Quote from: SteveD on 12/05/2015 04:32 amQuote from: rfmwguy on 12/04/2015 07:39 pm"NEGATIVE experimental results are under-reported or not reported at all. Negative results at a University (by Zellerium) is not even reported in the EM Drive wiki ( http://emdrive.wiki/Experimental_Results ), the last time that I checked. Several experimenters that started to report their build up, stopped reporting and are practically unheard of. At least one of the experimenters asked -in these threads- not to report the negative experimental results.An effort has to be made to report and properly document ALL NEGATIVE resultsAn effort has to be made to be objective."Certainly worth repeating. I did ask an experimenter if this could be labeled Null. Believe this is what you are referring to. I videotaped and posted 2 or 3 flight tests I considered Null before FT#2B and reconfirmed displacement changes in FT#2C, So 3 of 5 flight tests were Null on NSF-1701. I only documented the last one on the wiki page.Didn't Zellerium say he was getting something with a frustum, null with cylindrical and that he was discontinueing as he didn't think he could finish prior to his expected graduation date, and was doing this as a senior project?I really wish we could get a better analysis of FT 1. Would really like to see if having the rf feed on the small base produced a null (and since most of the heat is coming off the magnetron housing a large different between 1 and later tests would help rule out thermal currents as the source of thrust).Zellerium is a busy boy right before graduation, so I don't think he will respond. Where I thought he left it was Null on symmetrical cavity, no time for frustum. I do not recall reading a test report but seems to me I saw some pics posted here. Hope Kurt can get back here sometime and fill us in.
Quote from: rfmwguy on 12/04/2015 07:39 pm"NEGATIVE experimental results are under-reported or not reported at all. Negative results at a University (by Zellerium) is not even reported in the EM Drive wiki ( http://emdrive.wiki/Experimental_Results ), the last time that I checked. Several experimenters that started to report their build up, stopped reporting and are practically unheard of. At least one of the experimenters asked -in these threads- not to report the negative experimental results.An effort has to be made to report and properly document ALL NEGATIVE resultsAn effort has to be made to be objective."Certainly worth repeating. I did ask an experimenter if this could be labeled Null. Believe this is what you are referring to. I videotaped and posted 2 or 3 flight tests I considered Null before FT#2B and reconfirmed displacement changes in FT#2C, So 3 of 5 flight tests were Null on NSF-1701. I only documented the last one on the wiki page.Didn't Zellerium say he was getting something with a frustum, null with cylindrical and that he was discontinueing as he didn't think he could finish prior to his expected graduation date, and was doing this as a senior project?I really wish we could get a better analysis of FT 1. Would really like to see if having the rf feed on the small base produced a null (and since most of the heat is coming off the magnetron housing a large different between 1 and later tests would help rule out thermal currents as the source of thrust).
"NEGATIVE experimental results are under-reported or not reported at all. Negative results at a University (by Zellerium) is not even reported in the EM Drive wiki ( http://emdrive.wiki/Experimental_Results ), the last time that I checked. Several experimenters that started to report their build up, stopped reporting and are practically unheard of. At least one of the experimenters asked -in these threads- not to report the negative experimental results.An effort has to be made to report and properly document ALL NEGATIVE resultsAn effort has to be made to be objective."Certainly worth repeating. I did ask an experimenter if this could be labeled Null. Believe this is what you are referring to. I videotaped and posted 2 or 3 flight tests I considered Null before FT#2B and reconfirmed displacement changes in FT#2C, So 3 of 5 flight tests were Null on NSF-1701. I only documented the last one on the wiki page.
Quote from: rfmwguy on 12/05/2015 01:22 pmQuote from: SteveD on 12/05/2015 04:32 amQuote from: rfmwguy on 12/04/2015 07:39 pm"NEGATIVE experimental results are under-reported or not reported at all. Negative results at a University (by Zellerium) is not even reported in the EM Drive wiki ( http://emdrive.wiki/Experimental_Results ), the last time that I checked. Several experimenters that started to report their build up, stopped reporting and are practically unheard of. At least one of the experimenters asked -in these threads- not to report the negative experimental results.An effort has to be made to report and properly document ALL NEGATIVE resultsAn effort has to be made to be objective."Certainly worth repeating. I did ask an experimenter if this could be labeled Null. Believe this is what you are referring to. I videotaped and posted 2 or 3 flight tests I considered Null before FT#2B and reconfirmed displacement changes in FT#2C, So 3 of 5 flight tests were Null on NSF-1701. I only documented the last one on the wiki page.Didn't Zellerium say he was getting something with a frustum, null with cylindrical and that he was discontinueing as he didn't think he could finish prior to his expected graduation date, and was doing this as a senior project?I really wish we could get a better analysis of FT 1. Would really like to see if having the rf feed on the small base produced a null (and since most of the heat is coming off the magnetron housing a large different between 1 and later tests would help rule out thermal currents as the source of thrust).Zellerium is a busy boy right before graduation, so I don't think he will respond. Where I thought he left it was Null on symmetrical cavity, no time for frustum. I do not recall reading a test report but seems to me I saw some pics posted here. Hope Kurt can get back here sometime and fill us in.Rfmwguy,Was he the one who said he had someone who wanted to continue, as some sort of class project?
