EM Drive Developments - related to space flight applications - Thread 7

[rfmwguy]

This is a thread - Thread 7 in the series - focused on objective analysis of whether the EM Drive (a cavity resonating at microwave frequencies) reported "thrust force" is an experimental artifact or whether it is a real propulsion effect  that can be used for space applications, and if so, in discussing those possible space propulsion applications.

Objective skeptical inquiry is strongly welcome.   Disagreements should be expressed politely, concentrating on the technical, engineering and scientific aspects, instead of focusing on people.   As such, the use of experimental data, mathematics, physics, engineering, drawings, spreadsheets and computer simulations are strongly encouraged, while subjective wordy statements are discouraged. Peer-reviewed information from reputable journals is strongly encouraged.  Please acknowledge the authors and respect copyrights.

Commercial advertisement is discouraged.

In order to minimize bandwidth and  maximize information content, when quoting, one can use an ellipsis (...) to indicate the clipped material.

Only use the embed [img ]http://code when the image is small enough to fit within the page. Anything wider than the width of the page makes the page unreadable as it stretches it (we're working on auto reduction, but different browsers work different ways, etc.)

This link

http://math.typeit.org/

enables typing of mathematical symbols, including differentiation and integration, Greek letters, etc.

--

Links to previous threads:

Thread 1:
http://forum.nasaspaceflight.com/index.php?topic=29276.0

Thread 2:
http://forum.nasaspaceflight.com/index.php?topic=36313.0

Thread 3:
http://forum.nasaspaceflight.com/index.php?topic=37642.0

Thread 4:
http://forum.nasaspaceflight.com/index.php?topic=38203.0

Thread 5:
http://forum.nasaspaceflight.com/index.php?topic=38577.0

Thread 6:
http://forum.nasaspaceflight.com/index.php?topic=39004.0

--

Entry level thread:
http://forum.nasaspaceflight.com/index.php?topic=37438.0

Baseline NSF Article:
http://www.nasaspaceflight.com/2015/04/evaluating-nasas-futuristic-em-drive/

This is the link to the EM Drive wiki that users are encouraged to contribute to, edit for accuracy, and build as a knowledge resource for the EM Drive:

http://emdrive.wiki
http://rfdriven.com

Chris note: Please note all posts need to be useful and worthwhile or they will be removed via moderation. This subject has large interest, with over 3.5 million thread reads and 850,000 article reads. Most people are reading and not posting, so when you post it is in front of a very large audience.

Also, and it should go without saying, amateur experiments are discouraged unless you have gained educated and/or professional advice for safety reasons.

For clarity, the following is an (evolving) chart which attempts to place the "EMDrive" within a larger class of propellantless propulsion concepts. Special thanks to Dr. Jose Rodal.

[rfmwguy]

Welcome to T7, let the conversations continue!

[Rodal]

Re: EM Drive Developments - related to space flight applications - Thread 6

Thought about this a lot. Incidental reflections off the side wall don't seem nearly as critical as the endplate reflections. For me, just as in optics, the flatness, thickness and parallel alignment of the endplates is the critical factor.

Of course, I'm 1 DIYer and the beauty of this thing is there is no wrong way to build it; as the right way has not yet been formally disclosed...just hinted at.

It is arguably contrary to present knowledge that <<reflections off the side wall don't seem nearly as critical as the endplate>>.

This can be  shown using the well-known Snell's laws and the coefficients for reflection, as used in microwave engineering for countless applications.

It also runs counter to all the simulations shown in previous EM Drive threads, using:

Meep
Comsol
Feko

The recent analysis using FEKO by Monomorphic being most pertinent in this respect.

[TheTraveller]

Re: EM Drive Developments - related to space flight applications - Thread 6

Thought about this a lot. Incidental reflections off the side wall don't seem nearly as critical as the endplate reflections. For me, just as in optics, the flatness, thickness and parallel alignment of the endplates is the critical factor.

Of course, I'm 1 DIYer and the beauty of this thing is there is no wrong way to build it; as the right way has not yet been formally disclosed...just hinted at.

It is incorrect to state that <<reflections off the side wall don't seem nearly as critical as the endplate>>.

This can be trivially shown using the well-known Snell's law and the coefficients for reflection, as used in microwave engineering for countless applications.

