Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
Dr Rodal, part of the fun of following emdrive experimental outputs is in interpreting contradictions and reading tea leaves
Thanks for answer. I have no time right know to dig into the plots, my interpretation at a glance is that when rfmwguy states "4 cycles of 50% power at about 1.5 minutes each" on a plot subtitled "1 Minute Warmup and Four 50% Power Cycles" he refers to each cycle being the 4 on/off pulses cluster, the first "cluster" (continuous actually) being the "warmup" (hence you see 5 where he sees 1 + 4). This is consistent with, err, a consistent use of horizontal axis numbering (short of consistent unit indication) and also much more consistent with expected periods in torsion oscillations, as there is no longer period to explain that the natural ones in torsion (well, the latest plot does exhibits probably more than one component, maybe coupling to other axis, but still). To me this simpler explanation (awkward phrase to indicate what is a "power cycle" leading to confusion, units in seconds from plot one, no mysteriously long period) prevails until further clarification by experimenter.
It is a boon to the sleuths here that a lot of those (less than ideally documented) pendulum experiments have been done in a (more or less) under-damped mode so that second order time constant is still apparent without sophisticated analysis. It would be much harder to see those interpretation problems with a critically damped or over-damped system. So what is a less than ideal tuning (around the axis of interest the pendulum should be critically damped if possible, not under-damped) ends up being of some help. BTW I wonder if by using a cylindrical shape immersed in oil bath (instead of plates or vanes), sharing axis of torsion wire, it could be possible to have a largely over-damped response for other axis (swings... there are 4 other degrees of freedom on top of the torsion one, much shorter periods probably) while leaving the torsion axis just critically-damped. Maybe that could reduce spurious couplings and allow cleaner signals. Also of interest is the height of center of mass of moving assembly below the attach point.
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
By the way, what is rfmwguy's latest assessment of the quality factor Q in his experiments. When he was here, after spending a lot of time polishing his EM Drive fustrum, my recollection is that he was initially claiming that his Q was over 200,000, which I stated was right down impossible. I could believe 5,000, 10,000 perhaps up to 50,000, but over 200,000 didn't make sense.
Since the quality factor Q is a most important variable in all formulas (Shawyer, McCulloch and Notsosureofit) and he certainly should be able to measure the Q, it would be nice if you could give this information (or ask him if you are a member of his closed group rfdriven.com).
By now he may have had enough runs that he may have a better estimate for the Q in his experiments.
Thanks
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
Here is that same chart with beam displacement in mm and showing a amplified rate of change. Note the very rapid rate of change at the start of every power on period, which slows as the twisting wire adds angular torque load to the accelerating EmDrive and slows the rate of change. Which is exactly what is expected.
It does not help that as the maggie temperature increases, the maggie freq drops, which may be or not in lock with the frustum, so force generation is not constant.
However what is constant in EVERY test Dave has done since the 1st he published, is there is very rapid initial acceleration generated just as soon as the maggie locks to the frustum.
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
Here is that same chart with beam displacement in mm and showing a amplified rate of change. Note the very rapid rate of change at the start of every power on period.
That is called the "transient response" of a dynamic system.
Depending on the magnitude of the damping coefficient, the initial transient may exceed the steady-state response.
It is purely due to the dynamics of the pendulum (mass, rotary inertia, torsional spring constant and damping). Not to something electromagnetic.
For example, it is known that the response to an impulse, of a dynamic system, is twice the magnitude of the static elastic response (the response of a dynamic system when a force is applied at a very low magnitude of acceleration).
This is why frobnicat wrote:
It is a boon to the sleuths here that a lot of those (less than ideally documented) pendulum experiments have been done in a (more or less) under-damped mode so that second order time constant is still apparent without sophisticated analysis.
as the dynamics of the torsional pendulum are clearly seen in the response
This was the benefit of rfmwguy using a measuring instrument with well behaved and well known response characteristics. Rfmwguy is to be applauded for using a torsional pendulum in his measurements
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
By the way, what is rfmwguy's latest assessment of the quality factor Q in his experiments. When he was here, after spending a lot of time polishing his EM Drive fustrum, my recollection is that he was initially claiming that his Q was over 200,000, which I stated was right down impossible. I could believe 5,000, 10,000 perhaps up to 50,000, but over 200,000 didn't make sense.
Since the quality factor Q is a most important variable in all formulas (Shawyer, McCulloch and Notsosureofit) and he certainly should be able to measure the Q, it would be nice if you could give this information (or ask him if you are a member of his closed group rfdriven.com).
By now he may have had enough runs that he may have a better estimate for the Q in his experiments.
Thanks
Dave has stated he believes the loaded Q is at least 10,000, this is from his VNA S11 rtn loss analysis.
