This is exactly the kind of content I come here for. Definitely adding "DIY linear motor based CNC mill" to the list of things I'd like to do but probably never will 😊
I am desperately awaiting the rest of this series! I have been wanting to use linear motors for a while. I just feel like without seeing one built and control figured out from start to finish I won't understand well enough to integrate it into a CNC machine and actually benefit from the linear motors. This is a fantastic video and you are terrific at explaining the motor side of it. I am already pretty well versed in electrical motors and how they work mechanically its just control wise I have a hard time with understanding coding for control loops. Obviously this just used a pre programmed timed switching on and off of coils to move it along rather than hall sensors or an encoder but I can tell its coming and I am STOKED! I'm going to be 30 minutes late for work because I saw this video and couldn't resist!
4:56 just a Tip: this center drill is not designed to be used for starting holes for drills but for the live center for lathes. you should get a spot drill with a higher tip angle than the drills you are using.
An explanation on *why* it's wrong would go a long way here. But from what I can find, it's because the sharper tipped ones cause the initial contact with the next bit to be used to deepen it, is harsh on the edges and more prone to causing them to fracture? I feel like a 90 would give better location results than the flatter 120 that is more like a traditional bit.. likely just avoid a 60 degree one, his looks like a 90, but maybe not.
@@UncleKennysPlace that makes more sense thinking into it, an edge getting caught means the drill will immediately try to 'walk' from the first spot it catches. Got it! Thank you for providing usable information.
@@Roobotics Center drill = 60*, regular drill = 118*, spot drill =120* (or larger). A 118* point will center on a 120* hole. A smaller angle will always self-center into a larger angle. Not the other way around.
@@chip-load that is fair, I think my brain has dyslexia when thinking about angles but as long as it's a 'needle' hitting a 'cone' it will tend to center itself, and the other way around can cause the cones edges to engage the sides of the hole and apply side-loads.
Pretty killer first video, im looking forward to your future here. Have you looked in to the Kern machines for inspiration? IIRC they use linear motors in some of their machines x and y axes. Paired with hydrodynamic slideways it creates a very rigid and low static friction system. Diy hydrodynamic ways would be a very cool addition, but certainly adds way more work than linear rails.
Well yes but it is really expensive and hard to do but maybe igus or pbclinear plastic bearings make sens for these machines. I would just stick pre-loaded bearings and rails for this project.
This has got to be one of the most successful first videos I've ever seen, when I finished this I went to your channel to go binge a bunch of videos and to say the least I'm astounded this is your first video. The production quality is amazing!
Excellent start. For the stator magnets, you will increase field strength (and therefore axis stiffness, ultimately) if you use a Hallback array configuration. That basically means additional magnets between the ones you already have, but with the field oriented along the axis, not perpendicular to it.
I've been keeping an eye on linear motor videos for a while and now this pops up. RUclips algorithm knows what i need 😅 Cant wait for the next episode 👍👍
Your english is fine. In fact better than most! Playback speed at 75% and subtitles makes a big difference. The issue is you need a better microphone! Far LESS base and more treble will make your voice much clearer. Great 1st post and looking forward to interesting engineering.
Hi thank you for your constructive feedback. That really helps to improve the content. In fact I bought a rode mic in the final hours of editing and after hearing a sample I redid the whole VO 😅. I’ll change the equalizer settings and try to speak slower and more clearly😁
@@NeverGetTiredEngineering OK! You have a pleasant voice. Just Speak s l o w e r . . . please. Anyway, very interesting content, I'm looking forward to seeing how it all turns out. I'm rather certain you need iron core in those coils. Even if it is solid iron rather than the preferred laminated form - at least for prototype. Good fortune with your projects!
Interesting design you did there! I woud suggest to increase the thickness of your steel plate. They are typically about 5mm thick for that size of motor. Generelly id orient myself on an existing motor in therms of propostions, magnet size and pitch and so on. Look at technotion, akribis or hiwin for example. Maybe the technotion tl6 or tl9 is a good starting point. You may wanna reduce magnetic pitch and increase the amount of poles to 6 or 9. Also please share more of your design so people can make suggestions. So far ive seen that you wanna use 4 rails for the Z axis? That is a heavy overconstrain. How do you plan to solve that? Flexibility in the frame?
Hey, thank you! I will take the thickness into account and gather more information about some commercial products. You're right! I somehow totally overlooked that 😅. I think I should really do an FE analysis to determine what's necessary. With a high enough second moment of area, I can probably dump two rails. I will also make everything open source as soon as possible.
Most linear motors use a mag-spring or counter-weight for vertical applications to reduce static load. Take a look at a LinMot solution for inspiration. Looks great so far tho!
