Hi again. I need help!
Some time ago an explanation of the thrust occurred to me. As I have been working closely with Shell making meep runs, I shared it with her and we are fleshing out a paper but it needs more supporting data (so what else is new?) . I have been running low resolution meep numerical experiments but we agree that better resolution is needed. I now have available a supper duper whiz bang machine on which to run the experiments. It is running Ubuntu 16.04 but it doesn't have meep installed There seems to be a problem between this version of Ubuntu and meep. The reason I think that is because I updated my old machine to Ubuntu 16.04 and now meep doesn't run there either.
Unfortunately I know very little about software beyond how to use it. I need the help of a systems administrator familiar with Linux. Does anyone here want to help? PM me to volunteer
Thanks,
aero
Thank you! I like your response so I spent some hours to debug your code. Yes I found the bug. Now I have updated my review as attached. Good luck!
cos()^2 caused all the issue!
And now, Voilą! we have a simulation that proves conservation of momentum.
Thank you for spending some of your time on this Mr. Li.
Apologies if too far off topic. This is for all who are dealing with magnetrons and their idiosyncracies.
A small microwave oven/food heater based on LDMOS transistors is apparently being introduced. Power is 200 Watts, although it's not clear whether that refers to microwave power delivered to the load or power draw of the total device. Price will be US $199.
If true, this is a very interesting development for any technology that needs appreciable microwave power at low cost. This may make it easier to build microwave power sources that are more stable and controllable than with magnetrons.
Company's site is:
http://www.wayvtech.com/#wayvI have no connection with the company at all, just thought it might be of interest to some here.
Hope nobody minds that I'm detailing out this build, I was asked politely to maybe post a little more. 
Back from testing the test stand torsional wire. It's quite stable and very sensitive even with the thicker wire I'm using for pretesting, although it still wants to exhibit oscillations that will potentially add error to the readings. I think I'm going to extend the hollow tube to the ends of the wire like in my drawing and keep the circular osculations to a minimum. It will still be highly sensitive with the thin piano wire <.008" through the brass end caps with .010 holes and keep the jiggling (scientific term 
) to a minimum.
Shell
Um, the last I heard you were engaged in a tear down and rebuild of a vertical test rig. Would I be correct in presuming that these changes indicate something other than the device testing null on prior rigs?
Hope nobody minds that I'm detailing out this build, I was asked politely to maybe post a little more.
Back from testing the test stand torsional wire. It's quite stable and very sensitive even with the thicker wire I'm using for pretesting, although it still wants to exhibit oscillations that will potentially add error to the readings. I think I'm going to extend the hollow tube to the ends of the wire like in my drawing and keep the circular osculations to a minimum. It will still be highly sensitive with the thin piano wire <.008" through the brass end caps with .010 holes and keep the jiggling (scientific term ) to a minimum.
Shell
Is the purpose of the tubes to stop the beam from rolling? I added a graphic of an airplane to define roll, yaw, etc. If the Center of Mass of the beam is below the attachment point of the wire there will be less tendency for the beam to want to roll or pitch. Any yaw oscillation can be damped with an eddy current device. You would need to have a thick Copper or Aluminum disk mounted to the beam, with the wire going through its center. The magnet could be on an arm and fixed to the frame. There may be interactions with your cables.
To clamp the wire you could use a short length of Aluminum tubing. Just drill a hole through it and tap it for a 4-40 screw. Run the cable through the tube and tighten the screw. When you are happy with the setup, put some Locktite on the threads.
Careful with eddy current damping on magnetron driven devices! As I illustrated in a previous post, just ONE magnetron magnet can result in 30 degree rotation of a torsion pendulum as it seeks orientation with the Earth's magnetic field.
Clamping music wire is a bit tricky. It's a bit like clamping a glass fiber. The slightest nick severely compromises the tensile strength of the wire, which is why I suggested pin vises in an earlier post. Shell's method of using a guitar tuner is excellent, as the wire is wrapped around a large radius bend, not clamped. The downside is losing some control of the wire axis. It needs to be pretty darn vertical (relative to local gravity).