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
By the way, what is rfmwguy's latest assessment of the quality factor Q in his experiments. When he was here, after spending a lot of time polishing his EM Drive fustrum, my recollection is that he was initially claiming that his Q was over 200,000, which I stated was right down impossible. I could believe 5,000, 10,000 perhaps up to 50,000, but over 200,000 didn't make sense.
Since the quality factor Q is a most important variable in all formulas (Shawyer, McCulloch and Notsosureofit) and he certainly should be able to measure the Q, it would be nice if you could give this information (or ask him if you are a member of his closed group rfdriven.com).
By now he may have had enough runs that he may have a better estimate for the Q in his experiments.
Thanks
Not a member of rfdriven Dr. Rodal, just a few emails. I'll ask but I'm not sure he is around this 4th. I'll ask.
I did a inverse FFT of just the deflection data and can see a couple harmonics AFTER power off, Power on is very apparent.
http://sooeet.com/math/online-fft-calculator.phpShell
Added a standard FFT.
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
Here is that same chart with beam displacement in mm and showing a amplified rate of change. Note the very rapid rate of change at the start of every power on period.
That is called the "transient response" of a dynamic system.
Depending on the magnitude of the damping coefficient, the initial transient may exceed the steady-state response.
It is purely due to the dynamics of the pendulum (mass, rotary inertia, torsional spring constant and damping). Not to something electromagnetic.
For example, it is known that the response to an impulse, of a dynamic system, is twice the magnitude of the static elastic response (the response of a dynamic system when a force is applied at a very low magnitude of acceleration).
I suggest it is due to the EmDrive being free to accelerate (no initial torque load) and then as torque load increases, the rate of change decreases as would be expected.
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
I'm no expert in OpenOffice Calc, but in Excel, you can clean up the weird x-axis labels by doing a scatter plot. That would make the data a lot easier to read.
Sent from my SM-T710 using Tapatalk
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
By the way, what is rfmwguy's latest assessment of the quality factor Q in his experiments. When he was here, after spending a lot of time polishing his EM Drive fustrum, my recollection is that he was initially claiming that his Q was over 200,000, which I stated was right down impossible. I could believe 5,000, 10,000 perhaps up to 50,000, but over 200,000 didn't make sense.
Since the quality factor Q is a most important variable in all formulas (Shawyer, McCulloch and Notsosureofit) and he certainly should be able to measure the Q, it would be nice if you could give this information (or ask him if you are a member of his closed group rfdriven.com).
By now he may have had enough runs that he may have a better estimate for the Q in his experiments.
Thanks
Not a member of rfdriven Dr. Rodal, just a few emails. I'll ask but I'm not sure he is around this 4th. I'll ask.
I did a inverse FFT of just the deflection data and can see a couple harmonics AFTER power off, Power on is very apparent.
http://sooeet.com/math/online-fft-calculator.php
Yes, now that we have sorted out the time domain, it is clear that the oscillations are due to the torsional period of oscillation of his torsional pendulum. Score that as a plus for using torsional pendulums !
Still, the long term exponential decay seems to me that can only be explained as due to a thermal effect, like heating of the air by the magnetron at the big end.
Clearly one cannot explain that long term exponential decay with electromagnetics from the microwave drive or with Lorentz forces, I think
He should move the magnetron to the side, in line with the axis of the pendulum's beam, instead of being perpendicular to it.
In addition anything he can do to avoid heating the air would be a plus.
Suggestion: can he run the experiment with the magnetron heating the air at the big end but without exciting the EM Drive? For example by installing a copper plate inside the EM Drive right after the magnetron, to close off any excitation of the EM drive, but allow the heating to take place.
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
Here is that same chart with beam displacement in mm and showing a amplified rate of change. Note the very rapid rate of change at the start of every power on period.
That is called the "transient response" of a dynamic system.
Depending on the magnitude of the damping coefficient, the initial transient may exceed the steady-state response.
It is purely due to the dynamics of the pendulum (mass, rotary inertia, torsional spring constant and damping). Not to something electromagnetic.
For example, it is known that the response to an impulse, of a dynamic system, is twice the magnitude of the static elastic response (the response of a dynamic system when a force is applied at a very low magnitude of acceleration).
I suggest it is due to the EmDrive being free to accelerate (no initial torque load) and then as torque load increases, the rate of change decreases as would be expected.
see this, for example:
http://bit.ly/29jIKAo
Look I understand you have not altered your beliefs since you 1st post on this forum and it would appear that nothing anybody can show will cause that to alter. I could be wrong.
As someone who has seen, 1st hand, several EmDrive builds generate thrust, all I can say is you are mistaken in your current beliefs.
