-
#2020
by
Kenjee
on 08 Oct, 2017 14:44
-
Here are all the sims on bell curve.
-
#2021
by
giulioprisco
on 08 Oct, 2017 14:53
-
-
#2022
by
Mulletron
on 08 Oct, 2017 15:33
-
From the article:
If a pilot wave does explain the thrust behind the device, then it could also lead to a way to make the propulsion system even more powerful in future, and it's as simple as tweaking the shape.
"We have seen that the effect could be enhanced using a different shape for the frustum," said Castro. "In fact a trumpet exponential form is expected to increase the thrust."
What the heck is a trumpet exponential form? What does it look like?
Would it be where the frustum walls curve inward, like near the mouth of a trumpet?
Technically, the mouth of a trumpet is called the "bell" - an ironically appropriate name - and one with a better ring to it than "frustum".
Heh - so even a propellantless rocket still needs a bell, huh? Especially when it uses non-local hidden variables to satisfy Bell's theorem? 
Hey Monomorphic, TheTraveler, or whoever - is it possible to run this shape through your analytical software and produce a graphical render from it?
From http://www.krynaglobal.com/product/eprop.html :
That trumpet shape looks like an Euler spiral in 3D. Parts of the curve are used to control jerk. It would be interesting to see how this relates from another perspective about how these shapes effect group velocity, acceleration, and jerk. I've been trying figure out how these two shapes (current EMdrive vs this shape) effect the jerking motion of a partial standing wave. I was thinking that an infinite jerk might be better to have than a linear change. I'm still undecided.
http://dynref.engr.illinois.edu/avt.htmlhttps://en.m.wikipedia.org/wiki/Euler_spiralhttp://iopscience.iop.org/article/10.1088/0143-0807/37/6/065008
-
#2023
by
sghill
on 08 Oct, 2017 16:00
-
Here are all the sims on bell curve.
Can you run it with an inverted curved top and bottom bottom instead of a flat bottom? Not a neutral curved bottom, a parabola or catenary curve...
ThatOtherGuy had asked for this back in March, but the discussion moved on when Paul March commented on another topic of discussion.
https://forum.nasaspaceflight.com/index.php?topic=41732.msg1655717#msg1655717"I wonder if and how the fields would change replacing the two (top/bottom) flat panels with curved (parabolic or catenary ?) ones, also, it may be interesting to have the bottom side curved up (out to in) and the top side (smaller one) curved up too (in to out)"
-
#2024
by
Kenjee
on 08 Oct, 2017 16:16
-
If you could draw a sketch I`ll try to sim. I don`t understand what`s inverted from what.
-
#2025
by
ThatOtherGuy
on 08 Oct, 2017 16:29
-
If you could draw a sketch I`ll try to sim. I don`t understand what`s inverted from what.
Well, imagine the frustum sitting on the larger end; the larger end will be concave (curve going "inside" the frustum) while the small end is convex (curve going "outside" the frustum); clear enough now
?
[edit]
in such a config, the antenna used to inject the signal should possibly be placed near/at the focus of the small plate
-
#2026
by
OnlyMe
on 08 Oct, 2017 17:49
-
There has been something confusing me about the whole convex/concave end plate issue... and even to some extent the waveguide/antenna placement.
Early on it seems that guidance attributed to Shawyer:
Recommended the large end being concave and the small end convex, perhaps intending to minimized bouncing photons off the walls and keep them bouncing back and forth between the end plates???
Later it seems, there was a design that incorporated a super conducting large end plate that remained flat (to difficult to construct a convex superconducting plate?), while the small end turned around into a convex shape with an antenna centered, in the small end. Maybe I missed an explanation for the change in shape of the small end or other than theoretical reasoning for moving microwave insertion from a sidewall to the center of the small end.
Then again attributed to Shawyer, a (superconducting) wedge shaped design reverting to large end convex and small end concave.., and returning to a waveguide in a side wall.
