-
#3460 by keithpickering on 02 Jul, 2016 22:58
-
...
OK, point taken, Q can't drop all the way to zero. But it should drop very low, i.e., lower than a frustum. As a photon approaches the point, it's going to be doing a LOT of bouncing off the walls, each bounce being another possibility of dissipation. And that effect accelerates the closer you get to the point. So yeah, a lot of photons are going to get lost down there. Bad Q = bad thrust.
It doesn't work like that.
First we are dealing with a cavity on the order of the wavelength, so "bounces" isn't a particularly meaningful model at this scale.
In the first place, photons are points. It's the EM field that has wavelength. And in the second place, we're not necessarily dealing with any particular scale.Second, there would be no particularly large number of bounces happening even if you assumed a bounce model (plotted the path of a small laser beam). You can try to draw out some examples yourself. I would do it, but I feel that you would learn more by working it out yourself.
Q is defined as the inverse of dissipation that occurs in a given cycle of angular frequency, and the inverse of frequency is wavelength. So the more bounces that occur in a given length, the greater the possibility of dissipation. As the walls get closer, the number of bounces in that length gets greater and greater, so the dissipation per wavelength must increase. Which means Q must decrease. -
#3461 by keithpickering on 02 Jul, 2016 23:02
-
NO the wave will be reflected in the opposite direction.
If that is indeed the case then MiHsC does not predict infinite thrust. So your original point is moot. It's the same as putting a small end at the cutoff point.
-
#3462 by Monomorphic on 03 Jul, 2016 01:03
-
Build update: More work continues on the test stand. I purchased the same laser displacement sensor as Rfmwguy, thanks Dave for finding that sweet deal on ebay! I am in the process of integrating the sensor into the test stand.
I am also removing all ferromagnetic material from the vicinity of the emdrive. This includes replacing all zinc screws with brass. and moving the computer monitor to the far right.
I am also including an image of the test stand support with sorbothane pads.
-
#3463 by SeeShells on 03 Jul, 2016 01:08
-
Build update: More work continues on the test stand. I purchased the same laser displacement sensor Rfmwguy has, thanks Dave for finding that sweet deal on ebay! I am in the process of integrating the sensor into the test stand.
That's some beautiful work! I noticed your magnetron is now at the bottom, smart move.
I am also in the process of removing all ferromagnetic material from the vicinity of the emdrive. This includes replacing all zinc screws with brass. and moving the computer monitor to the far right.
I am also including an image of the test stand support with sorbothane pads.
Shell -
#3464 by SeeShells on 03 Jul, 2016 02:13
-
Build update: More work continues on the test stand. I purchased the same laser displacement sensor as Rfmwguy, thanks Dave for finding that sweet deal on ebay! I am in the process of integrating the sensor into the test stand.
When they built your garage did they use wood studs or metal studs? Metal you can find with a magnet.
I am also removing all ferromagnetic material from the vicinity of the emdrive. This includes replacing all zinc screws with brass. and moving the computer monitor to the far right.
I am also including an image of the test stand support with sorbothane pads.
Shell -
#3465 by Monomorphic on 03 Jul, 2016 02:44
-
When they built your garage did they use wood studs or metal studs? Metal you can find with a magnet.
My lab/workshop is underground in the basement. The wall near the test stand is cinder block. -
#3466 by RotoSequence on 03 Jul, 2016 02:53
-
When they built your garage did they use wood studs or metal studs? Metal you can find with a magnet.
My lab/workshop is underground in the basement. The wall near the test stand is cinder block.
A properly made cinder block wall will have steel rebar in the hollows. -
#3467 by Monomorphic on 03 Jul, 2016 03:01
-
A properly made cinder block wall will have steel rebar in the hollows.
It's actually a firewall between my town-home and the neighbors. I've not seen any rebar in the hollows. This is in the southeast USA - built in the early 1980's. Doubt there's any rebar in there at all!
But if there were, I guess I would need a metal detector to find out. -
#3468 by RotoSequence on 03 Jul, 2016 03:04
-
A properly made cinder block wall will have steel rebar in the hollows.
It's actually a firewall between my town-home and the neighbors. I've not seen any rebar in the hollows. This is in the southeast USA - built in the early 1980's. Doubt there's any rebar in there at all!
Well, probably wouldn't hurt to double-check. Convenient for the sake of experiment, but boo hiss to un-reinforced masonry!
-
#3469 by Superfastjellyfish on 03 Jul, 2016 05:02
-
Build update: More work continues on the test stand. I purchased the same laser displacement sensor as Rfmwguy, thanks Dave for finding that sweet deal on ebay! I am in the process of integrating the sensor into the test stand.
I am also removing all ferromagnetic material from the vicinity of the emdrive. This includes replacing all zinc screws with brass. and moving the computer monitor to the far right.
I am also including an image of the test stand support with sorbothane pads.
