PCB Motor 2-Wheeled Robot (Part 2/3)
ฝัง
- เผยแพร่เมื่อ 23 ธ.ค. 2021
- In this video I optimized the rotor for my PCB Motor and fixed the wheels for my robot!
Get a free trial of Altium Designer - carlbugeja.com/altium
PCB Manufacturer - www.pcbway.com
Help me make more projects by supporting me on Patreon - / carlbugeja
This project is Open-Source and licensed with Creative Commons Attribution Share Alike 4.0 International License
My Electronic Kits - flexar.io
Tools I use: (some of these are affiliate links)
Main Camera - amzn.to/30tRgY7
Second Camera - amzn.to/2ZLm2wy
Light Setup - amzn.to/3jtqckA
Heating Plate - amzn.to/36qYzD4
Ultrasonic Cleaner - amzn.to/3nfUGHY
Magnifying Lamp - amzn.to/3jrcaQv
Vernier Caliper - amzn.to/3eTju3I
Multi-meter - amzn.to/3hq6B2R
Digital Milligram Scale - amzn.to/30HIw1X
Digital Microscope - amzn.to/3hyzoFt
3D Printer - shop.prusa3d.com/#a_aid=0550494M
LCR Meter - cutt.ly/Wlmg0b0
Power Supply - amzn.to/3u4umF7
Social Media:
TH-cam Channel - / carlbugeja
Instagram - / bugeja.carl
Twitter - / bugejacarl
Facebook - / bugeja.carl
Linkedin - / carl-bugeja-0b922a135
Website - carlbugeja.com
Music:
Love Aside - Patrick Patrikios (TH-cam Audio library)
Engine Fire - Silent Partner (TH-cam Audio library) - วิทยาศาสตร์และเทคโนโลยี
Use bushing instead of ball bearing, they are easy to find or make some with PTFE
It would seem that bushings would be a problem above a certain RPM. But the speeds he’s currently operating the wheel set at, it wouldn’t seem to be a problem.
Yes, oiled bronze bushing is better than BB for low-speed stuff like that. If an intial lubrication isn't sufficient, he can 3D print a tiny chamber for an oiled wad of cloth (packing). He'll need to break each one in with with a fine polish compound. ...Works for instruments.
Could use one of the Iglidur bushings that Igus produce, they are light, available in a massive range of sizes and types and Igus have a really nice sample program.
@@deepsleep7822 it wouldn't be a problem anyway as the bushing is holding the inner bearing shaft solid & static, only the outer bearing part is moving.
the end cap bushing is just to give the shaft rigidity through the pcb.
hell he could even use stock 2mm shaft, copper electro plate it, then solder it directly to through hole pads on both sides of the pcb ,then push the bearing assembly etc over that solid pin, simples. hell 2 just put a bunch of through hole pads in a circle then solder real insulated copper in the overlapping 1+3 pattern winding.. as rob says more copper= more volts/amps.
fundamentally there's only ever 3 things at play/to tweak here ,the amount of copper length, the magnetic fields cutting that copper, -only copper cutting at 90%, parallel copper is just waste that adds nothing up/down rather than right down, right,up , etc, every right is waste, minimise it- & speed/current.
as in the real basis for a real generic 1 to whatever pcb Cycloidal Gearbox say in a 1inch or 2inch pcb Cycloidal Gearbox video's with sub's & clearer english
came here to say just that! ahahhahaha I would not recommend ptfe bushings for something so weak and expect good speed. go for small bronze disc bushings, use one on each side, and light oil, you should be fine. It is, after all, what every small motor use as a bearing element.
sorting out the slop on the bearings will greatly improve the torque as well, as once you have a rigid bearing assembly, you can make the gap between the magnets and the coils smaller.
Can I suggest you make some small oil lite bearings, the loads and speeds your are using don't really call for ball races and oil lite bearings will remove your wobble. Keep up thd great work.
Hi Carl. I really enjoy your videos, and your enthusiasm for the subject. If you ever want some custom machined parts (such as shoulder screws or bushings) give me a shout, I would love to help out. Merry Christmas.
Merry xmas
It's really impressive to see you ordering custom made parts!
Man that tiny robot looks awesome , I think it needs some weight in the middle because it is not moving fast , nice work
What you want is either a brass bushing for these low RPMs or something similar to the bearings in a PC fan. Although the second option could be expensive, I think.
Great implementation of the rotor improvements discussed on previous videos!