Most of this jumping around on unverified design recommendations, sometimes seeming to be contradictory, tends toward questionable credibility.
So far where has anyone seen or even claimed credible results involving any curved endplates? Monomorphic is working on that now. Has there been any other real example of a build with curved end plates? It does seem reasonable to explore the curved endplates, given the resources and assets Monomorphic has and has been willing to commit to the project. Still...
Right now, unless a builder is working with unpublished credible experimental results, should not the goal be to just produce a basic design that returns some real, even if small level of thrust? And since most if not all of the designs that initiated this search involved flat end plates and either a magnetron mounted directly on the large end plate or (it seems preferable) sidewall microwave insertion, by waveguide or antenna, shouldn't the basic design remain within those general limits until some credible experimental data suggests otherwise.
There is no credible theory of operation at present and cannot be one until there is at least a basic functional device to work from. And yet it seem that much of the design is being driven by one or another unconfirmed theory...
-
#2027
by
Peter Lauwer
on 08 Oct, 2017 17:54
-
First S11 measurements on the new frustum cavity.I performed some first measurements on a frustum-shaped cavity I recently made (picture), with the Windfreak Technology SynthNV signal source/scalar network analyzer + a directional coupler.
As coupling loop I have made one according to Zhang et al., 2013 [1], see the attached picture (it is not easy, soldering a small wire to a 1-mm thick copper sheet. I needed my 300 W soldering iron to do that and it is rather clumsy). The remaining plastic cover has been removed from the copper plate, of course.
The results are not bad, I would say. Please see the attached example, two adjacent peaks at ~3036 MHz and ~3073 MHz.
The frustum (inner) dimensions are: BD = 202.0(5) mm, SD = 122.0(5) mm, h = 165(1) mm. Copper plate thickness: Big endplate and side wall = 0.5 mm, small endplate = 1 mm.
Coupling loop parameters: 1-mm wire, r = 11.0(5) mm, w = 14.0(5) mm, theta = ~45° (see the artcle by Zhang et al.).
More later.
Peter
[1] H. Zhang et al.,
Research on Novel Loop Antenna in Microwave Cavity Measurement of Permittivity, Int. J. of Information and Electronics Eng., Vol. 3, No. 4, July 2013, pp. 396-398.
-
#2028
by
ThatOtherGuy
on 08 Oct, 2017 17:56
-
You are right, by the way, at this stage, asking (say) monomorphic to change his design would be detrimental to the experimenting progress, yet, using such a design to feed a simulation program may be interesting imVVHo
-
#2029
by
Kenjee
on 08 Oct, 2017 18:40
-
Here are all the sims on bell curve.
Can you run it with an inverted curved top and bottom bottom instead of a flat bottom? Not a neutral curved bottom, a parabola or catenary curve...
ThatOtherGuy had asked for this back in March, but the discussion moved on when Paul March commented on another topic of discussion.
https://forum.nasaspaceflight.com/index.php?topic=41732.msg1655717#msg1655717
"I wonder if and how the fields would change replacing the two (top/bottom) flat panels with curved (parabolic or catenary ?) ones, also, it may be interesting to have the bottom side curved up (out to in) and the top side (smaller one) curved up too (in to out)"
Mono`s new inverted.
-
#2030
by
ThatOtherGuy
on 08 Oct, 2017 18:46
-
Thank you
!
Would you mind trying different curves? e.g. top different from bottom or both the same but deeper... and the like ?
-
#2031
by
Peter Lauwer
on 08 Oct, 2017 19:07
-
Frustum resonance measurementsIn my previous posting [https://forum.nasaspaceflight.com/index.php?topic=42978.msg1733603#msg1733603] I described the frustum and coupling loop.
During the measurements I discovered that using 0.5 mm copper sheet for the big diameter endplate was not such a good idea (though in doing so, I managed to keep the total weight below 1 kg). I first measured with the cavity in vertical position and discovered that it is very sensitive to putting some pressure on it, the endplate deforms easily and resonance peaks shift by the order of 1 MHz.
