed, since the beginning. If we don't know why (or even if) it is working, and taking out a possibly functional piece (air) of the initially measured system decreases the thrust measurement by a substantial amount, test the whole system, which includes gas on both the inside and outside. To me, this is a no-brainer.
No one has done this, and it is for this reason alone I keep considering spending the required time to perform this test myself. I keep not doing it, simply because I cannot find the time, but I really want to see this test.
Edit: What I mean to advocate is; get a working system, and then put it in a hermetically sealed container and see if the thrust signal remains.
I would steer clear of this isolated feed approach to testing the EMdrives due to the complaint that if any element of the RF source is mounted in the laboratory frame of reference, the argument can be made that any unbalanced forces developed by the frustum are just leveraged off the RF power supply and its mounts to the lab via its RF feed lines. The only convincing way to demonstrate these EMdrives is to treat them as "free flyers" with the controls, RF source and battery flying WITH the frustum as they would in free space. That recommendation came out of the July 2014 Eagleworks (EW) Blue Ribbon PhD panel and that was the primary reason we built the Integrated Copper Frustum Test Article (ICFTA) and Cavendish Balance test article the way we did.
Best, Paul M.
Dear Paul,
I don't really see this problem (at least, not as being a big problem).
Of course, you have to perform all kind of tests to show this way of feeding does not impose disturbing forces itself: dummy-load instead of frustum, cylindrical cavity instead of frustum, etc.
But I see everyone struggling with the 100+ W of heat generated on the measurement device when only measuring some tens of micronewtons. It is just not convincing. I think the problems are much less with this 'isolated feed approach' (and the 'convincing power' stronger).
Anyway, it is how I am going to do it. It will surely be useful if several ways are tried.
There is merit to performing tests with the microwave cavity in a different frame from the microwave source; however, until a test is performed that has the entire system on one side of a torsion balance producing meaningful thrust levels, either in vacuum, or hermetically sealed, I do not think it will be sufficiently convincing for space testing. Any efforts to characterize such a system to a sufficient degree to be convincing are less than the efforts required to put the whole thing on a rotating rig.
In no way do I mean to discourage such testing, it has great merit in development. However, as has been shown, it has limitations in convincing power (with good reason).
All that matters is whether or not this system produces an effective thrust greater than a photon rocket per energy input, without what we would consider an obvious reaction mass (such as ejected photons, ejected or thermally excited air, etc.). If it does this, it's... YUGE.
Give me a thermally uniform, hermitically sealed container all on one side of a rotating rig with 5 sigma thrust an order of magnitude greater than a photon rocket, and I will (well, someone will) give you a Nobel prize. It doesn't take much. An order of magnitude can change the world as we know it. -
#493 by dustinthewind on 05 Jul, 2017 07:29
-
If the differences of Minkowski's and Abrahams' momenta of light have anything to do with the Emdrive, the transparent medium (the probably ionized gas) inside of the cavity would be a significant part of the effect.
Which means an Emdrive on a near full vacuum tested here on Earth would be noticeably less efficient than one on an atmosphere, but probably not null, because in any "vacuum" we can make on Earth there are some traces of gas left and the cavity could sustain some out-gassing during its function.
But if we tested it in a near perfect and self-replenishing vacuum (like in deep space), there the thrust could perfectly go to zero or become negligible/undetectable. Unless we were smart and prepared some test article where the cavity is airtight and contains some inert gas.
I have been an advocate of an EMDrive container system that is designed to have an ~uniform thermal radiation signature, and is hermetically sealed, since the beginning. If we don't know why (or even if) it is working, and taking out a possibly functional piece (air) of the initially measured system decreases the thrust measurement by a substantial amount, test the whole system, which includes gas on both the inside and outside. To me, this is a no-brainer.
No one has done this, and it is for this reason alone I keep considering spending the required time to perform this test myself. I keep not doing it, simply because I cannot find the time, but I really want to see this test.
Edit: What I mean to advocate is; get a working system, and then put it in a hermetically sealed container and see if the thrust signal remains.
I would steer clear of this isolated feed approach to testing the EMdrives due to the complaint that if any element of the RF source is mounted in the laboratory frame of reference, the argument can be made that any unbalanced forces developed by the frustum are just leveraged off the RF power supply and its mounts to the lab via its RF feed lines. The only convincing way to demonstrate these EMdrives is to treat them as "free flyers" with the controls, RF source and battery flying WITH the frustum as they would in free space. That recommendation came out of the July 2014 Eagleworks (EW) Blue Ribbon PhD panel and that was the primary reason we built the Integrated Copper Frustum Test Article (ICFTA) and Cavendish Balance test article the way we did.
