-
#1320 by as58 on 08 Oct, 2016 19:22
-
His paper being wrong is no more a matter of opinion than 1+1=2 being a matter of opinion.
I must admit that I've been skimming the thread a bit as of late due to the rate and volume of updates; have you gone into the specifics of the mistakes contained in the Woodward paper in prior posts?
Woodward's paper looks really bizarre. When he quite correctly arrives at a contradiction (even obviously wrong) with conservation of energy when he assumes constant thrust with constant power, he doesn't seem to even consider the possibility that it's because his theory just doesn't work. -
#1321 by WarpTech on 08 Oct, 2016 19:34
-
... The equation also shows that we want the lowest order mode harmonic...
That makes sense. The lowest fundamental frequencies should have the highest amplitudes.
All of this agrees with @Notsosureofit's model as well.
Higher mode shapes like TE013 instead of TE011 have been emphasized by some following Shawyer. One justification for that is that (if one does not have other means to provide an asymmetry, like electrostrictive inserts or ferromagnetic ends) TE013 has the maximum energy density closer to the small end than TE012 (which has the maximum energy density towards the middle of the longitudinal direction).QuoteTE013 has the maximum energy density closer to the small end than TE012 (which has the maximum energy density towards the middle of the longitudinal direction).
[/color]
This is one reason I'm pursuing TE013.
I've reached the limits in the home lab with vibration and noise and EM interference. I'm starting on a program over the next 6 weeks to set up a new lab. Separate power, EMI shielded and vibration isolated along with some new and better equipment.
Before I can put up this cute little building for the Space Lab I have my work cut out for me.
This attached pic is the area next to the house I'm going to be clearing off to set this lab. I've got to Split and move all the wood, grade and level the ground and pour the concrete. Erect the building, insulate it, wire it, EMI shield and heat it. Then comes the real work, moving in and setting up the test equipment and start to test. I guess you could ask why I'm doing this? Simple, error bars and to get myself from two prancing 135 pound Great Pyrenees and on stable ground.
Oh thank goodness I have a tractor with a loader and can grade. I want this to be done by Thanksgiving by adding some sweat and blisters.
Back to work!
My Best,
Shell
PS: Many more details in the construction and lab isolation I've not included.
Suggestion: Before you pour the foundation pad, put a #2/0 bare copper ring around the whole pad. Due to frost level, I would put it at least 18" deep, and put ground rods in all 4 corners, then ground that to the rebar in the pad. Leave a copper tail sticking up through the concrete. Then purchase one of those outdoor metal sheds, and set it on the concrete and ground it. Viola! You have a cheap faraday cage. Then put insulation and wiring inside. Steel is not as good as mu-metal, but it's better than nothing and it's cheaper than construction permits.
-
#1322 by rfmwguy on 08 Oct, 2016 19:50
-
Woodwards Theory has their own thread: http://forum.nasaspaceflight.com/index.php?topic=31037.0His paper being wrong is no more a matter of opinion than 1+1=2 being a matter of opinion.
I must admit that I've been skimming the thread a bit as of late due to the rate and volume of updates; have you gone into the specifics of the mistakes contained in the Woodward paper in prior posts?
Woodward's paper looks really bizarre. When he quite correctly arrives at a contradiction (even obviously wrong) with conservation of energy when he assumes constant thrust with constant power, he doesn't seem to even consider the possibility that it's because his theory just doesn't work.
Some may consider EmDrive and Woodwards devices similar, but we're really dealing with Microwave frequencies versus Ultrasonic frequencies on his device. No expert on it at all, but I mentioned earlier that it seemed too close to things like this: http://waset.org/publications/10000733/cleaning-performance-of-high-frequency-high-intensity-360-khz-frequency-operating-in-thickness-mode-transducers
Don't think his theories blend well with EmDrive theories, in fact they may obscure some theoretical advances made in EmDrive work recently. -
#1323 by meberbs on 08 Oct, 2016 20:11
-
Exactly.His paper being wrong is no more a matter of opinion than 1+1=2 being a matter of opinion.
