Author Topic: EM Drive Developments - related to space flight applications - Thread 4  (Read 1245621 times)

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275


The Eggcelerator
« Last Edit: 08/09/2015 12:07 am by deltaMass »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5911
  • USA
  • Liked: 6124
  • Likes Given: 5564
I love this thread and am totally addicted, can't put it down.  Thanks for all the work and information!  Discussions about Q, oxidation, broad / narrow spectrum and the shape of the ends have got me thinking about an alternate configuration.  It has many of the same characteristics but might be slightly easier to build / find parts / work with.  It is an ordinary cylinder with a cone inside.  Maybe something to consider once folks are enjoying consistent thrust (lifting small objects and pets, etc...  ;)

Keep up the great work!

WELCOME to the thread.  :)

That looks similar to an early patent by Shawyer, as I recall Shawyer had a cylinder with a dielectric insert inside it in the shape of a cone.  Maybe somebody can find it.
« Last Edit: 08/09/2015 12:08 am by Rodal »

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2442
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 3186
  • Likes Given: 2708
Todd



I opted out of a summer end party tonight mainly because Monday is a milestone birthday. Think I will celebrate with the sense that the next generation will question everything and challenge the status quo. Here's to nonconformity...

Many years ago my first electronics teacher upon graduation from high school and before I set off to collage told me something I thought was kind of silly (you know 17 and so) HE said: "The is the golden age of mediocre conformity, never let yourself be trapped".  As I got older I realized what a gift of wisdom he had said.

Happy Birthday to a very non-mediocre person. :)

Shell

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2442
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 3186
  • Likes Given: 2708
Something has been bothering me since last night and I couldn't help but watching the modes change and flip in meep.

How can you calculate seriously any Q in a cavity that simply jumps around from one T mode decaying and building into another in such a short time? And they all do it, every simulation with varying speeds.

Just something to mull over on a day away from the shop.

Shell

Added: Back to lurk mode and I'll be quiet.

Excellent question.  I hope that you don't shy away from asking such questions because it is only this way that one can understand what is being output.

My understanding (aero to confirm) is that the quality factor is calculated by aero using the routine Harminv.

Please notice that Meep has this disclaimer for using Hamrinv to calculate the quality factor Q:

http://ab-initio.mit.edu/wiki/index.php/Meep_Reference#Harminv

Quote
Important: normally, you should only use harminv to analyze data after the sources are off. Wrapping it in (after-sources (harminv ...)) is sufficient.

Thus, the Quality factor and the frequency (also obtained by Harminv) should properly be obtained after the sources are OFF, not when they are on. 

This takes us back to the whole discussion about Q quality factor in experiments.  The problem is not only how to best experimentally measure and report Q, but it seems not ideal to me to discuss and report a Q (as first done by Shawyer and then imitated by all other EM Drive experimenters) with the source ON.

It seems to me that proper measurement of Q also in experiments should be done upon turning the source off and examining the decay (as posted by Frobnicat in a separate post).

With the sources OFF, the definition of Q (inverse to damping) is well-posed.

With the sources ON, the meaning of Q is tricky.  It seems to me that when people are measuring Q with the sources on they are assuming a well-posed problem with "nice" properties (symmetry, etc.) that may not be fulfilled.  This is particularly contradictory with TheTraveller: who rejects Finite Element solutions (using COMSOL or ANSYS) and exact solutions of the problem saying that only a solution that calculates a force can properly calculate the frequency and at the same time Shawyer "measures" Q with the RF feed ON which assumes a well posed nice solution amenable to presentation of Q as if it would be the same Q as the one calculated with the RF feed off.

The way that the Q is being measured experimentally by Shawyer and others using S11 and the 3db width is similar to a common method in structural vibration analysis.  Of course, this presupposes a steady-state response.  In reality with the RF feed ON, the measurement may be a transient instead.

??????????????????????????????????????????????????????????

QUESTION TO aero:  have you been calculating the quality factor Q with Harminv, and if so, have you been doing so with the sources ON or OFF.  Have you been wrapping with  (after-sources (harminv ...)) ?

I've been researching this today and I've about reached one days brain drain.

Dr. Rodal, it's not only in meep that it's happening it's in the cavity as well, under power and I know It's driven by Maxwell's equations for a specific time set with in a finite computational defined area.