Great questions and good advise.
I'll answer them as best I can.
Careful with eddy current damping on magnetron driven devices!
My magnetron is in the Electronics box on top sealed away from the drive.
Shell's method of using a guitar tuner is excellent
Thanks and it will keep the thin wire under tension without nicking.
Shell
Hope nobody minds that I'm detailing out this build, I was asked politely to maybe post a little more.
Back from testing the test stand torsional wire. It's quite stable and very sensitive even with the thicker wire I'm using for pretesting, although it still wants to exhibit oscillations that will potentially add error to the readings. I think I'm going to extend the hollow tube to the ends of the wire like in my drawing and keep the circular osculations to a minimum. It will still be highly sensitive with the thin piano wire <.008" through the brass end caps with .010 holes and keep the jiggling (scientific term ) to a minimum.
Shell
Um, the last I heard you were engaged in a tear down and rebuild of a vertical test rig. Would I be correct in presuming that these changes indicate something other than the device testing null on prior rigs?
Ummmm, Very correct.
In some cases the data conflicted with itself with 3 different drive designs and it ... no, everyone deserves better and cleaner data. I'll publish later, but I'm going to fine tune, re-run with a better test rig. Use this current rig can test as a torsional pendulum, a teeter todder balance beam and power off beams and power within a fully enclosed beam.
I learned NOT to publish more O.M.G. moments, we deserve better than that.
Shell
Hope nobody minds that I'm detailing out this build, I was asked politely to maybe post a little more.
Back from testing the test stand torsional wire. It's quite stable and very sensitive even with the thicker wire I'm using for pretesting, although it still wants to exhibit oscillations that will potentially add error to the readings. I think I'm going to extend the hollow tube to the ends of the wire like in my drawing and keep the circular osculations to a minimum. It will still be highly sensitive with the thin piano wire <.008" through the brass end caps with .010 holes and keep the jiggling (scientific term ) to a minimum.
Shell
Um, the last I heard you were engaged in a tear down and rebuild of a vertical test rig. Would I be correct in presuming that these changes indicate something other than the device testing null on prior rigs?
Ummmm, Very correct.
In some cases the data conflicted with itself with 3 different drive designs and it ... no, everyone deserves better and cleaner data. I'll publish later, but I'm going to fine tune, re-run with a better test rig. Use this current rig can test as a torsional pendulum, a teeter todder balance beam and power off beams and power within a fully enclosed beam.
I learned NOT to publish more O.M.G. moments, we deserve better than that.
Shell
Shell, the discussion that your OMG post resulted in was valuable. This being a forum for amateurs as well and it is inspiring to see the lengths you are going to get clean data. Since I started reading NSF less than a year ago, the experimental techniques described have been an education it would be hard to get anywhere else.
Sometimes I doubt the wisdom of sharing my incipient theoretical notions here but it is the only place I am likely to find folk willing to disagree. Please keep up the reporting, every post helps me to get my own next build right. Thank you, JMN..
Hope nobody minds that I'm detailing out this build, I was asked politely to maybe post a little more.
Back from testing the test stand torsional wire. It's quite stable and very sensitive even with the thicker wire I'm using for pretesting, although it still wants to exhibit oscillations that will potentially add error to the readings. I think I'm going to extend the hollow tube to the ends of the wire like in my drawing and keep the circular osculations to a minimum. It will still be highly sensitive with the thin piano wire <.008" through the brass end caps with .010 holes and keep the jiggling (scientific term ) to a minimum.
Shell
Um, the last I heard you were engaged in a tear down and rebuild of a vertical test rig. Would I be correct in presuming that these changes indicate something other than the device testing null on prior rigs?
Ummmm, Very correct.
In some cases the data conflicted with itself with 3 different drive designs and it ... no, everyone deserves better and cleaner data. I'll publish later, but I'm going to fine tune, re-run with a better test rig. Use this current rig can test as a torsional pendulum, a teeter todder balance beam and power off beams and power within a fully enclosed beam.
