You could increase the torque even more, like, A LOT, by putting a piece of iron on top of the rotor, right on the back of the magnets. Brushless motors have so called "back iron" which is the iron ring where the magnets stick to. Yeah that's not alluminium. The magnetic circuit otherwise closes behind the magnets in the air which has a very high reluctance. You can cut such a piece from the iron of an old transformer with big scissors. Since this piece of iron will provide structure you could save part of the gained weight by slimming the aluminium part. This is my field of expertise.
Would it add iron losses? If not, what's the deal with "back ironless" motor designs? And what about Halbach arrays? Also, would sandwiching the stator between two rotors be a good idea?
@@CarlBugeja I agree with nasone32's general idea. Part of the problem is that the magnetic flux generated by one of your windings is weak because most of the path from the north to south of the winding is through the air. So in concept, I think what you want is to equip each winding with a "C" shape iron piece, where the winding and the rotor sit in the gap of the C. The C supplies a much more susceptible path for the coil field, hence the flux is higher. So, a total of six C pieces, one for each winding. This should also mean that when the rotor is spinning, it generates a far higher back-emf, resulting in a lower effective EMF applied to heating the coil. Just the fact that your coils are getting very hot is a sign that a major proportion of the energy you're putting into the coils is not being used to drive the motor. (However, If the motor's stalled, then the windings will get hot by straight I^2 x R anyway, because there's no back emf.)
@@revimfadli4666 it would add very minor losses, the magnetic field at the back of the rotor doesn't change much at all. coreless/ironless motors have no material in the stator which sees a very rapidly changing magnetic field, unlike the field in the rotor. a halbach array would work, but packaging in such a tight space would be problematic.
@@bgdwiepp I see, thanks! What about adding another rotor so the stator gets "sandwiched"? Would that improve the performance? And would using a back iron mean adding a Halbach array afterwards basically does nothing?
The delta would be usable with half the voltage. Scaling up motors to the industrial scale, 3 phase motors come as user configurable as Star or Delta. In Australia, the single phase power is 240Vac while 3 phase is 415Vac. When using a single phase to 3 phase inverter to drive a motor, it is normal to connect the motors in the delta config. but when used on the 415Vac 3 phase, we connect them in the star config. In your demo, you were using 5V, so there is no reason why it shouldn't work at 2.5V in the delta config. Nice job by the way.
Carl, your work is really amazing, especially because you do innovations. Infact your videos, compared with the others, are really mind stimulating. Thank you so much ☺️
Cool stuff. The first thing that comes to mind in terms of practical applications when seeing this version is the LSU polygon motor in a laser printer. It's a motor that spins up a polygon mirror attached to it. The mirror then deflects the printer's laser beam to make it cycle through the page in a predictable manner. I think you'll need at least 15000 RPM for this motor to be practical in that role, and that speed needs to be sustained and incredibly stable. On the upside - no need for high torque whatsoever. With your setup, the mirror and the magnet housing might as well be the same piece, so the only thing the motor would have to spin up is itself. The same type of mirror unit can be used in a variety of laser devices, such as laser projectors or LIDAR-like sensors.
Have you considered making increases to the rotor field strength? Would help with coupling and certainly improve efficiency ( = less heating! :D) For example, you could put magnets on both side of the stator, or use a Halbach array like Cristoph Laimer's 3d-printed motor.
Very good progress! As stated by others, you need back iron to couple the magnets and complete the magnetic circuit. This will also keep the magnets from being demagnitized as easily.
Very cool work. Just wondering if the switch to using aluminum for the rotor causes eddy currents, and if that might reduce the efficiency of the motor?
Thanks! Hmm 🤔 very good question.. I don't think eddy currents will reduce the efficiency with regards to speed and torque.. But in reality they might be also getting generated in the bearing and shaft... That is one of the reasons why I prefer this over the old design because the bearing is now inside the rotor and not directly underneath the rotor's magnets
@@CarlBugeja I actually did research in electrodynamic wheels(basically a wheel with circumferencially oriented magnets in a halbach array). We found that all metallic objects created a large eddy current that opposes the field of the magnets. The magnetic field generated in the aluminum will have a phase to the magnetic field of the magnets passing over it, such that as your RPM increases, you will double up or divide your magnetic field strength. Using PVC is a much better idea. blogs.nvcc.edu/intercom/2014/03/24/majewskis-students-win-research-competition/
Oh good thinking there, it’s probably why the magnets got so hot. I’d expect the rotor to remain cold. Perhaps make a silicone mould of the milled rotor and pour in some epoxy, maybe with glass beads or fibres for reinforcement.
@@TurkishLoserInc Yeah the aluminum will definitely produce eddies from the magnetic fields and generate heat + counter-torque from their induced currents, this is why motors are produced in laminated sheets. The only question is how much drag and heat is being generated at these magnetic flux levels. You could figure out how much drag there was experimentally by changing the rotor back to plastic or milled carbon fiber and measure the RPM differences between it and the aluminum rotor, and you could test if the heat was the result of eddies or the PCB by literally heating the PCB with DC current (no eddies) through the coils while the rotor is stationary, and see if that heat can make its way through the bearings and shaft and get the rotor past 80°C. If it can't, then heating must be coming from eddy current losses.
@@CarlBugeja I can see that it might not have much effect at this scale but if scaled up it might factor in. also something to note is that the magnetic fields in the permanent magnets are a lot stronger than the electromagnets, and the fields from the electromagnets are the ones that will induce the eddy currents. so they should not be a concern.
2:01 "GIMME PCB!!" Ehi for the heat, what about a little fan (maybe in plastic) on the back of the motor pcb? The height increases, but the performance will decrease less. Or maybe a different design of the rotor: fan shaped, so it will cool itself without any other extra parts.
You can re magnetize the magnets by heating them and exposing them to a magnetic field then cooling them. You can also add a second pcb on the other side of the rotor to double the power of your motor. Just add a longer shaft. That is the beauty of an axial motor.
