Yeah, then in a few months you will see them on eBay for at least 3k..free shipping from China. This is a fully working and well made project, much respect for that, and also for allowing the whole project open source. Sad though that this kindness and generosity will be stolen and sold commercially as someone else's work. I have nothing against giving this sort of knowledge and work away, some people do. It shows good character and something more people should have. If I ever decide to use this project then a worthy gift in return would be offered.
I'm only 2 minutes into this video and I gotta stop and tell you how impressed I am by this. Amazing. As an industrial programmer myself, this is awesome. I'm gonna watch the rest now.
THIS was the reason I purchased a 3D printer. I'm fiddling for small stepper motors and micro servos at the moment. I printer a my arm and I'm just assembling it using an old PC power supply. Thanks for the video and the time it took! Now I'm excited all over again. :)
I started writing the software probably a year maybe a little longer ago when I did the first robot, it's been a work in progress. this robot took me 3 or 4 months to design picking away at it in the evenings.
I'm 15 and have been searching for 3D printed arms. This is by far the most precise and low cost quality thing I've seen. thanks! I can't wait for the parts list to come out.
Beautiful work on so many aspects. I was a fan of the old Radio Shack "Armatron" back in the day, your project shows how far we've come. The availability of 3D printing, online learning & collaboration, Arduino, etc is really the beginning of a new era.
Hi robert i want to learn this thing i use Arduino for simple project but this its loking hard to mee programing by paython with Arduino chipset if there is a Chaîne or vidéos on TH-cam to explain the relation with paython and thanx bro
Absolutely awesome! I've built a couple robot arms in past years, with 3D printers, neither anywhere as good as this one. So I totally appreciate and salute this accomplishment. Hats off to Chris Annin!
Chris, every once in a while someone comes along to give you an ooh-ahh day. Congratulations, you're it for this week! Looking forward to making my own 3 axis bed CNC and cutting a robotic arm to go on the cross slide to replace the single axis up /down cutter head movement. Put a collet chuck where your grab arm is and with router bits / drills / milling cutters we have an inexpensive 3d subtractive workstation. Put a 90 degree gear on the end and with a spindle in, place a saw blade and you have a grooving station for sheet goods. Put a grinding wheel on the shaft and you have a surface grinder. Wow, a whole workshop in one tool. Can't wait!
Yea I wonder sometimes... I thought it was kind of a cool project to share. I suppose people run across it who aren't into stuff like this but you would think they would just keep scrolling rather than go to the dark side.... trolls? Nothing better to do? Got me.
The video of the new aluminum version is here th-cam.com/video/AeCLbhPHltw/w-d-xo.html - You can find all the software, instructions and print files here: github.com/Chris-Annin/AR2
Hello chris. I download the files and i just started to review. IS there any possibility to build it aluminium instead of power abs. What you think about the stepper power is enough?
We have the first half of the robot built out of aluminum - there are some minor differneces using aluminum so we are working out the bugs on the part drawings. If it all works and testing goes well I may make an aluminum parts kit available. It will take some time but im working on it.
Amazing work! The dedication to see a project like this from concept to actually making it work is incredible. Not to mention producing a Bill of Materials and an Assembly Manual and sharing all this for free. Not to mention all the skill sets needed, mechanical engineering, electrical engineering, computer programming, CAD modeling, 3D printing, then playing mechanic, electrician, robotics, to assemble the components. I hope you get paid very well for whatever you do for living because not many folks are capable of this. - Thanks for sharing Chris
holy shit. your giving all this away? your a good person and embody the spirit of the internet. I'm just nerding out on robotics videos with only a yearning of actually doing it. but thank you.
Just found myself watching this again for the nth time. Regarding your comments here about moving the kinematic calculation into the arduino, that sounds like quite a hassle. Isn't the arduino just there to relay step/direction info to the motors (and listen to the limit switches of course)? The kinematic calculation strikes me as being perfectly suited to a computer with some proper processing power, and more easily modifiable setup and tweaking and potentially also graphical visualization, especially if you intend to release this for others to use on their own builds.
I believe you are correct on all accounts - I think its current form is the simplest and has a lot of value for others to use and modify which of course was my goal to keep it as simple as possible - also considering this was a learning process for me as well so it started simple and I built it to a point where it is now. There are probably multiple ways to do this but my thinking is that if I wanted to get into complex path interpolation like perfectly straight lines, arcs or other 3d generated paths I would need the arduino to execute a series of positions along the arc or line seamlessly with zero delay between moves - it would be needing to read ahead and be headed to the next point as it finishes the last with no delay. Currently the computer sends serial commands to the arduino and there is a limit as to the string length that can be sent. Im sending one move at a time and there is a delay between serial commands. unless there is a way to establish real time communication between the PC and the arduino you would end up with a choppy line or arc movement. as it is now the PC sends 6 joint positions to the arduino - I think I need to be able to send Cartesian position to the arduino so that it could then do the kinamatics and figure out joint angles and calculate multiple points along the line or arc path from its current position to its destination position so that it could execute many 10's or 100's of positions along that line or arc seamlessly with no delay. Im not certain but I think being able to send the arduino coordinate positions instead of joint angles would open alot of doors for CNC, printing or other types of applications. but as it is now its simple, easy to use and it fulfills its intended purpose of a simple material handling robot that can move from point A to point B - Its also good for education and hobby purposes. I dont think its best suited for material application or CNC applications but maybe thats something I can work on in the next version - perhaps a servo driven version. Let me know what you think or if you have any ideas. Im learning as I go here.....