I then put the frustum in horizontal position (picture) and remeasured some resonances. Some of them I tabulate below, with estimated Q-values (determined from the -3 dB width):
3036.5 MHz Q = ~14k
3037.7 MHz Q = ~15k
3679.7 MHz Q = ~16k
The TE013 mode should be at ~3600 MHz for these dimensions.
I found that one very pronounced peak around 3490 MHz was not visible now (see picture Frustum1_3490MHz peak_no pressure.jpg). But, by putting some pressure on the big endplate, it appears again (see Frustum1_3490MHz peak_WITH pressure.jpg).
This mode seems to be very sensitive to the right dimensions:
3489.15 MHz Q = ~10k
Who can identify these modes?I think about soldering a nut in the center of the big endplate and with a bridge etc. I can then use it to tune. Maybe pulling is better, as the big endplate gets a bit spherical then.
In the coming weeks, I hope to check these results with a professional network analyzer (VNA).
-
#2032
by
ThatOtherGuy
on 08 Oct, 2017 19:28
-
You may just add a bar (or two crossed bars) to the larger end, using them to support a screw with a spherical (or ball) tip which may then be used to apply a pressure to the (center of the) larger plate and deformate it, that should be easier to build/apply to the frustum
-
#2033
by
Peter Lauwer
on 08 Oct, 2017 19:32
-
Frustum resonance measurements
...some resonances. Some of them I tabulate below, with estimated Q-values (determined from the -3 dB width):
3036.5 MHz Q = ~14k
3037.7 MHz Q = ~15k
3679.7 MHz Q = ~16k
So, coming back to some earlier discussions, it seems that with copper walls that are not specially polished, one can still achieve good Q-values. The Q's I've measured now are in the range 10k-20k. Not the best one can reach, but I think the influence of the seam in the side wall and imperfectness of the coupling loop are more important. The inner walls of my cavities are from untouched (I handle them with gloves), commercially available sheets. Shining brightly, but not extra polished.
-
#2034
by
X_RaY
on 08 Oct, 2017 19:46
-
Frustum resonance measurements
In my previous posting [https://forum.nasaspaceflight.com/index.php?topic=42978.msg1733603#msg1733603] I described the frustum and coupling loop.
During the measurements I discovered that using 0.5 mm copper sheet for the big diameter endplate was not such a good idea (though in doing so, I managed to keep the total weight below 1 kg). I first measured with the cavity in vertical position and discovered that it is very sensitive to putting some pressure on it, the endplate deforms easily and resonance peaks shift by the order of 1 MHz.
I then put the frustum in horizontal position (picture) and remeasured some resonances. Some of them I tabulate below, with estimated Q-values (determined from the -3 dB width):
3036.5 MHz Q = ~14k
3037.7 MHz Q = ~15k
3679.7 MHz Q = ~16k
The TE013 mode should be at ~3600 MHz for these dimensions.
I found that one very pronounced peak around 3490 MHz was not visible now (see picture Frustum1_3490MHz peak_no pressure.jpg). But, by putting some pressure on the big endplate, it appears again (see Frustum1_3490MHz peak_WITH pressure.jpg). This mode seems to be very sensitive to the right dimensions:
3489.15 MHz Q = ~10k Who can identify these modes?
I think about soldering a nut in the center of the big endplate and with a bridge etc. I can then use it to tune. Maybe pulling is better, as the big endplate gets a bit spherical then.
In the coming weeks, I hope to check these results with a professional network analyzer (VNA).
Hi Peter,
here is a fast 2D spreadsheet result. Altrough it is less precise than a simulation this may help to identify some of the modes.
I am busy making sims for someone else, therefore my sim PC is not available yet.
