this entire project is so awesome, the fact that this is all made with the limitations of 3d printers really shows how crazy the tech is. Imagine how simplified a metal-machined version of this could be! Serious innovation

The level of dedication to design and print each individual part of this build… over several iterations!

@IBoughtItMyself

2 жыл бұрын

Also shows how brilliant James is

This is ridiculously satisfying to see how it all works. Insanely underrated

2:00 Thinking quickly, James constructs a home made speed controller using only some string, a squirrel, and a speed controller.

This is an amazing project, one thing that I noticed is that you could have used spiral bevel gears to reduce the backlash, by using the helical gears you would achieve a constant contact between the gears, although there is the problem that the friction would increase, it is not something that I imagine would be harder to 3D print.

@LegoDork

2 жыл бұрын

Backlash, backlash, backlash... This isn't the first, nor will it be the last, while backlash is continuously kicking his ass. Personally, I don't think he wants to do the finish work on the parts.This is prototyping, but part of that of that is refinement. Clean the edges (deburr), and lubricate, Spiral bevel and herring bone FTW!

Awesome project! Not to be an annoying armchair engineer but try putting a 10k ohm pulldown resistor between the analog pin of the potentiometer and ground if you are using an Arduino to read the potentiometer data. This should hopefully clear most of the electrical noise because some cheaper Arduinos don't include them in their circuitry. I ran into the same problem a while back.

@jamesbruton

2 жыл бұрын

Thanks!

@eDoc2020

2 жыл бұрын

Any noise reduction achieved from this would likely come from the reduced output impedance, but it would have the downside of changing the potentiometer's taper. Another option would be using a smaller-value potentiometer. I would suggest a 1 k-ohm potentiometer.

I seriously love how the shoulder and elbow joints use the same clutches and gears to move. The fact you control them by moving them in the same direction or inverse of each other is brilliant.

You're honestly fair bonkers mate. Talent beyond comprehension.

I was waiting for a new video for two weeks!! This is one of my favourite projects of yours.

It's amazing how you've got the arm working in so many different axis using only 1 motor

So happy to see another upload thanks for blessing my feed

To account for the clutch dead spot with software you could use an observer style position control. You could manually measure the dead zone and max-min points and then pass the requested movement into a function that calculates the actual clutch output movement (instead of assuming a directly proportional servo-pos to clutch output relationship). This could constrain the requested servo/clutch movement and calculate the actual arm movement speed to drive synchronization of the inverse kinematics of the next stage. The second arm stage controller could use the same observer system but with the offset of the clutch (servo) movement such that the output will still be within the servo/clutch movement range.

Thats impressive you have that much control James. Brilliant project

Awesome job! It is so satisfying to see everything move off one motor.

Great job James, impressive.

An example of a robust cvt is a snowblower's transmission. However, to actually change speeds you do need to momentarily disengage the clutch from the motor.

Another awesome guild, thanks!

very interesting. i would be amazed to see what this kind of concept would be best used for. i think this kind of arm robot isnt the best use as you found out, but in the right application it would be amazing.

Hey James, quick feedback on your position control for the base. Since you are using a simple proportional controller, you will always have a position error that is inversly proportional to the gain of the controller. You could get accurate positioning by adding an integral term to the controller. Great video!

seriously though, hats off. You are amazing!

Works just about as good as the Radioshack Armatron from the 80's, Still a really cool project though, well done James, BTW haha good one "Simple trigonometry" math is anything but simple for a lot of people lol

Awesome! It'd be super interesting to see a similar approach with eCVT, planetary gearbox like hybrid cars use

If I ever have any sort of formal event or party, I want to invite James and his robots.

I remember waaaay back in your channel when there was talk of an iron man suit up gantry. With this project i thought you were finally going there. Maybe consider that if you ever decide to revisit this to work on the transmission etc

You could use mono-directional tension from a constant force spring located at the shoulder joint to reduce the majority of the backlash. This would press the arm in one direction so that the geartrain is always actively meshed. This would be largely the same thing done by gravity already, but to a greater and more controllable degree.

Mükemmel çalışmalar

This is exactly the kind of engineering challenge I’ve been thinking about! It’s like having an engine that drives all the power and you just change how it gets diverted via gears, pulleys, and chains

It exists a version of backlash free gear, in essence two concentric gear join by a spring, you charge the spring before installing them, might be something to explore like the motor

This is cool to see how far the mechanical side can be pushed. Can you please comment on the idea of using open dog lower leg on a shoulder type joint so it can lay flat on the ground and the upper leg rotates through 360 while the lower leg remains still for “emergency mode” to get through deep soft sand snow or slippery.

This is really inspiring, and also the colors are so nice. But I have some serious work and studying to do as a beginner, including learning how to use the design software. Just seems overwhelming. I aspire to design and make medical devices/technologies professionally.. The potential of 3D printers though for making at home prototypes of such is probably huge.

Mesmerizing mechanical motion :o

If anyone's interested in getting more into the part where the controller is setting the motor values to get the desired behavior, i's called Control Theory. By knowing how your motors and sensors interact with the arm, you can get great performance even under non-ideal conditions. A PID controller is a very powerful control method, but there are a lot of other things you can try in order to get the results you want.

keep an eye on this guy... he is one cat away from becoming a super villain...