Very well made video as well as great project, really looking forward to the next one. A couple recommendation for the design, with the industrial linear motors that I have seen on a few pick and place machine have a gap between the forcer and the stator is only a couple hundreds of a mm (plus the 0.1mm coating over the magnets and the coil). This should help with the motor power (which will likely be critical on such a small motor). Another recommendation as well, would be some kind of water cooling for the coil, as heat is going to be a major problem with a small forcer of that size. The industrial unit use an aluminum extrusion profile as the main cradle for the forcer (milled on the back as well too for improved parallelism). The magnet tracks have a ground steel back and then they mount onto the machine frame with very uniform height. As for machining if you want to make sure that your linear rail and magnet track is very parallel I recommend for simplicity that you machine them both on the mill in a single mounting. Mind that when dealing with plate stock like this the best way would be to not use a vise as the vise will jaw lift the part as well as bend the part which will result in the top and bottom not being flat or parallel. What I would do is mount a sacrificial plate or blocks, that will lift the part high enough for machining and then you face them to insure its parallel with the axis travel. Then you just drill mounting holes into the blocks/plate and bolt the plate down with countersink or counterbored screws. That will allow you to machine both the top as well as the perimeter of the part (then flip the part and deck the bottom). While this does not guarantee the part will stay flat when unbolted it will guarantee uniform thickness, which when bolted onto a precise mounting surface will flatten the part. It will also be a more solid mounting which will reduce the chatter due to the part bending under the cutting force (as well as allow for larger parts), so I would really recommend doing it that way. Sorry for the wall of text and I hope it helps!
This should be a fun project. Not sure how you are going to get the torque needed for milling but it would work great for a non-contact machine like a 3d printer. Great start for a channel!
Cool project ! Did you already find a linear encoder to use? -Btw, commercial linear motors use a stainless steel backplate, mostly for magnetic shielding. I'm wondering if aluminium works the same.
Big thanks 😁 To be honest, I haven't really looked yet. If you have a recommendation, I'd love to hear it. Otherwise, I would have used an incremental encoder with a belt or borrowed a linear encoder from my mill. Aluminum has minimal influence on static magnetic fields. The magnets are mounted on a steel sheet, which should both shield them and reduce magnetic resistance.
This is epic. Are you planning to open source your design? I think your biggest challenge is actually going to be achieving thr required precision, not the forces. The magneto-x printer has custom driver boards for the linear motors.
Hi and thanks for the feedback, I never thought that this project would skyrocket like that. I will publish all my CAD files and so on as soon as possible. I’m also curious about both precision and force.
Im really interested to see how you handle encoding. If you're spinning your own firmware, SimpleFOC could be a good reference library for understanding the mechanics of FOC if you haven't done it before. I think people have used it for linear motors before as well. @NeverGetTiredEngineering
@@NeverGetTiredEngineering Looks like the pro's have a single piece laminated ferrite core that magnetically connects all 3 coils (can probably make with laser cutting: ruclips.net/user/shortsQSZ_Pyt_HJo And it looks like the backing plate for the magnet rail is made from mu-metal :O (I have dissassembled a lot of old harddrives to recover their magnets and that is what mumetal looks like. ruclips.net/user/shortsVswyNilqWP8
Hey @@Peter-898, I planned using the motor control SDK from ST (www.st.com/en/embedded-software/x-cube-mcsdk.html). It is more or less plug and play. But SimpleFOC might be a good alternative. I'll try both
Hey! Thanks for subscribing and for the kind words! 😊 Things are pretty hectic at work right now. But don’t worry, I’m doing my best to get the next video out soon. Stay tuned, it’s coming! 🎬
Haha, yes, my very first video and my first day as a content creator. It's definitely a steep learning curve, but I'm taking it very seriously and I'm happy that the video is being so well received. Big thanks!
I wonder if it's possible to print the core using resin that has had iron powder mixed into it? It may not be the most efficient core, but it could be a quick and dirty solution for initial prototypes.
While it’s an option to try and get silicon steel laminations to make your magnetic out of, you could also try to cast your own ferrite epoxy. But you’d need to calculate if the strength would be acceptable. You could even design water cooling channels to be embedded within the magnetic core, if you want to maximise your force output. I’d recommend whipping up a jig to measure the peak force output of your setup and at what current, and extrapolating from there how much current you’d need in order to get the force required to move a CNC mill without running into chatter or skipping steps. My guess is that direct drive just won’t be able to output enough power with a motor of that size, there’s a reason most CNC mills use screw reductions. Don’t forget to use field-oriented-control to maximise force and minimise force-ripple. Using an excessively high supply voltage will mean higher switching losses, but the FOC controller (I recommend the TMC4671, coupled to a TMC6200) will be able to use that extra voltage to force extra current into the motor when it’s needed to maintain position or speed. Putting silicon steel or ferrite behind the magnets might make a difference too, though less so if you’re using a halbach array. Which you should be using. Any plans for a linear encoder?
