I Made a Robot Arm to Hold My Camera [$500]

[@3DprintedLife]

You and your company can try Onshape for free at www.Onshape.pro/3dprintedlife Subscribe to make sure you don't miss my next awesome project! kzbin.info

[@AgentCain]

Have you seen, that RobotAnno has a 3d Printed Robot for educational use ?

[@szabolcs__]

11:12 It would be nice to blance aout the arm like a crane , for softwere Freecad robot part. For the end use a worm gear type power trans mission.

[@chaseing_chase]

Thank you so much for making this video, im a senior at Widener University and studying Robotics Engineering and all the stuff you showed is exactly what we have and are learning about!

[@danielczech560]

I would be excited to know more about the re-design you mentioned before the github link :) Does it make sense to wait a bit longer if i would like to give your project a try soon ?

[@MrAussieTeen]

the goat is back

[@eaglenebula9080]

🐐

[@jdhasfun7087]

Ohhh yeah

[@marsgizmo]

impressive progress! great work 👏😎

[@Uncle_Farts]

Sup I m subed

[@3DprintedLife]

Thank you! Cheers!

[@moji3812]

Yeah👌

[@aleksjenner677]

Sick bro. The filmmaking potential with motion control is massive. Makes compositing a piece of cake!

[@aleksjenner677]

@viewz Seems you don't really know what you're talking about, Bub. Software has no effect on accuracy, motion control has been used for filmmaking since the first star wars movies in the 70's. As long as a motor knows where it is, and can return to a set position then it'll work. Only thing that will ruin accuracy is mechanical slop in the system.

[@musikSkool]

Thanks, now all I can think of is a slow moving camera dancing around a piece of cake and showing it from every angle with a green-screened background. You know, compositing a piece of cake.

[@musikSkool]

Most of the cost of high-tech things isn't the materials. It is the countless hours of tinkering done in a back room somewhere to get it to work. And then twice that amount of time again designing the tools to mass produce it in one tenth the time and guaranteeing that the mass produced ones still work as well as the prototype.

[@CruzMonrreal]

A suggestion. Even if you switch over to PETG or another material, all of the encased stepper motors will still overheat and keep causing problems. They at least need ports for ambient air to reach them, maybe even a heatsink too

[@BL4S7ER]

ABS or Nylon?

[@CruzMonrreal]

@@BL4S7ER Changing material won't change how some of the motors are completely enclosed. Motors that overheat are motors that will lose steps at best, and fail at worst.

[@MoppelMat]

I disagree. In my voron 3 stepper sit in 60°C and print fast in there. I can not imagine this to be more than an inconvenience at best for this design.

[@molomono9481]

@@MoppelMat What are you disagreeing with? That thermal damage causes failures for the motors? (check the datasheet it's in there +60 degree c is within the normal operation range). That stepper motors need to be able to vent? (If they are drawing sufficient current they get hot enough to be a problem) Just add a grill pattern to the motor housing part and or skeletonize the fixture, if they still run hot add a small fan. CruzMonrreal is just giving good advice.

[@MoppelMat]

@@molomono9481 I said the motors should and can cope with 60° env temps.

[@sublucid]

Very impressive! You can dramatically increase the smoothness of the movement by setting the high-level acceleration/speed/position parameters of the Trinamic drivers rather than using step/dir commands (assuming you have it hooked up).

[@3DprintedLife]

Ahh yeah I do have that hooked up, that should help. Thanks!