Quote from: OnlyMe on 12/05/2015 01:54 pmQuote from: rfmwguy on 12/05/2015 01:22 pmQuote from: SteveD on 12/05/2015 04:32 amQuote from: rfmwguy on 12/04/2015 07:39 pm"NEGATIVE experimental results are under-reported or not reported at all. Negative results at a University (by Zellerium) is not even reported in the EM Drive wiki ( http://emdrive.wiki/Experimental_Results ), the last time that I checked. Several experimenters that started to report their build up, stopped reporting and are practically unheard of. At least one of the experimenters asked -in these threads- not to report the negative experimental results.An effort has to be made to report and properly document ALL NEGATIVE resultsAn effort has to be made to be objective."Certainly worth repeating. I did ask an experimenter if this could be labeled Null. Believe this is what you are referring to. I videotaped and posted 2 or 3 flight tests I considered Null before FT#2B and reconfirmed displacement changes in FT#2C, So 3 of 5 flight tests were Null on NSF-1701. I only documented the last one on the wiki page.Didn't Zellerium say he was getting something with a frustum, null with cylindrical and that he was discontinueing as he didn't think he could finish prior to his expected graduation date, and was doing this as a senior project?I really wish we could get a better analysis of FT 1. Would really like to see if having the rf feed on the small base produced a null (and since most of the heat is coming off the magnetron housing a large different between 1 and later tests would help rule out thermal currents as the source of thrust).Zellerium is a busy boy right before graduation, so I don't think he will respond. Where I thought he left it was Null on symmetrical cavity, no time for frustum. I do not recall reading a test report but seems to me I saw some pics posted here. Hope Kurt can get back here sometime and fill us in.Rfmwguy,Was he the one who said he had someone who wanted to continue, as some sort of class project?Yes, that is where he left off IIRC.<edit> found it: http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=37642.0;attach=837146
Didn't Zellerium say he was getting something with a frustum
I thought he left it was Null on symmetrical cavity, no time for frustum.
Let me summarize this, as I've read what Woodward thinks about the EmDrive: he does not believe that an EmDrive without a dielectric within can work at all; and while he thinks an EmDrive with an internal electrostrictive dielectric could work because of some Mach effect, he denies the reality of any quantum vacuum plasma-based propulsion.
Quote from: rfmwguy on 12/05/2015 02:31 pmQuote from: OnlyMe on 12/05/2015 01:54 pmQuote from: rfmwguy on 12/05/2015 01:22 pmQuote from: SteveD on 12/05/2015 04:32 amQuote from: rfmwguy on 12/04/2015 07:39 pm"NEGATIVE experimental results are under-reported or not reported at all. Negative results at a University (by Zellerium) is not even reported in the EM Drive wiki ( http://emdrive.wiki/Experimental_Results ), the last time that I checked. Several experimenters that started to report their build up, stopped reporting and are practically unheard of. At least one of the experimenters asked -in these threads- not to report the negative experimental results.An effort has to be made to report and properly document ALL NEGATIVE resultsAn effort has to be made to be objective."Certainly worth repeating. I did ask an experimenter if this could be labeled Null. Believe this is what you are referring to. I videotaped and posted 2 or 3 flight tests I considered Null before FT#2B and reconfirmed displacement changes in FT#2C, So 3 of 5 flight tests were Null on NSF-1701. I only documented the last one on the wiki page.Didn't Zellerium say he was getting something with a frustum, null with cylindrical and that he was discontinueing as he didn't think he could finish prior to his expected graduation date, and was doing this as a senior project?I really wish we could get a better analysis of FT 1. Would really like to see if having the rf feed on the small base produced a null (and since most of the heat is coming off the magnetron housing a large different between 1 and later tests would help rule out thermal currents as the source of thrust).Zellerium is a busy boy right before graduation, so I don't think he will respond. Where I thought he left it was Null on symmetrical cavity, no time for frustum. I do not recall reading a test report but seems to me I saw some pics posted here. Hope Kurt can get back here sometime and fill us in.Rfmwguy,Was he the one who said he had someone who wanted to continue, as some sort of class project?Yes, that is where he left off IIRC.<edit> found it: http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=37642.0;attach=837146The attached link (http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=37642.0;attach=837146) shows that:Concerning the question QuoteDidn't Zellerium say he was getting something with a frustum , NO, the document does not report getting any force measurements using the frustum of a cone. The attached document is a short document proposing measurements. No experimental measurements with a frustrum of a cone are reported.Essentially my recollection is identical to rfmwguy:QuoteI thought he left it was Null on symmetrical cavity, no time for frustum. Except that "symmetrical" is not a precise, correct, statement, because any cavity (including cavities with uniform cross-section for example circular or rectangular cross-section) containing an unsymmetrically placed dielectric are effectively unsymmetric for the purposes of electromagnetic wave propagation (as there is a power loss in the dielectric but effectively insignificant power loss in the empty section, and as the group and phase velocities are different in the dielectric section than in the empty section of the cavity).The unsymmetry produced by an unsymmetrically placed dielectric in a cavity with uniform cross-section have been discussed by Prof. Woodward, Paul March, Notsosureofit, Mulletron and even Roger Shawyer (in his first EM Drive patent).The negative result of Zellerium with such a cavity is important to be documented (it is still not listed in the EM Drive wiki).
