-
#2480 by Monomorphic on 16 May, 2016 14:29
-
Wait...if you are deciding either/or, get youself the miniVNA tiny/VNA/J and not the sig gen. You will need a spec an though.
So the $330 RF Instruments SA0314 spectrum analyser + the $560 miniVNA tiny + the free VNA/J = ~$900 -
#2481 by Monomorphic on 16 May, 2016 14:38
-
That's the route I took, yes.Wait...if you are deciding either/or, get youself the miniVNA tiny/VNA/J and not the sig gen. You will need a spec an though.
So the $330 RF Instruments SA0314 spectrum analyser + the $560 miniVNA tiny + the free VNA/J = ~$900
Which did you get first? If I space them out it doesn't hurt as much! That's $200 more than I got official approval for.
-
#2482 by TheTraveller on 16 May, 2016 15:07
-
If this works out, Shawyer may prove to have had the right stuff "inventor's intuition". What I don't understand is why Shawyer dismisses other explanations like McCulloch's, White's etc: Edison, Westinghouse and Tesla didn't debate theories with Hertz: they invented and experimented.
Even Dr. White did not discuss his theory in his 2004 AIAA paper, which concentrated on the experiments.
Roger shared with me his theory has been developed with the help of several academics and has been debated on both side of the Atlantic.QuoteHi all
I guess by now you have all measured thrust from your experimental EmDrives. The measured results are likely to vary considerably with stability of input match, the type of measurement system etc. However it is likely that your experiments can enable verification of some aspects of the classic theory of EmDrive operation to be tested. This work has been carried out in a number of research organisations, and in each case the predictions of classic theory have been demonstrated.
I have taken the liberty of attaching a presentation of basic theory. This is a “high level” presentation aimed at non-specialists but it summarises the results of many years of discussion, argument and study amongst professional Physicists, Mathematicians and Engineers on both sides of the Atlantic.
In particular two aspects can be noted.
Inside the cavity Einstein rules, EM waves travel at relativistic velocities and the EmDrive can be considered an open system. This was first pointed out many years ago by an eminent UK government scientist working at Farnborough, the late David Fearn. If the cavity is operated below cut-off frequency, travelling waves are distorted and thrust approaches zero.
Outside the cavity Newton rules, thus action and reaction are equal and opposite. A fully constrained EmDrive thruster will therefore not exhibit thrust, whilst an unconstrained thruster will accelerate in the opposite direction to the thrust vector. Any static thrust measurement system will apply some constraint to the thruster and depending on spring constants etc, it can be set up to measure thrust, reaction force or zero force.
Unfortunately I am not now able to enter into public disclosure of anything other than general theory and the experimental work already published, but the information in the attachment has been widely discussed and feel free to distribute it and discuss it as you wish.
Good luck with your work, I hope that successful public experiments, such as yours, will eventually release the floodgates of information from the more constrained organisations that I work with.
Best regards
Roger
I don't see a lone wolf ignoring the comments and input of others.
-
#2483 by TheTraveller on 16 May, 2016 15:22
-
...
That has been done. And many issues have been pointed out. Recently an issue with the conservation of energy slide was asked very nicely to Shawyer in an e-mail exchange with rfmwguy, and Shawyer ignored the question entirely. (It was just a few pages back if you missed it)
Maybe review the attached.
As I just pointed out and then Rodal provided a link to where Mike McCulloch pointed out the same last year, the arrows on the Conservation of momentum slide indicate Shawyer's confusion with the direction of forces. Cullen's equations if applied blindly and the side walls are ignored result in a force (applied to the cavity walls) in the direction of the "thrust" arrow on that slide. Because from the definition of force, acceleration of an object is in the same direction as the force, the acceleration arrow should also be to the left.
Rather than reposting things that provide no new information, could you address any of the self-contradictions with this theory? Failing that, (since I can think of no way to resolve these contradictions other than simply saying Shawyer is wrong) you could just focus on experiment results and let other people consider theories that at least have a slim chance of explaining any anomalous forces instead of insisting people accept a theory that is equivalent to 1=0.
