Brake discs would make great well balanced masses - even from the junkyard. Great bot!

@celeron55

3 жыл бұрын

These days they tend to have cooling fins which would waste power and make unnecessary airflow to mess things up. Find a model that doesn't have those.

@user-fe1od7yg8v

3 жыл бұрын

Brake discs have a lot of excessive mass, because of their moment of inertia, specifically mass, that is to close to the axis. Much better would be, to use some form of rings. And also brake discs would have turbulent air, which will mess up balance.

@ryanokeefe12

3 жыл бұрын

@@celeron55 Maybe on 5% of cars (performance vehicles).. the other 95% have standard solid discs.

@ryanokeefe12

3 жыл бұрын

@@user-fe1od7yg8v A flywheel off of any motor... Flywheels are specifically designed to be as light as possible, for overall weight, whilst having as much MOI as required to keep the engine rotating smoothly.

@cambridgemart2075

3 жыл бұрын

@@ryanokeefe12 Almost all modern cars have vented discs.

Just having the robot lean should make it turn. It's not going to turn on a dime and I don't know how easy it would be for the robot to maintain and recover from a turn. The reaction wheel solution does remind me of riding a unicycle and doing quick turns. It's like a quick twist of your hips and then correcting the over/undershoot.

@sofuckingannoying

3 жыл бұрын

Look no further than cats. Cats know how to turn in mid-air with no fulcrum, reaction control thrusters, or aerodynamic forces, while also preserving the angular momentum. Would be a very cool project for James.

@jakobgschwind4722

3 жыл бұрын

If he uses a wheel that is wider and round on the outside that should do the trick. Maybe it would also be better if he put the gyroscopes on the sides of the wheel.

@RichardBronosky

3 жыл бұрын

Yes. He already has his answer in this video with the Electric UniCycle. On an EUC, you do all your turns by leaning and using the smaller diameter of the edge of the wheel which causes you to turn much like slowing the inner track of a tank causes it to turn. I don't know why he even asked the question.

@ilyasofficial1617

3 жыл бұрын

@@RichardBronosky maybe the complicated one is because leaning forward could make it fall. i suggest maybe put a limit on how much lean angle is permitted (tolerance), say 15, when press forward. if not pressing forward, then activate balance mechanism in video.

@RichardBronosky

3 жыл бұрын

@@ilyasofficial1617 I understand what you are thinking, but that is not how it works. It's not [directly] an issue of angle. It's about maximum torque and maximum speed. These balance devices are most comparable to balancing a broom on the palm of your hand. You keep your hand under the center of gravity and you and the broom stay relatively still. If you want to walk with it, you pull your hand back slightly and it begins to fall forward. As long as you can walk/run fast enough to get back under the center of gravity, you can keep it stable and even stop it. If not it falls to the ground. The same is true on Electric UniCycles, OneWheels, etc. The same would be true here. While there is vicariously a maximum lean angle, it's determined by the mass being balanced and the torque required to raise it back to vertical. Search for "OneWheel nosedive" to learn more.

This whole video I was on the edge of my seat hoping James is going to reveal that he is making "Gizmo Duck" from Duck Tales/ Darkwing Duck.

@jameswalker199

3 жыл бұрын

I'm not a huge disney fan, so the closest I came is thinking it looked like Mo fro, Wall-e (which still is disney, I guess, but I've seen Wall-e and I haven't seen Duck Tales)

@suncat530

3 жыл бұрын

@@jameswalker199 Wall-e is Pixar, which is a company that was bought by Disney, so what they make is quite different to "normal" Disney content

Shame on you guys for telling James that his gyros would break up and kill him - lol

@kirkc9643

3 жыл бұрын

I laughed when he said that. Anyone who has seen at least a few of these videos should know better

@bizzehdee

3 жыл бұрын

Yep, that is the sort of thing that any man would want to learn on his own, the hard way. Speaking for myself, i would rather something fail and learn why it failed, than to never push the limit of something in the first place

"I'm gonna make bigger and scarier version in the future." My man!