Awesome idea! This could also be really interesting to make a linear motor on a PCB. I figure PCB milling should be quite precise enough as a crude linear guide. You could make the coils wider horizontally, add a chip to only power the active coils and maybe add some iron to increase the force. I don't really have much knowledge about motors though. This could be cool for something like a stewart platform. 6 legs just with their distance to the ground would define the 6dof position of the robot. So with 8 legs that can freely swing you should be able to make a simple robot that can lift two legs while the other 6 keep is fixed, then move all platofmr, lift 2 legs and let them swing back to vertical.
I wonder if you could sandwich the rotter between 2 PCBs on top and 2 PCBs on bottom. This will trap your spindle between 2 PCBs on either side (higher stability) . Then use the outer exposed disk for mounting blades to instead of shaft. Might be interesting to do push pull between top and bottom PCBs due to heat. Also if u cut small curved vain into the aluminum you can have it self cool to a point.
Would probably really benefit from a ferromagnetic core in the windings somehow. Maybe with vias and a ground plane? Connecting all the magnets together in a ground plane, to create a closed loop magnetic field. Not having one is like putting resistors in your magnetic field, the field needs something to kling on to.
Very nice!, except for adding iron to define magnetic path, maybe also try rotor with no indentation to minimize air resistance or drag, the loss is related to drag and square speed.
With a proper driver and commutation using hall sensors, the basics involved are this: Current = Torque. Speed = Voltage. A proper motor driver will reduce current when the commutation is in phase with no lag, so low heat and high efficiency at all speeds. At higher velocity, the current for the torque remains constant if you ignore friction and induced current loss in the conductive magnets. This motor control can be performed with an inefficient class AB analog amplifier/driver to feed the commuted 3 phase drive to the motor using 3 hall sensors and 4 quadrant analog multipliers to generate drive signals to the amplifier which is configured not as a voltage amplifier, but a voltage to current amplifier. Motor drag, is responded with a change of current to maintain the correct speed by adjusting current to change torque. As the motor runs slow or faster, the voltage to current amplifier automatically adjusts voltage based on motor back EMF to apply the correct committed torque at all speeds. For higher efficiency, a high speed PWM voltage to current driver can be used instead to produce the 3 phase currents required. If you change from a 3 phase design to a 90 degree 2 phase design, cheap stepper motor drivers can be used to run constant current, thus constant torque and constant heating in the motor at all speeds. Use higher supply voltages for higher maximum speed. To grasp the concepts, stepper motors are inexpensive in smaller sizes along with off the shelf stepper motor drivers. Play with and learn stepper motor driver theory and technology to transfer to your PCB motor designs. You can buy 3 phase versions, but they are less common and more costly. www.amazon.com/dp/B00PNEQ79Q?tag=duckduckgo-d-20&linkCode=osi&th=1&psc=1 Stay away from the arduino stepper H Bridge drivers. They are made for slow speeds such as 3D printers. To learn motor drive and view waveforms, get a real CNC style stepper motor driver. www.amazon.com/Usongshine-Stepper-Controller-Leadshine-2-Phase/dp/B07PHLJM22/ref=sr_1_36?dchild=1&keywords=stepper+motor+controller&qid=1592375778&s=electronics&sr=1-36 You set the desired current in dip switches, and your maximum speed is limited by your maximum voltage used. If 1 Amp is too much, you can buy smaller motor drivers.
Very interesting! I like the direction this is going. Have you considered putting thermal epoxy as a heatsink on rear of pcb where the coil areas are? Also you can get magnets that are rated to 125+ celcius. "Polymagnets" can generate greater force than standard magnets, Correlated Magnetics make polymagnets.
thermal epoxy is there just to fill the gaps between a heat source and an alloy heat sink (cause perfect flat surfaces, that could touch each other with 100% of their expanse, don´t exist), and doesn´t help dissipate heat itself!
@@klausbrinck2137 yes an actual heatsink would be superior with thermal epoxy to bond and provide optimal thermal connection. But thermal properties can be improved with just thermal epoxy, it still wicks heat away fairly quickly and increases surface area. Have you seen microcontrollers on a pcb under a pool of set thermal epoxy? Thermal epoxy can be cast or overmolded into shapes like normal resin.
> Have you seen microcontrollers on a pcb under a pool of set thermal epoxy? Do you mean the black blobs common in very cheap applications? This isn't for thermals, if's for protecting the die.
There's the right way to set a bearing with a nice press that doesn't put undue stress on the bearing, and then there's they way everyone actually does it, which is HAMMER :D
Any room for heatsinking the back side of the PCB? If you mill another magnet rotor, maybe cut some grooves / fins into it so there is a slight cooling effect without much if any extra drag.
The low resistance of the delta configuration is not a problem, it is necessary for higher speeds but you will need a smarter driver that has current limiting
But he would have to put them in parrallel since else the current would be half as much leading to the same torque since B=μ*N*I/L if N stays the same and I is getting half and L is staying the same the magnetic field per PCB would be half as much. And due to 2 PCBs the magnetic field would stay the same. But if he put it in parrallel there would be a magnetic field being stronger.
neutronen stern Series would mean a higher total impedance, which would make the coils easier to drive with less losses in the driver, or prevent the coils overheating at 100% duty-cycle at the fixed voltage like we were seeing earlier. Not that he should be forced to run the coils at 100% duty-cycle, that is.
1-2 pcb on each side of the rotor? more surface for heat dissipation, for the higher-current case (delta?) ? anyway not all pcb´s on 1 side, cause the small magnetic fields (of he windings) would never reach through to the rotor-magnets!!
There might also be an advantage in sandwiching the magnets between two PCB's in being able to direct air across the sandwich of PCB's and magnets and thereby increase cooling efficiency.
I like to use such motor in a cassette tape housing (on a custom PCB inside) that turns the wheels or use one of them as generator to drive the second wheel. The magnets are in the surroundings of the wheels and coils are around the holes. The reason I want to do this is to make BT-adapter without noisy rotating gears. Some cassettedecks requires the motion of both wheels, that is why there are gears inside such adapter. It would be fantastic when this can be done without batteries at all, so when one wheel is driven by the cassettedeck, the other automaticly spins (slower of course due energy loss). That would be amazing! A nice application, right? (or stupid idea?)
