ฝัง
- เผยแพร่เมื่อ 12 พ.ค. 2024
- Thanks to www.PCBWay.com/?from=atomic for supplying the PCBs used in this video.
We're getting somewhere with the PCB motor - it spins pretty fast - but we're more interested in torque. There's been an interesting question from the viewers - why use a wedge-shaped coil? Surely a spiral coil is better.
Let's try and answer this question with a bit of science!
0:00 Why are we here?
1:13 Why not just use a spiral - the intuitive explanation
1:47 What kind of forces are we trying to generate?
2:00 Simulating the magnetic field from our coils
4:25 Simulating the force produced by the magnet on our coils
---
If you'd like to help support the channel then I have a Patreon page: / atomic14
And if you're shopping on AliExpress or Amazon then if you click on these links I will get a little bit of affiliate money:
Aliexpress - s.click.aliexpress.com/e/_Ac05mF
Amazon - www.amazon.com/?tag=atomic14-20
You can also join us on Discord here: / discord
And we've got an exciting new newsletter! makernews.substack.com/ - วิทยาศาสตร์และเทคโนโลยี
![BOWKYLION - ส่วนต่าง (do it without me) [Official MV]](http://i.ytimg.com/vi/7x-pwcMopzE/mqdefault.jpg)
PCBway is probably the best sponsor on TH-cam, they provide an actual useful quality service (unlike 99% of sponsors) and actually sponsor good content
Very true
I think you are missing out on a lot of force/efficiency by making the traces the same size as the spacing. This maximizes the number of turns, but not the current density. If you keep the same minimum widths, but make the traces thicker, you will have less loops, but with a much higher current capacity, for an overall higher density. For ex, if you made the traces 3x thicker, you'd have 1/2 the number of loops, with 0.5 * 3 = 1.5x more current density. Right now you're effectively only using 50% of the copper available to make the coil!
So the goal is to maximize magnetic flux right? So ya maximizing change in current makes sense, so maximizing the peak current ability makes sense :)
@@nikilragav The magnetic field strength doesn't depend on the rate of change.
@@deang5622 oh yea. Not dI/dt but just I
Magneto motive force is generated by N x I, where N is the number of turns in the coil and I is the current, so reducing the coils by some factor, means you have to INCREASE the current by the same factor to get the same force.
In electronics, increasing the current required usually comes at extra cost - typically in power losses and / or the size required for the parts to drive the circuit.
For example, for the wires leading to the motor, losses are R x I^2, where R is the wire resistance and I is the current. So if the coil current is increased by a factor of 3, losses in that wire go up by a factor of 9. For small motors like this that can easily be solved with a bigger wire, but as your motor gets larger, that starts adding cost and weight.
@@Cynthia_Cantrell ..ya...i think so too, the solution is not to use bigger wires but thinner wires and more loops, and ofc ferros, as i understand it the experiment is to get as much torke out as possible... also in the initial experiment he uses magnets...he should be using coils instead...as i understand it hes not trying to reach a high rpm but torke...the rotor doesent even have to be of a metal...it can be another board sandwiched between two boards...just with an axis...if each 3xboards is 4mm thick its possibly stack a lage ammount of sequenced (sandwiched) boards to increse torke and lower the current....its even possible rotete each section to have an applied force almost constant with enugh sections... such a motor can also be used to deaccelerate almost immedietly (like a modern electric truck with electric breaking)
Nice video. Thinking about xyz fields is key. You can see this just playing with two disk magnets, how the pole to opposite pole interaction is so different from the edge to flipped edge interaction. It's why it's so much easier to slide magnets apart than pull them apart!
if you want to stress your bearings less, put another board on the other side. not sure if you can just tie them both together or not, but I think so, as long as you either have the top one upside down or something.
I agree with this comment. You could even put another rotor on the back side of the PCB and have them share the same axel. I wonder if it will allow for more torque and or efficiency.
If the opposite sides of the board are both attracting or repelling the rotor the shaft won't be torqued in the wrong direction.
