-
#1080 by Rodal on 09 Apr, 2016 15:47
-
NSF-1701A Update - Suggestion Needed
OK, been looking into a Cavendish Torsion conversion on my teeter-totter test stand. Pretty simple actually. One big question:
Q: How would you suggest to calibrate in horizontal displacement?
Teeter-totter was easy enough, place a calibrated weight direct above 4 point harness attachment to beam, directly above emdrive.
Thanks,
Dave
How about using a wire with well-specified material and uniform diameter, hence well specified torsional stiffness?
for example:
http://www.tribology-abc.com/calculators/properties_of_common_spring_materials.pdf
then:
http://mathworld.wolfram.com/TorsionalRigidity.html
θ is the angle of twist in radians
T is the applied torque
L is the wire length between supports
J (or "K" as used by Wolfram) is the torsional constant
G is the torsional or shear modulus of the wire
where D is the diameter of the wire
You have to calculate your setup ahead of time anyway, to chose the desired wire diameter and length, anyway, before attempting any calibration. If the wire has a uniform diameter and well specified material, its torsional stiffness should be well-defined, since the above formula is well verified.
This is one of the advantages of the torsional pendulum: that the torsional stiffness is well specified.
To calibrate (because you want to verify that the diameter is indeed constant, etc.) you can use the period oscillation of the torsional pendulum when you hang an object with known moment of inertia:
http://vlab.amrita.edu/?sub=1&brch=280&sim=1518&cnt=1
http://scienceworld.wolfram.com/physics/TorsionalPendulum.html
http://www.colorado.edu/physics/phys1140/phys1140_sp05/Experiments/M4Fall04.pdf -
#1081 by rfmwguy on 09 Apr, 2016 16:50
-
Thanks. This was one of the things that held me back from a torsion setup, no direct weight calibration in the X axis, although it seems to be able to be calculated with reasonable accuracy.
Can see this being superior against thermal pluming, but direct weight calibration is not practical with all the push-pull force gauges I looked at...not enough resolution.
-
#1082 by Rodal on 09 Apr, 2016 16:53
-
Thanks. This was one of the things that held me back from a torsion setup, no direct weight calibration in the X axis, although it seems to be able to be calculated with reasonable accuracy.
You can calibrate (as described in the links above) using the period oscillation of the torsional pendulum when you hang an object with known moment of inertia. That is just as good as calibrating something using a known mass. If you don't have a good clock, perform the calibration (with little or no damping, to maximize the number of oscillations) using several periods of oscillation and calculate the average
Can see this being superior against thermal pluming, but direct weight calibration is not practical with all the push-pull force gauges I looked at...not enough resolution.
Here is a recent article using a torsional pendulum to measure nanonNewton forces !
http://arxiv.org/pdf/0806.3300v1.pdf
Nanonewton force generation and detection based on a sensitive torsion pendulum
Sheng-Jui Chen and Sheau-Shi PanQuoteIn this paper, we introduce the experiment based on a
sensitive torsion pendulum for measuring and calibrating
small forces at nanonewton scale. The force standard for
calibration is the universal gravitation between four masses
separated by known distances. It is realized by two test
masses suspended as the part of torsion pendulum and
two source masses on a rotation table. Two force generation
mechanisms, optical force from radiation pressure
and electrostatic force by capacitive actuation unit, are designed
and will be calibrated by the gravitation force. We
present our recent results of radiation pressure measurements,
and describe the design of capacitive displacement
sensing/actuating unit.
Notice: they use the gravitational force as a calibration -
#1083 by Rodal on 09 Apr, 2016 19:21
-
It will be interesting if you do it both ways: with a teeter-totter and with a torque pendulum. Prof. Tajmar conducted experiments with a teeter-totter and a torque pendulum but got different answers. He reported the experiments in different media (one at ambient pressure and the other one in a vacuum chamber) with different instruments. He should also have reported measurement with both instruments both in ambient conditions, for better reference and for thoroughness.
Perhaps you will be the first one to report measurements with both instruments under same ambient conditions
-
#1084 by Tellmeagain on 09 Apr, 2016 23:32
-
rfmwguy,
I think you are making a good move. Thank you!
How do you calibrate your torsion balance? First of all, the calibration is not necessarily accurate at all. An approximate calibration is good enough. Why? Because if you obtain an force of (1+-0.1%)*3mN/kW，it is as good as an force of (1+-0.1%)*10mN/kW, because they are both good to 3-sigma (edit: may not be 3-sigma, but you can get the idea). So what's important is your angular displacement measurement; Make sure they are accurate and precise. The calibration factor, either 3mN/kW or 10mN/kW，does not make a big difference. We adopted this principle in the Lorentz force paper. We reached 5-sigma, but we calibrated with a piece of 1cmx2cm paper leaning on the vertical wall of the balance.
