-
#1800 by rfmwguy on 16 Jan, 2016 22:20
-
Doc, its only a gut feel and I have no evidence, but the poor Q may be an indication of a mechanical discontinuity...iow, a large physical chunk of the cavity was malformed. Occams razor leads me to believe the wr340 had no coupling port...it could have been wide open, therefore grossly overcoupled. The 200 degree C magnetron tells me a big mismatch occured...such as a waveguide opening directly into a cavity. Since the assembly is small, not too many choices where this could happen.
-
#1801 by RFPlumber on 16 Jan, 2016 22:33
-
…
This design addresses the shortcomings of the prior test in phasing the two incoming RF signals. I talked about this method very lightly about 6 months ago using two phase locked magnetrons, but I found way too many design hurdles to overcome using the direct injection method. One was the massive amount of excess heat from the magnetrons that would be strapped onto the cavity, swamping any measurements of thrust.
If the burning probe matchstick test failed because of design limitations (which it kind of did, got a large force displacement on the digital scales before it fried itself) I had plan number 2. And I'll present it here for the first time and invite comments.
Getting together pieces and parts to build, I've had to go surplus because it's not cheap. From the undocumented power on test results of the first one I know it's the way I need to go to assure a stable clean high power high Q mode control into my drive.
…
PS: RfPlumber this is why I hadn't answered your last question, simply I've gone beyond it and I'd welcome any inputs you or anyone may have.
I wish my alias would have been “RFProfessor” or at least “RFDoctor”… being what it is though here’s some feedback (I am not suggesting all these points need to be addressed, just something to think about):
1. This will likely cost on the order of $10K+ to build if using new parts.
2. I am not sure you need 2 separate phase shifters (assuming the splitter gives you the initial 180, I don’t know if it is the case though). Otherwise you do.
3. You still aren’t measuring no RF power anywhere… So you will be turning the knobs on those phase shifters looking for what exactly? For thrust?
4. I am assuming this will use an off-the-shelf magnetron to waveguide launcher?
5. What is the Q factor of your frustum? Sorry, I am repeating this, but it looks like magnetron frequency drift is on the order of a few MHz over the first few minutes of operation. Hence anything with Q over a couple thousand will be out of resonance very shortly.
Google for “magnetron phase stability”. The very first link:
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwib6e-ulK_KAhUOyGMKHQagCxAQFggiMAA&url=https%3A%2F%2Fwww.ll.mit.edu%2Fmission%2Faviation%2Fpublications%2Fpublication-files%2Fatc-reports%2FLabitt_1977_ATC-74_WW-15318.pdf&usg=AFQjCNFpB3-SOv1pT9PHkLY5WYdUSS-Tvg
So it may not be the thermal frustum expansion which ruins the resonance mode, it may be the magnetron freq drift (or both). Without measuring anything how do you know it is out of resonance? (No thrust? )
6. Not immediately related to this RF plan, but why are you only worried about (and trying to prevent) frustum controlled thermal expansion in one direction? Its resonance frequency depends not just on its central length, but on both end diameters as well… So it will still be changing.
7. What power supply are you using for your magnetron? Is it really a true 1 kW+ 4KV DC monster?
Overall, my gut feel is that this is a bit of an overkill, but it will most likely achieve the task. What is the most expensive way of inducing a TE mode? This could be a winner.
Other thoughts:
Assuming there is indeed something special about magnetron in regard to producing new effects, it then seems that the more one is trying to bring magnetron & feed to that of the perfect pure RF source + ideal coupler, the less is the chance of seeing any new effects. If the effect is indeed specific to magnetron RF output, when the important differences to keep in mind appear to be these:
1. Frequency is not stable (A few kHz difference over a microsecond timeframe).
2. Frequency drift at start-up (A few MHz over the first few minutes).
3. Pulsed (~10-30 ms?) RF output if using microwave oven power supply. (The built-in 1 uF capacitor is not enough to deliver full DC power over the half-cycle of 60 Hz).