I learned NOT to publish more O.M.G. moments, we deserve better than that.
Shell
Shell, the discussion that your OMG post resulted in was valuable. This being a forum for amateurs as well and it is inspiring to see the lengths you are going to get clean data. Since I started reading NSF less than a year ago, the experimental techniques described have been an education it would be hard to get anywhere else.
Sometimes I doubt the wisdom of sharing my incipient theoretical notions here but it is the only place I am likely to find folk willing to disagree. Please keep up the reporting, every post helps me to get my own next build right. Thank you, JMN..
I love this place, because we are all after the goal of getting to the truth. No name calling like you're a crackpot nut job for even testing the drive or thinking outside of the box with theories.
Thanks for being here.
Shell
I purchased one of these 2.45Ghz 250 watt LDMOS transistors a few months ago. If I get any results from a magnetron, I will begin working on a new RF source that uses these chips.
http://www.nxp.com/products/rf/rf-power-transistors/rf-cooking/2450-mhz-250-w-cw-32-v-rf-ldmos-transistor-for-consumer-and-commercial-cooking:MHT1003N
Good luck with the maggie.
My experience is dealing with high Q frustums can be soul destroying. They can be beasts. My advise as a DIYer is you need to be able to do single freq stepping 1kHz at a time, have forward & reflected output and the ability to control Rf output Watts over a wide range.
From my and another experimenters results:
VNA peak rtn loss freq is not lowest powered VSWR freq is not best thrust freq. VNA rtn loss bandwidth & curve is not the same as thrust bandwidth, which is much narrower.
So to get it right, you need to be in solid control & have good measurement of values that can't happen with a maggie. Sure a maggie can work but only if you are lucky and there is little data other than thrust available.
These new chips may really assist good thrust generation as they really drop the $/watt cost and allow precise freq control. Just need to add programmable attenuator & forward/reflected power monitoring to make a good EmDrive Rf system.
I purchased one of these 2.45Ghz 250 watt LDMOS transistors a few months ago. If I get any results from a magnetron, I will begin working on a new RF source that uses these chips.
http://www.nxp.com/products/rf/rf-power-transistors/rf-cooking/2450-mhz-250-w-cw-32-v-rf-ldmos-transistor-for-consumer-and-commercial-cooking:MHT1003N
Good luck with the maggie.
My experience is dealing with high Q frustums can be soul destroying. They can be beasts. My advise as a DIYer is you need to be able to do single freq stepping 1kHz at a time, have forward & reflected output and the ability to control Rf output Watts over a wide range.
From my and another experimenters results:
VNA peak rtn loss freq is not lowest powered VSWR freq is not best thrust freq. VNA rtn loss bandwidth & curve is not the same as thrust bandwidth, which is much narrower.
So to get it right, you need to be in solid control & have good measurement of values that can't happen with a maggie. Sure a maggie can work but only if you are lucky and there is little data other than thrust available.
These new chips may really assist good thrust generation as they really drop the $/watt cost and allow precise freq control. Just need to add programmable attenuator & forward/reflected power to make a good EmDrive Rf system.
Am I wrong, or would those chip cook themselves in fairly short order if there's any significant amount of reflected power?
I purchased one of these 2.45Ghz 250 watt LDMOS transistors a few months ago. If I get any results from a magnetron, I will begin working on a new RF source that uses these chips.
http://www.nxp.com/products/rf/rf-power-transistors/rf-cooking/2450-mhz-250-w-cw-32-v-rf-ldmos-transistor-for-consumer-and-commercial-cooking:MHT1003N
Good luck with the maggie.
My experience is dealing with high Q frustums can be soul destroying. They can be beasts. My advise as a DIYer is you need to be able to do single freq stepping 1kHz at a time, have forward & reflected output and the ability to control Rf output Watts over a wide range.