Best, Paul M.
Dear Paul,
I don't really see this problem (at least, not as being a big problem).
Of course, you have to perform all kind of tests to show this way of feeding does not impose disturbing forces itself: dummy-load instead of frustum, cylindrical cavity instead of frustum, etc.
But I see everyone struggling with the 100+ W of heat generated on the measurement device when only measuring some tens of micronewtons. It is just not convincing. I think the problems are much less with this 'isolated feed approach' (and the 'convincing power' stronger).
Anyway, it is how I am going to do it. It will surely be useful if several ways are tried.
There is merit to performing tests with the microwave cavity in a different frame from the microwave source; however, until a test is performed that has the entire system on one side of a torsion balance producing meaningful thrust levels, either in vacuum, or hermetically sealed, I do not think it will be sufficiently convincing for space testing. Any efforts to characterize such a system to a sufficient degree to be convincing are less than the efforts required to put the whole thing on a rotating rig.
In no way do I mean to discourage such testing, it has great merit in development. However, as has been shown, it has limitations in convincing power (with good reason).
All that matters is whether or not this system produces an effective thrust greater than a photon rocket per energy input, without what we would consider an obvious reaction mass (such as ejected photons, ejected or thermally excited air, etc.). If it does this, it's... YUGE.
Give me a thermally uniform, hermitically sealed container all on one side of a rotating rig with 5 sigma thrust an order of magnitude greater than a photon rocket, and I will (well, someone will) give you a Nobel prize. It doesn't take much. An order of magnitude can change the world as we know it.
A problem with testing a microwave source that is disconnected from the cavity, such that it can exist in another frame is that light would travel between the two frames. This would likely cause mutual repulsion between the two frames similar to a Photonic laser thruster: https://en.wikipedia.org/wiki/Photonic_laser_thruster . These have already been shown to work. Disconnecting the frames may fundamentally change what it is. -
#494 by Peter Lauwer on 05 Jul, 2017 07:44
-
I agree with most of your assessment although if I may add a few thoughts on what I did and why, to provide a high power stable 2.45GHz narrow band frequency to the frustum.Again, thanks a lot for the sims. But the frequencies you got out of it, 3.50 and 3.35 GHz, do not really come close to the measured frequencies (3.60 and 3.25 GHz).
Does it due to, apart from the influence of the connectors etc., the fact that one of the endplates in not in contact with the wall?
Peter
Added: I guess it is best to paint the cavity and take pictures with an IR camera to determine what modes it is running.
I think the difference is likely both, but more because of the connectors and coax inside the cavity. It is a fairly small cavity with a lot of clutter inside. Unless I modeled the exact fittings and curve of the coax, there is likely to be quite a difference between the sim and measured resonance.
My workaround to this problem was to fabricate the simplest antenna that could excite the TE modes and mount it close to the end-plate so only a very small portion of the connector extended into the cavity. This yielded measurements that were very close to the simulations.
For the 'EMDrive frustum' that is a very good solution. But for the 'coupling cavity' it is too narrow bandwidth. In order to be usable, the coupling cavity needs to be rather broadband (at least a few MHz) since trough it, you have to feed the frustum (with a shifting resonance frequency due to temp change).
As most here know I did my own clean variable DC power supplies driving a thermally stabilized copper lined water jacket magnetron with an radiator heat exchanger. This drives a magnetron>antenna> cavity much like the one you did, but with a tuning endplate captured with a quartz rod through the center that allows thermal expansion in the resonate cavity. It maintains the frequency and keeps the mode locked. The magnetron has the ability to "lock" to the resonate frequency of the waveguide when driving this arrangement. This gave me a stable RF source that was stable, variable in power and some flexibility in frequency tuning.
I use the output of this waveguide to drive into a frustum that utilizes the same thermally stabilized design. A Quartz tuning rod for stabilizing the thermal expansion and contraction as the Drive cavity fills with RF. *attached image
My Very Best,
Shell
opps... 2.45GHz
This looks like a very elegant setup. Neat copper work as well.
But is the microwave signal also coupled in a non-contact way? -
#495 by Peter Lauwer on 05 Jul, 2017 07:47
-
A problem with testing a microwave source that is disconnected from the cavity, such that it can exist in another frame is that light would travel between the two frames. This would likely cause mutual repulsion between the two frames similar to a Photonic laser thruster: https://en.wikipedia.org/wiki/Photonic_laser_thruster . These have already been shown to work. Disconnecting the frames may fundamentally change what it is.