(http://ab-initio.mit.edu/wiki/index.php/Meep)
Resonant modes and frequencies — by analyzing the response of the system to a short pulse, one can extract the frequencies, decay rates, and field patterns of the harmonic modes of a system (including waveguide and cavity modes, and including losses).
--------------------
I feel there is much more going on that meep cannot show because of some of the inherent limitations. What I'd like to see is a round tube ends capped off as a resonate chamber with a RF input, measure the Q in meep.  Would you look at the CSV files from aero and do a simple compare of the Q and stress values between the two?

Maybe it's nothing but this old gal has a bone and maybe you have another way of looking at it. I think simply the decaying and mode switching in the frustum are a great red flag.

Back to lurk
Shell

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5911
  • USA
  • Liked: 6124
  • Likes Given: 5564
...I feel there is much more going on that meep cannot show because of some of the inherent limitations. What I'd like to see is a round tube ends capped off as a resonate chamber with a RF input, measure the Q in meep.  Would you look at the CSV files from aero and do a simple compare of the Q and stress values between the two? ...
Shell
Two possible ways to calculate Q from Meep, and none are possible at the moment with the output available:

1) To calculate the Q one needs to have access to all the Meep output: all the fields at all the nodes.  This is necessary to calculate the energy over the whole volume (which would be divided by the power loss on the copper surface). That output information is not available for any of the Meep runs.

2) Alternatively one could calculate the Q from the time decay with the RF feed off.  This cannot be done at the present time either because there has not been any runs with the RF feed off.  All the runs are with the RF feed ON, for a total of 0.013 microseconds (if my memory is correct).  During this time period there is no decay, on the contrary there is exponential magnification, so the calculated Q would be an imaginary number.
« Last Edit: 08/09/2015 01:40 am by Rodal »

Offline aero

  • Senior Member
  • *****
  • Posts: 3629
  • 92129
  • Liked: 1146
  • Likes Given: 360
Something has been bothering me since last night and I couldn't help but watching the modes change and flip in meep.

How can you calculate seriously any Q in a cavity that simply jumps around from one T mode decaying and building into another in such a short time? And they all do it, every simulation with varying speeds.

Just something to mull over on a day away from the shop.

Shell

Added: Back to lurk mode and I'll be quiet.

Excellent question.  I hope that you don't shy away from asking such questions because it is only this way that one can understand what is being output.

My understanding (aero to confirm) is that the quality factor is calculated by aero using the routine Harminv.

Please notice that Meep has this disclaimer for using Hamrinv to calculate the quality factor Q:

http://ab-initio.mit.edu/wiki/index.php/Meep_Reference#Harminv

Quote
Important: normally, you should only use harminv to analyze data after the sources are off. Wrapping it in (after-sources (harminv ...)) is sufficient.

Thus, the Quality factor and the frequency (also obtained by Harminv) should properly be obtained after the sources are OFF, not when they are on. 

This takes us back to the whole discussion about Q quality factor in experiments.  The problem is not only how to best experimentally measure and report Q, but it seems not ideal to me to discuss and report a Q (as first done by Shawyer and then imitated by all other EM Drive experimenters) with the source ON.

It seems to me that proper measurement of Q also in experiments should be done upon turning the source off and examining the decay (as posted by Frobnicat in a separate post).

With the sources OFF, the definition of Q (inverse to damping) is well-posed.

With the sources ON, the meaning of Q is tricky.  It seems to me that when people are measuring Q with the sources on they are assuming a well-posed problem with "nice" properties (symmetry, etc.) that may not be fulfilled.  This is particularly contradictory with TheTraveller: who rejects Finite Element solutions (using COMSOL or ANSYS) and exact solutions of the problem saying that only a solution that calculates a force can properly calculate the frequency and at the same time Shawyer "measures" Q with the RF feed ON which assumes a well posed nice solution amenable to presentation of Q as if it would be the same Q as the one calculated with the RF feed off.

The way that the Q is being measured experimentally by Shawyer and others using S11 and the 3db width is similar to a common method in structural vibration analysis.  Of course, this presupposes a steady-state response.  In reality with the RF feed ON, the measurement may be a transient instead.