From my and another experimenters results:
VNA peak rtn loss freq is not lowest powered VSWR freq is not best thrust freq. VNA rtn loss bandwidth & curve is not the same as thrust bandwidth, which is much narrower.
So to get it right, you need to be in solid control & have good measurement of values that can't happen with a maggie. Sure a maggie can work but only if you are lucky and there is little data other than thrust available.
These new chips may really assist good thrust generation as they really drop the $/watt cost and allow precise freq control. Just need to add programmable attenuator & forward/reflected power to make a good EmDrive Rf system.
Am I wrong, or would those chip cook themselves in fairly short order if there's any significant amount of reflected power?
Data sheet says they can handle VSWR 10:1. Ok hard to believe but that is the stats.
I purchased one of these 2.45Ghz 250 watt LDMOS transistors a few months ago. If I get any results from a magnetron, I will begin working on a new RF source that uses these chips.
http://www.nxp.com/products/rf/rf-power-transistors/rf-cooking/2450-mhz-250-w-cw-32-v-rf-ldmos-transistor-for-consumer-and-commercial-cooking:MHT1003N
Good luck with the maggie.
My experience is dealing with high Q frustums can be soul destroying. They can be beasts. My advise as a DIYer is you need to be able to do single freq stepping 1kHz at a time, have forward & reflected output and the ability to control Rf output Watts over a wide range.
From my and another experimenters results:
VNA peak rtn loss freq is not lowest powered VSWR freq is not best thrust freq. VNA rtn loss bandwidth & curve is not the same as thrust bandwidth, which is much narrower.
So to get it right, you need to be in solid control & have good measurement of values that can't happen with a maggie. Sure a maggie can work but only if you are lucky and there is little data other than thrust available.
These new chips may really assist good thrust generation as they really drop the $/watt cost and allow precise freq control. Just need to add programmable attenuator & forward/reflected power monitoring to make a good EmDrive Rf system.
Phil, for potential reference down the road, can you characterize the thrust bandwidth, its reference to peak RL frequency wise, its effective bandwidth and confirmation that accelerometer is the best way to measure tuning? Thanks in advance...
As you know Dave, I'm not the only experimenter stating that happens, plus Roger has given us his insight on why that happens.
Exciting times, successful EmDrive builders discussing operational parameters and effects with each other and with the Inventor Roger Shawyer. Effect that are only revealed when real thrust generating drives start to be parameterised as we search out the edges of operational characteristics. Of course doing this helps to widen out the margins so to generate more tolerant designs.
Yes, its all new to me since I haven't characterized a thrust bandwidth, but am dealing with a high-Q beast in my new design. Will be interesting to see how much of the mag's splatter extends to resonance. I am also going to watch for the cavity to "help" lock the mag. IOW, will the cavity's resonance perhaps pull or push the mag to oscillate there. The thing is almost a free-running oscillator, so it's load could possibly impact its resonance. I've seen this happen is simple, low-power oscillator circuits...not sure it translates up to high power though.
Maggies can be freq pulled as you saw with your old frustum & spectrum scans. However the freq pulling is done at the expense of powered delivered as the freq pulling introduces phase shift or in effect reflected power back to the maggie.
Freq pushing is done by varying maggie anode current, which is why poorly regulated or rippled DC supply causes freq splatter. With a very tightly regulated and ripple free supply, maggie freq can be electronically adjusted by varying anode current but the anode current change will also vary maggie power.
I purchased one of these 2.45Ghz 250 watt LDMOS transistors a few months ago. If I get any results from a magnetron, I will begin working on a new RF source that uses these chips.
http://www.nxp.com/products/rf/rf-power-transistors/rf-cooking/2450-mhz-250-w-cw-32-v-rf-ldmos-transistor-for-consumer-and-commercial-cooking:MHT1003N
Good luck with the maggie.
My experience is dealing with high Q frustums can be soul destroying. They can be beasts. My advise as a DIYer is you need to be able to do single freq stepping 1kHz at a time, have forward & reflected output and the ability to control Rf output Watts over a wide range.