This channel is amazing and I love all of these robots. I am inspired to do the same later on in life because of this channel

@manojsk6189

3 жыл бұрын

Yes Im also loved this channel

For the steering, maybe (when one of the batteries is the front) making the robot lean to the side you want it to steer it in with the gyroscopes while its driving forward can make it turn, and then when the gyroscopes moves back it will just straighten up the robot, not turn the other way.

@PhilCoati

3 жыл бұрын

This may take a wheel redesign to work but I think it could.

Next step: One-ball wheel self-balancing robot. GO!

If you make it rotate those gyroscopes on another axis it would be able to turn on the spot that way instead of having to lean to one side while moving forward.

Nice, one step closer to making a real gizmoduck cosplay.

@franksonjohnson

3 жыл бұрын

O great gizmoduck reveal your secrets, how do you turn??

You're constant output of complex and unique builds never stops amazing me!

To make it turn, replace the mechanical gearing 4:16 by two servo, when turning one servo needs to turn a bit more or less than the 1:1 ratio... it'll sure be tricky

This one looks kinda badass James!! Nice work!

Awesome work James and always a pleasure to follow your creations!!!

I love it! Can't wait to see it stand still with, rotate in place and drive around while turning!

Turn this one into "Claptrap" and your channel will explode my friend.

@ClickItYT

3 жыл бұрын

Like his channel isn't already near the 1 million subs. :}

@HoRiOnS

3 жыл бұрын

@@ClickItYT Yeah, I know it's gotten big. But with the new Borderlands movie comming up, and Claptrap being such an iconic character, I could see that doubling or more. If this movie is a succes, the person who has a real claptrap robot is gonna draw a lot of attention.

@blakelowrey9620

3 жыл бұрын

Yes pls build ai claptrap

@filval387

3 жыл бұрын

@@HoRiOnS "Borderland movie" I beg your pardon?

Love your videos James!

Hmmmmm the full rotation gyro idea sounds interesting, but I feel like there has to be a super elegant solution that I’m too dumb to come up with. Excited to see what you end up going with!

Loving the beer socks! Keep up the good work.

Awsome! Wack a reaction wheel on Top!

One wheeled robot........ClapTrap from Borderlands? That would be pretty cool.

You took my advice on the one wheel. Great.

You should build one to look like the TV bots from fallout! Would be super cool 👍

@AlphaLegionTacticoolShitposts

2 жыл бұрын

That's what brought me to this video

This is a great start to a Claptrap

Congratulations on one million subscribers! You inspire me to work on my ambitions.

Excellent job

To implement turning you could have loaded springs that suddenly release and the sudden angular momentum of the moving piece would be canceled by the turning of the robot, then if you slowly re arm the spring static friction will stop it from turning back

So close to 1 mil!!!

Can't you use the gyros to hold it in a lean and that would cause it to turn?

@Snapum

3 жыл бұрын

I agree. I feel like you could do something like you did with the BB8 droid; when you basically moved the balance position to make it move. IIRC? So if you move forward and adjust the left/right balance offset to the left or right, you'd get a lean, and thus a curve. But then - it might lead to an exponential failure and it just falls down. XD. Good luck ~balancing~ that! :)

@MichaelAPede

3 жыл бұрын

Yeah, like a one wheel motorcycle. It would need to move to turn, which is ok in my books.

@willjhcave

3 жыл бұрын

Yup. This 👆

@CDRaff

3 жыл бұрын

The only way to get them to make it lean is to mount them horizontally like he did in the last one. Then the problem becomes that if you set it to lean one way the second gyro(which is spinning the opposite direction) would counteract that force. You could spin them the same way, but then that would make the whole bot spin in place unless it was turning.

@paulsaunders5555

3 жыл бұрын

Ok, I may be oversimplifying but if you put a pivot and a servo below the vertical giros would that not tilt the bottom wheel as you want but keep the geniral structure the same. Probably a hideous idea. 🤭

I've been working on doing this with reaction wheels although I've been leaning towards gyros or a combination after your last videos. Looks cool.

You could have taken your two wheel robot, added a third but have them as hubs sitting inside a large wheel.

