Author Topic: EM Drive Developments - related to space flight applications - Thread 3  (Read 1878669 times)

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
Just to note that there have been dozens of experiments on The Woodward Effect which have returned a result compatible with the null hypothesis. This is not something that one normally hears about, though.

Many nails in that coffin, sad to say.
I would be interested in knowing of any nullification experiments for Woodward's hypothesis.  The only one I had seen was the one by Brito Marini and Galian but a number of Mach Effect supporters came from other threads and posted that I should ignore it because that report dealt with what they call Mach Lorentz thruster (MLT) and not the Mach Effect thruster (MET) (or something to that effect MET dealing with piezoelectric effects instead of electromagnetic effects for MLT, of something like that, I don't recall the gory details).  Any other nulllifiication reports, please
To be fair they were largely MLTs. I kinda lost interest after that.

Offline deltaMass

  • Full Member
  • ****
  • Posts: 955
  • A Brit in California
  • Liked: 671
  • Likes Given: 275
If output power of a generator = m v a, how can power be extracted at constant speed? For then, a=0. What are the relevant equations?
P = F v comes in somewhere.

ETA OK I get it. There is no need to cycle acceleration/deceleration phases in the over-unity power generator. Constant speed operation will work fine.  It goes like this:

- The speed v is set to something higher than effective power breakeven speed (i.e. including the efficiency (e) term): v = n/k, where n > 1/e
- The dissipative load exerts an equal and opposite force to the driving force, so acceleration = 0 and
F = k  Pin
- The dissipated power PD = F v = k Pin *  n/k = n Pin

And that's all folx.
« Last Edit: 07/12/2015 03:35 AM by deltaMass »

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2335
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2982
  • Likes Given: 2601
Been a long day and I'm wore out. I've got plenty to digest in the morning when the head is working and quiet dominates. I see you hacked the perforated copper up one side and down the other, good thoughts but GN all.

Shell

Offline Ricvil

  • Member
  • Posts: 69
  • Liked: 66
  • Likes Given: 24
I'm just looking for a possible hybrid mode on cavity. But it is not so easy.
There are some possibles candidates on Nasa's paper.
Some formulas of sensitivity can be used for adjust the dimensions of cavity and to control the frequencys.
Very cool.
By the way. In corrugated waveguides, hybrid modes have very low losses.  In cavity, perhaps  they produce more higher Q.

Interesting. I'd like to see a simulation of what a doppler shift does to the RF in the cavity, forward/reverse paths mode-split or something. If it takes an hour to FDTD simulate a few dozen nanoseconds, is it 100 years to simulate a dozen milliseconds?

Oh well.

There's another graph that was posted here, showing frequency and modes "Shawyer Conical Resonant Cavity Modes-2 (4).jpg". Couldn't find it using search, here it is again:



Oh oh oh!!
Nice. The graph shows the mixing of the modes as a function of cavity shape.
There is a choice of frequency, and dimension on graph. Any specific motivation?

I'm looking for a configuration when there is one hybrid mode, followed by any other mode very close.
Why? Because I think there is another thing interacting with the electromagnetic field of cavity.
This thing probally will have a very small coupling with the electromagnetic field.
To enhance this coupling I need:
- A mode excited by a source puting the max energy on cavity
- At least one second mode with frequecy very close to the first but not excited by the source.
In this situation, when a small region of the cavity has its electromagnetics properties changed to anything different of vacuum (epslon0 and mu0), then this little "scatter" region triggers a very strong perturbation called "ghost mode".
In waveguides "ghost modes" are caused by deformations or imperfections on waveguide, but in principle, any "pertubation" can cause this effect.
This "ghost modes" can in some situations, reflect almost all power flux in the waveguide, and the "scatter" will be under very strong radiation pressure.
I don't know if this case can be considered also a type of Fano resonance, but I think if I want some type of interaction of the field inside of cavity with some "other thing", I would try  to maximize this interaction with this setup.