1st thing is for you to accept the EmDrive works as claimed and then to start exploring why Roger's theory does explains the engineering characteristics of the frustum that give rise to the predicted thrust generation.
As an engineer, I have followed, to the letter, as much as I could, all of Roger's papers, email breadcrumbs and his theory, which to date has always provided the correct force and very close to the correct thrust freq for the measured frustum and excitation conditions. BTW my measured thrust bandwidth and that of another EmDrive builder is much narrower than the rtn loss bandwidth and the peak thrust freq is not the peak rtn loss freq. Roger has recently shared why that can happen and it makes engineering sense as does everything he breadcrumbs. Guess we engineers know how to speak to each other in our common engineering lingo and belief systems that have seen, over the years, too many theory types get it very wrong.
When someone can come up with another theory that can generate the detail of engineering data that Roger's theory and breadcrumbs provides, well I might listen to them. But I'll not hold my breath for it to happen. -
#2484 by Rodal on 16 May, 2016 15:22
-
...I don't see a lone wolf ignoring the comments and input of others.
OK, it would help if the "Physicists, Mathematicians and Engineers on both sides of the Atlantic" he refers to would speak in public and write in support of Shawyer's theory, to address several issues. They were absent to defend Shawyer's theory when the New Scientist article was published. (See: https://www.newscientist.com/blog/fromthepublisher/2006/10/emdrive-on-trial.html )
At the moment we would appreciate an answer to the following question: when Shawyer refers to Kinetic Energy to address conservation of energy: how does he measure the velocity to use in his Kinetic Energy expression ? (the kinetic energy changes drastically with frame of reference), and how does he reconcile this frame of reference choice with Special Relativity?
For rockets using propellant exhaust (even relativistic rockets) this issue has been addressed, and there is universal agreement: the effective exhaust velocity of the propellant. The problem for any propellant-less device like the EM Drive is what is the reference velocity?
Has this been addressed?
Thanks
PS: The problem has to do with the "free energy" machine. McCulloch and White's theory imply that extra energy would come from the Quantum Vacuum (which would need to be a degradable, mutable vacuum, as in the Multiverse Quantum Vacuum that has several levels, instead of a single zero point). Shawyer, on the other hand posits that his device satisfies conservation of energy without needing anything external. -
#2485 by TheTraveller on 16 May, 2016 15:36
-
The miniVNA PRO is on sale for $490. Strange that normally the tiny is a little more expensive, but the PRO has more features, such as bluetooth, battery operation, more dynamic range, 1Hz vs 10Hz, 0dBm vs -6dBm. I guess you are paying for the miniaturization? http://www.hamradio.com/detail.cfm?pid=H0-012128
I too started with the mini/tiny VNA but soon the data from the Rf amps forward and reflected real time outputs and the ability to get my Rf gen to do freq sweeps turned it into a backup tool.
As I see it now there are 3 levels to getting the ideal thrust excitation freq:
1) Rough VNA S11 measurement to be sure you are in the right freq range and the rtn loss > loaded Q data looks acceptable as other go back to polishing. You can bet for sure Dave was grinning a wide smile when he saw the VNA data on his new frustum. For sure, for sure. Low rtn loss / loaded Q means you have issues to fix before going forward to powered tests.
2) Using real time forward and reflected signals to adjust the freq to get lowest VSWR at operational power. This tunes the entire system, Rf amp, leads, connectors, antenna coupler and frustum. If best VSWR, generated from Rf drive system forward and reflected power signals, is above say my max limit of 1.6 then something is wrong and you need to fix it before going further to acceleration or force based feedback freq tuning.
3) Using real time force or acceleration feedback to tune for max thrust. Don't be surprised if that freq turns out to be to be some distance away from best VNA rtn loss freq and best VSWR freq.
This is how basic EmDrive R&D works. Not quick, simple or low cost but this process does work if you put in the resources and time to make it work.
-
#2486 by TheTraveller on 16 May, 2016 15:43
-
...I don't see a lone wolf ignoring the comments and input of others.