A way to tell if something is brass or bronze is to use a magnet. It's not very magnetic, but it is. Probably too small to matter, but there has to be a non magnetic metal that can be used. Just my armchair opinion. Experiment on! -
#3470 by X_RaY on 03 Jul, 2016 07:58
-
NO the wave will be reflected in the opposite direction.
If that is indeed the case then MiHsC does not predict infinite thrust. So your original point is moot. It's the same as putting a small end at the cutoff point.
Please take a look at McCulloch's equation
F = P Q L/c * (1/Ds - 1/Db)
This formula alone does not imply a minimum plate size.
If Ds is infinite small 1/Ds is infinite also.
All other terms in the equation are second order in this situation because they are small compared to the infinite term.
It follows when Ds=∞ --> F=∞
My orginal statement was:
It shows the limit for the equations not that this is possible in physical nature.
All known different thrust equations for the EM-Drive predict that beside a false infinite thrust generation for a pointed cone.
The only restriction was given by TT and Sawyer to their magic 0.82 cutoff rule for TE01p modes.
One must take the nature of light waves into account i.e. microwaves in our case.
On the other hand if you set Ds=0 it's not solvable. If you set 1/Ds=0 you get F = P Q L/c * (1/Db) and the force only depends on the size of the big diameter plate and the length of the cone.
This approximation formula only deals with the total amount of power inside the cavity and the inverse damping factor, not the microwave component for example the waveform(mode) and the related physical effects.
I would be careful to use the cut off diameter calculation rule of a open waveguide for a closed cavity resonator. Total reflection take place at a significantly smaller diameter than the calculated cut off diameter for an open waveguide as shown by Dr.Rodal for TE01 modes in his paper "Cut off of resonant modes in truncated conical cavities". -
#3471 by X_RaY on 03 Jul, 2016 10:01
-
X_Ray, The attached geometry #1 should resonate in TE012 at +/-2.45 GHz with extruded HDPE dielectric resonator/disks (same as NASA frustum). The question I have is: will TE012 be excited in this geometry by steady clean maggie flow image #2 (attached) as demonstrated by Shell. And in what location would/should the maggie RF be "fed into" ...top, bottom, side of the frustum? As you can see in image #3 the antenna is located on the side, and I calculate it to be located 1/4 of the cavity height from the bottom of the cavity. Agreed? Frustum size can be made slightly larger to accommodate 2.43189 GHz if needed. Pickering comments below suggested I calculate the Q of this "smaller" RC...Am coming up with something +/- 14,000-15,000 + dimensions not showing as well as I'd like in the schematic, therefore Small diameter= 12.1841 cm, Height= 17.545 cm and Bottom diameter= 21.444 cm Frequency +/- 2.45 GHz
Congrats, with HDPE it is TE012
HDPE dielectric disks (top of frustum) dimensions = 12.027 x 4.145 cm
Using this dimensions with HDPE I get a Q of ~19000.
The resonant mode is gouverned by the overall dimensions of the resonator, the antenna will be more or less able to excite the field pattern. About the "best" antenna shape, location and direction there was a lot of discussion in this threads. In fact there are several possibilities for each mode to excite it. Hard to say whats the best way. In general, you get good coupling if the field vectors of the antenna at its location are equal/close to the vectors of the resonant mode in the cavity.
-
#3472 by Monomorphic on 03 Jul, 2016 13:20
-
Using this dimensions with HDPE I get a Q of ~19000.
Try switching to copper instead of the perfect electric conductor. See how much that lowers Q. -
#3473 by X_RaY on 03 Jul, 2016 13:29
-
I didn't use feko to compute the Q.Using this dimensions with HDPE I get a Q of ~19000.
Try switching to copper instead of the perfect electric conductor. See how much that lowers Q.
-
#3474 by TheTraveller on 03 Jul, 2016 14:43
-
All known different thrust equations for the EM-Drive predict that beside a false infinite thrust generation for a pointed cone.
Roger developed a Df or Design factor that deals with the way the big and small ends behave. It is part of his thrust equation F = (2 Qu Pwr Df) / c. As the small end diameter decreases, the guide wavelength increases and the Df increases. As the big end diameter increases, the guide wavelength decreases and the Df increases. Max Df = 1. When small end is at or below cutoff diameter, the Df = 0. Design goal is to make the small end diameter as small as possible while operating just above cutoff diameter and for the big end diameter to be as big as possible.
MiHsC gives the same prediction: to maximize thrust, make the big end bigger and the small end smaller.
The only restriction was given by TT and Sawyer to their magic 0.82 cutoff rule for TE01p modes.
With Roger's equation F = (2 Qu Pwr Df) / c, when the small end is below cutoff diameter, the Df = 0 so make the small end too small and there is no thrust. -
#3475 by meberbs on 03 Jul, 2016 14:54
-
First we are dealing with a cavity on the order of the wavelength, so "bounces" isn't a particularly meaningful model at this scale.