To me this thing is the axion field/particle. To others can be particles from "quantum vacuum" or a space-time flutuactions, but the result of the ghost mode arising is the same,  change the incidence of electromagnetic radiation on the walls of cavity.
« Last Edit: 07/12/2015 02:43 PM by Ricvil »

Offline dumbo

  • Member
  • Posts: 15
  • I believe eleph^H^H^H^H^HEMDrives can FLY
  • Tianjin, China
  • Liked: 9
  • Likes Given: 77
Ok, so from previous posts by aero, I was able to find the Bradycone3D.ctl file used for the meep run. I have now also contacted the FEFF project requesting access to their MEEP AMI.

Assuming they agree to this request, I can then start to run the simulation in an Amazon EC2 instance where the concerns of wives and whatnot will not influence the possible computation length, and so will be able to run over thousands of cycles.

Offline aero

  • Senior Member
  • *****
  • Posts: 2786
  • 92129
  • Liked: 724
  • Likes Given: 249
Ok, so from previous posts by aero, I was able to find the Bradycone3D.ctl file used for the meep run. I have now also contacted the FEFF project requesting access to their MEEP AMI.

Assuming they agree to this request, I can then start to run the simulation in an Amazon EC2 instance where the concerns of wives and whatnot will not influence the possible computation length, and so will be able to run over thousands of cycles.

It seems you are serious. So use the right model, the one for which Dr. Rodal has already performed a significant analysis on the short time runs. It is a model of rfmwguy's NSF-1701 cavity. It is attached, and is set to produce output.

A few points.
Line 76 names the output directory
Line 230 "cc" defines the number of cycles to run currently set to 32. We need to decide how much data is needed.
Line 251 starts the output data flow
followed by 6 lines naming the field types and output frequency.

I have not yet discovered how to turn the output off, after turning it on but that is something that you will likely need. A 13 cycle .h5 data set of 14 time slices of one field component, (1.4 cycles) is 1.9 GB, so the six components weigh in at about 11.4 GB. Not really something you'd want to upload to many times but needed every so often.
Retired, working interesting problems

Offline apoc2021

  • Member
  • Posts: 18
  • Liked: 37
  • Likes Given: 27
Assuming they agree to this request, I can then start to run the simulation in an Amazon EC2 instance where the concerns of wives and whatnot will not influence the possible computation length, and so will be able to run over thousands of cycles.

It seems you are serious. So use the right model, the one for which Dr. Rodal has already performed a significant analysis on the short time runs.

Let me know if/when you need an AWS account.

Offline demofsky

  • Full Member
  • *
  • Posts: 166
  • Liked: 96
  • Likes Given: 1671
Assuming they agree to this request, I can then start to run the simulation in an Amazon EC2 instance where the concerns of wives and whatnot will not influence the possible computation length, and so will be able to run over thousands of cycles.

It seems you are serious. So use the right model, the one for which Dr. Rodal has already performed a significant analysis on the short time runs.

Let me know if/when you need an AWS account.

The other nice thing about this is that you don't have to upload anything to the cloud.  Just use the AWS version of google drive.

Offline dumbo

  • Member
  • Posts: 15
  • I believe eleph^H^H^H^H^HEMDrives can FLY
  • Tianjin, China
  • Liked: 9
  • Likes Given: 77
It seems you are serious. So use the right model, the one for which Dr. Rodal has already performed a significant analysis on the short time runs. It is a model of rfmwguy's NSF-1701 cavity. It is attached, and is set to produce output.

A few points.
Line 76 names the output directory
Line 230 "cc" defines the number of cycles to run currently set to 32. We need to decide how much data is needed.
Line 251 starts the output data flow
followed by 6 lines naming the field types and output frequency.

I have not yet discovered how to turn the output off, after turning it on but that is something that you will likely need. A 13 cycle .h5 data set of 14 time slices of one field component, (1.4 cycles) is 1.9 GB, so the six components weigh in at about 11.4 GB. Not really something you'd want to upload to many times but needed every so often.

