An Axial Flux BLDC Motor Build 

eeee

Did you measure the resistance of the phases? it is possible that maybe there's a short in one of the coils.

@@RoycoNL The resistance of the phases is the same as far as my multimeter can tell (0.1ohm resolution).

Hey, is there any way I could reach you? (E-mail, discord, Instagram...) Maybe my team could help you.

It may be short-circuit in one of phase coils or some issues with coil winding resistance (some wire conductivity imperfections or such). Just my guessing. So anyway it is an interesting video, thank you )

Thanks! Why I don't really want to release the 3d files is because the motor requires quite a bit of other "stuff" as well to work (the washers, magnets, wire, bearings, shaft, etc...); so for anyone to really be able to replicate it, I would certainly have to make a proper manual for it. I think this particular design is not quite worth the time making such a manual. Whenever I get to making the next motor however, I will also make an assembly manual and definitely release any files :)

Well reasoned. I was going to ask a similar question but instead I’ll wait for the next impressive motor you come up with :)

I wonder if the halbach array was the key

Hi. RPM does not necessarily mean higher power. Power = distance (RPM) * torque. This motor has a low kV (low rotation speed) but a high torque instead. Next year I might want to try making an e-bike motor, and e-bike motors need to have a lot of torque but lower RPM. The ESC plays little role here - it will give you enough power until it burns itself or shuts itself off completely to protect itself, neither of which happened here. I now have a sensored ESC for my next motor project though, and that would increase the performance of the motor a little bit as well, if it had hall effect sensors fitted.

Aah, the motor was never meant for anything but learning and demonstration of working principle... I would have to do a lot more research into manufacturing if I wanted to make something that would be a useful prototype... maybe some day.

Iron cores for the coils lead to problems with cogging. I also think it is a problem to combine the Halbach design with a backplate. For a Halbach design there is also too much space between the magnets. I guess the use of the normal switched polarity magnets in combination with backplates should work best. In this case it is also easy to calculate the flux density, because the air gap mainly determines the parameters of the magnetic circuit.

Yeah, I was actually quite surprised that the sound did give a real feel of a V2 haha.

@@BirdbrainEngineer Put a gearbox on it and you could make a good sounding electric car like this one: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-YJ0CU0-PzQc.html

@@swecreations Also whoops, I replied to it from my comments tab from youtube studio, and I misjudged the video this comment was made on - thought it was on the V2 air engine video haha

@@BirdbrainEngineer Look at the video I linked still though, I think you'll find it interesting

@@swecreations I did, sounds wicked!

Yeah, I'd like to measure the inductance... still haven't gotten myself to buy a compatible multimeter or something more specialized for that. All the phase-pairs measured 0.7ohm if memory serves... Would a... say 0.1ohm difference between phase-pairs cause that great of a difference in kv? Should look into that... the actual underlying math looked kinda complicated on the surface though, not gonna' lie haha.

Realistically, due to there always being some unevenness in the print bed and 3d printed plastics being... well... plastics and so a little malleable, you are not really able to get an airgap smaller than about 1mm. At such scales though, the difference in efficiency is nearly negligible.

Hi! The solid steel washers are not "cores" but "back-irons", which serve to give the unexposed permanent magnet poles an easier flux paths between one-another, which in turn slightly increases the magnetic force exhibited by the permanent magnets. The magnetic flux generated by the electromagnets is not able to influence the back-iron due to the permanent magnets and their much more dense magnetic flux being in the way. Therefore, the back-iron does not experience any eddy currents (there are no *changing* magnetic fields inside the back-irons) and cause little-to-no drop in efficiency. Yes, a back-iron is less needed in a Halbach array rotor, so it *may* have been better to not have it but it certainly would not have been a significant difference either way. Have you ever tried finding electrical steel laminations for sale? Especially for your specific project? Electrical steel laminations are for products that actually end up getting mass manufactured or sell for large amounts of money per article. I am quite poor, and most definitely can't afford 10000$ on a single project! I have entertained the idea of finding a product with suitable electrical steel laminations and simply cannibalizing the electrical steel from there and cut them to shape to fit my purpose, but so far I have always just opted to use mild steel, which for many of the common blends have maybe about 20x worse magnetic characteristics than most blends of electrical steel... but still 50x better magnetic characteristics than just air... The efficiency of diy motors is difficult to get very high, especially when 3d printing components out of plastic, because the tolerances simply aren't there. The air-gap between the rotors and the stator for this motor are between 1 and 2 mm... (in production motors this gap is always less than 1mm, usually on the order of 0.1 to 0.2mm). Add to that the use of non-ideal materials, inefficiencies from things like rotor inbalance and less uniform windings... I think that a diy plastic FDM 3d printed motor will not reach even 80% efficiency for the next decade. Hall effect sensors can be used, yes... in-fact they were necessary for my next motor project which I will release probably within a month. Once again, this is diy, made cheap... *literally* in a commie-block apartment room, I'm not sure what kind of manufacturing quality can be expected unless I order like... CNC-milled stuff, which would be too much of a monetary cost for me to take right now. Thrust bearings should be added if the motor shaft is expected to get pushed or pulled like when a propeller is attached to it. Having said that your motor needs to be pretty darn big and powerful for this to be really necessary. Not a single drone motor (even those crazy ones that supposedly have a 2hp peak power output) I have seen, uses thrust bearings in their construction. Of course there are formulae. I'd contest that such formulae are used in conventional ways nowadays any more though... and experimentation and optimization is the gold standard of the industry nowadays. Simply that the experimentation and optimization is done in software using finite-element-analysis models... Unfortunately, the best ones, such as EMworks, are part of larger CAD suites, which are prohibitively expensive for diy-ers to use. If you know of a good and intuitive free software for magnetic, or indeed, electric- and magnetic analysis, then do let me know! I have been looking for one on and off for some time now.