What you are seeing in the rate of change plots, is an EmDrive getting freq lock from the maggie and rapidly generating 18mN or so of force and as the initial torque load is zero, it achieves it's highest acceleration. Then as the wire twists, the increasing torque load reduces the rate of change as would be expected until either the maggie drops lock and thrust stops or the torque load matched the accelerative torque and further deflection stops, or the maggie power stops before reaching max torque load and any remaining kinetic / velocity in the system generates further non powered on deflection until finally the twisting wire has absorbed all the energy delivered by the EmDrive to the load.
What I see in this plot from Dave's 1st data share (vertical scale has been converted into mm of deflection) is the EmDrive powered on and maintained an averaged rate of change of 130um / 0.8 sec for 30 seconds. Note Rate Of Change is amplified 100 times so 13mm on the left scale means 130um Rate Of Change / 0.8 sec time interval.
Impulse? Constant for 30 seconds? I think not.
Please note that for the 1st 30 seconds, full maggie power was flowing over the wires and the maggie antenna INSIDE the frustum was heating it with 900Wrf and yet there is no significant movement. Ok there is a slight probably Lorentz kick as the maggie current 1st starts to flow. Sort of kicks the Lorentz and thermal sources are the cause ideas to the curb.
...Look I understand you have not altered your beliefs since you 1st post on this forum and it would apprear that nothing anybody can show will cause that to alter....
For my it never is a question of beliefs but it is always a question of mathematical analysis and experimental facts. Certainly I can change my opinion and will, but based on mathematical analysis and experimental facts, not on beliefs.
For example, I just agreed upon examination of the facts concerning the time scale, that the oscillations are purely due to the time period of the torsional oscillations of the torsional pendulum.
Can you explain the long term exponential decay after power is off ?
------------------
He should run the experiment with the magnetron heating the air at the big end but without exciting the EM Drive. For example by installing a copper plate inside the EM Drive right after the magnetron, to close off any excitation of the EM drive, but allow the heating of the big end air to take place.
Then compare results.
...Look I understand you have not altered your beliefs since you 1st post on this forum and it would apprear that nothing anybody can show will cause that to alter....
For my it never is a question of beliefs but it is always a question of mathematical analysis and experimental facts. Certainly I can change my opinion and will, but based on mathematical analysis and experimental facts, not on beliefs.
Can you explain the long term exponential decay after power is off ?
------------------
He should run the experiment with the magnetron heating the air at the big end but without exciting the EM Drive. For example by installing a copper plate inside the EM Drive right after the magnetron, to close off any excitation of the EM drive, but allow the heating of the big end air to take place.
Then compare results.
Dave, with the help of other experienced EmDrive builders, is working to further understand the dynamics. I believe there may be 2 oscillatory periods in action. One from the tangential thrust pulling the centre of mass off the normal rotation axis and the other as the axis of rotation and the axis of centre of mass through the centre of thrust are not at a right angle. Plus there is twisted wire hanging some vertical and some trailing away, which will cause alterations in the twist up and twist down of the fairly short piano wire section. Will find the tests setup photo and mark it up to shown
What I can tell you is the same rate of change curve is visible in EVERY EmDrive test Dave has done, since the 1st I published here.
Dave said I can share the data from his runs. This is his datasheet with labels. Maybe this can help to clear up the questions in his dataset.
Shell
By the way, what is rfmwguy's latest assessment of the quality factor Q in his experiments. When he was here, after spending a lot of time polishing his EM Drive fustrum, my recollection is that he was initially claiming that his Q was over 200,000, which I stated was right down impossible. I could believe 5,000, 10,000 perhaps up to 50,000, but over 200,000 didn't make sense.
Since the quality factor Q is a most important variable in all formulas (Shawyer, McCulloch and Notsosureofit) and he certainly should be able to measure the Q, it would be nice if you could give this information (or ask him if you are a member of his closed group rfdriven.com).
By now he may have had enough runs that he may have a better estimate for the Q in his experiments.
Thanks
Not a member of rfdriven Dr. Rodal, just a few emails. I'll ask but I'm not sure he is around this 4th. I'll ask.
I did a inverse FFT of just the deflection data and can see a couple harmonics AFTER power off, Power on is very apparent.
http://sooeet.com/math/online-fft-calculator.php
Yes, now that we have sorted out the time domain, it is clear that the oscillations are due to the torsional period of oscillation of his torsional pendulum. Score that as a plus for using torsional pendulums !
Still, the long term exponential decay seems to me that can only be explained as due to a thermal effect, like heating of the air by the magnetron at the big end.