...* Writing about science fiction flying cars and short-trips to Pluto (or even the Stars ! ) that violate conservation of energy by Shawyer, instead of providing new, strong experimental evidence that can be independently verified...are not positive
Quote from: Rodal on 12/05/2015 02:46 amQuote from: aero on 12/05/2015 02:23 amQuote from: Rodal on 12/05/2015 01:17 amQuote from: aero on 12/05/2015 01:10 am...Dr. Rodal - I uploaded these progressively longer runs of the Yang-Shell 6 degree model just about the time than you went dark. I did notify you but have no indication that you received my notification, so I repeat it here. The csv data is here. Please read the data description there.https://drive.google.com/folderview?id=0B1XizxEfB23tRm41bVFtM1pVYlU&usp=sharingThese runs range in length from 32 cycles to 2048 cycles of the drive frequency. That is not a long time, I guess something like 2048cycles/2.45 GHz, or close to a microsecond of simulated time, but something like 2 days wall clock time for each of the 2048 cycle runs. On the other current subject, I did modify the copper conductance model after seeing the same information from you and three other sources, two of which you linked. The model code now reads as.(material (make medium (epsilon epsilon_r) (D-conductivity CU-D-conduct)))where epdilon_r = 1 numerical value of CU-D-conduct is: (printed from a run log file) CU-D-conduct = 43393352.18305066old-CU-conduct= 4904277351.366935 ratio of new over old = 0.008848062430840282As I gain some meep experience in working with this new copper model, Shell will know and I'm sure data will be posted. As for now I have learned that it cuts Q a lot and Q is strongly dependent on meep resolution with this new model, not in the same way as for the older models. The other data hinted at is ... Well as I've only ran one cavity case, cylindrical at that, I should keep my suspicions under my hat until I have at least some confirmation.And if that is not the correct numerical value as you understand it, please tell me promptly. A simple "still not right, aero" would have me double checking again though the correct numerical value would be helpful. Rest assured, I will not change the model until I satisfy myself that it needs to be changed.Thanks, that is the correct ratio of new to old conductivity: 0.0088480= 1/113.019...The new one should be 113.019 times lower.Remember that this is for high-content copper, as intended by DeltaMass.Other materials (silver, brass, gold, etc.) need to be ratioed by their corresponding conductivities (or their inverse, resistivities).Ok, here is a very nice data set of conductivities/resistivities:http://eddy-current.com/conductivity-of-metals-sorted-by-resistivity/DeltaMass' conductivity number was 3.25E+8 determined from considerations of the electron cloud, as I recall. As Murphy would dictate, that number does not appear in this data list. I can search further but I think it is safe to assume that 3.25E+8 will not appear elsewhere exactly. What is the recommended way to adjust ratios to maintain consistency, or would it be less complex to simply use the meep units conversion wiki, as I did for this modification?aeroNo !. 3.25E+8 was not the conductivity. 3.25E+8 is the number in SI units that corresponds to the expression used in Meep: it is not the conductivity, it is instead the imaginary part of the relative complex permittivityTherefore 3.25E+8 corresponds instead to epsilon"/epsilon_o = 0.00288/epsilon_o(The number really is 3.252698....E+8)The numerical value of the imaginary part of the permittivity: epsilon"=0.00288 was given by DeltaMassThe conductivity in SI Units that corresponds to epsilon"=0.00288 is:conductivity = omega * epsilon" = 2 Pi frequency 0.00288 = 2 Pi 2.4E+9 * 0.00288 = 4.342937 E+7which is almost 10 times smaller than 3.25E+8Please also recall that DeltaMass was conscious that it is incorrect to take 0.00288 as a constant !That value is a function of frequency.What is approximately constant in this regime is the conductivity itself.DeltaMass gave you explicit instructions to keep the conductivity constant, at other frequencies: for example, the conductivity at 1 GHz is also 4.342937 E+7, so at 1 GHz you should input into Meep (3.25...E+8 ) *2.4 = 7.8 E+8, for exampleSo, for pure Silver, for example (from the table in the link in your post), Conductivity = 6.090E+07So, instead of 3.25...E+8 for copper, you have to use ( 6.090E+07/4.342937 E+7 ) *3.25...E+8 at 2.4GHzin other words, at 2.4 GHz, your input to Meep for pure Silver should be 1.402276 times higher than for the copper value given by DeltaMass. About 40% higher, whether in SI units or in Meep units.In other words, everything else being the same, the quality of resonance (Q) should be about 40% higher with pure silver than with copper.
