-
#2360 by Rodal on 20 Oct, 2014 21:47
-
Further factual, numerical evidence against the importance of the dielectric in NASA's Brady et.al. tests.
The importance of the dielectric is based on this paragraph in the report:
<<We performed some very early evaluations without the dielectric resonator (TE012 mode at 2168 MHz, with power levels up to ~30 watts) and measured no significant net thrust.>>
I have noted before:
-very early evaluations
-no Q was reported during these "very early evaluations" without the dielectric
-at 2.168 GHz instead of 1.880 GHz (2.168 GHZ is way off to the right off scale in both the COMSOL Finite Element calculations and in the S21 plot)
What I had not pointed out is that Brady et.al. describe this evaluations at 2.168 GHz to be exactly the same mode shape (TE012) as the mode shape excited at 1.880 GHz which did result in substantial thrust.
The problem is evident: if both the 1.880 GHz and the 2.168 GHz are the same mode shape TE012, then one frequency (2.168 GHz) must be considerably further away from the peak amplitude . It is evident from the S21 plot that the peak amplitude occurs near 1.880 GHz. Therefore the 2.168 GHz "very early evaluation" with and without the dielectric were conducted at a frequency way away from the narrow bandwidth corresponding to the Q=22,000 measured at resonance at 1.880 GHz.
Therefore, the tests conducted at 2.168 GHz are evidently way off the resonance peak frequency for mode TE012 and hence completely unrepresentative. No wonder that they did not report a Q at 2.168 GHz, since that frequency was too far away from resonance.
I attach the COMSOL Finite Element predicted amplitude vs frequency (on top) and the S21 actual measurement (on the bottom) of amplitude vs. frequency. Did they test initially at 2.168 GHz based on the COMSOL Finite Element analysis predicting peaks to occur at higher frequencies ?. It is obvious from comparison of the COMSOL FE results to the actual S21 plot that Brady's COMSOL FE result is inaccurate for these purposes, as it predicts peak amplitudes at a different frequency than the measured frequency. In my assessment this is due to a modeling problem:was a fine enough mesh used? (apparently not), was convergence investigated with an increasing number of Finite Elements to show convergence of the results?. Besides the finite element mesh not being fine enough to demonstrate convergence, was the finite element type used able to achieve convergence ? (what interpolation function did the finite element use? what finite element shape?).
From the S21 measurements they must have realized that the actual peak for mode TE012 was at a considerably lower frequency than the >2 GHz predicted by the COMSOL Finite Element analysis. This may explain why Brady et.al. then measured at 1.880 GHz instead of 2.168GHz.
-
#2361 by aero on 20 Oct, 2014 21:55
-
Doesn't that show the use of less than the ultimate care is construction of the cavity?
-
#2362 by Rodal on 20 Oct, 2014 22:05
-
Doesn't that show the use of less than the ultimate care is construction of the cavity?
No it is not related to care in physical construction of the cavity. It just shows that they did not conduct a computational convergence study of their Finite Element analysis investigating finer and finer meshes as well as other types of Finite Elements (with higher order interpolation polynomials) until achieving convergence. This could be due to lack of time and urgency for the project to complete in a specified amount of time and not having the time to conduct a convergence study of their Finite Element results. It could also be due to project priorities: perhaps they prioritized the experimental measurements over the computational analysis. To Brady's credit they conducted a Finite Element analysis and they properly identified the mode shapes involved. They deserve kudos for that. I have not seen the mode shapes identified for the experiments conducted in other countries. I very much appreciate that Brady et.al. conducted the Finite Element analysis and reported the mode shapes.
-
#2363 by aero on 20 Oct, 2014 22:11
-
In other words they built the cavity first then tried to figure out how to excite it.
-
#2364 by Rodal on 20 Oct, 2014 22:18
-
In other words they built the cavity first then tried to figure out how to excite it.
Their Finite Element analysis shows amplitude peaks occurring at very different frequencies and with different amplitudes than occurs in actuality. So there is no doubt that their Finite Element model is not good enough to represent the actual cavity at the level of frequency resolution needed to conduct the experimental tuning. It would not be unusual in a fast project in industry to not be able to conduct a finite element mesh study (and finite element interpolation polynomial study) to convergence.
It would not make sense to me that they would build a different model that they modeled or that they would model a different geometry than they built (although it is possible). So the question is on the finite element mesh and the finite element interpolation type, but the question is not on the geometry. Due to the FE model being not converged at the level of frequency resolution needed to conduct the experimental tuning, they had to spend some time investigating at what frequency resonance takes place for different modes. I do appreciate that Brady et.al. had the initiative to conduct the Finite Element analysis and reported the mode shapes. I don't recall the studies from other countries showing what mode shapes are being excited. -
#2365 by RotoSequence on 20 Oct, 2014 23:13
-
What good does it do to try and sell an electric SSTO space plane and interstellar probes, when they haven't even put together a tabletop demonstration whose effect is visible without high sensitivity instrumentation?
-
#2366 by IslandPlaya on 20 Oct, 2014 23:47
-
What good does it do to try and sell an electric SSTO space plane and interstellar probes, when they haven't even put together a tabletop demonstration whose effect is visible without high sensitivity instrumentation?
The *good* it does is to point out how game changing a positive result would be.
