Magnetic Levitation Anamoly Needs Your Explanation 1

[Auburnate]

I have a short video of a magnetic levitator I built and am trying to explain a curious effect.

http://www.youtube.com/watch?v=fWRJL3Nbb-0

Notice the washer oscillating side to side without touching anything? Can someone please explain what this can be attributed to?


Theory 1. Flux density: If the return path for the flux is biased in one direction (i.e., through the
back of the visible setup), then the washer tends to rotate so as to minimize
the path length in air.

Theory 2. Irregularities in the surface of the washer cause the sensors to see it in different spots, causing it to oscillate like a limit cycle.

Theory 3. Video is faked. Change career to film editor immediately.


Thanks,

Nathan

Short explanation of operation:

When a current passes through a wire, a small magnetic field is generated. When that wire is coiled around a ferromagnetic core, a very strong magnetic field is created. This magnetic field provides the lifting force to cause ferric metal objects (e.g. a steel ball) to rise towards the electromagnet. The IR diode and IR phototransistor measure the location of the ball. With the essential phase lead circuitry, a control signal is generated that drives a power transistor. The power transistor turns on or off the electromagnet. The timing of when to turn off or on the electromagnet causes the ball to rise and fall so fast as to appear to be levitating. Just before the ball is outside the controlling distance of the electromagnet, the current is sent through the coil and the ball is pulled back up.

 

[btrueblood]

Theory 1 sounds close to correct; the presence of the sensors creates an asymmetry in the magnetic field, the washer wants to align itself to this assymmetry. It oscillates in the weak potential "null point", just like a torsion pendulum.

 

[IRstuff]

The point of levitation is a metastable point, since you must drive it back into position with the feedback controller. However, your sensor is optical, and is therefore sensitive to the optical crossection of the object. Since there is nothing to constrain the washer from rotating, it would otherwise simply continue to rotate from being hit by air currents, etc. However, as it rotates, it changes when your sensors respond, thereby causing a different field strength generated from the control loop.

You should be using a ball bearing to avoid this type of behavior.

TTFN

FAQ731-376

 

[MacGyverS2000]

To combine the two replies... paint a bearing with a dot. If the bearing doesn't stop the spinning, your optical sensor is not the issue.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[logbook]

It is not clear from the video how the electro-magnet is supported. Also the drive path to the electromagnet is not visible. Either could create a magnetic bias to favour a specific orientation of the washer. If there is an initial moment on the washer, there is so little damping in the system that it may well just oscillate for a very long time. Have you left it running for several hours or does it over heat and require shutting down?

To test the theory you could attached a light piece of tissue to the washer to try to increase the air damping effect. If the oscillations get less with time the point is made. If you can estimate the mid-point of the oscillation, you could try starting it at this position, ie giving it less of a moment in the first place.

I’m not buying IRstuff’s theory because the sense circuit/feedback is vertical and does not apply any sort of torque to the system.

 

[VEBill]

I believe that IRstuff is thinking roughly the same thing as me, but he used a couple of really big words ;-) ...

Putting it simply: Align the coil (probably move it forward a tiny bit) so that its axis exactly passes through the axis of the optical sensor.

The torque arises quite naturally from the offset combined with the shape of the washer.

 

[DickEE]

Just noticed that when the washer is nearly parallel to the light beam, it appears to oscillate laterally in translation, seemingly in the plane of the washer. I can't tell whether this happens during other portions of the rotation cycle. A time-varying lateral force must be causing this. Ferrous hardware in the optical fixtures could produce a lateral force. Whenever this force vector is not radial to the center of mass, rotation would be imparted. It is not sufficient for this force to vary statically with some function of the rotation angle - this would only change the equilibrium position. It must have a phase lag or lead (depending on whether it's an attractive or a repulsive force) in order to build and sustain the oscillation. The combination of two oscillation modes might have a part in this, but cannot explain it completely if the oscillation builds up from zero and continues indefinitely. To see if this hypothesis has any merit, I suggest moving the IR fixtures much further apart (if possible) and seeing whether this has any effect. If it does, try replacing any ferrous hardware in them with non-ferrous. If this stops the phenomenon, put it back the way it was - it's much more interesting this way

 

[logbook]

The problem with the "ferrous sensor" theory is that the washer is 90 degrees off in its semi-stable position. A transverse flux would align the washer with the sensor axis, not perpendicular to this axis.

 

[DickEE]

There's another experiment that should be tried. Rotate the whole apparatus 90 degrees and see whether the midpoint of the rotation moves with it. If the bolt (the electromagnet) is not exactly plumb, or part of the return flux path is through nails in the frame, then there's a bias in the field pattern which would tend to align the washer with the strongest part of the return flux path (Auburnate's Theory 1; logbook's 1st post). It would also cause a slight assymetry in the magnetization of the washer, which would make it somewhat responsive to the geomagnetic field. These biases would combine with any bias from the optics fixture hardware to collectively determine the equalibrium position, like btrueblood's torsion pendulum (1st reply). Also, is it really certain that any force between the optics fixture hardware and the washer would be attractive versus repulsive? The top of the washer is much closer to the pole piece than the point on the edge of the washer nearest the optics. The latter might be in the return flux path. The tops of the optics fixtures are near the downward pole. The result could be repulsion.