OK, it would help if the "Physicists, Mathematicians and Engineers on both sides of the Atlantic" he refers to would speak in public and write in support of Shawyer's theory, to address several issues. They were absent to defend Shawyer's theory when the New Scientist article was published.
At the moment we would appreciate an answer to the following question: when Shawyer refers to Kinetic Energy to address conservation of energy: how does he measure the velocity to use in his Kinetic Energy expression ? (the kinetic energy changes drastically with frame of reference), and how does he reconcile this frame of reference choice with Special Relativity?
For rockets using propellant exhaust (even relativistic rockets) this issue has been addressed, and there is universal agreement: the effective exhaust velocity of the propellant. The problem for any propellant-less device like the EM Drive is what is the reference velocity?
Has this been addressed?
Thanks
The EmDrive works as claimed.
It is not a rocket engine so please try to stop trying to make it work like a rocket engine. The "expelled particle" that carries away lost EmWave momentum (big end momentum - small end momentum) is the momentum gained by the accelerating cavity.
Work backward from that to find what you seek.
The accelerative CofE issue will be soon subjected to experimental data via continual acceleration on rotary test tables as Roger did back in 2006. That data will reveal if the gained angular kinetic energy properly relates to consumed energy - losses or not.
Both QV and MiHsC suggest there are outside energy sources involved to maintain accelerative CofE. Roger says CofE is maintained but not as we perceive and there are no outside energy sources involved nor needed. We should have some prelim data by end 3rd qtr.
Will be exciting to see how this unfolds. Popcorn anyone? -
#2487 by Rodal on 16 May, 2016 16:00
-
...
OK, I admit I like your experimental approach to verify conservation of energy and verify operation of the EM Drive, if you perform your experiments with batteries, so that the center of momentum and the center of energy are in the moving platform, instead of using power cords.
The accelerative CofE issue will be soon subjected to experimental data via continual acceleration on rotary test tables as Roger did back in 2006. That data will reveal if the gained angular kinetic energy properly relates to consumed energy - losses or not.
Both QV and MiHsC suggest there are outside energy sources involved to maintain accelerative CofE. Roger says CofE is maintained but not as we perceive and there are no outside energy sources involved nor needed. We should have some prelim data by end 3rd qtr.
Will be exciting to see how this unfolds. Popcorn anyone?
While I look forward to such experimental results, I wonder: why did it take an outsider, TheTraveller , to realize that these experiments were needed, and why have not these experiments been performed before?
Perhaps we will also get in 2016 similar experimental results from NASA, and Cannae (battery operated, and using a torsional pendulum, under partial vacuum to eliminate thermal convection effects), and this issue will be put to bed... If that works, an experiment in Space should follow
-
#2488 by X_RaY on 16 May, 2016 16:12
-
Sorry this is not really helpful to get answers to fix the addressed "holes" in (t)his theory
The EmDrive works as claimed.
...
Work backward from that to find what you seek.
snip...
-
#2489 by Monomorphic on 16 May, 2016 16:14
-
I too started with the mini/tiny VNA but soon the data from the Rf amps forward and reflected real time outputs and the ability to get my Rf gen to do freq sweeps turned it into a backup tool.
What equipment are you using now for RF gen? It would be helpful for DIYers to know so they aren't wasting time and money on tools that aren't good enough in the long-run. -
#2490 by TheTraveller on 16 May, 2016 16:26
-
Hi Guys,
Before my amp died some of the tests I did indicated the best thrust freq was not the same as the best rtn loss freq. I believe this is due to frustum manufacturing errors introducing phase distortion in the multiple reflected travelling waves being compensated by a slightly off freq excitation such that the say negative phase distortions introduced by the freq are compensated by positive phase distortions caused by the manufacturing defects resulting in the min overall phase distortions over the life of the travelling wave & thus generating a higher thrust than operating at best rtn loss freq.
I also found the thrust bandwidth was very much narrower than the rtn loss bandwidth.
This info is preliminary & needs to be confirmed by a lot more tests when I get my amp back.
