Part 3. Axial Flux Alternator Stator Assembly and application options

Рет қаралды 43,118

7 жыл бұрын

Detailed assembly of the stator for the electric bike integrated axial flux alternator. Scaling options for alternative applications.

Пікірлер: 63

@TheBillythekid2010 7 жыл бұрын

I totally agree with no gaps in the coils. And what you've shown with your layout and your construction of a gapless phases are impressive and mathematical equations are accepted. thanks for sharing and application will be interesting.

@YesEinstein 7 жыл бұрын

Thanks for your valued comment. Good luck with your project. More should follow; time and space permitting.

@trevortrevortsr2 7 жыл бұрын

It is such a delight to hear someone talking articulately

@gman101185 4 жыл бұрын

lovin that polka

@YesEinstein 4 жыл бұрын

Join the club!

@noutram1000 2 жыл бұрын

'tis a thing of beauty -I wish Faraday could see it....

@YesEinstein 2 жыл бұрын

Paradox should he not so behold?

@Made2hack 6 жыл бұрын

Hi YE, You mentioned creating the litz wire assembly yourself using 0.1mm wire. Are these individual strands that you then roll together to form a litz assembly? Or are they pre-bundled litz magnet wire that you sometimes see used elsewhere? (I believe in radio)? Also, any book or links for information regarding calculations and design specifications?

@YesEinstein 6 жыл бұрын

Dear Made 2 Hack, Yes, with aid of a hand drill; available prebundled lengths too costly for one-off windings. Radio frequency Litz has very fine sub wires to avoid skin-depth effect, unnecessary in our case. Search "Litz wire types" for information on bundling. Part 1 video appends my expression for losses assuming perfect bundling. Regards, YE

@NicozStrat 7 жыл бұрын

Hi man, can you explain me please how the overlapped winding works? If you could give me a name of good electromechanical built book i'll appreciate

@TheBaconWizard 6 жыл бұрын

This looks excellent. I was wondering what gauge the litz wire is you've used here?

@YesEinstein 6 жыл бұрын

Dear The Bacon Wizard, Windings in the video were four bundles of fifty 0.1mm solderable enamelled copper wire; each bundle twisted to about 2 turns per magnet length. The bundles were similarly twisted together and tightly bound, opposite-hand, with fine polyester thread. Fine wire minimises bulk eddy loss. Double bundling helps equalise simultaneously generated voltage on all wires for minimum end-effect ohmic loss. Ideally, the sub bundles would be serially connected at their phase ends, in this case for four tmes the effective passes around the machine and minimum end effect loss. Regards, YE

@TheBaconWizard 6 жыл бұрын

Thank you very much for a detailed reply, I intend to put this to good use!

@mazekbatani8148 2 жыл бұрын

Thanks for your informative video. What configuration did you use for connecting the 3 phase wires?

@YesEinstein 2 жыл бұрын

Dear Maz Ekbatani Simplest option is basic Star. Sets of 4 pins are wired together to form successive start points and common end of the three phases. A 3 phase rectifier (6 diodes) produces the output. Regards YE

@mazekbatani8148 2 жыл бұрын

@@YesEinstein thank you very much.

@robmad7657 6 жыл бұрын

Where can we find that drawing or how to make it? I asume it has a relation with magnet dimensions, also I noticed it has 15 degrees separation. Regards

@YesEinstein 6 жыл бұрын

Dear Roberto Madriz, It may be drawn on Photoshop with sufficient accuracy. Magnets are 30x10x(5)mm, pitched at 15 linear mm for ~360 mm inside circumference. 3, phase Radii drawn per pitch. Other circles indicate wiring limits and attachment holes. Printed to unity scale on an HP A3 printer. Regards YE

@natashamcgill1041 3 жыл бұрын

I am learning to design and build AC axial flux alternators for green energy. And part of my focus is energy densification. I am totally intrigued by your stator winding scheme. Unfortunately, I cannot for some reason find any really decent books on stator windings that include your style shown. Can you referance books for me? As many as you please, if you have more than one. Also, can you describe to me how to create the litz wire? From start to end? Every litz I have created so far, wants to twist itself up into really nasty bindings and pretty much ruins itself. And last of all, is there a way to scale this down to be a 12V alternator at lower speeds? Say, for charging a car battery in emergency situations? If you can respond, I would be soooo grateful, that I would reference you as my inspiration when I finalize a product!