It also runs counter to all the simulations shown in previous EM Drive threads, using:

Meep
Comsol
Feko

The recent analysis using FEKO by Monomorphic being most pertinent in this respect.

As I see it, if the end plates are not highly parallel and at a right angle to the frustum length axis, the reflections will walk off toward the widest separation of the end plates and bunch up against the side wall at that point.

It may be possible to put a very slight concave curve into each end plate, which will try to stop the walk off even if the end plates are not highly aligned. Or maybe not.

[TheTraveller]

Re: EM Drive Developments - related to space flight applications - Thread 6

Thought about this a lot. Incidental reflections off the side wall don't seem nearly as critical as the endplate reflections. For me, just as in optics, the flatness, thickness and parallel alignment of the endplates is the critical factor.

Of course, I'm 1 DIYer and the beauty of this thing is there is no wrong way to build it; as the right way has not yet been formally disclosed...just hinted at.

It is incorrect to state that <<reflections off the side wall don't seem nearly as critical as the endplate>>.

This can be trivially shown using the well-known Snell's law and the coefficients for reflection, as used in microwave engineering for countless applications.

It also runs counter to all the simulations shown in previous EM Drive threads, using:

Meep
Comsol
Feko

The recent analysis using FEKO by Monomorphic being most pertinent in this respect.

All the sims are done with perfect geometry and alignment. Maybe misalign / non parallel the end plates in the sim and see what is produced?

[Rodal]

Quote from: Rodal on Today at 01:40 PM
...

It is trivial  to show (with equations routinely used by Aerospace Engineers to design thin shell structures for rockets and spacecraft) that a shell structure made of copper 2 mm thick and 0.2 m length and diameter, would be easy to deform by more than 13 micrometers just by handling.  (Several DIY reported early on that they could change the natural frequency of their EM Drive DIY just by squeezing the EM Drive with their hands).

So, if Shawyer really means that a tolerance of 13 micrometers is necessary, he must have used much thicker (than 2 mm) copper shell structure for the Flight Thruster.

If anything even close to 13 micrometers tolerance is needed, it is easy to show that the thickness of copper used by rfmwguy, SeeShells, Berca,  NASA and others for the EM Drive conical walls are way too thin.

Also, it is well known by Aerospace Engineers for a long time that to provide shell structure stiffness with minimum weight, this can be done with a sandwich wall construction, for example using honeycomb between fascia, for example, such that the metal (aluminum or copper or silver) fascia are thin and such that shear stiffness is provided by the honeycomb and bending stiffness is provided by the distance between the metal (aluminum or copper or silver) fascia in the sandwich wall construction.

If weight savings is not an issue, it should be easier to provide high tolerance for roundness by machining techniques of a thick metal form, than to achieve high tolerance using a very thin metal fascia.

[rfmwguy]

Re: EM Drive Developments - related to space flight applications - Thread 6

Thought about this a lot. Incidental reflections off the side wall don't seem nearly as critical as the endplate reflections. For me, just as in optics, the flatness, thickness and parallel alignment of the endplates is the critical factor.

Of course, I'm 1 DIYer and the beauty of this thing is there is no wrong way to build it; as the right way has not yet been formally disclosed...just hinted at.

It is incorrect to state that <<reflections off the side wall don't seem nearly as critical as the endplate>>.

This can be trivially shown using the well-known Snell's law and the coefficients for reflection, as used in microwave engineering for countless applications.

Sidewall deformations are not ideal, but I (just me) am speculating that the fields produced by the alignment of the endplates are more critical than sidewall perfection.

If a sidewall reflection eventually leads to the large plate in a conical scheme, I surmise the large and small plates are what needs to be aligned perfectly. A dented microwave horn will still resonate and radiate, just produce a slightly distorted pattern.

Others are welcome to their own opinions, of course. When it comes to emdrive, I've learned there are few absolutes 

[Rodal]

Re: EM Drive Developments - related to space flight applications - Thread 6

Thought about this a lot. Incidental reflections off the side wall don't seem nearly as critical as the endplate reflections. For me, just as in optics, the flatness, thickness and parallel alignment of the endplates is the critical factor.

Of course, I'm 1 DIYer and the beauty of this thing is there is no wrong way to build it; as the right way has not yet been formally disclosed...just hinted at.

It is incorrect to state that <<reflections off the side wall don't seem nearly as critical as the endplate>>.