Or power several coils at once to cancel out the imbalance. That's already how brushless motors work. With 12 coils, that's 3 sets of 4 coils, giving you 3 input wires.
or put another magnet rotor on the other side of the board? this should double the effect of the coils, making the motor more efficient, i think.
Have you considered arranging your magnets in a Halbach array for more efficient use of the magnetic field? Halbach arrays have all of the magnetic field on one side of the array, and almost zero field on the other. This means for the same mass of magnets (or perhaps a bit more, since you have the interstitial magnets that force the field onto one side) you double the usable magnetic strength.
See if you can take advantage of the Halbach array concept to make a better PCB motor.
This is a very clear video with very good information that I have not seen anywhere on youtube. I will have to replay this a few times and get back to you with my thoughts. Job well done fella too.
Had to Like and subscribe for the detailed explanation on the unique characteristics of magnetic fields, how to find them, and the wonderful utilization of a PCB to contain the actual working components of a motor!
Beautifully made, and very educational! I felt 3Blue1Brown vibes - but in a hands-on application to a real-life problem. Excellent.
Thank you - very flattering to be even mentioned in the same sentence as 3Blue1Brown!
Great video! I'm trying to design a pcb motor and this was a great explanation, thanks!
I wonder how many layers we would need, if we scaled this up to a ring with 29" and used it to drive the wheel of a bicycle. Use this on both sides to sandwhich the rim, which acts as the rotor. Essentially building an axial flux synrm motor. I'm guessing it would need to spin at around 500 rpm and create a torque of around 150 Nm. Which would translate to around 50 km/h for a 29" wheel and the torque that a time trialist typically creates at the hub when sprinting.
I once drew on a paper napkin that instead of coils, the 3 phases would snake in and out like a Celtic knot, along the flat stator. The magnetic fields would be equivalent to the spiral coils, but with no loop-back, only a loop-forward to the opposite coil.
I never said it was a better idea 😮
The flat spiral coil traces should be thinner near the center and bold near the edge… then only need to cool the center.
Now try feeding it to Maple or a field simulation and then advancing the rotor. Plenty of 3 phase machines to draw on if you get stuck ofc.
What you said at 3:24 caught my attention. Are there any motor designs out there where the strongest magnetic flux direction is roughly in line with the tangential direction of rotation in order to provide more torque?
subscribed on the thumbnail and title alone. Watching now. Please keep it up. The world needs more like this
Awesome video! seriously! One point of feedback, if I may, would be to give units on all of your graphs :) ; it makes it easer to follow your calculations that you include in your simulation. Again, thanks for the video!
I have noticed a lot of the more compact VHS belt mechanisms use a setup similar to the pizza slice PCB traces to drive the tape reels. However, in those instances, the coils are not the PCB traces but are 2mm tall loops of copper soldered and glued directly onto the PCB in the pizza slice pattern. you might be able to do something similar, to test even more shapes and loop thicknesses without the need of multiple PCBs. I love the video, keep it up!
Yep, and there is steel disc's on the magnets and under the coils to focus flux
i think you can get tiny bit more efficiency by changing angles inside coils.. for example inner triangle have lines that are tilted slightly, you can move them closer eliminating dead space in the middle and distribute it by creating many thin triangles if you do not care about lines being parralel
great video, great analysis. great demonstration of how it's totally worth to spend time building a tool to make your life easier through the entire project.
but i don't understand why not make the poles the magnets and go in the same direction as the motion and place the coils to push/pull them as such. isn't it the most force creating configuration?
I need more on this… please keep working on pcb motors 🙏🏻
It’s going on in the background. Not much to report at the moment though…
Interesting. If you look at commercial generators/motors, the 'coils' are laid out such that most of the copper is running parallel (or skewed only slightly) to the shaft. This puts most of the field in a position to exert force at right angles to the shaft, maximizing torque. The 'end turns' are outside of the iron and only help to route the current to another slot to form a field of opposite polarity some distance away.