@YesEinstein 3 жыл бұрын

Dear Natasha McGill, In the context I have only read, and referenced earlier, Professor Carter's 1954 book, The Electromagnetic Field In Its Engineering Aspects. I have since read some of the many articles on the internet, so far with no obvious conflict. The Fem2 field analysis program is freely obtainable. Low spec Litz needed practice; try twisting and folding under tension. I have already indicated scaling options. In view of your and similar comments maybe I have to video a method of low speed high power design. Automatic efficient matching of one machine to different speeds and varying loads was a feature of the early bike series and can be developed for the permanent magnet machine. Thanks for the kind offer! Regards YE

@natashamcgill1041 3 жыл бұрын

@@YesEinstein I would most definitely would appreciate a video of a low speed high output design to be shown. So far my focus seems to be energy identification. Aka overlapping stators. Concentrated coil or wave wound. Also, I am focusing on extraction of the most amount of emf via reducing the thickness of the air gap. And a diameter of a decimeter also. I don't know why , but, I am really stuck on creating a design of 10 cm diameter, with minimum 8 poles, and up to 16 coils double layered, or even tripple layered. Or windings like yours. And with a performance high enough to charge a couple of 18650 batteries . And by wind also. So about maybe 900 to 1200 rpm. It's a challenge no less lol.I have had some rudimentary successes so far. Ultimately , I want to create a stator winding that basically has inductors in every single mm for maximum flux capture. This endeavor I have undertaken is most difficult lol. I do say, I really really admire your stator . Its so neat and tidy.

@jackmahkimetas8694 7 жыл бұрын

How is cooling achieved with the windings encased in all that resin?

@YesEinstein 7 жыл бұрын

Dear Greg. This is a major consideration. Primary defence is to minimise the only dissipation (which is in the stator) when the machine is operated at its 1 KW rated output at 6000 RPM; hence accent on figure of merit (FOM) in the Part 2 and 3 videos. The stator is only 3mm thick including 2mm plus of wire within a 5mm total gap. My guess and observation is that the indicated loss of 20-30 W comprising copper ohmic and eddy should be well dissipated by windage. It would be interesting if some kind soul did an analysis based on this type of machine. The loss is basicly proportional to the machine's output power squared and inversely to RPM squared. Attempt to take high power from this size of machine at low RPM would bring certain disaster. This cannot happen in the bike application where engine power is limited at low speed.

@natashamcgill1041 3 жыл бұрын

Would you get the same amount of power out of your design if you used regular magnet wire?. Say, about 10 strands @ 0.5 mm magnet wire instead of the litz?

@YesEinstein 3 жыл бұрын

Dear Natasha McGill. Thanks for the progress report. About 10 strands of 0.5mm per magnet crossing would maintain the baseline output voltage and winding resistance. Calculation indicates however an additional 200 Watt copper eddy loss at 6000rpm. Operation at 1000 rpm should give 5Volt and 100Watt at 90% efficiency. Around 5Watt eddy loss may remain. If required, finer strands with more bunch crossings would elevate the voltage while further eliminating eddy loss. Note the baseline winding can be connected, essentially this way, for 1, 2 or 4 times output voltage without efficiency loss. Analysis has shown that the gap between rotors should be about the same as a magnet depth; 5mm in the baseline. Design should pack maximum possible copper into this space. Note that wiring should ensure, as far as possible, all strands develop the identical voltage! Regards YE

@RuddyHerdiansyah 4 жыл бұрын

intersting tutorial, i have 3 questions about the axial flux permanent magnets... #1 how much length do we need from coil winding, between magnet to magnet?? 2# how to calculate voltage output and ampere? 3# is there a book to learn for all of this?? thankyou.. and your video is a great.

@YesEinstein 4 жыл бұрын

Dear Ruddy, thanks for your comment. More of an unvetted exposition. #1 Overhang outside the magnets is minimised by close magnet spacing. i.e. lots around the circumference. Too close and flux leakage negates the benefit. #2 Voltage:- Flux density times length under the flux times speed per crossing per wire. Current:- limited by tolerable dissipation. e.g. If the resistance of the load is less than the winding resistance more than half the input will be spent heating the windings. #3 For basic and applied theory, I have already referenced my "bible". There are lots of worthy internet reports about "coreless axial flux".