Ball bearings are available with different levels of play (named MC1 to MC5 for some of the manufacturers). In this case, as no extra mechanical stress is applied on the bearings, the least play version is the best option.
Just curious, would bearings with less play inherently have more friction and require more torque to turn? Seems like there would have to be a trade off for removing the play in the bearing races.
@@ttgydddsd6305 In this application there will be about no difference in friction for as long the bearings have some play. A very slight difference in friction could be caused by the grease that has to be pushed through narrower gaps when little play. But that is easily solved by just dissolving the grease. From as axial forces are applied (is the case for most applications, but not here), the story is different. The least play bearings then have quite larger friction because the contact angle between balls and circumference is much steeper and results in much higher contact pressure and more friction.
Sintered bronze bushings preloaded with oil. Cheap, reliable, can support both axial and radial load, low friction. They are used everywhere for these types of applications. Plus they are lighter than the ballbearings you are using.
Didn't know PCBWay manufactured these types of things, awesome!
oiled bronze bushings (journal bearings) might be a better option than a simple ball bearing. Other bushing options would be PTFE and UHMW-PE.
If you need help just ask, use FOC to improve the control. Takes 5 minutes to implement and tune but will make a massive difference.
Wobble will also be gone
Keep going, don't stop to perfect a piece, ride the wave of the art
The laminated core can be superseded by a Halbach array, then you can just make use of your aluminum core and forget about the flux.
Aa some have said, bushings would be acceptable for your setup, also using thinner bearings you could get 2 into the rotor without adding thickness. they may be unshielded though which would need cleaning more frequently
if you add a third wheel sticking out the middle, the whole assembly will be secured in place while driving, eliminating the need for balance checking and letting the 'bot move more smoothly
Great work my friend, and sure lots to learn from your content
Could also yoke the magnetic field on the non rotor side of the coil. This will further increase the magnetic field in the gap between stator and rotor, as well as provide thermal heatsink for the coils.
Keep up the good work! Love this project! Merry Christmas!
Hello Carl, I am from Turkey. I like your videos and projects. So I suggested your channel to my students at universty.
Carl, your projects are so cool!! Keep it up
Great work. A common technique to remove the wobble is to use two bearings and preload the bearings. The wobble is now as small as the precision of the bearing race. Can you make a stackable version of the pancake motor.
I like that you are designing for manufacturing.
Maybe add a tail to the middle, to prevent the middle from spinning.
Things to try: Higher precision bearings with cylindrical rollers , sleeve bearings (Delrin comes to mind), self lubricating bronze bushings or a regular bearing with two internal rolling tracks (double ball bearing in one) Or even a trust bearing (sometimes called PANCAKE bearing). They all fix your problem.
With two wheel robots a tail helps the motors turn the wheel instead of the body. It adds a small amount of weight and ground friction but for a robot like that I would just use a thin wire.
I think if you add some kind of motion sensor and implement some simple control, you can implement this with the slop in the wheels
Would recommend adding a piece out front with a third wheel or even a ball. You can prevent the whole assembly from moving which would enable more smooth movement of the robot in general. Love your rotor designs so far and looking forward to continued improvements!
i think the point is to make a two wheeled microdrone style robot, but I would agree that a tiny servo arm or some third point of contact in the center slightly in front of the bot would hugely increase speeds and stability possibly something retractable that only extends when "forward" is required.
Do it like the rainbow 6 drone
That looks very solid!
Idea for diy bushing. Using a syringe needle, and a pin. The smaller diameter means less area for friction, and light oil between the two would make them slide.
Fab! You're making some good progress on this :)
man , you are great, love your skills and Invention. Thank you
Haha the Ai subtitler transcribed the sponsor "PCB way" as "pissing away". Marketing's gotta love that.
Something I'd like to see before You remake the robot is have the battery be on the back end so it would lean nose up and have a rigid tail dragging on the table to keep it from tipping backwards
you could also use a through shaft and put the bearings in that axial. or try looking into gem bearings as they should work for this
I had the same idea, put 2 bearings inside the center section and add an axle that the wheel attaches to.
you could try thrust bearings.... thinking about a usable configuration, but with them you could use 3 (preloaded) for a pair of wheels instead of 4....