Much like your explanation of the trapezoidal winding, it's the current conductor portion that is radial to the central shaft that imparts the torque. The inner and outer peripheral current doesn't really help.
Are the stators your referring to the ones that look basically like straight lines radiating from the center point? Very interesting.
I agree, though I think he was working on an axial torque motor, vs the traditional radial torque motor.
my thought was that he could make one out of stacked rotor and stator plates for more torque, you could keep the plate spacing small, above a certain speed, they would float off of each other via the Bernoulli effect.
@@davidconner-shover51this is with both axial and radial flux machines but it is possible to use most of the wire by zip zagging it but then the wire is at an angle so it's a trade off
Also in normal stators, the coil is wound around the "hammers" of the stator, so that as much of it as possible runs along the motor´s length, and not perpendicular to it... ideally, the hammers are as long and thin as possible, so to totally avoid the perpendicular coil-portion (but the ideal solution cannot exist in reality). The perpendicular portion will be re-formed in a half-oval-shape, so that it has a longitudinal portion too, and not solely a perpendicular portion. For that, the hammers have to be increasingly thinner at the stator´s ends, and that´s very expensive, so that only experimental motors have it (beside a single stencil to usually stamp out all laminates for the stator, now you need countless additional ones, to stamp out the laminates for the increasingly thinner ends, since a single laminate is as thin as 0,1mm). Normally, the stator will be a casual one, which means, that to avoid the perpendicular portion a bit, something else is used, that´s not of iron, for the half-oval-shape, and as it sn´t of iron, it cannot concentrate the magnetic field as good... Otherwise, ohmic resistanse is lower too, since the half-oval-portions are shorter than the perpendicular ones, resulting to less overall cable-usage, lowering resistance
This really really good project. Please not lost this !!
Good explanation!! Impressed, Can you share the script for the simulator? There are some follow up questions:
- how can the motor best be shielded, is it required to be shielded due to safety requirements or can this be sold as is?
- there are articles online where people add ferrites and further enhance the field,
- you can use ferrites to pin the rotor
the motor is shielded by distance. the magnetic field drops sharply a few inches away
Finally! I've been waiting for this, after seeing all your short teasers.
Next step will be to actually measure the torque.
To stop the load on the bearing due to the magnet disc twisting, put a second board above it in the same polarity. That way it's being pushed or pulled equally in both directions.
Get that working well enough and you might be able to do away with the shaft and bearing and have the disc levitating between the two boards. Design the disc to pull air in at the edges, and out in the centre, creating a pillow of pressurized air between the disc and board to further reinforce the levitation.
I don't know how useful that would be, but it would be cool to do.
Which software tools did you use for design works?
I like the way your coils design it covers more room
Try using diametrically magnetized magnets with the diameter line separating N/S running radially and the diameter equivalent the size of your wedge coil. That way both sides of the wedge will generate a force tangential to the motor axis.
Thanks for the analysis.
Then why not ovalize the part of the wedge coil which is not in the useful direction ?
I would like to know, is core saturation a problem here if you don't have a core ? Will these coils saturate ? Would it be a good idea to insert a small oriented grain silicon steel via in the middle of the coil ?
Great analysis, and really excellent graphics. How do you do this in such a short time?
I've been cheating and using Co-Pilot with the manim library from @3Blue1Brown. It has taken about a month to get this video done though - so not that quick...
Very informative, thank you!
Great info. Thank you.
in that last image you showed, what would happen if you replaced the less useful sections of the wedge coil with solid copper sheet? will the "coils" work, if they are parallel lines? working from this idea, what would happen if you built a board that was purely radial lines, and then switched each of those lines? is it possible to refine the generated field, in this way?
Would it be possible to arrange the coils sideways by wrapping them back and forth alternating PCB layers/sides? What effect would that have on the force applied to magnets at different orientations and positions?