Maybe there's a way to sandwich rotor between two pcb stators which should increase the torque (probably).. So for drone it might be possible to have 2 pcb stators on both sides of a rotor. but what if combine two sandwiches together? 2 rotor 3 stator configuration.. not sure what the benefit of that would be.
Try to add magnets at different side of pcb for more torque. And sure avoid the high temperature. Also try to insert copper inside of pcd. It's a multilayer pcb like used in graphical card where we have more than 7 layers.
he already has multiple layers connected with vias. You are right though about stacking. He should stack a stator on the other side of the rotor though so that the magnets are acted on from both sides.
To get the most out of your magnets you need to confine their flux to pass through the stator as much as possible. To achieve this you should use a magnetically permeable material (a steel washer perhaps?) to back the rotor and make the rotor double sided. If you could source them, halbach arrays would be ideal. Not sure how realistic that is at this scale.. Trapezoidal or sector magnets would also let you use the space under the disc more efficiently. Maybe you can get the HDD magnets from an oldschool iPod or something. Given the choice of using more of the space you have for magnets or just a higher grade, more magnetic material is usually the way to go. imho.
What about creating the rotor with fins, to aid cooling & prevent overheating of the magnets? Also, have you looked into graphite as a washer option?...
consider changing the magnet housing to a non-conductive material (other than plastic of course) or use laminated sheets of aluminium epoxied together perpendicular to the pcb. the eddy currents in the housing is giving issues with heat and sapping a lot of the power. adding iron would make this worse. iron is only good for flux coupling and since the magnets are already so close the coils there is no point. any additional metal on the rotor would do little to the flux linkage other than leak power. iron is usually used around coils to help concentrate the flux into a particular axis. around magnets not much happens. consider stiffer thermo-setting plastics such as delrin if possible. delrin is easy to machine even on small mills and lathes.
Have you thought about designing one with an attachable heat sink? There is an article that seems to talk about how motor performance can be negatively impacted by high heat. Maybe you put as many vias in the extra room of the PCB as possible to also help with cooling.
Great video and loads of interesting comments. But was I the only one alarmed at the figure of 80C in the spec of the magnets? I'm aware of the Curie point and that hot magnets lose their power, but never thought it was that low. If I simmer a neodymium magnet in water at near 100C for an hour will it become useless?
Couple questions. How difficult would it be to implement a halbach array in a design this thin? I realize this would require magnets with more optimal shapes than what off the shelf circular magnets provide. Have you also tried using a rotor on each side of the stator? Most axial flux motors, which this motor is, uses that design; I also know some designs, both single and dual sided, use an iron core on the rotors, where other designs do not, so that might be another thing to look into.
All N-type magnets should be 80 °C, higher temp magnets have a M for 100 °C, H , SH, UH, EH and AH types have even higher max temps :) I love those tiny motors you make, have to build one myself just for fun one day ;) Edit: I just found a better list but they are only reference values: type, max usage temp, curie point N 80 °C * 310 °C (*N52 for example max 65°C ) M 100 °C 340 °C H 120 °C 340 °C SH 150 °C 340 °C UH 180 °C 350 °C EH 200 °C 350 °C AH 230 °C 350 °C
You could probably increase torque by sandwiching the rotor between two pcb, not just one on the one side? Another possible improvement: maybe through-holes as air vents
Eddy currents in the aluminium will definitely reduce the efficiency. If you want something that can take the heat, has good mechanical properties and is non conductive use GFRP or in other words: Mill out a thick FR4 PCB. It will be lighter, stronger and more efficient than aluminium if done right.
Congratz on your v3 motor! Yeah that's a hot board. I don't think it was mentioned, but were you using 2oz copper or 1oz over the substrate? Doubling your thickness could quarter your resistance. Something you could try at no additional cost that will improve resistance & cooling is to *remove your soldermask layer over your coils!* 1) At the voltage you're running at (5V), there should be more than enough isolation between turns. 2) Removing the soldermask layer would allow the boardhouse to add surface finish to your coil traces, adding thickness and reducing the resistivity. 3) Removing the soldermask layer will reduce the thermal resistance between air and your traces, improving the heat flux and keeping your board cooler. The aluminum magnet holder could also have some low-profile fins incorporated into the milled aluminum design to add some convection cooling. If you did some 2-sided milling, you could try a 6/8/12 fin design at maybe 1mm depth and 0.5mm thickness (these are arbitrary numbers) then flip the board and add your magnet slots. 1)You'd sacrifice some torque and reduce temperature by having the finned surface facing the coils, and have the magnets sit behind the fins + a very small wall thickness 2)Or You could keep the magnets as close to one side of the board as possible, and add a second milled aluminum piece to the other side of the PCB Lots of directions to take, anyway I look forward to the developments! Great production quality as always. edit: punctuation
Thx for the tip! 🙂 Copper thickness is currently at 1oz.. To increase its resistance I can go for a 0.5oz thickness but I need to check whether my pcb supplier can do that as it is not currently listed on there website.. Regarding the no-soldermask I was very tempted do it for this version.. I also think it would improve the thermal performance but it also remove insulation and risks short circuiting the phases.. But it is something that I am still considering especially for the scenario where the motor is fixed and enclosed in a housing.. The fins idea I am not so sure about as that might cause drag when spinning
@@CarlBugeja That's a pretty interesting idea, but I want to caution against increasing your coil resistance. Increasing resistance might help make control simple but is ultimately holding your torque targets back with the lower current. IMO, controller design and low coil resistances is really going to be what takes your design even further. Look at it this way. Your coil is acting as both a resistor and an inductor, but only the latter generates useful torque. You could achieve the same results by decreasing coil resistance to 0 and adding all the series resistance in off-board resistors to regulate current. The coil would generate identical magnetic flux and the heat would be localized in the resistors where it can be managed without damaging the motor. Since your design uses an air core, you're not going to be facing any magnetic saturation if you increase current, so the cooler you can get your coils, the higher you can push the envelope with your current and generate greater torque. If you added a current sense resistor and sampled the voltage drop, you could add some current control by regulating about a current setpoint using PWM, add some overcurrent protection against shorted coils (if you go with the no-soldermask option), and have lower coil temps for the same torque and start pushing the envelope even higher. A little PTC chip on the board could add further protection and allow you to throttle the coil current when temperatures get too high to protect your magnets. And yeah, having a fan will definitely add drag as you use some of that mechanical power for cooling. The best option is to just reduce the coil resistance as much as you can so you don't even need to resort to cooling strategies.