If you have a lever attached to the main body, then it could limit the main body's rotation relative to the ground, thus allowing wheels to easily push robot
Having a third wheel for equilibrium will make it rest on 3 points and be much muvh more stable and thus faster
To decrease the wobble seen at 4:50 you should look into implementing Field-Oriented-Control (FOC) for the motor. Maybe back-EMF will do but I highly recommend adding some kind of encoder to know the state of the rotor. Then I would design a PID controller to keep the center PCB level using a gyro sensor. Whatever is currently controlling the motors (and indirectly PCB angle) is inducing way too much oscillation. FOC will use smooth sinusoidal coil currents and also the most energy efficient way to drive electric motors. I'm not sure if you have experience in Control Theory but I am certain that good controller design could compensate for the limitations given the hardware.
I think this was more like a quick test, because the design itself isnt final anyways
Or just add a third wheel in the middle lol
Nice work!
It would be interesting to see how it works with fixing a sled to the axle so that the axle remains fixed relative to the table surface.
I am sure one of these days, these robots will be used for surveillance! Merry Christmas and Happy New Year Carl!
Your maximum torque is limited by the balancing nature of the 2-wheel design. While it’s an option to lower the COM as low as possible, a third wheel would multiply the possible torque quite a lot.
That’s what I was thinking too, one trailing wheel to prevent over-rotation.
Just do it like rainbow 6 siege or the ThrowBot
Do you know about brushless gimbals for cameras?
They use 3 phase brushless motors but not in a motor configuration but rather like a positioning drive. They do it by driving them without BEMF or sensor feedback, by supplying them with 3 constant currents following a sinus shape with 120° phase angle to one another.
That way, they can drive them incredibly smooth from a standstill. I once built a little RC car with direct drive, that operated its motor in that kind of configuration.
No idea if you're already using that technique, but maybe it helps.
That's pretty cool bro 👍
If you want the angular play to go away, you'll need a second bearing. If you shim the outer races apart just a bit, you'll produce a back to back angular contact set, which will eliminate the play and be much stiffer.
How about an inverted pendulum bot? :)
great job, looks like ur getting that to work pretty well. Maybe some PID tuning to remove some of the stuttering? Are these motors strong enough to move a potentiometer? That would be really cool in something like a modular synthesizer application (envisioning all of the knobs moving to give visual to the sound coming out... really cool).
You need to increase the weight of the center assembly and move the center of mass further from the axle. At the moment the CoM appears very close to the axle, which is th factor that is limiting your torque.
A way to reduce wobble could be to press for a shaft into the rotor. Then you could cut a pocket into the rotor a larger diameter than the bearing OD And make a cup that holds the outside of the bearing, then attache that to the pcb. Place another cup and bearing on the other side with the bearing over the shaft. You can even place another rotor on the end of the shaft. This should keep it compact and would also allow you to utilize two rotors to increase torque.
Your videos just keep getting better and better!
Maybe you could add a third wheel to help keep it stable by keeping the board in the middle stationary?
The point of the project is to have as few wheels as possible.
@@gordonlawrence1448 then maybe a lever arm or something. It seems like it loses a good amount of energy due to the rocking of the axel (the main board)
Good job, go ahead.
On the inner side use a rotor with magnets with the diameter smaller then the wheel, just enough for the connections between PCBs.
Move the ball bearings under the main PCB and use 2 axels rigidly mounted on the rotor(s), use 2 BB per axel to remove wobble.
or
Make the hole in middle big enough for the ball bearings so that you can use one BB on each rotor while keeping it flat. The torque is influenced more by the outer diameter anyway...
You don't need two bearings just preload the bearings inner ring with a washer spring and the wobble is gone. Obviously you would have to need some thing to regulate the tention of the washer spring
Can literally see this guy growing older
You should get a mini lathe.
It will open worlds of possibilities for your projects.
Like custom screws and bearings/bushings.
Bushings would probably be the right choice for this project as they can be made with less play)
The bearing fit could be improved of your shoulder screw clamps the inner race of the bearing
Can you source 2 thinner bearings? This should already improve the slack.
Also you might want to get started on the "self balancing" motor driver code, so it can smoothly drive, at least for better visuals 😁
Mag-lev bearing could present challenges, but virtually no friction may get you up that ramp. 🙂 Also, one stator PCB with two rotors, one each side, may give more torque? Just a thought. I'm really enjoying your work, thank you for sharing it!
Looks like you are driving the motors faster than they can go at the end there. You might have more success if you slowly ramped up the speed you drive them at until a target speed is reached, so the wheels have time to spin up and gain momentum
Maybe just go with igus filament to print a bushing? You could also make that conical to bear axial and radial load.
may i suggest that you add some more weight at a lower centre of gravity on the body of the machine so that the motion from the motors is more efficiently transferred to the wheel. at the moment the motors and programming are fighting each other. once that is sorted the efficacy will shoot up.
