Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1800169 times)

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1227
  • Do it!
  • Vista, CA
  • Liked: 1293
  • Likes Given: 1745
After seeing that vid I'm adding another artifact candidate - steam jet.
I agree, that's why I want to do a lot of thermal testing of hotspots. The machine screw head was way hotter than surrounding metal. I suspect some rf leakage between copper clad end plates. Simple fix I'll mess with over the weekend. Thermal issues need to be addressed well ahead of torsion or balance testing imho.

Use a heatsink that has radial fins, so orientation doesn't matter. As it is now, the fins are in the wrong direction. The fins are flat with the openings on the sides. The opening should be up and down, so convention will cool it, pulling cool air in from the bottom. The airflow is blocked.

Granted, this will never work in a vacuum chamber without liquid cooling or a very large mass to dump heat into.
Todd

Offline aero

  • Senior Member
  • *****
  • Posts: 2740
  • 92129
  • Liked: 704
  • Likes Given: 237
Just a random thought @rfmwguy.

I know you don't want to blow air over the fins, but would you be able to calibrate the force balance if you used your wife's vacuum sweeper hose to suck air across them? Maybe a longer hose so it exhausted outside? Maybe an inlet of some sort to channel the moving  air to minimize stray air currents?

Maybe a liquid cooling loop.  ;D

The lag in the temperature readings after turn-off seemed to indicate that the center of the magnetron got much hotter than the fins where you were measuring temperature.
« Last Edit: 08/01/2015 03:27 AM by aero »
Retired, working interesting problems

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Actually I got the impression from the vid that the heat capacity of the magnetron is quite low - it lost heat very rapidly after turn off. Oddly perhaps, the same could not be said of the frustum.

Or perhaps these observations are due to vid editing messing with the observed timeline?
« Last Edit: 08/01/2015 03:33 AM by deltaMass »

Offline TheTraveller

...
2) His remarks about buying the microwave oven and the cooking oil were in jest.  I doubt he had the time to do that personally.  Maybe the guy that is on vacation was the one that bought the cooking oil.
...

Isn't putting cooking oil in an experiment involving vacuum asking for troubles ? I mean, it likely is not low vapour pressure (like would be vacuum pump oil), has high water content...
Isn't purposely testing an EM Drive with a Q lower than 50 and with a huge bloody hole from a huge bloody waveguide asymmetrically placed on one side also asking for trouble ?  It is par for the course...

Totally agree. Then putting a 200C magnetron INSIDE a sealed enclosure with a 50 Q EMDrive is just asking for a measurement disaster. When Shawyer did his sealed enclosure testing the magnetron was OUTSIDE the enclosure so he did not have to deal with such a massive heat source inside his enclosure.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline Silversheep2011

  • Member
  • Posts: 85
  • Austraila
  • Liked: 40
  • Likes Given: 258
rfmwguy.

I'm thinking replace the magnetron [see attachment below looks the same] with a RF transistor as used here

http://www.zdnet.com/article/freescales-radio-frequency-oven-the-end-of-the-microwave/

maybe it runs with less heat loss?
and runs cleaner r.f. waves
http://www.freescale.com/webapp/video_vault/videoSummary.sp?code=RF-SAGE-VIDEO

 see 2.37

Offline TheTraveller

Another error that Shawyer has in that chart (besides the direction of the force in NASA's experiments) is the specific force for Yang's experiments.  Shawyer has only 0.288 N/kW for Yang.  Yang reported 1 N/kW.  You now say that it was really 4 N/kW. Regardless, why is Shawyer reporting only 0.288 N/kW and making Yang's experiments to have lower specific force than Shawyer's Flight Thruster ?  (it looks like the peer reviewers for Shawyer's paper missed all of this ...)

Explanation:

Specific thrust to power ratio is given for a specific thrust measured at a specific input power, which does not vary linearly a priori (but this last statement is not true, as explained below).

288 N/kW is Yang's specific thrust measured for the maximum force recorded (720 mN) at the maximum power she used (2.5 kW).

1 N/kW is a contrario the maximum mean specific thrust Yang recorded at an input power of 300 W, where she measured a thrust of 310 then 270 mN (2013 Yang paper, page 7 and 8).

But after she corrected the dip in the thrust versus power curve due to the magnetron to cavity bandwidth mismatch (see same paper, page 8) we can indeed calculate an average specific thrust comprised between 2 and 4 N/kW (and even an -erroneous?- 14N/kW for the first recorded value):



As Yang concludes, this relation shows that the EM thrust monotonously increases with the practical power augmentation.