@@InfinionExperiments That is a very good list of design improvements with rationales! I guess PWM will be used anyways (it makes no sense controlling speed by phase frequency alone) so minimizing the coil resistance and lowering the PWM duty cycle will increase efficiency. An extra upside would be that there is torque headroom for short bursts if needed.
Did you consider making the rotor out of sheet steel to work as a back connector for the magnetic field? Or using a double rotor with the pcb sandwiched inbetween to build an axial flux motor? Both variations might result in stronger magnetic fields and thus higher torque
The steel option I believe would reduce the speed of the motor (which was the main goal) but as you said it will also improve the torque.. So I might consider it for some other applications
When designing a motor, you want the winding resistance to be as low as possible in order to minimize losses. Since the motor runs on AC, the current will be limited by the winding inductance. You could drastically increase the efficieny and torque of your motor by using an iron core and rotor, thereby decreasing the magnetic resistance and increasing the inductance.
I think thats a good Idea to try. And its Not the first time people Point it Out in PCB Motors. The question ist how to add the Iron core. On think is you have the place normally used by the vias connecting one Side of the coil. Other Problem that could you use as Iron core.
Hi Karl, you have magnets, winding, bearing, but you do not have a cores which is key in motors. basically iron to close and guide the magnetic field and use the best of your coils, imagine an electromagnet without core, it would be many times less strong ! At least it will improve the torque. PM me if you don't see what I mean.
Can you add a second rotor like the 'dual rotor axial flux motor'. That should improve torque. They are already using this design in YASA and MAGNAX. Just out of curiosity ;-) [ Fan of your work. Keep Improving ]
Is there any way to add a fan or directional air flow holes into the aluminum to cool the PCB? Maybe in the divots or area between the magnets. I know nothing of this just trying to help.
why not consider an aluminium (or alike) heatsink on the backside of the PCB? (also, weighing magnets might result in unpredictable results, as the magnets may pull toward possible metal inside the scale, altering the measurements thereof)
Why not make a bigger pcb with bigger coils so they can carry more current and generate more torque? Also stack the pcbs on a single shaft so torque is improved, maybe some heatsink as well to cool the rotor. If major cause of heating is the eddy currents, go ahead and use stacked style in trnasformers.
WOW PCB motor is great idea for making switched reluctance motor. It does not contain magnets. It could be 2 PCBs. One you have and second with coils, capacitors and possibly some logic. It would run at much higher temperatures without magnets, processing much more energy. good for quad.
Very interesting and fascinating. I personally would like to see some simulation of the generated magnetic field. Maybe one can throw an optimizer at the shape of the spirals to optimize the generated field? From the previous video it looked like you choose the shape quite arbitrary? I would be surprised if the shape wouldn't have a big influence over the power output of the motor. But i'm just a commenter on the internet that doesn't really know much about that topic, so...
You could increase the torque even more, like, A LOT, by putting a piece of iron on top of the rotor, right on the back of the magnets. Brushless motors have so called "back iron" which is the iron ring where the magnets stick to. Yeah that's not alluminium. The magnetic circuit otherwise closes behind the magnets in the air which has a very high reluctance. You can cut such a piece from the iron of an old transformer with big scissors. Since this piece of iron will provide structure you could save part of the gained weight by slimming the aluminium part. This is my field of expertise.
Thanks for the tip! I will research this topic further 🙂
Would it add iron losses? If not, what's the deal with "back ironless" motor designs? And what about Halbach arrays?
Also, would sandwiching the stator between two rotors be a good idea?
@@CarlBugeja I agree with nasone32's general idea. Part of the problem is that the magnetic flux generated by one of your windings is weak because most of the path from the north to south of the winding is through the air. So in concept, I think what you want is to equip each winding with a "C" shape iron piece, where the winding and the rotor sit in the gap of the C. The C supplies a much more susceptible path for the coil field, hence the flux is higher. So, a total of six C pieces, one for each winding. This should also mean that when the rotor is spinning, it generates a far higher back-emf, resulting in a lower effective EMF applied to heating the coil. Just the fact that your coils are getting very hot is a sign that a major proportion of the energy you're putting into the coils is not being used to drive the motor. (However, If the motor's stalled, then the windings will get hot by straight I^2 x R anyway, because there's no back emf.)
@@revimfadli4666 it would add very minor losses, the magnetic field at the back of the rotor doesn't change much at all.
coreless/ironless motors have no material in the stator which sees a very rapidly changing magnetic field, unlike the field in the rotor.
a halbach array would work, but packaging in such a tight space would be problematic.
@@bgdwiepp I see, thanks!
What about adding another rotor so the stator gets "sandwiched"? Would that improve the performance?
And would using a back iron mean adding a Halbach array afterwards basically does nothing?
The delta would be usable with half the voltage.
Scaling up motors to the industrial scale, 3 phase motors come as user configurable as Star or Delta.
In Australia, the single phase power is 240Vac while 3 phase is 415Vac.
When using a single phase to 3 phase inverter to drive a motor, it is normal to connect the motors in the delta config. but when used on the 415Vac 3 phase, we connect them in the star config.
In your demo, you were using 5V, so there is no reason why it shouldn't work at 2.5V in the delta config.
Nice job by the way.
Your work is super cool Carl. Keep it up!