Really? Photonic thrust - recalling from mind - gives you only 3.3 nanonewton/watt. Not something to worry about, isn't it? -
#496 by Monomorphic on 05 Jul, 2017 13:17
-
A problem with testing a microwave source that is disconnected from the cavity, such that it can exist in another frame is that light would travel between the two frames. This would likely cause mutual repulsion between the two frames similar to a Photonic laser thruster: https://en.wikipedia.org/wiki/Photonic_laser_thruster . These have already been shown to work. Disconnecting the frames may fundamentally change what it is.
Really? Photonic thrust - recalling from mind - gives you only 3.3 nanonewton/watt. Not something to worry about, isn't it?
Depends on the number of reflections. With the photonic laser thruster, 1000 reflections of 1W comes to 3.3uN. -
#497 by Peter Lauwer on 05 Jul, 2017 14:09
-
A problem with testing a microwave source that is disconnected from the cavity, such that it can exist in another frame is that light would travel between the two frames. This would likely cause mutual repulsion between the two frames similar to a Photonic laser thruster: https://en.wikipedia.org/wiki/Photonic_laser_thruster . These have already been shown to work. Disconnecting the frames may fundamentally change what it is.
Really? Photonic thrust - recalling from mind - gives you only 3.3 nanonewton/watt. Not something to worry about, isn't it?
Depends on the number of reflections. With the photonic laser thruster, 1000 reflections of 1W comes to 3.3uN.
But that's not very likely the case in this setup, 1000 reflections, isn't it? Actually, what are we talking about in my setup? Reflections between what?
Even in case of the cited plt I don't believe in 1k reflections at astronomical distances. No way, not even 100, not even 10. -
#498 by SeeShells on 05 Jul, 2017 14:41
-
Thanks for the complements, I try.
I agree with most of your assessment although if I may add a few thoughts on what I did and why, to provide a high power stable 2.45GHz narrow band frequency to the frustum.Again, thanks a lot for the sims. But the frequencies you got out of it, 3.50 and 3.35 GHz, do not really come close to the measured frequencies (3.60 and 3.25 GHz).
Does it due to, apart from the influence of the connectors etc., the fact that one of the endplates in not in contact with the wall?
Peter
Added: I guess it is best to paint the cavity and take pictures with an IR camera to determine what modes it is running.
I think the difference is likely both, but more because of the connectors and coax inside the cavity. It is a fairly small cavity with a lot of clutter inside. Unless I modeled the exact fittings and curve of the coax, there is likely to be quite a difference between the sim and measured resonance.
My workaround to this problem was to fabricate the simplest antenna that could excite the TE modes and mount it close to the end-plate so only a very small portion of the connector extended into the cavity. This yielded measurements that were very close to the simulations.
For the 'EMDrive frustum' that is a very good solution. But for the 'coupling cavity' it is too narrow bandwidth. In order to be usable, the coupling cavity needs to be rather broadband (at least a few MHz) since trough it, you have to feed the frustum (with a shifting resonance frequency due to temp change).
As most here know I did my own clean variable DC power supplies driving a thermally stabilized copper lined water jacket magnetron with an radiator heat exchanger. This drives a magnetron>antenna> cavity much like the one you did, but with a tuning endplate captured with a quartz rod through the center that allows thermal expansion in the resonate cavity. It maintains the frequency and keeps the mode locked. The magnetron has the ability to "lock" to the resonate frequency of the waveguide when driving this arrangement. This gave me a stable RF source that was stable, variable in power and some flexibility in frequency tuning.
I use the output of this waveguide to drive into a frustum that utilizes the same thermally stabilized design. A Quartz tuning rod for stabilizing the thermal expansion and contraction as the Drive cavity fills with RF. *attached image
My Very Best,
Shell
opps... 2.45GHz
This looks like a very elegant setup. Neat copper work as well.
But is the microwave signal also coupled in a non-contact way?
I tried both, waveguides to the frustum and stress relieved high flex coax cabling with the Torsion Pendulum design.
I used a larger air gap non-contact waveguide freely rotating center much like what you did and a more compact design (which proved too demanding in machining with my current tools). I plan to revisit the tiny design when I get my lathe.
Both designs tended to leak some at the air gap which I don't like and had to put a small local floating Faraday cage around them. On another note the power levels in both designs would vary to the frustum during rotations of the drive on the wire pendulum. This was due the free hanging arm on the wire and imperfections in the cavity build, antenna etc made the torsion arm flex around and the floating top plate could hit the sides.
Where I am now.
