-
#2800 by Rodal on 23 Feb, 2017 15:24
-
Monomorphic -
Now I realize that your rig is free to pivot in all directions about the suspension point, I can see how it can sustain a mode of relatively un-damped oscillations despite serious damping!
The answer is that the rig is free to pivot about an axis through the pivot point and the damper. I think you've already mentioned that damping at the suspension point produces poor damping - same with damping on any axis through the damper. You will see that mode of oscillation predominate eventually as other modes are damped out.
The solution is very simple - multiple dampers, probably weaker to keep the same total effect. Any two dampers on the bench will suppress all modes of rigid gravity forced oscillation on any axis through the suspension point. Any three non-co-linear dampers will suppress all such motion about any axis at all. You will still be left with non-rigid vibrations of the rig, some obvious if you only have two dampers.
I understand that Monomorphic is using a "U channel" paddle that provides damping (mostly) in the direction of rotation and also in the vertical direction, perpendicular to the plane of rotation.
Hence there is only one other Cartesian direction left: along the longitudinal axis of the U channel.
Although one would expect this motion (directed along the radius of rotation, along the axis of the beam) would be negligibly small, it would be easy to fix: just close with a plate the U channel to increase damping along that radial direction too. -
#2801 by Rodal on 23 Feb, 2017 15:31
-
If it is ground noise, then how do you explain this? Static test with no electrical components powered.
What I will do next is repeat this test with each of the main components powered individually. See which is interfering most, mitigate it and move to the next. Repeat until the trace is as flat as possible.
IMHO Jamie you're not seeing very much seismic activity. When I lived in California my partner and I belonged to the seismic net and had string pendulums of different lengths hung from the ceiling. With these simple weights, on a string, we could deduce what direction we were seeing a propagating P or S wave, how strong and which direction a tremor or quake was was. We even wrote a program in Basic running on our XT-PC to semi-automate the process.
With your power off your beam oscillations disappear.
I think your seeing a Lorentz motor action from some "harmonic beat" magnetic fields created in your hardware that's causing your beam to fluctuate, doesn't matter that your unit is made from Aluminum, it still will respond to Time-varying magnetic fields.
If this is the case where it's your hardware shielding with mu-metal could help to minimize it. http://www.ebay.com/itm/Mumetal-Nikel-Permalloy-Magnetic-shielding-foil-Sheet-Mu-metal-0-1T-30-45cm-/171173026902?hash=item27dab51c56:g:T24AAOxyyjpRxam8
My Best,
Shell
Relative magnetic permerabilty of aluminum = 1.000022 (barely distinguishable from free space)
of air = 1.00000037
of Mu-metal = 20,000
of electrical steel = 4,000
Since the noise displays random behavior, I would suspect instead of the aluminum being influenced by magnetic fields, that the noise is due to thermal excitations from the electronic equipment (damping heat).
We will see
-
#2802 by Monomorphic on 23 Feb, 2017 15:33
-
Very good news. The kangaroo mini-pc is not the cause of a significant amount of noise. This means the primary culprit is something else and major changes to the "brains" of the pendulum are not necessary.
Anyone want to make bets on what is the major cause?
1. USB spectrum analyser
2. Signal generator
3. USB Hub
4. Pre-amp
5. Second pre-amp
6. Main amp
7. Wireless keyboard dongle
8. Main battery
9. Secondary battery
10. multi-sensor
Or they could all add up evenly, but I doubt that scenario.
-
#2803 by Rodal on 23 Feb, 2017 15:36
-
Very good news. The kangaroo mini-pc is not the cause of a significant amount of noise. This means the primary culprit is something else and major changes to the "brains" of the pendulum are not necessary.
Anyone want to make bets on what is the major cause?
1. USB spectrum analyser
2. Signal generator
3. USB Hub
4. Pre-amp
5. Second pre-amp
6. Main amp
7. Wireless keyboard dongle
8. Main battery
9. Secondary battery
10. multi-sensor
Or they could all add up evenly, but I doubt that scenario.
I bet on whatever component is damping most heat.
Can you scan the components with an infrared pyrometer and ascertain which one is damping most heat ? -
#2804 by RERT on 23 Feb, 2017 15:39
-
Monomorphic -
Now I realize that your rig is free to pivot in all directions about the suspension point, I can see how it can sustain a mode of relatively un-damped oscillations despite serious damping!