I can smell a quadcopter business card on the horizon
Hmm is getting better and better! Good job!
Oh my electronoobs on a Maltese video wow wow 😮
Only 6 likes after 6 months
Kinda weird
Carl, your work is really amazing, especially because you do innovations. Infact your videos, compared with the others, are really mind stimulating. Thank you so much ☺️
Thanks so much 😊
Cool stuff. The first thing that comes to mind in terms of practical applications when seeing this version is the LSU polygon motor in a laser printer. It's a motor that spins up a polygon mirror attached to it. The mirror then deflects the printer's laser beam to make it cycle through the page in a predictable manner.
I think you'll need at least 15000 RPM for this motor to be practical in that role, and that speed needs to be sustained and incredibly stable. On the upside - no need for high torque whatsoever. With your setup, the mirror and the magnet housing might as well be the same piece, so the only thing the motor would have to spin up is itself.
The same type of mirror unit can be used in a variety of laser devices, such as laser projectors or LIDAR-like sensors.
That was excellent, and I appreciate your sense of humor integrated with the technical information. I wish I found your channel sooner.
I like the way you build your motors. *HAMMERS EVERYTHING*
I have watched a few of your videos, and I absolutely love them😃 You make videos and projects like no one else keep it up!
You are such a pleasure to watch and learn from. We send our best wishes to you.
This is much look better than last time
This channel is underrated.
Excellent work. Top of the world technology development. Keep rolling.
Have you considered making increases to the rotor field strength? Would help with coupling and certainly improve efficiency ( = less heating! :D)
For example, you could put magnets on both side of the stator, or use a Halbach array like Cristoph Laimer's 3d-printed motor.
Very good progress! As stated by others, you need back iron to couple the magnets and complete the magnetic circuit. This will also keep the magnets from being demagnitized as easily.
love the fact that you use your skills to create original new stuff, i can relate.
and the coolest diy i've ever seen
Very cool work. Just wondering if the switch to using aluminum for the rotor causes eddy currents, and if that might reduce the efficiency of the motor?
Thanks! Hmm 🤔 very good question.. I don't think eddy currents will reduce the efficiency with regards to speed and torque.. But in reality they might be also getting generated in the bearing and shaft... That is one of the reasons why I prefer this over the old design because the bearing is now inside the rotor and not directly underneath the rotor's magnets
@@CarlBugeja I actually did research in electrodynamic wheels(basically a wheel with circumferencially oriented magnets in a halbach array). We found that all metallic objects created a large eddy current that opposes the field of the magnets. The magnetic field generated in the aluminum will have a phase to the magnetic field of the magnets passing over it, such that as your RPM increases, you will double up or divide your magnetic field strength. Using PVC is a much better idea.
blogs.nvcc.edu/intercom/2014/03/24/majewskis-students-win-research-competition/
Oh good thinking there, it’s probably why the magnets got so hot. I’d expect the rotor to remain cold. Perhaps make a silicone mould of the milled rotor and pour in some epoxy, maybe with glass beads or fibres for reinforcement.
@@TurkishLoserInc Yeah the aluminum will definitely produce eddies from the magnetic fields and generate heat + counter-torque from their induced currents, this is why motors are produced in laminated sheets. The only question is how much drag and heat is being generated at these magnetic flux levels.
You could figure out how much drag there was experimentally by changing the rotor back to plastic or milled carbon fiber and measure the RPM differences between it and the aluminum rotor, and you could test if the heat was the result of eddies or the PCB by literally heating the PCB with DC current (no eddies) through the coils while the rotor is stationary, and see if that heat can make its way through the bearings and shaft and get the rotor past 80°C. If it can't, then heating must be coming from eddy current losses.
@@CarlBugeja I can see that it might not have much effect at this scale but if scaled up it might factor in. also something to note is that the magnetic fields in the permanent magnets are a lot stronger than the electromagnets, and the fields from the electromagnets are the ones that will induce the eddy currents. so they should not be a concern.
2:01 "GIMME PCB!!"
Ehi for the heat, what about a little fan (maybe in plastic) on the back of the motor pcb? The height increases, but the performance will decrease less.
Or maybe a different design of the rotor: fan shaped, so it will cool itself without any other extra parts.
Superb video as always 🤩. Waiting for its final prototype. Keep it up sir
You can re magnetize the magnets by heating them and exposing them to a magnetic field then cooling them. You can also add a second pcb on the other side of the rotor to double the power of your motor. Just add a longer shaft. That is the beauty of an axial motor.
Awesome idea! This could also be really interesting to make a linear motor on a PCB. I figure PCB milling should be quite precise enough as a crude linear guide. You could make the coils wider horizontally, add a chip to only power the active coils and maybe add some iron to increase the force. I don't really have much knowledge about motors though.
This could be cool for something like a stewart platform. 6 legs just with their distance to the ground would define the 6dof position of the robot. So with 8 legs that can freely swing you should be able to make a simple robot that can lift two legs while the other 6 keep is fixed, then move all platofmr, lift 2 legs and let them swing back to vertical.
I wonder if you could sandwich the rotter between 2 PCBs on top and 2 PCBs on bottom. This will trap your spindle between 2 PCBs on either side (higher stability) . Then use the outer exposed disk for mounting blades to instead of shaft. Might be interesting to do push pull between top and bottom PCBs due to heat. Also if u cut small curved vain into the aluminum you can have it self cool to a point.
Excellent job
Consider that aluminum will have eddy currents through it and consider halbach array of magnets too! Just a thought, don't judge :)
I really like this project. I'm going to be thanks about tall the possible variations and potential design changes all day now.🤯
i love that miller!
Would probably really benefit from a ferromagnetic core in the windings somehow. Maybe with vias and a ground plane? Connecting all the magnets together in a ground plane, to create a closed loop magnetic field. Not having one is like putting resistors in your magnetic field, the field needs something to kling on to.
Very nice!, except for adding iron to define magnetic path, maybe also try rotor with no indentation to minimize air resistance or drag, the loss is related to drag and square speed.