My main concern was that the robot definition (number and orientation of axes, limb segment lengths etc) would be on the arduino, which would be a nuisance to have to keep flashing a new sketch to set up and tweak. You could perhaps store the definition in EEPROM and update it from the PC instead of re-flashing the sketch every time, but that would have its own complications. Interpolations will produce quite a lot of info to deal with and developing them with the primitive debugger-less feedback from the arduino doesn't sound like fun. I also suspect that the kind of memory needed to store a decent straight line interpolation would be beyond the spec of the arduino - sure you could limit it to whatever movement distance it can actually handle, but this is just another hassle that the PC would not have. Other than that, from an open-sourcing point of view it would be advantageous to break down the calculation stages into more discrete 'modules' that flow from higher to lower levels of control. For example, somebody might like to use their own path interpolation algorithm, which would be a lot easier if there was a clearly defined communication interface for their stage to issue commands to the next lower level (ie. which motors move where). Taking it a step further in the other direction, you might have an even higher-level stage (eg. computer vision to identify parts on a table) that would give commands to the path interpolator. I think the easier each stage can be swapped out or worked on without knowledge of other stages, the more takers you'll have for collaboration. It sounds like your main concern at this point is the speed of communication between the PC and the arduino? I wasn't suggesting that the PC would issue step pulses directly, just in case you got that impression. The arduino would still be given the desired joint positions and be required to figure out the necessary steps on its own, but it would be given those many 10's or 100's of interpolated positions you mentioned. This I think is an achievable data rate, especially if you make the communication interface a little more efficient. I looked over your sketch and I don't see any obvious reason it would be particularly slow, especially at faster baud rates. But using binary data instead of strings would help to decrease the data volume and time spent parsing strings. To have seamless movement with no delay you'll want to buffer some points on the arduino, which means you'll need to have the arduino acknowledge completed points so that the PC knows when to send more. I guess it would be helpful if acknowledgments like that worked their way back to the other higher level stages too, so that none would get too far ahead of the others, although buffering is less of a problem on the PC side. In any case, I'm pretty sure the PC to arduino connection would be able to send positions faster than the arm can physically carry out the movement. As a rough estimate, a 'packet' with 6 joint positions of 16 bit integers each, plus a few bytes for desired speed, some flags, packet ids, and other housekeeping might come to around 140 bytes. A 115200 baud hardware UART should transfer about 100 such packets per second. So the resolution of the interpolation for that case would be the distance moved per second divided by 100, eg. moving 20cm per second would be done with a resolution of around 2mm. I'm starting to get together some parts to build a little robot arm myself. I have plenty of programming experience but only just started with steppers and actuation recently when I built a little CNC machine. The inverse kinematic calculation stuff gives me a headache just to look at, so I am thinking to fudge it with a 2D physics library for games (eg. Angry Birds, Hill Climb Racing). It supports joints with limits and would allow setting the position of the two arm ends and simply reading off the angles of the joints. That would handle the main arm joints in their 2D plane, for a typical arm the other degree of freedom is cylindrical rotation which should not be too hard. Since this library is normally used in games with dozens of joints and constraints and typically running at 60 frames per second I have no doubt it would be able to keep up speed-wise. I'm not sure how the accuracy of this method would be though.
oops, just noticed my packet size calculation was using 16 byte integers, not 16 bit, I was thinking that number came out a little too large heheh. So a typical packet (with 32 bit integers this time) would be more like 56 bytes or 250 packets per second, ie. about 1mm resolution at 25cm/s
Very exciting stuff! I can't wait to see more. What are you planing to do with the mechanical design and software? A side from releasing it on get hub, which is a very nice thing for you to do. But, do you have an application you are working toward? Welding bot or machine tending?? Best Matt
thank you. I would say the project has a couple goals - primarily just something Ive wanted to do for a while, I work with industrial robot systems for a living - the discounted integrator pricing for most of the smaller robots is $25K up to $100K for larger ones (the price is higher if your the end user). Even the small lab grade servo robots are around $15k which is out of reach for most. Ive always thought it would be cool if there was something affordable that the small business or garage shop could have for small, low complexity machine tending applications. I was hoping to come up with something people could make themselves. This robot is not a finshed - polished product but its a start or at least a feasible concept.... and its a fun project either way. I was hoping to come up with something that would only cost someone a few thousand dollars to make - at this point im at $1080 in materials for the robot in the video. I have a colleague who owns a machine shop and we are working on an aluminum version, hoping to put it to work in his shop doing machine tending. Im hoping we can continue to refine the quality and programming and perhaps at some point have a polished build your own robot package that is useful. we'll see how it goes - right now im swamped at work and trying to find the time to get all the files finished up so I can post it and there are still a few bugs im working on. future steps on this project might be closed loop steppers or servos, moving the path calculations from the PC into the arduino and develop path planning (arcs, lines ect), tool frame kinamatics and a few other offset functions. thanks for the feedback. I'll get it posted as soon as I can find some time away from work.
I agree. You have a wonderful dream to chase. And a noble one. Robots are expensive and very useful if that issue is solved it would change a lot of lives. I will continue to follow your progress and I might extend some of my manufacturing expertise in casting to aid you with your project if you wish. I will be in touch. Best Matt
Amazing Work!! I will definitely be following your progress & hopefully building one for myself, once the details are released. Thanks for sharing your work. i believe this is the best 6 axis 3D printed robot arm I have seen on TH-cam.
Figuring out the kinamatics have been the most difficult part for me. When I made the video I had the robot jogging in XYZ and reorienting the wrist ok with a few minor issues but I found later when I had J4 and J6 at certain positions it would make unexpected changes in orientation which is no good. Long story short im still working on figuring out where I screwed up. Id like to have it figured out before I post all the files but the program will always be a work in progress. hopefully something thats always improving and adding functionality. I think making a video on the kinamatics is a good idea - all the information ive gone through so far to get where im at has been difficult to follow given that its always written from the perspective of someone who already understands linear algebra and frame transformations and most of it is geared toward a variety of types of robots and not alot of examples of this configuration. Once I get it finished up I will make a video for the rest of us thats easy to understand. good idea thank you.
That is called a singularity, you cant move lineary with the 4, 5 and 6 axis aligned. You should move with a axis movement instead to avoid that. I work with industrial robots and they have the same problem.
Approximate cost, at the time he compiled the materials list is 935 USD (plus some screws and nuts). Cheapest printer that would be suitable for this is roughly 500 USD, plus the plastic thread, soldering kit, computer and multimeter.
Awesome Project This by far is the most detailed description of how too, I've ever seen. You Sir! are amazing! Thank you for not only putting this together, but also sharing this with everyone! OUTSTANDING!.. CONGRATS!...
What an exceptional project. I instantly subscribed! Thankyou for posting this. Now I'm on the edge of my seat waiting for news of any progress! Absolutely a brilliant effort! Well done.
Wow, that is awesome dude. I'm so pleasantly pleased that there are people like you out there. Smart people(s); Gets my faith in humanity back. Nicely done!