Quote from: aero on 12/05/2015 02:23 amQuote from: Rodal on 12/05/2015 01:17 amQuote from: aero on 12/05/2015 01:10 am...Dr. Rodal - I uploaded these progressively longer runs of the Yang-Shell 6 degree model just about the time than you went dark. I did notify you but have no indication that you received my notification, so I repeat it here. The csv data is here. Please read the data description there.https://drive.google.com/folderview?id=0B1XizxEfB23tRm41bVFtM1pVYlU&usp=sharingThese runs range in length from 32 cycles to 2048 cycles of the drive frequency. That is not a long time, I guess something like 2048cycles/2.45 GHz, or close to a microsecond of simulated time, but something like 2 days wall clock time for each of the 2048 cycle runs. On the other current subject, I did modify the copper conductance model after seeing the same information from you and three other sources, two of which you linked. The model code now reads as.(material (make medium (epsilon epsilon_r) (D-conductivity CU-D-conduct)))where epdilon_r = 1 numerical value of CU-D-conduct is: (printed from a run log file) CU-D-conduct = 43393352.18305066old-CU-conduct= 4904277351.366935 ratio of new over old = 0.008848062430840282As I gain some meep experience in working with this new copper model, Shell will know and I'm sure data will be posted. As for now I have learned that it cuts Q a lot and Q is strongly dependent on meep resolution with this new model, not in the same way as for the older models. The other data hinted at is ... Well as I've only ran one cavity case, cylindrical at that, I should keep my suspicions under my hat until I have at least some confirmation.And if that is not the correct numerical value as you understand it, please tell me promptly. A simple "still not right, aero" would have me double checking again though the correct numerical value would be helpful. Rest assured, I will not change the model until I satisfy myself that it needs to be changed.Thanks, that is the correct ratio of new to old conductivity: 0.0088480= 1/113.019...The new one should be 113.019 times lower.Remember that this is for high-content copper, as intended by DeltaMass.Other materials (silver, brass, gold, etc.) need to be ratioed by their corresponding conductivities (or their inverse, resistivities).Ok, here is a very nice data set of conductivities/resistivities:http://eddy-current.com/conductivity-of-metals-sorted-by-resistivity/DeltaMass' conductivity number was 3.25E+8 determined from considerations of the electron cloud, as I recall. As Murphy would dictate, that number does not appear in this data list. I can search further but I think it is safe to assume that 3.25E+8 will not appear elsewhere exactly. What is the recommended way to adjust ratios to maintain consistency, or would it be less complex to simply use the meep units conversion wiki, as I did for this modification?aeroNo !. 3.25E+8 was not the conductivity. 3.25E+8 is the number in SI units that corresponds to the expression used in Meep: it is not the conductivity, it is instead the imaginary part of the relative complex permittivityTherefore 3.25E+8 corresponds instead to epsilon"/epsilon_o = 0.00288/epsilon_o(The number really is 3.252698....E+8)The numerical value of the imaginary part of the permittivity: epsilon"=0.00288 was given by DeltaMassThe conductivity in SI Units that corresponds to epsilon"=0.00288 is:conductivity = omega * epsilon" = 2 Pi frequency 0.00288 = 2 Pi 2.4E+9 * 0.00288 = 4.342937 E+7which is almost 10 times smaller than 3.25E+8Please also recall that DeltaMass was conscious that it is incorrect to take 0.00288 as a constant !That value is a function of frequency.What is approximately constant in this regime is the conductivity itself.DeltaMass gave you explicit instructions to keep the conductivity constant, at other frequencies: for example, the conductivity at 1 GHz is also 4.342937 E+7, so at 1 GHz you should input into Meep (3.25...E+8 ) *2.4 = 7.8 E+8, for example
Quote from: Rodal on 12/05/2015 01:17 amQuote from: aero on 12/05/2015 01:10 am...Dr. Rodal - I uploaded these progressively longer runs of the Yang-Shell 6 degree model just about the time than you went dark. I did notify you but have no indication that you received my notification, so I repeat it here. The csv data is here. Please read the data description there.https://drive.google.com/folderview?id=0B1XizxEfB23tRm41bVFtM1pVYlU&usp=sharingThese runs range in length from 32 cycles to 2048 cycles of the drive frequency. That is not a long time, I guess something like 2048cycles/2.45 GHz, or close to a microsecond of simulated time, but something like 2 days wall clock time for each of the 2048 cycle runs. On the other current subject, I did modify the copper conductance model after seeing the same information from you and three other sources, two of which you linked. The model code now reads as.(material (make medium (epsilon epsilon_r) (D-conductivity CU-D-conduct)))where epdilon_r = 1 numerical value of CU-D-conduct is: (printed from a run log file) CU-D-conduct = 43393352.18305066old-CU-conduct= 4904277351.366935 ratio of new over old = 0.008848062430840282As I gain some meep experience in working with this new copper model, Shell will know and I'm sure data will be posted. As for now I have learned that it cuts Q a lot and Q is strongly dependent on meep resolution with this new model, not in the same way as for the older models. The other data hinted at is ... Well as I've only ran one cavity case, cylindrical at that, I should keep my suspicions under my hat until I have at least some confirmation.And if that is not the correct numerical value as you understand it, please tell me promptly. A simple "still not right, aero" would have me double checking again though the correct numerical value would be helpful. Rest assured, I will not change the model until I satisfy myself that it needs to be changed.Thanks, that is the correct ratio of new to old conductivity: 0.0088480= 1/113.019...The new one should be 113.019 times lower.Remember that this is for high-content copper, as intended by DeltaMass.Other materials (silver, brass, gold, etc.) need to be ratioed by their corresponding conductivities (or their inverse, resistivities).Ok, here is a very nice data set of conductivities/resistivities:http://eddy-current.com/conductivity-of-metals-sorted-by-resistivity/DeltaMass' conductivity number was 3.25E+8 determined from considerations of the electron cloud, as I recall. As Murphy would dictate, that number does not appear in this data list. I can search further but I think it is safe to assume that 3.25E+8 will not appear elsewhere exactly. What is the recommended way to adjust ratios to maintain consistency, or would it be less complex to simply use the meep units conversion wiki, as I did for this modification?aero