INPUT:
SD=122mm; BD=202mm; LN=165mm (Theta=13,6270°)
DATA:
Mode f in GHz
TE010 2,3071296094
TE011 2,4827188408
TE012 2,9442770717
TE013 3,5802976604
TE110 1,1086047692
TE111 1,437059387
TE112 2,1330277241
TE113 2,9461103615
TE210 1,8390081526
TE211 2,0547141204
TE212 2,5922559871
TE213 3,2955359988
TE310 2,5296056173
TE311 2,6907846997
TE312 3,1221661595
TE410 3,2017843398
TE411 3,3307029934
TE120 3,2101477373
TE121 3,3387444774
TM010 1,4479816863
TM011 1,713173527
TM012 2,3293690825
TM013 3,0919367053
TM110 2,3071356306
TM111 2,4827244383
TM112 2,5590355956
TM113 3,192965652
TM210 3,0922412936
TM211 3,2255268039
TM212 3,3466429397
TM020 2,0073982454
TM021 1,5188121105
TM022 2,7148953009
TE122 2,6643401366
TM122 3,1397732738
TM014 3,9167132186
TM123 3,7433479825
TE123 3,1695131874
Your frustum looks quite good
Best Regards
ADD:At least one can exclude some modes far away in frequency.
... But, by putting some pressure on the big endplate, it appears again ...
The phenomenon you describe by pressing on the plate sounds like contact issues which implies a TM mode.
-
#2035
by
Peter Lauwer
on 08 Oct, 2017 20:04
-
Hi Peter,
here is a fast 2D spreadsheet result. Altrough it is less precise than a simulation this may help to identify some of the modes.
At least one can exclude some modes far away in frequency. The phenomenon you describe by pressing on the plate sounds like contact issues which implies an TM mode.
I am busy making sims for someone else, therefore my sim PC is not available yet.
INPUT:
SD=122mm; BD=202mm; LN=165mm (Theta=13,6270°)
DATA:
Mode f in GHz
TE010 2,3071296094
TE011 2,4827188408
TE012 2,9442770717
TE013 3,5802976604
TE110 1,1086047692
...
Thanks! I will compare them to my measured values.
Peter
-
#2036
by
R.W. Keyes
on 08 Oct, 2017 20:16
-
I've seen the use of dielectrics in various emdrive models. However, any dielectric material is going to have loss, which increases greatly with frequency. This loss would lower Q, undermining all of the other efforts which have been made to increase Q. Yes, the use of a dielectric will allow more freedom in the physical structure of the frustum, but at what cost? Is high Q necessary for high performance in an emdrive, or not? Perhaps I am looking at design from a different viewpoint than those who use dielectrics: some want to build a simple structure which proves an emdrive possible, and such experimenters have equipment to accurately measure very small amounts of force, whereas my priorities would be to gain as much emdrive effect as possible through precise and somewhat difficult construction, and be able to utilize far less sensitive equipment in the detection of force.
-
#2037
by
Peter Lauwer
on 08 Oct, 2017 20:36
-
... But, by putting some pressure on the big endplate, it appears again ...
The phenomenon you describe by pressing on the plate sounds like contact issues which implies a TM mode.
The plate seems to be soldered well. And that resonance has a good Q (10k).
-
#2038
by
X_RaY
on 08 Oct, 2017 20:49
-
... But, by putting some pressure on the big endplate, it appears again ...
The phenomenon you describe by pressing on the plate sounds like contact issues which implies a TM mode.
The plate seems to be soldered well. And that resonance has a good Q (10k).
OK this was just a guess. If it is soldered the big plate should not be the reason (maybe its the small side where the contact is less what increases in this position when preasure is applied, whatever) the other possibility is there is a mode very sensitive to the impedance between antenna and cavity (again a TM mode to first order). Since you take measurements with a scalar network analyzer rather than a VNA it is hard to tell what happens exactly. I general higher order modes (higher integer) lead to greater change in frequency when alter the frustum dimension such as shorter the z dimension (pressing on the plate).
I will do some sims when i have time and resources.
-
#2039
by
Kenjee
on 08 Oct, 2017 21:12
-
?