Paul March, do you have the ability to tune your excitation freq as against thrust generated in real time? If not might be worth a thought as to how to do this to both tune freq for max thrust as well as to map the thrust versus freq bandwidth as against rtn loss versus freq bandwidth. I'm sure there is gold to mine in that data.
Roger send me an email reply to this post:QuoteHi Phil
From your NSF posting, I understand that you and another builder have measured thrust bandwidth. The reason why the thrust bandwidth is smaller than the measured cavity bandwidth is because the measured cavity Q is loaded by the Q of the input circuit. In an ideal match, the loaded Q is half the unloaded Q of the cavity.
Classic EmDrive theory however predicts that thrust is dependent on the cavity unloaded Q, as well as the power actually input into the cavity, which is dependent on the match.
Looks like you are both on the right track.
Good to hear your health is improving.
Best regards
Roger
Of course this is EmDrive "yes it does work and here is another breadcrumb" engineering talk and a long way from "it can't work" theory talk. -
#2491 by TheTraveller on 16 May, 2016 16:52
-
I too started with the mini/tiny VNA but soon the data from the Rf amps forward and reflected real time outputs and the ability to get my Rf gen to do freq sweeps turned it into a backup tool.
What equipment are you using now for RF gen? It would be helpful for DIYers to know so they aren't wasting time and money on tools that aren't good enough in the long-run.
I'm using a low cost USB freq gen that steps in 1kH increments. As the output is low, I use a 100mW WiFi booster to drive the Rf amp but only over a limited freq range. As the Rf amp has a 31 dBm stepable attenuator in 1dBm steps and I can directly measure forward and reflected, I'm not that fussed with the exact Rf amp drive level as long as it is high enough to get good stable operation.
Signal generator:
http://www.ebay.com/itm/138M-4-4G-USB-SMA-signal-source-signal-generator-simple-spectrum-analyzer-/
I'm developing laptop software to directly control this unit without using the supplied software so I can implement a semi automatic / automatic freq control of best VSWR or best acceleration and best frustum freq tracking or what even I need to do with the freq.
100mW WiFi amp:
Can't find it on ebay at the moment. Will try to find something equivalent. This units dBm output depends on the min Rf amp input drive dBm requirements, so your Rf amp may have different drive requirements to my amp. -
#2492 by TheTraveller on 16 May, 2016 17:08
-
Hi Phil, glad you're feeling better. I think we should probably leave the correct theory to a future date. Mr Shawyer helped and inspired many people. White, McCullough and others took a different theory direction, but we're all on the same team...an obscure and very interesting device that has tremendous ramifications if all the pieces of the puzzle align.
I'll leave the theory to the experts. You and I can build and test to whichever design(s) we like...that's part of the enjoyment. WE pick and choose, not a lab supervisor
Dave,
So far only Roger's theory lead breadcrumbs reveals has lead to my successful build. If I tried to follow what is proposed here, well there would be no thrust. OK hard words but true.
As an example FEKO is a good get you close tool but it has serious limitations as it failed to show a smaller than cutoff small end would not work. It showed nice mode images that were incorrect and would never gen thrust.
Roger shared with me this is an issue with all the commercial analysis packages he knows of and is why SPR wrote their own in house analysis package to help to get it right before cutting metal. I can tell you I would not be spending $7k on my spherical frustum based on any other analysis data from this forum. The issue is no one here has had real world experience in obtaining thrust and then to feed that data back into making the analysis package model the reality of the build and the thrust.
What is presented here by many well intentioned contributors is best guess work, without any EmDrive thrust feedback and can be wrong. No nice way to say that.
SPR has the tools and EmDrive experience to do designs and builds right the 1st time. Roger's breadcrumbs have helped me to get it right the 1st time. What he shares with me, I share here.
So sorry to disagree but the theory need to support the engineering or there will be no thrust. -
#2493 by zen-in on 16 May, 2016 17:17
-
So, I thought I would sweep the frustum BEFORE soldering closed...yeah I know, its not a good idea. So, looks like center is around 2.474 GHz, higher than the model, but within mag range. Return Loss 43dB which is outstanding without solder. If I measure Q from 3dB off best RL, it would be crazy high. Not going to go there yet, things will change after it gets soldered. Bottom line, almost ready to button up...
