He really does do an impressive amount. However I've also noticed that he seems to only finish about 80% of the project. In real life, the last 20 percent of the project takes 80% of the time. At 80% done, you have something that is mostly functional, and mostly there. Good enough for a video, not really good enough to ship.
He's just using the wrong soap. Real dishwasher soap doesn't foam. Real dishwasher soap is a type of acid that just dissolves the food. (so is clothes washing soap)
Last video I mentioned what an inspiration this channel has been for me. Since I think you liked it, more specifically, I had no clue it was possible to print so many large parts and fit them together into complex mechanisms. My imagination couldn't even conceive of it I guess. But after seeing your work, and starting working on exos at grad school, I've almost completed an entire wearable exo arm with an entirely new intended motion pickup system. That's actually what my thesis is. I would have almost certainly never had the confidence to do this without your videos. I'm genuinely grateful
The center of mass is a bit high. An improvement that would require a ground-up redesign would be to widen the vertical space between the two parallel beams between the wheels and put the motors there. Also, with a small cylindrical gap in the base between the wheels, there would be a great place to keep the electronics and batteries that would make it basically impossible to tip over.
Dude, you literally invented a thing. I watch a lot of "maker" content creators but rarely do they actually invent a unique thing. Absolutely amazing mate
Would casting the outer part of the wheel out of silicone be a feasable solution for getting more traction? Or even just stretching an air balloon over the print? This whole "print more sections out of TPU" / "print with dual extruder" seems a bit to overcomplicated.
In university (25+ years ago) I spent an inordinate amount of time messing around with mobile robots. At several conferences, I saw a lot of RWI's robots, and was severely impressed. This robot platform reminds me more than a little bit of their robot base, but the team that developed those were larger than your one-man-band and the mechanics required to drive their wheels and achieve the non-holonomic drive like you are getting here were several orders of magnitude more complicated, and likely required a long time with a mill to create. Major respect.
This is probably one of the best use cases for such a wheel, and it seems like it can excel at this task! What it loses in having a complex wheel design with many parts, it makes up for with it's simplicity in driving and it's overall capabilities in a home environment.
I too noticed that it suffered from the trike instability when stopping. Five spokes/wheels as used on (better) office chairs would help, especially with a high CoG, and may even help with obstacle climbing by having more driven wheels in contact with the "flat" surface. Control is complicated somewhat however.
@@PiPArtemis i would go for 5, you have basically no edge to fall off because there's always a wheel in the way, they don't have to be as large as the three, there's a reason office chairs have 5 double wheels
This really reminds me of the way in which characters are moved in video games, particularly first person ones. the left joystick translates the character while the right controls pitch and yaw of the viewpoint. it would be interesting to see the robot controlled by only looking through a first person view. i imagine it would be very intuitive.
This is really amazing. This robot can really be used for practical applications in automation. Nice work! I hope you build a simple robot like this which can run over curb sidewalks easily.
Just No. If you've ever used a luggage trolly at Heathrow you'll understand why this is a terrible idea. For some reason they have the only trollies I've ever come across with four swivelling wheels. Makes cornering way more difficult than it needs to be. I'm convinced this is done to entertain the security staff monitoring the cameras.
@@ChrisCameronPhoto I’ve seen similar carts at IKEA, they’re easy to push but hard to rotate. I wonder if a circular cart would work because rotation wouldn’t matter.🤔
I think the 4 swivelling wheels they currently have are almost as good as these, and I'm pretty sure they're easier (and cheaper) to make. It would be really cool, but also very unnecessary :)
@@ChrisCameronPhoto shopping carts here in Germany all have 4 swiveling wheels, hardware stores tend to have fixed wheels at the back for the bigger carts for heavy appliances and materials etc.
@@noahluppe May be ok in your shopping centres, but when you're trying to negotiate heavy luggage around a hard u turn on a steep ramp at Heathrow and there's a security camera on the apex you get the feeling someone has got to be having a laugh.
Most people I know still dont think 3d print is useful ¯\_(ツ)_/¯ And peope dont think space is useful, sometimes im just baffled by what people refuse to see value in
I wonder how tread would work on omni-wheels. Tread on tires usually redirects water in the direction of the wheel rotation, but here there is no given direction of wheel rotation. This is not something I would have thought of without seeing this prototype driving over other objects. Awesome work James!