[@justinc2633]

do you mean to use the uart option? my tmc2209 has that, but ive never looked into its benefits over step/dir

[@Barnaclebeard]

Wow. This is really impressive. Really great work, Inanimate Carbon Rod!!! 3:43

[@TD-er]

Just wondering, why don't you use a "mechanical" input first? Thus make a very simple clone with the same ratio for distances between hinges and then simply add a potentiometer on each hinge. Then you only need to 'copy' the angles on those hinges first to your motors. If you get this communication smooth, you know any computed intermediate step will also work on your arm too. This will also help you find any issues where each hinge may struggle to keep up with rotational speeds on the input. As a bonus, you can also log a few basic movements you can make by hand so you can then try to experiment on the needed resolution for how many intermediate steps are needed and whether this may depend on the angle and weight of the moving hinge. It may also make it a lot easier for you to try movements from point A to B where you can easily see up front what would be a better bend shape of the arm. For example I can imagine it may appear more smooth when the movement from point A to point B isn't a straight line, when for example some hinge may need to lift considerably more weight and needs to accelerate beyond what it can handle, when moving in a straight line.

[@creativefun1758]

Just an idea i use on my projects add a small camera to the bottom of your robot head that way it keeps track points and you can use that to make the movement smoother and possibly write a program that does it for you

[@PiefacePete46]

The resultant video is much better than you might expect when you look at the phone in motion... the potential is very encouraging! Thanks for sharing your project. Like & Subscribed.

[@dexterousmuffin]

6:29 lmao subscribed, great work. Also, many interesting solutions. Keep up the great work man!

[@ndragon47]

Heck yeah! I Just watched the other 2 parts of this the other day! Glad you see that you could finish it!

[@chrisgenovese8188]

This is badass. Now my life feels empty without a desktop 5 axis robot arm.

[@thomasaddison4268]

This guy is criminally under-rated

[@hudsonreynolds4349]

Love this project. I defo want to build one. Is the updated project going to be released in the next six months?

[@3DprintedLife]

Unlikely. I have way too much on my plate so this project keeps getting bumped :/

[@FlyGamingChannel]

This is the first video of yours that I’ve seen and it deserves an instant sub. Interesting and entertaining - thank you.

[@Roy-K]

I’m in my senior year of Mechatronics engineering technology, and I’m taking a class all about robot kinematic and mathematical definitions - I totally feel your pain lol

[@Roy-K]

All I can say though - as much as matrix math sucks, the equations really help in terms of actually getting smoother performance. The trick is to just define the initial configuration (when everything is at zero) and define your joints (which will take a while for sure). After that, program in the transformation equations like Power of Exponentials and let the computer do all the hard work The best thing humans ever did was tricking rocks into doing math for us lol

[@SP-ny1fk]

You could program it to track your gaze, and hold the phone at an exact distance from your face, so that it occupies your entire field of vision - this way you can entirely do away with a large monitor.

[@ScottJWaldron]

Wow! Fantastic project and video.

[@mariusj8542]

Im struggling with much of the same, I love stepper motors, but when it comes to torque and smoothness when handling mass, servo motors are having value. You can get a driver that handles electrical wave construction, like ie. sinusoidal or trapezoid and off course the accuracy of an encoder for feedback. Holding torque can ie. be simulated via the encoder, just pushing the other way if the error rate between expected position and actual position starts to drift. In addition and even though it’s more expensive, current (A) will help with torque. Anyway, love your project!

[@SaitoGray]

That awesome. I was contemplating building my own too but i lack the knowledge to do so. That amazing work, can't wait to see it finish, and who know, printing one myself one day.

[@sourmcnuggets6178]

You're such an underrated channel! Just found you and have really enjoyed watching your antics.

[@centurialinc]

I would try closed loop stepper motors. I use them on my cnc mill and lathe and they work well. It's also a simple system. The stepper drive gets the command and simply confirms that it made it to that position. You choose the amount of error in the software and specify how much the drive needs to try to reach or hold position. If you buy those it will likely change the size of the robot as the motors come with the encoders attached and it will definitely increase costs. But it would speed up the completion of the project. The software is the main issue with these types of cnc machines. I would focus primarily on that. The AR3 robot project by Chris Annin is very well documented and lots of software examples and tutorials are available you should check them out if you haven't already. Good work keep it up and I'll be following along.

[@Drankoii]

That's amazing! For a long time I'm planning to make robot with 2 manipulators and place it on rails under the ceiling, to help me with home chores. Glad to see that something similiar already available to make at home.