At this time only EW knows (or at least has a theory about) which of these are required to produce thrust. Unfortunately they are not sharing.
-
#1802 by Rodal on 16 Jan, 2016 23:24
-
All the details (including r1, r2 and theta, assuming Tajmar measured your definition of height) are in this post: https://forum.nasaspaceflight.com/index.php?topic=39214.msg1477307#msg1477307Dr. Rodal - Please check your axial length numbers posted above. They are the same. I suspect you copied/pasted the same data in both places intending one set to be axial length and the other to be lateral or slant length.
As I use axial height, it really is equal to the slant height times the cosine of the cavity half angle, that is:
AXhi = (r2 - r1 ) * cos (θ ) as shown in the poorly annotated image attached.
Here, (r2 - r1 ) is the slant height of the frustum.
Tell me the value of r1 and r2 and knowing the values of the radii, I can work with that.
To make the long story short, I think that it is most likely that what they measured are the external dimension of (r2 - r1) and that's what Tajmar calls height, because what you call height cannot be readily and accurately measured (it is much easier to measure external r2 - r1) .
Therefore, I think that these are the dimensions:
bigR = 0.0541 meter;
smallR = 0.0385 meter;
axialLength = 2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
Giving the following r1, r2 and theta:
halfAngleConeRadians = ArcSin[(bigR - smallR)/axialLength];
halfAngleConeDegrees = (180/Pi)*halfAngleConeRadians
= 6.82876 degrees
r1 = axialLength/((bigR/smallR) - 1)
= 0.323795 meter
r2 = axialLength/(1 - (smallR/bigR))
= 0.454995 meter
and resulting in the followng:
EXACT SOLUTION
first natural frequency (mode shape TM010 (*)) = 2.47432 GHz
second natural frequency (mode shape TM011 (*)) = 2.98405 GHz
(*) not really TM010 because the mode is not constant in the longitudinal direction, but it is the truncated cone analog mode of TM010 in a cylinder -
#1803 by Rodal on 16 Jan, 2016 23:26
-
Doc, its only a gut feel and I have no evidence, but the poor Q may be an indication of a mechanical discontinuity...iow, a large physical chunk of the cavity was malformed. Occams razor leads me to believe the wr340 had no coupling port...it could have been wide open, therefore grossly overcoupled. The 200 degree C magnetron tells me a big mismatch occured...such as a waveguide opening directly into a cavity. Since the assembly is small, not too many choices where this could happen.
We agree !
That's why I said they should have used an iris, which has been standard convention since WWII.
It looks like they just connected the waveguide to the cavity with a large opening
Hard to believe they would not use an iris, but that's what it looks like ...
Grossly overcoupled ... grossly done -
#1804 by aero on 16 Jan, 2016 23:34
-
@SeeShells - Set antenna length to 0.030 meters (1/4 wl) -- done
@Dr. Rodal - I can work with that, Spherical ends coming up, just maybe not today.
@Dr. Roda; - Is an "iris" the same thing as SeeShells refers to as a "z-choke?"
aero -
#1805 by Rodal on 16 Jan, 2016 23:41
-
@SeeShells - Set antenna length to 0.030 meters (1/4 wl) -- done
@Dr. Rodal - I can work with that, Spherical ends coming up, just maybe not today.
@Dr. Roda; - Is an "iris" the same thing as SeeShells refers to as a "z-choke?"
aero
This is a picture of a choke connection from Wikipedia:
article: https://en.wikipedia.org/wiki/Waveguide_flange#Choke_connection
An iris is a choke that can be made bigger or smaller at will to control the amount of coupling (like the iris of our eyes or the iris of a photo camera to acurately control the amount of light coming in):
Original paper by Nobel Prize winner Hans Bethe on coupling (1944): http://server.physics.miami.edu/~curtright/Diffraction/Bethe1944.pdf
(After receiving security clearance in December 1941, Bethe joined the MIT Radiation Laboratory, where he invented the Bethe-hole directional coupler, which is used in microwave waveguides such as those used in radar sets.)