I must admit that I've been skimming the thread a bit as of late due to the rate and volume of updates; have you gone into the specifics of the mistakes contained in the Woodward paper in prior posts?
Woodward's paper looks really bizarre. When he quite correctly arrives at a contradiction (even obviously wrong) with conservation of energy when he assumes constant thrust with constant power, he doesn't seem to even consider the possibility that it's because his theory just doesn't work.
I have specific criticisms scattered throughout a few posts, and Tellmeagain wrote a paper detailing the issues, which can be found here.
-
#1324 by SeeShells on 08 Oct, 2016 20:17
-
Good ideas and you betcha I see the reasoning, you're a very sharp guy. Also you got to see George's presentation at the Advanced Propulsion Workshop as well and the many details needed to reduce the error bars. Did you notice that in the MEGA and MET data there were no error bars? That bothers me. (Dr. Rodal, you have a answer?)... The equation also shows that we want the lowest order mode harmonic...
That makes sense. The lowest fundamental frequencies should have the highest amplitudes.
All of this agrees with @Notsosureofit's model as well.
Higher mode shapes like TE013 instead of TE011 have been emphasized by some following Shawyer. One justification for that is that (if one does not have other means to provide an asymmetry, like electrostrictive inserts or ferromagnetic ends) TE013 has the maximum energy density closer to the small end than TE012 (which has the maximum energy density towards the middle of the longitudinal direction).QuoteTE013 has the maximum energy density closer to the small end than TE012 (which has the maximum energy density towards the middle of the longitudinal direction).
[/color]
This is one reason I'm pursuing TE013.
I've reached the limits in the home lab with vibration and noise and EM interference. I'm starting on a program over the next 6 weeks to set up a new lab. Separate power, EMI shielded and vibration isolated along with some new and better equipment.
Before I can put up this cute little building for the Space Lab I have my work cut out for me.
This attached pic is the area next to the house I'm going to be clearing off to set this lab. I've got to Split and move all the wood, grade and level the ground and pour the concrete. Erect the building, insulate it, wire it, EMI shield and heat it. Then comes the real work, moving in and setting up the test equipment and start to test. I guess you could ask why I'm doing this? Simple, error bars and to get myself from two prancing 135 pound Great Pyrenees and on stable ground.
Oh thank goodness I have a tractor with a loader and can grade. I want this to be done by Thanksgiving by adding some sweat and blisters.
Back to work!
My Best,
Shell
PS: Many more details in the construction and lab isolation I've not included.
Suggestion: Before you pour the foundation pad, put a #2/0 bare copper ring around the whole pad. Due to frost level, I would put it at least 18" deep, and put ground rods in all 4 corners, then ground that to the rebar in the pad. Leave a copper tail sticking up through the concrete. Then purchase one of those outdoor metal sheds, and set it on the concrete and ground it. Viola! You have a cheap faraday cage. Then put insulation and wiring inside. Steel is not as good as mu-metal, but it's better than nothing and it's cheaper than construction permits.
One end of the mini-lab will be setup, isolated physically (vibration isolation) from the rest of the little lab and fully enclosed in a grounded Faraday cage with screen and backed with sheet steel walls, much like you describe. The test stand will be surrounded with Mu-Metal sections with non-bent overlapping sections (very important).
My goal is to reduce the error bars and increase the measurement of thrust as far as I can in air and it's going to take attention to detail. My goodness, I'd not be dropping all this money and sweat if I didn't think it was worthwhile to pursue.
Best to you,
Shell -
#1325 by Rodal on 08 Oct, 2016 21:37
-
...Good ideas and you betcha I see the reasoning, you're a very sharp guy. Also you got to see George's presentation at the Advanced Propulsion Workshop as well and the many details needed to reduce the error bars. Did you notice that in the MEGA and MET data there were no error bars? That bothers me. (Dr. Rodal, you have a answer?)...