Many thanks for the file aero, and for the additional clarifying comments you have made. Given the sizes of the time-slices, I am thinking a possible space-saving shortcut would be to have meep compute the Poynting-vector directly, instead of after the computation has finished. So basically just replacing:
(after-time (* 30.7 T_meep)              ; to use, uncomment this and the closing parin.   
        (to-appended "ex" (at-every (* .1 T_meep) output-efield-x))  ; time evolution of fields.
        (to-appended "ey" (at-every (* .1 T_meep) output-efield-y))
        (to-appended "ez" (at-every (* .1 T_meep) output-efield-z))
        (to-appended "hx" (at-every (* .1 T_meep) output-hfield-x))
        (to-appended "hy" (at-every (* .1 T_meep) output-hfield-y))
        (to-appended "hz" (at-every (* .1 T_meep) output-hfield-z)) )
with
(after-time (* 30.7 T_meep)
       (to-appended "poynting" (synchronized-magnetic output-poynting)))

This will save us from shuffling terabytes of data around. Please let me know if you have any objections to this idea, maybe I am misunderstanding something about how meep works.

Offline TheTraveller

I would suggest to every EMDrive builder that having a Vector Network Analyser and understanding what it is telling you is critical to your success.

Doing 1 cable SLOW sweeps can test your excitation antenna design, impedance matching, find your cavity resonances, peak return loss dBs, VSWR, reflection coefficient, Q and cavity bandwidth.

Without a VNA, it is hard to understand how any build could achieve Force generation as there are so many VNA determinant / measurable variables that need to be all as close as possible to optimum values.

My soon to arrive 100W Rf amp has the ability to be throttled from 50dBm (100W) output to 19dBm (79mW) output and has a real time analogue VSWR output pin, which I plan to use to generate a DIY VNA as knowing the VSWR, Rf amp impedance and the frequency will enable real time calculation of the return loss db, the reflection coefficient, how well the cavity is impedance matched to the Rf amp, the cavity unloaded Q and Q bandwidth.

http://cgi.www.telestrian.co.uk/cgi-bin/www.telestrian.co.uk/vswr.pl

Seems a lot of information to get from 1 pin of analogue information, but the Rf amp is driving the cavity at it's impedance, output frequency and reporting on what is happening via the VSWR analogue output. Additionally doing it this way ensures each EMDrive is impedance matched as close as possible to it's 100W Rf amp.

Comments from the experienced microwave guys that have used 1 cable VNA measurements most welcome.
« Last Edit: 07/12/2015 12:21 PM by TheTraveller »
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline rfmwguy

  • EmDrive Builder (retired)
  • Senior Member
  • *****
  • Posts: 2165
  • Liked: 2681
  • Likes Given: 1124
I would suggest to every EMDrive builder that having a Vector Network Analyser and understanding what it is telling you is critical to your success.

Doing 1 cable SLOW sweeps can test your excitation antenna design, impedance matching, find your cavity resonances, peak return loss dBs, VSWR, reflection coefficient, Q and cavity bandwidth.

Without a VNA, it is hard to understand how any build could achieve Force generation as there are so many VNA determinant / measurable variables that need to be all as close as possible to optimum values.

My soon to arrive 100W Rf amp has the ability to be throttled from 50dBm (100W) output to 19dBm (79mW) output and has a real time analogue VSWR output pin, which I plan to use to generate a DIY VNA as knowing the VSWR, Rf amp impedance and the frequency will enable real time calculation of the return loss db, the reflection coefficient, how well the cavity is impedance matched to the Rf amp, the cavity unloaded Q and Q bandwidth.

http://cgi.www.telestrian.co.uk/cgi-bin/www.telestrian.co.uk/vswr.pl

Seems a lot of information to get from 1 pin of analogue information, but the Rf amp is driving the cavity at it's impedance, output frequency and reporting on what is happening via the VSWR analogue output. Additionally doing it this way ensures each EMDrive is impedance matched as close as possible to it's 100W Rf amp.

