'levitating alternator' electricity generator question

In summary: I'm not sure if you're talking about the rotor and the base or the magnets within the coils, but if you are then you might want to consider adding more magnets to the rotor. This will increase the magnetic repulsion and hopefully allow the electricity to be generated.
  • #1
hey all, i recently had an idea to combine the concepts of a magnetic bearing i.e. using the repulsion of magnets to lift a rotating object so that the object may rotate without mechanical friction , as well as an axial-flux alternator where coils of wire are exposed to a regularly fluctuating magnetic field by means of magnets placed on a rotor

here is the rotor in question
the outer magnets are all aligned the same way, and the inner magnets are all aligned the opposite way, and the 'star' is made of mild steel which acts as the magnetically permeable material, and thus the flux flows from one magnet to the next, and the effect should be somewhat similar to a 'horseshoe' magnet which zig zags around a circle ... the 'magnet face' of the star can lift a steel tool from several inches away, while the 'back face' can barely lift a nail

this is my base
the inner magnets are also all aligned in the opposite way to the outer magnets, and the large disc is also made of magnetically permeable material... but here, the magnets are supposed to give as uniform a magnetic field as possible, ideally this should have been 2 large concentric ring-shaped magnets , instead of rings of smaller magnets... the magnetic field should approximate a half torus

this is the magnetic levitation in action
the rotor is being held up by the repulsion of the magnets... at least mechanically, the magnetic field of the base is uniform enough for the rotor to spin without any 'bumps' or 'dips'... in actual use, the rotor would be loaded down and the gap between it and the base would be less

thus, it is working as a magnetic bearing, allowing rotation without mechanical friction
however, the actual purpose of the star-shaped arrangement of the rotor magnets is to cause a regularly fluctuating magnetic field, somewhat like an axial flux alternator... the rotor magnets should distort the half torus magnetic field of the base into a 6-sided star shape, and as the star rotates, stationary points along the path of the magnets should experience a fluctuating magnetic field
and in order to generate electricity from this magnetic field fluctuation, i have wound these coils:
there are 3 coils for every magnet, thin enough to fit in the lift gap, and they are glued onto a piece of paper printed with a template of the positions, so while their positioning isn't exactly perfect, they should be within a tolerable margin of error, and all the magnets should pass over a coil pretty much simultaneously ... and all the coils which do simultaneously have a magnet over them are wired in series, the inner coils are wound in the opposite direction from the outer coils

and thus, this creates 3 separate circuits of twelve coils each... one end of each circuit is joined together while the other is exposed to be connected to the electrical systems receiving the electricity output , hence this should be a 3 phase alternator, with 6 cycles for every revolution of the rotor

the problem is that this doesn't actually generate any voltage when i spin the rotor ... i am thinking that the magnetic field from the base magnets are 'shielding' the coils from the rotor magnets... where the rotor magnets arent causing enough of a perturbation in the magnetic environment of the coils to produce a measurable voltage

my idea for correcting this is to get nails which have the same length as the thickness of the coils, and with a flared butt end ... the nails would be glued in with the butt end facing upwards (in order to catch as much flux from the rotor magnets as possible) and the pointy ends downwards (to present as little ferromagnetic material in close proximity to the base magnets as possible), and so these nails will act as cores for the coils, concentrating the flux from the rotor magnets into the coils... i will lose abit of lifting force due to the rotor magnets being drawn towards the nails and risk the rotor magnets colliding with the coils, but i can always stack more magnets on the rotor to compensate with an increased amount of magnetic repulsion

does my idea have a sound physical basis? or should i just scrap everything...
Last edited by a moderator:
Engineering news on Phys.org
  • #2
I think your problem is just like you said, the opposing fields are "shielding" the coils. The only generators I've ever seen have a magnet inside the coils, not outside.
  • #3
First of all, major kudos for actually building something. I see to many people want to bury their noses in books, learning a bunch of theory and never actually using physics to engineer something. I like you.