@TheOne-jg7fd 2 жыл бұрын

@@YesEinstein nice

@natashamcgill1041 2 жыл бұрын

how would you make a jig to make this kind of winding with out gluing to paper? I have a 3d printer and cant figure out a simpler way to create this coil

@YesEinstein 2 жыл бұрын

A more sophisticated one-off assembly method might be developed. Evenly attach 72 (24x3) thin rectangular “spokes” to the initial fibreglass sheet, between magnet inside and outside radii and up to the magnet thickness. Attach two solid rings nominally twice the winding depth leaving radial space for the winding arcs and open hub. Successively wind each phase, starts two slots apart. Given a 3D printer, the frame might be fabricated in one piece adding encapsulant and top sheet after winding. Regards YE

@acesias 3 ай бұрын

I am fascinated by this design. Thank you for your very good explanation. Is it ok if, instead of 30x10x5mm, I should try to use 30x10x10mm magnets, for a greater magnetic flux between coils?

@YesEinstein 3 ай бұрын

Yes. See my Parts 4 and 4A. Meant something like 3 times rotor diameter and magnet volume for acceptable output power and and efficiency at the lower speed of a directly coupled wind turbine.

@acesias 3 ай бұрын

@@YesEinstein Hello and thank you for your response. I want to try to build one. I had to choose between 40x20x5mm and 40x10x10mm magnets, both N52 with the same price. I do not know if I chose corect, but I will get 40x10x10mm, because of greater magnetic field at 1mm distance from the magnet (4693 gauss vs 2335 gauss). I will use 24 magnets on each rotor, pitched at 15 linear mm. For the coils, I will use 40 twisted 0.12 enameled wire to form one strand, that will be equivalent to 0.759 mm wire. Then, I will twist 40 of those strands, so they will have a diameter of approx. 4.8mm. Two strands will be put in parallel (for an equivalent of 1.073mm wire) and 20 of those in series for a total of 20 turns per coil leg. I do not know the math behind, I hope this will work. If you have time, please let me know if it is ok. Thank you again for the help!

@YesEinstein 3 ай бұрын

@@acesias Suggestions: Decide number of turns needed. Reference should be around 1.5 v per magnet crossing at 6000 rpm. Note bundle width geometrically limited to 5mm per phase. Optimum axial gap also looks around this value when filled by copper. Litz-type windings ideally need to ensure every sub filament crosses the same net flux and has the same impedance. A good composition might be four wound sub bundles each comprising three sub subs for 1, 2 or 4 x connection options. Shame 5mm magnets not available; double the cost and mass for not that much computable advantage. Best regards YE

@acesias 3 ай бұрын

@@YesEinstein I have found 40x10x5mm N52 magnets. As you said, half the cost and weight. But, the magnetic field at 1mm distance for them is 3331 gaus vs 4693 gaus for 40x10x10mm. That difference in magnetic field will not be so important for the same 5mm gap between magnets? I am targeting a slow rpm, at between 500 - 1000rpm and I am looking at getting 24 - 30 coil turns per magnet

@YesEinstein 3 ай бұрын

@@acesias Capability may be judged by square of field across the gap divided by the winding resistance. N52 should expect around 0.6 Tesla at optimum gap of around half the magnet thickness. We might thus expect 10 mm to double power capacity. My analysis however, indicated less than 50% improvement. Twice the copper mass is needed. For your low speed, best Litz may just be 3 wound subs wound 4 times around the stator per phase.

@MrPounal 7 жыл бұрын

teacher will it generate current at low rpm?

@YesEinstein 7 жыл бұрын

Dear Kevin Pounal Not tested but the machine as is should produce some 75 Watt at around 15 volt at 500 rpm plus approximately 11 Watt dissipation in the windings. Lower speeds will produce proportionally less output.

@MaxChillen Жыл бұрын

Is it possible to make this into a wind turbine? with more phases or more laps?

@YesEinstein Жыл бұрын

Dear Max Chillen More laps would increase open circuit voltage but not affect power handling capability. Three phase is optimum. For a low speed wind turbine, dimensional scaling by 3 might give acceptable capability. Regard YE.

@anirudh_pednekar 2 жыл бұрын

You can also 3d print a core like i have

@YesEinstein 2 жыл бұрын

Dear RoboTronicsMotors . You appear to have a coreless central stator (i.e. no iron) containing three-phase windings. For maximum efficiency, design aims should be 1. to maximise the magnetic energy in the stator volume (not clear here), and 2. to fill it with as much active copper as feasible. In this respect any separators should be as thin as practical as advised. You clearly have lots of lovely machinery at your disposal. Regards YE

@kevinroberts5092 4 жыл бұрын

YesEinstein, noting that you posted this video in 2016 I was wondering if you would be willing to do an email exchange on this subject I am currently investigating low cost wind turbine designs for manufacture and sale in Ethiopia the principle issue being to get as much power as possible at the lowest cost, the Halbach array combined with progressive wave coil may offer some potential to reduce copper content and therefore cost, would you be willing to discuss ?