You can adopt our method (not accurate at all), or EW's method (see their paper), or Dr. Rodal's method, or Emmett Brown's method that showed 2mN effect (a few days ago), or RFPlumber's method. There should be many more... -
#1085 by otlski on 09 Apr, 2016 23:52
-
Music wire is a good choice for wire; high tensile strength, uniform, etc. As Dr Rodal said, use something with a known MOI. A flat steel disk or bar is ideal because it is easy to find the center so it hangs straight - easy to calculate. Pick a disk or bar that is close to the same mass as you expect your payload to be. The wire will stretch under load and change its torsional constant. Similar mass is better. Time with an ordinary camera set to 30 fps video mode. In post processing, using Virtual Dub or similar software, count off twenty oscillations. Record the starting crossing time stamp and the same for the 20th crossing time stamp. Divide by 20 to get the single oscillation period.
-
#1086 by Monomorphic on 09 Apr, 2016 23:59
-
Regarding Tajmar's COMSOL results. I'm having difficulty replicating TM010 inside the frustum using FEKO. It seems most of the resonance is in the waveguide. I've tried a monopole in the location of the magnetron and a dipole antenna inside the frustum. Both results are the same. I remember an NSF user, I think it was XRay (edit: maybe it was aero), saying he thought Tajmar's frustum was resonating in the waveguide, and not the frustum. These results may confirm that.
-
#1087 by Tellmeagain on 10 Apr, 2016 00:02
-
Music wire is a good choice for wire; high tensile strength, uniform, etc. As Dr Rodal said, use something with a known MOI. A flat steel disk or bar is ideal because it is easy to find the center so it hangs straight - easy to calculate. Pick a disk or bar that is close to the same mass as you expect your payload to be. The wire will stretch under load and change its torsional constant. Similar mass is better. Time with an ordinary camera set to 30 fps video mode. In post processing, using Virtual Dub or similar software, count off twenty oscillations. Record the starting crossing time stamp and the same for the 20th crossing time stamp. Divide by 20 to get the single oscillation period.
This may be overly complicated. Just hang his experiment payload as the experiment goes. Measure the oscillation period. Add a small mass (say, 1g) at a place L cm on the beam from the center. Measure the new period. I am sure a formula for the calibration( mN/degree) can be derived from these two periods.
Edited for grammar -
#1088 by otlski on 10 Apr, 2016 00:22
-
This may be overly complicated. Just hang his experiment payload as the experiment goes. Measure the oscillation period. Add a small mass (say, 1g) at a place L cm on the beam from the center. Measure the new period. I am sure a formula for the calibration( mN/degree) can be derived from these two periods.
Edited for grammar
Yes it can be derived. Your method might not have the timing sensitivity to detect a small mass. It might end up increasing the gross MOI less than 1%. A better way to do it would be to add two identical masses equal distance and opposite from each other - out on the ends where they are most effective. Heavy enough to increase the MOI say 20% might be about right.
In the end your previous post is likely right. It is simpler to just use a calculated torsion rate. Close enough to start with. If one were to get real thrust, one could re-characterize the apparatus to their heart's content. -
#1089 by Rodal on 10 Apr, 2016 00:34
-
Regarding Tajmar's COMSOL results. I'm having difficulty replicating TM010 inside the frustum using FEKO. It seems most of the resonance is in the waveguide. I've tried a monopole in the location of the magnetron and a dipole antenna inside the frustum. Both results are the same. I remember an NSF user, I think it was XRay, saying he thought Tajmar's frustum was resonating in the waveguide, and not the frustum. These results may confirm that.
Yes, that Tajmar experimental report is a real mess. I spent a lot of time modeling it with an exact solution and the dimensions he gives (even the corrected ones) don't quite do it. It looks like there is something wrong with their COMSOL model: that it was run with parameters not quite equal to the dimensions they gave us.
The huge opening into the waveguide explains their terribly low Q: completely wrong coupling.
aero may have further recollections, as he also tried to model this with MEEP.
What is amazing is that Tajmar et.al. list Shawyer as their adviser on how they run this experiment, and there were media reports reporting that Shawyer was very happy with Tajmar's experiment.
-
#1090 by frobnicat on 10 Apr, 2016 00:55
-
...Breaking my silence for once. A "over unity" scheme does not need constant acceleration with constant power, it needs constant force with constant power input. Because acceleration is relative (what you say of acceleration in SR depends on reference frame, from the spacecraft itself there is no decrease in acceleration at constant force...) there is always a great confusion with interpretation in delta kinetic energy. And a force can be applied at constant velocity, no acceleration, as in a tractor pulling a plow at constant velocity (wrt to ground), as in a turbojet thrusting for a plane moving against aerodynamic drag at constant velocity (wrt to air mass). ...