Shell, please remember, for example, the graph by Woodward showing measured force vs. voltage, for different voltages, which included the range of values for each plotted point.
Shell
In any case, "error bars" (as often used by the Quality Control, and the Social Sciences community) often represent one standard deviation. From a rigorous statistical viewpoint it is inappropriate to use standard deviation as a measure of dispersion for a process with undetermined statistical distribution because the use of one standard deviation presumes a Gaussian Normal distribution. For experiments where people are still debating what is being measured, including what is the physical process responsible, there may not be a basis upon which to assume a Gaussian distribution (which is often assumed by Quality Control people and people involved in the Social Sciences because a) it is the simplest distribution that allows simple, explicit, formulas for calculation and b) many random processes in Quality Control and Social Sciences approximately obey it (*), but that is not applicable for a variety of well known processes). The appropriate approach would be to plot histograms to assess the proper distribution law, but unfortunately many experimental papers lack the number of statistical samples to make such determinations.
Personally, when I showed plots comparing experimental data with calculations, I plotted all the experimental points . This is the approach advised by great statisticians like Tukey and Mosteller, which I think is much superior to anyone that would hide the number of experimental points (the sample population) behind a "standard deviation error bar" that is not justified statistically when there is a small number of experiments and the distribution is normal, or when the distribution is not normal (for any number of statistical samples), or when the distribution is unknown (displaying standard deviation error bars in those cases would constitute a misuse of statistics).
______________
(*) And many other processes do not, foremost upon which is the "Dismal Science" (Economics), where normal distributions have been assumed for economic processes that have been shown to not justify such an assumption (fat tails).
Also see: https://en.wikipedia.org/wiki/Misuse_of_statistics for the misuse of statistics.
(**) George never had the time to properly discuss statistical methods. He only had a chance to show a bullet point about it as a title saying that of course statistical analysis is recommended. I doubt that George would advocate the use of standard deviation error bars based on sample data for experiments where a statistical distribution is unknown and cannot be justified by histograms or even for the case of a known underlying normal distribution when the number of experiments is too small. -
#1326 by WarpTech on 08 Oct, 2016 22:43
-
Let the mud fly!
EM DRIVE THEORY - GRAVITY IN A CAN (1st Draft, comments welcome!)
https://www.researchgate.net/publication/308948407_EM_DRIVE_THEORY_-_GRAVITY_IN_A_CAN
References:
AN ENGINEERING MODEL OF QUANTUM GRAVITY (SLIDES)
https://www.researchgate.net/publication/308891927_AN_ENGINEERING_MODEL_OF_QUANTUM_GRAVITY_SLIDES
AN ENGINEERING MODEL OF QUANTUM GRAVITY
https://www.researchgate.net/publication/305501551_AN_ENGINEERING_MODEL_OF_QUANTUM_GRAVITY
We need to update the Wiki.
Thanks,
Todd
-
#1327 by SeeShells on 09 Oct, 2016 00:40
-
There are a few things that are not related to the DUT and those are related to the test stand. What is the max/min resolution of the device and are the deviations linear or not. Are there thermal errors that can be shown? These relate to even when running one or just a few tests....Good ideas and you betcha I see the reasoning, you're a very sharp guy. Also you got to see George's presentation at the Advanced Propulsion Workshop as well and the many details needed to reduce the error bars. Did you notice that in the MEGA and MET data there were no error bars? That bothers me. (Dr. Rodal, you have a answer?)...
Shell, please remember, for example, the graph by Woodward/Fearn showing measured force vs. voltage, for different voltages, which did include the range of experimental results for each plotted point, and hence included a statistical assessment of the data.