The answer is that the rig is free to pivot about an axis through the pivot point and the damper. I think you've already mentioned that damping at the suspension point produces poor damping - same with damping on any axis through the damper. You will see that mode of oscillation predominate eventually as other modes are damped out.
The solution is very simple - multiple dampers, probably weaker to keep the same total effect. Any two dampers on the bench will suppress all modes of rigid gravity forced oscillation on any axis through the suspension point. Any three non-co-linear dampers will suppress all such motion about any axis at all. You will still be left with non-rigid vibrations of the rig, some obvious if you only have two dampers.
I understand that Monomorphic is using a "U channel" paddle that provides damping both in the direction of rotation and in the vertical direction, perpendicular to the plane of rotation.
Hence there is only one other Cartesian direction left: along the longitudinal axis of the U channel.
(RERT removes picture)
Although one would expect this motion (directed along the radius of rotation, along the axis of the beam) would be negligibly small, it would be easy to fix: just close the U channel to increase damping along that direction too.
With the posited mode of oscillation, the paddle will 'twist' in the pot, not move along any of the Cartesian axes. I think Monomorphic has already said that that produces poor damping. Consider leverage, for example: if the beam at the LDS is moving a few microns, the maximum movement of the twisting paddle in the damping fluid might be a tenth of that, as it is closer to the axis of rotation. Further, in a twisting motion, shear will be reduced as the fluid can actually undergo rigid rotation.
Honestly, I think a second damper can't hurt (modulo effort), and might make a big difference. -
#2805 by Monomorphic on 23 Feb, 2017 15:44
-
I bet on whatever component is damping most heat.
Can you scan the components with an infrared pyrometer and ascertain which one is damping most heat ?
That is likely to be one of the amps. It could also be a moderate power consumer with its power cable coiled in the middle.
I have an IR camera that I have used extensively. The USB spectrum analyser also damps a lot of heat.
-
#2806 by PotomacNeuron on 23 Feb, 2017 15:49
-
Very good news. The kangaroo mini-pc is not the cause of a significant amount of noise. This means the primary culprit is something else and major changes to the "brains" of the pendulum are not necessary.
Anyone want to make bets on what is the major cause?
1. USB spectrum analyser
2. Signal generator
3. USB Hub
4. Pre-amp
5. Second pre-amp
6. Main amp
7. Wireless keyboard dongle
8. Main battery
9. Secondary battery
10. multi-sensor
Or they could all add up evenly, but I doubt that scenario.
My bet is your ground loop DC current interacting with external changing magnetic field. The external changing magnetic field may be generated by a transformer nearby or by leaking magnetic field of your home main. It might first show up as your RF amp causing the problem. But the true culprit might later be found to be the DC current supplied to your RFamp. -
#2807 by PotomacNeuron on 23 Feb, 2017 16:04
-
Very good news. The kangaroo mini-pc is not the cause of a significant amount of noise. This means the primary culprit is something else and major changes to the "brains" of the pendulum are not necessary.
Anyone want to make bets on what is the major cause?
1. USB spectrum analyser
2. Signal generator
3. USB Hub
4. Pre-amp
5. Second pre-amp
6. Main amp
7. Wireless keyboard dongle
8. Main battery
9. Secondary battery
10. multi-sensor
Or they could all add up evenly, but I doubt that scenario.
My bet is your ground loop DC current interacting with external changing magnetic field. The external changing magnetic field may be generated by a transformer nearby or by leaking magnetic field of your home main. It might first show up as your RF amp causing the problem. But the true culprit might later be found to be the DC current supplied to your RFamp.
The leaking magnetic field of your main may be from ground loops too. Your home water main may carry amperes of AC current. This can be caused by electricity trying to find the lowest resistance route to return to the transformer ground rod. Yes, I have seen that. -
#2808 by RERT on 23 Feb, 2017 16:05
-
(snip)
The solution is very simple - multiple dampers, probably weaker to keep the same total effect. Any two dampers on the bench will suppress all modes of rigid gravity forced oscillation on any axis through the suspension point. Any three non-co-linear dampers will suppress all such motion about any axis at all. You will still be left with non-rigid vibrations of the rig, some obvious if you only have two dampers.