EDIT: This was already pointed out by TheTraveller several days ago.

Yes, but apples should be compared to apples.  Why does Shawyer use in the chart a figure for Yang at a much higher thrust than the one achieved by the Flight Thruster.  It doesn't make sense.  What would make sense is to compare the specific forces for the same thrust, or to compare the highest specific force, or some other fair comparison, but not to make the present arbitrary comparison.  Shawyer should report something like "Mean Specific Thrust" or "Highest Specific Thrust" etc.

The present comparison is making the Shawyer Flight Thruster look better than Yang's which is not accurate as Yang reported the highest thrust, and the highest specific thrust than any other EM Drive, including the Flight Thruster.

Why?

I suggest there is a lot of protecting IP and not revealing the true Specific Force numbers going on.

As one example, if one takes the Flight Thruster numbers and works them backward, the Q was well over 100k despite the reported Q being 50,000. I mean what the neck it had spherical end plates? The earlier Demonstrator with flat end plates achieved a reported Q of 45,000. So go figure.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline TheTraveller

Hmmm, If Shawyer and Yang are using non standard methodology to figure out Q. What does this non standard methodology tell us. For example is there a difference in result between the Shawyer methodology and the standard methodology? If so, with what parameters does the Q result begin to diverge? What is the implication of this divergence?

Also, and most interestingly. Why has Shawyer and Yang opted for this non standard methodology?

Doing S11 return loss measurements to determine unloaded Q on a 1 port resonant system is NOT non standard. This method is the standard way to measure EMDrive cavities unloaded Q. As used by EWs, Shawyer, Prof Yang, Tajmar and myself.

What others are talking about is doing 2 port loaded Q measurements using S21 methodology. No one in the EMDrive world is interested in loaded Q values or doing 2 port (2 holes in the cavity) S21 measurements.

Here is a paper describing how to do unloaded Q measurements using S11. It is not correct to say using S11 to measure unloaded Q on a 1 port cavity is non standard.

2 documents attached to support that opinion.

Hmmm.. this is an interesting paper. The power delivered to the load, PL implies that in a tapered waveguide where w0 is a variable. More power is being delivered to the end of the frustum where w ~ w0, and less power will be delivered to the end where these frequencies have a larger difference. So if the frequency is tuned to match the small end, then more power will be delivered to this end, and less will be delivered to the big end. The difference must either be stored energy, thrust or lost as heat.
Todd

Using return loss / VSWR to tune to the best frequency, you are tuning to the entire load as what you want is to get the min reflected power and the most absorbed forward power inside the cavity. As for data collection knowing the real forward power will also allow you to correctly calc Specific Force as you will know how much power the cavity has absorbed, instead of using raw power output as is done in the magnetron case.

This means that even if the inside atmo becomes vacuum and the resonance frequency changes (as I suspect it will), the ideal Rf frequency will always be found as VSWR or reflected power versus forward power or return loss is king to freq track a EMDrive cavity.

Of course if using a magnetron, you have no control over the frequency and no ability to track and adjust for either cavity resonance changes nor magnetron frequency changes with magnetron temperature changes.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline TheTraveller

...
Sure it will resonate at some lower frequency because of the traveled distance. What is interesting is the modes it makes and watching the waves decay into Evanescent decay I'd assume. I was wondering if a CSV file was around for this.

Interesting chart observation on the TE013 mode.

Shell

1) No there are NO csv files to analyze for that case. I am 100% sure of that.  Aero would have to run it again, and would have to excite it at a lower frequency than 2.45 GHz to take into account the slightly lower natural frequency for the same mode.

2) Looking at my table, the frequency is lowered only a little with the extension towards the apex, for the same mode, and after a given amount it practically does not change much as one extends it.  This is due to the fact that in a tapered cavity the mode does not travel a longer distance once it reaches a transition zone: it stops travelling altogether in the section that is close to the small base, which becomes purely evanescent.

3) Note that in the aero runs he did NOT excite the mode that becomes purely evanescent near the small base.  Instead he excited a higher mode with higher p.  This was due to the fact that he did not lower the frequency perhaps or perhaps was due to antenna placement.

4) Now that I have found out that Yang in her 2010 paper in Chinese (not in the butchered translation) explicitly mentions the dielectric losses inside her cavity, this means to me that Yang may have used a dielectric for her experiments in 2010.  Why include an equation for dielectric losses lowering the Q if Yang did not use a dielectric inside her cavity?

http://forum.nasaspaceflight.com/index.php?topic=37642.msg1411922#msg1411922

A dielectric lowers the natural frequency.  Therefore the Yang/Shell geometry will be a WRONG model of Yang's geometry if Yang used a dielectric, because the Yang/Shell geometry was calculated on the basis that there was no dielectric inside her cavity.