Shell
"Error bars" (as often used by the Quality Control, and the Social Sciences community) often represent one standard deviation. From a rigorous statistical viewpoint it is inappropriate to use standard deviation as a measure of dispersion for a process with undetermined statistical distribution because the use of one standard deviation presumes a Gaussian Normal distribution. For experiments where people are still debating what is being measured, including what is the physical process responsible, there may not be a basis upon which to assume a Gaussian distribution (which is often assumed by Quality Control people and people involved in the Social Sciences because a) it is the simplest distribution that allows simple, explicit, formulas for calculation and b) many random processes in Quality Control and Social Sciences approximately obey it (*), but that is not applicable for a variety of well known processes). The appropriate approach would be to plot histograms to assess the proper distribution law, but unfortunately many experimental papers lack the number of statistical samples to make such determinations.
Personally, when I showed plots comparing experimental data with calculations, I plotted all the experimental points . This is the approach advised by great statisticians like Tukey and Mosteller, which I think is much superior to anyone that would hide the number of experimental points (the sample population) behind a "standard deviation error bar" that is not justified statistically when there is a small number of experiments and the distribution is normal, or when the distribution is not normal (for any number of statistical samples), or when the distribution is unknown (displaying standard deviation error bars in those cases would constitute a misuse of statistics).
______________
(*) And many other processes do not, foremost upon which is the "Dismal Science" (Economics), where normal distributions have been assumed for economic processes that have been shown to not justify such an assumption (fat tails).
Also see: https://en.wikipedia.org/wiki/Misuse_of_statistics for the misuse of statistics.
(**) George never had the time to properly discuss statistical methods. He only had a chance to show a bullet point about it as a title saying that of course statistical analysis is recommended. I doubt that George would advocate the use of standard deviation error bars based on sample data for experiments where a statistical distribution is unknown and cannot be justified by histograms or even for the case of a known underlying normal distribution when the number of experiments is too small.
Shell
-
#1328 by Rodal on 09 Oct, 2016 00:46
-
...There are a few things that are not related to the DUT and those are related to the test stand. What is the max/min resolution of the device and are the deviations linear or not. Are there thermal errors that can be shown? These relate to even when running one or just a few tests.
Yes of course, but that is not a requirement for the use of standard deviation error bars in plots of results where one cannot justify assuming a Gaussian distribution. (or where the number of experiments is too small so that the sample standard deviation is unrepresentative of the true statistical standard deviation). Using standard deviation error bars for presentation of such data in such cases would constitute a misuse of statistical methods.
Shell
An error bar is really a shorthand way to parameterize a Probability Density Function (PDF). Most often this means pretending the PDF is Gaussian. Many experimenter's sins with error bars come from incorrectly assuming Gaussian distributions or having too few data points.
Again, in such a case it is better to plot all the experimental points (as I presented) instead of using standard deviation error bars. -
#1329 by RotoSequence on 09 Oct, 2016 01:00
-
Let the mud fly!
EM DRIVE THEORY - GRAVITY IN A CAN (1st Draft, comments welcome!)
https://www.researchgate.net/publication/308948407_EM_DRIVE_THEORY_-_GRAVITY_IN_A_CAN
References:
AN ENGINEERING MODEL OF QUANTUM GRAVITY (SLIDES)
https://www.researchgate.net/publication/308891927_AN_ENGINEERING_MODEL_OF_QUANTUM_GRAVITY_SLIDES
AN ENGINEERING MODEL OF QUANTUM GRAVITY
https://www.researchgate.net/publication/305501551_AN_ENGINEERING_MODEL_OF_QUANTUM_GRAVITY
We need to update the Wiki.
Thanks,
Todd
So what should happen if a high Q-factor cavity builds up a lot of energy inside, and the resonant state is disrupted by some mechanism such as cavity deformation? Would the sudden reduction in ground state create a large impulse? -
#1330 by WarpTech on 09 Oct, 2016 01:47
-
Let the mud fly!