I finished the rig and now I’m making a core like you suggested. I’m also using the halbach array and some steel for reluctance forces. The software from ST features FOC. Tweaked the design a little. Ran a couple FEMMs and the minimum force output should be ~170N @2A. I found a rotary encoder in my drawer so I’m using that for now and later switch to a linear one.
Uff so many coils 😅 maybe I’ll give it a shot in the future and compare both systems. But I think done correctly and disconnecting unused coils we could distribute the power losses over a larger area and pump more current through. Great idea
Very Nice! This is the kind of stuff I like to play around with myself, always love finding someone else who is just as interested! take a like and sub!
Nice video, is just I want to become in few years, I just need the space for all the machines and tools, I hope next year moving around to a new place with more space. Now I just have a 3d printer, but I want to build a CNC machine, spooler, hot plate for soldering, and a plastic filament recycler.
@@NeverGetTiredEngineering and heat the windings with a heatgun to ca. 60°-80° Celsius before dropping on the epoxy. That way the epoxy gets water thin on contact and fills any tiny voids, and it hardens faster 😀
Thanks! Good idea. I will try to make an iron core from sheets, but it would also be interesting to see the performance of an iron powder & epoxy core.
Have you thought about the max force the motor is able to produce? I somehow have the gut feeling that this could be not enough for the cnc machine? Otherwise great video!
Yes, I did some very basic and idealized calculations to estimate the force. If the magnetic field isn't homogeneous, which is the case in this setup, the situation gets complex quite quickly. To calculate the force, I used the Lorentz force equation, but we can and should also consider leveraging magnetic reluctance. I plan to experiment with various magnet and steel configurations to optimize the setup. Currently, my biggest challenges are power losses and cooling. So it's hard to say how much current I can pump through the forcer without setting it on fire. Whether the force generated will be sufficient for the CNC machine depends on its specific mechanical requirements and load conditions. Thanks for your input and for watching the video!
You mean the spacing between the magnets? As I use a halbach array in one of the following videos it would be 0. But I can say that the distance between each pole center would be 20mm
Using off the shelf motor controller defies the purpose of linear motors in CNC, which is to prevent vibrations. Ideally you'd have to make an analog driver that provides linear current changes, not a chopped DC one like in all those drivers. In other words you need an audio amplifier on the output & and at least 3 function generators on the input which are 120 degrees out of phase. This can be done digitally, with proper filtering to smooth out the output current. But the power amplifier has to be linear. At least you could try to use some low pass filters to smooth out the chopped DC from the driver. But then this might mess up the current sensing those drivers use to orient their phases.
Hi i get your point. However there is a big drawback when using analog linear amplifiers. Massive power losses in the semiconductors. Fortunately the inductors smooth out the current (of course we cannot fully eliminate the ripples). But we can reduce them by increasing the PWM frequency or the inductances. Additionally the motor‘s mass also dampens these ripples. So the motor is a PT2. I guess that for a whole CNC mill the we would have a massive cooling issue when using linear amplifiers 😬
I sugest you make it tubular linear. make the stator of widing couper so the cable dont have to move and create friction. if dont get to hot you can put the cable inside the tube. and the rotor will be an holow tube with magnets using halbach aragment in a sandwich. so the magnetic field will be focus inside and the outside(outer ring side) no magnetism. powerful and easy. Maybe i need to hink a litel bit more on the stator because is not valid having it as sandwich widings because the exterior has less magnetic force the inside where cable goes. To make it easy can be used as sheet as yours plane instead as tube.but coper down and magnets up and use halbach array aragement to avoid magnetism on top on machinery and empower interaction with the stator. And to see where it is position just lector hall senzor in one side with track detection. i am not goo in electronics but i can see the mechanical deasing.
Нас выращивали денно, Мы гороховые зерна. Нас теперь собрали вместе, Можно брать и можно есть. Но знайте и запоминайте: Мы ребята не зазнайки, Нас растят и нас же сушат Не для того, чтоб только кушать.
It'll be way stronger if you put an iron core in the coils instead of plastic - either laminated or powdered, not just solid iron or eddy currents will be a problem
Hey yeah I am currently building different cores with powdered iron/epoxy, small iron pellets/epoxy and laminated iron core :) Edit: I ran a couple FEMMs and we should expect worst case ~170 N at 2A
Comming soon xD. the last days i finished the test stand and also did a couple simulations and tweaked the design "a little". Next week I have time to start filming and cutting so stay tuned :D
Why didn’t they want to repeat the stepper motor in an expanded form? two rows of plates with teeth and a magnet between them - the lower part, and the same two rows of teeth on top with a coil. Was it really difficult?
if you replace the iron core which conduct magnetics with 3D printed part which assume the magnetic potential greatly, I'm afraid the output force would decrease and the speed and dynamic reaction would become bad sharply, like a Coreless Motor, . so I suggest you should design the magnetic circuit carefully and concentrate the magnetic potential to the air gap between the coil and the permanent magnets, just like a real motor
hey, I just used this 3D printed part because I was lazy 😅. Of course I will make a iron core. One will be made from laminated iron and for the other i'll use epoxy and iron powder/granulate.