I like this. . . its satisfying somehow

Very cool content. Thank you for a great video experience!

If you control three or four control moment gyroscopes independently in a roof or pyramid configuration you can generate torque around all three axis. This would let you use them for stabilisation and rotation. CMGs are a fascinating topic, thanks for making videos about them. For anyone interested in them I can highly recommend 'Spacecraft Momentum Control Systems' by Leve, Hamilton and Peck. If you or others have questions, give me a shout. I worked with CMGs during my PhD.

This is sick!!!

You have excellent quality content on your channel! Keep up the awesome work!

I don't know for sure but you may use two separate servos for the gyro and operate them accordingly to turn

Eventually you are gonna have a working Claptrap unit from borderlands

Damn, so close to 1 mil subs

Finally a relevant sponsor that's not a scam.

It would be cool if it could rotate with a reaction wheel, so it would work with 3 different systems, with gyroscopic precession, inertia of the device itself and with the reaction wheel.

A way to make the robot turn is not to create a way to move the robot in a yaw rotation but instead use the gyros to turn the robot in the roll rotation This would turn the robot as yaw is a secondary motion to roll.

It is possible to have the gyros move 360 degrees. If you imagine two perpendicular bevel gears, one large (attached as part of the gyro) and one small. The small gear is attached to the gyro motor, and the big gear is held by a sort of carrier and can be rotated/orbited around the small gear. this does add a slight orbital component to the rotation, but there are ways around it I'm pretty sure. Alternatively you could have two large bevel gears engaged on opposite sides of the small bevel gear. This would give you two counter rotating gyros as well as canceling/evening out the orbital motion.

very nice build :)

What about putting it on a single ball shaped wheel instead of this regular barrel one, in order to go sideways ?

@CrazyStranger11

2 жыл бұрын

I think that would put the gyroscopic precession off axis. All the balancing would be done by the wheel.

Im amazed at how many sponsors there are here I cant even get mad at it thats good business

Super cool! Interesting you mentioned that the weight needs to be low for the giros but high for the self balancing wheel. Would be cool if you would share your design process for these robots!

What if you varied the speed of the gyros? Would spinning one of them faster or slower cause it to turn?

@MrDianGamble

3 жыл бұрын

That was my 1st thought

@bruceneely4859

3 жыл бұрын

I was thinking the same thing

@gingermanc

3 жыл бұрын

Yeah I thought this might work because a drone can yaw without pitching or rolling using a similar effect

@MadScientist267

2 жыл бұрын

@@gingermanc it's not gyro effect in a drone however that causes this... It's simply torque imbalance. Slower motors impart less torque than the faster motors, and the drone yaws as a result

Suggestion for how to turn: a weight that shifts side to side to change the center of gravity, which in turn changes the wheel from running on the center to running in the edge, which would cause it to turn as the wheel rotated. The tricky bit would be getting the gyros to balance the robot at an angle instead of just balancing to vertical.

@LastRoseStudios

3 жыл бұрын

Maybe mount the batteries in front of and behind the drive motor, and shift them left/right to change the center of gravity?

Now make one that can balance on a super thin wheel

wow good job

Genius. Love the beer socks. Maybe it's as simple as pneumatic rods that push down on one side or another with little casters?

One way to make it stable would be to make a proper physical model of the system and tune the stability that way... (in some software like matlab)

very interesting and very clever

Yum, that's some lovely gyros

Someday im gonna download this *3d printing music* just for the nonstalgia

Awesome!

Impressive !

I remember Disney labs made a trapeze robot with gyros so it always lands on its feet.