I understand that Monomorphic is using a "U channel" paddle that provides damping both in the direction of rotation and in the vertical direction, perpendicular to the plane of rotation.
Hence there is only one other Cartesian direction left: along the longitudinal axis of the U channel.
(RERT removes picture)
Although one would expect this motion (directed along the radius of rotation, along the axis of the beam) would be negligibly small, it would be easy to fix: just close the U channel to increase damping along that direction too.
With the posited mode of oscillation, the paddle will 'twist' in the pot, not move along any of the Cartesian axes. I think Monomorphic has already said that that produces poor damping. Consider leverage, for example: if the beam at the LDS is moving a few microns, the maximum movement of the twisting paddle in the damping fluid might be a tenth of that, as it is closer to the axis of rotation. Further, in a twisting motion, shear will be reduced as the fluid can actually undergo rigid rotation.
Honestly, I think a second damper can't hurt (modulo effort), and might make a big difference.
OK, thanks for the explanation, rotating around the axis of the U channel is being damped by the horizontal flat surface of the U and the lateral vertical surfaces, but if that is not enough, Monomorphic can put a flat rectangular plate under the U to further dampen that rotation ( essentially flaps coming out of the U). Is that the rotation you have in mind damping further ?
He can actually have several such flaps come out of the U to dampen that rotation
Suppose Monomorphic switches to two dampers, half the size of the current one, each the same distance as the present one on opposite sides of the beam. Damping about the vertical suspension axis would be the same. Damping about the Horizontal axis through the pivot along the beam would be the same.
The second pot would provide high damping about the axis through the first (&vv), because the paddle must move much farther through the pot for a given angular displacement in that mode of vibration. Damping in any pot of twisting about an axis through the pot is always much lower, because of the reverse: the travel of the paddle through the damping fluid drops to zero at the axis of rotation.
Changing the shape of the paddle might help, but adding a second damper is simpler and will obviously be effective. -
#2809 by Rodal on 23 Feb, 2017 16:19
-
...
I can see the need for this when Monomorphic had a damping ratio =13.35 (according to your calculations), and a strange initial undamped motion:
Suppose Monomorphic switches to two dampers, half the size of the current one, each the same distance as the present one on opposite sides of the beam. Damping about the vertical suspension axis would be the same. Damping about the Horizontal axis through the pivot along the beam would be the same.
The second pot would provide high damping about the axis through the first (&vv), because the paddle must move much farther through the pot for a given angular displacement in that mode of vibration. Damping in any pot of twisting about an axis through the pot is always much lower, because of the reverse: the travel of the paddle through the damping fluid drops to zero at the axis of rotation.
Changing the shape of the paddle might help, but adding a second damper is simpler and will obviously be effective.
But now with damping ratio = 0.23 this may not be as important as taking care of the noise source. Do you agree?
The damping is however something important to take into account if Monomorphic later decides to test also with a higher damping ratio near, equal to or exceeding critical damping. -
#2810 by jmossman on 23 Feb, 2017 16:31
-
Very good news. The kangaroo mini-pc is not the cause of a significant amount of noise. This means the primary culprit is something else and major changes to the "brains" of the pendulum are not necessary.
Anyone want to make bets on what is the major cause?
1. USB spectrum analyser
2. Signal generator
3. USB Hub
4. Pre-amp
5. Second pre-amp
6. Main amp
7. Wireless keyboard dongle
8. Main battery
9. Secondary battery
10. multi-sensor
Or they could all add up evenly, but I doubt that scenario.
I bet on whatever component is damping most heat.
Can you scan the components with an infrared pyrometer and ascertain which one is damping most heat ?
Hi Monomorphic,
Any chance of getting a small styrofoam box and dumping all of the electronics inside for a quick power-on test? (i.e. hang/attach/duct tape styrofoam box to pendulum) I'd propose not connecting the amp to the frustum during this test to avoid any wires having to exit the styrofoam box. In lieu of a stryofoam box, perhaps a simple plastic enclosure closed up (the pressure buildup for a few minutes of runtime shouldn't be too large, and a true air-tight seal should be possible with a caulk/sealer/glue if desired).