In the first place, photons are points. It's the EM field that has wavelength. And in the second place, we're not necessarily dealing with any particular scale.
No, photons are not points, they are EM field waves. None of the discrete nature of photons is in any way relevant to an EMDrive, because EMDrives are a particular scale. The modes people are talking about mean that the cavity is on the order of the wavelength, if you scale the cavity, you have to scale the frequency so this remains true.
The number of bounces don't get greater, the angle of the walls means that a ray trace can reverse direction back towards the big end without hitting the small end. Why don't you try drawing it like I suggested?Second, there would be no particularly large number of bounces happening even if you assumed a bounce model (plotted the path of a small laser beam). You can try to draw out some examples yourself. I would do it, but I feel that you would learn more by working it out yourself.
Q is defined as the inverse of dissipation that occurs in a given cycle of angular frequency, and the inverse of frequency is wavelength. So the more bounces that occur in a given length, the greater the possibility of dissipation. As the walls get closer, the number of bounces in that length gets greater and greater, so the dissipation per wavelength must increase. Which means Q must decrease.
Also, you didn't pay attention to what you were doing when you reworked Q to have it as "dissipation per wavelength", this doesn't have the meaning you seem to think it does. It seems like you think that means dissipation per z-axis distance traveled, but there is nothing special about the z-axis, and you would need to measure slant path for this to be relevant, which is a roundabout way of measuring time. As I stated above you should draw out some of these paths, and see that there will not end up being a particularly large number of bounces. Also, if you do this, don't take a "nearly parallel" limit. That is changing a different variable than just extending one of the typical EMdrive shapes to be a cone rather than a frustum. -
#3476 by zen-in on 03 Jul, 2016 16:06
-
Build update: More work continues on the test stand. I purchased the same laser displacement sensor as Rfmwguy, thanks Dave for finding that sweet deal on ebay! I am in the process of integrating the sensor into the test stand.
I am also removing all ferromagnetic material from the vicinity of the emdrive. This includes replacing all zinc screws with brass. and moving the computer monitor to the far right.
I am also including an image of the test stand support with sorbothane pads.
Excellent build! I admire your workmanship. Brass is slightly diamagnetc while type 303 stainless steel has very low ferromagnetic properties. You should not see any magnetic attraction between 303 hardware and a strong magnet. -
#3477 by Bob Woods on 03 Jul, 2016 16:28
-
I am also including an image of the test stand support with sorbothane pads.
I agree with Zen, your work is very impressive. Keep it up.
-
#3478 by FattyLumpkin on 03 Jul, 2016 17:38
-
X_Ray, thanks for those sims! Can you tell me where you located the antenna, type of antenna and are you able to sim RF from a maggie with your version of FEKO? If so, would you attempt a sim with maggie? So many builders are using magnetrons, I think it would be a worth while experiment to see how well or "not" magnetrons work given a frustum that is known to work with an antenna + dielectrics. As mentioned in my previous post the NASA folks had difficulty keeping the frustum in TE012 and could not "get" much power (Watts) into it. It is assumed that this is because TE012 is close to other modes in this configuration, and they were experiencing "mode jumping" perhaps also due to antenna position. I don't know if you watched the presentation, but Sonny indicated that they would employ "phase lock loop(ing)" in their next test campaign to control for this problem. Unfortunately, it is the results et al. from that campaign that are now in peer review. Thanks once again for your hard work! , K
-
#3479 by X_RaY on 03 Jul, 2016 20:08
-
X_Ray, thanks for those sims! Can you tell me where you located the antenna, type of antenna and are you able to sim RF from a maggie with your version of FEKO? If so, would you attempt a sim with maggie? So many builders are using magnetrons, I think it would be a worth while experiment to see how well or "not" magnetrons work given a frustum that is known to work with an antenna + dielectrics. As mentioned in my previous post the NASA folks had difficulty keeping the frustum in TE012 and could not "get" much power (Watts) into it. It is assumed that this is because TE012 is close to other modes in this configuration, and they were experiencing "mode jumping" perhaps also due to antenna position. I don't know if you watched the presentation, but Sonny indicated that they would employ "phase lock loop(ing)" in their next test campaign to control for this problem. Unfortunately, it is the results et al. from that campaign that are now in peer review. Thanks once again for your hard work! , K
1) First of all its hobby not hard work.
2) The "antenna" during this feko calculation was based on two dipoles located near the big base as shown in the pic, locations and directions are clear shown.
3) I don't think a standard magnetron with the given antenna is a good choice to try to excite TE01 modes. (Maybe thats why Rfmwguy switch to TM013?)
4)I cannot simulate the full magnetron using feko lite, also not the big and noisy signal bandwidth.
5) I think a PLL controled source is a great choice to get a stable low bandwidth HF-source.