Thanks for the open source Genious. I really appreciate it very much ... and when i say that I believe that i am speaking in behalf of all those have watched your project. Godspeed in your journey Professor. Saluts from Algeria
Hi Chris, Thanks for your great videos. Thanks to you I used same stepper motors for my project and they worked just great. (videos of the build on my channel) I keep following your channel; the progress you have made is just amazing !
awesome work! I've been playing the software a bit. I've got it to run on a Raspberry Pi and it seems to do just fine. Had to workshop the GUI a bit to get it to play nice, but looks good now and it ticks like a clock! I think an "While...if...break" generator based on Inputs would be an amazing improvement. Can't wait to see the kinematics addition to this! You are the man!
I've got J1 completed, working on J2 and controls. There is an error in the photo of the stepper drivers connections to the arduino. Two photos, one shows drivers reversed! There's also an M8 screw not listed in parts, but is in manual. Don't know if its actually needed yet. But awesome project! Next version will be aluminum!
Hey Chris, First off, not complaining, love the project and that you give it out to people like myself that can't do this by ourselves! What I'm looking at is the Electrical Enclosure instructions. On page 12 it shows the Stepper Drivers mounted from top to bottom as J4, J1, J2, J3, J5, J6 and a Track if needed. On page 20, it shows the Stepper Drivers from top to bottom listed as J4, J6, J5, J3, J2, J1, and then the Track Stepper Driver if needed. I thought the enclosure may have been flipped upside down, but based on the Arduino and it starts as J4, then either J1 or J6, I assumed the first picture on page 12 is correct. What really messed me up is the ribbon wires start AND end with a black and white wire! I also looked at the wiring diagram which helped. I'm also thinking of printing out a modified base plate that has the additional supports already part of the print, then just glue them to the J1 Turret Housing after it's assembled. Again, THANK YOU for getting this together. If there's any way I can help let me know.
Its very nice! But we can see that robot is not very precise because bigest problem is that wobbling due to oscilation of movements - how to fix it? I believe that all plastic printed parts must be replaced with metal ones then wobbling could be reduced till minimum and increased precission same time.
Chris, mine is done to the point of J5! J6 is giving me wiring problems, but other than that it's working great! Still waiting on the 3mm rods, tomorrow maybe.
Yeah, man. Impressive!! I like the wiring on the control board - there's some system going on there and the Python Desktop App's also interesting. Well done, keep it up. I'm thinking about getting one of those Braccio Robotic Arms to play with first up. Thanks for the clip.
What is the reason to build it in Aluminium? Or what do you expect to improve with aluminum? Thanks for your answer - and thanks for sharing it with the world!
Mostly rigidity. the printed version is pretty good but the aluminum version doesn't have some of the vibration or bounce. Also strength - I'm afraid to fully test lifting capacity for fear of breaking the J5 carrier or post. My hope is that people can build whichever model suits their needs. obviously printing it will be lower cost - part of my goal is to create the lowest cost robot people can build at home so I hope people find this project useful. Thank you.
thank you. I started out with 1 stepper motor on a table with and arduino and google in front of me. its just a matter of starting small, work on one improvement after another, dont get overwhelmed just focus on learning one thing at a time.
Have you tried to benchmark how much weight it's capable of holding? Apologies if you did mention it in a video, sometimes I'm watching at work without sound! Do you still plan to machine the parts in aluminum?
Ive tested the plastic version with 25oz but havnt fully tested for fear of breaking the J5 carrier. with the economy series gearboxes the calculation comes out at just over 3lb limit. On the aluminum version I have tested it and it can easily carry 3lbs. Before I put out a video and release updated software and instruction I want to switch the aluminum version to a couple of the precision series gear drives from stepper online. with these I should be able to get closer to 5lbs but I will likely keep its rated payload at 3 lbs for an aluminum version. and 1.5lbs for a plastic version.
Thanks, that's very cool. I'm a second year Computer Science student at my local university currently and I'm probably going to make one of these for use in my startup company! Our goal is automation in 3D Prinitng so I'm interested to see what we can accomplish with it! The 3D Print club is also interested in building some so I look forward to helping them as well. Thanks again!
really nice. look forward to the aluminum version. maybe you could consider some velocity valves to the gripper so they don't slam open or close. Mahalo!
Excellent design and development work. Thank you for taking the time to spell everything out for us. I am not clear on how the part numbers in your BOM change from regular to precision. Can you please give the part numbers for all six precision motors in the aluminum version? Also I see there are differences in the motor BOM between the plastic and aluminum versions. I assume that is due to the extra weight of the aluminum parts. Will the precision motors for the aluminum version fit into the plastic parts? If so, I can make the plastic one first and then go to the metal without buying more motors. Thank you in advance for your help. P.S. Please let me know when the kickstarter goes online.
Hello, There were some changes from the original design on the first 3-D printed robot that I build to the current version. The current version uses precision series motors for the first three joints and the same motors for joints forfive and six. The biggest difference in the parts themselves is that the spindle border for joint two is 14 mm and the motor mount for joints 12 and three are different. If you look at the bill of materials and I think the manual for the electrical enclosure and the manual for the wiring harness all the part numbers should be listed. The folder with the print files includes models to build a robot either way.I am currently right in the middle of revising all the manuals and showing pictures of the robot being built with aluminum parts the new manuals should clear everything up. The folder with all the print file includes models for both iterations using the Percision series motors and the economy series motors.I should have all the manuals ready and posted within three weeks I think. And I will be launching the kick starter shortly after that. Please feel free to email me if you have any questions and I will try to answer them.
I am curious to see how you are controlling 6 steppers with the Arduino. I was planning to try to build a 4 axis arm using an old CNC shield I have. But now that I've seen yours, I am jaleous. Hats off Sir. That is some nice work.
I am highly tempted to try and build this. Currently pricing it all up, wondering if some nice chinesium bearings might bring the cost down a bit. The steppers dont look any cheaper on aliexpress than the listed stepperonline site which is interesting.
Nice. Why does it shake? Is it something that can be improved with s/w ? Maybe it stops too suddenly. Maybe s/w could make it decelerated gradually to a stop?
I have acceleration and deceleration controls programmed and It works really well. The shaking is largely due to the plastic material, some of it is due to backlash in the belt tension given plastic parts, some of it is due to backlash in the gear drives, and some of it is due to the shaky plastic table I had it on. I'm currently working on an aluminum version which will be much better. But I'm not too disappointed with this one given it's a 3d printed low-cost robot.
It's to much wiggle room so it's inaccurate in some areas so the aluminium one will be more robust and accurate, i also recommend to gets some higher precision and higher torque motors so your able to lift more will more accuracy, but it's an awesome homemade project! Nice work indeed!