Advanced designs incorporated into the drive required more than just the RF source to the frustum, those required a battery source that "rides" on the pendulum arm. This dramatically increased the load bearing requirements of the pendulum wire and the complexity of the build to the point that it became a Rube Goldberg nightmare and not realistic. Unlike monormorphic who is pursuing 1-25 watts in a pure RF drive I need more power. (Love that line
)
I'd recommend those who are building a device review what the industry recommends in evaluating micro-thrusters in several test beds to achieve credible results and not have to reinvent the wheel.
PDF warning...
http://hpepl.ae.gatech.edu/papers/2013_IEPC_Polk.pdf
Because I've seen more than just the small m/N thrust anomalies that need to be characterized I'm rebuilding the hanging wire torsion pendulum by using flexure bearings.
I have several sets left over from building semiconductor equipment that required extreme precision rotational capabilities.
When I was building a 4 axis machine a couple years ago.
By luck I still had a few of the parts left over. I'm using them to build the flexure torsional pendulum with some mods.
This will allow measurements into the µN- to mN-level thrust and uto N-level impulses and provide support to the higher power requirements I need.
Also I could use Galinstan, a Gallium/Indium/Tin contacts for sensors and still use the high power non-contact microwave waveguide. But I believe I have a better fix to get away from the Galinstan issues and using it altogether.
This isn't a air bearing rotational stand (which has issues and is costly) or a Torsional Wire Pendulum (which reached limits in the design). The Flexure bearing design follows many of the industry guidelines for thruster testing and gives me the wide range of measurements I need.
My Very Best,
Shell
-
#499 by mikegem on 05 Jul, 2017 16:03
-
I agree with most of your assessment although if I may add a few thoughts on what I did and why, to provide a high power stable 2.45GHz narrow band frequency to the frustum.Again, thanks a lot for the sims. But the frequencies you got out of it, 3.50 and 3.35 GHz, do not really come close to the measured frequencies (3.60 and 3.25 GHz).
Does it due to, apart from the influence of the connectors etc., the fact that one of the endplates in not in contact with the wall?
Peter
Added: I guess it is best to paint the cavity and take pictures with an IR camera to determine what modes it is running.
I think the difference is likely both, but more because of the connectors and coax inside the cavity. It is a fairly small cavity with a lot of clutter inside. Unless I modeled the exact fittings and curve of the coax, there is likely to be quite a difference between the sim and measured resonance.
My workaround to this problem was to fabricate the simplest antenna that could excite the TE modes and mount it close to the end-plate so only a very small portion of the connector extended into the cavity. This yielded measurements that were very close to the simulations.
For the 'EMDrive frustum' that is a very good solution. But for the 'coupling cavity' it is too narrow bandwidth. In order to be usable, the coupling cavity needs to be rather broadband (at least a few MHz) since trough it, you have to feed the frustum (with a shifting resonance frequency due to temp change).
As most here know I did my own clean variable DC power supplies driving a thermally stabilized copper lined water jacket magnetron with an radiator heat exchanger. This drives a magnetron>antenna> cavity much like the one you did, but with a tuning endplate captured with a quartz rod through the center that allows thermal expansion in the resonate cavity. It maintains the frequency and keeps the mode locked. The magnetron has the ability to "lock" to the resonate frequency of the waveguide when driving this arrangement. This gave me a stable RF source that was stable, variable in power and some flexibility in frequency tuning.
I use the output of this waveguide to drive into a frustum that utilizes the same thermally stabilized design. A Quartz tuning rod for stabilizing the thermal expansion and contraction as the Drive cavity fills with RF. *attached image
My Very Best,
Shell
opps... 2.45GHz
A thought from somebody who has built high power microwave systems for CVD.
I noticed in the figures your sliding plate with flexible beryllium gasket at its periphery. If this is a finger stock type of seal, and if it is at a point where skin currents are large, local high temperatures can oxidize the finger stock elements. Then high resistance, higher temperatures, then local finger melting. Also, copper is inherently "sticky" and tends to score or gall with finger stock motion.
To eliminate these effects in moving finger stock seals engaged with both copper and aluminum walls and in stub tuners, I used a thin layer of silver paste spread over the expected range of motion. Silver is electrically conductive and inherently lubricious. It eliminated scoring, galling, and the other rubbing phenomena that caused my finger stock seals and tuning stubs to fail.
Eventually, I silver-plated every surface that carried skin currents. Problem solved. But the paste was an effective interim solution.
If your flexible beryllium seal is a non-finger stock design, what I've written may be irrelevant, please ignore.
?
)