Comments from the experienced microwave guys that have used 1 cable VNA measurements most welcome.
Agreed but a 1 port measurement is simply return loss or an antenna measurement not a bandpass cavity where a sample port is needed for S21. call it reflected S11 Q not forward S21 Q. This will give unnaturally high numbers leading to confusion. If u get the chance, ask shawyer for S21 plots since he had a sample port. This should illustrate assymetrical bandpass response.

Offline TheTraveller

I would suggest to every EMDrive builder that having a Vector Network Analyser and understanding what it is telling you is critical to your success.

Doing 1 cable SLOW sweeps can test your excitation antenna design, impedance matching, find your cavity resonances, peak return loss dBs, VSWR, reflection coefficient, Q and cavity bandwidth.

Without a VNA, it is hard to understand how any build could achieve Force generation as there are so many VNA determinant / measurable variables that need to be all as close as possible to optimum values.

My soon to arrive 100W Rf amp has the ability to be throttled from 50dBm (100W) output to 19dBm (79mW) output and has a real time analogue VSWR output pin, which I plan to use to generate a DIY VNA as knowing the VSWR, Rf amp impedance and the frequency will enable real time calculation of the return loss db, the reflection coefficient, how well the cavity is impedance matched to the Rf amp, the cavity unloaded Q and Q bandwidth.

http://cgi.www.telestrian.co.uk/cgi-bin/www.telestrian.co.uk/vswr.pl

Seems a lot of information to get from 1 pin of analogue information, but the Rf amp is driving the cavity at it's impedance, output frequency and reporting on what is happening via the VSWR analogue output. Additionally doing it this way ensures each EMDrive is impedance matched as close as possible to it's 100W Rf amp.

Comments from the experienced microwave guys that have used 1 cable VNA measurements most welcome.
Agreed but a 1 port measurement is simply return loss or an antenna measurement not a bandpass cavity where a sample port is needed for S21. call it reflected S11 Q not forward S21 Q. This will give unnaturally high numbers leading to confusion. If u get the chance, ask shawyer for S21 plots since he had a sample port. This should illustrate assymetrical bandpass response.

As my Rf gen frequency can be stepped in 1kHz increments and the Rf amps power dropped to 79 mWs, simple to then initially find the biggest return dB loss frequency (lowest VSWR as measured by the Rf amp and converted to rtn loss dBs) and then step the frequency +- away from the centre frequency until the return loss peak drops 3dB (Rf amp measured VSWR increases, so rtn loss dB decreases) from the max return loss value frequency at the higher and lower frequency sides of the slope.

Do that for both sides and I then have the cavity bandwidth and unloaded Q.

Correct?

From your description, what I'm measuring is reflected S11 Q. Why is that not a valid Q measurement of what is going on inside a non accelerating cavity? What I'm proposing to measure is what the Rf amp will be seeing when it drives my EMDrive cavity.

As far as I know, the Shawyer, Chineses and Eagleworks sample ports were used as feedback into the Rf frequency tracking system to keep the externally generated frequency always locked to the dynamically real time changing cavity resonance frequency.

Another thought. Just maybe Shawyer's Force equation F = (2 Df P Q) / c is based on the Q being measured as reflected S11 Q and thus the high values are correct as far as the equation is concerned? Will ask him if the Qs he and the Chinese quote are based on reflected S11 Q or forward S21 Q?

For now I will assume Shawyer's and the Chinese Qs are based on reflected S11 Q, because that seems to fit the measured and quoted Q data.
« Last Edit: 07/12/2015 01:58 PM by TheTraveller »
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2335
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2982
  • Likes Given: 2601
I would suggest to every EMDrive builder that having a Vector Network Analyser and understanding what it is telling you is critical to your success.