Professor Frobnicat, one of the great though-models I remember you made was the one of a road, and that the EM Drive does not have a road with friction for a vehicle to travel on.
I like you to consider Quantum Vacuum Friction. If it would exist, then, such friction could be exploited for propulsion:
https://www.newscientist.com/article/mg20927994-100-vacuum-has-friction-after-all/
http://physicsworld.com/cws/article/news/2009/mar/17/quantum-friction-does-it-exist-after-all
https://www.researchgate.net/publication/230902810_Quantum_vacuum_friction
http://arxiv.org/pdf/1504.02204.pdf
Your opinion on this hypothesis ?
Dr Rodal, per your suggestion I reproduce here PM I sent you with my initial impressions :
Your post https://forum.nasaspaceflight.com/index.php?topic=39772.msg1512764#msg1512764 links to interesting predicted phenomenon indeed. I only skimmed briefly. There is no such thing (experimentally shown so far) as an absolute frame of reference for translational movements in vacuum (real vacuum, not the vacuum of space that's full of CMB and dark matter and stuff above ground level...) and hence no intrinsic 0 translational velocity, but there is such a thing as an intrinsic 0 rotational velocity.
Intuitively it seems logical it makes possible a "friction" coupling able to despin an object wrt the "background". Thermodynamical aspects/consequences not clear to me, I'll have to think hard about that for some time. Will post on the forum when/if I feel I have relevant comments. Superficially at least I think there is a potential from there to discuss seriously "propellantless rotational propulsion" that looks more theoretically sound than propellantless translational deltaV, so this is exciting, maybe.
You remember that my charge of professor is not in physics. Even if I do have background on some undergrad physics, any comments would be limited by lack of expertise on more advanced topics like QFT and the such ( you know, waves etc...
)
End of PM (identity removed)
Few hours of sleep later and still not in depth reading...
Considering the second link, they seem to imply a kind of lateral force or "friction" in a Casimir coupling between parallel plates. No idea who is right or wrong but from the summaries in the article this looks like it's still debated ? Beyond the interest in understanding fundamental issues with quantum vacuum and its interpretation (you were among contributors indicating that Casimir effect doesn't "need" virtual particle formalism but rather be better modeled as retarded van der Waals forces ?) unless I missed some subtlety, the prospect of application to propulsion application seems weak, there is already plenty of means of exchanging forces between two parts of a system, at much larger distances, as with the BAE recycled photons system for instance. But this is not "propellantless" propulsion of a single spacecraft as understood by EMdrive proponents, ie that don't require some sacrificial counter part soaking the acquired momentum in the opposite direction. Side note : the beauty of BAE system vision is that the momentum equilibrium counter part is not sacrificed but reused, at the cost of needing a whole infrastructure/network of momentum exchange. Anyway, while it (lateral Casimir forces discussed in the physicsworld article) isn't looking promising as far as propellantless propulsion is directly concerned, we see some familiar outcome of a (probably broken) model that allows some form of quantum vacuum engineering : "extracting unlimited energy from the quantum vacuum" (please consult the article for a context to the argument).
Third link is paywalled mostly, haven't full access. Seems to deal with exchanges between solids, part A & part B through QV, not part A alone to QV. As above I don't see much propulsion application. Slow down of rotation of black holes is also mentioned, that would be part A (and what a part !) against QV alone... But I'll leave the vicinity of event horizons to more adventurous minds. Likewise the fourth link, I'm sorry dr Rodal, even comparatively tame magnetars and the required formalism crush my intuition down to sub µm altitude. This is again about spindown of an isolated object, alone in vacuum.
The first link on newscientist article is also about spinning bodies, dust specks now ! It is more accessible, but frustratingly so.
The publication behind the newscientist article is paywalled also, but there is some similar arxived paper by same authors. Being recalled there is (was ?) a "heated debate" with lateral Casimir (article is 2010... is it settled since ?). And again failing at following the formalism as soon as bottom column 1 of page 1. Vacuum at temperature T0 ? What do they mean by that ? Thermal bath ? Isn't a thermal bath defining a preferred rest frame (like CMB does) ? Why does a QV effect, supposed relativistic, "needs" such an above zero point medium ? Isn't the proposed dissipation of angular momentum simply some kind of coupling/scattering with such a momentum carrying medium (photons of the thermal bath) ? The effect look very very weak, especially with low vacuum temperatures. Can't we spindown a rotating body, more efficiently, and still "propellentlessly" by radiating away with synchrotron radiation ?