This can be trivially shown using the well-known Snell's law and the coefficients for reflection, as used in microwave engineering for countless applications.

Sidewall deformations are not ideal, but I (just me) am speculating that the fields produced by the alignment of the endplates are more critical than sidewall perfection.

If a sidewall reflection eventually leads to the large plate in a conical scheme, I surmise the large and small plates are what needs to be aligned perfectly. A dented microwave horn will still resonate and radiate, just produce a slightly distorted pattern.

Others are welcome to their own opinions, of course. When it comes to emdrive, I've learned there are few absolutes 

In your present proposed experiment I understand that you are going with much thicker end plates than the thickness of the copper used for the EM Drive walls.

The thin copper used for your sidewalls is easy to deform out-of-round.

You can perform experiments, if you like, measuring the Q quality of resonance, with EM Drive's made with sidewalls that are not round, to convince yourself that to have a high Q you must have round cross-sections and hence how important it is to have round sidewalls for reflection purposes and hence for high Q.


Alternatively, the effect on Q of having out-of-round sidewalls vs. out-of-flat end plates can be analyzed with numerical methods like Meep, Comsol or FEKO.


*************

Concerning the EM Drive "experiments" this discussion reveals once again how un-rigorous have been the reports on EM Drive experiments up to now:

* Yang has not provided dimensions, we have no idea what is the thickness and dimensions she used

* of institutional reports only NASA has provided full information on dimensions, including thickness

[rfmwguy]

I've learned a lot about metal working the past couple of weeks, regarding flatness. If it looks flat...it really isn't.

The 1/8 copper disk (small diameter) looked great from the supplier until I put it on a lapping plate with a backing disk. Progressing from 60 to 2000 grit sandpaper, I was amazed at the material I had to remove to make it close to "flat".

At 2000 grit last night, serious striations are gone but there is still a micron level "unflatness" even after about 4 hours of hand sanding. What this taught me is there is almost no way DIYers can ever achieve perfection in sidewalls. I agree its desirable to maintain shape and symmetry, but I think its best not to worry about it too much. A fully machined and polished frustum is well beyond our budgets. I received a quote for leveling and polishing the small endplate and it was $675 

[RERT]

TT has reported Roger Shawyer's remarks on required tolerances, but we don't know what Shawyer's goals are.

In other words, if he's trying to make something with extremely high Q (presumably for high thrust) his comments might make sense.

Even if he is right, that doesn't really speak to whether DIY builds can generate measurable thrust with less demanding tolerances.

Maybe TT has more context for RS's remarks which he can share.

R.

[TheTraveller]

TT has reported Roger Shawyer's remarks on required tolerances, but we don't know what Shawyer's goals are.

In other words, if he's trying to make something with extremely high Q (presumably for high thrust) his comments might make sense.

Even if he is right, that doesn't really speak to whether DIY builds can generate measurable thrust with less demanding tolerances.

Maybe TT has more context for RS's remarks which he can share.

R.

Roger stated with the tolerance quoted, if should be possible to achieve around 75% of the theoretical max Q and to do it in an affordable manner.

... If you aim for something like 75% of theoretical Q you will still get viable levels of thrust within a reasonable budget ....

I suggest that someone with a lot more sim model making experience than myself can start to model how far out of parallel the end plates must be to start to get degraded Q and mode.

[Rodal]

...

I suggest that someone with a lot more sim model making experience than myself can start to model how far out of parallel the end plates must be to start to get degraded Q and mode.

The numerical simulations would have to be conducted with much, much finer mesh that used up to now.

The simulations that have been shown in this thread using Meep, and FEKO have a mesh too coarse to accurately show the effect on Q of such tolerance.  The number of nodes is way to small to get close to the tolerances being discussed.   It doesn't appear that this can be really done with the computer resources they have available (certainly they cannot do it to show anything close to the 13 micrometers tolerance being asked by Shawyer).

Still, simulations could show that the side walls are just as important or more important than the end plates, since I really question rfmwguy's assertion regarding that the end plate tolerance is more important than the roundness of the side walls.

[rfmwguy]

Good to see you back Phil...Yes, I'd love to see the modeling of endplate tolerances...a mm off may drastically change things. Wondering if all feko/meep/comsol models have been assuming perfect parallel plates.