Quote from: aero on 12/05/2015 01:10 am...Dr. Rodal - I uploaded these progressively longer runs of the Yang-Shell 6 degree model just about the time than you went dark. I did notify you but have no indication that you received my notification, so I repeat it here. The csv data is here. Please read the data description there.https://drive.google.com/folderview?id=0B1XizxEfB23tRm41bVFtM1pVYlU&usp=sharingThese runs range in length from 32 cycles to 2048 cycles of the drive frequency. That is not a long time, I guess something like 2048cycles/2.45 GHz, or close to a microsecond of simulated time, but something like 2 days wall clock time for each of the 2048 cycle runs. On the other current subject, I did modify the copper conductance model after seeing the same information from you and three other sources, two of which you linked. The model code now reads as.(material (make medium (epsilon epsilon_r) (D-conductivity CU-D-conduct)))where epdilon_r = 1 numerical value of CU-D-conduct is: (printed from a run log file) CU-D-conduct = 43393352.18305066old-CU-conduct= 4904277351.366935 ratio of new over old = 0.008848062430840282As I gain some meep experience in working with this new copper model, Shell will know and I'm sure data will be posted. As for now I have learned that it cuts Q a lot and Q is strongly dependent on meep resolution with this new model, not in the same way as for the older models. The other data hinted at is ... Well as I've only ran one cavity case, cylindrical at that, I should keep my suspicions under my hat until I have at least some confirmation.And if that is not the correct numerical value as you understand it, please tell me promptly. A simple "still not right, aero" would have me double checking again though the correct numerical value would be helpful. Rest assured, I will not change the model until I satisfy myself that it needs to be changed.Thanks, that is the correct ratio of new to old conductivity: 0.0088480= 1/113.019...The new one should be 113.019 times lower.Remember that this is for high-content copper, as intended by DeltaMass.Other materials (silver, brass, gold, etc.) need to be ratioed by their corresponding conductivities (or their inverse, resistivities).
...Dr. Rodal - I uploaded these progressively longer runs of the Yang-Shell 6 degree model just about the time than you went dark. I did notify you but have no indication that you received my notification, so I repeat it here. The csv data is here. Please read the data description there.https://drive.google.com/folderview?id=0B1XizxEfB23tRm41bVFtM1pVYlU&usp=sharingThese runs range in length from 32 cycles to 2048 cycles of the drive frequency. That is not a long time, I guess something like 2048cycles/2.45 GHz, or close to a microsecond of simulated time, but something like 2 days wall clock time for each of the 2048 cycle runs. On the other current subject, I did modify the copper conductance model after seeing the same information from you and three other sources, two of which you linked. The model code now reads as.(material (make medium (epsilon epsilon_r) (D-conductivity CU-D-conduct)))where epdilon_r = 1 numerical value of CU-D-conduct is: (printed from a run log file) CU-D-conduct = 43393352.18305066old-CU-conduct= 4904277351.366935 ratio of new over old = 0.008848062430840282As I gain some meep experience in working with this new copper model, Shell will know and I'm sure data will be posted. As for now I have learned that it cuts Q a lot and Q is strongly dependent on meep resolution with this new model, not in the same way as for the older models. The other data hinted at is ... Well as I've only ran one cavity case, cylindrical at that, I should keep my suspicions under my hat until I have at least some confirmation.And if that is not the correct numerical value as you understand it, please tell me promptly. A simple "still not right, aero" would have me double checking again though the correct numerical value would be helpful. Rest assured, I will not change the model until I satisfy myself that it needs to be changed.
Just a reminder to everyone the math typeit web page is really handy for entering equations that use Greek symbols.http://math.typeit.org/Instead of:CONDUCTIVITY = omega * epsilon" = omega * tanDelta * epsilon' = omega * tanDelta * epsilon0*epsilon_r = omega * 3.25*(10^8)σ = Ω × εΩ = 2πf = 2 * π * 2.4 GHz
Except that "symmetrical" is not a precise, correct, statement, because any cavity (including cavities with uniform cross-section for example circular or rectangular cross-section) containing an unsymmetrically placed dielectric are effectively unsymmetric for the purposes of electromagnetic wave propagation (as there is a power loss in the dielectric but effectively insignificant power loss in the empty section, and as the group and phase velocities are different in the dielectric section than in the empty section of the cavity).. . .The negative result of Zellerium at a US University with such a cavity is important to be documented (it is still not listed in the EM Drive wiki).