If thermal is the primary source of noise, placing them in a closed box for a few minutes while powered on should give a pretty clear change in the noise profile/bandwidth. Removing the lid and rerunning would presumably allow most of the thermal noise to return. Actually, I'm wondering if a simple air-tight garbage bag wouldn't be sufficent to quickly check for broad brush-stroke differences in the thermal noise profile/bandwidth (although melting the bag might be of concern at certain locations, a towel or two could likely mitigate for a quick couple of minutes test).
If no change in noise was observed with and without the lid, then an EM interaction would be the primary source. Right now we don't have a ball-park estimate of how much thermal noise vs EM/Lorentz/ground loop/etc noise the setup has to deal with.
Also, were the LDS sensors powered-on during both the "electronics payload on" and "electronics payload off" tests? If yes, perhaps testing with the LDS powered-off while "electronics payload on" might also be worthwhile. -
#2811 by RERT on 23 Feb, 2017 18:35
-
...
I can see the need for this when Monomorphic had a damping ratio =13.35 (according to your calculations), and a strange initial undamped motion:
Suppose Monomorphic switches to two dampers, half the size of the current one, each the same distance as the present one on opposite sides of the beam. Damping about the vertical suspension axis would be the same. Damping about the Horizontal axis through the pivot along the beam would be the same.
The second pot would provide high damping about the axis through the first (&vv), because the paddle must move much farther through the pot for a given angular displacement in that mode of vibration. Damping in any pot of twisting about an axis through the pot is always much lower, because of the reverse: the travel of the paddle through the damping fluid drops to zero at the axis of rotation.
Changing the shape of the paddle might help, but adding a second damper is simpler and will obviously be effective.
(RERT snips image)
But now with damping ratio = 0.23 this may not be as important as taking care of the noise source. Do you agree?
(RERT snips image)
The damping is however something important to take into account if Monomorphic later decides to test also with a higher damping ratio near, equal to or exceeding critical damping.
I think it would be great to find and eliminate the noise source. Whatever it is, it is exciting and oscillation with a period of around 40 seconds, which is probably a natural mode of the rig - the noise itself seems very unlikely to have that characteristic period. The noise would probably be greatly reduced if that mode was properly damped.
I think moving to multiple dampers would probably also eliminate the noise, and should be tried if the source can't be found or can't be eliminated. In general, if designing such a setup from scratch (with 20/20 hindsight) one would not have a system where the level of damping was highly mode-dependent, and would probably go for 3 non-co-linear dampers.
Actually I suspect multiple dampers would make it possible to lower the overall damping to produce a more responsive system, and/or one with less noise. So I don't think it is more applicable in a more highly damped case. -
#2812 by Rodal on 23 Feb, 2017 22:12
-
...
I can see the need for this when Monomorphic had a damping ratio =13.35 (according to your calculations), and a strange initial undamped motion:
Suppose Monomorphic switches to two dampers, half the size of the current one, each the same distance as the present one on opposite sides of the beam. Damping about the vertical suspension axis would be the same. Damping about the Horizontal axis through the pivot along the beam would be the same.
The second pot would provide high damping about the axis through the first (&vv), because the paddle must move much farther through the pot for a given angular displacement in that mode of vibration. Damping in any pot of twisting about an axis through the pot is always much lower, because of the reverse: the travel of the paddle through the damping fluid drops to zero at the axis of rotation.
Changing the shape of the paddle might help, but adding a second damper is simpler and will obviously be effective.
(RERT snips image)
But now with damping ratio = 0.23 this may not be as important as taking care of the noise source. Do you agree?
(RERT snips image)
The damping is however something important to take into account if Monomorphic later decides to test also with a higher damping ratio near, equal to or exceeding critical damping.
I think it would be great to find and eliminate the noise source. Whatever it is, it is exciting and oscillation with a period of around 40 seconds, which is probably a natural mode of the rig - the noise itself seems very unlikely to have that characteristic period. The noise would probably be greatly reduced if that mode was properly damped.
I think moving to multiple dampers would probably also eliminate the noise, and should be tried if the source can't be found or can't be eliminated. In general, if designing such a setup from scratch (with 20/20 hindsight) one would not have a system where the level of damping was highly mode-dependent, and would probably go for 3 non-co-linear dampers.