A quick estimate says the width of the notch 3 dB up from the max. deflection is maybe 100 MHz. I don't know how that would be used to estimate the Q. I think a stronger case for estimating the Q can be made for the absolute maximum deflection of the trace. Any setup will show a narrow 3dB notch but only the best network analyzer will show how deep the notch is. The notch may also be wider with a better instrument. There are many different things that can influence the S11 trace: quality of connectors, cable, and instrument used are some. If the cable is jiggled does the shape of the trace change? Are all the connectors and cable first quality cables and have the connectors all been tightened to the correct torque? The quality of the instrument doing the measurement is unknown if calibration tests are not done. Among the prized possessions of the RF engineers I know are a set of calibration connectors. They include a dead short, an open connector, and terminations of different impedances. Using these calibration connectors the measurements from a NIST certified instrument can be compared to a low cost instrument and corrections made. It's also a good idea to test a filter that has performance plots from the manufacturer on the instrument you are using and see how that measurement compares to what the manufacturer supplied.
-
#2494 by Rodal on 16 May, 2016 17:35
-
Hi Phil, glad you're feeling better. I think we should probably leave the correct theory to a future date. Mr Shawyer helped and inspired many people. White, McCullough and others took a different theory direction, but we're all on the same team...an obscure and very interesting device that has tremendous ramifications if all the pieces of the puzzle align.
You had asked to quantify the effect of the power cord on the EM Drive experiments.
I'll leave the theory to the experts. You and I can build and test to whichever design(s) we like...that's part of the enjoyment. WE pick and choose, not a lab supervisor
Unfortunately such quantification depends on the nature of the anomalous force: according to classical physics the EM Drive should not self-accelerate. So according to classical physics, the power cord should only make a difference regarding experimental artifacts: Lorentz forces, thermal expansion of the power cord, stiffness of the power cord, etc.
However, we can make the following clear statements:
1) All theories of operation (Shawyer, McCulloch, NotSoSureOfIt, etc.) posit that the anomalous force is due to the photons inside the EM Drive, that experience a change in wavelength along the longitudinal direction in which the EM Drive is tapered. The theories differ as to where does the exterior energy required to satisfy conservation of energy and conservation come from.
2) Photons do not have a rest mass (which is the invariant measure of mass in Relativity) and their relativistic mass is undefined (since zero rest mass times infinite gamma (because they travel at speed c), is undefined). Therefore it is not satisfactory to discuss EM Drive anomalous force as being due to a shift in center of mass. No, what photons have, according to universal consensus is energy and momentum. Therefore any satisfactory discussion should be in terms of center of energy and center of momentum. Not center of mass.
3) Power cords introduce both energy and momentum into the EM Drive. This cannot be ignored: if the power is not on, there is no anomalous force. The anomalous force is, as shown by experiments, due to the power being on.
4) The feeding point of the energy and momentum also cannot be ignored: experiments have shown that it makes a difference whether the energy and momentum are fed in the middle of the EM Drive vs. being fed at one end or the other.
5) Therefore it is evident that the source of energy and momentum (the power cord) cannot be ignored. An experiment run with batteries in the moving platform has the center of energy and the center of momentum in the moving platform. Conversely, an experiment run with power cords with energy from a stationary supply does not have the center of energy and the center of momentum at the same place as the center of mass of the EM Drive: the center of energy and momentum may be outside the moving platform.
6) I have shown that coaxial power cables do not at all solve this issue. On the contrary coaxial cables deliver practically 100% of the electromagnetic momentum along their length. Running the coaxial cable from different directions does not make a difference, because the electromagnetic momentum gets delivered along whatever path you use. Once the coaxial cable RF momentum is fed, for example at the middle of the EM Drive, then that electromagnetic momentum and power become the photons that have increased momentum towards the big end and decreased momentum towards the small end. For these purposes (center of energy-momentum) it does not matter how you run the coaxial cable from one end to the other. What matters is whether the power source is stationary, with respect to the moving EM Drive, and where in the EM Drive is the RF energy and momentum injected.