An Idea to solve your problem may be to create more space between your 2 sets of magnets. The magnets providing the bearing lift need to provide as uniform a field as possible but the magnets providing the induction need to provide an alternating field, these are contradictory aims. Suppose you support your armature shaft horizontally instead of vertically. You would have one magnetic bearing supporting each end of a shaft. You could put your alternating magnets on a disk around the center of the shaft where they would be well away from the bearing magnets. If your shaft is made of a non-magnetic material the amount of interaction between bearing and induction magnets will be limited by the half-length of the shaft so a long enough shaft would make the interaction negligible.
  • #4
hey thanks for the replies - i really should have studied the problem a little bit more before i irreversibly glued the coils in place, now i am stuck with a pretty expensive physics problem... the design requires that the shaft is vertical, that the weight of the rotor is bore on a thurst bearing, and that the rotary motion generates electricity - hence this 'levitating alternator' that i am trying to do

any thoughts on that nails idea? I am about to go out and buy some, if i can find the right lengths...
the idea is to make the nails 'penetrate' the 'shielding' effect of the base magnets, but being not from a physics or magnetics background, i am still abit wary about the whole thing...

::edit:: I've got the nails glued in the middle of the coils, so they act like 'transformer cores'... if they did increase the magnetic fluctuation that the coils are exposed to, it still wasnt enough to exceed 1 volt total

my next idea is to scrap all the coils I've done, and do a new and neater single ring of coils, as well as buy an additional 12 magnets to create a new middle ring where the magnets are arranged in an alternating manner i.e. it will be a true alternator ... this would solve the 'shielding' effect since the coils are no longer directly above the base magnets, and it would also take advantage of the 6 sided star i have
Last edited:
  • #5
thank you mods for moving this to a more approprate section of the forums

let me elaborate a little bit about 'using the star' ... the major problem with my current configuration isn't just that the coils are too close to the magnets, but the flux from both the base and the rotor is in a direction which is horizontal and perpendicular to the circle, hence there is very little change in flux due to the movement of the rotor
i am going to be buying 12 new bar magnets to place on the straight sections of the star , where a pair of straight sections define an 'arm' of the star

so the magnets are going to be aligned in pairs , one 'arm' of the star will have south downwards, then the next arm will be north downwards, and so on
this creates flux which is tangential to the circle in the gaps between the arms, and each gap will have flux in the opposite direction from the next gap
so my coils will be wound as sections of a torus, hopefully i should finally get usable voltage out of my alternator this way

for today i really need to look for some stainless steel tools, having to force the magnets into place is hard enough , but having the entire tool be sucked into the magnets just makes trying not worth it...
  • #6
What are your voltage goals? What are the strengths of these magnets?
Are you actively sampling/taking field/flux measurements?
Have you used any sw tools for modeling analysis? Have you tried or
succeeded with a scaled down version of this rotor? Are you
familiar with the problems/sensitivities of the Tesla induction motor?
  • #7
voltage goal - 12v nominal ...
all the magnets are n45 as far as i know
i do not have the tools for flux measurements, and i don't know of any software which allows me to model it in 3d,this is the prototype, there is no scaled down version
i am familiar with the concept of an asynchronous AC motor, but i do not see how it relates to my alternator because i am using magnets and also i am not trying to build a motor...
the only relationship i can think of is pretty off topic, like how an induction motor, when spinning above a certain speed, can also be used as a generator... which ties in with certain types of wind turbines which can capture useable amounts of windpower once they reach a certain RPM
Last edited:
  • #8
Think about Lenz's Law and Faraday's Law.

Draw an animation of how the flux intersects with the coils and draw the flux pattern of the coils with an induced emf.

The induced emf is proportional to the rate of change of flux intersecting the coils and also to the angle of incidence between the flux and the coil, as well as the other physical properties of the coils (turns per metre, gauge of wire, conductivity of wire etc.). Optimise these parameters and you will maximise the induced emf.

If you don't have a background in this kind of thing it's worth noting that even though the emf is measured in Volts it is misleading to think of it as a potential difference between two arbitrary points as the emf is present at all points in the circuit intersected by the magnetic flux and created by the Lorentz force as a consequence of flux moving relative to charge carriers in the coil.Also take a look at Faraday Discs, also known as unipolar or homopolar motors and make sure you know what the direction is of the flux (think about how the fields are compressing one another - the flux is possibly radial and tangential, not axial, where the coils are...) as it intersects the coils - maybe you need to turn the coils; a bit counter-intuitive but it's a thought. But, as said before, maximum respect for going ahead and building something.
Last edited:
  • #9
For a useful and cheap graphical tool for troubleshooting/debugging check out ww.openmodelica.org or www.scicos.org - this is free download sophisticated industrial modeling tools used by Ford and others for research and design of complex electrical/mechanical systems. You'll get real numbers and can test out the effect of
your nails without having to compromise your development efforts.
  • #10

Wow! I've just spent the last hour checking out modelica libraries.

Amazing resource.
  • #11
i don't even know where to start with these types of simulation software... in particular i don't understand how the block-based system used in scisos is applicable to simulating a 3d problem like mine?
Last edited:
  • #12
heres the rotor, with some new magnets and the polarities facing towards the camera labelled... the magnets labelled with red are the 'lift' magnets, the yellow labels are the 'alternator' magnets... the four extra rectangular magnets working as lift are there to compensate for the loss of lift due to the alternator magnets
here are the coils, with their phases labelled... each coil is 200 turns, all wound the same direction, and with the same wire as my last attempt at coils
[/URL]and this is how they should have worked together
well guys, I am stumped, and that's why these pictures are online now ... when i turn the stator , and i connect 2 of the three leads to a multimeter , it cannot detect any voltage or current... i know that, because of my geometry, the resulting waveform won't be a perfect sinusoid, it won't even be symmetrical, but i was expecting some electrical activity at all due to the magnetic field fluctuating in a perpendicular direction relative to the coils... all the coils in the same phase should be exposed to the same mangnetic environment as far as the rotor is concerned

it cannot be the shielding from the base magnets anymore because the coils are no longer directly above the base magnets, and they are also between the rotor lift mangets, i can push it down so that the rotor lift magnets are below the level of the coils, and it still would read 0.00v on the multimeter
my only explanations are that , either the non-uniformity of the base magnetic field is detrimental to the alternator aspect of the arrangement by causing successive coils in each phase to cancel each other out (there are 9 inner and outer base magnets, and there are 18 coils, and the phase repeats every third coil, so every succesive coil in each phase will be exposed to the opposite magnetic environment from the base magnets compared to the last and next coil), or that there is some intrinsic property which is preventing the voltage from being generated...

i am leaning towards the second explanation, because i don't think its likely that the magnetic fields from the base magnets are causing alternate coils in each phase to cancel each other when the rotor is in motion , and even if it is cancelling each other, its even less likely that the coils will exactly cancel each other out and give exactly 0 volts ,since the magnets arent placed very precisely, and the coils are crudely wound...
Last edited by a moderator:
  • #13
i hope that proper physicists specialized in magnets come by this section of the forums often, because i have a question for them ... do magnetic flux lines tend to exit a ferromagnetic material perpendicular to its surface?

::edit:: perhaps a more general question would be , is it possible to 'shape' the magnetic field with specially shaped pieces of ferromagnetic material, so that there is a specific region which is devoid of magnetic flux

or is that only possible by using additional magnets to directly influence the magnetic field?

what i am trying to do now is to work by taking my existing arrangement of magnets as a constraint, and cause my coils to have maximum magnetic fluctuation from the rotation of the rotor
Last edited:
  • #14
i have tested the alternator by using the short circuit detection system of a multimeter, and there is a periodic 'break' in the circuit (due to current going the opposite way of the multimeter's test current) as the rotor is turned... the periodic nature is consistent with fluctuation of magnetic field experienced by the coils, but the minute amount of electricity generated is characteristic of the shielding problem which still exists

i think i have a way of 'shielding' the coils from the shielding of the base magnets (abit of a convoluted concept) to allow the rotor magnets to influence them more

first, i will be getting a large ring of mild steel which goes over the circle of the large outer magnets, this should 'blend' out the irregularities in the magnetic field which was causing the rotor to 'cog' i.e. there are bumps and pits in the magnetic repulsion due to the gap between the large magnets
then i am going to get a disc-shaped aluminum plate which completely covers the area of the base, and it goes on top of this mild steel ring... on the aluminum plate will be a smaller ring of mild steel as well as 9 bar magnets, forming a 'wall' shielding a region of space above and between the magnets... the coils will then be placed in this region of reduced magnetic field

here are 2 pictures I've attached... they show a semi-accurate 'cross section' of the magnetic environment representing the radial components of the magnetic field, and they are showing the left half of the alternator i.e. the left side is the outer side, the right side is the inner side... one picture features the 'ring wall' like i have described, and the other without to show the current magnetic environment that the coils are exposed to

it seems that if i replace the magnet with just a simple piece of mild steel, the magnetic void becomes more intense... i have attached a third picture, with the same colour notations , red = mild steel, light blue = aluminum ,but with a new 'wall' position, to show the simulated effect... i forgot which contrast setting i used so pardon the irregularity of the colours, but you can see that the mild steel rectangle in the middle has a much higher concentration of flux relative to the magnet... all the magnets were specified as '1 Tesla' , except the lower right one, which has a value twice as big to approximate how close the inner magnets are to each other

therefore , it seems, i need 3 mild steel rings and one aluminum disc... i do not have the capability to simulate the entire system in 3d, but i hope that the mild steel rings will , in addition to shielding from the radial component of the magnetic fields, also act in an analogous way to the ferromagnetic cores in transformers, i.e. attract and concentrate the magnetic flux lines through the coils... in this case i am hoping that the rings will amplify the tangential component of the magnetic fields


  • magnets.jpg
    95.2 KB · Views: 548
  • magnets2.jpg
    92.1 KB · Views: 589
  • magnets3.jpg
    92 KB · Views: 532
Last edited:
  • #15
also, it seems that this is one time when the presence of ferromagnetic cores in the coils will be beneficial

attached are 2 pictures...

one is a 'top down' approximation of a section of the rotor, the upper side represents the outer ring of magnets and lower side represents the inner ring of magnets, and how it interacts with discrete pieces of ferromagnetic material placed between them... you can see that the cores make up a 3 phase system, although you can see the simulation isn't perfect because its not an actual ring
also note that the new small rectangular magnets are left out of the simulation- it should be possible to use the lifting magnets by themselves, without any specifically alternating magnet arrangements, to cause a good fluctuation in the magnetic fields experienced by the coils, and generate a voltage that way

the other is a 'side' profile approximation showing the alternating magnets of the rotor at the top, the coil cores in the middle, and the 'shield-shielding ring' at the bottom... as you can see the alternation has negligible effect on the continuous piece of ferromagnetic material at the bottom , and it is mostly on the left and right edges, so if this was an actual ring there should be no magnetic activity in italso i have omitted the 'wall' because it would have also blocked the coils from the influence of the rotor magnetsso to recap... the 'shield-shielding' ring will take care of the radial flux, while the ferromagnetic cores will concentrate the alternating tangential flux into the coils

anyone have anything to say?


  • magnets4.jpg
    96.9 KB · Views: 584
  • magnets5.jpg
    95.9 KB · Views: 523
Last edited:
  • #16
simulating different shapes of the middle mild steel ring

the ring is laser cut from a flat sheet of metal, so it will look like a rectangle in its cross section

but i can use a Dremel to file down the edges and create a slope, and the top of the alternator's rotor can be used as a platform for the ring to spin around as i work on it, so that the slope can be made uniform across the whole ring

the positions of the magnets have also been updated to match their radius positions on the base and the rotor, the aluminum plate has been omitted, but i have moved the middle ring to the bottom side of the aluminum plate

due to the inner diameter of the ring, if it is left in its original rectangular profile, it creates a large interference on the repulsion , especially of the inner magnets, and there will be a siginificant drop in the lifting ability
so by creating a slope which 'points at' the lower inner magnet (because that magnet is the closest), it presents the least surface area for the flux lines to enter, and it no longer causes a major interference to the repulstion

if i also slope the outer edge of the ring and give it a triangle cross section, it will also capture less flux from the rotor as you can see from the higher flux density in upper region... as the rotor's magnetic field is alternating, this should mean that the coils are exposed to more fluctuation from the rotor

the precise shape of the triangle can almost certainly be optimized to give more lift and more power output, but because i do not have the tools or the means to precisely work on the ring, i will just keep it simple and make it symmetrical


  • magnets6.jpg
    92 KB · Views: 541
  • magnets7.jpg
    92.6 KB · Views: 573
  • magnets8.jpg
    93 KB · Views: 530
  • #17
this is going to work, right?
  • #18
well i was having a bit of a 'due diligence' moment and i found this video

clearly this 'class' of alternators is feasible (also i am working on a wind turbine as well) , its just the little details that make it matter... like are they actually using a dedicated set of magnets for levitation, and have a separate alternator set of magnets

the big question is if it is possible to use the levitation magnets as the alternator magnets as well, in a feasible way...the efficiencies in terms of mechanics (less cogging, adequate support to stop vibrations/high order movement, etc) , magnetics (eddy currents in any ferromagnetic and conductive pieces used to shape the magnetic flux) , and electrics (turns vs coil resistance , voltage relationship between the phases) will have to be worked on

in my case i have been told that the harmonics in the electricity will be the biggest problem , which seems obvious due to the discrete pieces of magnets on the base, i am trying to compensate for it by using less poles, and then compensating for that by using more turns per coil
Last edited:
  • #19
as an aside, i have installed the outer mild steel ring, and it really does smooth out the magnetic pattern... a negligible loss in lift, but virtually eliminated all the cogging
  • #20
middle ring has been fabricated
Last edited by a moderator:
  • #21
finally i got an instrument responsive enough to directly measure the AC voltage, as well as sensitive enough to respond to the fractions of a volt that each of the 3 phases is giving out at angular velocities which the rotor will spin at

So now i am thinking of switching to a 1-phase design , basically a coil for every magnet on the rotor
because i am using a ferromagnetic core for the coils, i think that the distance from the ends of the ferromagnetic core to the magnets on the rotor is related to the magnetic flux, and thus the voltage, by a power of three ... hence by using longer ferromagnetic coil cores which span all the way between magnets of opposite poles , i am hoping to be able to raise the voltage to usable levels
  • #22
I don't understand this enough to comment on the specifics, but I hope it works for you!
  • #23
i guess a simple question i could ask ... is what i said about the distance and the flux density being related by a power of three , true?
  • #24
carmatic said:
i guess a simple question i could ask ... is what i said about the distance and the flux density being related by a power of three , true?

I *think* so, but I'm not sure. Maybe someone else would know.
  • #25
i have measured the output at ~2 volts when I am turning it at the rpm's i expect it to work at, and it goes up to around 5 volts if i turn it as fast as my hands can turn them... this is measured by 'eyeballing' an analogue voltmeter as i turn it, i should probably get a rectifier bridge and a capacitor to get a true measurement

well today i have rearranged the magnets, the central alternating magnets have been removed and incorporated into the 'lift' magnets
my coils are hot-glued into position... in my previous 2-volt arrangement , the coils stayed in place, but with the new all-lift arrangement, the magnetic pull on the nails i use as coil cores was actually strong enough to rip a coil loose... this should mean a higher magnetic flux though the nail, and therefore a higher voltage
i think what was happening was the middle ring of alternating magnets were interfering with and 'diluting' the overall magnetic pattern... and now that all the magnets are in the 'vertex' positions of the 6-sided star, the magnetic field is more focused spatially... the magnetic fields are also focused on the ends of the nails, so when these two 'focii' meet, there is a large magnetic flux going through the nails, and the resulting force and torque was enough to break the coils free of the hot glue
  • #26
3 volts at the RPM's I'm working with ... its not going to solve any energy crises, but i am already thinking of trying a 'version 2' of the alternator
  • #27
Sweet. Hope it goes well!
  • #28
an interim update
  • #29
Was there supposed to be sound?
  • #30
no, it is a silent recording... the only sounds you would have heard was a scraping sound due to the slip ring i was trying to build which was going to send power up to the barrels to power some decorative lights, and the air conditioning in the background anyway...
Last edited:
  • #31
Ah ok. I didn't know if there was some dialogue or something. :)
  • #32
a further update
the counter rotation is now a lot more stable

a short description ... this is a 5 phase x 3 poles = 15 coils alternator, with the upper (primary) rotor acting as 2 phases and the bottom (secondary) rotor acting as 3 phases, thus having 6 and 9 pairs of magnets respectively ... you can get a better view here https://www.facebook.com/media/set/?set=a.10150196515871197.341138.502061196&type=1

how it works is ... picture a triangle (3 phases) rotating at a rate of 2x, and a line (2 phases) rotating in the opposite direction at the rate of 3x , they will overlap at 5 regular points (a pentagon)
the 'T' shaped elements are ferromagnetic, they will focus the magnetic flux so that a magnet on one side of the element will greatly repel the magnet on the other side... this would be equivalent the points midway between the vertices of the pentagon
hence, the magnets will be forced to overlap where the edges of the coils are, and the flux will flow through the coils

this design can be applied to other things...

as an example, an ocean wave energy buoy ... rather than relying on the vertical reciprocal motion of the entire buoy, a new buoy design could use flippers extending outside the main hull of the buoy

these flippers are connected by gears and one way bearings to a counter-rotating flywheel setup analogous to what i have built here... the flywheel which rotates slower will have more mass, so that both flywheels will have equal angular momentum, but in opposite directions so that their gyroscopic reactions will cancel out and keeping the buoy pointed upwards

as the buoy tips side to side on the waves, the flippers will actuate and the flywheels will spin, causing a gyroscopic stability which resists the tipping motion and the buoy will stay straighter... this in turn causes the flippers to actuate over a greater angle, causing the flywheels to spin faster, and creating a greater gyroscopic effect, which in turn makes the flippers actuate over an even larger angle, etc etc... the limit is when the buoy only tips minimally on the waves
  • #33
I can't say I understand the mechanics behind it all, but it looks like a pretty cool idea! I'm glad it's coming along well.
  • #34
it can be mathematically described as such

suppose that a complete circle is described as the number 1

a line, representing 2 phases, would have nodes at 0 and 0.5

a triangle, representing 3 phases, would have nodes at 0 , 1/3 and 2/3start off with both the line and the triangle's nodes overlapping in the 0 position ... suppose that the line rotates clockwise, and the triangle rotates counter clockwise

if for every 3x traversed by the line, the triangle traverses -2x , the next pair of nodes to overlap would be the 0.5 node of the line with the 1/3 node of the triangle...
and they shall overlap at the 0.4 position ... the line's node would have traveled by 0.1 , while the triangle's node would have traveled by -0.0666666...

0.1/-0.6666666... = 3/-2
so it satisfies the ratio

repeat this, and your overlap points would be 0.4, 0.8 , 0.2, 0.6 , and 0 ... it would trace out a 5 sided star , hence the pentagonin the case of what I've built, I've repeated it 3 times around the circle, so instead of a line, a triangle and a pentagon, you would have a hexagon, a nonagon and a pentadecagon
  • #35
about the inner/outer rings, and the tangential arrangement of coils... in the above post, i talked about nodes on a circle... if these nodes are magnets with the same orientation, then there will be an equivalent set of magnetic poles facing the opposite direction, in the middle between each node...
for my purpose of producing maximum repulsion as in a magnetic bearing, these opposite pole, inter-nodal positions are taken up by actual magnets which form a concentric ring, resulting in the 12 sided star shape of the primary rotor and the 18-magnet arrangement of the secondary rotor
more importantly, there will also be twice as many overlap points for the magnets: the original nodal or 'like' overlaps, and an equivalent set of 'opposite' overlaps which exist between each nodal overlap... accordingly, there would be 30 overlap points on the coil ring

the coils are arranged tangentially, such that their 'ends' (where flux enters and exits the coils) coincide with the 'nodal' overlaps, where 2 magnets of like polarity are in the closest proximity, forming the highest flux density at that point... this flux will then be focused through the nails which form the core of the coils, before coming out the other side of the coil and into a region of almost equally high flux density where an overlap of the opposite polarity is either about to occur, or has just occurred (since only one of the 5 phase overlaps occur at any time)

the middle of each coil would therefore be equivalent of the 'inter-nodal' overlap, where the magnets on each side have the opposite polarity... or, as seen from my setup, where a magnet would overlap with a gap on the other side... this is when the magnetic flux does not do any useful work because it goes straight from the primary rotor to the secondary rotor without going through the coils
the T-shaped elements are used here to assist and control the overlapping, as a magnet approaches one side of the element, the flux would be focused through it and emerge on the other side, thereby repelling the magnet on that side, and ensuring that the magnets only overlap at the edges of the coils

Similar threads

  • General Engineering
  • Electromagnetism
  • Electrical Engineering
  • General Engineering
  • Electromagnetism
  • General Engineering