Actually I suspect multiple dampers would make it possible to lower the overall damping to produce a more responsive system, and/or one with less noise. So I don't think it is more applicable in a more highly damped case.
I understand that you are saying that the vertical motion (up like a rocket) with apparently negligible damping in the image below:
is due to rotation of the torsional pendulum beam around the horizontal longitudinal axis of the beam, while the long-time overdamped motion is due to rotation of the torsional pendulum around the vertical axis of rotation of the pendulum.
This presents issues of measurement , and not just issues of damping (see attached figure)
1) I thought that Monomorphic was using a laser to optically measure the displacement. If there would be a rotation of the beam around its horizontal longitudinal axis, it looks to me that the reflection would be at another angle and hence not confused by the readout as being due to rotation of the torsional pendulum around its vertical axis of rotation . (I look for Monomorphic's feedback on this)
2) For argument's sake, let's say that you are right, that Monomorphic's optical measurement is somehow interpreting the rotation of the beam around its horizontal longitudinal axis as a displacement in the same direction as the rotation of the torsional pendulum around its vertical axis of rotation(see figure). In that case, this would be not only an issue of damping such a motion but most importantly also it would involve a measurement issue, possibly needing 2 optical measurements to resolve
-
#2813 by Tcarey on 24 Feb, 2017 07:05
-
If the noise is coming from the electronics then I have to wonder what the mechanism is that is coupling it to the apparatus. If it is a heat effect I would expect it to be either the beam dimensions changing due to local heating or from some aerodynamic effect from rising air.
If it is from heating coupled to the beam then I would expect that to hit an equilibrium and see the noise level drop.
A possible mechanism could be local spot heating of the beam causing it to warp slightly. It is is from rising air I suppose that it could be turbulent air flow and not reach an equilibrium.
The argument advanced that it might be a ground loop interacting with some local magnetic field, say mains induced, is problematic for me. A steady field created by DC interacting with 60Hz AC is not going to generate the long period oscillations seen in the data.
If the DC current levels are changing then an interaction with a stable magnetic field could impart motion to the beam.
As far as isolation of the test stand from the environment I think of something similar to what Shell proposed with a few minor changes. An air mattress between the table and the bottom of the apparatus should do well with the vertical vibration in terms of not coupling the higher frequency vibrations very effectively. I suspect it would do even better in isolating horizontal vibrations and torsional direction vibrations. Of course the suggestions made to put a significant mass on top of the first isolation material and then a second layer of isolation material could easily and cheaply be done with two air mattresses and a slab of marble reject from a kitchen counter installer or manufacturer.
Perhaps there is something I am not understanding (quite likely) about how vertical vibrations are causing the rotational movement of the beam or the fixed apparatus relative to the beam. That suggests to me that the vibrations that are creating noise, if they are external, are likely to be horizontal or torsional or more likely a combination of axises.
I applaud Monomorphic's efforts and willingness to share the process and results with us. My best to you.
-
#2814 by RERT on 24 Feb, 2017 09:32
-
Dr. Rodal -
Forgive me, I'll come back in a few minutes on your latest post, but I just want to air this further clarification I've put together.
I realized this morning that the effectiveness of a damper depends on the couple generated contrary to the rotation about an axis. The couple consists of the damping force, times the distance from the axis through which the force acts. Where the axis is close to the damper paddle, not only is the force lower because the velocity through the damping fluid is lower, but the distance from the axis is smaller.
Doing the integration, one gets the following chart for the relative damping constant as a function of the distance between the paddle centre and the rotation axis, for a rectangular paddle perpendicular to the axis. It no longer surprises me that Monomorphic reports that axial damping is very poor!
For fun I've included the damping from two half paddles separated by 3 paddle lengths, as a function of distance from the rotation axis to one paddle.
-
#2815 by RERT on 24 Feb, 2017 10:39
-
I understand that you are saying that the vertical motion (up like a rocket) with apparently negligible damping in the image below:
(RERT removes image)
is due to rotation of the torsional pendulum beam around the horizontal longitudinal axis of the beam, while the long-time overdamped motion is due to rotation of the torsional pendulum around the vertical axis of rotation of the pendulum.