Assuming perfect solid construction materials (infinite rigidity and 0 density) it seems possible within SR mechanics alone to despin a body to an arbitrarily small remnant rotation with an arbitrarily low sacrificed mass : just extend an arm (with the magic perfect material) from the body, with a small lump of reaction mass on the end that moves away (from center of rotating body). Extend arm long enough so that the end reaches high enough velocities (wrt center of rotating body) until enough of the (conserved total) angular momentum is concentrated in the small lump. And then release arbitrarily small lump. This is as propellantless as possible, and unlike the linear change of momentum, where same argument of concentrating an arbitrarily high momentum into an arbitrarily small sacrificed mass can be made, but at the expense of energy. And energy weighs. Reaching c ejection velocity imposes a limit to the mass efficiency (as a measure of "propellantlessness" of a propulsion concept, matter/anti-matter powered photon rocket being the best, theoretically so far within uncontroversial frameworks). On the other hand, in the case of de-spinning there is no need to artificially power the acceleration of the reaction mass, since the energy is given by the intrinsic rotational energy of the spinning body (and that is precisely what we want to get rid of)/
Will try to substantiate those ideas quantitatively when time permits. -
#1091 by Monomorphic on 10 Apr, 2016 00:56
-
Yes, that Tajmar experimental report is a real mess. I spent a lot of time modeling it with an exact solution and the dimensions he gives (even the corrected ones) don't quite do it. It looks like there is something wrong with their COMSOL model: that it was run with parameters not quite equal to the dimensions they gave us.
Even if I try and match the comsol model's shorter and offset waveguide, I get something different.
-
#1092 by RotoSequence on 10 Apr, 2016 01:09
-
Yes, that Tajmar experimental report is a real mess. I spent a lot of time modeling it with an exact solution and the dimensions he gives (even the corrected ones) don't quite do it. It looks like there is something wrong with their COMSOL model: that it was run with parameters not quite equal to the dimensions they gave us.
Even if I try and match the comsol model's shorter and offset waveguide, I get something different.
This might be a dumb question, but is there a preset for waveguides in COMSOL that's being used in the Tajmir build sim? It's almost like the COMSOL simulation is making assumptions about where and which parts to simulate. -
#1093 by Monomorphic on 10 Apr, 2016 01:37
-
Iulian Berca dimensions. 2.45Ghz. Definitely not TM212.
-
#1094 by Monomorphic on 10 Apr, 2016 01:46
-
NSF-1701. 2.45Ghz
-
#1095 by Rodal on 10 Apr, 2016 01:49
-
Iulian Berca dimensions. 2.45Ghz. Definitely not TM212.
Looking at the top of your model it has 4 poles, so m=2
Also, longitudinally there are two wavepatterns so that, p=2
It seems like a malformed mode, (the top 1/2 different from the bottom half).
Could be n=1, giving m,n,p=212
What mode do you think it is?
It looks like a face, with two big eyes at the top and a big mouth at the bottom
-
#1096 by Monomorphic on 10 Apr, 2016 01:55
-
#1097 by Rodal on 10 Apr, 2016 02:26
-
What mode do you think it is?
TM something?
Both Iulian Berca and Dave Distler used magnetrons (I don't recall whether Tajmar used a magnetron, but given the fact that the experimental wiki lists 2.44 GHz, it seems probable that Tajmar also used a magnetron, somebody should check), so to be sure what they excited you need to run the full spectrum of the magnetron, encompassing 2.45 GHz +/- 0.03 GHz= 2.42 GHz to 2.48 GHz:
Need to consider the peaks at 2.43 GHz and 2.47 GHz as well, since they are very important in the magnetron spectrum.
Iulian Berca's seems reasonably close to TM212, considering the spectrum
-
#1098 by SeeShells on 10 Apr, 2016 04:47
-
What mode do you think it is?
TM something?
[/quote
Both Iulian Berca and Dave Distler used magnetrons (I don't recall whether Tajmar used a magnetron, but given the fact that the experimental wiki lists 2.44 GHz, it seems probable that Tajmar also used a magnetron, somebody should check), so to be sure what they excited you need to run the full spectrum of the magnetron, encompassing 2.45 GHz +/- 0.03 GHz= 2.42 GHz to 2.48 GHz:
Need to consider the peaks at 2.43 GHz and 2.47 GHz as well, since they are very important in the magnetron spectrum.
Iulian Berca's seems reasonably close to TM212, considering the spectrum
Dr. Rodal,
A clean DC driven power supply for the magnetron will provide a good narrow band width RF source.
Shell
-
#1099 by X_RaY on 10 Apr, 2016 07:27
-
Regarding Tajmar's COMSOL results. I'm having difficulty replicating TM010 inside the frustum using FEKO. It seems most of the resonance is in the waveguide. I've tried a monopole in the location of the magnetron and a dipole antenna inside the frustum. Both results are the same. I remember an NSF user, I think it was XRay (edit: maybe it was aero), saying he thought Tajmar's frustum was resonating in the waveguide, and not the frustum. These results may confirm that.
Yes the waveguide dimensions is not negligible(with respect to this small cavity dimensions) when the coupling is realized directly into the cavity, But it is possible they used a kind of iris for coupling.
)