Arguing about how many millions of years it will run for seems to be missing the forest for the trees. IMO the real message here is, holy crap magnetic bearings are amazing, now how do we get them into everything else?
By making significant improvements in the material science and manufacturing of paramagnetic and diamagnetic materials. Magnetic bearings are limited by Earnshaw's theorem [1] which states (roughly) that you can't stabilize some regular magnets using only their charge - they have to dynamically respond ti each other or they'll all just fly off. Usually this is done with electromagnets and a control system like in jet engines but its expensive and difficult. The alternative is passive para/diamagnetic materials whose magnetic fields change based on the other fields around them but those are usually rather exotic and have to be precisely engineered. Without that stabilization, the bearings would be worse that useless since they would never be able to prevent the rotor from touching the stator.
> Spinning ferromagnets (such as the Levitron) can—while spinning—magnetically levitate using only permanent ferromagnets
> Pseudo-levitation constrains the movement of the magnets usually using some form of a tether or wall. This works because the theorem shows only that there is some direction in which there will be an instability. Limiting movement in that direction allows levitation with fewer than the full 3 dimensions available for movement (note that the theorem is proven for 3 dimensions, not 1D or 2D).
We already have magnetic bearings in real life fans - the Corsair ML range of computer fans.
They're fantastic, move a lot of air and high static pressure, but the only noise that comes off them is the moving air, they're virtually silent. They also have a huge lifespan compared to fluid dynamic bearings, rifle bearings and sleeve bearings because there's no friction leading to wear on bearings.
I won't say you're wrong, the losses are small. But I guess they remind me of other technologies in the past that were by themselves incremental improvements, but as cost came down and they were put together with other incremental improvements, our capabilities advanced significantly & in ways we may not have expected.
For one, I think bearings wearing out is the main failure mode of most electric motors.
Great example, also from rotating systems, is the harmonic drive gearbox -- a very compact & high gear ratio device that (curiously) uses a deformable metal "cup" (flexspline) rather than a 100% rigid piece of metal like a traditional gear. Huge improvement that is now commonplace in robotics.
This is an opportunity to look at the problem from a different perspective.
Currently, in all mechanical designs bearing losses are small, because if they were not the design would probably be unworkable or unfeasible. But consider, maybe there are new things that could come into existence which are not feasible today because of bearing losses.
The internet, for example, didn't make it newly possible to send mail to people, but it made it so much easier that it changed what it meant to send mail.
When talking times like that, lots of 2nd-, 3rd- and 4th-order effects become important. Like flexing of the mechanism due to Coriolis forces, or coupling between the magnet and the earth's magnetic field. Heat losses that will stop the motor in years, not millennia.
It's a cool project, but (as many other commenters noticed) most of the battery's energy is being dissipated through the resistors, so it's not going to last for millennia as built.
Actually, I think it could last for millennia as built. Just not the full 89,000,000 years as mentioned. Even if you ditch the theory, there's a good chance this will keep running as built for several decades and that's pretty crazy. Makes me want to build one and let it run.
The mechanical aspect of this has nothing whatsoever to do with why it wouldn't last for millennia - it's the fact that it's coupled to a resistive load that will dissipate power from the battery much faster than the motor itself uses, and act to brake it over time as the battery voltage drops.
The electronics and heaters aboard each nearly one-ton Voyager spacecraft can operate on only 400 watts of power, or roughly one-fourth that used by an average residential home in the western United States.
A set of small thrusters provides Voyager with the capability for attitude control and trajectory correction. Each of these tiny assemblies has a thrust of only three ounces. In the absence of friction, on a level road, it would take nearly six hours to accelerate a large car up to a speed of 48 km/h (30 mph) using one of the thrusters.
The Voyager scan platform can be moved about two axes of rotation. A thumb-sized motor in the gear train drive assembly (which turns 9000 revolutions for each single revolution of the scan platform) will have rotated five million revolutions from launch through the Neptune encounter. This is equivalent to the number of automobile crankshaft revolutions during a trip of 2725 km (1700 mi), about the distance from Boston,MA to Dallas,TX.
The Voyager gyroscopes can detect spacecraft angular motion as little as one ten-thousandth of a degree.
JPL knows. BUt more importantly, when the original thrusters broke after 37 years, they were able to switch the backups and it worked fine. To me, remotely fixing a space probe beats a spinning magnetic bearing on earth for amazing engineering any day.
They are only claiming 89,000 years, not 89 million. But yeah, probably a few decades is plausible, the battery isn't going to last thousands of years.
Funny you should mention that. Hi-vac turbo-molecular pumps often use maglev bearings. This is the pump that takes a chamber from low vacuum (like 1 torr) to high vacuum (like 1e-6 torr). It's a turbine that resembles a small jet engine.
"The pitch drop experiment is a long-term experiment which measures the flow of a piece of pitch over many years. 'Pitch' is the name for any of a number of highly viscous liquids which appear solid; most commonly bitumen. At room temperature, tar pitch flows at a very low rate, taking several years to form a single drop."
At 8:15 in the video, he mentions a solar cell used while prototyping is under a "dark blue plastic box" to keep the thing from spinning out of control. There's probably not too much UV getting in there. You could probably hide the cell under the base and it would go until the printed parts crumble.
Modern high quality monocrystalline silicon PV cells, 156mm size, are commonly warranted for 80% or up to 84% of their original STC rating, after 25 years.
PV cells definitely do degrade, but in terms of spinning a nearly frictionless thing in a vacuum, if a 5W rated single cell degrades to 2.5W over 75 years of being exposed to direct sunlight inside a glass vacuum jar, it's still way more than the required energy to keep the thing spinning.