Super awesome work 💯. SUGGESTION: Instead of the wheel at the end of ball wheel, a freely movable ball will work much better and it will hopefully remove the jerks in the travel of the bot.
A captive ball might actually impair certain performance. The wheel only exists to cover the edge case of movement exactly along the axle. The wheel is indistinguishable from the rest of the solid surface of the ball in all other cases. If you replace the cap wheel with a captive ball, it could create slip in that part of the travel. I could maybe see a ball working better in an unpowered omniwheel, but that just begs the question of why not just use a captive ball the size of the omniwheel in the first place.
It needs to be a wheel, or when you use that ball for movement it will have two periods of zero traction. Remember the drive motor pushes the wheel and rolls along the ball, of it gets to a freewheeling ball then it will slip for the entire duration of that when it is trying to accelerate. The little captive wheels NEED to be wheels, and NEED to be oriented along the axel. It is fundamental to the design.
Ball shaped wheels would be an amazing experiment for a 3-4 wheeler. Their meant for tough terrain so making the wheels omnidirectional would probably be a huge benefit.
The math for this is impressive. I've done some similar stuff creating game AI/telling objects where to point/move in 3d space- confusing stuff. Nice work!
11:00 the ball joint design would be perfect to eliminate the gap, and also placing and spacing the actuators inside the joint for better and more precise control, and increased traction/torque. A reverse trick that's been in use on computer mouses and trackballs.
Looks awesome! I remember looking into that same tech back when I was designing ball drives. Just thought you might find this project I did a while back pretty interesting, I solved a lot of the problems to do with making a drivetrain with actual solid spheres: th-cam.com/video/j09PzmVmzJw/w-d-xo.html
OH, I remember that video vividly. Pretty much sent it to all of my FTC contacts. My favorite ball drive I’ve seen so far is this: th-cam.com/video/INIBjexEA9Q/w-d-xo.html
@@Duckeasily oh wow that’s quite unique and cool, I’ve not seen that video before! I’m getting a lot of ball balancer vibes from it with the vertical configuration. That’s also by far the best minimally intrusive ball retention system I’ve seen. Thanks for sharing! My design with 4042 was loosely inspired by this one grabcad.com/library/basketball-riding-omnidirectional-robot-drivebase-1
@@emberklavins9567 My love of ball drives were also founded by an FRC design. I found this randomly one day on chief Delphi: www.chiefdelphi.com/uploads/default/optimized/3X/0/4/0486456db177cce51fa73dd4fd18917606b57695_2_690x500.jpeg Not too sure on which post I found it on, as it was years ago.
stabilty and control issues when encountering small light objects, I suggest you move the motors down on the side of the black wheels bracket holders so each wheel has the motor at 60 degrees upwards instead of straight vertical, (obviously shorter belts) that lowers the center of gravity by a huge factor, liberating the top for other applications, on top I would place a heavy wide surface batery (not tall, just wide), also good for authonomy, but moslty for lowering the total center of gravity, in such configuration it should climb the lid allright and drive over small light objects without dragging them. If you increase the wheels diameter to 100%, keep TPU stripes slim and place them almost touching to eachother, you get a slightly bigger base, maybe 60% larger, but more fit for a robot that is tall and unstable to mount atop, while it still can pass through the doors allright, thou it might not be able to climb stairs.
it was my 3 year old project when I got idea from the same good year video but couldn't worked out. But watching it now working and runing like that made me feel happy 😍😍.
In addition to replacing the round shaft with a flat plate, you could probably actually widen the gap and put TPU "brushes" lining the edges so that they can spread around the axle as the wheel rolls but also provide decent support and traction in the other axes when the wheel is rolling.
Seems like it might benefit from some kind of differential, suspension and/or gyroscope. 1. Differential do that the wheels can go at different speeds. 2. Suspension to help the robot handle rough terrain. 3. Gyroscope to help prevent the robot from falling over. Also, I think the robot should have a full 6-axis IMU, plus a 3D compass. This way it can more easily self-correct its own movement when something is just slightly miscalibrated.
Amazing! Maybe you could replace those white treds with slanted roller contacts also. Making a "Split sphere Mecanum Wheel" or perhaps an omni wheel design with 3d printed linear TPU barrels. It would reduce a lot of friction and would love to see the batteries using less.