EM DRIVE THEORY - GRAVITY IN A CAN (1st Draft, comments welcome!)
https://www.researchgate.net/publication/308948407_EM_DRIVE_THEORY_-_GRAVITY_IN_A_CAN
References:
AN ENGINEERING MODEL OF QUANTUM GRAVITY (SLIDES)
https://www.researchgate.net/publication/308891927_AN_ENGINEERING_MODEL_OF_QUANTUM_GRAVITY_SLIDES
AN ENGINEERING MODEL OF QUANTUM GRAVITY
https://www.researchgate.net/publication/305501551_AN_ENGINEERING_MODEL_OF_QUANTUM_GRAVITY
We need to update the Wiki.
Thanks,
Todd
So what should happen if a high Q-factor cavity builds up a lot of energy inside, and the resonant state is disrupted by some mechanism such as cavity deformation? Would the sudden reduction in ground state create a large impulse?
Absolutely! Just like popping a balloon. I had the idea of using a shutter a long time ago, but I think that reaction would be more like a photon rocket impulse than a gravity impulse. We could think of Zeta as the leakage, if the big end were say, partially transparent and partially reflective, like the output mirror of a laser. The thrust should increase a tiny bit due to the photon rocket effects, and then it becomes a Gravity Assisted Photon Rocket, or GAP Drive.
We'd need a new Thread.
-
#1331 by RERT on 09 Oct, 2016 02:58
-
Well, I found the error I predicted in my post above, and I've fixed it. Overall conclusion is not altered.
-
#1332 by ThinkerX on 09 Oct, 2016 04:58
-
Warp Tech - the math in your paper is beyond me.
However, I believe it is 'frustum,' not 'frustm.'
Might I humbly suggest you check the equations for similar typo type errors? -
#1333 by WarpTech on 09 Oct, 2016 05:13
-
Warp Tech - the math in your paper is beyond me.
However, I believe it is 'frustum,' not 'frustm.'
Might I humbly suggest you check the equations for similar typo type errors?
Thank you for catching the typos. As I said, "1st Draft"... I'm sure there's more to come.
-
#1334 by Bob Woods on 09 Oct, 2016 06:58
-
Todd, you keep doing what you're doing.
-
#1335 by masterharper1082 on 09 Oct, 2016 08:21
-
Great work! Two more nitpicks inside the quotation marks, since my reader doesn't support more elaborate formatting...Warp Tech - the math in your paper is beyond me.
However, I believe it is 'frustum,' not 'frustm.'
Might I humbly suggest you check the equations for similar typo type errors?
Thank you for catching the typos. As I said, "1st Draft"... I'm sure there's more to come.
1. Once again, simplify by "substituting", ωs~c0χm,nRs,
2. If the rate of decay at each end is asymmetrical, such that the big end absorbs and dissipates more power than the small end", t"he stored energy will have a tendency to "fall" toward the big end.
Since your model appears to rely on differential dissipation, could there be a double benefit by making the small side superconducting, and allowing the big side to do all of the dissipation?
Your 1/f relationship indicates the potential for more force in a smaller package by going to higher frequencies, holding the resonant mode shape constant. The trick is to keep the dissipating end cool through good heat transfer practices, and to produce the resonant mode at the smaller distance scale. Am I understanding your predictions correctly?
Once we have repeatable test setups with known error bounds, these predictions should be testable to confirm the validity of your model.
mh -
#1336 by Monomorphic on 09 Oct, 2016 13:53
-
Let the mud fly!
"Performance also improves with the lowest order mode frequency"
So your theory predicts improved thrust for TM010, TE111, TM011, and TE011 over the often discussed TE013? Would TM010 have the greatest thrust since it is the lowest order mode?
It doesn't look like you've taken spherical end-caps into account. There is a significant difference in the internal volume when using spherical end-caps. I want to work up a FEKO model that incorporates your "ratio of the radii, Rb/Rs = 1.500." and "The length L should be no longer than Rs is wide, shorter is better and the performance improves as 1/L." Should I use flat or spherical end-caps?