You have remade the drones from "Dr Brain Action Reaction" videogame from 1999. I think you should keep developing this. I ride an EUC over 30 MPH regularly. It's amazing and it makes me see there is so much potential with one wheel designs!

you could use a reaction wheel to steer

Damn this is cool

Nice work, very cool! FYI these are called CMGs (Control Moment Gyroscopes) and are commonly used on satellites/spacecraft to perform attitude control without the need to eject mass. To maximize control effectiveness you should be looking to maximize angular momentum in your flywheels, where angular momentum = moment of inertia x rotational speed. Since moment of inertia is effectively mass x radius^2 you have 3 ways to get more bang for your buck: (1) increase mass by using copper/brass instead of steel (tungsten would be great but $$$), (2) distribute that mass as far out as you can (e.g. wind copper all the way around the rim instead of using discrete BBs, (3) get the RPM up as high as you dare. I reckon, from quick visual inspection, that you could at least double your control effectiveness doing this. Keep it up!

Very very cool

Could make an interesting re-visit of the BB-8 project, no need for drive wheels and worrying about friction if the internal carriage is driven by gyroscopic procession.

Next episode James makes a 3D printed RCS

cant wait for 1 mil

Dang. Thats cool.

Awesome 👌

If the round section wheel is soft enough the differences in the circumference of the wheel over the flat spot during a side-to-side lean will generate a twisting force when driven forwards or backwards, causing a turn into the lean.

This is exactly like an EUC! The only significant difference is you use gyros for side to side balance and an EUC uses the shifted weight of a human and the cochlea in our ears. You created a fully self balancing EUC! I think you have a marketable product concept.

It seems like 2 gyros that spin opposite each other on a frame so their axels are locked parallel could be used as a balanced resistance to perpendicular rotation. So you would mount this frame in another frame that can be driven on an axis perpendicular to the gyros rotation. The drive against the gyros resistance would allow balancing on one axis. Other gyro pairs could be used for other axes.

You'll either create our weapons to fight skynet, or will create skynet itself! Impressive.

The steering can be done with a small rod, hold in the middle, which can be shifted to the left or the right. The shift of weight will make the robot fall over, and now you can make a turn in either direction. And this is just an example, the idea is just to shift the weight to left or right.

That is an awesome robot, would love to see a bigger version you can ride. And have you seen the Abenix: active ball joint mechanism? I think that would be right up your alley with the Ultron build if you could figure out the coding for it. Cheers

For your robot: Use a spherical wheel to give it greater degrees of freedom on the ground, and also use your ball-CVT thing to allow slight variations in the speed of each gyro. For other things: I would love to see a belt-drive adapted for use on a derailleur, or a derailleur adapted for a belt drive. For ranting: I hate "hover boards" because they don't actually hover. They aee also only "boards" in the loosest sense.

Encorporating fans to steer would be super cool! Maybe adjustable pitch propellers or moving ducts

For steering: Just use spring loaded poles which you can extend to the ground. Another idea. Use a big pole with weights on the ends parallel to the ground, and 90 degrees to the wheel on the ground. Moving it from one side to the other, you could lean your robot left and right by moving it. With it you should be able to make turns, although not on the spot.

Mount the gyro on a slide that you can control position to bias weight distribution to control lean. You can use it to induce motion or to bleed off inertia of a reaction wheel if you add that. Cool bot!!

I was going to say you need an accelerometer to keep it from walking off, but an encoder would work as well. Really any method for monitoring x and y movement will work. To steer, you could tilt forward until it starts moving and then use the gyros to lean sideways. Like riding a motorcycle or unicycle.

It should actually be possible to steer without any additional hardware. The trick is to use the gyros to let the robot _oscillate_ side-to-side, rather than standing as still as possible. The effect is that the center of mass will alternately be on the left or the right of the tyre. Then you need to sync the motor-balancing so it e.g. accelerates while the CM is on the left, and decelerates while it's on the right, in order to get a net left-turning torque.

What would the force be if you use the gyros as reaction wheels in the current configuration. e.a. increase and decrease the speed of one of the gyros slightly. Would it be possible to steer that way?

@andylindsaytunes

3 жыл бұрын

Ida know; using gyros would make such a mess, even if he didn't put tzatziki on it first. And the pita would tear and would probably fall off immediately anyway.

@melody3741

3 жыл бұрын

I think reaction wheels are just not fast enough and waste energy.

@charliesteiner2334

3 жыл бұрын

That would rotate the robot forward/back, rather than side to side. But maybe you're on to something! If the gyros were horizontal this could totally work. Only problem is they would saturate super fast.