7) Running an experiment with a power cord is therefore similar to running an experiment with a normal rocket using propellant coming from a stationary source. Such a test may be OK once one knows that rockets work, and one just wants to test their thrust, but it is unsatisfactory for a propellant-less device that is very much under question whether it can work in Space to self-accelerate.
No propellant ejection for a normal rocket = no thrust
No RF power fed into an EM Drive = no anomalous force
To conclusively show that rockets worked, Goddard conducted experiments with the propellant and everything else self-integrated in the rocket and showed lift-off in his experiments. Similarly, to conclusively show that an EM Drive, that works with photons that do not have mass, and only have energy and momentum, can work in outer space, the most conclusive test is the one where batteries and everything else are self-integrated in the moving platform, and not one where electromagnetic energy and momentum are fed from a stationary source.
An experiment with a power cord from a stationary source can perhaps show that one can create an anomalous force on a tapered resonant cavity being fed from a stationary source. It does not prove that there is an anomalous force acting on the whole integrated system (battery+EMDrive+etc.) when the energy and momentum are self-integrated together.
-
#2495 by TheTraveller on 16 May, 2016 18:03
-
One example of limitations in existing analysis packages is, as far as I know, there is no ability to plot the guide wavelength and related EmWave momentum changes from the big to small end of the frustum. As the momentum in the EmWave is directly related to the inverse of the guide wavelength change, being able to model it aids understand how the EmWave momentum varies from the Big to Small end.
Adding this to my Excel model was one of my 1st tasks.
Attached is the guide wavelength plot for my 1st build. As the EmWave momentum is directly related to the inverse of the guide wavelength (longer guide wavelength = smaller EmWave momentum), the reverse of the attached curve shows how EmWave momentum drops as the guide wavelength lengthens as it approaches the small end.
In Roger's thrust equation F = (2 Qu Pwr Df) / c, Df is the complex relationship between the big and small end guide wavelengths as attached.
It is not possible to do this modeling without reference to the EmDrive theory developed by Roger and SPR over the years.
-
#2496 by TheTraveller on 16 May, 2016 18:16
-
The 3dB up points from -43dB are probably about a MHz or less wide. The 3dB down points from insertion are wider. Everything was variable as the top and bottom plates were not even soldered, so I didn't get too excited. Just wanted to see generally where peak was.So, I thought I would sweep the frustum BEFORE soldering closed...yeah I know, its not a good idea. So, looks like center is around 2.474 GHz, higher than the model, but within mag range. Return Loss 43dB which is outstanding without solder. If I measure Q from 3dB off best RL, it would be crazy high. Not going to go there yet, things will change after it gets soldered. Bottom line, almost ready to button up...
A quick estimate says the width of the notch 3 dB up from the max. deflection is maybe 100 MHz. I don't know how that would be used to estimate the Q. I think a stronger case for estimating the Q can be made for the absolute maximum deflection of the trace. Any setup will show a narrow 3dB notch but only the best network analyzer will show how deep the notch is. The notch may also be wider with a better instrument. There are many different things that can influence the S11 trace: quality of connectors, cable, and instrument used are some. If the cable is jiggled does the shape of the trace change? Are all the connectors and cable first quality cables and have the connectors all been tightened to the correct torque? The quality of the instrument doing the measurement is unknown if calibration tests are not done. Among the prized possessions of the RF engineers I know are a set of calibration connectors. They include a dead short, an open connector, and terminations of different impedances. Using these calibration connectors the measurements from a NIST certified instrument can be compared to a low cost instrument and corrections made. It's also a good idea to test a filter that has performance plots from the manufacturer on the instrument you are using and see how that measurement compares to what the manufacturer supplied.
The 3dB bandwidths have been argued a lot. My test report tried to resolve this by characterizing the shape factor of return loss by creating a ratio of insertion 3dB to peak 3dB bandwidths. This forces a measurement of both bandwidths to avoid confusion. I think Yang measured off peak, EW measured off insertion.