Not really: I wasn't discussing this phenomenon. I'm discussing the slow (circa 40 second period) noise signal in the latest measurements, and actually in the previous powered tests. I think that's indicative of a specific underdamped mode of oscillation of the rig, about an axis through the damper and the pivot point.
As regards the case you cite, the simple model for beam motion about the suspension wire has only two state variables: the position and velocity of the beams horizontal LDS number. Once the initial transient noise has reached a v=0 position, it would start to decay in the slow relaxation mode like the rest of the curve, if the measurement was due merely to angular rotation of the beam about the pivot vertical. The fact that it doesn't proves that something else is happening.
In fact if you look closely at the transient, it is well damped, though less than critically damped.
To explain the correlation between vertical and horizontal measurements, I postulated that the rig rotated about a horizontal axis parallel to the beam and through the pivot. If you work through the geometry, you will see that the motion about that axis produces horizontal and vertical displacements:
h=R tanθ+O(θ²) and v=εtanθ+O(θ²)
Where R is the distance from the axis of rotation to the beam centre, and ε is the horizontal distance from the vertical LDS sensor to the axis of rotation. Hence the ratio of v to h is small, but they move instep, as observed. If
ε is zero, there is still an effect at second order.
I think it's obvious that the rig will have a mode of oscillation as I described, although I could have missed something in the walk around.QuoteThis presents issues of measurement , and not just issues of damping (see attached figure)
The measurements are working, so we can conclude that the angles are not large enough for this to matter. It isn't easy to see how the LDS could distinguish between an off-centre reflection and an absorbing surface.
1) I thought that Monomorphic was using a laser to optically measure the displacement. If there would be a rotation of the beam around its horizontal longitudinal axis, it looks to me that the reflection would be at another angle and hence not confused by the readout as being due to rotation of the torsional pendulum around its vertical axis of rotation . (I look for Monomorphic's feedback on this)
Just to repeat, I wasn't talking about the rotation of the beam about it's horizontal axis, but another parallel axis through the pivot.Quote2) For argument's sake, let's say that you are right, that Monomorphic's optical measurement is somehow interpreting the rotation of the beam around its horizontal longitudinal axis as a displacement in the same direction as the rotation of the torsional pendulum around its vertical axis of rotation(see figure). In that case, this would be not only an issue of damping such a motion but most importantly also it would involve a measurement issue, possibly needing 2 optical measurements to resolve
Again, not that axis, but I sort of agree. I didn't want to suggest Monomorphic added an LDS. For one thing, if I were him, I would probably have punched someone by now - possibly RERT. But yes, another horizontal sensor at the pivot point would show if other rotations were happening. The angular displacement of the beam would then be a function of the difference between the two horizontal displacements. -
#2816 by Peter Lauwer on 24 Feb, 2017 11:31
-
Yes, let's recall that a number of builders started to run experiments with teeter-totters with damping.
Monomorphic,They are known as Lab Jacks - often the platforms and scissor arms are aluminum, but the hinge bars and threaded parts are usually steel.
Thanks. Anodized aluminum and stainless steel will work fine. Gotta love Amazon free same day delivery. I have this one coming by the end of the day.
I'm not trying to tell you how to run your tests that decision is truly up to you, although I will try to provide you with some info with my build that helped me.
Dr. Rodel is correct in your oil dampner is over damped and data gleaned from it will be hard to evaluate. You want your pendulum to be able to move quicker than the oil bath dampened speed to be able to see the acceleration and when it responds to removal of the RF drive signal.
I tried oils several kinds but they proved problematic as the viscosity and therefor the dampening factor would change over temperature. That is why I settled for anti-freeze and water mix, as it maintained it's viscosity over temperature. The key to increasing the damping was simple, just increase of decrease the paddle area by moving the paddle up or down in the stabilized anti-freeze mixture.
My Best,
Shell
Corrected speeelings
Teeter-totter testing was criticized because the (EM Drive) force was opposed (besides the inertial force) only by the damping force (since in a teeter-totter there is no elastic stiffness), and the damping force is difficult to characterize because damping force in viscous liquids is a function of viscosity and viscosity of many liquids is sensitive to temperature changes from room temperature.