For a low speed omnidirectional robot I can see this working really well. But for a car I'm a little sceptical to be honest. Because it is a ball it has a contact area which will probably be terrible for breaking performance. I'm also concerned about the rotation speed (and efficiency) when the wheel is rotating almost on the tips. Also, if a car drives at 100km/h and it rolls towards the tip it will pick up a lot of speed, but then it rolls over the little wheel at the tip and when it rolls even further it has to suddenly rotate in the other direction at its fastest speed. I think this will cause a lot of wear and possibly even accidents. But as I said, for low speed omnidirectional robots these wheels may be perfect :D
Amazing content! Really excited to see whatever comes next. It really moves incredibly well. Slap some treads on those bad boys, and you'll be all set!
great job, I'd assume you could make i more stable by leaving more distance from the center of the bot and the wheels, and in that increased gap, you could put some of the component to lower the center of gravity. if the brackets arm for the wheels are also flexible... you've got a basic suspension system.
Excellent! If you put the motors to the side instead of above, your cog will be better. I'm not worried about the dishwaser, but I do worry about tipping over if you build this up higher, the wheelbase is also small.
A garbage can that follows you is such a great idea for people with big work areas. I half expect you to upscale this and making yourself a rolling throne 😁
I think that the gap between the two hemispheres could be removed completely by using three sphere caps instead of two hemispheres and, consequently, a triangle shaped beam going through the center with three branching beams going through the sphere caps and a little wheel on the end of each of them. Geometrically the three caps can be constructed by cutting from a sphere a prism that goes through the center of the sphere and has a regular triangle base which can be inscribed in the circumference of the sphere. The main trasmission mechanism would stay the same, with the pulley acting on the base of the center beam, but there would be an extra rotating part and another wheel in the middle of it, which, as it turned out from your prototype, is acceptable. The nice thing is that the sideways rotation would go along a (ideally) perfect circumference, without any (big) gaps. if you don't understand, sorry for my bad english
I am on a robotics team and we made a very similar ball drive. To make the ball drive be able to climb over obstacles we printed hemispheres out of flexible filament. I would love to see a version of the ball drive with squishy wheels made by some one with as much expirence as you.
This seems like a good application for a deep learning A.I to figure out optimizations for the system of movement, i.e vibration reduction, g force reduction, wear balancing, etc...
Very cool project! One concern while watching the video is that it seems a hard deceleration almost tips it over, maybe add metal beads inside the hemispheres of the wheels to lower the center of mass would also be a useful improvement? It seems quite high at the moment with the big motors up and the batteries up. And if you add more stuff on top it will be even worst. :/
I’m curious as to the longevity of this design as compared to traditional wheels or other omni-directional designs. I’m wondering if the fact that to go anywhere instead of just rotate you need to use opposite torques to generate forward force might lead to some amount of bending of the axles if they have to deal with some load, like say if the robot it’s carrying is quite heavy,
this is awesome, a really easy way to pull off omnidirectional travel, and surprisingly easy to code for by the sounds of it! I have a suggestion for improving the ball wheels though, It came to me after watching that the driven axel really has no reason to need to be a cylinder you could use some 2 mill flat metal plate and it really wouldn't matter, it would mean you could get those hemispheres super close.
I love where you are going with this. I invented a way to drive the spherical wheel without interference in engineering class and have had it on the backburner ever since. This was actually before goodyears tire design came out but theirs inspired me to keep innovating my design .I wish to invest in a 3d printer so i can test my hypothesis .
OHHHH I LIKE THE WAY YOU WENT WITH THE AXLE WHEELS!!! i had the idea i suggested on your last omni wheel ball vid that perhaps you could take the same concept as anolog mouse trackin sensors of a ball that you than put into a divit that is a bit more than a half circle with room for the ball to rotate so it cant be pushed out but allows easy rotaion in all directions although the way you went is intresting too of from the little sinppet at the begining im guessing that its a normal wheel but its mount allows it to spin in a 360 degree circle and the premise is when it gets on the axle side the flat of the whell will allow it straighen to the direction it needs to to move
I was thinking about the gap between the two semi spheres.. My first idea was to fill it with two sets of link treads, following the shape of the sphere and returning along the threaded axis back. It does not need to be connected to spheres, only to axis and free-rolling. The thing is - it is easy to brainstorm such things, but not easy to make it work. Anyway, nice concepts.
potential exp. given each drivebelt is located near the wheel, adding suspension components to each knuckle(im assuming the black prints) that bot could haves some incredible articulation in each strut making it a very adept offroad/ rough terrain vehicle. love your work!