Clearly one cannot explain that long term exponential decay with electromagnetics from the microwave drive or with Lorentz forces, I think
He should move the magnetron to the side, in line with the axis of the pendulum's beam, instead of being perpendicular to it.
In addition anything he can do to avoid heating the air would be a plus.
Suggestion: can he run the experiment with the magnetron heating the air at the big end but without exciting the EM Drive? For example by installing a copper plate inside the EM Drive right after the magnetron, to close off any excitation of the EM drive, but allow the heating to take place.
In the FFT you can clearly see the harmonics of the system, it's been years since I've played around with FFT but some things stay.
The long sloped decay in red you see the torsional wire non-linear response on the 26 pounds of beam and drive weight. In other words as the beam gets closer to center the force it puts on the beam decreases.
Shell
Shell
Look I understand you have not altered your beliefs since you 1st post on this forum and it would appear that nothing anybody can show will cause that to alter. I could be wrong.
...
You have consistently ignored major, blatant contradictions that have been pointed out to you, and you have refused to consider any other perspectives or explanations other than your own (or those you get from Shawyer).
Rodal has stated quite frequently what kind of experiment it would take to convince him there is a useful force here. On the other hand, you have never stated anything that could convince you that this force is not real.
Before accusing someone of refusing to change their minds who has repeatedly asked for a specific type of experimental data, why don't you offer at least as much consideration for their perspective? Or acknowledge that Shawyer's claim of a self accelerating cavity with no external applied force (breaking the definition of conservation of momentum) is incompatible with Shawyer's claim of following existing physics (which is based upon conservation of momentum) before expecting people to trust you as a good judge of whether or not something is an experimental artifact.
He should run the experiment with the magnetron heating the air at the big end but without exciting the EM Drive. For example by installing a copper plate inside the EM Drive right after the magnetron, to close off any excitation of the EM drive, but allow the heating of the big end air to take place.
Then compare results.
Yes, a dummy test with just heating on the magnetron's side would be great for figuring out if heating is the only source of thrust (or just part of it).
What you are seeing in the rate of change plots, is an EmDrive getting freq lock from the maggie and rapidly generating 18mN or so of force and as the initial torque load is zero, it achieves it's highest acceleration. Then as the wire twists, the increasing torque load reduces the rate of change as would be expected until either the maggie drops lock and thrust stops or the torque load matched the accelerative torque and further deflection stops, or the maggie power stops before reaching max torque load and any remaining kinetic / velocity in the system generates further non powered on deflection until finally the twisting wire has absorbed all the energy delivered by the EmDrive to the load.
What I see in this plot from Dave's 1st data share (vertical scale has been converted into mm of deflection) is the EmDrive powered on and maintained an averaged rate of change of 130um / 0.8 sec for 30 seconds. Note Rate Of Change is amplified 100 times so 13mm on the left scale means 130um Rate Of Change / 0.8 sec time interval.
...
Is this theory based on measurements you have taken while Dave was performing his experiments or is it your opinion? The whole purpose of building an em-drive and mounting it on a torque pendulum is to perform enough experiments under different conditions to be able to eliminate noise and error sources. We need to separate real data from the noise. I am still waiting for someone to use just heat as a driving force. Stick a big resistor inside the fustrum and take measurements. Referring to aircraft motion, torque pendulum experiments are instrumented to measure yaw. There is also pitch and roll. These other movements have their own time constants. The oscillations seen in Dave's data may be from pitch or roll. If data was collected where the magnetron was not powered this oscillation may be present. This is another experiment that could be done to isolate contributions to the waveform seen when the magnetron is energized.
...
In the FFT you can clearly see the harmonics of the system, it's been years since I've played around with FFT but some things stay.
The long sloped decay in red you see the torsional wire non-linear response on the 26 pounds of beam and drive weight. In other words as the beam gets closer to center the force it puts on the beam decreases.
Shell
Shell
1) FFT's are symmetric, the plot should be shown with only the positive numbers, thus avoiding negative frequencies.
2) FFT is showing the frequencies in the horizontal axis, FFT deals with the frequency domain, not the time domain.
3) In the FFT the frequency bins are linear. For example, if we have a bin width of 43 Hz (the result of dividing Nyquist frequency by the FFT frame size), then one has bins from 0 Hz to 43 Hz, 43 Hz to 86 Hz, 86 Hz to 129 Hz, and so on. I don't understand why you are discussing nonlinearities in a FFT plot:
The long sloped decay in red you see the torsional wire non-linear response on the 26 pounds of beam and drive weight
4) The FFT is showing the frequency spectrum response with an exponential time decay. What is being shown certainly does not explain the long term exponential decay with time. It is FFT that one would expect from an exponential time decay.