Nice video
You need an offset weight with it’s own wheel to counterbalance the torque.
I think you need a third wheel to keep the board from turning and losing momentum
Bravo !!
I am intersted in linear motors and would love to see your ideas building them.
I am interested in linear shaft travel internal combustion driven generators: Coil around a shaft with a piston at each end & no rotating parts inside the case... Very few parts overall. Seems brilliant?
see all things m.th-cam.com/video/OI_HFnNTfyU/w-d-xo.html Professor Eric Laithwaite: Magnetic River 1975
the father of the linear motor
I bet you could try electroplating some nickel or something onto the bolts to get a more precise fit and reduce the bearing wobble. It might also be a good idea to use a bit of lubricant in there
Super cool...
i love this
You could use a graphite air bearing! Just have to sort out where to fit the air compressor LOL!
what if , the bearing is as big as the wheel, and the motor magnets are inside that bearing ? one side of the bearing connected to the main body, the other to the magnet assembly ?
I thinnk with a bit of teflon and oil you could seal the back and banish the wobble
Another awesome experiment. Thank you for sharing Carl!
About the bearings, what if you press fit the axel into the Roa tasting part of the wheel and gave your bearings/bushings more room by placing them inboard, maybe even make their mounts into PCB features?
If u add a third wheel you can use the sandwiched magnet design
I'd try to keep the wheel diameter to a minimum to increase torque and use 4 wheels with tank steering.
Damn... Amazing project ! What is your goal for the project ? Having a diy kit, or maybe a full rtr product would be cool, but if you're just want to learn cool things, it's still amazing choice, definitely a really cool way to do it !
Thanks! 😊 for now I'm just doing it as a hobby & to learn more about the potential of pcb robots, as this field has not been explored that much
Obviously the goal is make a drone from Rainbow Six: Siege. Just need to add a camera, WiFi, and jump actuator.
Carl can you make one with piezoelectric Motors
Hi Carl,
Another bearing type you can use is angular contact bearings.
Thanks for the tip! but unfortunately the minimum size for angular contact bearings is 10mm which is too big for this rotor - the one i'm currently using has a 5mm outter diameter
Would having a trailing flexible leg help in stabilizing the body of the robot and help turn that fluttering motion back into torque? Great project Carl, keep those clever ideas coming!
I like your channel 👍
Why not put the bearing around the magnets in place of the center? Take a look at the THK RAU bearing series, which is a cross roller design. Also it may be easier to find a needle bearing that could work. You may have to glue it to mount it?
I’m really enjoying your projects lately. Where did you get the magnets made (and how much did that cost) if you don’t mind me asking?
It's shivering a lot, please make it covering, it might be feeling cold, poor little bot.
What about gyroscope in top of the main pcb?
And a bush shaft that way the wobble is less ?
if your using a single bearing, then you Need to make a long End Cap Bushing to support its circlip retainer + inner shaft & so take out the generic slop.
End Cap Bushing as you can make the pcb hole to fit the lip for a tight push through pcb support fit.
a diy micro version of 'Board Rod End Cap Foosball Bushing' if you will.
guessing custom sized magnets is a waste, just get round mags of generic size, then just chop them in to quarters ,grind etc then re-electroplate the ends or other seal them against the oxidising air
I guess sintered bronze bushings will be may be better at this size instead of bearings.
seeing how the 'robot' shakes all the time, i suggest to add more delay between the reading of the sensor(s) and actually applying power to the electric motors.
Also a way softer start and softer switching power between the coils that move the magnets.
How about adding a gyro? just one electric motor that will spin at high rpm to stabilize everything end kill the wobble?
Tilt the wheels inward so the axles are slightly hanging. Adds friction, but it will keep the wheels more stable relative to the chassis. Negative camber.
Do u likt the water bottle air up ? Bc i got one and i loved it
Perhaps try a Hard drive Bearing assembly will stop the wiggle ;p
Add negative camber to wheels, weight will pin wobble inward.
Add one more groundwheel motor to get the gyro effect. Then the center won't swing that much.
having the wheels at 20 degrees angle would improve stability and turning but would limit forward and back
Ever consider axial flux faulhaber wound coils on a PCB?
nice
You should have one more wheel, but perpendicular to another two, and spin fast. Then you will cancel the momentum like a gyroscope.
what about an active counterweight to control the inertia?
Try to put some weight on the body itself so that torque from the motor won't rotate the main body instead of wheels. This way I think Robot will be more responsive to sudden change in acceleration.