Update: I spent the weekend polishing the cylinder and it looks much better (couldn't get some of the bigger scratches out unfortunately). However it definitely improved the VNA plots. Today we tested the attached resonance, seems to be matched very well at 2.434 GHz. But we still don't have a spectrum analysis of our magnetron, so we aren't sure how the power is distributed and how that changes over time. The EMF strips arrived to hopefully improve electrical connection between the waveguide and cylinder without the messy Ox Gard. We heard some arcing when the magnetron fired up which seemed to stop (or at least quiet down) as the test continued. We definitely have some deflection but it could easily be due to asymmetric current flow to ground. The x deflection is actually a torque on the cylinder (clockwise if looking from above). This may also be from misalignment of the cylinder axis. The y deflection is in the right direction (towards the dielectric) and we calculated the force causing it is about 1 mN. Unfortunately our noise level was fairly high so we will need to repeat this test early in the morning when we can get less noise. I'll post the raw data if anyone is interested.The biggest problem is obtaining symmetric current flow: we've shown that we can get significant deflections on resonance with asymmetric grounding which disappear off resonance. When fields don't build up surface currents are too small to cause a disturbance. We may need to weld our movable plate near a resonant position and manufacture some tuning fixture. However at high power we fear a tuning screw would not work (would probably arc). Are there any other relatively easy analog tuning methods for high power out there?The simulations are coming along, I've modeled a coax to waveguide to frustum with a VSWR of 1.15 and quality of ~22,000. Changing the length of the coax shifts the resonant frequency (non-linearly which surprised me) but maintains the quality. One question I have yet to answer: how can I model a coax to waveguide bought from a manufacturer without their proprietary information? (i.e. penetration depth and back wall distance) I was able to determine dimensions that give me a VSWR of 1.2 (their maximum quoted value) and used the dimensions for an RG142 coax. Instead I could model the system with a plane wave excitation, but how do I know the phase of the wave exiting the purchased waveguide? Maybe we should attempt to manufacture this waveguide on our own? I've emailed the manufacturer back with these questions, perhaps I am missing something or over analyzing...We've come to the conclusion that a microwave oven magnetron is far from ideal for this application. (As many of you have said before) The wide bandwidth and unpredictable shifting max power make designing any high quality resonator near impossible. Using a circulator could deliver power at the right frequency but would still filter out the majority of the power. And then we don't know how much power is actually being delivered to the cavity... NWPU must have overcome this challenge with a higher tolerance magnetron with a more stable frequency distribution. Anyone have any suggestions, comments, questions? -Kurt
I looked at the wiki page a few times and I don't believe the data accurately shows the outcome of diy experiments. Part of the problem is people like to "data reduce" and cherry pick data. A more scientific method is to establish a protocol to follow for each experiment and to record everything in exacting detail as it is done in a lab notebook. Each experimental run then has a complete paper record that can be referred to later on. Video recording is fine but I don't think it is suitable for recording an experiment. There are too many measurements, changes to the apparatus, etc. that are not available in a video. It appears the EW tests did follow this method but I don't believe they provided data on all their tests. Another aspect of testing is repeatability. If someone does a test they should try to set up the experiment again under the same conditions and run it again. These would be recorded as repeats of one experiment where the protocol has not changed. It may be necessary to allow time for everything to cool down to room temperature and to fiddle with the apparatus so it is setup the same each time. After several runs have been done - maybe 10 - 20 then run some statistical data analysis. What is the standard deviation of the "thrust" that was measured? What is the mean? Plot the results on a bar graph so the whole set of experiments can be seen. Now change some part of the experiment and repeat.