Yep, that is something I will try to look into in the future when making my next motor(s).

You can run a BLDC motor in "sensored" mode as well (as long as the electronic speed controller supports it), but in that case you need 3 hall effect sensors. This motor is not sensored, and does not require any sensors to work. Unsensored motors and ESC-s (like the vast majority of drone motors for example) only expose the three motor phases and nothing else. The ESC figures out when to change the phase based on looking at the voltage induced on the disconnected phase. The reason for why this motor doesn't start up properly is because one of the coils in one of the phases is connected up in the wrong direction... to this day I am not sure how I managed to do it but I have not bothered to fix it.

The advantage of axial flux motors as opposed to radial flux ones in EV-s is the higher power density and consequently higher power to weight ratio. This is indeed highly sought after in EV-s as it allows you to put more batteries on the vehicle, prolonging its range. Axial flux motors are much more difficult to manufacture though, as any automatic winding machines are much more complex, and simply manufacturing a good performing stator with iron cores essentially mandates the use of milling - contrast that to radial flux's need of just stacking electrical steel laminations to create the stator, it's a massive price and complexity difference. Indeed, another problem is the need to use rare earth magnets, usually neodymium magnets, which are made from expensive, environmentally polluting and politically contentious resources. I also looked into the iron nitride magnets you hinted at, and while its flux density of 1 to 1.5T is impressive, the coercivity is pretty low, meaning they will be quite prone to demagnetizing over time... and forget about making any Halbach array, where magnets are constantly trying to demagnetize each-other. In addition, it seems like the coercivity drops off faster with temperature than rare-earth magnets. All of which is to say, I can definitely see iron nitride replacing rare-earth in some applications, but for EV-s it would require careful cooling design considerations, which might complicate the whole vehicle design enough for manufacturers to not bother in many cases.

On an axial flux motor, the only possible place to put an iron backing is on the rotor(s). This is a very common thing to do on axial flux motors. The iron backing only acts as a low-reluctance path for the magnets to couple to each-other and ever-so-slightly increase the magnetic flux density in the airgap of the motor, as well as minimize the magnetic field outside the useful working area of the motor. I am fairly certain the problem is in one of the coils (maybe i ended up winding one coil the wrong way... though i did quickly check for that and at first glance it didn't seem like anything was wrong). For the next axial flux motor I will take more care to triple-check every assembly move I do.

supermagnete

That's an idea for if I ever get some land and build a homestead heheh

It's just fairly difficult to fit iron cores in an axial flux motor, and indeed, it's just very difficult to source a good electrical steel material. I was thinking that maybe old transformer cores would work if cut into shape, but I am not sure about their high-frequency characteristics - maybe it changes something.

@@BirdbrainEngineer failing that you might just try protopasta's iron-filled PLA. It's definitely not as good as pure iron but it would probably be better than no core at all

@@0hypnotoad0 I believe I saw someone testing the magnetic permeability of "magnetic pla"-s and they don't have a magnetic permeability greater than 5 to 10, which is really really bad. Put ontop of that the fact that PLA melts at a low temperature, it's just not a good idea to use that in a motor one would expect to be "useful" imho. (The motor made in this video was never meant to be "useful" for anything, rather just to lean how to make one, hence the heavy use of PLA). I think in the future if I 3d print parts for a motor then they will be printed with ABS... or if it is possible with my printer, then maybe even polycarbonate.