Sei que vou ser mais um a perguntar... As partes desse braço foi feito em impressora 3d ? Vc teria os stls pra compartilhar, assim com o projeto em si ?
Hi Chris,Very comprehensive description of your project.I'm starting to get the parts together (I'm based in the UK). At the moment I can only find a 23H22-2804S-PG47 (rather than the 23HS30-2804S-PG47). Is this a problem? I think that the difference is operating voltage but as it is the most expensive motor/gb I want to get it right. Also is it possible to use PLA rather than ABS for the plastic parts?You set the bar really high with the quality of your instructions, explanations and downloads. This gave me the confidence to embark on this project. Thanks again Keith
Either motor should work fine, the one you are looking at is a shorter body motor at 56 mm and I believe it's 1.25 N meters of holding torque. The one I used has a longer body motor at 76 mm and it has 1.9 N meters of holding torque but both are 2.8 A and use the same gearbox so it should work fine. I selected the larger motor just so that I had more torque but I don't know that you really need it with the printed version. The aluminum version with the larger motor can handle 3lbs. But I'm working on upgrading to stepper online's new precision series motors for J1, J2 and J3 on the aluminum version robot before I release the next video and software. The plastic version will stay the same using the economy series motors. The aluminum version will have minor differences at least on a few of the motors but the software will be universal and only require a couple settings be changed.
When you get a chance... please update the description with the link for the plans, etc. I would like to build this robot... perhaps even do a video series on setting up and running a 6-DOF robot from LinuxCNC. Thanks! (Subscribed)
yes, absolutely will. I have a couple more things to finish up and I'll post all the files and I was planning on a setup or how to video. im not as far along as I thought I would be but i'll get it all posted as soon as I can.
Chris Annin wonderful. don't sweat it... I also have a TH-cam channel... I'm backed up at least 5 videos. sometimes life gets in the way. I'll probably be ready to start the printing etc. around September at the earliest. Thanks!
Do you consider to go open source and cooperate , for example, with freecad and Blender, and Blender cam that is for cnc , and make an opensource kit platform , that people can buy and add ideas , and 3d priting upgrades, also more hackerspaces has got lathe and metal machinery... so you can build a disen community even pick and place or use it on bigger cube type 3d printer
Dear respected sir, Hope you are in good health with success and happiness. I am very happy for your awesome work. I'm alos working on robot arm. I'm totally begainer. Here I have seen you are using belt and pinion. But could run the arises by Direct motors instead of using belt pulleys? That's means, could I make Direct motor articulated robot arm?
I'm doing it! The documentation is awesome! I've ordered parts and printed the base and J1 parts, had to split the base! And I'm terribly lazy, wish there had been an aluminum kit for under $3k, but what the heck! Thanks!
You're doing tinkerers, hobbyists, and students such an amazing service here. We are all INCREDIBLY thankful. Dont forget how incredible this is.
MrGuyCali thank you, I’m happy to help out.
This is top drawer. I admire your mindset especially that you are going totally open source for users. well done.
thank you.
Holy cow you set the bar high!
SUBSCRIBED!!!!!
Chris Annin No thank you! Great work and thanks for sharing!
Yeah, then in a few months you will see them on eBay for at least 3k..free shipping from China.
This is a fully working and well made project, much respect for that, and also for allowing the whole project open source. Sad though that this kindness and generosity will be stolen and sold commercially as someone else's work.
I have nothing against giving this sort of knowledge and work away, some people do. It shows good character and something more people should have.
If I ever decide to use this project then a worthy gift in return would be offered.
I'm only 2 minutes into this video and I gotta stop and tell you how impressed I am by this. Amazing. As an industrial programmer myself, this is awesome. I'm gonna watch the rest now.
Was you classed as a keyworker during the quarantines?
THIS was the reason I purchased a 3D printer. I'm fiddling for small stepper motors and micro servos at the moment. I printer a my arm and I'm just assembling it using an old PC power supply. Thanks for the video and the time it took! Now I'm excited all over again. :)
hi garry i am strugling with couple of parts.....can you help me ??
This is the best 3d printed 6 axis robot arm I've seen on youtube, mad props. Subbed!
thank you
how long did it take you to design the robot and write software?
I started writing the software probably a year maybe a little longer ago when I did the first robot, it's been a work in progress. this robot took me 3 or 4 months to design picking away at it in the evenings.
Hats down.... incredible! This is simply awesome!
Thank you
I'm 15 and have been searching for 3D printed arms. This is by far the most precise and low cost quality thing I've seen. thanks! I can't wait for the parts list to come out.
Thank you. i will get it posted - probably 2 -3 more weeks depending how things are going at work.
Chris Annin awesome
Beautiful work on so many aspects. I was a fan of the old Radio Shack "Armatron" back in the day, your project shows how far we've come. The availability of 3D printing, online learning & collaboration, Arduino, etc is really the beginning of a new era.
Thank you
Wow - this is impressive - I've spent 15 years programming Fanuc Robots and to see what you've built is amazing. You are quite the engineer 👍🏻
Thank you very much
Hi robert i want to learn this thing i use Arduino for simple project but this its loking hard to mee programing by paython with Arduino chipset if there is a Chaîne or vidéos on TH-cam to explain the relation with paython and thanx bro
Absolutely awesome! I've built a couple robot arms in past years, with 3D printers, neither anywhere as good as this one. So I totally appreciate and salute this accomplishment. Hats off to Chris Annin!
Chris, every once in a while someone comes along to give you an ooh-ahh day. Congratulations, you're it for this week! Looking forward to making my own 3 axis bed CNC and cutting a robotic arm to go on the cross slide to replace the single axis up /down cutter head movement. Put a collet chuck where your grab arm is and with router bits / drills / milling cutters we have an inexpensive 3d subtractive workstation. Put a 90 degree gear on the end and with a spindle in, place a saw blade and you have a grooving station for sheet goods. Put a grinding wheel on the shaft and you have a surface grinder. Wow, a whole workshop in one tool. Can't wait!
With all the work put into the arm, the video, the links etc.
How can so many people have disliked this video?
Yea I wonder sometimes... I thought it was kind of a cool project to share. I suppose people run across it who aren't into stuff like this but you would think they would just keep scrolling rather than go to the dark side.... trolls? Nothing better to do? Got me.
This is an excellent project Chris, thanks for the effort, appreciate it.