Did it take long to tune for the first motorized shot at 7:06? It seems like one of those things that so straightforward I would be baffled if it actually worked the first time
A bigger apex roller with spherical ends to match the hemispheres might help a bit too. Give better ground clearance for when it's on the ground. Might also be able to eliminate the gaps between hemispheres by using three hemisphere/apex roller assemblies.
When James mentioned robotic bin, thought of the little robots in Zorg's office, in the film the Fifth Element, that cleaned up when he smashed the glass.
Wow you have a dish/floor washer combo.
probably why he made this robot. he'll attach a mop to it lol
😁 Never put regular soap in dishwashers. Also dishwashers are not watertight.
Ignore that fake James Bruton post, that's not him. Please report it as spam.
Thats so dumb.......note to self: patent that
XD
This guy does more projects in a month than I have done over my lifetime.
*have done=will do
He has his passion as well as 💲💰💰💵
Amazing what happens when TH-cam gives you enough money to self-survive and you can just explore science
He really does do an impressive amount. However I've also noticed that he seems to only finish about 80% of the project. In real life, the last 20 percent of the project takes 80% of the time. At 80% done, you have something that is mostly functional, and mostly there. Good enough for a video, not really good enough to ship.
James, are you putting dish detergent for the sink in your dish washer? Made that mistake once
8:00 let's discuss our definitions about fine, JAMES
Maybe take a break from engineering robots and engineer a new gasket for that dish washer. lol
@@connorbailey308 or just use dishwasher soap, instead of regular detergent. Dishwasher soap doesn't foam up like that
Eh mine uses a cable and buckle wrapped around it to stay closed
And a wooden platform to align the door and frame.
@@WLxMusic what? you are saying if I stop using shampoo, i won't have to mop the floor every day?
It is amazing how chaotic the omni-directional wheels look when they are all in motion but the movement of the robot is so smooth.
Definitely seems like it has some complicated kinetics in comparison to regular wheels...
Ah, this will make a perfect office chair for an evil genius.
Or a better electric wheelchair.
I was imagining it as a camera mount that would lock on to something and move freely while always focused on the subject.
But my office chair can already move in any direction...
I was imagine too,.. This wheels for the caddie in mall😀
This is what the Hoveround wanted to be when it grew up.
lmao "my dishwasher is fine thanks"
Man builds robots for a living you’d think he’d know how to fix his dishwasher 😂😂
It doesn't look fine
He's just using the wrong soap. Real dishwasher soap doesn't foam. Real dishwasher soap is a type of acid that just dissolves the food. (so is clothes washing soap)
@@Scott_C uhh..pretty sure they're both alkaline. True though that he's using the wrong kind.
@@punkdigerati You're probably right. I don't know enough.
Is your dishwasher OK, James? We're all very concerned.
I don’t think it’s feeling well
“The cobbler's children are the worst shod.”
Yes. What is going on there !!!!!
@Robo Cop Humor you know.....
I'm glad I wasn't the only one who was concerned. I kept waiting for him to mention it but nothing so I came straight to the comments to check.
Only three symmetrical wheels, this is an incredibly great design!
Last video I mentioned what an inspiration this channel has been for me. Since I think you liked it, more specifically, I had no clue it was possible to print so many large parts and fit them together into complex mechanisms. My imagination couldn't even conceive of it I guess. But after seeing your work, and starting working on exos at grad school, I've almost completed an entire wearable exo arm with an entirely new intended motion pickup system. That's actually what my thesis is. I would have almost certainly never had the confidence to do this without your videos. I'm genuinely grateful
Why did TH-cam algorithm take so long to recommend me the robot master's channel? This is awesome.
"What is my purpose?"
"You carry my kitchen bin"
“Oh my god”
Welcome to the club pal .