@YesEinstein 4 жыл бұрын

Dear Kevin Roberts. YE responds to public constructive comments, and queries relevant to the immediate videos. Your contribution in this mode would be welcome. Regards YE

@faratec4367 4 жыл бұрын

Dear YesEinstian . Thanks for the educational content What is the name of this type of coil? I want it for research

@YesEinstein 4 жыл бұрын

Dear Fara Tec, Progressive Wave. Regards YE

@faratec4367 4 жыл бұрын

@@YesEinstein Thanks for the reply

@faratec4367 4 жыл бұрын

I tried to make a sample similar to your model and did all the things, but with two differences 1- In my model, instead of rotating your magnets or routers. I turn the coil for less weight 2- Instead of using a fuel engine, I used an electric spindle motor, and of course I turn the rotation speed to the same 8000 rpm. But unfortunately, the result for me is that it doesn't even turn on a 30-watt light bulb Can you help me make a 1 kW sample like the one you made yourself? Thanks

@YesEinstein 4 жыл бұрын

It would help to show how you made and commutated your moving coils, Litz and magnets assembly, winding and air gap thicknesses , actual output waveform and amplitude. etc. YE

@faratec4367 4 жыл бұрын

@@YesEinstein My handmade Litz wire is made of four strings of 60 types (AWG 38). My type of magnet 30 * 10 * 5 with 42 degree and 48 on both sides of the router is exactly the same as what you made. The distance between the two routers is 8 mm and the distance between the coils and each router is about 2 mm . I can't offer anything about the domain because I don't have an oscilloscope, but each phase has an output of about 3 volts for me. I can email you more details with the photo if you wish Thanks

@NicozStrat 7 жыл бұрын

How many poles?

@YesEinstein 7 жыл бұрын

Dear NicozStrat Overlapped coils work exactly like the more conventional non overlapped, individually wound, except there are effectively twice the number around the stator. This nominally halves windings dissipation at the same output power. There are 24 magnets (poles) around each rotor. Their circumferential spacing is arranged to give a smooth (sinusoidal) axial magnetic field distribution around the air gap. Design and performance analyses are done on the equivalent basis of flux cutting by the wires within coil limbs. The current presentation Parts 1, 2 and 3 cover the design in more detail. Suggest you sift the many workshop based articles on the internet for relevant practical details but, as the New York taxi driver said, you need to practice to get to Carnegie.

@NicozStrat 7 жыл бұрын

Thank you for answer sir, you make two pass for each phase, so does it mean that number of turn of each coil is two? Which voltage generate your motor with speed increase? Thank you, very nice project and of course building

@daviddavids2884 6 жыл бұрын

thanks for posting. nice work. a few very late thoughts. i do not see any COILS forming as the conductor is glued down. i have experimented with a similar not-coil design. i have no idea how you have connected these pseudo-phases; or, what the optimum number of magnet poles would be. in my case, i could not get good performance. without an actual coil, the flux and fields do not couple, sufficiently. i guess. i have a theory about the main cause of back-emf in a coil. my conclusion is that, while back-emf is a source of internal impedance, etc., it is the main cause of the self-inducting properties of a conventionally wound coil. see my channel, if it suits. in the case of a high-speed application, connecting the stator coils of a pma delta, will result in output rising with rpm, amps over volts. (so, 300 watts, might be, 10 volts at 30 amps.) a suitable voltage-doubler or tripler would then give voltage, in the 'battery charging' range. you could try winding flat, single layer stator coils. the profile of the coil is just one layer thick (except in the area where one lead comes out.) cheers

@YesEinstein 6 жыл бұрын

Dear David Davids, Thanks for your interest. There is just one coil per phase unfolded to a single layer encompassing all magnets. Upward legs cross North-South; downs cross South-North for additive coupling. Number of magnets only needs to be even. The three coils can be simply star connected for rectification. ( The composite wire is splittable to give up to four times the turns per coil.) Induced voltage on each element of wire is given by v = Crossing flux density times length times velocity. This is mathematically the same as rate of change of total flux through the coil area. Coreless inductance and armature effects should be small compared that of an iron cored machine. Regards YE

@YesEinstein 6 жыл бұрын

Dear David Davids, Winding resistance is the dominant source of loss in the coreless machine. Inductance has a secondary effect, not necessarily detrimental. It will be near proportional to ( magnet face area x number of turns divided by (winding thickness plus 2 times one rare earth magnet thickness); thus much smaller than in an iron-cored machine. How the winding is layered pretty irrelevant. Regards YE