Yet another (of several) Nobel Prize winners that worked on microwave guides and resonant cavities during WWII.
The coupling strength can be changed by changing the size of the coupling (at will with an iris) or by changing the longitudinal location of the waveguide with respect to the cavity, or by changing the location of the terminating short of the waveguide itself.
I attach below a great slide from David Alesini (LNF, INFN, Frascati, Italy)
-
#1806 by aero on 17 Jan, 2016 00:01
-
@SeeShells - Set antenna length to 0.030 meters (1/4 wl) -- done
@Dr. Rodal - I can work with that, Spherical ends coming up, just maybe not today.
@Dr. Roda; - Is an "iris" the same thing as SeeShells refers to as a "z-choke?"
aero
This is a picture of a choke connection from Wikipedia:
article: https://en.wikipedia.org/wiki/Waveguide_flange#Choke_connection
An iris is a choke that can be made bigger or smaller at will to control the amount of coupling (like the iris of our eyes or the iris of a photo camera to acurately control the amount of light coming in):
The coupling strength can be changed by changing the size of the coupling iris or by changing the longitudinal location of the waveguide with respect to the cavity, or by changing the location of the terminating short of the waveguide itself.
Ok - Then for my purpose in meep, an iris is a circular choke with it's radius being a parameter to set/adjust. -
#1807 by Rodal on 17 Jan, 2016 00:04
-
,,,,
It does not need to be circular
Ok - Then for my purpose in meep, an iris is a circular choke with it's radius being a parameter to set/adjust.
Take a look at cat's eyes !
An iris is a choke (with any geometrical shape, including non-circular) that you can control and adjust to larger or smaller opening area
-
#1808 by aero on 17 Jan, 2016 00:40
-
Here is the cavity with the doubled slant height. It no longer looks like the photo of the prototype. Note that the inside height of the WR 340 waveguide section is correctly scaled to documented dimensions from multiple sources.
Although after some scrutiny, neither does my original cavity ...
-
#1809 by Rodal on 17 Jan, 2016 00:43
-
...I wish my alias would have been “RFProfessor” or at least “RFDoctor”…...
1) Due to your great contributions to this forum, you have been awarded, by acclamation, the title of RFProfessor (Honoris Causa) ages ago
2) Your choice of your alias "Plumber" is shared with very illustrious company: famous Physicist Leonard Susskind began working as a plumber at the age of 16, taking over from his father who had become ill:
https://en.wikipedia.org/wiki/Leonard_Susskind
The man who proved Steven Hawking wrong: Leonard Susskind -
#1810 by SeeShells on 17 Jan, 2016 00:44
-
…
This design addresses the shortcomings of the prior test in phasing the two incoming RF signals. I talked about this method very lightly about 6 months ago using two phase locked magnetrons, but I found way too many design hurdles to overcome using the direct injection method. One was the massive amount of excess heat from the magnetrons that would be strapped onto the cavity, swamping any measurements of thrust.
If the burning probe matchstick test failed because of design limitations (which it kind of did, got a large force displacement on the digital scales before it fried itself) I had plan number 2. And I'll present it here for the first time and invite comments.
Getting together pieces and parts to build, I've had to go surplus because it's not cheap. From the undocumented power on test results of the first one I know it's the way I need to go to assure a stable clean high power high Q mode control into my drive.
…
PS: RfPlumber this is why I hadn't answered your last question, simply I've gone beyond it and I'd welcome any inputs you or anyone may have.
I wish my alias would have been “RFProfessor” or at least “RFDoctor”… being what it is though here’s some feedback (I am not suggesting all these points need to be addressed, just something to think about):
I very much want to thank you for taking the time in replying back, your feedback is very welcome. I know my way around somethings and can read and watch youboob to learn but there is nothing like hearing some advise from someone who has their hands in this field.