Atonomy of the comment section
90% talking about his floor- sorry, dish washer
10% genuine comments on the robot
The center of mass is a bit high. An improvement that would require a ground-up redesign would be to widen the vertical space between the two parallel beams between the wheels and put the motors there. Also, with a small cylindrical gap in the base between the wheels, there would be a great place to keep the electronics and batteries that would make it basically impossible to tip over.
I think the dishwasher is jealous - look at it - it's DROOLING!
You mean frothing 😎
If you do design a robot body for this, might I suggest Rosie from "The Jetsons". That would be amazing to see.
I truly like this idea - drape it with a Rosie body.
He can borrow the head from this project th-cam.com/video/8CH7B6zuqAk/w-d-xo.html
Really impressive combination of mathematics, design, building and overall creativity. Well done!
don't use dish soap in a dish washer, use dishwasher detergent, dish soap is meant for handwashing and creates a lot of foam.
The only reason I know that is because I did it once
Looks like James ran out of dishwasher soap and tried using dish soap instead. He now has a bubble making machine!
And fill up both cartiges
Dude, you literally invented a thing. I watch a lot of "maker" content creators but rarely do they actually invent a unique thing. Absolutely amazing mate
Would casting the outer part of the wheel out of silicone be a feasable solution for getting more traction? Or even just stretching an air balloon over the print?
This whole "print more sections out of TPU" / "print with dual extruder" seems a bit to overcomplicated.
Might as well make use of the free material from the sponsor.
In university (25+ years ago) I spent an inordinate amount of time messing around with mobile robots. At several conferences, I saw a lot of RWI's robots, and was severely impressed. This robot platform reminds me more than a little bit of their robot base, but the team that developed those were larger than your one-man-band and the mechanics required to drive their wheels and achieve the non-holonomic drive like you are getting here were several orders of magnitude more complicated, and likely required a long time with a mill to create. Major respect.
This is probably one of the best use cases for such a wheel, and it seems like it can excel at this task! What it loses in having a complex wheel design with many parts, it makes up for with it's simplicity in driving and it's overall capabilities in a home environment.
A stool with 3 feet never wobbles, but falls over in a hurry, the higher you build it i think acceleration and deceleration will need to be limited
I too noticed that it suffered from the trike instability when stopping. Five spokes/wheels as used on (better) office chairs would help, especially with a high CoG, and may even help with obstacle climbing by having more driven wheels in contact with the "flat" surface. Control is complicated somewhat however.
Gyroscope.
Wider base..
or you have to make the base increasingly wider to compensate
@@PiPArtemis i would go for 5, you have basically no edge to fall off because there's always a wheel in the way, they don't have to be as large as the three, there's a reason office chairs have 5 double wheels
I really really love the way this robot looks when moving. So mesmerizing. Great video! That was really cool
"my dishwasher is fine thanks"
Your dishwasher is having a stroke and dying
Brilliant platform for developing a robot upper.
But James, go fix your dishwasher!
I feel, in the force, the there with be a video about dishwasher/washing machine maintenance coming up soon! :-)
I'm amazed how much you get done in one week! Great stuff.
This really reminds me of the way in which characters are moved in video games, particularly first person ones. the left joystick translates the character while the right controls pitch and yaw of the viewpoint. it would be interesting to see the robot controlled by only looking through a first person view. i imagine it would be very intuitive.
probably the most consistent and creative youtubers out there, good luck
Very interesting and impressive as always James. You deserve so many more subs
This is really amazing. This robot can really be used for practical applications in automation. Nice work! I hope you build a simple robot like this which can run over curb sidewalks easily.
They should use this for the trolley wheels in grocery stores.
Just No. If you've ever used a luggage trolly at Heathrow you'll understand why this is a terrible idea. For some reason they have the only trollies I've ever come across with four swivelling wheels. Makes cornering way more difficult than it needs to be. I'm convinced this is done to entertain the security staff monitoring the cameras.
@@ChrisCameronPhoto I’ve seen similar carts at IKEA, they’re easy to push but hard to rotate. I wonder if a circular cart would work because rotation wouldn’t matter.🤔
I think the 4 swivelling wheels they currently have are almost as good as these, and I'm pretty sure they're easier (and cheaper) to make. It would be really cool, but also very unnecessary :)
@@ChrisCameronPhoto shopping carts here in Germany all have 4 swiveling wheels, hardware stores tend to have fixed wheels at the back for the bigger carts for heavy appliances and materials etc.