So, the focus, for good reason, was shifted towards conducting EM Drive experiments with the instrument used for sensitive measurements of gravitational forces, for hundreds of years: the torsional pendulum as in the Cavendish pendulum.
In a torsional pendulum one seeks an instrument where the force (in this case the putative EM Drive force) is opposed by (in addition to the inertial force by) an elastic force like the torsional stiffness of the spring, which is well characterized and defined, as the torsional stiffness is dependent on the cross-sectional area and length of the spring, and the modulus of elasticity (E) of the wire, which, for steel, is pretty insensitive to temperature variations (at temperatures near room temperature, much lower than the softening temperatures of steel).
Therefore, the whole purpose of using a torsional pendulum is to run the instrument in the regime where the elastic stiffness plays the most important role, and where viscosity forces play a minor role: the underdamped torsional pendulum.
Ideally, one wants the torsional pendulum to be underdamped enough so that the period of torsional oscillations is pretty insensitive to damping. This may require a fluid with small damping, for example SeeShells used water with anti-freeze.
Conducting an experiment with a torsional pendulum with high damping such that the torsional pendulum is near critical damping, or much worse, it is overdamped, defeats the whole purpose of the torsional pendulum: as when the pendulum is overdamped, the viscosity plays the major role, instead of the well-characterized elastic stiffness of the torsional wire.
Present state for Monomorphic's experiments: some people recommended to increase the elastic stiffness. Monomorphic said that he has #18 wire. There is so much that Monomorphic will be able to do by increasing elastic stiffness because the higher the stiffness of the wire: the smaller the motion. The smaller the motion the more error he will encounter in his optical measurement.
A much more productive way to go would be to decrease the amount of damping, so that the pendulum is underdamped enough that the elastic stiffness plays the major role: damping should be small enough so that there is a well defined period of torsional oscillation.
(trying to catch up after 1 week of holiday; almost 10 pages NSF to work through)
I am using an electronic version of the torsion balance, after Crandall, 1983 (attached). In this way, the torsion wire is only used as low friction pivot (torsion moment <1% of servomechanism moment).
It is in principle possible to let the electronics do the damping. However, I found it beneficial, in previous setups of the balance, to use liquid damping as well (I used silicone oil, btw).
I wonder how important it is to have exactly a critical (or actually, a little under-) damped balance. Of course, if you want to monitor the time behaviour of the EMdrive as good as possible it might be. But we assume it should be a constant force (at resonance).
-
#2817 by Peter Lauwer on 24 Feb, 2017 11:46
-
3) Are you going to use wire #18 that will increase the stiffness (higher frequency) at the expense of reducing the motion?
If I can avoid having to switch out the wire that would be preferable. It is totally doable, I would just prefer the higher sensitivity of the thinner wire - and to avoid the extra work.
Here is another run with an even larger dampening paddle. This looks like an improvement to me and perhaps I could use an even larger paddle. However, with a sensitivity on the order of a few uN, this should be plenty good for powered runs.
Magnetic tap tests will be noisier than powered runs as I am moving around in the room quite a bit.
Btw, you can increase the damping by drilling holes in the damper. Because of the more turbulent flow, more energy is absorbed by the liquid then. -
#2818 by Peter Lauwer on 24 Feb, 2017 11:50
-
I've never been happy with the dampening paddle attached to the counterweight. I think it will work better if I separate the two like this.
I'm only on for a second. I also never liked the fluid dampner at the end of the beam.
Maybe you could consider a center placed fluid dampner.
MY Best,
Shell
Or use two dampers, symmetrical distributed. [added later: I see RERT already discussed this] -
#2819 by Monomorphic on 24 Feb, 2017 12:29
-
I think I have traced the source of a majority of the noise to three bundles of wires associated with the USB hub. Working theory now is that these bundles of wires are creating EM fields that are either interacting with each other, causing the wires to flex slightly -- transferring that force to the beam, or interacting with the Earth's magnetic field for the same effect. Next step is to cut the cables to minimum lengths. USB cables contain 4 wires each, power, ground, and two for data. That's soldering plus I have to add shielding where removed and then heat shrink - so this may take a day or two.
Magnetic tap tests will be noisier than powered runs as I am moving around in the room quite a bit.