-
#1337 by dustinthewind on 09 Oct, 2016 15:14
-
Great work! Two more nitpicks inside the quotation marks, since my reader doesn't support more elaborate formatting...Warp Tech - the math in your paper is beyond me.
However, I believe it is 'frustum,' not 'frustm.'
Might I humbly suggest you check the equations for similar typo type errors?
Thank you for catching the typos. As I said, "1st Draft"... I'm sure there's more to come.
1. Once again, simplify by "substituting", ωs~c0χm,nRs,
2. If the rate of decay at each end is asymmetrical, such that the big end absorbs and dissipates more power than the small end", t"he stored energy will have a tendency to "fall" toward the big end.
Since your model appears to rely on differential dissipation, could there be a double benefit by making the small side superconducting, and allowing the big side to do all of the dissipation?
Your 1/f relationship indicates the potential for more force in a smaller package by going to higher frequencies, holding the resonant mode shape constant. The trick is to keep the dissipating end cool through good heat transfer practices, and to produce the resonant mode at the smaller distance scale. Am I understanding your predictions correctly?
Once we have repeatable test setups with known error bounds, these predictions should be testable to confirm the validity of your model.
mh
I think you mean to use lower frequencies to maximize thrust/larger cavities. where g~1/f^2 or something like that. I think it seems similar to what Rodal was saying about using larger cavities to lower the frequency and maximize thrust. Both notsureofit's and WarpTech's math seem to have this. -
#1338 by WarpTech on 09 Oct, 2016 15:42
-
Let the mud fly!
"Performance also improves with the lowest order mode frequency"
So your theory predicts improved thrust for TM010, TE111, TM011, and TE011 over the often discussed TE013? Would TM010 have the greatest thrust since it is the lowest order mode?
It doesn't look like you've taken spherical end-caps into account. There is a significant difference in the internal volume when using spherical end-caps. I want to work up a FEKO model that incorporates your "ratio of the radii, Rb/Rs = 1.500." and "The length L should be no longer than Rs is wide, shorter is better and the performance improves as 1/L." Should I use flat or spherical end-caps?
My model uses a cylinder with flat end-caps. Spherical end caps should increase the Q and will increase the thrust, provided the differential dissipation is the same and favors the big end.
Regarding mode "shapes", I'm still mulling that over but the model says the lower the frequency and shorter the length, the better the performance, like a tapered pill-box.
-
#1339 by WarpTech on 09 Oct, 2016 15:46
-
Great work! Two more nitpicks inside the quotation marks, since my reader doesn't support more elaborate formatting...Warp Tech - the math in your paper is beyond me.
However, I believe it is 'frustum,' not 'frustm.'
Might I humbly suggest you check the equations for similar typo type errors?
Thank you for catching the typos. As I said, "1st Draft"... I'm sure there's more to come.
1. Once again, simplify by "substituting", ωs~c0χm,nRs,
2. If the rate of decay at each end is asymmetrical, such that the big end absorbs and dissipates more power than the small end", t"he stored energy will have a tendency to "fall" toward the big end.
Since your model appears to rely on differential dissipation, could there be a double benefit by making the small side superconducting, and allowing the big side to do all of the dissipation?
Your 1/f relationship indicates the potential for more force in a smaller package by going to higher frequencies, holding the resonant mode shape constant. The trick is to keep the dissipating end cool through good heat transfer practices, and to produce the resonant mode at the smaller distance scale. Am I understanding your predictions correctly?
Once we have repeatable test setups with known error bounds, these predictions should be testable to confirm the validity of your model.
mh
Thanks for the corrections! Your conjectures are correct.
According to the model, size scales linearly with frequency, but thrust only scales with increased power. IMO, a smaller device will have a more difficult time keeping cool and staying asymmetrical. My thoughts are toward larger devices running at MHz, so they can get higher power input and have the surface area to dissipate the heat.
We'd need a new Thread. 