From my and another experimenters results:
VNA peak rtn loss freq is not lowest powered VSWR freq is not best thrust freq. VNA rtn loss bandwidth & curve is not the same as thrust bandwidth, which is much narrower.
So to get it right, you need to be in solid control & have good measurement of values that can't happen with a maggie. Sure a maggie can work but only if you are lucky and there is little data other than thrust available.
These new chips may really assist good thrust generation as they really drop the $/watt cost and allow precise freq control. Just need to add programmable attenuator & forward/reflected power to make a good EmDrive Rf system.
Am I wrong, or would those chip cook themselves in fairly short order if there's any significant amount of reflected power?
Data sheet says they can handle VSWR 10:1. Ok hard to believe but that is the stats.
VSWR of 10:1 is for sure a huge reflection. The cooling conditions/specification must be satisfied. In regard to the efficiency of "only" 59% good cooling is necessary in general.
Efficiency: 59%
Voltage: 32V
P1dB: 263W
263W/32V=8,21875A for the µW
8,21875A/59*100= 13,93A total current
13,93A - 8,21875A =5,71125A
5,71125A*32V= 182,76W waste heat (in a almost reflection free case)
Smith 3.10 calculation program download link
http://www.fritz.dellsperger.net/smith.html
My experience is dealing with high Q frustums can be soul destroying. They can be beasts. My advise as a DIYer is you need to be able to do single freq stepping 1kHz at a time, have forward & reflected output and the ability to control Rf output Watts over a wide range.
Specs say it operates at 2.4 - 2.5Ghz. I wonder if these chips can freq step if my signal generator can.
My experience is dealing with high Q frustums can be soul destroying. They can be beasts. My advise as a DIYer is you need to be able to do single freq stepping 1kHz at a time, have forward & reflected output and the ability to control Rf output Watts over a wide range.
Specs say it operates at 2.4 - 2.5Ghz. I wonder if these chips can freq step if my signal generator can.
It is an Rf amp on a chip. Quoted bandwidth of 100MHz spanning 2.4 to 2.5GHz. So drive it with any freq inside it's bandwidth and should work.
The output is 250W ~54dBm, with a min gain of 15dB. Suspect there is another chip in the set that gens the freq and provides the ~39dBM (~8W) of required drive.
Apologies if too far off topic. This is for all who are dealing with magnetrons and their idiosyncracies.
A small microwave oven/food heater based on LDMOS transistors is apparently being introduced. Power is 200 Watts, although it's not clear whether that refers to microwave power delivered to the load or power draw of the total device. Price will be US $199.
If true, this is a very interesting development for any technology that needs appreciable microwave power at low cost. This may make it easier to build microwave power sources that are more stable and controllable than with magnetrons.
Company's site is: http://www.wayvtech.com/#wayv
I have no connection with the company at all, just thought it might be of interest to some here.
Mikegem, this could be very handy for quick amateur builds. The circuitry is ready to go with battery integrated, which could minimize the work required to set up a basic experiment. Nice link, thank you, JMN..
This comment/question is for Shell and Dave - but anyone else feel free to chime in.
Dampening. I've noticed that my tortional pendulum is sensitive to me moving around in the room or opening a door. Fortunately my workshop can be completely sealed off. But even so, it takes 15 or 30 minutes for the pendulum to come to a near stop. Entering the room to start the laser and other equipment starts the pendulum oscillating again.
The way I see it I have two options. I can dampen the pendulum with oil, but that adds complexity and will decrease sensitivity. Or I could rig up webcams, wireless switches and extension cords for the laser and other equipment - and perform the experiment from another room in the house.
Any thoughts on which way you plan on going?
Monomorph, the advantage of oil damping is that, that is all it does. It is a tried and tested method. The added complexity of remote control devices should maybe not be added unless necessary. Plexiglass draught exclusion and faraday cage sound to me like essentials.
) to a minimum.