I'm excited to see how this goes, nice to see a DIY CNC that's not just aluminium extrusion and lead screws
Yeah, I wanna make it as sturdy as possible with a compact design. It will take some time, but it will be legendary. So stay tuned 😁
Marco Reps, built a high end CNC mill, actually he is still building it.
@@Convolutedtubules It's never going to be done because Marco is the good kind of engineer.
Wow - You managed to make it into my home feed with your first video :D
I never thought it would happen so quickly 😆
Yes me too.
This is exactly the kind of content I come here for.
Definitely adding "DIY linear motor based CNC mill" to the list of things I'd like to do but probably never will 😊
This is one heck of a first video. Fantastic work!
I am desperately awaiting the rest of this series! I have been wanting to use linear motors for a while. I just feel like without seeing one built and control figured out from start to finish I won't understand well enough to integrate it into a CNC machine and actually benefit from the linear motors. This is a fantastic video and you are terrific at explaining the motor side of it. I am already pretty well versed in electrical motors and how they work mechanically its just control wise I have a hard time with understanding coding for control loops. Obviously this just used a pre programmed timed switching on and off of coils to move it along rather than hall sensors or an encoder but I can tell its coming and I am STOKED! I'm going to be 30 minutes late for work because I saw this video and couldn't resist!
Hey it is comming :D
I am just very busy at work right now 😅
I’d recommend setting up a brushless closed loop motor with SimpleFOC based on your comment.
I'm lucky to find this channel on its first video
This is completely unhinged and I love it.
😐
What a great prototype! I thoroughly enjoyed seeing the process and learned something at the same time! Cheers!
Glad you liked it, cheers!
4:56 just a Tip: this center drill is not designed to be used for starting holes for drills but for the live center for lathes. you should get a spot drill with a higher tip angle than the drills you are using.
Good to know! I’ll buy one.
An explanation on *why* it's wrong would go a long way here. But from what I can find, it's because the sharper tipped ones cause the initial contact with the next bit to be used to deepen it, is harsh on the edges and more prone to causing them to fracture? I feel like a 90 would give better location results than the flatter 120 that is more like a traditional bit.. likely just avoid a 60 degree one, his looks like a 90, but maybe not.
@@UncleKennysPlace that makes more sense thinking into it, an edge getting caught means the drill will immediately try to 'walk' from the first spot it catches. Got it! Thank you for providing usable information.
@@Roobotics Center drill = 60*, regular drill = 118*, spot drill =120* (or larger). A 118* point will center on a 120* hole. A smaller angle will always self-center into a larger angle. Not the other way around.
@@chip-load that is fair, I think my brain has dyslexia when thinking about angles but as long as it's a 'needle' hitting a 'cone' it will tend to center itself, and the other way around can cause the cones edges to engage the sides of the hole and apply side-loads.
It’s your first vid and I’m already subscribed in the first 45 seconds. I’m excited to see where this goes.
Just popping in to say that I watched your first video before you have a million subs 😎😂 dig the vid
Great content and I liked a lot your work with video editing, music and comments.
Many thanks 😁
awesome video - love that you're taking on the challenge of building your own linear motors - can't wait for more :D
I'm glad to see other people building cnc mills with linear motors. I've been buying parker motors for my mill
Pretty killer first video, im looking forward to your future here. Have you looked in to the Kern machines for inspiration? IIRC they use linear motors in some of their machines x and y axes. Paired with hydrodynamic slideways it creates a very rigid and low static friction system. Diy hydrodynamic ways would be a very cool addition, but certainly adds way more work than linear rails.
Hi thank you, your comment made my day xD. Noted. I'll dig into that
Well yes but it is really expensive and hard to do but maybe igus or pbclinear plastic bearings make sens for these machines. I would just stick pre-loaded bearings and rails for this project.
This has got to be one of the most successful first videos I've ever seen, when I finished this I went to your channel to go binge a bunch of videos and to say the least I'm astounded this is your first video. The production quality is amazing!
Wow, thank you so much! 😄 I wish I had a whole playlist ready for you to binge, but I’m working on it. Stay tuned, more is on the way! 🎥✨
UR A HERO I HAVE BEEN RESEARCHING LINEAR MOTORS FOR WEEKS AND CANT FIND ANY USEFUL INFORMATION
Excellent start. For the stator magnets, you will increase field strength (and therefore axis stiffness, ultimately) if you use a Hallback array configuration. That basically means additional magnets between the ones you already have, but with the field oriented along the axis, not perpendicular to it.
dude you are amazing i just found you at 11 pm but this is fantastic!