The total span of this image was 2.4 to 2.5 GHz I believe. Insertion return loss should improve once its all soldered. I am using an 8 inch piece of silver plated coax with an SMA from the BeO mounted monopole to the VNA.
Yes, you're correct, a calibrated connector set, cal kits and cables are worth their weight in gold. Especially when working up to and beyond 18 GHz. If my observation tests looks good, I can run down to my old company and have an NIST calibrated sweep done for accuracy. For me, it all hinges on if I can scale the observe effect up giving much more attention to the cavity. If it does, I'll do more. If not, I will retire from building emdrives but will continue to follow developments.
Roger has shared with us how SPR measure frustum loaded Q and from that how to calculated unloaded Q as used in his thrust equation.
I do agree with Zen-In that your and mine low cost VNA only do an approximate job and the result from a high price VNA will probably be very different.
However what I have found from my single freq excitation work is it is good enough to get an approximate result, which is good enough to strongly suggest your maggie should trigger bursts of resonance inside the frustum bandwidth and that your polish and general construction accuracy should result is some nicely measurable thrust.
As you now have solid side walls, do a gravity stack, which is as far as I got and using your VNA, try applying some significant weight on different parts of the outer end plate edge to see if this shifts peak rtn loss resonance or changes rtn loss depth. With my 0.5mm thick side walls and 1mm thick end plates, I did see changes, which to me suggests I was altering the parallelism of the end plates by very slightly compressing the side walls at various points as Dr. Rodal suggest can be done with such thin side walls.
You are getting to the fun time as you have a new and very shiny toy to play with and gather good data from. Enjoy. I'm jealous as I can't yet play again with my shiny toy. Well not so shiny at the moment. -
#2497 by X_RaY on 16 May, 2016 18:56
-
The factor of 2 between Q and Q0 in the case of critical coupling is well known.
The 3dB up points from -43dB are probably about a MHz or less wide. The 3dB down points from insertion are wider. Everything was variable as the top and bottom plates were not even soldered, so I didn't get too excited. Just wanted to see generally where peak was.So, I thought I would sweep the frustum BEFORE soldering closed...yeah I know, its not a good idea. So, looks like center is around 2.474 GHz, higher than the model, but within mag range. Return Loss 43dB which is outstanding without solder. If I measure Q from 3dB off best RL, it would be crazy high. Not going to go there yet, things will change after it gets soldered. Bottom line, almost ready to button up...
A quick estimate says the width of the notch 3 dB up from the max. deflection is maybe 100 MHz. I don't know how that would be used to estimate the Q. I think a stronger case for estimating the Q can be made for the absolute maximum deflection of the trace. Any setup will show a narrow 3dB notch but only the best network analyzer will show how deep the notch is. The notch may also be wider with a better instrument. There are many different things that can influence the S11 trace: quality of connectors, cable, and instrument used are some. If the cable is jiggled does the shape of the trace change? Are all the connectors and cable first quality cables and have the connectors all been tightened to the correct torque? The quality of the instrument doing the measurement is unknown if calibration tests are not done. Among the prized possessions of the RF engineers I know are a set of calibration connectors. They include a dead short, an open connector, and terminations of different impedances. Using these calibration connectors the measurements from a NIST certified instrument can be compared to a low cost instrument and corrections made. It's also a good idea to test a filter that has performance plots from the manufacturer on the instrument you are using and see how that measurement compares to what the manufacturer supplied.
The 3dB bandwidths have been argued a lot. My test report tried to resolve this by characterizing the shape factor of return loss by creating a ratio of insertion 3dB to peak 3dB bandwidths. This forces a measurement of both bandwidths to avoid confusion. I think Yang measured off peak, EW measured off insertion.
The total span of this image was 2.4 to 2.5 GHz I believe. Insertion return loss should improve once its all soldered. I am using an 8 inch piece of silver plated coax with an SMA from the BeO mounted monopole to the VNA.