Doing 1 cable SLOW sweeps can test your excitation antenna design, impedance matching, find your cavity resonances, peak return loss dBs, VSWR, reflection coefficient, Q and cavity bandwidth.

Without a VNA, it is hard to understand how any build could achieve Force generation as there are so many VNA determinant / measurable variables that need to be all as close as possible to optimum values.

My soon to arrive 100W Rf amp has the ability to be throttled from 50dBm (100W) output to 19dBm (79mW) output and has a real time analogue VSWR output pin, which I plan to use to generate a DIY VNA as knowing the VSWR, Rf amp impedance and the frequency will enable real time calculation of the return loss db, the reflection coefficient, how well the cavity is impedance matched to the Rf amp, the cavity unloaded Q and Q bandwidth.

http://cgi.www.telestrian.co.uk/cgi-bin/www.telestrian.co.uk/vswr.pl

Seems a lot of information to get from 1 pin of analogue information, but the Rf amp is driving the cavity at it's impedance, output frequency and reporting on what is happening via the VSWR analogue output. Additionally doing it this way ensures each EMDrive is impedance matched as close as possible to it's 100W Rf amp.

Comments from the experienced microwave guys that have used 1 cable VNA measurements most welcome.
Agreed but a 1 port measurement is simply return loss or an antenna measurement not a bandpass cavity where a sample port is needed for S21. call it reflected S11 Q not forward S21 Q. This will give unnaturally high numbers leading to confusion. If u get the chance, ask shawyer for S21 plots since he had a sample port. This should illustrate assymetrical bandpass response.

What are your thoughts on this solution?

http://www.ebay.com/itm/138M-4-4G-SMA-signal-source-generator-simple-spectrum-analyzer-Tracking-source-/111493176997?pt=LH_DefaultDomain_0&hash=item19f582dea5
Add a couple of SMA 3dB attenuators for mismatch.
Add a sampling port to the frustum with the software and a laptop it should do what needs to be done.

(much thanks to someone that has been a wonderful source of information saving me hours of research)

Shell

Offline TheTraveller

I would suggest to every EMDrive builder that having a Vector Network Analyser and understanding what it is telling you is critical to your success.

Doing 1 cable SLOW sweeps can test your excitation antenna design, impedance matching, find your cavity resonances, peak return loss dBs, VSWR, reflection coefficient, Q and cavity bandwidth.

Without a VNA, it is hard to understand how any build could achieve Force generation as there are so many VNA determinant / measurable variables that need to be all as close as possible to optimum values.

My soon to arrive 100W Rf amp has the ability to be throttled from 50dBm (100W) output to 19dBm (79mW) output and has a real time analogue VSWR output pin, which I plan to use to generate a DIY VNA as knowing the VSWR, Rf amp impedance and the frequency will enable real time calculation of the return loss db, the reflection coefficient, how well the cavity is impedance matched to the Rf amp, the cavity unloaded Q and Q bandwidth.

http://cgi.www.telestrian.co.uk/cgi-bin/www.telestrian.co.uk/vswr.pl

Seems a lot of information to get from 1 pin of analogue information, but the Rf amp is driving the cavity at it's impedance, output frequency and reporting on what is happening via the VSWR analogue output. Additionally doing it this way ensures each EMDrive is impedance matched as close as possible to it's 100W Rf amp.

Comments from the experienced microwave guys that have used 1 cable VNA measurements most welcome.
Agreed but a 1 port measurement is simply return loss or an antenna measurement not a bandpass cavity where a sample port is needed for S21. call it reflected S11 Q not forward S21 Q. This will give unnaturally high numbers leading to confusion. If u get the chance, ask shawyer for S21 plots since he had a sample port. This should illustrate assymetrical bandpass response.