@@noahluppe May be ok in your shopping centres, but when you're trying to negotiate heavy luggage around a hard u turn on a steep ramp at Heathrow and there's a security camera on the apex you get the feeling someone has got to be having a laugh.
It is crazy to me that people can just 3d print actually useful parts now. I remember when people thought it wasn't gonna be useful.
Most people I know still dont think 3d print is useful ¯\_(ツ)_/¯
And peope dont think space is useful, sometimes im just baffled by what people refuse to see value in
God forbid your home ever caught fire. Nobody would understand what they'd see when they walked into the mess of melted plastic 🤣
I wonder how tread would work on omni-wheels. Tread on tires usually redirects water in the direction of the wheel rotation, but here there is no given direction of wheel rotation. This is not something I would have thought of without seeing this prototype driving over other objects. Awesome work James!
The wheels honestly look like something you would see in a sci-fi show/movie.
so glad youtube recommended the first episode of this two parter to me. consider me subscribed! good quality content.
omg what is going on with that dish washer ahahahaha Omni directional wheel platform seems like it has alot of promising future applications
Great vids James, love the fact you 3d print EVERYTHING
Be very cool to see this become an amazing wheelchair design for somebody...
In the veritable sea of youtubers with poor ideas, you channel is a shining beacon of really cool, functional, content!
Super awesome work 💯.
SUGGESTION: Instead of the wheel at the end of ball wheel, a freely movable ball will work much better and it will hopefully remove the jerks in the travel of the bot.
Aka a trackball?
A captive ball might actually impair certain performance. The wheel only exists to cover the edge case of movement exactly along the axle. The wheel is indistinguishable from the rest of the solid surface of the ball in all other cases. If you replace the cap wheel with a captive ball, it could create slip in that part of the travel. I could maybe see a ball working better in an unpowered omniwheel, but that just begs the question of why not just use a captive ball the size of the omniwheel in the first place.
It needs to be a wheel, or when you use that ball for movement it will have two periods of zero traction.
Remember the drive motor pushes the wheel and rolls along the ball, of it gets to a freewheeling ball then it will slip for the entire duration of that when it is trying to accelerate.
The little captive wheels NEED to be wheels, and NEED to be oriented along the axel. It is fundamental to the design.
Ball shaped wheels would be an amazing experiment for a 3-4 wheeler. Their meant for tough terrain so making the wheels omnidirectional would probably be a huge benefit.
"My dishwasher is fine thanks" 🤣🤣🤣🤣 I love it!!!!
The math for this is impressive. I've done some similar stuff creating game AI/telling objects where to point/move in 3d space- confusing stuff. Nice work!
8:05 If you say so, James. You're the engineer afterall 🤔😕
11:00 the ball joint design would be perfect to eliminate the gap, and also placing and spacing the actuators inside the joint for better and more precise control, and increased traction/torque. A reverse trick that's been in use on computer mouses and trackballs.
8:35 this is what it feels playing with tank controls.
This was really the best robot part so far. It could really be a practical design with very little change in form, just better materials needed.
What if, hear me out, you replaced the wheels with more balls?
Hmm... So let's say, tracks but with omni wheels?
This works awesomely, man. I didn't expect such performance out of them while you were building, to be honest.
Looks awesome! I remember looking into that same tech back when I was designing ball drives. Just thought you might find this project I did a while back pretty interesting, I solved a lot of the problems to do with making a drivetrain with actual solid spheres: th-cam.com/video/j09PzmVmzJw/w-d-xo.html
OH, I remember that video vividly. Pretty much sent it to all of my FTC contacts. My favorite ball drive I’ve seen so far is this: th-cam.com/video/INIBjexEA9Q/w-d-xo.html
@@Duckeasily oh wow that’s quite unique and cool, I’ve not seen that video before! I’m getting a lot of ball balancer vibes from it with the vertical configuration. That’s also by far the best minimally intrusive ball retention system I’ve seen. Thanks for sharing! My design with 4042 was loosely inspired by this one grabcad.com/library/basketball-riding-omnidirectional-robot-drivebase-1
@@emberklavins9567 My love of ball drives were also founded by an FRC design. I found this randomly one day on chief Delphi: www.chiefdelphi.com/uploads/default/optimized/3X/0/4/0486456db177cce51fa73dd4fd18917606b57695_2_690x500.jpeg
Not too sure on which post I found it on, as it was years ago.