_______________________________________________________

QUESTION TO aero:  have you been calculating the quality factor Q with Harminv, and if so, have you been doing so with the sources ON or OFF.  Have you been wrapping with  (after-sources (harminv ...)) ?

   (set! sources  drivesrc-Gaus)
    (run-sources+ (* gc T_meep)  ; This time, # peroids, is for non-resonant frequencies dissapation (start-up).
    ; Lower Q, shorter time, higher Q, longer time before measurement made. Take your best guess.
        (after-sources (harminv Ez (vector3 0.05 0.05 0.05) fmeep BW 5)) )
    (exit)
Retired, working interesting problems

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2442
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 3186
  • Likes Given: 2708
A couple of key points I found about Harminv in this description of operational programing of Harminv in wiki.
---------------
(Quote)
Important: normally, you should only use harminv to analyze data after the sources are off. Wrapping it in (after-sources (harminv ...)) is sufficient. (end Quote)

Also.
Can we access this in meep to analyze?

(harminv-err result)


---------------
(Quote)
http://ab-initio.mit.edu/wiki/index.php/Meep_Reference#Harminv

Harminv
The following step function collects field data from a given point and runs Harminv on that data to extract the frequencies, decay rates, and other information.
(harminv c pt fcen df [maxbands])
Returns a step function that collects data from the field component c (e.g. Ex, etc.) at the given point pt (a vector3). Then, at the end of the run, it uses Harminv to look for modes in the given frequency range (center fcen and width df), printing the results to standard output (prefixed by harminv:) as comma-delimited text, and also storing them to the variable harminv-results. The optional argument maxbands is the maximum number of modes to search for; defaults to 100.
Important: normally, you should only use harminv to analyze data after the sources are off. Wrapping it in (after-sources (harminv ...)) is sufficient.
In particular, Harminv takes the time series f(t) corresponding to the given field component as a function of time and decomposes it (within the specified bandwidth) as:
f(t) = \sum_n a_n e^{-i\omega_n t}
The results are stored in the list harminv-results, which is a list of tuples holding the frequency, amplitude, and error of the modes. Given one of these tuples, you can extract its various components with one of the accessor functions:
(harminv-freq result)
Return the complex frequency ω (in the usual Meep 2πc units).
(harminv-freq-re result)
Return the real part of the frequency ω.
(harminv-freq-im result)
Return the imaginary part of the frequency ω.
(harminv-Q result)
Return dimensionless lifetime, or "quality factor", Q, defined as -\mathrm{Re}\,\omega / 2 \mathrm{Im}\,\omega.
(harminv-amp result)
Return the complex amplitude a.
(harminv-err result)
A crude measure of the error in the frequency (both real and imaginary)...if the error is much larger than the imaginary part, for example, then you can't trust the Q to be accurate. Note: this error is only the uncertainty in the signal processing, and tells you nothing about the errors from finite resolution, finite cell size, and so on!
For example, (map harminv-freq-re harminv-results) gives you a list of the real parts of the frequencies, using the Scheme built-in map.

Offline aero

  • Senior Member
  • *****
  • Posts: 3629
  • 92129
  • Liked: 1146
  • Likes Given: 360
Yes, analyse away: Data table is to wide to format nicely here, so a spread sheet is attached.
Retired, working interesting problems

Offline TheTraveller

My Frustum dimensions, resonant frequency and Df are now locked and loaded:

Frustum big diameter       m   0.40000
Frustum small diameter   m   0.15900
Frustum centre length      m   0.24070
External Rf                      Hz   2,450,250,000
Calculated Df                   Df   0.857


The SPR and my resonance models agree to with-in 4.6MHz or 0.178% which is close enough for me.
« Last Edit: 08/09/2015 06:38 am by TheTraveller »
It Is Time For The EmDrive To Come Out Of The Shadows

Offline TheTraveller

I love this thread and am totally addicted, can't put it down.  Thanks for all the work and information!  Discussions about Q, oxidation, broad / narrow spectrum and the shape of the ends have got me thinking about an alternate configuration.  It has many of the same characteristics but might be slightly easier to build / find parts / work with.  It is an ordinary cylinder with a cone inside.  Maybe something to consider once folks are enjoying consistent thrust (lifting small objects and pets, etc...  ;)

Keep up the great work!