I’m really glad you like it. 😁
what an amazing start! already love your videos, looking forward what you'll do next
I've been keeping an eye on linear motor videos for a while and now this pops up. RUclips algorithm knows what i need 😅
Cant wait for the next episode 👍👍
I'm glad to hear that. 😁 Comming soon
Very curious to see how this goes. Great Video 👍
Your english is fine. In fact better than most! Playback speed at 75% and subtitles makes a big difference. The issue is you need a better microphone! Far LESS base and more treble will make your voice much clearer. Great 1st post and looking forward to interesting engineering.
Hi thank you for your constructive feedback. That really helps to improve the content. In fact I bought a rode mic in the final hours of editing and after hearing a sample I redid the whole VO 😅. I’ll change the equalizer settings and try to speak slower and more clearly😁
@@NeverGetTiredEngineering OK! You have a pleasant voice. Just Speak s l o w e r . . . please.
Anyway, very interesting content, I'm looking forward to seeing how it all turns out. I'm rather certain you need iron core in those coils. Even if it is solid iron rather than the preferred laminated form - at least for prototype.
Good fortune with your projects!
Interesting design you did there! I woud suggest to increase the thickness of your steel plate. They are typically about 5mm thick for that size of motor. Generelly id orient myself on an existing motor in therms of propostions, magnet size and pitch and so on. Look at technotion, akribis or hiwin for example. Maybe the technotion tl6 or tl9 is a good starting point. You may wanna reduce magnetic pitch and increase the amount of poles to 6 or 9.
Also please share more of your design so people can make suggestions. So far ive seen that you wanna use 4 rails for the Z axis? That is a heavy overconstrain. How do you plan to solve that? Flexibility in the frame?
Hey, thank you! I will take the thickness into account and gather more information about some commercial products. You're right! I somehow totally overlooked that 😅. I think I should really do an FE analysis to determine what's necessary. With a high enough second moment of area, I can probably dump two rails.
I will also make everything open source as soon as possible.
Most linear motors use a mag-spring or counter-weight for vertical applications to reduce static load. Take a look at a LinMot solution for inspiration. Looks great so far tho!
oh boy this is gonna be interesting! subscribed
Very well made video as well as great project, really looking forward to the next one.
A couple recommendation for the design, with the industrial linear motors that I have seen on a few pick and place machine have a gap between the forcer and the stator is only a couple hundreds of a mm (plus the 0.1mm coating over the magnets and the coil). This should help with the motor power (which will likely be critical on such a small motor).
Another recommendation as well, would be some kind of water cooling for the coil, as heat is going to be a major problem with a small forcer of that size. The industrial unit use an aluminum extrusion profile as the main cradle for the forcer (milled on the back as well too for improved parallelism).
The magnet tracks have a ground steel back and then they mount onto the machine frame with very uniform height.
As for machining if you want to make sure that your linear rail and magnet track is very parallel I recommend for simplicity that you machine them both on the mill in a single mounting. Mind that when dealing with plate stock like this the best way would be to not use a vise as the vise will jaw lift the part as well as bend the part which will result in the top and bottom not being flat or parallel.
What I would do is mount a sacrificial plate or blocks, that will lift the part high enough for machining and then you face them to insure its parallel with the axis travel. Then you just drill mounting holes into the blocks/plate and bolt the plate down with countersink or counterbored screws. That will allow you to machine both the top as well as the perimeter of the part (then flip the part and deck the bottom).
While this does not guarantee the part will stay flat when unbolted it will guarantee uniform thickness, which when bolted onto a precise mounting surface will flatten the part. It will also be a more solid mounting which will reduce the chatter due to the part bending under the cutting force (as well as allow for larger parts), so I would really recommend doing it that way.
Sorry for the wall of text and I hope it helps!
Hey thanks a lot for the great advice 😁
It's great that at least someone took up this issue
Very excited for CNC machine.
👁🗨👁🗨
This should be a fun project. Not sure how you are going to get the torque needed for milling but it would work great for a non-contact machine like a 3d printer.
Great start for a channel!
Looking forward to your next video. Subscribed!
Awesome! Thank you!
Please continue this series. There's almost no good videos on YT showing full complete builds of linear motors.
Hey new video is almost finished. 😁
this is basically the same motor they use to launch roller coaster, lsm motors, i love it!
Boss level geekiness! Love it ...
I'm looking forward to how this will turn out 😁
Very cool, looking forward to seeing where this goes. Subscribed.
Subbed. Good luck with your channel. Looking forward to the next one.