With a proper driver and commutation using hall sensors, the basics involved are this:
Current = Torque. Speed = Voltage.
A proper motor driver will reduce current when the commutation is in phase with no lag, so low heat and high efficiency at all speeds. At higher velocity, the current for the torque remains constant if you ignore friction and induced current loss in the conductive magnets.
This motor control can be performed with an inefficient class AB analog amplifier/driver to feed the commuted 3 phase drive to the motor using 3 hall sensors and 4 quadrant analog multipliers to generate drive signals to the amplifier which is configured not as a voltage amplifier, but a voltage to current amplifier. Motor drag, is responded with a change of current to maintain the correct speed by adjusting current to change torque. As the motor runs slow or faster, the voltage to current amplifier automatically adjusts voltage based on motor back EMF to apply the correct committed torque at all speeds.
For higher efficiency, a high speed PWM voltage to current driver can be used instead to produce the 3 phase currents required. If you change from a 3 phase design to a 90 degree 2 phase design, cheap stepper motor drivers can be used to run constant current, thus constant torque and constant heating in the motor at all speeds. Use higher supply voltages for higher maximum speed. To grasp the concepts, stepper motors are inexpensive in smaller sizes along with off the shelf stepper motor drivers. Play with and learn stepper motor driver theory and technology to transfer to your PCB motor designs. You can buy 3 phase versions, but they are less common and more costly.
www.amazon.com/dp/B00PNEQ79Q?tag=duckduckgo-d-20&linkCode=osi&th=1&psc=1
Stay away from the arduino stepper H Bridge drivers. They are made for slow speeds such as 3D printers. To learn motor drive and view waveforms, get a real CNC style stepper motor driver.
www.amazon.com/Usongshine-Stepper-Controller-Leadshine-2-Phase/dp/B07PHLJM22/ref=sr_1_36?dchild=1&keywords=stepper+motor+controller&qid=1592375778&s=electronics&sr=1-36
You set the desired current in dip switches, and your maximum speed is limited by your maximum voltage used. If 1 Amp is too much, you can buy smaller motor drivers.
Very interesting! I like the direction this is going.
Have you considered putting thermal epoxy as a heatsink on rear of pcb where the coil areas are?
Also you can get magnets that are rated to 125+ celcius. "Polymagnets" can generate greater force than standard magnets, Correlated Magnetics make polymagnets.
thermal epoxy is there just to fill the gaps between a heat source and an alloy heat sink (cause perfect flat surfaces, that could touch each other with 100% of their expanse, don´t exist), and doesn´t help dissipate heat itself!
@@klausbrinck2137 yes an actual heatsink would be superior with thermal epoxy to bond and provide optimal thermal connection. But thermal properties can be improved with just thermal epoxy, it still wicks heat away fairly quickly and increases surface area.
Have you seen microcontrollers on a pcb under a pool of set thermal epoxy?
Thermal epoxy can be cast or overmolded into shapes like normal resin.
> Have you seen microcontrollers on a pcb under a pool of set thermal epoxy?
Do you mean the black blobs common in very cheap applications? This isn't for thermals, if's for protecting the die.
You could also stack 2 coil pcb's on top of each other, increasing the torque by 175%?
Nice idea. He should defently try it. U could use it as the main structure of the drone as well. I can already see the design :)
I was thinking the same thing. 1 on each side of the magnet. Or 2 on each side of the magnet!
If you set rotor whis ferromagnetic or magnets on opposite side of coils pcb, it increases torque more then two times.
There's the right way to set a bearing with a nice press that doesn't put undue stress on the bearing, and then there's they way everyone actually does it, which is HAMMER :D
You have a really charming style! Keep it up!
#60.
These motors are very cool and have so many uses. It's hard to find good info about them, so Thanks for making these vids.
Any room for heatsinking the back side of the PCB?
If you mill another magnet rotor, maybe cut some grooves / fins into it so there is a slight cooling effect without much if any extra drag.
If you put a lot of coils onto a lager pcb, could you make something like a marble track? Or a clock, where the small metal ball shows the time.
The low resistance of the delta configuration is not a problem, it is necessary for higher speeds but you will need a smarter driver that has current limiting
Much Respect 💪
Good research Carl 👏👏
Solder two or three on top of each other for more effective windings! Would make the delta more usable.
In that case maybe go with more layers?
But he would have to put them in parrallel since else the current would be half as much leading to the same torque since B=μ*N*I/L
if N stays the same and I is getting half and L is staying the same the magnetic field per PCB would be half as much. And due to 2 PCBs the magnetic field would stay the same.
But if he put it in parrallel there would be a magnetic field being stronger.
neutronen stern
Series would mean a higher total impedance, which would make the coils easier to drive with less losses in the driver, or prevent the coils overheating at 100% duty-cycle at the fixed voltage like we were seeing earlier. Not that he should be forced to run the coils at 100% duty-cycle, that is.
1-2 pcb on each side of the rotor? more surface for heat dissipation, for the higher-current case (delta?) ? anyway not all pcb´s on 1 side, cause the small magnetic fields (of he windings) would never reach through to the rotor-magnets!!
There might also be an advantage in sandwiching the magnets between two PCB's in being able to direct air across the sandwich of PCB's and magnets and thereby increase cooling efficiency.
I like to use such motor in a cassette tape housing (on a custom PCB inside) that turns the wheels or use one of them as generator to drive the second wheel. The magnets are in the surroundings of the wheels and coils are around the holes. The reason I want to do this is to make BT-adapter without noisy rotating gears. Some cassettedecks requires the motion of both wheels, that is why there are gears inside such adapter. It would be fantastic when this can be done without batteries at all, so when one wheel is driven by the cassettedeck, the other automaticly spins (slower of course due energy loss). That would be amazing! A nice application, right? (or stupid idea?)
I know this is an old vid... however, try to incorporate fan blades into the rotor to try cooling the PCB
Can you please tell about Calculations regarding BLDC Motor design and windings?
It looks more stable than the last one!