The next axial flux motor will have a more unique design for me too - this one is basic. Slotless iron core? That sounds more like an iron backing rather than iron cores. From your channel banner it would appear that the core is essentially a donut-shaped hunk of metal? (Correct me if I am wrong) If so, then I would imagine your motor suffers from substantial eddy losses?

Your right that my axial flux motor does have an iron backing, that's what makes it slotless quite similarly to radial flux slotless. The iron provides a low reluctance path through part of the magnetic circuit. But my motor doesn't suffer as much eddy currents as you think. The iron torus backing is made from steel strapping used to secure heavy loads to pallets. The enamel coated steel strapping is wound up in a coil where in each layer is electrically phase isolated from the others above and below so that eddy currents can not form radially, and thus significantly reduces the irons capability to form eddy currents overall. If i were using electrical grade steel with a high silicon content, the steel's higher resistance would conduct even less eddy current.

And I assume higher eddy currents?

There are no iron cores in this motor. Only iron backplates for the rotors. There will be some eddy currents from the electromagnets but really not that much, as in the ideal case the produced electromagnet poles are moving at the same speed as the rotor. Reality is not ideal though, so yes, there will be some very small eddy currents in the iron backings but most likely not enough to actually hurt the performance.

If you ask about something specific then I may be able to help you out. In terms of calculations made for this particular motor...? Well there were extremely few because the design was mostly dictated by the materials I had on hand. The 20x10x5mm and 20x5x5 neodymium magnets, The wire gauges I had on hand, etc... So in a lot of cases, only back-of-the-napkin precision calculations were necessary to find the best combination possible with the materials I had on hand.

@@BirdbrainEngineer okay..! Thanks

Indeed, it would be better. However, making a true sine-wave is actually not that easy to do. For example, the 3-phase induction motors that are extremely common in industry simply use 3-phase power off the grid directly. And the motor controller for a synchronous reluctance motor costs many hundreds or thousands of dollars. The cheap way to make an approximation of a sine-wave is to use transistors and switch them with pulse-width-modulation for an approximation of a certain voltage. (Usually a capacitor is also used to smooth out ripples in the "signal")

@@BirdbrainEngineer look up VESC Project, there's some vesc based escs out there that are really cheap and works with most BLDC motors

@@stealther1401 Thanks for the tip! I had planned to look into vesc based stuff once I start working on my e-bike project. FOC would be pretty nice to have.

Most freely available steel is either some low carbon steel or stainless steel... out of both categories, most blends will be suitable to act as iron cores for the electromagnets, but they are not really that great at that role. And indeed, if you are not careful, then you may actually end up losing performance due to increased eddy currents and cogging. I am working on a motor project that uses steel though, so I definitely haven't ruled out the possibility

@@BirdbrainEngineer what about silicon steel as used in transformer cores ?

@@heartobefelt Silicon steel is the most common type of steel used for electromagnet cores. Usually the exact magnetic characteristics of the electrical steel (different name for silicon steel in electric applications) are specifically manufactured to suit a task... May it be as motor electromagnet cores, transformer cores, noise filters, etc... I have thought about using old transformer cores as electromagnet cores for a motor, but I haven't really looked much into it yet. But yes, it is on my radar as a possibility for sure.

The main magnets are 20x10x5mm, and the sideways magnets are 20x5x5mm

Agreed, that was certainly a compounding factor. Though I think the other factors, potentially the more impactful ones were that the kV was messed up, so the ESC would have had a harder time with picking up the BEMF, and that the rotor has a lot of inertia as well as an amount of resistance, which is certainly higher than the ESC was really "designed" for.

@@BirdbrainEngineer As an improvement you could use a BLheli ESC. They are more expensive but also have better startup torque than the noname ESC that you used.

Iron cores essentially extend the generated magnetic field all the way to the airgap of the motor. So yes, generally it is a good idea to have iron cores in motors, however it's simply difficult to accomplish for axial flux motors, and ontop of that... most steel products that you are able to find easily do not make for good electromagnet cores due to their geometry and magnetic properties.

Probably. Would need a redesign of the chassis though. Also would need to be designed for a lot higher power draw, since a 35W motor would have a hard time getting an e-bike going even with a large gear reduction. In the future I will probably attempt to make a diy electric motor for an e-bike conversion, because it feels like it would be a pretty fun larger project to undertake. I'd aim for around 250-500W motor then though.