What are your thoughts on this solution?

http://www.ebay.com/itm/138M-4-4G-SMA-signal-source-generator-simple-spectrum-analyzer-Tracking-source-/111493176997?pt=LH_DefaultDomain_0&hash=item19f582dea5
Add a couple of SMA 3dB attenuators for mismatch.
Add a sampling port to the frustum with the software and a laptop it should do what needs to be done.

(much thanks to someone that has been a wonderful source of information saving me hours of research)

Shell

I'll let rfmwguy answer your and my questions.

You may find this microwave engineers discussion of S parameters of interest:
http://www.microwaves101.com/encyclopedias/438-s-parameters-microwave-encyclopedia-microwaves101-com
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline WarpTech

  • Full Member
  • ****
  • Posts: 1313
  • Do it!
  • Vista, CA
  • Liked: 1351
  • Likes Given: 1817
@aero et al - regarding longer MEEP runs
...

Thanks. If someone has done it, that means running meep on the cloud is possible. Good to know.
...

I have the opposite request if you would please. I would like to see the reflections of short pulse, 1 or 2 cycles, and nothing after it until it goes dark, starting at the antenna. When the antenna sends out the pulse, I would like to see what happens after the end of the pulse reaches the big end. It may require more time slices I think, but not such a long time.

We can see the reflections from the antenna now, moving toward the back end. What I want to see is what happens when the antenna turns on and shuts off. How does it evolve from there? That will give us the full path of the wave, forward and backward. I have a hunch that it does not persist too long, but it will persist for a while before it dissipates and the whole time the vector dS/dt will still be in the same direction, exerting thrust forward. It will show us the evolution of a decaying evanescent wave, rather than a growing evanescent wave which we have now.

Thanks!
Todd
« Last Edit: 07/12/2015 04:06 PM by WarpTech »

Offline aero

  • Senior Member
  • *****
  • Posts: 2786
  • 92129
  • Liked: 724
  • Likes Given: 249
We've been talking about the 50% duty cycle of magnetrons causing the cavity Pin to switch on ond off every half cycle of the 60 cycle wall plug power.

If anyone is interested, meep does allow the source to be switched off. I could make a slightly different run so we could look at the field images immediately after power-off if that would be of interest. Maybe a normal of 32 cycles with data collection turned on at 32 instead of 30.7 cycles as before, but instead of stopping the run at 32 cycles, just switch the source off at 32 cycles and let the run continue to, say, 33.4 cycles. That would give 14 time slices with the first being the same as the last one of previous runs, but the remainder showing the immediate transient after power off.

If that would provide helpful information, let me know.

Add: Its funny in a way - we cross posted the same thought.
« Last Edit: 07/12/2015 04:14 PM by aero »
Retired, working interesting problems

Offline TheTraveller

...

Sent you a PM about your Flight Thruster 3D files. Please check your private messages. Thanks
« Last Edit: 07/12/2015 04:25 PM by TheTraveller »
"As for me, I am tormented with an everlasting itch for things remote. I love to sail forbidden seas.
Herman Melville, Moby Dick

Offline SeeShells

  • Senior Member
  • *****
  • Posts: 2335
  • Every action there's a reaction we try to grasp.
  • United States
  • Liked: 2982
  • Likes Given: 2601
To hone my test equipment and my building skills I'm going with two first designs. First is the crash dummy low power, made from OTS perforated copper sheet. the second will be after that the one using O2 free copper and a little thicker but still with holes.
https://concordsheetmetal.com/store/perforated-copper/

I might for Myth Busters sake go full power on the first one video and record the results. ;)
Wish I had a Tech here to tell them to push the button!

I am planing to build more than just two basic designs, I have on the drawing board injected microwaveguide phase locked magnetrons that the power supply has been modified to narrow the hash and be able to sweep and phase lock the to cavity and vary the duty cycles.

And that may prove interesting, as it currently is a 50% duty cycle on the standard microwave magnetron seems to be pointing st some form of action in the cavity.
This isn't a simple on the weekend hang from the shower curtain build this is to narrow down the actions that produce the measured thrusts and add that data to the theories of how it works.