Hence a bit more funding. Hence we can get a bit closer to the *truth*
-
#2367 by frobnicat on 20 Oct, 2014 23:49
-
What good does it do to try and sell an electric SSTO space plane and interstellar probes, when they haven't even put together a tabletop demonstration whose effect is visible without high sensitivity instrumentation?
This overselling applications before much more convincing and detailed and open reproducible experimental design was already complained about. That said, starting to claim hundreds of milliNewtons this should no longer be in the realm of high sensitivity. A good air cushion table should be able to show the thing zipping around (if only it were self contained, including power), and hell, a turnstile also : time to see how the thing goes when approaching the velocity where it gives more energy than it takes. Shawyers semi-classical attempts at mitigating this "problem" with "velocity corrections" are less than convincing to me. Anyone understands the slide 3 of latest presentation and how it's supposed to deal with energy conservation, when the claimed consequences slide 9 & 10 are still out of sensible energy budget (and yet supposedly with efficiency <1 slide 11) ?
good night.
-
#2368 by IslandPlaya on 20 Oct, 2014 23:54
-
Maybe the EM frustum setup is a good test for Axions at least...
-
#2369 by aceshigh on 21 Oct, 2014 00:34
-
Terminal velocity = 204,429 km/s = 0.68 c
Even if you can get to 68% of the speed of light, how do you slow down at that point?
reminds me of the book Tau Zero -
#2370 by JohnFornaro on 21 Oct, 2014 00:35
-
This is something where JohnFornaro and AutoCAD could really help ...
I just plopped
"AnomalousThrustProductionFromanRFTestDevice-BradyEtAl.jpg"
into ACAD, and arbitrarily scaled it. Nobody's given me a wide end dimension, so, for the purposes of discussion, I just scaled it to your *cough* typical EM-Drive diameter *cough* of 28 cm.
Interesting similarities in the proportions.
Got any more where that came from?
-
#2371 by JohnFornaro on 21 Oct, 2014 00:36
-
What good does it do to try and sell an electric SSTO space plane and interstellar probes, when they haven't even put together a tabletop demonstration whose effect is visible without high sensitivity instrumentation?
The *good* it does is to point out how game changing a positive result would be.
Hence a bit more funding. Hence we can get a bit closer to the *truth*
Unfortunately, however, they are selling sizzle before they have a steak. -
#2372 by Rodal on 21 Oct, 2014 00:39
-
This is something where JohnFornaro and AutoCAD could really help ...
I just plopped
"AnomalousThrustProductionFromanRFTestDevice-BradyEtAl.jpg"
into ACAD, and arbitrarily scaled it. Nobody's given me a wide end dimension, so, for the purposes of discussion, I just scaled it to your *cough* typical EM-Drive diameter *cough* of 28 cm.
Interesting similarities in the proportions.
Got any more where that came from?
Thank you!
Very professional job. We even got pdf's for everybody to look at!
And everything in metric units as well !
Now, would it be too much to ask to also get cheese with the excellent wine?
We know that the (Faztek) beam (at the bottom of the picture, shown in cross-section) has a square cross section of 1.5 inch by 1.5 inch
Knowing that, what are the dimensions? -
#2373 by IslandPlaya on 21 Oct, 2014 00:39
-
The smell of the sizzle is often good enough to want a burger though...What good does it do to try and sell an electric SSTO space plane and interstellar probes, when they haven't even put together a tabletop demonstration whose effect is visible without high sensitivity instrumentation?
The *good* it does is to point out how game changing a positive result would be.
Hence a bit more funding. Hence we can get a bit closer to the *truth*
Unfortunately, however, they are selling sizzle before they have a steak.
-
#2374 by JohnFornaro on 21 Oct, 2014 01:29
-
I gladly pay you Thursday for a cheeseburger today.
-
#2375 by JohnFornaro on 21 Oct, 2014 01:30
-
Here's the "small" shawyer device. Correct my titles.
-
#2376 by aero on 21 Oct, 2014 01:30
-
This is something where JohnFornaro and AutoCAD could really help ...
I just plopped
"AnomalousThrustProductionFromanRFTestDevice-BradyEtAl.jpg"
into ACAD, and arbitrarily scaled it. Nobody's given me a wide end dimension, so, for the purposes of discussion, I just scaled it to your *cough* typical EM-Drive diameter *cough* of 28 cm.
Interesting similarities in the proportions.
Got any more where that came from?
Thank you!
Very professional job. We even got pdf's for everybody to look at!
And everything in metric units as well !
Now, would it be too much to ask to also get cheese with the excellent wine?
We know that the (Faztek) beam (at the bottom of the picture, shown in cross-section) has a square cross section of 1.5 inch by 1.5 inch
Knowing that, what are the dimensions?
We also know that the diameter of the chamber is 0.762 meters so the radius is 0.381 meters. Maybe that will help. -
#2377 by aero on 21 Oct, 2014 01:35
-
Here's the "small" shawyer device. Correct my titles.
But we know from the documentation that the "small" Shawyer device big end was 16 cm diameter.
Other than that, I really like the way your software presents results. -
#2378 by JohnFornaro on 21 Oct, 2014 01:51
-
Bitch and moan. All I ever hear around here.
-
#2379 by Rodal on 21 Oct, 2014 01:55
-
Bitch and moan. All I ever hear around here.
Outstanding job!