[@burggerbig102]

A bit of PID control would allow for more smooth movement than the stuttering at the end. But this is so impressive, can't wait to find a time to make one of my own

[@justinc2633]

how would you add pid control to an open loop system?

[@OwenBuckingham11]

your solution to IK is actually entirely valid, I went to school for robotics and they taught us the matrix version but also the geometric way to solve for IK. So nice job it's pretty legit :)

[@PCBWay]

Welcome BACK! Do more uploads pls

[@Gee-Dee-q1e]

I enjoyed that video way more than I thought I would. Great job with its comedic aspects. I'll probably need to rewatch this a couple of times, but outside of being enjoyable, it was very informational.

[@3DprintedLife]

Thanks, glad you enjoyed!

[@speedracer9132]

I just subbed to you for this arm from your earlier videos! Glad to see this video!

[@thijsvanliempd2824]

finally; a youtuber who needlessly makes his own stuff because the existing stuff is too expensive BUT i don't have to feel incompetent because this youtuber's design dicisions (oopsies) and production speed are actually realistic to what a viewer might achieve. Ah, the sweet bliss...

[@gearscodeandfire]

Your honesty about your roadblocks both gives me vicariious anxiety while reassuring me that this happens to everybody. Great work

[@freakinccdevilleiv380]

Amazing production quality on this video 👍

[@zincfive]

Wow this is brilliant. Interesting project, good production.

[@Atmos_Glitch]

*_Mom & Dad_* Really came in clutch huh???

[@Rcmike1234]

If you ever want some easy weight savings without changing much, ABS or ASA is ~ 1.04 g/cm^3 vs PLA at 1.24g/cm^3. Not sure how much that'll save you over the entire arm but thought it couldn't hurt to mention.

[@JPToto]

This in incredibly well done! Funny, I've never seen anyone call out the IK specifically before but I see how important it is. Looking forward to more!

[@caramelzappa]

Awesome project. Next step, mount a hotend on it so it can be it's own 3d printer and replicate itself.

[@paperplane627]

I love the Oversimplified videos playing in the background!

[@UNVIRUSLETALE]

Insane design with mub and cf rods, very nice work

[@timondalton8731]

Your channel is insane. For an engineering student, I get the same feeling when I watch stuff made here. The channel name is cool, but I think something else would help you reach a wider audience.

[@3DprintedLife]

You’re not the first person to bring this up, definitely will consider it!

[@ma.donnabelramos4350]

eventually it all snapped into place and I started learning how to add all the effects, titles, motion text. It was pretty cool to see my

[@datblitz9927]

Hello, i mean or i would pick the coordinates into a buffer, and use an ISR to get it smooth and calculate the next (Upslope downslope and syncro ) position.

[@slatetechnologies1724]

I think this video deserves a celebration because it’s been so long-awaited.

[@adamcboyd]

Patiently waiting for the final revision. Will happily print one and then figure out how to get a PTZ control working.

[@rokketron]

know your basics course is just what I needed. Taking it a day at a ti so I can fully grasp the core of soft.

[@protomake1349]

This is sweet! I'd love to build one if I can ever get caught up on my current projects. I wonder if you could add a 3 axis gyro/acc to each axis and use a feedback loop to ensure the target angle has been achieved instead of encoders. Could also investigate smoothing like the Klipper firmware?

[@DanielSanPedro]

You sir, a genius madman. Well done!

[@user-eh9jo9ep5r]

You could print absolutelly everything, really interesting

[@LeoDaLionEdits]

Love the project. Can't wait to see how far you can improve it

[@jitheshkk8447]

Excellent work and a very nice video. Great !!

[@Randomdude-i8x]

It sounds like a killer robot from the future. Very very impressive that you got this far.

[@ThePrintHouse]

This is great work. I would love to make one for my own channel!

[@lavamelon9478]

Helical cut gears instead of straight cut would allow for smoother movements, same concept in vehicle transmissions

[@mtbmike]

beautiful, keep up the R n D. you could actually sell a "lego" style build kit

[@bronzwikgk1496]

Great job man. Any plans to reduce the jerks in the motion?

[@davepreiss]

Looks like you could make a pretty legit crust cutting robot with this setup 🙃

[@3DprintedLife]

Damn, that cuts deep

[@itayeshkarsaban731]

I love your harmonic drive gear box's, but I can also suggest planetary gears.

[@braunsch4901]

I'd recommend ASA or CF nylon for any gear and motor brackets. Definitely printing this to make an assembly line when you finish the redesign

[@Mbow]

I am a roboticist and the robot you made is really exceptional

[@madeintexas3d442]

This video literally changed my life.

[@tylouww.1915]

very nice, but the problem is you can only reach 5 degrees of Freedom and not all 6, because you have only have 5 joints. You need another twisting joint between your current joint 3 and 4 to be able to reach every position. So the first 3 joints are for translation (xyz positioning) and the joints 4 to 6 are for the orientation of the endeffector. The simplest way to design it, is to model the joints 4,5,6 in a way that their rotation axis meet in one point. That point is then also the point that is translated in space by the first 3 axis. You can search online for like "robot with spherical wrist" for more information. Also a good open source control software is from Chris Annin, he has a yt channel with robots and how to control them :D

[@g_gaming2893]

Absolutely incredible man!

[@paralanx626]

If you don't want to write the IK stuff and GUI yourself, I'd suggest ROS Moveit using the ROS robotics framework. It has a great visualizer and can just do the IK for you if you feed it an arm model

[@zobdog205]

so rad dude. great work

[@Mulakulu]

Welcome back buddy. We've missed you

[@JayarBass]

did you try changing the locking settings for the nema 17? in my grbl cnc, if i set the "Step Idle Delay (msec)" to 200, the stepper motors lock enough to hold that weight... i cant manually push any axis...

[@johanjotun1647]

one of these strong enough to grip a laptop, celing mounted to a lighting fixture would be handy, you could power it and have it in any position W/O blocking the air vents, everyone who had mobility issues would get one.

[@Micreathonn]

I respect the Oversimplified video in the background, I do the same thing!

[@joao_1337]

Thank you so much for all these tutorials bro. So much valuable knowledge

[@wayne8797]

This video is so underrated. It should be a crime!

[@MegaSteamfreak]

Maybe consider placing the stepper drivers as close as possible to the motors for better performance. And also a slip ring for the USB connection or even an optical interface between the base and the arm and the slip ring only for power?

[@evanmccloy5643]

ay bro i was gonna do this for the exact same thing even using the automaticaddison stuff! I'm glad someone else got it working first haha

[@teitgenengineering]

Somehow i didn't see this video i thought you had given up on this im glad i found it

[@TheDarkLord-qg5lr]

I dont know if this would be possible because of wobble but mounting it on the ceiling of a box putting a drill on it and using it as a homebuilt CNC-Machine would be sick

[@3DprintedLife]

Technically “possible” but it would be a really really bad cnc

[@tymoteuszkazubski2755]

it wouldn't be stuff enough.

[@shrimpinpat]

This would be dope for product shots. I need something like this!

[@iambrianlevy]

Beautiful work! I have been following along and have a few of your actuators printed out. I look forward to testing them, soon!

[@emanuell6672]

Amazing! Hoping to see more

[@nhibbs3]

i love it! really impressed by your dedication and open sourcing everything! i just subbed

[@conorstewart2214]

Does your phone have some form of optical image stabilisation? The video from the phone looks a lot less shaky than the video of the arm holding the phone. Another option is something that is used for FPV drones to smooth out the video and it is called “ReelSteady”, I have never used it but I have heard of it being used a lot for freestyle FPV.

[@mothballz99]

Is there space on the left side of the MUB, furthest from the corner where the two connectors are kind of situated, to perhaps add a through-hole for a screw so that it can be secured down to whatever mainboard you're using (with the help of a standoff) to make the connection stronger, avoiding potentially tearing the connectors? Just a thought

[@3DprintedLife]

If there’s room for a through hole connector then I messed up and should have made the whole board smaller haha. The board to board connectors I use are actually quite strong, just not strong enough to handle my very bad decisions.