5) Thus, more discussion is needed: did Yang use a dielectric for her experiments ?

Prof Yang, according to Shawyer did not use a dielectric. She never mentions a dielectric. So it seems no dielectric was used by Prof Yang.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1227
  • Do it!
  • Vista, CA
  • Liked: 1293
  • Likes Given: 1745
Hmmm, If Shawyer and Yang are using non standard methodology to figure out Q. What does this non standard methodology tell us. For example is there a difference in result between the Shawyer methodology and the standard methodology? If so, with what parameters does the Q result begin to diverge? What is the implication of this divergence?

Also, and most interestingly. Why has Shawyer and Yang opted for this non standard methodology?

Doing S11 return loss measurements to determine unloaded Q on a 1 port resonant system is NOT non standard. This method is the standard way to measure EMDrive cavities unloaded Q. As used by EWs, Shawyer, Prof Yang, Tajmar and myself.

What others are talking about is doing 2 port loaded Q measurements using S21 methodology. No one in the EMDrive world is interested in loaded Q values or doing 2 port (2 holes in the cavity) S21 measurements.

Here is a paper describing how to do unloaded Q measurements using S11. It is not correct to say using S11 to measure unloaded Q on a 1 port cavity is non standard.

2 documents attached to support that opinion.

Has anyone else looked at this yet? UnloadedQmeasurement.pdf from TT's post above. They give an equation for PL, which is the power delivered to the load. I'm referring to the first and second occurrence, where it is shown power by Ohm's Law, I^2 R and they derive power exactly, in terms of the impedance function Z. (Pity they didn't use equation numbers.)

Anyway, PL is maximum at resonance, when w=w0. When it is not at resonance, energy is stored as reactive power, out of phase with the current. Either current lags voltage or vise versa. It is easily shown that the current has a variable drift velocity on a cone. In other words, as I said about Zeng and Fan, there is a Power Factor that is less than 1 and approaches 1 at the small end. When there is a power factor less than 1, not all of the power does work. Some of it just circulates.

So this is new. At a constant frequency w, the big end has a cut-off far from w, while the small end has a cut-off at w. The power factors are different because there is a phase shift between E and H. When the energy is at the small end, Then the power factor is close to 1, the Watts (Real power that can do work) is maximized. At the big end, there is a power factor < 1, so there is stored energy as VARS and less energy as Watts, which means less power to do work. This is exactly what the phase diagrams (Beta/kr) show in Zeng and Fan.

So in any of these equations where F ~ P/v, the P term is in Watts, not VA and therefore the power is as asymmetrical as the force! Their equation for PL, precisely shows that the power depends on the dimensions of the waveguide, on L & C as w0.

(So does gravity BTW.)
Todd



Offline TheTraveller

Hmmm, If Shawyer and Yang are using non standard methodology to figure out Q. What does this non standard methodology tell us. For example is there a difference in result between the Shawyer methodology and the standard methodology? If so, with what parameters does the Q result begin to diverge? What is the implication of this divergence?

Also, and most interestingly. Why has Shawyer and Yang opted for this non standard methodology?

Doing S11 return loss measurements to determine unloaded Q on a 1 port resonant system is NOT non standard. This method is the standard way to measure EMDrive cavities unloaded Q. As used by EWs, Shawyer, Prof Yang, Tajmar and myself.

What others are talking about is doing 2 port loaded Q measurements using S21 methodology. No one in the EMDrive world is interested in loaded Q values or doing 2 port (2 holes in the cavity) S21 measurements.

Here is a paper describing how to do unloaded Q measurements using S11. It is not correct to say using S11 to measure unloaded Q on a 1 port cavity is non standard.

2 documents attached to support that opinion.
You are not understanding me Mr T. The methodology EM Drive people are using is not the norm. Yet, they had placed sample ports in the device to make a standardized 2-port measurement. The fact they did not use the second port is indicative of a desire to uniquely define their Q...thousands of times the normal range of measurement.

There is no justification in EM drive Q measurements 3dB above best return loss. Again, any RF engineer will scoff at this "invented" measurement technique.

Since I am scoffing at this myself, I am committed to a 2 port Q measurement on my build as the RF industry would accept...those not using conventional methodogy are advertising their design weakness, IMO.