If you have not done this already, take a look at a 3.5" floppy disk motor. The latest versions are really optimized.
can you rotate each layer of the coils by one quarter coil arc and stagger their activation? it would yield more steps resolution though it would also mean 4x the amount of connections necessary.
The radial lines being particularly important makes me wonder if you could make a better coil by stitching back and forth on two layers, so that the coil is rotated 90 degrees to the PCB, and the length of the non-radial sections is reduced to the length of vias.
I have a headache so in case that doesn't make sense, what I mean is run a trace radial outward, via to a different layer (can't think well enough to judge what layer pairings would be good), run the trace back, via back to the first layer, repeat. On a four layer board you'd be able to fit two coils, either stacked or nested (or interleaved?), I just can't fit all the details in my head to resolve if it'd be better for layer pairs to be close or far. Of course this whole plan would benefit greatly from blind vias so you don't have to fit four sets of vias doing the same pattern in the same space.
The big advantages would be more radial trace length since you don't lose the triangle of... axial? traces and you don't have a hole in the middle where it's hard to fit the spiral geometry. So you should be able to sweep radial traces of the same length over the entire wedge.
Maybe instead of 2 depth-first coils it'd make more sense to spiral across all 4 layers and then step in depth, so each coil-layer has 2 rotations instead of 1
Coils (with current flowing in them) approximate magnets. So rotating a coil by 90 degrees produces a similar field to a magnet that's been rotated 90 degrees. There are specific cases where this is desirable (look up halbach arrays) but it's generally better to have the flux lines pointed orthogonal to the rotor/stator, rather than parallel to it.
Exactly how early pancake motors were made.
@@paradiselost9946 Any chance you know of any text about how that worked out and why it where it went? (I don't know where it went)
Awesome work!
Pretty disappointed that none of the mentioned articles, simulation tools, or references were sited.
I am curious why the SHARP turn on the center connection of the coils? Wouldn't a smoother corner reduce noise and allow avoiding a hot-spot in that sharp turn?
Do you series each winding for higher input voltage or do you parallel each for higher current?
Designed a pancake coil on a PCB recently (used svg2shenzhen because couldn't get that jupyter notebook for that plugin to pump out a json for that plugin), but sadly that pancake coil only got hot when about 1A was running through that. Wondering if I should make another with wider traces. I am already pissed I spent 7 DOLLARS on the first one that didn't function properly though, so I want to MAKE SURE the next one functions properly.
Very good presentation. Can you tell me about my knowledge, what about wire thickness size use and number of layers. It's helpful me. And i want One other comparison in magnet flux density of neodymium magnet vs Iron nitrate magnet for better understand. Thanks NAMASTE 🙏
I think the simplest way to explain it is that magnetic induction - in other words, the strength of the magnetic field - is proportional to the number of loops of current that constitute the electromagnet.
So, as wedge shaped coils allow you to fit more loops in the same amount of space, they are indeed more efficient.
Which programs did you use to make the graphs of the magnetic forces? Was it on python?
Well done!
The more turns the more inductance, therefore limiting top speed and its torque.
If you want more speed use less turns = less inductance, just need to use more amps, also this has more copper, lowering the resistance increasing torque at speed.
Can this design work for an axial flux synrm or switched reluctance design? Or must you use permanent magnets?
This is very interesting however im curious how a helix coil would compare to wedge and spiral coils.
Thanks. What software do you use to simulate?
This makes me wonder, what if instead of coils following a simple wedge pattern, the perimeter were serrated so that the coils would overlap a little? The overall surface area covered by the coils would be identical, but they'd have more influence inside their neighbor's regions. I suspect the operation would be smoother and quieter, as you wouldn't be jumping as abruptly between magnetic fields. But I can't visualize what it would do to the motor's torque.
First thought: if the lateral lines of the coils are all we care about, why not shield the middle / vertical lines so they dont produce a pushing force on the wheel? I imagine the amount of surface area of the vertical coils shielded would need to be small not covering them completely, and probably rod-shaped ?