[@Andreas-gh6is]

I think the best way to join two PLA parts is using a soldering iron tempered to regulated to something like 200-230 degrees. It can be ugly, but the joint can also often be hidden or retouched...

[@boydkane5469]

3:50 earned my subscription :D

[@mounirsaidi5358]

Next goal: making it smooth and redesigning some parts. Once you make it smooth it will be completly done, the cheapest robot arm you should definetly start selling

[@Hclann1]

Great video, this just showed up in my feed. In your opinion, If you wanted the arm to be less noisy, would you use servos, or microstepper controllers?

[@Mailzas]

If you planing to do playback anyway, instead of all the calculations, you should use potentiometers, to read position. then you save 4 numbers. Then you hand move to another position, save 4 numbers again. To move programatically back to xyz position, you move until your potentiometers show same.

[@Eilaithen1]

iPhone did a great job of stabilising that footage pretty impressive

[@osmanpasha_diy]

This looks very cool! However, I want to point out some issues with the design: 1. Weight distribution. Most industrial 6DoF robots put motors in closest positions to base as possible - 1 motor in the base, 2 motors right above the base, 3 at the end of waist (take a look at kuka/abb/fanuc robots). This way moving mass is minimized, so faster acceleration, bigger speed, smaller motors. There are very few commercial robots that instead put motors near their joints (UR5 and KUKA iiwa come to mind). 2. The problem with shakiness is caused by previous problem - large moving mass at the end of robot, the motors just cannot compensate for all that inertia quickly enough (although they can lift it no problem as you've shown), hence the shaking. Redesigning transmission and strengthening the body may help, but the cause of the problem will still be there. 3. Does your code do smooth acceleration? Moving camera for real filming purposes will most definitely require that. That will also reduce shakiness. BTW, I'd suggest you take existing code (grbl-5x or marlin for example), at least for start. That will take away the need to reinvent proper motor control, pulse generation, acceleration etc. And you can concentrate on IK. (Edit: ah, you are using TMCs with integrated controller, so you probably don't need to do all that low-level stuff)

[@Plastic3D]

Super inspirational man!

[@dibblemcbibble2787]

It's jiggling because your PID control needs tuning. Phenomenal build

[@arenspringvloed9020]

Hot Damn. Thats pretty amazing! Well done. :)

[@capoworks4530]

You may want to take a look at simplefoc, it’s a lil bit crazy to convert everything but it would make your arm one of the smoothest around, anyway, impressive job!!

[@roboage1027]

Great job 👍 👌. Such videos inspire me keep up with my own robotic arm project! Thank you!

[@Perceivedshift]

Man, looking forward to this one day being an open-source project with many collaborators making it incredibly easy to build and run. You are treading new ground man!

[@ingmarm8858]

Have you seen Skyentific channel? Smooth 3d printed robotics...

[@dhruvagarwal]

not sure if this will help with shakiness but someone on yt built a similar working bot and used tension band to provide resistance to motion o each joint which required more power from motors but made the motion a lot smoother, hope this helps

[@Thefedmog_Challenge]

Damn that's huge. Great work 👍.

[@oneil9615]

Your work is really awesome

[@matheuscampos8097]

The Iphone image stabilization made some miracles right there

[@foobars3816]

And yet the part of this video that made me smile was the cat winking. Impressive work though. ;-)

[@3DprintedLife]

Thanks! and great 😃

[@weirdsciencetv4999]

You can solve the IK between start and end world coordinates. Then compute world coordinates at the midpoint of the start end end joint coords. Then use forward kinematics to see where it would be in world coordinates. See if the difference between the world coordinate midpoints and actual is above some threshold. If it is, subdivide that segment of the trajectory. Repeat recursively. Then send the series of joint coordinates computed to be linearly interpolated by your microcontroller.