Quote from: Rodal on 12/05/2015 03:07 pmExcept that "symmetrical" is not a precise, correct, statement, because any cavity (including cavities with uniform cross-section for example circular or rectangular cross-section) containing an unsymmetrically placed dielectric are effectively unsymmetric for the purposes of electromagnetic wave propagation (as there is a power loss in the dielectric but effectively insignificant power loss in the empty section, and as the group and phase velocities are different in the dielectric section than in the empty section of the cavity).. . .The negative result of Zellerium at a US University with such a cavity is important to be documented (it is still not listed in the EM Drive wiki).QuoteUpdate: I spent the weekend polishing the cylinder and it looks much better (couldn't get some of the bigger scratches out unfortunately). However it definitely improved the VNA plots. Today we tested the attached resonance, seems to be matched very well at 2.434 GHz. But we still don't have a spectrum analysis of our magnetron, so we aren't sure how the power is distributed and how that changes over time. The EMF strips arrived to hopefully improve electrical connection between the waveguide and cylinder without the messy Ox Gard. We heard some arcing when the magnetron fired up which seemed to stop (or at least quiet down) as the test continued. We definitely have some deflection but it could easily be due to asymmetric current flow to ground. The x deflection is actually a torque on the cylinder (clockwise if looking from above). This may also be from misalignment of the cylinder axis. The y deflection is in the right direction (towards the dielectric) and we calculated the force causing it is about 1 mN. Unfortunately our noise level was fairly high so we will need to repeat this test early in the morning when we can get less noise. I'll post the raw data if anyone is interested.The biggest problem is obtaining symmetric current flow: we've shown that we can get significant deflections on resonance with asymmetric grounding which disappear off resonance. When fields don't build up surface currents are too small to cause a disturbance. We may need to weld our movable plate near a resonant position and manufacture some tuning fixture. However at high power we fear a tuning screw would not work (would probably arc). Are there any other relatively easy analog tuning methods for high power out there?The simulations are coming along, I've modeled a coax to waveguide to frustum with a VSWR of 1.15 and quality of ~22,000. Changing the length of the coax shifts the resonant frequency (non-linearly which surprised me) but maintains the quality. One question I have yet to answer: how can I model a coax to waveguide bought from a manufacturer without their proprietary information? (i.e. penetration depth and back wall distance) I was able to determine dimensions that give me a VSWR of 1.2 (their maximum quoted value) and used the dimensions for an RG142 coax. Instead I could model the system with a plane wave excitation, but how do I know the phase of the wave exiting the purchased waveguide? Maybe we should attempt to manufacture this waveguide on our own? I've emailed the manufacturer back with these questions, perhaps I am missing something or over analyzing...We've come to the conclusion that a microwave oven magnetron is far from ideal for this application. (As many of you have said before) The wide bandwidth and unpredictable shifting max power make designing any high quality resonator near impossible. Using a circulator could deliver power at the right frequency but would still filter out the majority of the power. And then we don't know how much power is actually being delivered to the cavity... NWPU must have overcome this challenge with a higher tolerance magnetron with a more stable frequency distribution. Anyone have any suggestions, comments, questions? -KurtI'd hold off on calling this null for a bit. Doesn't seem that the expirement progressed to the point that a conclusion could be reached.
Quote from: zen-in on 12/05/2015 06:55 pmI looked at the wiki page a few times and I don't believe the data accurately shows the outcome of diy experiments. Part of the problem is people like to "data reduce" and cherry pick data. A more scientific method is to establish a protocol to follow for each experiment and to record everything in exacting detail as it is done in a lab notebook. Each experimental run then has a complete paper record that can be referred to later on. Video recording is fine but I don't think it is suitable for recording an experiment. There are too many measurements, changes to the apparatus, etc. that are not available in a video. It appears the EW tests did follow this method but I don't believe they provided data on all their tests. Another aspect of testing is repeatability. If someone does a test they should try to set up the experiment again under the same conditions and run it again. These would be recorded as repeats of one experiment where the protocol has not changed. It may be necessary to allow time for everything to cool down to room temperature and to fiddle with the apparatus so it is setup the same each time. After several runs have been done - maybe 10 - 20 then run some statistical data analysis. What is the standard deviation of the "thrust" that was measured? What is the mean? Plot the results on a bar graph so the whole set of experiments can be seen. Now change some part of the experiment and repeat.Zen-in, I agree with most of what you say above. In a perfect world it would be good to have every last detail logged and recorded. I expect that within reason the DIYs are keeping such logs. I don't know anything about what is involved in editing Wiki pages and I am not sure that posting every last detail for every experimental run or rerun, on the Wiki page, would be a good idea. For a large part of the Wiki audience it would be just noise. It would be good to have that information available for those with credible experience and the desire to review it.Where I disagree is on the value of the video record. These are DIY adventures that rely in part (large or small) on the donation of $$, by mostly interested lay persons. Both the Wiki page and especially the videos go a long way toward encouraging that kind of support.I cannot say how well any of the DIYs have been at keeping an unpublished log of their efforts. Still I have been impressed by their efforts and achievements.
Quote from: OnlyMe on 12/05/2015 07:29 pmQuote from: zen-in on 12/05/2015 06:55 pmI looked at the wiki page a few times and I don't believe the data accurately shows the outcome of diy experiments. Part of the problem is people like to "data reduce" and cherry pick data. A more scientific method is to establish a protocol to follow for each experiment and to record everything in exacting detail as it is done in a lab notebook. Each experimental run then has a complete paper record that can be referred to later on. Video recording is fine but I don't think it is suitable for recording an experiment. There are too many measurements, changes to the apparatus, etc. that are not available in a video. It appears the EW tests did follow this method but I don't believe they provided data on all their tests. Another aspect of testing is repeatability. If someone does a test they should try to set up the experiment again under the same conditions and run it again. These would be recorded as repeats of one experiment where the protocol has not changed. It may be necessary to allow time for everything to cool down to room temperature and to fiddle with the apparatus so it is setup the same each time. After several runs have been done - maybe 10 - 20 then run some statistical data analysis. What is the standard deviation of the "thrust" that was measured? What is the mean? Plot the results on a bar graph so the whole set of experiments can be seen. Now change some part of the experiment and repeat.Zen-in, I agree with most of what you say above. In a perfect world it would be good to have every last detail logged and recorded. I expect that within reason the DIYs are keeping such logs. I don't know anything about what is involved in editing Wiki pages and I am not sure that posting every last detail for every experimental run or rerun, on the Wiki page, would be a good idea. For a large part of the Wiki audience it would be just noise. It would be good to have that information available for those with credible experience and the desire to review it.Where I disagree is on the value of the video record. These are DIY adventures that rely in part (large or small) on the donation of $$, by mostly interested lay persons. Both the Wiki page and especially the videos go a long way toward encouraging that kind of support.I cannot say how well any of the DIYs have been at keeping an unpublished log of their efforts. Still I have been impressed by their efforts and achievements.Please note that some of the experimental work being discussed has been conducted at Universities and it is not DIY in a home or garage work. There is the work that Zellerium completed on the uniform cross section unsymmetrically placed dielectric cavity, and the yet still-to-be-heard of results of research work at a Canadian University using MegaWatts of power, as part of a student's project for example.