[Monomorphic]

Thanks for reminding me about the end-plate to end-plate bolts used in Shawyers emdrive. Compressing the frustum is an easy way to tweak the internal dimensions by small fractions. I'm using aluminum plate here to help ad rigidity to the end plates. I think others are using ceramic.

[TheTraveller]

From my calcs of the internals of the Flight Thruster as below:

BD: 0.2314m
SD: 0.1257m
Len: 0.1386m

the internal area is:  0.137670m^2. Stated mass: 2.92kg

Assuming no flanges and end plates the diameter of the end of the frustum side walls and constant thickness overall, the mass is then:

2mm Copper: 2.46kg
7mm Aluminium: 2.51kg

A 7mm thick end plate and side wall Aluminium Flight Thruster should be fairly stable.

[TheTraveller]

Thanks for reminding me about the end-plate to end-plate bolts used in Shawyers emdrive. Compressing the frustum is an easy way to tweak the internal dimensions by small fractions. I'm using aluminum plate here to help ad rigidity to the end plates. I think others are using ceramic.

Those external bolts, nuts and attachment flanges are there for a reason.

[zellerium]

I designed a frustum that should be able to be made within at least .001 in on an auto lathe. Attached is the drawing for the TE113 mode.
Unfortunately this method get much more pricey for a larger frustum because its turned from a solid piece of Al, but a 9 inch diameter 9 inch tall slug of 6061 should run ~$500 to $600.
 
A few updates I haven't made to these drawings since latest sims:
Aperture size: .875 in by 1.005 in
top radius 1.4 in
bottom radius 4.4 in
waveguide input should be at center, not 1/4 length

Also, I haven't really figured out how to secure the waveguide without welding it in case it needs to be replaced. Anyone have a simple solution?

[aero]

Good to see you back Phil...Yes, I'd love to see the modeling of endplate tolerances...a mm off may drastically change things. Wondering if all feko/meep/comsol models have been assuming perfect parallel plates.

Yes, all of my meep simulations assume perfect parallel plates.

For a cavity 0.2 m in length, with resolution of 200 gives 1000 voxels per meter. That is one mm resolution and turns out to be a computer challenging but doable resolution. For meep, both cpu cycles and memory requirements preclude directly simulating 10 micron resolution at 2.45 GHz.

Perhaps someone can imagine an effective approximation.

[SeeShells]

Thanks for reminding me about the end-plate to end-plate bolts used in Shawyers emdrive. Compressing the frustum is an easy way to tweak the internal dimensions by small fractions. I'm using aluminum plate here to help ad rigidity to the end plates. I think others are using ceramic.

Yes, I'm using ceramic plates <.05um flatness, they provide these very critical build and design points.

Force the endplate flat. The copper 101 O2 free .032" from the factory is better then .3um flat. (best I can measure) When bonded I don't need to do extreme lapping to get high flatness and plus it makes it easier to lap.

Keep deformations to a minimum even during a build from soldering or handling.

They maintain the flatness of the plates during cavity heating.

They make sure the plates stay together via the Quartz rod during tune and maintain parallelism during runs and thermal heating.

Allow the sidewalls expand up and past the small endplate which isn't captured but has a beryllium gasket.

Makes a stiffer overall frustum.

Shell

[rfmwguy]

I designed a frustum that should be able to be made within at least .001 in on an auto lathe. Attached is the drawing for the TE113 mode.
Unfortunately this method get much more pricey for a larger frustum because its turned from a solid piece of Al, but a 9 inch diameter 9 inch tall slug of 6061 should run ~$500 to $600.
 
A few updates I haven't made to these drawings since latest sims:
Aperture size: .875 in by 1.005 in
top radius 1.4 in
bottom radius 4.4 in
waveguide input should be at center, not 1/4 length

Also, I haven't really figured out how to secure the waveguide without welding it in case it needs to be replaced. Anyone have a simple solution?

Fantastic dwgs! THIS is what I've been looking for and most familiar visual with from back in the day. Guess I've gotta get a student version somehow 

Waveguide launch into the side is quite challenging. Shell might have the answer. Consider this: direct injection of magnetron monopole requires 1 hole and an adapter ring that can be pressure fitted. A 4 point frame would be needed to mount magnetron assembly.

Think it might be easier than a waveguide weld.