1. This will likely cost on the order of $10K+ to build if using new parts.
Saying I got sticker shock was an understatement when I was getting some price quotes, you're not far off. Being in the semiconductor fabs field for years I still have some good used equipment dealers I know and we are actively looking for the hardware I need for this. Also hoping for some contributions from a few companies.
2. I am not sure you need 2 separate phase shifters (assuming the splitter gives you the initial 180, I don’t know if it is the case though). Otherwise you do.
For now I think I'll design for 2 phase shifters.
3. You still aren’t measuring no RF power anywhere… So you will be turning the knobs on those phase shifters looking for what exactly? For thrust?
, no will not be turning knobs looking for WWVB AM. I have 2 spectrum analyzers and a VNA that I plan on using to monitor the frustum, SOP. I'll be using the VNA to sweep the cavity starting from the frustum and working back to where the magnetron enters to set the phases and the small endplate Fo for the mode which is adjustable.
4. I am assuming this will use an off-the-shelf magnetron to waveguide launcher?
Yes, I'll not build this one like in the first test, I will use a standard WR340 magnetron>waveguide launcher.
On the OTS magnetron, it's no different than the magnetrons that are used by John Gerling of Gerling Applied Engineering, Inc for their designs. I had a very nice chat with him asking him about magnetrons and the Inverters and explained what I had done to correct the splattering and drifting of the output and his answer to me was he could do no better and the corrections I applied to the inverter would provide a clean and stable output.
5. What is the Q factor of your frustum? Sorry, I am repeating this, but it looks like magnetron frequency drift is on the order of a few MHz over the first few minutes of operation. Hence anything with Q over a couple thousand will be out of resonance very shortly.
Google for “magnetron phase stability”. The very first link:
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwib6e-ulK_KAhUOyGMKHQagCxAQFggiMAA&url=https%3A%2F%2Fwww.ll.mit.edu%2Fmission%2Faviation%2Fpublications%2Fpublication-files%2Fatc-reports%2FLabitt_1977_ATC-74_WW-15318.pdf&usg=AFQjCNFpB3-SOv1pT9PHkLY5WYdUSS-Tvg
So it may not be the thermal frustum expansion which ruins the resonance mode, it may be the magnetron freq drift (or both). Without measuring anything how do you know it is out of resonance? (No thrust? )
A probe is needed into the waveguide into the frustum to be able to monitor the VWSR.
You are correct it's the first 4 seconds of power on that the magnetron sputters before stabilizing. Much of the splattering is caused by the heater in the magnetron and I have a high voltage vacuum tube switch on a delay that shuts off the heater after 4 seconds.
The other issue in the magnetron is the OTS power supplies that are designed to have a 50% duty cycle pulsing the magnetron, the modified inverter does not.
Q is going to be highly dependent of the ability to tune the frustum and with the micrometer I can inch it up and down ~15um. Some figures areo has run hit 80-100k, I know I'll never reach those levels but I'll be happy with 5k.
6. Not immediately related to this RF plan, but why are you only worried about (and trying to prevent) frustum controlled thermal expansion in one direction? Its resonance frequency depends not just on its central length, but on both end diameters as well… So it will still be changing.
I couldn't agree more that the endplates will want to heat up and expand. This is the reason that the endplates are bonded onto ceramic plates that will restrict any thermal deformations, the side walls will want to expand in the long direction which is ok as long as it can slide past the top small endplate which has a beryllium gasket sliding on the side wall of the tune chamber. The very slight change in the angle of the fustrum walls in this expansion will not be a issue.
The first design consideration was the heating of the frustum and what factors will cause the frustum to deform or grow out of resonance.
Endplate deformation, addressed by ceramic plates
Emdplate length changes, capturing both endplates with a quartz tuning rod running through the center of the frustum
Sidewalls, not a issue as long as they are free to expand and not deform.