You might be able to get it to turn by leaning it like everyone else is saying, but a reaction wheel would make it so so awesome.

Nice

How about a centrifugal expanding moment arm? Imagine a pair of weights attached to a spinning pivot via an extending (possibly elastic) linkage. When the pivot rotates, the masses fling out and increase their moment of inertia, when it slows down, the linkages retract in. Alternatively, but much more complicated build, would be to actuate the retraction/expanding of the masses, to have full control over the angular momentum of the thing.

It would certainly be interesting to see one of these that was safe to stand on. The human rider might be able to solve your conflict of one system likes a low centre of mass and the other wants a higher centre of mass.

Amigo eres increíble me encanta ver tus vídeos... un fuerte abrazo desde españa

suggestion... split the wheel. It would still be a single wheel contact surface but when turning one edge would move slower than he other. it would require another motor and encoder and drive controller. or figure out how to gear the difference between the 2 halves. but this would be much more complex. just my two sense. awesome work.

Make the wheel a pair of cones with a flat strip in the middle. The flat part in the middle is for driving straight back and forth. You then lean the robot to the side, on to the section of the wheel that is angled, causing the robot to turn. You can also do this with a more rounded shape, but the math to get the right level of turning gets more complicated, but using a gyroscope sensor to get the tilt of the whole robot can be used to form a PID loop to control it. As for turning in place, I think you might need to simply make the wheel capable of 360 degree movement using a ring and contact system to transmit the power and sensor data. You could also limit the turning to just over 90 degrees and use coiled and wrapped wires like the ones that power the limbs of OpenDog. Getting the motor inside the the hub of the wheel and using compact cycloidal drives or planetary gears to replace the gear reduction belt would also be an idea, but it would raise mass and complexity. This might also require a third or even fourth balancing gyro for stability when traveling perpendicular to the gyros.

he passing by viewers leave a like for James hardwork

@tanjiro3285

3 жыл бұрын

edit : on the video 😅

i suggest using an upright lever with a significant weight on one end of it connected to a servo in the front should provide enough momentum to force the gyro to counter steer to the left or right. Hope it works.

Absolutely love this build and how each build is built upon the previous! Love your stuff. I gives me inspiration for my Claptrap build. Just out of curiosity, where did you find your MPU6050 with pins on both sides? The ones I find only have them on one side.

You could use electrical slip rings to drive the gyroscopes through infinite rotation.

If you get the standstill working a rapid turn to the uniwheel independently in the direction you want to move and then a slower catch up movement to align the gyro may be worth a try. The split wheel idea does seem good too.

You'd want to turn by tipping the robot to one side. You'd kick the rear wheel around to change the facing angle. A bit... very tricky but I'm pretty sure you could program the idea. To turn left, you tip left and speed up the wheel to compensate the tipping force with... momentum. To stop turning, the device untips.

I love that thing :)

To get the robot to turn you could use the balancing gyros to tilt the robot while moving. Much like when riding a bike, the self correcting nature of the moving wheel would cause the robot to turn. Turning in place would not be possible, but a series of backwards and forwards turning arcs could turn the robot around (like a 3 point turn in a car)

wheely cool!

A weight that moves in a semicircle perpendicular to the ground. Move the weight to the left side to move left, to the right side to move right.

A possible way to turn would be to tilt to the side you want to turn towards. By balancing on the edge of the rounded wheel while driving fowards or backwards, the radius of the wheels section that is touching the ground is smaller on the side it is tilted towards, thus travelling a smaller distance in that direction, which results in the Robot turning. I think to make that effect work well the wheel needs to be wide with a small curvature or even a V shaped cross section with a small angle, so that a lot of surface is touching the ground when turning. A welcome side effect would be, that when tilting, the center of mass is shifted towards the direction you want to turn in, which shifts the center of inertia and for that reason would help turning. The main problem is, that your Robot isn't able to tilt for a sustained amount of time as the gyros will reach their endstop after a while, so it would have to turn in small intervals, stopping to reset the gyros a couple of times.