Yes, you're correct, a calibrated connector set, cal kits and cables are worth their weight in gold. Especially when working up to and beyond 18 GHz. If my observation tests looks good, I can run down to my old company and have an NIST calibrated sweep done for accuracy. For me, it all hinges on if I can scale the observe effect up giving much more attention to the cavity. If it does, I'll do more. If not, I will retire from building emdrives but will continue to follow developments.
Roger has shared with us how SPR measure frustum loaded Q and from that how to calculated unloaded Q as used in his thrust equation.
I do agree with Zen-In that your and mine low cost VNA only do an approximate job and the result from a high price VNA will probably be very different.
However what I have found from my single freq excitation work is it is good enough to get an approximate result, which is good enough to strongly suggest your maggie should trigger bursts of resonance inside the frustum bandwidth and that your polish and general construction accuracy should result is some nicely measurable thrust.
As you now have solid side walls, do a gravity stack, which is as far as I got and using your VNA, try applying some significant weight on different parts of the outer end plate edge to see if this shifts peak rtn loss resonance or changes rtn loss depth. With my 0.5mm thick side walls and 1mm thick end plates, I did see changes, which to me suggests I was altering the parallelism of the end plates by very slightly compressing the side walls at various points as Dr. Rodal suggest can be done with such thin side walls.
You are getting to the fun time as you have a new and very shiny toy to play with and gather good data from. Enjoy. I'm jealous as I can't yet play again with my shiny toy. Well not so shiny at the moment.
I think to work with thicker walls is a good idea due to possible deformation issues.
Is your frequency tracking system complete or planned for later?
-
#2498 by Rodal on 16 May, 2016 19:12
-
...It is not possible to do this modeling without reference to the EmDrive theory developed by Roger and SPR over the years.
It is not possible to calculate Shawyer's force without Shawyer's equation.
HOWEVER:
Expressing the wavelength as per longitudinal location is no problem with:
1) the exact solution to the problem. This can be readily shown from Greg Egan's solution to the problem. It has to do with the fact that the exact solution: spherical Bessel functions have a non-uniform wavelength, as I have shown in my post previously. This was known much prior to Roger Shawyer. It was actually proven by the great Russian-American engineer Dr. Schelkunoff at Bell Labs in the 1930's (maybe before Shawyer was born ?).
Picture of Schelkunoff showing frustums of a cone waveguides in the 1930's (prior to many people in this forum being born?)
There is no need to approximate with an integration function in a spreadsheet (although such Excel spreadsheet as the one from TheTraveller do indeed play a very useful role in engineering design because they are much easier to run than a program using Spherical Bessel functions and Legendre functions).
The reason why it is important to remark this is because the fact that the wavelength in a tapered cavity changes along the length is not a subject of dispute: there is uniform agreement between Shawyer and Greg Egan on this (whose post is to debunk the EM Drive).
The disagreement between Egan and Shawyer is not anything to do with the wavelength changing along the length, but is instead whether the EM Drive can self-accelerate as a result of this.
2) Meep which was written at MIT as an open source so that users themselves could write their own code for calculations such as this (Meep was not meant to be used as a "black box" like FEKO or COMSOL which are "black boxes" where the user can make only small or no changes). -
#2499 by Monomorphic on 16 May, 2016 19:54
-
The miniVNA PRO is on sale for $490. Strange that normally the tiny is a little more expensive, but the PRO has more features, such as bluetooth, battery operation, more dynamic range, 1Hz vs 10Hz, 0dBm vs -6dBm. I guess you are paying for the miniaturization?
I didn't check the frequency range. Looks like the miniVNA PRO only has a range of 100KHz to 200MHz (extendable to 1.5Ghz with an add-on). Not really good for emdrive research unless you're building a HUGE emdrive. I'm going to delete my original post so no one makes the mistake of purchasing this unit.
Suggest Dave delete the one where he recommended I purchase it since it was one sale.
The miniVNA TINY has a frequency range from 1MHz to 3Ghz. Which is good for emdrive RF testing. http://miniradiosolutions.com/54-2/
Suggest Dave delete the one where he recommended I purchase it since it was one sale.