Like carl sagan used to say: To build a cnc machine, you must first invent your own linear motor 😂😂
leiwandes video :)
Great video man! Impressive work!!
Glad you enjoyed it!
i think you need winding stator, but rotor should be permanent, and install halls to determine the position. It will be like bldc but linear.
I am using an incremental encoder (higher resolution than hall-effect sensors) for the test stand. Just finished it. Maybe I'll give it a try.
WTF? where does someone suddenly even show up in the scene with content that has such a factor of high quality?
Very interesting, can’t wait to see more on this project.
Love this project. In hppe for next episodes
Cool project ! Did you already find a linear encoder to use?
-Btw, commercial linear motors use a stainless steel backplate, mostly for magnetic shielding. I'm wondering if aluminium works the same.
Big thanks 😁
To be honest, I haven't really looked yet. If you have a recommendation, I'd love to hear it. Otherwise, I would have used an incremental encoder with a belt or borrowed a linear encoder from my mill.
Aluminum has minimal influence on static magnetic fields. The magnets are mounted on a steel sheet, which should both shield them and reduce magnetic resistance.
This is great! Looking forward to many more videos hopefully 😊
best ending ever, love it
This is epic. Are you planning to open source your design?
I think your biggest challenge is actually going to be achieving thr required precision, not the forces.
The magneto-x printer has custom driver boards for the linear motors.
Hi and thanks for the feedback, I never thought that this project would skyrocket like that.
I will publish all my CAD files and so on as soon as possible. I’m also curious about both precision and force.
Im really interested to see how you handle encoding. If you're spinning your own firmware, SimpleFOC could be a good reference library for understanding the mechanics of FOC if you haven't done it before. I think people have used it for linear motors before as well. @NeverGetTiredEngineering
@@NeverGetTiredEngineering Looks like the pro's have a single piece laminated ferrite core that magnetically connects all 3 coils (can probably make with laser cutting:
ruclips.net/user/shortsQSZ_Pyt_HJo
And it looks like the backing plate for the magnet rail is made from mu-metal :O (I have dissassembled a lot of old harddrives to recover their magnets and that is what mumetal looks like.
ruclips.net/user/shortsVswyNilqWP8
Hey @@Peter-898, I planned using the motor control SDK from ST (www.st.com/en/embedded-software/x-cube-mcsdk.html). It is more or less plug and play. But SimpleFOC might be a good alternative. I'll try both
Very cool, great editing skills for first video !
THX glad you liked it 😊
Will be awesome to follow the progress!
I instantly subscribed after watching this, just to see this is your only video :(
Hey! Thanks for subscribing and for the kind words! 😊 Things are pretty hectic at work right now. But don’t worry, I’m doing my best to get the next video out soon. Stay tuned, it’s coming! 🎬
wait this is your first video? very impressive. definitely subscribing.
Haha, yes, my very first video and my first day as a content creator. It's definitely a steep learning curve, but I'm taking it very seriously and I'm happy that the video is being so well received. Big thanks!
Please just make more videos ! You have a unique style for it.
Thank you ☺️. It’s coming. I'm trying to take a few days off to focus on the next video
Excellent work, subscribed
Now this was great, can't wait for the next part.
Das ist ein extrem Hochwertiges Video!
Schön einen weiteren Österreichischen Kanal zu finden!
Grüße aus Osttirol!
Griaß di! Danke freut mich sehr, dass es dir gefallen hat!
Und genau mein Musikgeschmack auf deinem Kanal. Werde ich durchhören! 😁
LG Tom
@@NeverGetTiredEngineering Perfekt! Viel Erfolg noch mit deinen Videos, werde sicher ab und zu reinschauen :)
Awesome 🤩
I doubt it get enought torque for a CNC but its an awesome project!
Can't wait to see your next video.
Coming soon!
Waiting for the continuation part
Well done, Subscribed.
Thanks :D
Also subscribed - I wanted to build something similar and I'm stocked to see your process 👍🏻
amazing!
Cool project. Keep it up man!
Thanks! I’m glad you like this project 😁
I wonder if it's possible to print the core using resin that has had iron powder mixed into it? It may not be the most efficient core, but it could be a quick and dirty solution for initial prototypes.
I'll try that. I want to compare this iron/resin core, non-ferrus core and laminated iron
While it’s an option to try and get silicon steel laminations to make your magnetic out of, you could also try to cast your own ferrite epoxy. But you’d need to calculate if the strength would be acceptable. You could even design water cooling channels to be embedded within the magnetic core, if you want to maximise your force output. I’d recommend whipping up a jig to measure the peak force output of your setup and at what current, and extrapolating from there how much current you’d need in order to get the force required to move a CNC mill without running into chatter or skipping steps. My guess is that direct drive just won’t be able to output enough power with a motor of that size, there’s a reason most CNC mills use screw reductions.