7. What power supply are you using for your magnetron? Is it really a true 1 kW+ 4KV DC monster?
Inverter modified from Panasonic. Free running it will peak over 6KV and it's a scary & be wary of thing. 50 years of poking around electrons and I've only got bitten a little just once, I'm very careful around this.
Overall, my gut feel is that this is a bit of an overkill, but it will most likely achieve the task. What is the most expensive way of inducing a TE mode? This could be a winner.
My plans are to increase the power levels and compensate for the thermal deformations in the cavity and this is the only way I could thing of. The task is to drive the levels out of the noise.
Other thoughts:
Assuming there is indeed something special about magnetron in regard to producing new effects, it then seems that the more one is trying to bring magnetron & feed to that of the perfect pure RF source + ideal coupler, the less is the chance of seeing any new effects. If the effect is indeed specific to magnetron RF output, when the important differences to keep in mind appear to be these:
1. Frequency is not stable (A few kHz difference over a microsecond timeframe).
2. Frequency drift at start-up (A few MHz over the first few minutes).
3. Pulsed (~10-30 ms?) RF output if using microwave oven power supply. (The built-in 1 uF capacitor is not enough to deliver full DC power over the half-cycle of 60 Hz).
At this time only EW knows (or at least has a theory about) which of these are required to produce thrust. Unfortunately they are not sharing.
EW=
But I understand the why of them not sharing,
1. The BW of the magnetron if it becomes a critical issue can be "made" to be wider and a little dirtier.
2. I have been looking at a temperature controlled cooler much like the ones on the CPUs to stabilize the magnetron. There are some stand alone and with some thermal compound and a little copper piping and small additional radiator you could have a very cheap and effective magnetron cooler.
http://www.ebay.com/itm/TWC-External-CPU-Water-Cooling-Kit-/281052857871?hash=item41700e460f:m:mv1M0gfZWyf9JuTlh3izlEQ
Wow, thanks for the feedback it is valued!
Shell
-
#1811 by Rodal on 17 Jan, 2016 00:55
-
Here is the cavity with the doubled slant height. It no longer looks like the photo of the prototype. Note that the inside height of the WR 340 waveguide section is correctly scaled to documented dimensions from multiple sources.
Although after some scrutiny, neither does my original cavity ...
See what happens if you don't use the factor of 2 here in this post: https://forum.nasaspaceflight.com/index.php?topic=37642.msg1410336#msg1410336
-
#1812 by Rodal on 17 Jan, 2016 01:06
-
And see https://forum.nasaspaceflight.com/index.php?topic=37642.msg1410193#msg1410193
for alternative dimensions giving 2.45 GHz
Assumed dimensions:
Big diameter = 0.1062 m = (2*0.0541m - 0.002 m)
Small diameter = 0.075 m = (2*0.0385 m - 0.002 m)
Axial Length = 0.100842 m = 0.735*2*0.0686 m
I subtracted 2 mm for copper thickness from the external dimensions, however this has a negligible influence on the results
The axial internal length is 73.5% of the exterior length(it is adjusted internally with a screw prior to testing)
Natural frequency = 2.446 GHz
The COMSOL FEA looks more like these dimensions, or your old ones
The photograph looks different, the EM Drive looks taller
-
#1813 by rfmwguy on 17 Jan, 2016 01:41
-
Good find doc. in hindsight, guess the standard mag antenna injection into the frustum is the simplest way to guarantee critical coupling to a cavity. That's how they were designed...to provide best match into a cavity and guaranteed longer tube life. Lots of old microwaves still work and think the legacy designs are well tested.
shells design is admirable but higher complexity will require specialized tweaking to obtain proper coupling...not too much, not too little. She can do this for sure.