By treating the field of only one coil at a time, you're missing out the effect adjacent coils have on each other. The long radial strips running in opposite directions create a stronger field in the X (tangential) direction.
Czy wytworzone pole magnetyczne jest innego kształtu przez nawijanie inaczej cewek ? czyli od zewnątrz do wewnątrz czy po szerokości lub długości gdyż dużo zmiennych mam w głowie
What about shape of the magnets? What shape is the best? Rectangle of the height of the coil? Narrow or wide? Or is the round magnet better?
Do you also run the half poles in your inner layers?
+A-c+B-a+C-b
hi, it is posible to get the pcb files and any other like a part list to make my own, please
Try the same simulations but sandwich the design.
Magnets on top and bottom side, and a steel backing plate behind both magnet discs to close their back-fields.
are you still work on pcb motor design? Is the pcb motor have to start super high speed? can it start at very slow or zero speed, but still keep the torque force?
Could this be modified into a uniform induction heating cooktop?
How did you create the plots?
How would this be applied in eddy current induction motors?
Fascinating. My chemistry brain really wants to think it's looking at pi* and pi bonding :). I wonder how a spiral with equal trace length would compare? I'll have another watch when my brain is more awake!
That’s one flaw with my analysis. The wedge coils have a longer length in total than the spiral coil - which means higher resistance and so less current… I’ve seen some examples where the horizontal (useless) traces are made thicker so that the total resistance of the coil is reduced. There are some amazingly complicated wiring schemes.
Спасибо за моделирование.
К подобному выводу приходишь, если сравниваешь силу Лоренца и Ампера у разных катушек. Но моделирование тоже полезно.
Thanks, very well explained , and this is an invitation to study electromagnetics laws I'm going there 😊
Have you seen the csiro pancake motor designed for solar car challenge?
Nice design and explanation. What about distributing the coils around the whole wheel, something like star or gear wheel shaped coils? Another improvement may be to let the PCB spin and have two sets of magnets on both sides, though that makes it much harder to power the circuit.
Why not turn the magnets upright into a cross-field configuration to maximise the x-plane / angular force (torque)..? Alternate the polarity of the stator coils so each magnet's being pushed and pulled simultaneously, and time the duty cycle so they cut out as the magnets pass their centers (the sticky spot) to minimise negative torque and cogging. The twisting / precessing moment (the magnets want to turn face-on to the coils) could be mutually cancelled 180° opposite sides of the motor to minimise off-axis stresses..?
Great Chanel Congrats my friend .
what ESC did you use
would it be feasible to incorporate a induction heating coil in a PCB?
Do you run multiple plates with one more stator than rotor to reduce the flex and end forces hence reduce the air gaps?
The entire thing is in the air. It the main reason this has no power. A typical motor has only 1mm air gap in the flux path.
@@aerobiotic
Toy car motors run a whole lot less than 1mm
The reason is efficiency. This is why all big iron core motors have multiple slots instead of 4 big slots for the coils.
Is a PCB Halbach Array possible? Now that would be interesting.
a second, thick, two layer pcb with purely radial trace coils offset from the main drive coils and fed the phase difference between the two adjacent drive coils.
though a ferrite disc on the back side would have a similar effect
If you make the corners too tight the electrons can fly out.
Mass of coil that is in the path of magnet in the zone of highest reluctance?
Simple answer: parallel wires/field (better distribution, layout and Force x Distance) have higher torque efficiency than radial ones.
what software is that?
Hmm, I think you're missing a lot of flux out the back side of the board. What if you somehow make the rotor wrap around the PCB so you can add some magnets to the back of it?
I just had a thought, what does the output look like in reverse. Why you ask ? because I'm collecting ideas to make a spinner for games such as Arkanoid, Tempest and of course good old breakout. My first idea is to take the motor and plater from an old hard drive it looks really cool after shaping pulses out with shmiitrigers and the like, I was disappointed with the speed of response and it's now on the shelf (we all 'the shelf' ) but just seen this and it might be an avenue to explore...any thoughts anyone ?...cheers.