Don’t forget to use field-oriented-control to maximise force and minimise force-ripple. Using an excessively high supply voltage will mean higher switching losses, but the FOC controller (I recommend the TMC4671, coupled to a TMC6200) will be able to use that extra voltage to force extra current into the motor when it’s needed to maintain position or speed.
Putting silicon steel or ferrite behind the magnets might make a difference too, though less so if you’re using a halbach array. Which you should be using.
Any plans for a linear encoder?
I finished the rig and now I’m making a core like you suggested. I’m also using the halbach array and some steel for reluctance forces. The software from ST features FOC. Tweaked the design a little. Ran a couple FEMMs and the minimum force output should be ~170N @2A.
I found a rotary encoder in my drawer so I’m using that for now and later switch to a linear one.
Did you consider a tubular linear motor design? It seems it might be easier to construct.
Sorry for the late reply, honestly didn‘t think about that, let me think about that
Might save a little time using an ODrive for control and a quadrature magnetic strip encoder like the Peopoly?
Thank you for the suggestion; I will take a closer look at the controller. Subsequently, I would like to use exactly this kind of encoder.
There are many smart people in Austria too.
You should make the coils stationary and have it move the magnets. Looks nice though.
Uff so many coils 😅 maybe I’ll give it a shot in the future and compare both systems. But I think done correctly and disconnecting unused coils we could distribute the power losses over a larger area and pump more current through. Great idea
@@NeverGetTiredEngineering Yep, Charlieplexed mosfets for the win.
Very Nice! This is the kind of stuff I like to play around with myself, always love finding someone else who is just as interested! take a like and sub!
Thanks! Looks like we're on the same wavelength! :D
Nice video, is just I want to become in few years, I just need the space for all the machines and tools, I hope next year moving around to a new place with more space. Now I just have a 3d printer, but I want to build a CNC machine, spooler, hot plate for soldering, and a plastic filament recycler.
Thanks! Sounds like you’ve got some awesome plans ahead. I'm excited to see your projects. 😁
Looking forward to more videos. What's the software you use to sketch with at the beginning?
Hey,
I'm using Leonardo
Great video! Could you share what software u use to sketch, looks quite intuitive.
Hi thanks, I’m using „Leonardo“ drawing software
Neuer Use superglue on enamelled wire, it rips off the Isolation and causes shorts. Use 2part epoxy
Hey thanks for your advice. Didn’t thought about that 😅
@@NeverGetTiredEngineering it's an aquired Know-how 😊
@@NeverGetTiredEngineering and heat the windings with a heatgun to ca. 60°-80° Celsius before dropping on the epoxy. That way the epoxy gets water thin on contact and fills any tiny voids, and it hardens faster 😀
@@wiju Awesome! I'll give it a shot 😁
Subscribed. I would suggest you make the iron core using epoxy resin and iron powder if you don't have steel laminations.
Thanks! Good idea. I will try to make an iron core from sheets, but it would also be interesting to see the performance of an iron powder & epoxy core.
Have you thought about the max force the motor is able to produce? I somehow have the gut feeling that this could be not enough for the cnc machine? Otherwise great video!
Yes, I did some very basic and idealized calculations to estimate the force. If the magnetic field isn't homogeneous, which is the case in this setup, the situation gets complex quite quickly. To calculate the force, I used the Lorentz force equation, but we can and should also consider leveraging magnetic reluctance.
I plan to experiment with various magnet and steel configurations to optimize the setup. Currently, my biggest challenges are power losses and cooling. So it's hard to say how much current I can pump through the forcer without setting it on fire. Whether the force generated will be sufficient for the CNC machine depends on its specific mechanical requirements and load conditions.
Thanks for your input and for watching the video!
Wait so what lateral spacing do you want for the stator and the magnets?
You mean the spacing between the magnets? As I use a halbach array in one of the following videos it would be 0. But I can say that the distance between each pole center would be 20mm
cool, you gained a new subscriber
Welcome aboard!
Using off the shelf motor controller defies the purpose of linear motors in CNC, which is to prevent vibrations.
Ideally you'd have to make an analog driver that provides linear current changes, not a chopped DC one like in all those drivers.
In other words you need an audio amplifier on the output & and at least 3 function generators on the input which are 120 degrees out of phase. This can be done digitally, with proper filtering to smooth out the output current. But the power amplifier has to be linear.
At least you could try to use some low pass filters to smooth out the chopped DC from the driver. But then this might mess up the current sensing those drivers use to orient their phases.
Hi i get your point. However there is a big drawback when using analog linear amplifiers. Massive power losses in the semiconductors. Fortunately the inductors smooth out the current (of course we cannot fully eliminate the ripples). But we can reduce them by increasing the PWM frequency or the inductances. Additionally the motor‘s mass also dampens these ripples. So the motor is a PT2.