Quote from: Rodal on 12/05/2015 07:47 pmQuote from: OnlyMe on 12/05/2015 07:29 pmQuote from: zen-in on 12/05/2015 06:55 pmI looked at the wiki page a few times and I don't believe the data accurately shows the outcome of diy experiments. Part of the problem is people like to "data reduce" and cherry pick data. A more scientific method is to establish a protocol to follow for each experiment and to record everything in exacting detail as it is done in a lab notebook. Each experimental run then has a complete paper record that can be referred to later on. Video recording is fine but I don't think it is suitable for recording an experiment. There are too many measurements, changes to the apparatus, etc. that are not available in a video. It appears the EW tests did follow this method but I don't believe they provided data on all their tests. Another aspect of testing is repeatability. If someone does a test they should try to set up the experiment again under the same conditions and run it again. These would be recorded as repeats of one experiment where the protocol has not changed. It may be necessary to allow time for everything to cool down to room temperature and to fiddle with the apparatus so it is setup the same each time. After several runs have been done - maybe 10 - 20 then run some statistical data analysis. What is the standard deviation of the "thrust" that was measured? What is the mean? Plot the results on a bar graph so the whole set of experiments can be seen. Now change some part of the experiment and repeat.Zen-in, I agree with most of what you say above. In a perfect world it would be good to have every last detail logged and recorded. I expect that within reason the DIYs are keeping such logs. I don't know anything about what is involved in editing Wiki pages and I am not sure that posting every last detail for every experimental run or rerun, on the Wiki page, would be a good idea. For a large part of the Wiki audience it would be just noise. It would be good to have that information available for those with credible experience and the desire to review it.Where I disagree is on the value of the video record. These are DIY adventures that rely in part (large or small) on the donation of $$, by mostly interested lay persons. Both the Wiki page and especially the videos go a long way toward encouraging that kind of support.I cannot say how well any of the DIYs have been at keeping an unpublished log of their efforts. Still I have been impressed by their efforts and achievements.Please note that some of the experimental work being discussed has been conducted at Universities and it is not DIY in a home or garage work. There is the work that Zellerium completed on the uniform cross section unsymmetrically placed dielectric cavity, and the yet still-to-be-heard of results of research work at a Canadian University using MegaWatts of power, as part of a student's project for example.Yes, and Paul March of NASA Eagleworks has posted some updates in the past. But none of those individuals/groups seem to be posting regularly here and wouldn't they be responsible for any of their own detailed information to be posted on Wiki?As I said I agree in principle. It just seems that it could be a burden in the case of most of those actively posting here.And my main point was that the photos and videos do play an important fund raising role for the DIYs who have been posting them.
...I've always been reluctant to update the wiki page on anyone other than my own experiment. Whats your thoughts Doc. I didn't start the wiki...
On repeatability, I've been thinking that the current balance beam setups might be a problem. Each one introduces a bunch of unknowns, and no tow beams are going to work alike. Simply sitting a frustum on a modern digital scale might cause RFI.So I got to thinking about the old tri-beam balances I used in high school. Why not measure on a completely analog scale? Dial in the weight of the frustum, let it stablize and if the thing moved that's interesting.While led me to find these things:And I got to thinking, put a frustum on one end and dial in the weight and you could:1. Place a level showing distance in mm on the other end, shoot a laser pointer at it, and measure the amount of displacement on the level with a digital camera.2. Put a laser displacement sensor under the unused side to measure very small movements.The upshot of this is that you can model the effect of lift and downward thrust by placing weights on either side of the balance.People here don't like the idea because of concerns about the magnetic dampening used. I still think this setup might be a useful tool in allowing tests to be done on a bench, and not requiring a large beam setup. Better yet it uses a measuring tool that can be easily acquired.Which brings me to my gold standard EMDrive test:1. Get a tripple beam balance rated to around 2600g with 0.1g resolution. I could be wrong, but I can't see any mention of any form of dampening on these things.2. Get a frustum believed to have at least 0.1 grams of thrust in a vacuum. Preferably an integrated test article.3. Point frustum so that the thrust is directed downward with the thing on the tripple balance beam scale in a vacuum chamber (testing in a vacuum).4. If the thing moves downward (especially on the integrated setup) then rebuild the setup with a balance made of non-ferrous polymers. If you're still getting 0.1 grams of force, then I'd be hard pressed to say that it's anything but thrust. No air, no electronics to get fouled, probably nothing magnetic to cause the effect.5. Flip and check to make sure the effect still exists inverted.