stabilty and control issues when encountering small light objects, I suggest you move the motors down on the side of the black wheels bracket holders so each wheel has the motor at 60 degrees upwards instead of straight vertical, (obviously shorter belts) that lowers the center of gravity by a huge factor, liberating the top for other applications, on top I would place a heavy wide surface batery (not tall, just wide), also good for authonomy, but moslty for lowering the total center of gravity, in such configuration it should climb the lid allright and drive over small light objects without dragging them.
If you increase the wheels diameter to 100%, keep TPU stripes slim and place them almost touching to eachother, you get a slightly bigger base, maybe 60% larger, but more fit for a robot that is tall and unstable to mount atop, while it still can pass through the doors allright, thou it might not be able to climb stairs.
it was my 3 year old project when I got idea from the same good year video but couldn't worked out. But watching it now working and runing like that made me feel happy 😍😍.
In addition to replacing the round shaft with a flat plate, you could probably actually widen the gap and put TPU "brushes" lining the edges so that they can spread around the axle as the wheel rolls but also provide decent support and traction in the other axes when the wheel is rolling.
Seems like it might benefit from some kind of differential, suspension and/or gyroscope.
1. Differential do that the wheels can go at different speeds.
2. Suspension to help the robot handle rough terrain.
3. Gyroscope to help prevent the robot from falling over.
Also, I think the robot should have a full 6-axis IMU, plus a 3D compass. This way it can more easily self-correct its own movement when something is just slightly miscalibrated.
Amazing! Maybe you could replace those white treds with slanted roller contacts also. Making a "Split sphere Mecanum Wheel" or perhaps an omni wheel design with 3d printed linear TPU barrels. It would reduce a lot of friction and would love to see the batteries using less.
For a low speed omnidirectional robot I can see this working really well. But for a car I'm a little sceptical to be honest. Because it is a ball it has a contact area which will probably be terrible for breaking performance. I'm also concerned about the rotation speed (and efficiency) when the wheel is rotating almost on the tips.
Also, if a car drives at 100km/h and it rolls towards the tip it will pick up a lot of speed, but then it rolls over the little wheel at the tip and when it rolls even further it has to suddenly rotate in the other direction at its fastest speed. I think this will cause a lot of wear and possibly even accidents.
But as I said, for low speed omnidirectional robots these wheels may be perfect :D
I would love to see this in a game of some kind. Absolutely lovely.
love this design its so simple and beautiful
Amazing content! Really excited to see whatever comes next. It really moves incredibly well. Slap some treads on those bad boys, and you'll be all set!
"Good all-rounder!" Love it!
This could make an excellent remote controlled camera mount as it can move in any direction and keep the camera in the correct orientation
I'd love to see this concept combined with the foot-wheel thing.
😍😍This is soo amazing. Not just this vid. The whole channel. I am incredibly impressed by all of it.
great job, I'd assume you could make i more stable by leaving more distance from the center of the bot and the wheels, and in that increased gap, you could put some of the component to lower the center of gravity. if the brackets arm for the wheels are also flexible... you've got a basic suspension system.
That really does move wonderfully James!
Excellent! If you put the motors to the side instead of above, your cog will be better. I'm not worried about the dishwaser, but I do worry about tipping over if you build this up higher, the wheelbase is also small.
A garbage can that follows you is such a great idea for people with big work areas.
I half expect you to upscale this and making yourself a rolling throne 😁
That's got to be the coolest robot you have made so far
I think that the gap between the two hemispheres could be removed completely by using three sphere caps instead of two hemispheres and, consequently, a triangle shaped beam going through the center with three branching beams going through the sphere caps and a little wheel on the end of each of them.
Geometrically the three caps can be constructed by cutting from a sphere a prism that goes through the center of the sphere and has a regular triangle base which can be inscribed in the circumference of the sphere.