Things I liked about her experiment is mag thermal isolation away from frustum, no potential lorentz force along the moment arm (dc wires) AND she reported higher equivalent micronewton force on the scale. That is a very good sign that something else is there. Her experiment minimizes lorentz and rf source thermals...the occams razor of beam displacement.
http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=39004.0;attach=1094495 -
#1814 by aero on 17 Jan, 2016 01:52
-
I snagged this image from the reference:
https://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_maschinenwesen/ilr/rfs/forschung/folder.2007-08-21.5231434330/ag_raumfahrtantriebe/JPC%20-%20Direct%20Thrust%20Measurements%20of%20an%20EM%20Drive%20and%20Evaluation%20of%20Possible%20Side-Effects.pdf
and counted pixels using a tool new to me, so the count was not very repeatable, but these are the numbers:
Small diameter = 134 px
big diameter = 184 - 187 px
slant height = 142 - 147 px
Using small diameter, Ds = 77 mm and calculating by ratio, gives big diameter ~ 105.7 mm to 107.5 mm and slant height ~ 81.6 mm to 84.5 mm
I don't know what that proves, but as a sanity check it doesn't seem all that close to anyone's numbers.
-
#1815 by Rodal on 17 Jan, 2016 02:03
-
I snagged this image from the reference:
Please take into account that the internal height is adjusted internally with a screw prior to testing, so the internal height at resonance can be much smaller than the actual exterior height of the cavity
https://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_maschinenwesen/ilr/rfs/forschung/folder.2007-08-21.5231434330/ag_raumfahrtantriebe/JPC%20-%20Direct%20Thrust%20Measurements%20of%20an%20EM%20Drive%20and%20Evaluation%20of%20Possible%20Side-Effects.pdf
and counted pixels using a tool new to me, so the count was not very repeatable, but these are the numbers:
Small diameter = 134 px
big diameter = 184 - 187 px
slant height = 142 - 147 px
Using small diameter, Ds = 77 mm and calculating by ratio, gives big diameter ~ 105.7 mm to 107.5 mm and slant height ~ 81.6 mm to 84.5 mm
I don't know what that proves, but as a sanity check it doesn't seem all that close to anyone's numbers.
-
#1816 by SeeShells on 17 Jan, 2016 02:39
-
Good find doc. in hindsight, guess the standard mag antenna injection into the frustum is the simplest way to guarantee critical coupling to a cavity. That's how they were designed...to provide best match into a cavity and guaranteed longer tube life. Lots of old microwaves still work and think the legacy designs are well tested.
You guys are so very good. I couldn't have done anything without all of your help. Thank you!
shells design is admirable but higher complexity will require specialized tweaking to obtain proper coupling...not too much, not too little. She can do this for sure.
Things I liked about her experiment is mag thermal isolation away from frustum, no potential lorentz force along the moment arm (dc wires) AND she reported higher equivalent micronewton force on the scale. That is a very good sign that something else is there. Her experiment minimizes lorentz and rf source thermals...the occams razor of beam displacement.
http://forum.nasaspaceflight.com/index.php?action=dlattach;topic=39004.0;attach=1094495
I thought I'd spring the no Lorentz forces a little later on. When I had this thought on doing direct waveguide injection the problems from the Lorentz force wasn't even on my mind. It was just when the paper came out on the Lorentz forces I went oh my, that will work.
Shell -
#1817 by aero on 17 Jan, 2016 02:50
-
I snagged this image from the reference:
Please take into account that the internal height is adjusted internally with a screw prior to testing, so the internal height at resonance can be much smaller than the actual exterior height of the cavity
https://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_maschinenwesen/ilr/rfs/forschung/folder.2007-08-21.5231434330/ag_raumfahrtantriebe/JPC%20-%20Direct%20Thrust%20Measurements%20of%20an%20EM%20Drive%20and%20Evaluation%20of%20Possible%20Side-Effects.pdf
and counted pixels using a tool new to me, so the count was not very repeatable, but these are the numbers:
Small diameter = 134 px
big diameter = 184 - 187 px
slant height = 142 - 147 px
Using small diameter, Ds = 77 mm and calculating by ratio, gives big diameter ~ 105.7 mm to 107.5 mm and slant height ~ 81.6 mm to 84.5 mm
I don't know what that proves, but as a sanity check it doesn't seem all that close to anyone's numbers.