Here's an example of a 30 year warranty for a fully assembled module (60 or 72 cells encapsulated in glass and back film), at >80% after 30 years.
Mineral coloration has literally a dozen different explanations. Most of them do not "wear out", but some of them change colors under irradiation.
Green glass is green because of iron (II) oxide (FeO), which is usually a naturally occurring contaminant of the sand used for glassmaking. Amber glass is amber because of FeS2. Neither change color under normal weathering conditions.
The duration of a protective coating depends on the coating, and its thickness. If you layer a glass pane under a synthetic sapphire pane, that would probably be better than a protective coating. The sapphire would protect the glass, and the glass would block UV-A.
How much would such a coating reduce the efficiency of the panel though? I’m seriously asking btw, I don’t know the answer or very much about PV panels.
That seems to be the fantastic part of this motor. Even if you traded off almost all efficiency in order for protectiveness of the material instead, this machine runs on such a low power you could make those trade offs. What you could use this motor for I am unsure.
The energy required is so small, I wonder if you could use a thermoelectric generator utilizing some tinfoil half in the sun and half in the shade. Maybe such a design could last longer than a solar panel based one.
Too cool. Hey, could these bearings be used in something like a water wheel or windmill? I realize the technology isn’t there yet, but if it could then really you’d could have lowered maintenance costs, slightly increased efficiency, and probably a lot less noise.
Not really, actually - hydrodynamic bearings already work on the "zero wear due to zero surface contact" principle. The reason this is using magnetic bearings is just because it needs super-super low friction, under super-tiny loads, in order to work.
>Probably a lot less noise
Bearings aren't generally noisy. If they were, they wouldn't be very good bearings!
Only about 5% of the Sun's energy, as filtered by Earth's atmosphere, is ultraviolet. You can block it all with negligible efficiency loss at your solar panel.
I'm not sure about UV filters that will be stable for 1000 years. Regular glass blocks most UV, but lets through about 25% of the frequencies close to blue.
For museum and artwork protection time frames acrylic is popular. It blocks almost all UV and is still stable after 25 years. I don't know how that extrapolates to 1000 years.
The glass isn't the thing that wears out, the PV cells are. And if you're keeping light from reaching those, then you're really defeating the purpose of the whole thing.
It's using so little power that this should not be a problem. In the accompanying video it is stated that
...the most surprising thing probably is that it was being powered off a small solar cell under this dark blue plastic
box. Any more power than that and the motor would spin up, lose control and fly out of stability. , https://youtu.be/wNcgnooayDc?t=129
It's a shame when comments like these are downvoted. You're spot on, magnetic bearings certainly have a very small drag, but it's not zero. Going over that reed switch will also produce a tiny amount of drag, it's miniscule by any measure, but it's still additional drag. Even just the fact that it's a spinning magnet passing by that conductive copper is going to induce a tiny drag force.
What would be cool is a version of this but instead of permanent magnets in the rotor make it an induction motor or an electrostatic motor.
He's being downvoted because that's completely irrelevant - the thing isn't relying on just coasting for millennia. That figure's based on the actual, measured power draw of the motor, which includes the bearing friction and other drag, and dividing by the energy stored in the battery.
Are there super capacitors with expected lifetimes measured in centuries / millennia?
Making a device that could keep running long enough for civilization to fall into the dark ages, and then grow again from scratch does something funny to my brain.
> Are there super capacitors with expected lifetimes measured in centuries / millennia?
No, certainly not. Supercapacitors self-discharge significantly faster than batteries do. Capacitance is proportional to the surface area of the electrodes and inversely proportional to the distance between them. Resistance is the opposite, so roughly speaking higher capacitance means higher self-discharge.
The current go-to for long lasting batteries is lithium thionyl chloried, used in remote areas, embedded electronics and things like portable defibrillators. Those are some of the most energy-dense batteries available (though not rechargeable) and can hold most of their charge for over a half century.
The Oxford Electric Bell[1] was linked in the comments under the post, and it has been ringing since 1840. It's not known what kind of battery it has, but its some kind of dry pile. Dry piles generate voltage via corrosion of metal (eg zinc) and have extremely high resistances between plates since there is no liquid electrolyte. As long as they are kept relatively dry they have incredibly long lifespans, although their power output is miniscule- orders of magnitude smaller than even this motor.
It might make more sense to tap into an extremely long-lived source of power, such as geothermal. Over millennia tectonic shift is only a problem across a fault. As long as you can set up a thermal gradient (eg by pushing a stainless steel wire deep into a hole), you can run a Peltier (Seebeck) generator. Some semiconductors have meaningfully limited lifespans and eventually fail under normal use, but many are effectively inert and last as near to forever as we can figure. I'm pretty sure most thermoelectric junctions are the latter, with exceptions for radiation. They have a standard MTBF of just under 23 years with frequent thermal cycling, which is the most damaging thing you can do. It wouldn't be hard to imagine it lasting millennia in a sufficiently stable environment.
multimeters of this sort cannot accurately measure such low voltages and amps since they by design inject low amounts of their own battery power into the system and measure what comes back.
multimeters of this sort cannot accurately measure such low voltages and amps
That's an interesting point. I've used similar multimeters for many years and was not aware of that limitation.
But for this particular design it would be trivial to determine both the voltage and the current to within about 1% accuracy.
V = I * R
Most of the voltage drop is across the 24 megohm resistor string. Measure the voltage across that, and you get the current flow, since you know the resistance is nominally 24,000,000 ohms. Most voltmeters have an impedance in the 1000+ megohm range, so measuring across the resistors won't introduce much inaccuracy.
The resistors themselves appear to be 5% tolerance. But by measuring the resistance of the string (open circuit) with the multimeter you should be able to determine their actual value to better than 1%.