I guess that for a whole CNC mill the we would have a massive cooling issue when using linear amplifiers 😬
Cogging can be reduced by carefully arranging the magnets.
Hey thanks for your advice. How would you recommend arranging the magnets to minimize cogging?
Greetings
if you have a milling machine maybe you should convert that to cnc instead ?
Hi! Cool Video! Whats the blue scetch program called?
Hi thanks! I’m using the Leonardo drawing software. But it’s 40 bucks 💸
Creative DIY! Please do more! Wondering if any custom PCBs, 3DP/CNC'd parts can help for further work? Would love to sponsor! (PCBWay zoey)
Wow thats awesome! Yeah that really would support me. I'll get in touch with you.
Many thanks!
@@NeverGetTiredEngineering Nice! My contact info is available in profile, looking forward to discussing more!
Would you start counting months in programming at 0 or 1?
Definitely from 0 😁
@@NeverGetTiredEngineeringso you're one of those... 🤪
@@aaaronme Feels more natural 😁
AWSSOOOMMEEEE!!!!!
Thank you very much! :D
hell yeah I'm in
Is that going to have good resolution? It looks like 1cm resolution to me.
Yeah way better than in the clip. Using an encoder the controller adjusts the current-vector so in theory resolution is depending on your encoder
Really good Job man! Dafür kriegste ein Abo
Freut mich, danke! 😁
I was wondering, why space apart the magnets instead of using a Halbach array?
I’m using a modified array in the next video.
You know this is basically a miniaturized underpowered railgun?
Yep, just trying to take over the world one day! xD
I don't think railguns use magnets or coils but that doesn't mean they can't weaponize it.
*coilgun
More like a maglev if you ask me
Anything is a railgun if you use it wrong enough.
I sugest you make it tubular linear. make the stator of widing couper so the cable dont have to move and create friction. if dont get to hot you can put the cable inside the tube. and the rotor will be an holow tube with magnets using halbach aragment in a sandwich. so the magnetic field will be focus inside and the outside(outer ring side) no magnetism. powerful and easy.
Maybe i need to hink a litel bit more on the stator because is not valid having it as sandwich widings because the exterior has less magnetic force the inside where cable goes.
To make it easy can be used as sheet as yours plane instead as tube.but coper down and magnets up and use halbach array aragement to avoid magnetism on top on machinery and empower interaction with the stator. And to see where it is position just lector hall senzor in one side with track detection. i am not goo in electronics but i can see the mechanical deasing.
When I finde some extra time i‘ll run one or two simulations and answer again
Нас выращивали денно,
Мы гороховые зерна.
Нас теперь собрали вместе,
Можно брать и можно есть.
Но знайте и запоминайте:
Мы ребята не зазнайки,
Нас растят и нас же сушат
Не для того, чтоб только кушать.
It'll be way stronger if you put an iron core in the coils instead of plastic - either laminated or powdered, not just solid iron or eddy currents will be a problem
Hey yeah I am currently building different cores with powdered iron/epoxy, small iron pellets/epoxy and laminated iron core :)
Edit: I ran a couple FEMMs and we should expect worst case ~170 N at 2A
more video please!!!
Comming soon xD.
the last days i finished the test stand and also did a couple simulations and tweaked the design "a little".
Next week I have time to start filming and cutting so stay tuned :D
@@NeverGetTiredEngineering Thanks. Huge Fan.
Why didn’t they want to repeat the stepper motor in an expanded form? two rows of plates with teeth and a magnet between them - the lower part, and the same two rows of teeth on top with a coil. Was it really difficult?
Enable playback on external websites please. We'd like to post this video on our homemade tool forum.
Done! :)
@@NeverGetTiredEngineering Thank you. We shared this video on our homemade tool forum last week 😎
Hi, coils don't need iron core?
Hi, to get the maximum power and strength, an iron core is indispensable. It will be featured in one of the upcoming videos.
Hi. Nice project, but i see Part 0 and Part 2....where is Part 1 ???
Hey thank you. Wanted to rename part 0 to 1 and forgot about they 😅
Feelin like im a little early here
if you replace the iron core which conduct magnetics with 3D printed part which assume the magnetic potential greatly, I'm afraid the output force would decrease and the speed and dynamic reaction would become bad sharply, like a Coreless Motor, . so I suggest you should design the magnetic circuit carefully and concentrate the magnetic potential to the air gap between the coil and the permanent magnets, just like a real motor
hey, I just used this 3D printed part because I was lazy 😅. Of course I will make a iron core. One will be made from laminated iron and for the other i'll use epoxy and iron powder/granulate.
Sooooo? Are we measuring precision or what?
We’ll measure force vs current and position, repeatability, precision with different forces and also reaction speed