See Shawyer's patent as attached.
It Is Time For The EmDrive To Come Out Of The Shadows

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2442
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 3186
  • Likes Given: 2708
Yes, analyse away: Data table is to wide to format nicely here, so a spread sheet is attached.
So what are we seeing? Do you need for me to dig into it?


Offline aero

  • Senior Member
  • *****
  • Posts: 3629
  • 92129
  • Liked: 1146
  • Likes Given: 360
Yes, analyse away: Data table is to wide to format nicely here, so a spread sheet is attached.
So what are we seeing? Do you need for me to dig into it?

You are seeing several cases of exactly what was described in your post above. If you want to analyse it, it is available. It tells me that the software thinks it is operating successfully, maybe it will tell you something different.
Retired, working interesting problems

Offline TheTraveller

I have a real concern with TheTraveller's Excel spreadsheet. The values I get from the first basic dimensions are inconsistent. I'm talking of the file EMDriveCalc20150617b.xls available from emdrive.wiki as well as TT's Gdrive.

Let's take know values, for example Eagleworks' frustum:
Db = 0.2794 m
Ds = 0.15875 m
Frustum length = 0.2286 m
cone half-angle = 14.78°

Input the first three values, and the spreadsheet returns a cone half-angle of 24.5° :(

Calculate the hypotenuse or draw the plan in a CAD software with the know values, you will easily get the frustum side length at 0.2364256 m. But the spreadsheet returns 0.2584848 m!

The formula for the cone half-angle (cell D8) in the spreadsheet is :
=DEGREES(ATAN((D3÷2)÷((D5×(D4÷2))+((D3÷2)−(D4÷2))+D5)))
Whereas it could use arccosine, frustum centre length (diameter center to diameter center) and frustum side length:
=DEGREES(ACOS(D5/D9)

Talking about the frustum side length (cell D9), its formula is wrong:
= SQRT(D5^2+(D3−D4)^2)
The correct formula should use end radii squared instead of end diameters squared:
= SQRT(D5^2+((D3−D4)÷2)^2)

How is the rest right or wrong? I can't even get Df right with the available spreadsheet. When inputing the Baby EmDrive data for example, Df becomes negative which is impossible (it should be comprised between 0 and 1) EDIT: my mistake, 24 GHz instead of 2.4 GHz resolved this issue.
Whatever, I don't get the same Df as TheTraveller for the same untouched spreadsheet and same input values. See fourth attachement below. Those differences are quite small, but everything else following in the spreadsheet gets very different values from those discrepancies.

@TheTraveller: can you please double-check those basic values in the spreadsheet, and upload a corrected version to the emdrive.wiki? This would be much appreciated by the EmDrive community :)

Below, I show two hypothesis for TT's EmDrive Mark 2, according to how the "Frustum centre length" is defined in the spreadsheet.
- The first with Frustum centre length = 208.71 mm has a cone half-angle (corrected formula) of 30°
- The second with Frustum centre length = 240.7 mm has a cone half-angle (corrected formula) of 26.6° (instead of 27.7° with the wrong angle formula).

What is important to note is that "Frustum centre length" as defined in the spreadsheet is the length between the centers of the two end diameters, and not the length defined by TheTraveller in his drawing where it is the apex r2-r1 length. All the misunderstanding comes from the difference in that drawing (attached in third position below).

Thanks for finding that. Fixed. There may be other such errors as a lot of that work was done when I was taking strong pain killing tables. These formula are not involved in the resonance calc.

My resonant calc matches the SPR model by 0.178%, so I'm happy with that. They were 4.6MHz higher.

The frustum length is the distance in meters between the centre points of the end plates. If the end plates are spherical then the frustum side wall length is also the centre separation length.

Have modified the spreadsheet to show Vertex length and spherical end plate radius as attached.

Also modified the drawing to show the same data.
« Last Edit: 08/09/2015 06:35 am by TheTraveller »
It Is Time For The EmDrive To Come Out Of The Shadows

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Yes, analyse away: Data table is to wide to format nicely here, so a spread sheet is attached.
No wonder it's too[sic] wide. You really don't want, or need, all those significant digits.