Maybe there's a way to sandwich rotor between two pcb stators which should increase the torque (probably).. So for drone it might be possible to have 2 pcb stators on both sides of a rotor. but what if combine two sandwiches together? 2 rotor 3 stator configuration.. not sure what the benefit of that would be.
Try to add magnets at different side of pcb for more torque. And sure avoid the high temperature. Also try to insert copper inside of pcd. It's a multilayer pcb like used in graphical card where we have more than 7 layers.
he already has multiple layers connected with vias. You are right though about stacking. He should stack a stator on the other side of the rotor though so that the magnets are acted on from both sides.
@@excitedbox5705 I saw only two layers by two sides of pcb plate. Need more)
It is actually 4 layers 🙂
yeah the new pcb motor look more promising. Good Job !
To get the most out of your magnets you need to confine their flux to pass through the stator as much as possible. To achieve this you should use a magnetically permeable material (a steel washer perhaps?) to back the rotor and make the rotor double sided.
If you could source them, halbach arrays would be ideal. Not sure how realistic that is at this scale.. Trapezoidal or sector magnets would also let you use the space under the disc more efficiently. Maybe you can get the HDD magnets from an oldschool iPod or something.
Given the choice of using more of the space you have for magnets or just a higher grade, more magnetic material is usually the way to go. imho.
You would need need "Water cooled PCB motor" 😂
What about creating the rotor with fins, to aid cooling & prevent overheating of the magnets? Also, have you looked into graphite as a washer option?...
consider changing the magnet housing to a non-conductive material (other than plastic of course) or use laminated sheets of aluminium epoxied together perpendicular to the pcb. the eddy currents in the housing is giving issues with heat and sapping a lot of the power. adding iron would make this worse. iron is only good for flux coupling and since the magnets are already so close the coils there is no point. any additional metal on the rotor would do little to the flux linkage other than leak power. iron is usually used around coils to help concentrate the flux into a particular axis. around magnets not much happens. consider stiffer thermo-setting plastics such as delrin if possible. delrin is easy to machine even on small mills and lathes.
So what if you placed two magnet arrays on either side of the PCB? Like a reversed axial brushless motor?
Yes! I've been waiting for another video on pcb motors. Still want to do a pcb drone.
Keep up the pcb motor!
An the great and helpful work.
What milling machine is that you are using? Also Great video
I wanna live like this guy.
Excellent video 👌
Impressive work!
Costs a bit more, but use a 4-layer (or more) PCB, and maybe some iron-covers around the backside of the coils.
Have you thought about designing one with an attachable heat sink? There is an article that seems to talk about how motor performance can be negatively impacted by high heat.
Maybe you put as many vias in the extra room of the PCB as possible to also help with cooling.
Awesome Idea :) , next step is to build smallest PCB step motor and next 3d printer from them, and next print small 3d printed engine :)
You get 100 likes from me..!!! Keep it up brother 🙂👌👌
Great video and loads of interesting comments. But was I the only one alarmed at the figure of 80C in the spec of the magnets? I'm aware of the Curie point and that hot magnets lose their power, but never thought it was that low. If I simmer a neodymium magnet in water at near 100C for an hour will it become useless?
Couple questions. How difficult would it be to implement a halbach array in a design this thin? I realize this would require magnets with more optimal shapes than what off the shelf circular magnets provide. Have you also tried using a rotor on each side of the stator? Most axial flux motors, which this motor is, uses that design; I also know some designs, both single and dual sided, use an iron core on the rotors, where other designs do not, so that might be another thing to look into.
All N-type magnets should be 80 °C, higher temp magnets have a M for 100 °C, H , SH, UH, EH and AH types have even higher max temps :)
I love those tiny motors you make, have to build one myself just for fun one day ;)
Edit: I just found a better list but they are only reference values:
type, max usage temp, curie point
N 80 °C * 310 °C
(*N52 for example max 65°C )
M 100 °C 340 °C
H 120 °C 340 °C
SH 150 °C 340 °C
UH 180 °C 350 °C
EH 200 °C 350 °C
AH 230 °C 350 °C
You could probably increase torque by sandwiching the rotor between two pcb, not just one on the one side?
Another possible improvement: maybe through-holes as air vents
You'd be better off sandwiching the pcb between two magnetic rotors, and adding back-iron to help the magnetic fields flow.
Wye and delta are completely equivalent except for the voltage which differs by a factor of 1,73 (sqrt(3))
Could you please share a video using PCB motors in order to achieve similar functionality as that of celluveyor
nice work, interesting research
You should use magnetics simulation such as JMAG or Maxwell (expensive) or MotorAnalysis (free) to simulate and optimize your design further.
Eddy currents in the aluminium will definitely reduce the efficiency. If you want something that can take the heat, has good mechanical properties and is non conductive use GFRP or in other words: Mill out a thick FR4 PCB. It will be lighter, stronger and more efficient than aluminium if done right.
Smash assembly, I love it!
On the metal thing that holds the magnets, couldn't you add some fins to the back of it when it spins to help cool it?
Super I like your project
Could you protect the magnets with a thin layer of thermal insulation? Maybe a paint with aerogel powder or insulating film.
Proset King :)
Congratz on your v3 motor!
Yeah that's a hot board. I don't think it was mentioned, but were you using 2oz copper or 1oz over the substrate? Doubling your thickness could quarter your resistance.
Something you could try at no additional cost that will improve resistance & cooling is to *remove your soldermask layer over your coils!*
1) At the voltage you're running at (5V), there should be more than enough isolation between turns.
2) Removing the soldermask layer would allow the boardhouse to add surface finish to your coil traces, adding thickness and reducing the resistivity.
3) Removing the soldermask layer will reduce the thermal resistance between air and your traces, improving the heat flux and keeping your board cooler.
The aluminum magnet holder could also have some low-profile fins incorporated into the milled aluminum design to add some convection cooling.
If you did some 2-sided milling, you could try a 6/8/12 fin design at maybe 1mm depth and 0.5mm thickness (these are arbitrary numbers) then flip the board and add your magnet slots.