If you removed the 2 mirrored horizontal shafts and replaced the gear with a double sided one you could reduce the amount of backlash. This would also make it easier to mount the feedback pot as the base of the arm would touch the sides directly.

Sir, you should attach a gripper, it should look on the CVT Robot Arm.

you should make a version 2 with the same base but using t5 pullies also you could use the clutch system to move it on a set of rails and add it to your great ping-pong ball contraption with the use of a dumping bucket or container that can carry many balls at once

I'm absolutely in love with this project, I've used the Tandy Armatron as the basis for many robot arm projects, although not as big as yours, my flashforge adventurer 3 prints the same size gears as the Armatron and to the same if not higher tolerance, is there a particular reason why it has to be so big, I am also curious as to why you want to control it via servo's, as it defeats the purpose of having it all run off a singular motor, why not just use the servo's to make the arm?

I'm looking forward to implementing a compound block and tackle cable system on this framework and making a Dexter's Laboratory style backpack exosuit with electrolysis hydraulics

It looks awesome ;) 👍

Oh man. You really do come up with a new project every week. Could you tell to other creators what is your trick to always keep so motivated? What do you do to never burn out? And ehat if you get stuck in a problem you can't solve, what is your trick to overcome and always produce this amazing content?

@jamesbruton

2 жыл бұрын

I just make it up as I go, but there's about 3/4 week planned ahead of me.

This would be perfect for a water mill. A medieval robot arm to pick up heavy sacks of grain.

This guy's gonna invent the first real general AI by accident.

You may want to take a look at "SOLIDWORKS SIMULATION OF IVT GEARS" it's basically a planetary gear but with 2 meshed planets, which makes it so that you can drive the middle with the input shaft and the outer ring with a one way cvt (even 0 speed is not necessary) and have backwards, stop and forward.

Awesome!

i think you should make a robot that uses a wireless optical mouse as a position transducer. two mice if you want orientation as well. or a transducer on a linear actuator. or on the rim of a wheel to measure rotation

Hi! Will the solution with single long shaft goong thru the whole arm with constant RPM (with cardan on every movement point), and steering noovement via cvt-ball-clutch with cykloidal transmiton? The whole system would be less complicated (i guess?). Especialy steering

Very cool

Hello sir ! Is it possible to make and program robot arm for cooking ? Just like molly robot. Could you please make one ?

Don't you want to become a professor for mechatronics? Come to my University in Germany. My people need you :)

Awesome.

Man explores this random idea that popped into his head and measures filament for that in spools. Meanwhile I feel bad having to reprint a ø60mm gear.

Next, do a robot that passes butter.

Radio Shack's Armatron did all this over 40 years ago using plastic gears.

Not to be nitpicky but a servo is a motor so it is actually a 3 motor CVT robot arm, but it is still an awesome project

Reminds me of the armatron from radio shack

rotational motion with cables inside slippery cable housing for cable routing, think about bike breaks but rotational, not linear

What if you put some springs or rubber bands to keep tension on the joins and cancel the backlash? Would it work or maybe it would make too much friction?

@jamesbruton

2 жыл бұрын

I considered it

Комментарий в поддержку канала и ролика, а также труда мастера.

Hi, The question may be silly How you doing the 3D designs ? What is the app name?

@alihms

2 жыл бұрын

Fusion 360. You can see that in minutes 0:25 to 0:40.

What about using that fishing line instead of the gears?

This old tony had à mill which ran 3 axis from one hydraulic motor. How about a hydraulic project?

Mate DARPA or NASA, job for life👍

La c est de l ingénierie !!!!! Cela fonctionne un peut comme un différentiel automobile

Servo also have motor so we can't say single motor robot??

At firs I read it as CBT, and was quite well.. amused.

I would like to see a hydraulic robot arm with one powerful motor to control all the joints like you described. Maybe it could have a gearbox reduction with 2-3 modes for insane power. The gripper could be powered by the same motor for crushing strength. I imagine a disadvantage of this design would be that the joints could not rotate continuously. Hydraulics might not be your area of expertise though.

@eDoc2020

2 жыл бұрын

Apparently there are hydraulic slip rings so continuous motion could actually be possible. However they appear to be expensive, especially if you want to pass lots of hydraulic lines.

you've built so many robots and yet have not used them to rob a single bank or anything

What if you control the joysticks with another 1 motor CVT robot arm 🤔

I'm counting 4 motors so far in your 1 motor arm

🔥💕👍

오오 한국사람 댓글은 내가 처음 ㅎㅎ

One motor plus the ones in the servos..

W

If you're using stepper motors as well as one driving motor then you're not using one motor really. Are you trying to save money on motors somehow? I'm not seeing the usual appeal. Servos have motors in them. What are the advantages of doing it this way?

@jamesbruton

2 жыл бұрын

As I said in the video, it's experimental - but what you describe is the same as Hydraulics. The proportional hydraulic valves in the Mantis cost around $2000-3000 each and there are 18 of them.

Adding my astonishment at your ability to realize concepts like this. Sure they need refinement but a functioning proof of concept takes you very far down the path. I think you underestimate the significance of a one motor design. These days every design is overly complex it seems, and "gutless wonders" which get the job done with a minimal amount of components are becoming scarce.

early gang!

I hope you're recycling old/failed prints. That's a lot of plastic.

Fifth

First

@nicholasballinger7429

2 жыл бұрын

With the formalities out of the way. Really quite inspiring work.