Offline Mulletron

  • Full Member
  • ****
  • Posts: 1150
  • Liked: 837
  • Likes Given: 1071

The Eggcelerator

Thanks for your eggcellent idea! Might want to run it through MEEP...JIK. It may crack interstellar flight.  8)

(Sorry couldn't resist, Good Morning!)
https://twitter.com/EggPuns/status/545618667481096193
Image: http://www.rt17.hr/teslas-egg-of-columbus/
« Last Edit: 08/09/2015 06:30 am by Mulletron »
And I can feel the change in the wind right now - Rod Stewart

Offline ThinkerX

  • Full Member
  • ***
  • Posts: 341
  • Alaska
  • Liked: 126
  • Likes Given: 63
Quote
The Eggcelerator

Dimensions?

Frequency?

And where does the magnetron go?

Hey, sort of the right shape and everything.  A run through with MEEP probably wouldn't hurt anything.

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Stranger things have happened. I looked at that "inverted cone" drawing and it sort of instantly generalised itself to an ovoid  :)

Offline Mulletron

  • Full Member
  • ****
  • Posts: 1150
  • Liked: 837
  • Likes Given: 1071
Stranger things have happened. I looked at that "inverted cone" drawing and it sort of instantly generalised itself to an ovoid  :)

Think that would do a better job of confining the fields to 1d. As per this:

http://www.pnas.org/content/111/29/10485.full (paper)
http://www.sciencedaily.com/releases/2014/07/140722091425.htm (newsy)
And I can feel the change in the wind right now - Rod Stewart

Offline ThinkerX

  • Full Member
  • ***
  • Posts: 341
  • Alaska
  • Liked: 126
  • Likes Given: 63
Delta Mass...and others.

I keep coming back to David Bae's photon recycling laser two platform system.

To me, it looks like a weird photon rocket variant - yet one that produces thousands of times the amount of thrust a photon rocket should be able to produce.  It also appears legit.

With Bae's device you get around over unity / conservation of energy via redshift and a statement that only a small fraction of each photonic bounce is used for propulsion.   Yet, at the same time, another implication is missed:

For that to be true, the total energy potential of a photon would have to be many thousands of time greater than what the standard photon rocket calculations allow for.  Anybody care to dispute or comment on this?

Therefor, while the calculations for maximum thrust for the classic photon rocket are correct, you HAVE to allow for the possibility of much higher thrust IF the photons greater energy can be tapped.  Bae's device shows one way of doing this.   

Something else I have been wondering ever since Doctor Rodal first posted it.  Specifically, from MEEP based calculations and a program of his own, he stated that if the cavity had the right angle, the forces within it did three things:

1 - they went exponential, at least for the duration of the MEEP run (roughly 1/1000th of a second?)

2 - did not sum to zero, at least within that time period.  (But sooner or later, to satisfy the laws of thermodynamics, they would have to sum to zero).

3 - and these forces were nowhere near the levels reported in the EM Drive experiments.

So I have been wondering.  The relevant law of thermodynamics is Time Averaged.  Suppose this device breaks or twists the average, enough to where the 'step down cycle' doesn't kick in for say....several seconds, or even a minute.  That the forces within the EM Drive continued to build exponentially in mostly one direction not for just a few thousandths of a second, but several actual seconds.  Would the exponential force by that point be on a par with the typically reported results?

A larger wave or cycle of some sort.

Two other relevant tidbits here:

Best I can recollect, no EM Drive test has lasted more than a couple minutes.

Something that has been pointed out by Rodal, Paul March, and others:  maintaining the relevant frequency to produce thrust is a severe pain.  Suppose, due to something we cannot see at the moment, the frequency is impossible to maintain for whatever reason for more than a minute or two?  And must be reset afterwards.
« Last Edit: 08/09/2015 07:36 am by ThinkerX »

Offline Mulletron

  • Full Member
  • ****
  • Posts: 1150
  • Liked: 837
  • Likes Given: 1071

If you really want to dive deep, look at Orbital Angular Momentum of circular radiation:

http://arxiv.org/ftp/arxiv/papers/0905/0905.0190.pdf

You think it is converting OAM to LM in there?

Didn't someone detect rotation in MEEP a few pages back?
« Last Edit: 08/09/2015 07:52 am by Mulletron »
And I can feel the change in the wind right now - Rod Stewart

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
0