Yes, consider that it can be shorter. But it can't be longer which kind of rules out axialLength = 2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm) -
#1818 by Rodal on 17 Jan, 2016 03:00
-
I snagged this image from the reference:
Please take into account that the internal height is adjusted internally with a screw prior to testing, so the internal height at resonance can be much smaller than the actual exterior height of the cavity
https://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_maschinenwesen/ilr/rfs/forschung/folder.2007-08-21.5231434330/ag_raumfahrtantriebe/JPC%20-%20Direct%20Thrust%20Measurements%20of%20an%20EM%20Drive%20and%20Evaluation%20of%20Possible%20Side-Effects.pdf
and counted pixels using a tool new to me, so the count was not very repeatable, but these are the numbers:
Small diameter = 134 px
big diameter = 184 - 187 px
slant height = 142 - 147 px
Using small diameter, Ds = 77 mm and calculating by ratio, gives big diameter ~ 105.7 mm to 107.5 mm and slant height ~ 81.6 mm to 84.5 mm
I don't know what that proves, but as a sanity check it doesn't seem all that close to anyone's numbers.
Yes, consider that it can be shorter. But it can't be longer which kind of rules out axialLength = 2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
Which means that
The dimensions I calculated a long time ago here: https://forum.nasaspaceflight.com/index.php?topic=39004.msg1477474#msg1477474
Axial Length = 0.100842 m = 0.735*2*0.0686 m
are very much in play ! -
#1819 by aero on 17 Jan, 2016 04:29
-
I snagged this image from the reference:
Please take into account that the internal height is adjusted internally with a screw prior to testing, so the internal height at resonance can be much smaller than the actual exterior height of the cavity
https://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_maschinenwesen/ilr/rfs/forschung/folder.2007-08-21.5231434330/ag_raumfahrtantriebe/JPC%20-%20Direct%20Thrust%20Measurements%20of%20an%20EM%20Drive%20and%20Evaluation%20of%20Possible%20Side-Effects.pdf
and counted pixels using a tool new to me, so the count was not very repeatable, but these are the numbers:
Small diameter = 134 px
big diameter = 184 - 187 px
slant height = 142 - 147 px
Using small diameter, Ds = 77 mm and calculating by ratio, gives big diameter ~ 105.7 mm to 107.5 mm and slant height ~ 81.6 mm to 84.5 mm
I don't know what that proves, but as a sanity check it doesn't seem all that close to anyone's numbers.
Yes, consider that it can be shorter. But it can't be longer which kind of rules out axialLength = 2*(0.0686 - 0.003) meter (since the wall thickness is 3 mm)
Which means that
The dimensions I calculated a long time ago here: https://forum.nasaspaceflight.com/index.php?topic=39004.msg1477474#msg1477474
Axial Length = 0.100842 m = 0.735*2*0.0686 m
are very much in play !
Welll... The next thing I did was to snag the COMSOL image from the reference, attached. It seems to me that this drawing should be to scale and should show the interior dimensions. Using the pixel measuring method, I find:
px px ref. 77 77
sd 244 241.9 77 77 ave
bd 341.1 339.2 107.6422131148 107.0426229508 107.3424180328
shi 211.4 206.8 66.712295082 65.2606557377 65.9864754098
shi 212.5 208.1 67.0594262295 65.6709016393 66.3651639344
66.1758196721
That is, with small diameter = 0.77 mm, by ratio the big diameter = 107.34 mm and the height = 66.176 mm. But at least that is close to one of the dimensions given previously.
, no will not be turning knobs looking for WWVB AM. I have 2 spectrum analyzers and a VNA that I plan on using to monitor the frustum, SOP. I'll be using the VNA to sweep the cavity starting from the frustum and working back to where the magnetron enters to set the phases and the small endplate Fo for the mode which is adjustable. 