Amazing to have people like you thank you so much for sharing your knowledge. When I finish my robot I will post a video with all credits to you.
Thank you.
The video of the new aluminum version is here th-cam.com/video/AeCLbhPHltw/w-d-xo.html - You can find all the software, instructions and print files here: github.com/Chris-Annin/AR2
Hello chris. I download the files and i just started to review. IS there any possibility to build it aluminium instead of power abs. What you think about the stepper power is enough?
im working on an aluminum version now. I'll post videos when its up and running.
can you publsih 3d model or 2d drawings for machining the aluminum.
We have the first half of the robot built out of aluminum - there are some minor differneces using aluminum so we are working out the bugs on the part drawings. If it all works and testing goes well I may make an aluminum parts kit available. It will take some time but im working on it.
you are awesome, thanks for posting these!
Amazing work! The dedication to see a project like this from concept to actually making it work is incredible. Not to mention producing a Bill of Materials and an Assembly Manual and sharing all this for free. Not to mention all the skill sets needed, mechanical engineering, electrical engineering, computer programming, CAD modeling, 3D printing, then playing mechanic, electrician, robotics, to assemble the components. I hope you get paid very well for whatever you do for living because not many folks are capable of this. - Thanks for sharing Chris
Thank you very much I appreciate your compliment and I hope this project becomes more and more useful in the future.
Awesome job! Also that is perhaps the best demo on capabilities for a diy job i have seen yet. keep up the great work!!!
thank you.
holy shit. your giving all this away? your a good person and embody the spirit of the internet. I'm just nerding out on robotics videos with only a yearning of actually doing it. but thank you.
It's more fun if it's a group effort, hopefully it's useful or turns into something people can have fun with. Thank you.
Very nice. Can't wait to see the aluminium one!
thank you
Just found myself watching this again for the nth time. Regarding your comments here about moving the kinematic calculation into the arduino, that sounds like quite a hassle. Isn't the arduino just there to relay step/direction info to the motors (and listen to the limit switches of course)? The kinematic calculation strikes me as being perfectly suited to a computer with some proper processing power, and more easily modifiable setup and tweaking and potentially also graphical visualization, especially if you intend to release this for others to use on their own builds.
I believe you are correct on all accounts - I think its current form is the simplest and has a lot of value for others to use and modify which of course was my goal to keep it as simple as possible - also considering this was a learning process for me as well so it started simple and I built it to a point where it is now. There are probably multiple ways to do this but my thinking is that if I wanted to get into complex path interpolation like perfectly straight lines, arcs or other 3d generated paths I would need the arduino to execute a series of positions along the arc or line seamlessly with zero delay between moves - it would be needing to read ahead and be headed to the next point as it finishes the last with no delay. Currently the computer sends serial commands to the arduino and there is a limit as to the string length that can be sent. Im sending one move at a time and there is a delay between serial commands. unless there is a way to establish real time communication between the PC and the arduino you would end up with a choppy line or arc movement. as it is now the PC sends 6 joint positions to the arduino - I think I need to be able to send Cartesian position to the arduino so that it could then do the kinamatics and figure out joint angles and calculate multiple points along the line or arc path from its current position to its destination position so that it could execute many 10's or 100's of positions along that line or arc seamlessly with no delay. Im not certain but I think being able to send the arduino coordinate positions instead of joint angles would open alot of doors for CNC, printing or other types of applications. but as it is now its simple, easy to use and it fulfills its intended purpose of a simple material handling robot that can move from point A to point B - Its also good for education and hobby purposes. I dont think its best suited for material application or CNC applications but maybe thats something I can work on in the next version - perhaps a servo driven version. Let me know what you think or if you have any ideas. Im learning as I go here.....
My main concern was that the robot definition (number and orientation of axes, limb segment lengths etc) would be on the arduino, which would be a nuisance to have to keep flashing a new sketch to set up and tweak. You could perhaps store the definition in EEPROM and update it from the PC instead of re-flashing the sketch every time, but that would have its own complications. Interpolations will produce quite a lot of info to deal with and developing them with the primitive debugger-less feedback from the arduino doesn't sound like fun. I also suspect that the kind of memory needed to store a decent straight line interpolation would be beyond the spec of the arduino - sure you could limit it to whatever movement distance it can actually handle, but this is just another hassle that the PC would not have. Other than that, from an open-sourcing point of view it would be advantageous to break down the calculation stages into more discrete 'modules' that flow from higher to lower levels of control. For example, somebody might like to use their own path interpolation algorithm, which would be a lot easier if there was a clearly defined communication interface for their stage to issue commands to the next lower level (ie. which motors move where). Taking it a step further in the other direction, you might have an even higher-level stage (eg. computer vision to identify parts on a table) that would give commands to the path interpolator. I think the easier each stage can be swapped out or worked on without knowledge of other stages, the more takers you'll have for collaboration.
It sounds like your main concern at this point is the speed of communication between the PC and the arduino? I wasn't suggesting that the PC would issue step pulses directly, just in case you got that impression. The arduino would still be given the desired joint positions and be required to figure out the necessary steps on its own, but it would be given those many 10's or 100's of interpolated positions you mentioned. This I think is an achievable data rate, especially if you make the communication interface a little more efficient. I looked over your sketch and I don't see any obvious reason it would be particularly slow, especially at faster baud rates. But using binary data instead of strings would help to decrease the data volume and time spent parsing strings. To have seamless movement with no delay you'll want to buffer some points on the arduino, which means you'll need to have the arduino acknowledge completed points so that the PC knows when to send more. I guess it would be helpful if acknowledgments like that worked their way back to the other higher level stages too, so that none would get too far ahead of the others, although buffering is less of a problem on the PC side. In any case, I'm pretty sure the PC to arduino connection would be able to send positions faster than the arm can physically carry out the movement. As a rough estimate, a 'packet' with 6 joint positions of 16 bit integers each, plus a few bytes for desired speed, some flags, packet ids, and other housekeeping might come to around 140 bytes. A 115200 baud hardware UART should transfer about 100 such packets per second. So the resolution of the interpolation for that case would be the distance moved per second divided by 100, eg. moving 20cm per second would be done with a resolution of around 2mm.
I'm starting to get together some parts to build a little robot arm myself. I have plenty of programming experience but only just started with steppers and actuation recently when I built a little CNC machine. The inverse kinematic calculation stuff gives me a headache just to look at, so I am thinking to fudge it with a 2D physics library for games (eg. Angry Birds, Hill Climb Racing). It supports joints with limits and would allow setting the position of the two arm ends and simply reading off the angles of the joints. That would handle the main arm joints in their 2D plane, for a typical arm the other degree of freedom is cylindrical rotation which should not be too hard. Since this library is normally used in games with dozens of joints and constraints and typically running at 60 frames per second I have no doubt it would be able to keep up speed-wise. I'm not sure how the accuracy of this method would be though.
oops, just noticed my packet size calculation was using 16 byte integers, not 16 bit, I was thinking that number came out a little too large heheh. So a typical packet (with 32 bit integers this time) would be more like 56 bytes or 250 packets per second, ie. about 1mm resolution at 25cm/s
A lot of work went into this. Your attention to detail shows. VERY Good job! VERY impressive!
Thank you
Oh man! This is awesome! Just came across this. I want to build one👍
That is a hard work, good documentation and open source. Hats off to you.
Thank you
Fantastic work! Thanks for sharing.
Wow, just checked the assembly manual illustrated with the aluminium version. This is an awesomely professional job! Congrats!
Thank you
Very exciting stuff! I can't wait to see more. What are you planing to do with the mechanical design and software? A side from releasing it on get hub, which is a very nice thing for you to do. But, do you have an application you are working toward? Welding bot or machine tending??
Best Matt
thank you. I would say the project has a couple goals - primarily just something Ive wanted to do for a while, I work with industrial robot systems for a living - the discounted integrator pricing for most of the smaller robots is $25K up to $100K for larger ones (the price is higher if your the end user). Even the small lab grade servo robots are around $15k which is out of reach for most. Ive always thought it would be cool if there was something affordable that the small business or garage shop could have for small, low complexity machine tending applications. I was hoping to come up with something people could make themselves. This robot is not a finshed - polished product but its a start or at least a feasible concept.... and its a fun project either way. I was hoping to come up with something that would only cost someone a few thousand dollars to make - at this point im at $1080 in materials for the robot in the video. I have a colleague who owns a machine shop and we are working on an aluminum version, hoping to put it to work in his shop doing machine tending. Im hoping we can continue to refine the quality and programming and perhaps at some point have a polished build your own robot package that is useful. we'll see how it goes - right now im swamped at work and trying to find the time to get all the files finished up so I can post it and there are still a few bugs im working on. future steps on this project might be closed loop steppers or servos, moving the path calculations from the PC into the arduino and develop path planning (arcs, lines ect), tool frame kinamatics and a few other offset functions. thanks for the feedback. I'll get it posted as soon as I can find some time away from work.
I agree. You have a wonderful dream to chase. And a noble one. Robots are expensive and very useful if that issue is solved it would change a lot of lives. I will continue to follow your progress and I might extend some of my manufacturing expertise in casting to aid you with your project if you wish. I will be in touch.
Best Matt
Seriously WTF! Are you for real!?!?!? Blown away, and truly amazed!
thank you very much
Wow, Great job, very impressive!!
Thank you
Best 3d printed roboter... It must have taken a lot of work and time... Respect...
Thank you.
I love it's industrial sound
TheEnragedGamer the sweet sound of stepper ramping up and down
Amazing Work!! I will definitely be following your progress & hopefully building one for myself, once the details are released. Thanks for sharing your work. i believe this is the best 6 axis 3D printed robot arm I have seen on TH-cam.
thank you. just posted all the files a couple mins ago. link in the description.
Very nice! Could you do a video on the kinematics? The joint to cartesian transformations must be hard!
Figuring out the kinamatics have been the most difficult part for me. When I made the video I had the robot jogging in XYZ and reorienting the wrist ok with a few minor issues but I found later when I had J4 and J6 at certain positions it would make unexpected changes in orientation which is no good. Long story short im still working on figuring out where I screwed up. Id like to have it figured out before I post all the files but the program will always be a work in progress. hopefully something thats always improving and adding functionality. I think making a video on the kinamatics is a good idea - all the information ive gone through so far to get where im at has been difficult to follow given that its always written from the perspective of someone who already understands linear algebra and frame transformations and most of it is geared toward a variety of types of robots and not alot of examples of this configuration. Once I get it finished up I will make a video for the rest of us thats easy to understand. good idea thank you.
Wonderful! Good luck with finding the bugs, I'm looking forward to your next videos!
That is called a singularity, you cant move lineary with the 4, 5 and 6 axis aligned. You should move with a axis movement instead to avoid that. I work with industrial robots and they have the same problem.
Just wanted to say congratulations on such a tremendous job! Very exciting work!
thank you
I've always found robotics interesting. I def would like to build and experiment with something at home. Thanks for your videos.
Approximate cost, at the time he compiled the materials list is 935 USD (plus some screws and nuts). Cheapest printer that would be suitable for this is roughly 500 USD, plus the plastic thread, soldering kit, computer and multimeter.
just wow! the amount of work into this is amazing! kudos for the great work!!
Thank you
very cool. Nice work
thank you
Andreas your robot also good
Very impressive. As a coder the control code is a decent project on it's own. With the hardware as well, it's awesome.
Thank you
It is possible to obtain the stl parts of the robot, electronics and software.
Yes I will be posting all the files in a couple weeks.
Yea just make some sand molds from the plastic ones and cast some aluminium
Beautiful, well designed & engineered Robot arm! Thankyou Chris!
Glad I can share it. Thanks for the compliment
But the real question is, can it choke the chicken?
Hi Chris,Thanks for the rapid response. I've now managed to source the higher torque motor.First parts arrived today! Thanks again, Keith
Amazing 🤘 I'd attach a small spindle on it and turn it into a Robotic Milling Machine.😉
Awesome Project
This by far is the most detailed description of how too, I've ever seen.
You Sir! are amazing!
Thank you for not only putting this together, but also sharing this with everyone!
OUTSTANDING!..
CONGRATS!...
OMG. this is the best 3d printed thing I have ever seen.
Dude, nicely done! This thing is badass.
What an exceptional project. I instantly subscribed! Thankyou for posting this. Now I'm on the edge of my seat waiting for news of any progress! Absolutely a brilliant effort! Well done.
Wow, that is awesome dude. I'm so pleasantly pleased that there are people like you out there. Smart people(s); Gets my faith in humanity back. Nicely done!
Thank you very much.
Thanks for the open source Genious. I really appreciate it very much ... and when i say that I believe that i am speaking in behalf of all those have watched your project. Godspeed in your journey Professor. Saluts from Algeria
CHAHROUR Abdenour thank you very much
Hi Chris,
Thanks for your great videos.
Thanks to you I used same stepper motors for my project and they worked just great. (videos of the build on my channel)
I keep following your channel; the progress you have made is just amazing !
thank you
awesome work! I've been playing the software a bit. I've got it to run on a Raspberry Pi and it seems to do just fine. Had to workshop the GUI a bit to get it to play nice, but looks good now and it ticks like a clock!
I think an "While...if...break" generator based on Inputs would be an amazing improvement.
Can't wait to see the kinematics addition to this! You are the man!
thank you
Great design, color, very noble from you to share everthing with us, ... awesome ! Big thank you!!!
I've got J1 completed, working on J2 and controls. There is an error in the photo of the stepper drivers connections to the arduino. Two photos, one shows drivers reversed! There's also an M8 screw not listed in parts, but is in manual. Don't know if its actually needed yet. But awesome project! Next version will be aluminum!
Glad you like it. Can you tell me what page showing the drivers you are referring to. Please email me Chris.annin@gmail.com
Hey Chris,
First off, not complaining, love the project and that you give it out to people like myself that can't do this by ourselves!
What I'm looking at is the Electrical Enclosure instructions. On page 12 it shows the Stepper Drivers mounted from top to bottom as J4, J1, J2, J3, J5, J6 and a Track if needed.
On page 20, it shows the Stepper Drivers from top to bottom listed as J4, J6, J5, J3, J2, J1, and then the Track Stepper Driver if needed.
I thought the enclosure may have been flipped upside down, but based on the Arduino and it starts as J4, then either J1 or J6, I assumed the first picture on page 12 is correct. What really messed me up is the ribbon wires start AND end with a black and white wire!
I also looked at the wiring diagram which helped.
I'm also thinking of printing out a modified base plate that has the additional supports already part of the print, then just glue them to the J1 Turret Housing after it's assembled.
Again, THANK YOU for getting this together. If there's any way I can help let me know.
Brian Mathis thank you for the detailed feedback, I’ll look at the manual when I get home tonight and see what I can do to clear that up.
Hi Chris, found the stepper controller settings under the electrical enclosure manual. I must look harder next time!Thanks, keith
Hey John Goodman is a maker!! Top notch work.
this is beyond 'hobby' level ! great stuff
maui waui thank you
I'm really impressed. Kudos on giving to the community!,
Thank you
Thank you for sharing all the knowledge, this is amazing!
Its very nice! But we can see that robot is not very precise because bigest problem is that wobbling due to oscilation of movements - how to fix it? I believe that all plastic printed parts must be replaced with metal ones then wobbling could be reduced till minimum and increased precission same time.
What a superb project, well done. Really interesting.
Thank you
Wow....outstanding work! Nice job
Thank you
Great job Annin, it seems to me that it have taken a lot of effort, money and time. But you have made it
thank you
Excellent work and thank you for sharing the files with other users!!
you bet, glad to make it available. thank you.
Chris, mine is done to the point of J5! J6 is giving me wiring problems, but other than that it's working great! Still waiting on the 3mm rods, tomorrow maybe.
Awesome job! Great project! Congratulations!
This is a wonderful project, I will do my best to build it next summer!
This is badass! You ever consider installing a hotend on the tip and mount a holder on the side for filament and make it a 6 axis 3D printer?
I've had a lot of requests for that and that is one of my goals to get the robot to read G-Code so I can do 3D printing and things of that nature
That's some fantastic work. Inspiring really. Keep working hard!
Thank you
This is like the coolest thing i've ever seen on earth!!
Its been fun to make and taken a while, Once I get it walking and talking maybe it could be the coolest ;). thank you for the complement.
No problem :D
Could you share the datasheet of this robot please ???
Yeah, man.
Impressive!!
I like the wiring on the control board - there's some system going on there and the Python Desktop App's also interesting.
Well done, keep it up.
I'm thinking about getting one of those Braccio Robotic Arms to play with first up.
Thanks for the clip.
What is the reason to build it in Aluminium? Or what do you expect to improve with aluminum? Thanks for your answer - and thanks for sharing it with the world!
Mostly rigidity. the printed version is pretty good but the aluminum version doesn't have some of the vibration or bounce. Also strength - I'm afraid to fully test lifting capacity for fear of breaking the J5 carrier or post. My hope is that people can build whichever model suits their needs. obviously printing it will be lower cost - part of my goal is to create the lowest cost robot people can build at home so I hope people find this project useful. Thank you.
You sir, are badass and I can only hope to aspire to achieve what you are doing some day.
thank you. I started out with 1 stepper motor on a table with and arduino and google in front of me. its just a matter of starting small, work on one improvement after another, dont get overwhelmed just focus on learning one thing at a time.
Have you tried to benchmark how much weight it's capable of holding? Apologies if you did mention it in a video, sometimes I'm watching at work without sound!
Do you still plan to machine the parts in aluminum?
Ive tested the plastic version with 25oz but havnt fully tested for fear of breaking the J5 carrier. with the economy series gearboxes the calculation comes out at just over 3lb limit. On the aluminum version I have tested it and it can easily carry 3lbs. Before I put out a video and release updated software and instruction I want to switch the aluminum version to a couple of the precision series gear drives from stepper online. with these I should be able to get closer to 5lbs but I will likely keep its rated payload at 3 lbs for an aluminum version. and 1.5lbs for a plastic version.
Thanks, that's very cool.
I'm a second year Computer Science student at my local university currently and I'm probably going to make one of these for use in my startup company! Our goal is automation in 3D Prinitng so I'm interested to see what we can accomplish with it!
The 3D Print club is also interested in building some so I look forward to helping them as well.
Thanks again!
this is beautiful, great work and thanks for sharing it!
thanks.
wow. You are amazing Chris.
thank you. I'm glad you like it.
Imagine the precision of this robot awesome
thank you
really nice. look forward to the aluminum version. maybe you could consider some velocity valves to the gripper so they don't slam open or close. Mahalo!
sweet! thanks a lot Chris.
Thanks for sharing man. Amazing amount of work.
thank you. Im glad you like it.
Excellent design and development work. Thank you for taking the time to spell everything out for us.
I am not clear on how the part numbers in your BOM change from regular to precision.
Can you please give the part numbers for all six precision motors in the aluminum version?
Also I see there are differences in the motor BOM between the plastic and aluminum versions.
I assume that is due to the extra weight of the aluminum parts. Will the precision motors for the aluminum version fit into the plastic parts? If so, I can make the plastic one first and then go to the metal without buying more motors.
Thank you in advance for your help.
P.S. Please let me know when the kickstarter goes online.
Hello, There were some changes from the original design on the first 3-D printed robot that I build to the current version. The current version uses precision series motors for the first three joints and the same motors for joints forfive and six. The biggest difference in the parts themselves is that the spindle border for joint two is 14 mm and the motor mount for joints 12 and three are different. If you look at the bill of materials and I think the manual for the electrical enclosure and the manual for the wiring harness all the part numbers should be listed. The folder with the print files includes models to build a robot either way.I am currently right in the middle of revising all the manuals and showing pictures of the robot being built with aluminum parts the new manuals should clear everything up. The folder with all the print file includes models for both iterations using the Percision series motors and the economy series motors.I should have all the manuals ready and posted within three weeks I think. And I will be launching the kick starter shortly after that. Please feel free to email me if you have any questions and I will try to answer them.
Thanks for the information and the quick reply. I will wait a month for all of your new material to come together.
I am curious to see how you are controlling 6 steppers with the Arduino. I was planning to try to build a 4 axis arm using an old CNC shield I have. But now that I've seen yours, I am jaleous. Hats off Sir. That is some nice work.
Stephen Remillard thank you
This is awesome! I was just telling my wife that there had to be a way to make robot arms simple and inexpensive. Good stuff man! 😃
Thank you
I am highly tempted to try and build this. Currently pricing it all up, wondering if some nice chinesium bearings might bring the cost down a bit. The steppers dont look any cheaper on aliexpress than the listed stepperonline site which is interesting.
Im sure lower cost bearings would work fine as its not moving at any speed. stepperonline is about as low cost as you are going to find.
Cheers Chris, I'll let you know how I get on!
This is awesome, I've loved robotic arms ever since I played with the Radio Shack Armatron when I was a kid.
Nice. Why does it shake? Is it something that can be improved with s/w ? Maybe it stops too suddenly. Maybe s/w could make it decelerated gradually to a stop?
I have acceleration and deceleration controls programmed and It works really well. The shaking is largely due to the plastic material, some of it is due to backlash in the belt tension given plastic parts, some of it is due to backlash in the gear drives, and some of it is due to the shaky plastic table I had it on. I'm currently working on an aluminum version which will be much better. But I'm not too disappointed with this one given it's a 3d printed low-cost robot.
This Man has several brains.
It's to much wiggle room so it's inaccurate in some areas so the aluminium one will be more robust and accurate, i also recommend to gets some higher precision and higher torque motors so your able to lift more will more accuracy, but it's an awesome homemade project! Nice work indeed!
You are a true amazing engineer, thank you very much!
thank you for the compliment. glad to share the project.
Sei que vou ser mais um a perguntar...
As partes desse braço foi feito em impressora 3d ?
Vc teria os stls pra compartilhar, assim com o projeto em si ?
Incredible build, great job!
thank you
Hi Chris,Very comprehensive description of your project.I'm starting to get the parts together (I'm based in the UK). At the moment I can only find a 23H22-2804S-PG47 (rather than the 23HS30-2804S-PG47). Is this a problem? I think that the difference is operating voltage but as it is the most expensive motor/gb I want to get it right. Also is it possible to use PLA rather than ABS for the plastic parts?You set the bar really high with the quality of your instructions, explanations and downloads. This gave me the confidence to embark on this project. Thanks again Keith
Either motor should work fine, the one you are looking at is a shorter body motor at 56 mm and I believe it's 1.25 N meters of holding torque. The one I used has a longer body motor at 76 mm and it has 1.9 N meters of holding torque but both are 2.8 A and use the same gearbox so it should work fine. I selected the larger motor just so that I had more torque but I don't know that you really need it with the printed version. The aluminum version with the larger motor can handle 3lbs. But I'm working on upgrading to stepper online's new precision series motors for J1, J2 and J3 on the aluminum version robot before I release the next video and software. The plastic version will stay the same using the economy series motors. The aluminum version will have minor differences at least on a few of the motors but the software will be universal and only require a couple settings be changed.
I haven't tried PLA but I think it would work, I believe the PLA is a little more brittle.
Chris, we need to have a beer so I can cry in it, realizing I could never top something this cool!
Thank you
When you get a chance... please update the description with the link for the plans, etc. I would like to build this robot... perhaps even do a video series on setting up and running a 6-DOF robot from LinuxCNC. Thanks! (Subscribed)
yes, absolutely will. I have a couple more things to finish up and I'll post all the files and I was planning on a setup or how to video. im not as far along as I thought I would be but i'll get it all posted as soon as I can.
Chris Annin wonderful. don't sweat it... I also have a TH-cam channel... I'm backed up at least 5 videos. sometimes life gets in the way. I'll probably be ready to start the printing etc. around September at the earliest. Thanks!
sinusoidal drives could be a thing to consider. Absurdly noisy. But very impressive... the interface, the construction, the function. wow. Crazy good.
Congratulation and thank you so much for making it public, it inspires me to do the same!
Simply put, Awesome work.
dude this is amazing. well done
Really great work, Bravo , waiting for your next post :)
thank you.
A true boss ! Thanks for the work you are a legend!
Thank you
Do you consider to go open source and cooperate , for example, with freecad and Blender, and Blender cam that is for cnc , and make an opensource kit platform , that people can buy and add ideas , and 3d priting upgrades, also more hackerspaces has got lathe and metal machinery... so you can build a disen community even pick and place or use it on bigger cube type 3d printer
impressive! u did all by yourself... awesome!
thank you
Dear respected sir, Hope you are in good health with success and happiness. I am very happy for your awesome work. I'm alos working on robot arm. I'm totally begainer.
Here I have seen you are using belt and pinion. But could run the arises by Direct motors instead of using belt pulleys?
That's means, could I make Direct motor articulated robot arm?
Super geniale Idee! Respekt!!! Danke fürs Zeigen.
I'm doing it! The documentation is awesome! I've ordered parts and printed the base and J1 parts, had to split the base! And I'm terribly lazy, wish there had been an aluminum kit for under $3k, but what the heck! Thanks!