The main trasmission mechanism would stay the same, with the pulley acting on the base of the center beam, but there would be an extra rotating part and another wheel in the middle of it, which, as it turned out from your prototype, is acceptable. The nice thing is that the sideways rotation would go along a (ideally) perfect circumference, without any (big) gaps.
if you don't understand, sorry for my bad english
I am on a robotics team and we made a very similar ball drive. To make the ball drive be able to climb over obstacles we printed hemispheres out of flexible filament. I would love to see a version of the ball drive with squishy wheels made by some one with as much expirence as you.
This seems like a good application for a deep learning A.I to figure out optimizations for the system of movement, i.e vibration reduction, g force reduction, wear balancing, etc...
Your creation is another level nothing to say 👍
absolutely smashed this design!
Literally reinvented the wheel
Amazing work. Im looking forward to part 3
Very cool project! One concern while watching the video is that it seems a hard deceleration almost tips it over, maybe add metal beads inside the hemispheres of the wheels to lower the center of mass would also be a useful improvement? It seems quite high at the moment with the big motors up and the batteries up. And if you add more stuff on top it will be even worst. :/
I’m curious as to the longevity of this design as compared to traditional wheels or other omni-directional designs. I’m wondering if the fact that to go anywhere instead of just rotate you need to use opposite torques to generate forward force might lead to some amount of bending of the axles if they have to deal with some load, like say if the robot it’s carrying is quite heavy,
this is awesome, a really easy way to pull off omnidirectional travel, and surprisingly easy to code for by the sounds of it!
I have a suggestion for improving the ball wheels though, It came to me after watching that the driven axel really has no reason to need to be a cylinder you could use some 2 mill flat metal plate and it really wouldn't matter, it would mean you could get those hemispheres super close.
Sweet project. Looks like you could lower the center of gravity too. Love the innovation.
I love where you are going with this. I invented a way to drive the spherical wheel without interference in engineering class and have had it on the backburner ever since. This was actually before goodyears tire design came out but theirs inspired me to keep innovating my design .I wish to invest in a 3d printer so i can test my hypothesis .
OHHHH I LIKE THE WAY YOU WENT WITH THE AXLE WHEELS!!! i had the idea i suggested on your last omni wheel ball vid that perhaps you could take the same concept as anolog mouse trackin sensors of a ball that you than put into a divit that is a bit more than a half circle with room for the ball to rotate so it cant be pushed out but allows easy rotaion in all directions although the way you went is intresting too of from the little sinppet at the begining im guessing that its a normal wheel but its mount allows it to spin in a 360 degree circle and the premise is when it gets on the axle side the flat of the whell will allow it straighen to the direction it needs to to move
So happy I stayed up late enough to be here for the new vid!
I'd love to see more of this. It's fascinating.
Amazing work as usual. Congrats!!
James @ 10:30 : *Taps the platform*
USB Cable: "I ATTAC!"
James: *Does not affraids*
I would love to see a head-to-head competition between these wheels and the wheels on the pepper robot.
I think you meant to say love love
You're just a stud like my friend was growing up. He's rich somewhere doing great projects like this.
I was thinking about the gap between the two semi spheres.. My first idea was to fill it with two sets of link treads, following the shape of the sphere and returning along the threaded axis back. It does not need to be connected to spheres, only to axis and free-rolling. The thing is - it is easy to brainstorm such things, but not easy to make it work. Anyway, nice concepts.
That worked better than I thought 😆 nice 1
This is really cool, James!
Excellent project
please never change the part in your outro with the dancing robot- it goes so well with the music that it just wouldn't feel the same
potential exp. given each drivebelt is located near the wheel, adding suspension components to each knuckle(im assuming the black prints) that bot could haves some incredible articulation in each strut making it a very adept offroad/ rough terrain vehicle. love your work!
Amazing James. Congratulations.
Did it take long to tune for the first motorized shot at 7:06? It seems like one of those things that so straightforward I would be baffled if it actually worked the first time
Ive printed 100% fill tpu for wheels. They work really well. Rigid enough to hold everything and still flexible so grippy and quiet.
Great work, very interesting. Looking forward to more
A bigger apex roller with spherical ends to match the hemispheres might help a bit too. Give better ground clearance for when it's on the ground. Might also be able to eliminate the gaps between hemispheres by using three hemisphere/apex roller assemblies.
When James mentioned robotic bin, thought of the little robots in Zorg's office, in the film the Fifth Element, that cleaned up when he smashed the glass.