Be out in the shop busy today... have a great day all.

Shell
« Last Edit: 07/12/2015 05:46 PM by SeeShells »

Offline Rodal

  • Senior Member
  • *****
  • Posts: 5895
  • USA
  • Liked: 6045
  • Likes Given: 5325
This was discussed back in May:


http://forum.nasaspaceflight.com/index.php?topic=36313.msg1369553#msg1369553

Quote from: Rodal
Paul March has addressed  and explained this as follows: Chinese (Prof. Yang) calculated loaded Q factors are much higher than the Q's reported by Shawyer and by NASA' Eagleworks because of the unorthodox way that the Chinese calculate their loaded Q factors.  Instead of using the S11 zero dB reference plane to measure their -3dB down bandwidths from, as is done elsewhere, the Chinese use the most negative dB S11 value located at the resonance frequency and measure up 3dB toward the S11 zero  dB plane.  Therefore, of course, the bandwidth figures used by the Chinese in this unorthodox calculation are going to be ridiculously small which yields correspondingly artificially large values of the calculated Q-factor. .
Here is where they went wrong...under no industrial RF standard does anyone measure Q on return loss, S11. It is done on S21, forward power in the frequency domain for cavities. I stand by my claim that "Specsmanship" was used to create an unnaturally large Q, either by unfamiliarity or intent.

Note that S21 requires a 2 port measurement, input and output (note the sampling port on the frustums will provide the output). I'd bet a six-pack of craft beer that realistic Qs are in the 4 digit range for both shawyer and yang. And yes Doc, Yang should have used the -3dB points below 0 insertion, not -3dB above best return loss...not RF types IMHO.

rfmwguy, a big thanks from me, and a big applause,  for clarifying this issue: you are 100% right. 

The Q should be measured using two port S21.

All Q measurements using S11 are suspect: everybody should take with a grain of salt the reported Q's from different EM Drive researchers, unless the procedure to measure the Q is detailed and they have used S21.
« Last Edit: 07/12/2015 05:51 PM by Rodal »

Offline mwvp

  • Full Member
  • **
  • Posts: 254
  • Coincidence? I think Not!
  • Liked: 169
  • Likes Given: 30
I would suggest to every EMDrive builder that having a Vector Network Analyser and understanding what it is telling you is critical to your success.
...
a real time analogue VSWR output pin, which I plan to use to generate a DIY VNA as knowing the VSWR, Rf amp impedance and the frequency will enable real time calculation of the return loss db, the reflection coefficient, how well the cavity is impedance matched to the Rf amp, the cavity unloaded Q and Q bandwidth.
...
Comments from the experienced microwave guys that have used 1 cable VNA measurements most welcome.

I've used VNA's to do antenna sweeps. I taught cellular engineers how to do antenna sweeps. The VNA was 2 port that went to a duplexer to make it 1 port.

Anyways, your transmitter VSWR isn't a VNA, it gives you a scalar, not vector (phase information) that is useful for determining Q. I've tuned a lot of stuff, but never needed to measure Q. BW, Ripple, insertion loss, return loss, but never was the Q of anything spec'd for the systems I worked on.

Anyways, doing the VNA thing means applying loads (open, short, 50ohm) and hitting the apt. calibrate buttons. Otherwise, how do you know your coax/waveguide isn't affecting the inferred reading you get from the return loss?

Putting a sample port on the frustrum isn't a bad idea anyways. You can make a very simple diode detector (resistor, cap, diode - costs $1) and see the peak on a $5 DVM.

Although, the ancient and venerable grid dip meter can  give you an indication of Q, but I think you'd have a heck of a time coupling to a closed frustrum. Just don't bother with that!

Best use a very lightly, barely, coupled sample port. There's a lot of power in your 100K+ Q cavity!

Tags: