'levitating alternator' electricity generator question

[Drakkith]

What benefits does your invention have over other options? Not a critique, just curious as to what it might be able to do.

 

[carmatic]

to be honest, i do not even know if it will work... it conceivably does, but it depends on the material properties of basically everything i have used, but especially the magnetic susceptibilities and permeabilities of the coil layer components, which i have not measured or tested

but basically , this is a low windspeed Savonius design which can work where the wind is not strong or stable enough for traditional propeller-like turbines to work...
a major point is the magnetic levitation, the only sealed mantainence-free bearings i could find are radial bearings , which carry load perpendicular to the axis of rotation, basically what i use to horizontally locate the rotors... thrust bearings, which would carry the load in the direction of the axis of rotation, are not sealed and they would get contaminated by dust , or corroded by water etc
a high quality thrust bearing which is large enough to support the weight of the barrels would also cost approximately as much as the magnets i am using, while a cheap 'lazy susan' style bearing has a lot of friction losses and is noisy, they would need to be replaced often, but the magnetic levitation is truly mantainence free ,and has perfect efficiency save for eddy currents in all the metallic components other than the coil wires, as well as unwanted flexing and movement of the paramegnetic elements

and if the winds do happen to be strong enough to be beyond the operating envelope of the Savonius turbine, you could swap the barrels for a professionally made Darrieus turbine which would spin at a higher RPM and extract more energy out of the wind, but this would lose the homemade aspect of it... unless someone knows how to make a homemade Darrieus turbine?

i wouldn't call this an 'invention' , it is a 'development'

 

[carmatic]

if you have studied the pictures, you would have noticed that the coil layer is itself levitating above the secondary rotor by an inner ring of rectangular magnets ... these same magnets draw the coil layer towards the primary rotor, and thus there is a spacer between the coil layer and the primary rotor bearing which keeps them apart

i have spent the whole day today reducing the thickness of this spacer, keeping it perfectly flat... because here , once the magnets are compressed by the weight of the windmill, we are looking at barely millimeters of clearance between the magnets and the coil layer's planar elements, and even the slightest tilt is enough to cause the coil layer's outer edge to collide with the magnets ...

however, once the coil layer has been brought close enough to the primary rotor, there is significant cogging... undoubtedly this is the effect of the magnets being attracted to the T-shaped elements,but i am hoping that some of the torque also comes from the flux going through the coils like i intended them to

recall that the primary rotor acts with 2 phases... the inner and outer ring each form a hexagon , and magnets 1,3,5 of the hexagon would be overlapping with the center of coils number 1,6,11 (hence axially polarizing their T-shaped elements) at the same time as magnets 2,4,6 would be overlapping with the space between coils 3-4 , 8-9 and 13-14 (hence tangentially polarizing those pairs of coils in opposite directions)... the design will be a success if the latter effect is more significant than the former effect

 

[carmatic]

the usage of the nails is not ideal because they have some ferromagnetic property, and the remnance and coercivity would cause a hysterisis loop which would be a source of energy loss, also their round cross section would encourage eddy currents to flow within them... however they are more easily polarizable lengthwise, and their arrangement shoud mean that the coil cores act with anisotropic permeability, where flux follows the direction of the nails

what are other magnetic parameters of the materials i should know about to properly design this?

 

[carmatic]

https://www.youtube.com/watch?v=S-MFvldaizA 5 phase output , rectified and connected in parallel to 4.7 μfarads of capacitance
as i have wrote in the Youtube description, the voltage readout on the voltmeter isn't the full story, because when i was touching the positive and negative output leads, i felt an electric shock... the only clue is that the 5 AC wires going to the rectifier have exposed solder joints, and by shorting 2 of the phases together, the voltage jumps to ~12v at the same rate of rotation, falls back to 3V , then jumps again when the short is removed... this only works either between neighbouring phases (1-2, 2-3 , etc), or phases one apart (1-3, 2-4, ...) , i don't know yet until i find out which AC lead is which phase ... and by continuously shorting and un-shorting the 2 phases, the high voltage can be mantained

the only explanation i can give is inductance in the coils giving voltage spikes during the current spike at the instant of the short, and the voltage then slowly drains from the capacitors... the coils have around 400 turns each, and one end of each coil basically meets the other end of the next coil... this would make them sensitive to each other's inductive effects, even at low currents...

i plan to mitigate this effect by reducing the stationary coil-layer magnets which were preventing the coil layer from 'collapsing' onto the secondary rotor... hopefully by bringing the coils closer to the magnet, there would be more influence from the alternating magnetic field compared to the inductive effect

 

[carmatic]

an update
i have converted my rectifier into a 3-stage voltage multiplier:
https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-snc7/485865_10150964175476197_502061196_11701521_952416425_n.jpg
as per http://www.celnav.de/hv/hv9.htm, it is the floating center Y design with 5 phases instead of three (notice the 3 'rectifier pentagrams')

this is just a straight voltage multiplication that makes no use of the inductive properties of the alternator, but the high voltage output should enable some form of automated shorting between the phases to drive the inductive voltage spiking, and by consistently shorting it only at the inter-phase voltage maxima, and unshorting it only at the voltage minima , more energy should be able to be harnessed, as opposed to wasted as current flow dissipating into heat

i think the way it works is that every coil is induced by the rotation of the magnets to an opposite EMF to the adjacent coils
hence , say that phase A and B , at an instant, are nearing their positive and negative minima respectively, hence develop a maximum voltage difference between them
upon shorting them at this instant, the sudden resistance drop causes and equally sudden current spike, which causes a strong magnetic field (which opposes the rotation of the magnets) to be suddenly formed...
the sudden increase in magnetic flux would induce the neighbouring coils/phases, E and C, to have an EMF which makes current in them flow the opposite way, and since the coils are all wound identically, this causes them to develop a stronger negative and positive voltage , respectively, and this is what is measured as the voltage spike

during the high current in A and B , the EMF which they induce would tend to oppose their own currents , but this would be somewhat countered by the still-changing magnetic field caused by the rotating magnets, hence keeping the current flowing longer and more magnetic energy to be transferred to E and C ... this would be the primary way of harnessing more energy out of the rotation of the magnets
suppose that the coils are induced in an 'alphabet-wise' manner, C would next be induced to create the next high-voltage pair between B and C , so the current that it has set up will start opposing opposing the rotation of the magnets, while E will be in a region of no magnetic induction, so its energy will simply contribute to the electrical output

and when it comes time to un-short the phases, the opposite would happen, the 'braking' effect on A and B ceases, but not only that, the sudden drop in current in A and B would induce C to have currents of its own which actually drives the rotor to spin faster, wasting the stored energy that way ...
hence, it is best to un-short the phases when the voltage difference between phases reaches the minimum within the cycle

if the time frame of the short-induction is quicker than that of the magnetic induction, then the short can be triggered every cycle ... but if it is slower, then the short should not be triggered until the magnetic induction comes back into effect
in any case, this would need to be electronically controlled, using electronics powered by boosted voltage out of the voltage multiplier, maybe a 555 chip along with diodes and transistors
then again, I am not really sure about any of what I've just said...

 

[carmatic]

there was abit of wind today so i decided to do some wind testing...it was dissapointing, partly because my voltage regulator for powering 5V usb devices only actually outputs 4.5v , and in any case i don't have a lot of things to plug into it anyway... but the main reason is that it seems that there needs to be more wind than i expect to overcome the cogging
due to the flux coupling between the main and counter-rotating rotor via the coils' iron cores, cogging is inevitable... here is a source of inefficiency of a levitating alternator, as the rapid changes in angular momentum would send energy via the imperfect mass, geometric and magnetic balance of the rotors into the imperfectly rigid support to be absorbed...
the cogging can be mitigated at high angular velocities by increasing the angular inertia of the counter-rotating rotor i.e. by adding weights to the outside perimeter of the rotor... since both the rotors' movements are coupled, the smoother the movement of the counter-rotator, the smoother too the movement of the entire wind turbine...
in effect it is 'dampening' the cogging, the reason that the lower rotor counter-rotates is because it is constantly settling into the lowest mechanical and magnetic flux energy state, and this is the same reason why cogging happens... but if the rotor has increased kinetic energy (due to increased mass) , the energy difference due to magnetic flux will be less significant in its movements... if you slow down the secondary rotor, it will cause a torque to slow down the primary rotor, and vice versa if you speed it up... so during a cog, the added angular inertia of the secondary rotor would cancel out more of the sudden speedups and slowdowns... this is also why the wind turbine can 'work through' the cogging by spinning faster, but having a heavier secondary rotor will reduce the post-cogging RPM, and reduce energy loss to structural vibrations

but adding more angular inertia isn't going to help if you are starting from no angular momentum, and the cogging is simply stopping your turbine from turning in the first palce... this would require a fundamental rework of the coupling between the primary and secondary rotors
the lowest energy, least-repulsion state when you overlay a square (the top rotor's 12 magnets divided by 3) over a pentagon (5 phases, 15 coils divided by 3) and a hexagon (the lower rotor's 18 magnets divided by 3) , in both cases, is when the square sits upright... in effect , the wind turbine would tend to 'drag' the alternator and the lower rotor together in circles, this was why i had to tie the alternator's wire to another spike in the ground
the square's vertices would be at 45, 135, 225 and 315 degrees, and the vertices of the pentagon will be at 0, 72, 144, 216, and 288 degrees... since we are trying to minimize the cogging between the primary rotor and the stator, they are what we will focus on
as you can see, the lower corners of both polygons are the closest to each other, with a difference of 9 degrees, hence provide the shortest pathway out of the cogging
i am thinking of splitting each of the pentagon's vertices into a pair of vertices, offset +9 and -9 degrees from the original vertex, basically having 2 pentagons offset by 18 degrees ...the coupling works by the attraction between the pentagon and the square, so by spreading the attraction to cover the gap between the primary rotor and stator, maybe the cogging can be reduced ... basically, on each coil, rather than the single iron protrusion in the middle as i have now (see https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-ash3/561840_10150877241271197_502061196_11576004_1032166382_n.jpg ), there should be a pair of protrusions at 3 / 8ths and 5 / 8ths of the length of the coil

 

[carmatic]

more ideas
in the current form of having an element in the middle of each coil , this is analogous to a single pentagon...

when you have a pair of points rotating concentrically over the pentagon, they will overlap with a vertex of the pentagon 10 times
therefore, there are 10 'peaks' and 'troughs' in the attraction between the pentagon and the rotating point pair ... this causes the cogging and it will require energy to overcome a 'peak'...
by splitting the pentagon into a pair of pentagons offset by 18 degrees, the resulting 'peaks' and 'troughs' which happen with a single pair of concentrically rotating points will be also offset from each other by 18 degrees ... but, the overlap point are 36 degrees apart (10 overlaps per rotation) , so when one set of overlap points is at their 'peaks' , the other set would be halfway between their own 'peaks' i.e. in their 'troughs' ... hence they should cancel out

however, I am not sure how this will affect the meshing between the primary and secondary rotors... the offset angle may need to be reduced from 18 degrees if the offset also cancels out the coupling between the rotors, or more importantly, cancels out the magnetic flux entering the coils

 

[carmatic]

today i have finally put together the ferrite-core rebuild of the alternator , where the only paramagnetic elements are the ferrite cores themselves... without the 'control elements' to ensure the counter rotation of the secondary rotor, the cogging has practically entirely dissapeared, it is turning much smoother than i anticipated ... more surprisingly, the counter rotating action is still there, solely by the influence of the ferrite cores... i have also used less turns with a thicker gauge wire, since the previous attempt had too much internal resistance to provide any power, and the higher material permeability and the larger volume of the cores should make up for the reduced number of turns
https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-snc7/425525_10151056090681197_544523935_n.jpg
https://fbcdn-sphotos-a.akamaihd.net/hphotos-ak-ash3/545120_10151062905946197_1457443095_n.jpg
hopefully i will get to test the power output in the next few days

 

[Drakkith]

Good luck with the test!