What, the spinner's a labor saving device for the joystick jockey?
Doe it have magnets on both side of the PCB ?. I hope having magnets on both side of the PCB will double the torque...
Wedge shape coils have more surface area than circles. You can fit a larger coil in a same space than circular coils.
Not watched it yet but I'm going to have a guess. Its to keep the rotation od the magnetic fields on both ends of the coil the same. As distance travelled per degree of arc is greater as you move away from the centre of rotation. Now to watch the video and see if my assumptions pan out.
I’ll be interested to see what you think.
@@atomic14
I think that we are thinking along similar lines but from different directions. My idea is that cutting the wires with the magnetic field is important for generation, so to arrange the wires as radials from the centre of rotation means that the field cuts all the wire at the same time. Were the curves of the loop mean that it takes more of a rotation to fully cut the outer most coils.
A lot of work and effort must have been put into making this video.
Also it’s a nice change to see someone working on making a motor instead of a free energy or greater than unity alternator.
Maybe I'll stumble across a perpetual motion machine...
JLC pcb can do up to 8 layers. Have you tried going from 4 layers to 8.
What do you get if make your coil perpendicular to the board instead of parallel with the board?
By that I mean run radial tracks on the top and bottom of the board then join them with vias. Will this make all/most of the track surface area 'useful'?
Yeah, I’ve seen some interesting alternative wiring arrangements. Still trying to understand them though!
My instincts tell me that the magnetic field would be less if configured as you decribe
i studied electronics for 3 years in highschool and again took physics of electronics in college... i still cant do what you do :) i am a pro at code, but not electronics lmao
How come youtube never recommend this video, I have watched a few of Carl Burjegas videos
would this not imply you would actually want diamond shaped coils? to have more of the current responsible for rotational torque?
I like watching you be right. Good Job. (yeah, that makes sense)
Все верно рассказываете. Я изучал это в институте.
Pentagonal wedges? That would explain a few things...?
why not ditch the magnets altogether and go with an induction motor?
you can do this by making the rotor a circular 4 layer board, use thin, tightly spaced radial traces connecting to a thicker inner and outer ring trace, the layers need not be connected. those traces being the inner and outer axial diameter of your stator coils, no holes needed except for the central axis. as long as you are feeding your stator coils 3 or more phases sequentially, the rotor will follow the coil phase propagation. torque will increase all the way to the point the rotor is locked.
in either case, PM or induction, if you could park a ferrite ring behind your stator coils, it would improve your transfer efficiency massively.
Maybe a dumb question, but could you gain additional torque by using both sides of the PCB?
He is using both sides, plus two additional layers inside the PCB. each visible coil has 3 identical coils below it, with the same turn direction, all wired in series. Since the magnets are only on one side, it could be possible to change the shape of the coil in each layer, in order to improve the shape of the magnetic field on the side where the magnets are. Another option is to have fewer turns and widere traces (with lower resistance) on the side facing away from the magnets, and more turns and thinner traces on the side of the magnets, assuming the generated heat would be spread through the board anyway, and the coils close to the magnets produce the most effect on the magnets.
yeah, the rotor is on both sides
Is it possible to make a single phase version of this design? If not can someone explain why like I'm 6 years old? thx!
Reminds me of end stage floppy drives.
You point out that the x component of the force is what's important for generating torque with the coils, but you use a magnet whose field is only in the z direction. How about using magnets whose axes are tangential to the coils on your board?
hasn't this been done in VCRs as far back as the 70's?
How come it can't be solid copper or solid copper with holes .
6:50 - I guess if (for some reason, like crankshaft like operation) your goal would be to exert radial forces rather than tangential ones, then the notions of "useful copper" and "less useful copper" would need to be flipped. - I wonder what the optimal geometry ten would look like …
I walked into Home Depot with a pink 18 650 rechargeable battery... They looked at me like I had three heads... I bet the mom-and-pop store might be interested? maybe they don't want to get burned down... With the inquisitorial technology