1)You'd sacrifice some torque and reduce temperature by having the finned surface facing the coils, and have the magnets sit behind the fins + a very small wall thickness
2)Or You could keep the magnets as close to one side of the board as possible, and add a second milled aluminum piece to the other side of the PCB
Lots of directions to take, anyway I look forward to the developments! Great production quality as always.
edit: punctuation
Thx for the tip! 🙂 Copper thickness is currently at 1oz.. To increase its resistance I can go for a 0.5oz thickness but I need to check whether my pcb supplier can do that as it is not currently listed on there website.. Regarding the no-soldermask I was very tempted do it for this version.. I also think it would improve the thermal performance but it also remove insulation and risks short circuiting the phases.. But it is something that I am still considering especially for the scenario where the motor is fixed and enclosed in a housing.. The fins idea I am not so sure about as that might cause drag when spinning
@@CarlBugeja That's a pretty interesting idea, but I want to caution against increasing your coil resistance. Increasing resistance might help make control simple but is ultimately holding your torque targets back with the lower current. IMO, controller design and low coil resistances is really going to be what takes your design even further.
Look at it this way. Your coil is acting as both a resistor and an inductor, but only the latter generates useful torque. You could achieve the same results by decreasing coil resistance to 0 and adding all the series resistance in off-board resistors to regulate current. The coil would generate identical magnetic flux and the heat would be localized in the resistors where it can be managed without damaging the motor.
Since your design uses an air core, you're not going to be facing any magnetic saturation if you increase current, so the cooler you can get your coils, the higher you can push the envelope with your current and generate greater torque.
If you added a current sense resistor and sampled the voltage drop, you could add some current control by regulating about a current setpoint using PWM, add some overcurrent protection against shorted coils (if you go with the no-soldermask option), and have lower coil temps for the same torque and start pushing the envelope even higher. A little PTC chip on the board could add further protection and allow you to throttle the coil current when temperatures get too high to protect your magnets.
And yeah, having a fan will definitely add drag as you use some of that mechanical power for cooling. The best option is to just reduce the coil resistance as much as you can so you don't even need to resort to cooling strategies.
@@InfinionExperiments That is a very good list of design improvements with rationales! I guess PWM will be used anyways (it makes no sense controlling speed by phase frequency alone) so minimizing the coil resistance and lowering the PWM duty cycle will increase efficiency. An extra upside would be that there is torque headroom for short bursts if needed.
Minor correction - 0.5oz will get a larger resistance than using 1oz
Did you consider making the rotor out of sheet steel to work as a back connector for the magnetic field? Or using a double rotor with the pcb sandwiched inbetween to build an axial flux motor? Both variations might result in stronger magnetic fields and thus higher torque
The steel option I believe would reduce the speed of the motor (which was the main goal) but as you said it will also improve the torque.. So I might consider it for some other applications
Could you put coils on the other side, too?
Maybe someone told you, but you didn’t think to place a small impeller on the shaft for active cooling of the motor?
just realized that this is impossible because pcb is a stator.
@@dimakrayushkin It is still possible to cut the rotor into a fan shape to blow air across the stator as it spins.
Having sort of holes inside the rotor would make a better job I think
That is great work)) would like to see something like this underwater 🤔
Try 2pcb circuit parallel to each other and magnet in between
I you are looking for interesting project ideas: I would love to see a miniature linear induction motor with a solid secondary
When designing a motor, you want the winding resistance to be as low as possible in order to minimize losses. Since the motor runs on AC, the current will be limited by the winding inductance.
You could drastically increase the efficieny and torque of your motor by using an iron core and rotor, thereby decreasing the magnetic resistance and increasing the inductance.
I think thats a good Idea to try.
And its Not the first time people Point it Out in PCB Motors.
The question ist how to add the Iron core.
On think is you have the place normally used by the vias connecting one Side of the coil.
Other Problem that could you use as Iron core.
Hi Karl, you have magnets, winding, bearing, but you do not have a cores which is key in motors. basically iron to close and guide the magnetic field and use the best of your coils, imagine an electromagnet without core, it would be many times less strong ! At least it will improve the torque. PM me if you don't see what I mean.
Great content! I am wondering how this design would work as a power generator? Maybe this is an idea for another video? Cheers
What about the Eddy currents in the aluminium part? Wouldn't they create drag?
Could u flip this and use it as a generator? Wonder how much to would put out?
I was wondering the same thing
Can you add a second rotor like the 'dual rotor axial flux motor'. That should improve torque. They are already using this design in YASA and MAGNAX. Just out of curiosity ;-)
[ Fan of your work. Keep Improving ]
Good genius! 👍
So amazing
Is there any way to add a fan or directional air flow holes into the aluminum to cool the PCB? Maybe in the divots or area between the magnets.
I know nothing of this just trying to help.
great job, keep going and waiting for dron
why not consider an aluminium (or alike) heatsink on the backside of the PCB? (also, weighing magnets might result in unpredictable results, as the magnets may pull toward possible metal inside the scale, altering the measurements thereof)
i will love to see a video of your setup like which software you use or which tool you recommand :)
Why not make a bigger pcb with bigger coils so they can carry more current and generate more torque? Also stack the pcbs on a single shaft so torque is improved, maybe some heatsink as well to cool the rotor. If major cause of heating is the eddy currents, go ahead and use stacked style in trnasformers.
WOW PCB motor is great idea for making switched reluctance motor. It does not contain magnets. It could be 2 PCBs. One you have and second with coils, capacitors and possibly some logic. It would run at much higher temperatures without magnets, processing much more energy. good for quad.
Very interesting and fascinating.
I personally would like to see some simulation of the generated magnetic field. Maybe one can throw an optimizer at the shape of the spirals to optimize the generated field? From the previous video it looked like you choose the shape quite arbitrary? I would be surprised if the shape wouldn't have a big influence over the power output of the motor. But i'm just a commenter on the internet that doesn't really know much about that topic, so...