EMDrive Force generating is based, in part on unloaded cavity Q, measured as 3bd down from the peak S11 return loss. If you wish to engage in doing predictions of the Force your cavity may be able to generate, then you need to measure the unloaded cavity Q as per how the other experimenters do it and now the predictive Force equations require it to be done.
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
I'm sure there's a conversion equation between real Q and "weird Q". It will involve geometrical terms probably.
« Last Edit: 08/01/2015 04:33 AM by deltaMass »

Offline TheTraveller

I'm sure there's a conversion equation between real Q and "weird Q". It will involve geometrical terms probably.

There are 3 types of Q that I know of:

1) unloaded (used in EMDrive Force equations)
2) loaded
3) external

as attached.

Unloaded Q, measured via S11 return loss at the -3dB points is not "Weird" Q. It may not be how some choose to measure it but it is valid as I have shown in the papers I posted and this attachment.

Time to move on.
« Last Edit: 08/01/2015 04:50 AM by TheTraveller »
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1227
  • Do it!
  • Vista, CA
  • Liked: 1293
  • Likes Given: 1745
rfmwguy.

I'm thinking replace the magnetron [see attachment below looks the same] with a RF transistor as used here

http://www.zdnet.com/article/freescales-radio-frequency-oven-the-end-of-the-microwave/

maybe it runs with less heat loss?
and runs cleaner r.f. waves
http://www.freescale.com/webapp/video_vault/videoSummary.sp?code=RF-SAGE-VIDEO

 see 2.37

They still require a heat sink. 3.7 deg. C/Watt is a lot of heat. These are also only good for about 200 to 400 mW per transistor, if you've got a good heat sink. If you want to get back up to 400W, you will need to parallel about a dozen of them on a well designed PC board, that is thermally conductive to the heatsink. At these frequencies, I have no idea how you would do that, or what proper design criteria are necessary for a microwave amplifier of sufficient power.
Todd

Offline SteveD

  • Full Member
  • **
  • Posts: 290
  • United States
  • Liked: 76
  • Likes Given: 10
rfmwguy.

I'm thinking replace the magnetron [see attachment below looks the same] with a RF transistor as used here

http://www.zdnet.com/article/freescales-radio-frequency-oven-the-end-of-the-microwave/

maybe it runs with less heat loss?
and runs cleaner r.f. waves
http://www.freescale.com/webapp/video_vault/videoSummary.sp?code=RF-SAGE-VIDEO

 see 2.37

They still require a heat sink. 3.7 deg. C/Watt is a lot of heat. These are also only good for about 200 to 400 mW per transistor, if you've got a good heat sink. If you want to get back up to 400W, you will need to parallel about a dozen of them on a well designed PC board, that is thermally conductive to the heatsink. At these frequencies, I have no idea how you would do that, or what proper design criteria are necessary for a microwave amplifier of sufficient power.
Todd

With X-Band HAM radio being at 10 Ghz some of the HAM guys might have addressed a problem similar enough to be useful.

Offline deltaMass

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

I'm thinking replace the magnetron [see attachment below looks the same] with a RF transistor as used here

http://www.zdnet.com/article/freescales-radio-frequency-oven-the-end-of-the-microwave/

maybe it runs with less heat loss?
and runs cleaner r.f. waves
http://www.freescale.com/webapp/video_vault/videoSummary.sp?code=RF-SAGE-VIDEO

 see 2.37

They still require a heat sink. 3.7 deg. C/Watt is a lot of heat. These are also only good for about 200 to 400 mW per transistor, if you've got a good heat sink. If you want to get back up to 400W, you will need to parallel about a dozen of them on a well designed PC board, that is thermally conductive to the heatsink. At these frequencies, I have no idea how you would do that, or what proper design criteria are necessary for a microwave amplifier of sufficient power.
Todd
I take issue with your use of the arithmetic division operator.

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1227
  • Do it!
  • Vista, CA
  • Liked: 1293
  • Likes Given: 1745
rfmwguy.

I'm thinking replace the magnetron [see attachment below looks the same] with a RF transistor as used here

http://www.zdnet.com/article/freescales-radio-frequency-oven-the-end-of-the-microwave/

maybe it runs with less heat loss?
and runs cleaner r.f. waves
http://www.freescale.com/webapp/video_vault/videoSummary.sp?code=RF-SAGE-VIDEO

 see 2.37

They still require a heat sink. 3.7 deg. C/Watt is a lot of heat. These are also only good for about 200 to 400 mW per transistor, if you've got a good heat sink. If you want to get back up to 400W, you will need to parallel about a dozen of them on a well designed PC board, that is thermally conductive to the heatsink. At these frequencies, I have no idea how you would do that, or what proper design criteria are necessary for a microwave amplifier of sufficient power.
Todd
I take issue with your use of the arithmetic division operator.