 

[RAH1234]

It stands to reason that as the coil turns on and off, the paper clip will oscillate with a pendulum motion as the hanging end is drawn, then released, by the on/off magnetic field.

The video shows the hanging end of the paper clip being drawn upward.

 

[12345guy]

Does the levitating object spin in the oposite direction if you change the polarity on the magnetic field? If so, I believe the answer lies in the historesis of the magnetic field. One or more may be happening: the object retains some magnetic field after the field is turned off and then back on again by our feedback loop. Their is no way to keep the object perfectly still in the levitating field so as the magnetic field turns on again, the moving magnet (i.e. the levitating object), receives a torque causing it to spin. The right hand rule should apply. Remember, the right hand rule applies to the flow of electrons, and electrons flow from negative to positive. Also see "Eddy currents".

 

[MaxMcPower]

Yes, 12345guy is right. This behavior is explained by Maxwell's equations. The DC current through the coil is switched on and off so it's essentially an AC current with a DC bias. Take a look at the shape of the waveform. Reverse the current and see if it spins the other way. The oscillating magnetic field causes an induced current and magnetic field in the object. You've built a great levitator, but it's also a very inefficient motor.

Since this was published in the September Design News you should expect a flurry of responses from both of the other readers.

 

[12345guy]

Thanks, MaxMcPower, for the better, clearer descripton of what I was trying to say. Not bad for a couple of ChemE's.

 

[1sailor]

Any chance the experiment has been attempted in places of different Lattitudes? Perhaps the sway has more to do with the "alternating" magnetic field in the earth's ~constant magnetic field with a little Coriolis thrown in?

 

[gcat122]

I built a similar device in '72. It was "Little Atlas" or "Lil' Atlas" from Popular Electronics. The construction article is still out there somewhere.
I love the theories, but start with the simple causes before going complex. Here are the possible keys to the motion explanation, that was much more obvious when the light receiver was a half inch wide.
1. The suspended object is going to start being moved by the slightest breeze or magnetic influence since it is a nearly friction free mounting.
2. The shadow cast by the suspended object varies as it rotates. If your light receiving area was wider you would see the washer visibly rise and fall.
3. Every time it reaches a certain angle where the shadow starts to obscure the receiver differently, you will see a change in coil current. When the shadow gets "larger" the washer would get a bump of extra pull to keep it from "falling". This causes an imperceptable rise in height and potential energy. The washer willingly drops back down this energy hill and changes direction.
4. Try a hemispherical plastic cover over the washer (1/2 a ping pong ball if IR opaque) and you can separate the optical from the magnetic influence.

Let me know how far wrong I am.

Thanks for the memories.

 

[MaxMcPower]

What could be simpler than the divergence of B equals zero? Or the curl of E equals minus the partial derivative of B with respect to time? Maxwell's equations are the essence of simplicity and they're on par with E=mc^2. What better approach to understanding the motion of a conductor in an electromagnetic field than to start with the equations that describe all electromagnetic behavior? Try it first before resorting to mysterious winds, luminiferous ethers, or shadowy humours. That would be like trying to explain the retrograde motion of Mars without something complicated like F=ma.

But where’s Nathan now? Has he learned anything about this in the last three months or is he content to collect his five hundred dollars and move on?

 

[frijoles]

It is on a fine clear fishing line!

 

[frijoles]

After looking at the video again, did you notice that the paper clip showed no angular momentum?

 

[kwing54]

here's another theory:

at the ruben h. fleet space theather science center in san diego, one of the displays is a large pendulum. the pendulum is a large brass ball at the end of about about a 4 meter filament. this pendulum will, over the course of several hours, precess around a circular pattern on the floor while swinging back and forth (forming a large flower petal pattern). the reason given for this complex motion is the earth's rotation. the rotation imparts a side-force that is higher on one side of the weight than the other, thus causing the precession motion about an inscribed circle.

the washer in question is suspended in the magmetic field, freeing it from the forces of gravity, but not those of the earth's rotation. with the earth's motion imparting vertical plane forces on sides of the washer, it begins to rotate about the vertical axis. however, the washer soon encounters weak magnetic flux line differential forces from the suspension system (probably due to the imperfections in the construction of the electromagnet and the presense of the optical sensors interfering with the lines of flux). these weak flux line forces overcome the earth's imparted rotational forces, swinging the washer back the other way. this equalibrium process is repeated at the other end of the vertical axis rotation, and the oscillation is sustained.

 

[davidbeach]

The rotation does not impart a side-force. The pendulum totally ignores the earth's rotation and continues to swing along the same plane while the earth turn underneath it.