Nick picking again...  oC/W. 

Offline TheTraveller

A thought experiment to show guide wavelength, group velocity and end plate reflected Force is different on each end plate of an EMDrive.

Please consider the attached where we have a tapered waveguide and 2 constant diameter circular waveguides at each end.

The smaller waveguide has inside a 1/2 wave at it's waveguides guide wavelength as determined by the cutoff wavelength at the TE01 excitation mode.

Likewise the larger constant diameter circular waveguide has inside a 1/2 wave at it's waveguides guide wavelength as determined by the cutoff wavelength at the TE01 excitation mode.

The Force on each flat end plate will be determined by the waves group velocity as set by the constant diameter waveguides cutoff wavelength at TE01 excitation mode.

While some here may deny any change in guide wavelength or group velocity inside the tapered waveguide section, in this setup it is very clear there will be a differential Force generated between the 2 end plates because the guide wavelength (group velocity) at each end plate is different.

I suggest this setup demonstrates there will be a differential Force generated on a non extended tapered waveguide's end plates as Shawyer and Prof Yang claim to happen.

« Last Edit: 08/01/2015 05:56 AM by TheTraveller »
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline deltaMass

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

I'm thinking replace the magnetron [see attachment below looks the same] with a RF transistor as used here

http://www.zdnet.com/article/freescales-radio-frequency-oven-the-end-of-the-microwave/

maybe it runs with less heat loss?
and runs cleaner r.f. waves
http://www.freescale.com/webapp/video_vault/videoSummary.sp?code=RF-SAGE-VIDEO

 see 2.37

They still require a heat sink. 3.7 deg. C/Watt is a lot of heat. These are also only good for about 200 to 400 mW per transistor, if you've got a good heat sink. If you want to get back up to 400W, you will need to parallel about a dozen of them on a well designed PC board, that is thermally conductive to the heatsink. At these frequencies, I have no idea how you would do that, or what proper design criteria are necessary for a microwave amplifier of sufficient power.
Todd
I take issue with your use of the arithmetic division operator.

Nick picking again...  oC/W.
Not that - that you say you need "about a dozen" 400 mW transistors to get 400 Watts.

Offline SteveD

  • Full Member
  • **
  • Posts: 290
  • United States
  • Liked: 76
  • Likes Given: 10
The embedded video did not load.  Probably not going to be able to get a look until I get back from vacation.  On heat issues two things come to mind.  1.  You could always call around and seem if you could rent/borrow and industrial freezer (one of the ones you can walk inside) for a day.  Might help if we're talking about temps around 100C.  2.  Some of the water/fluid cooling kits designed for computer equipment might have bearing (see Newegg) though I'm sure you're already aware of them.

Online Rodal

  • Senior Member
  • *****
  • Posts: 5836
  • USA
  • Liked: 5915
  • Likes Given: 5254
... She never mentions a dielectric. So it seems no dielectric was used by Prof Yang.
Wrong.  Yang explicitly mentions a dielectric in her original 2010 paper in Chinese.

Perhaps you did not read this post:  http://forum.nasaspaceflight.com/index.php?topic=37642.msg1411922#msg1411922, so I will make these points again one by one, and be more direct this time as to where you make mistakes:


1) The equation you posted for Q is wrong, it is not the equation that Yang used.   The equation you posted for Q does not make any sense.  It never appears in Yang's original 2010 paper in Chinese.

2) You posted an equation for Q that instead came from a very bad translation in Shawyer's website instead of coming from her original 2010 paper in Chinese. 

3) The original paper in Chinese by Yang has the correct equation for Q.

4) In the original paper in Chinese by Yang she uses tnδ instead of tnd.  tnδ means tangent delta (loss tangent) which is the material property for a dielectric loss, which reduces the Q.

5) In the original 2010 paper in Chinese by Yang she explicitly writes: "dielectric losses" instead of the badly translated "electric losses" that appears in the bad translation

I don't know whether Yang did use or did not use dielectrics in any of her experiments, but what is undeniable from reading her 2010 paper in Chinese (*) is that Yang clearly mentions dielectric losses explicitly and furthermore she goes further: she uses an equation for Q that takes into account dielectric losses. 

_______________
(*) instead of the very bad translation that appears in Roger Shawyer's website, that you used to refer to her Q expression
« Last Edit: 08/01/2015 01:01 PM by Rodal »

Tags: