That flex spine for the ring gear blew my mind; so simple but solves a complex problem using the strengths of 3d printing. Great video series, thank you so much for sharing!
I dont mean to be so offtopic but does anyone know of a trick to get back into an instagram account..? I was dumb forgot my password. I appreciate any tricks you can offer me.
@Jordan Andre i really appreciate your reply. I found the site thru google and im waiting for the hacking stuff now. Takes quite some time so I will get back to you later with my results.
What you have done here is really incredible and unbelievable. I have been following this design since you posted it on Thingiverse. I cannot wait to get started on mine but I was waiting for the design to settle down so that the parts wouldn't change since getting everything in Ecuador takes forever and is taxed. I am really looking forward to seeing this thing running with all axis using your new design and getting some specs like usable range and weight limits, etc. It is just amazing what you have accomplished so far. Thank you for making this available to all of us.
To be fair, Jeff Kerr posted it on Thingiverse. I ran with it because I thought what he did was so brilliant and made me actually believe a 3D printed robot arm was possible. I am working on the next actuators now that I feel like I've got the small ones dialed in really nicely now. It's easier to scale up this PCB design than scale it down, so I started on the small actuators first. I think the current PCB design nails it, so it's time to get more of the software layer going--particularly the synchronized Gcode motion.
One thing I'd like to do on the next rev of the Touch PCB is add icons to the dial buttons with a sort of Shift key approach, where you touch the center pad like it's a shift key, and then hit play, pause, record waypoint, delete waypoint, record all 6dof waypoint, etc. I'm finding the 1, 3, 6, and 9 second delay approach I took on the center button to be annoying. The dial would still work the way I showed it in the video for jogging, but this would give another 7 icon options for extended features.
I'm sooo happy with how that touch dial came out. I've never seen any other maker on the Internet go with a touch dial. ESP32 makes that possible, but there's not much code out there for it, if any, so had to build the code from scratch as well. Kinda got tricky. Had to build quite the algorithm to catch the sliding effectively.
Wow, this is really amazing. Watched all 3 videos. I can't wait to sit down and dig into the design, pcb and coding you used on this to see how it all comes together. Thank you immensely for documenting all of this. I've been building CNC Machines and using them for 6 years, and this is kind of the next natural progression of my obsession with robotics.
@@JohnLauerGplus you know, the next step is to attach a 3d printer head to the arm and write a slicer that can generate code for it. Just imagine: printing actual curves in space.
Super impressed! Thanks for making this video and working through this arm. I really want to build a robotic arm like this and try to get it to weld lol maybe a bit ambitious but you've done the leg work for the rest of us simple folks lol thanks man this is rad!
out of nowhere YT recommended this and wow, so much work must have gone into this. That's some nice solutions, the slip ring, the flex backlash solution, the touch dial (I was always wondering wtf to do with all these many touch inputs on the esp32 - this is great!). Have you considered printing your parts on a resin printer? You can get cheap ones for $300 like the anycubic photon. Or you find a FabLab near you with a resin printer. Mine got a Form 2 and it's great for printing parts like this. SUPER strong and durable. Glass fibre reinforced resin. And no pesky layer lines. It will look super professional.
@@JohnLauerGplus My FabLab has prints on display made with formlabs.com/store/form-2/materials/rigid-resin/ and they are gorgeous. If the Anycubic photon (s) or Elgoo mars can use similar resin then go for it.
I haven't been impressed with Form 2 Rigid resin strength. There's no doubt that all of the Formlabs resins are much weaker than PLA. And because it's acrylic-based, it's extremely brittle. Form printers are best suited for tight tolerances and precision, not structural strength.
I had PLA gears chew themselves up in no time leaving mushed up teeth that didn’t grip anymore and I doubt I would see the the same from a glass fibre reinforced resin
Just consider layer adhesion. An inherent problem of PLA printing. Absolutely 0 problem with resin printing. That is a huge difference in structural integrity.
Jeff, you started it all with your design. I never thought 3d printed gears would work, but actually, they're working! I did move to PETG in this latest build and I'm liking it better than PLA for the strength, albeit I can't get as clean of a look, but so what. I even find the flex spine to be a little nicer on the PETG.
BTW, the BOM on the parts for one small actuator sits around $15 including the ESP32 $4, a cheaper 35mm stepper from Aliexpress @ $3, custom PCB $1.50, TMP36 $1, etc. So this arm could come in under $100 to build.
@@JohnLauerGplus Interesting. The Thingiverse User MechEngineerMike did a blog post comparing materials for 3D printed gears. He seemed to indicate PLA had better wear properties than PETG, but I would think PETG would be better for impact. I've never actually tried PETG gears, though. Nylon, I think, would be best. At some point, I'd like to make this gearbox using SLS printed nylon.
Hey john! I’m back on this project again, a while back the cat jumped up and knocked my whole robot arm down on the ground and broke the base, it was my fault that I printed out a PLA! I’m gonna rip it down and see what components I need to upgrade to your current model as always thanks a lot for putting out this project I think it’s brilliant thanks James
Thanks James. I will say that I moved onto brushless motors as these stepper motors just didn't have enough power, even after the reduction gears. SimpleFOC is a project I recommend you go check out.
Have you considered counter weighting your arm? By adding a counterweight to the base axis you remove a lot of the strain. There are static weights or active weights that move to change the balance as the arm moves. With a static weight you can set t to offset 50% if the max extended weight allowing the arm to move much quicker. With an active counterweight you can even offset higher axis.
Love the project. In industry we use teach pendants alot which allows us to give a exact numerical value as well as free jog. You could set up something similar to connect via WiFi like you already plan or connect to your primary controller.
If you do end up getting it printed and assembled, it would be great for you to post a video of your build. I'm really curious how others fare getting their own build going.
You can address the vibration in the cycloidal design by using two of them 180 degrees out of phase so that the vibration from one is cancelled out by the other.
Thanks. I hope the slip ring holds up as a concept, but I am seeing little blips here and there. That would mean a power loss on the subsequent actuators. Wondering if the blips can be smoothed by a cap, super cap, or even rechargeable battery. Needs to handle 20v though.
@@JohnLauerGplus Hi John! I'll think, the copper-to-copper slip ring will cause more an more problems. What about graphite-coals, eg. like a drilling machine? Would it be possible to install the slip rings on the inner rotating part and the coals on the outside part? Like a motor? Then you could also attach several coals distributed on the circumference of the outer part. Sorry for my english, i'm from germany: O)
Slip ring idea is sweet, it looks like its already getting beat up. Would the voltage loss on magnetic induction work? Is there enough room between the plates for the magnetic fields for it without getting in the way of the Hall Effect sensor? Hmm, trying to think of some other ideas, you could modify the housing and use small conductive V-belts for low friction high surface area contacts, or like if you could get a tiny cable to have a rather seamless joint. A v-belt would be interesting, you might be able to get both polarities on one! I don't know if such a thing exists or not, or if it was made, how durable it would be.
I've been trying to use no belts at all on this project, so not sure a v-belt would do much. I am trying to think if there's a way to get one more conductive surface for actuator-to-actuator syncing using I2C or CAN bus since ESP32 has CAN bus built-in.
@@sinpwnzorz Only problem I see there is that to get magnetic induction you need decent surface area for the coils on both PCBs. I don't have much surface area to work with given those planetary gears. I could see a coil on the top part of the actuator maybe working embedded into the top surface, but the bottom part of the actuator I can't see where that would go? You need those coils to be within 1 mm of each other for good coupling. Also, not sure if 20V could work either because induction requires PWM'ing the voltage. Doubt there's good off-the-shelf stuff to do that, but don't know really.
@@JohnLauerGplus Valid space concerns, without looking at the CAD files I can't say for certain, just brainstorming as your design has inspired me. However, traditionally, magnetic induction coils are used on the "stacked ring" configuration you're familiar with. (Two rings on one finger). It should work by doing something like small ring inside larger ring, I've looked this up a bit, and I can't find anyone else configuring it like that, probably because it sucks. However it may suck just not enough to work here... an out of the box idea perhaps? Based on the strengths of the magnetic fields in that configuration there will be more exaggerated efficiency loss than a stacked configuration, but if you overcompensated for it and used voltage regulators you should be able to accommodate it. Based on the documentation below, one of the modules I use and it's data sheet, distance only affects amperage rather than voltage. (Up to a certain point I'm sure) That would still be an issue but easier to overcome than needing to transformer up voltage and regulate it constantly. You would likely need to print a jig, and make your own wire wraps to build custom induction rings, though I think it would be fairly doable. The only tricky part would be making custom transmitters / receivers for the coils, which I have no idea how to do... but am now curious so I'm going to do some digging. Example part: www.robotshop.com/ca/en/12v-600ma-wireless-charging-module.html Documentation: elecfreaks.com/estore/download/EF03092-User-specified.pdf
You are amazing, really huge efforts explains alot of knowledge and experience as well as nice design capabilities It really looks like industrial robot arms like Fanuc or ABB
You probably already know this but there's a Hall Effect sensor built into the ESP32. Of course trying to get the geometry to work with this sensor is probably not worth the bother. Thanks for sharing your progress with us. Your robot arm is looking great.
Finished up the three lower joints and wrote some custom software for them and my computer for controlling. Using the first kind of PCB you made. Got it working propperly and it can move to a desired x, y, z position now from the computer.
I have not gotten a chance to test torque yet. Where the actuator gives is the stepper motor will skip steps. You can crank the power to about 300mA on these 35mm steppers to get more out of them. They do heat up fast at that amperage, but the fan and heatsink actually really help compensate for that if you did a short move. These are not back drivable, so you can put the motor to sleep once you're done moving and it cools nicely. Running at slow steps and full steps, not micro steps, gives really good power. It's hard for me to stop with my fingers on the actuator, if not impossible. The plastic is super strong. The gears are super strong.
Hey John, since the closing of G+ I haven’t been able to catch up with you regularly. I’ve lost a number of folks I regularly keep up with. Are you using any particular forum like you did on G+ previously? I’ve had to just use some other ones on FB.
John Lauer Yes, I signed up for maker forum, but I don’t visit much since there’s little activity. I’ll just come back here. Nice to see you doing stuff on Fusion. Did you find all those designs for the electronics online or did you design each yourself? When I designed my new controller box it would take my computer forever to calculate any move of any component within the design. Very nice design on here!
@@georgeallen1751 Oh, I found almost every component online, whether it was GrabCAD or right from the manufacturer's website or now Digikey lets you filter results on whether CAD models exist. I did have to make the DC to DC model, but i think that was it.
Why did you move to acrylic instead of the airsoft bbs? You mentioned it runs smoother. Was is the diameter, or precision, or material? I'm making some slew bearings and I've been prototyping with airsoft.
Awesome! I’m getting a 3D printer soon, and hope to create this so that I can remove prints automatically. This would allow to to print multiple things, without any human interaction.
The arm itself looks a little crappy but the tech inside is brilliant! I'm blown away by the flexible ring gear, slip rings and the touch dial. You should seriously consider starting a robotics company. ;)
That's damn impressive. although for jogging each part, what do you think about a mini 4 axis joy stick attached to each part so you can just pull it around, or just have it log the positions its been manipulated into?
I don't have a resin 3D printer, but would love to try it. Do you have one? Would you be up for trying it? Is light cured resin really strong enough for gears?
@@cnc-ua True on resolution, but gears take a beating, so does light cured resin really hold up? The gears on these actuators are actually really good. It's the stepper motor that's the weakest part.
Sounds odd, but I think this is the perfect solution for making stop motion videos. I just don't know if the arm would be able to support and DSLR and lens?
Hello John, I am wondering if you are still working on this project. I am going to make this arm to integrate as a part of a kind of kinetic sculptural project, but I have some questions regarding the motors you have recommended, because these use different voltages and I do not know if this issue is handled. I'd be pleased if you can get in touch with me.
John, I have built the robot arm and I would like to upgrade it with your actuators, which is a real cool design btw. I have watched a bunch of your videos and am impressed by your skills. My problem is I am old school I worked in custom electronics in the 80s, I either did wire wrap or I laid out a PCB using a clear sheet and black layout tape at 4x scale then sent that off to the board makers. Having recently retired I am getting back into electronics and I am not sure what I need to do to have the pc boards you designed made. I have ordered all of the parts from the BOM, I am sure I need to have 3 boards made, the round pc board, the round touch sensor board and the small pc board. I am not sure which files I need to send to get the boards made. Thank you for any assistance you can offer.
I probably need to make a video of making the PCBs. I have also recently discovered that jlcpcb.com will even solder your basic parts to your board for you for a few extra dollars, so now I'm working on tweaking the design of the PCB to put all components on one side and switch to the standard parts in their library so that you get like a 70% assembled/soldered board in the mail. Since making that video could take several weeks until it's out, you could just watch a video on using jlcpcb to get through the ordering of the custom boards.
John Lauer thank you, I will check it out, my fingers aren’t as agile as they use to be 30 years ago so a pre assembled board with the small parts sounds good.
Welcome. I have a question about the universal robot arm, how is the mechanism of moving the arm by hand and then storing the movement, explain this please
Hello John, congratulations for the amazing job done. I'm looking forward to see new update. I had some problem of melting gear with pla, i'll try with ABS. I'm also working for add a magnetic encoder inside the actuator, however I think the static mechanic torque is pretty good. How is gone with the project? For the development of other actuator what do you think about using Mechaduino or makerbase closed loop servo ? King regards
I solved the melting plastic with the active cooling with the fan, heatsink, and code running on the ESP32. On the closed loop, I do think that would make sense, so the notion is to look at an encoder somewhere in the mechanism to know that steps are getting missed. Not sure best design. Would love to see if any others are able to get as far as I did and run some experiments.
www.blackmagic3d.com/Conductive-p/grphn-pla.htm I made some Nike sandals that I could plug my rife machine into instead of having to hold and step on pads. I just walk around with them plugged in! LOL
I sort of did, then tossed the idea out becuase there was nowhere good to place it on the actuator. After seeing the video, any thoughts where I could place it? I do think I still may run into issues in the future where you need to detect missed steps and the robot should know it's got too heavy of a load. An absolute encoder could be the trigger for that and solve the homing problem. Have you used the AS5600? With 12-bit DAC, is it really precise, or noisy and indeterminate such that you end up with about half of that resolution or worse for knowing your absolute position? I have one in my electronics stash but never really played with it much.
@@JohnLauerGplus I got my AS5600 from ebay a few days back... Still haven't opened the package... Will let you know how it goes when i get a chance to play with it... Which motor are you using for this project?
@@nakulsolanki The 35mm stepper motor is in the BOM on the Github repo, but really I have 3 different motors I've been using. The expensive one is here www.moonsindustries.com/p/35mm-permanent-magnet-stepper-motors/35pm048s8-08001-000004611120001819, but I needed some beater/test ones too and they seem to work just as well. This one is really strong and almost as good as the Moon's but you have to grind off the gear with a Dremel www.aliexpress.com/item/32975464158.html and then there's this super cheap one that works fairly well but isn't as strong and doesn't need any grinding www.aliexpress.com/item/2054451318.html
drive.google.com/open?id=1bSY4Je9cvTTw9S1i4J6flVdxAK5qOpen sorry i don`t have cad on my work - but something like on pic (kid pic but i think it more informative)
@@РодинаМояРоссия-ч3л I see what you mean. The balls in your diagram are very similar to the brushes in the 3D printed design, so I'm not sure they provide much benefit. I've considered springs behind the brushes, but the plastic acts as a spring with the shape of the brushes. Plus, it's easier to to just print, than have to buy parts and assemble.
can you give a cost estimate for this design. i am a computer engineering student building a mobile robot arm with robotic vision. Being from Pakistan cost are a huge factor in choice of design and components.
Oh, the dynamic model is that you train it on each axis to the position you want using the touch dial. Then record the point or all 6 points from each actuator. Then play it back. The play back just moves each actuator over the same length of time, so the longest actuator moves becomes the long pole in the tent. The shorter moves just slow down so they all end up at the end of the move at the same time.
I talk about the attempt to use the Trinamic driver in the video. It failed. The cost is too high at $5 per chip with zero benefits and only negative downsides. The chips blow like crazy, so you have to add 8 protection diodes, which don't really fit into the actuator and expand the amount of components you have to worry about. The sensorless homing does not work. The DRV8825's are actually fairly silent at 4 microsteps, which is what I'm running them at, so the silence of the Trinamic's isn't super compelling for the pain and cost. I am considering how to do absolute positioning, which would eliminate the need for the hall sensor, but there's no good solution given the format of the actuator. You want to measure the position on the reduced planetary gear output, not on the stepper motor shaft, so none of these magnetic chips do the job as far as I can find. Also, that chip you recommend is $12 so that's really expensive for this robot arm.
No, all of the sensors are done over wireless. Just the power is done over the split ring. I don't like it though as I get little blips of power loss, so will be abandoning it.
@@JohnLauerGplus Do you have an update of the changes that you made? Or how can you reduce the power loss? Would love to get your input on how to transmit power without wires!
@@cassandratom8986 I have not made any updates yet as I'm moving to a brushless motor for the actuator which has a slew of challenges. On the power, I'm not sending that wirelessly, I'm just running wires. It's the signals that were wireless, not the power. However, with moving to wired power, it may also be worth just running wires for the signals too.
I'd like to get it all metric too. I used Jeff Kerr's design and he used mostly imperial, but any new stuff I'm doing is metric as I think that should be the gold standard these days given we're a global ecosystem with so many parts out of China that we just have to settle on metric.
That flex spine for the ring gear blew my mind; so simple but solves a complex problem using the strengths of 3d printing. Great video series, thank you so much for sharing!
I dont mean to be so offtopic but does anyone know of a trick to get back into an instagram account..?
I was dumb forgot my password. I appreciate any tricks you can offer me.
@Denver Alan Instablaster ;)
@Jordan Andre i really appreciate your reply. I found the site thru google and im waiting for the hacking stuff now.
Takes quite some time so I will get back to you later with my results.
@Jordan Andre it worked and I now got access to my account again. Im so happy!
Thanks so much, you really help me out !
@Denver Alan glad I could help :)
I didn't understand a single thing, but that's really cool.
Wow, I'm baffled. You are making such beautiful work! And great explanations as well. Really looking forward to seeing your next update!
What you have done here is really incredible and unbelievable. I have been following this design since you posted it on Thingiverse. I cannot wait to get started on mine but I was waiting for the design to settle down so that the parts wouldn't change since getting everything in Ecuador takes forever and is taxed. I am really looking forward to seeing this thing running with all axis using your new design and getting some specs like usable range and weight limits, etc. It is just amazing what you have accomplished so far. Thank you for making this available to all of us.
To be fair, Jeff Kerr posted it on Thingiverse. I ran with it because I thought what he did was so brilliant and made me actually believe a 3D printed robot arm was possible. I am working on the next actuators now that I feel like I've got the small ones dialed in really nicely now. It's easier to scale up this PCB design than scale it down, so I started on the small actuators first. I think the current PCB design nails it, so it's time to get more of the software layer going--particularly the synchronized Gcode motion.
One thing I'd like to do on the next rev of the Touch PCB is add icons to the dial buttons with a sort of Shift key approach, where you touch the center pad like it's a shift key, and then hit play, pause, record waypoint, delete waypoint, record all 6dof waypoint, etc. I'm finding the 1, 3, 6, and 9 second delay approach I took on the center button to be annoying. The dial would still work the way I showed it in the video for jogging, but this would give another 7 icon options for extended features.
@@JohnLauerGplus I would like to talk to you on zoom if it is ok
The touch dail for the jogging is mind blowing ! Amazing work, very inspiring
I'm sooo happy with how that touch dial came out. I've never seen any other maker on the Internet go with a touch dial. ESP32 makes that possible, but there's not much code out there for it, if any, so had to build the code from scratch as well. Kinda got tricky. Had to build quite the algorithm to catch the sliding effectively.
Wow, this is really amazing. Watched all 3 videos. I can't wait to sit down and dig into the design, pcb and coding you used on this to see how it all comes together. Thank you immensely for documenting all of this. I've been building CNC Machines and using them for 6 years, and this is kind of the next natural progression of my obsession with robotics.
Wonderful, this is really high quality design
Well, thank you. Now I want to make it a BLDC instead of a stepper.
The quality of design and actual building is astonishing!
This guy should make a kit or a video series a really detailed one for newbies to make a robot just like this, it would be good.
I, personally, would love that
When will a Part 4 come out?? This is dope!
Trying to get a brushless motor approach going for more power, so major redesign.
aproved
@@JohnLauerGplus you know, the next step is to attach a 3d printer head to the arm and write a slicer that can generate code for it. Just imagine: printing actual curves in space.
@@JohnLauerGplus I hope it comes out soon, I could really use it
@@JohnLauerGplus I'm making an automated circuit board factory using this design. And I'm going to be selling the circuit boards on Amazon
Incredible job man. Your problem solving skills are top notch.
Super impressed! Thanks for making this video and working through this arm. I really want to build a robotic arm like this and try to get it to weld lol maybe a bit ambitious but you've done the leg work for the rest of us simple folks lol thanks man this is rad!
You and Jeff Kerr are such an amazing people, for both of you to make your work available to all is so selfless :)
out of nowhere YT recommended this and wow, so much work must have gone into this. That's some nice solutions, the slip ring, the flex backlash solution, the touch dial (I was always wondering wtf to do with all these many touch inputs on the esp32 - this is great!). Have you considered printing your parts on a resin printer? You can get cheap ones for $300 like the anycubic photon. Or you find a FabLab near you with a resin printer. Mine got a Form 2 and it's great for printing parts like this. SUPER strong and durable. Glass fibre reinforced resin. And no pesky layer lines. It will look super professional.
I have been wondering how strong the light cured resin is, but from what you're saying it really could work. I think I'm going to go get one.
@@JohnLauerGplus My FabLab has prints on display made with formlabs.com/store/form-2/materials/rigid-resin/ and they are gorgeous. If the Anycubic photon (s) or Elgoo mars can use similar resin then go for it.
I haven't been impressed with Form 2 Rigid resin strength. There's no doubt that all of the Formlabs resins are much weaker than PLA. And because it's acrylic-based, it's extremely brittle. Form printers are best suited for tight tolerances and precision, not structural strength.
I had PLA gears chew themselves up in no time leaving mushed up teeth that didn’t grip anymore and I doubt I would see the the same from a glass fibre reinforced resin
Just consider layer adhesion. An inherent problem of PLA printing. Absolutely 0 problem with resin printing. That is a huge difference in structural integrity.
Wow! So many outstanding solutions in one video. Never heard about flex spline, thanks a lot!
Holy crap - this is amazing what you've done! People are going to be clamoring to buy this as a product.
Jeff, you started it all with your design. I never thought 3d printed gears would work, but actually, they're working! I did move to PETG in this latest build and I'm liking it better than PLA for the strength, albeit I can't get as clean of a look, but so what. I even find the flex spine to be a little nicer on the PETG.
BTW, the BOM on the parts for one small actuator sits around $15 including the ESP32 $4, a cheaper 35mm stepper from Aliexpress @ $3, custom PCB $1.50, TMP36 $1, etc. So this arm could come in under $100 to build.
@@JohnLauerGplus Interesting. The Thingiverse User MechEngineerMike did a blog post comparing materials for 3D printed gears. He seemed to indicate PLA had better wear properties than PETG, but I would think PETG would be better for impact. I've never actually tried PETG gears, though. Nylon, I think, would be best. At some point, I'd like to make this gearbox using SLS printed nylon.
@@lobocnc Check out Igus' filaments, they are engineered for bearings!
@@JohnLauerGplus tagging you here too, see above
Hey john! I’m back on this project again, a while back the cat jumped up and knocked my whole robot arm down on the ground and broke the base, it was my fault that I printed out a PLA! I’m gonna rip it down and see what components I need to upgrade to your current model as always thanks a lot for putting out this project I think it’s brilliant thanks James
Thanks James. I will say that I moved onto brushless motors as these stepper motors just didn't have enough power, even after the reduction gears. SimpleFOC is a project I recommend you go check out.
Oh my god... I've never seen a project so advanced crafted by a single person... I would really like to see the part 4!!
Have you considered counter weighting your arm? By adding a counterweight to the base axis you remove a lot of the strain. There are static weights or active weights that move to change the balance as the arm moves. With a static weight you can set t to offset 50% if the max extended weight allowing the arm to move much quicker. With an active counterweight you can even offset higher axis.
Отличная работа, не забрасывай проект! Спасибо за труд.
Вообще норм
I absolutely adore this series!
Well, thank you.
it looks great. Thanks for explaining what worked and what didn't.
Love the project. In industry we use teach pendants alot which allows us to give a exact numerical value as well as free jog. You could set up something similar to connect via WiFi like you already plan or connect to your primary controller.
Thats some wonderful engineering work. Thank you very much for sharing the process
Hi, it was great to see this amazing project and i leant alot.
Just watched last part and hope you improve it and upload another parts.
Amazing job, you should have patreon! Waiting for further updates, I want to replicate this and share any improvements that I might find
If you do end up getting it printed and assembled, it would be great for you to post a video of your build. I'm really curious how others fare getting their own build going.
Woow this is an atual product, unlike a lot of youtubers that just make somethink bulky. Awesome
Your dedication is remarkable. Great work!
Nice job, it's nice that you seem to be enjoying it really much :)
You can address the vibration in the cycloidal design by using two of them 180 degrees out of phase so that the vibration from one is cancelled out by the other.
Great work! The slip ring idea, that’s awesome, can’t wait to see this in action!
Thanks. I hope the slip ring holds up as a concept, but I am seeing little blips here and there. That would mean a power loss on the subsequent actuators. Wondering if the blips can be smoothed by a cap, super cap, or even rechargeable battery. Needs to handle 20v though.
@@JohnLauerGplus Hi John! I'll think, the copper-to-copper slip ring will cause more an more problems. What about graphite-coals, eg. like a drilling machine? Would it be possible to install the slip rings on the inner rotating part and the coals on the outside part? Like a motor? Then you could also attach several coals distributed on the circumference of the outer part.
Sorry for my english, i'm from germany: O)
Amazing man, I am good on code and little electronic but with these mechanic parts I need a lot to learn, that is a really awesome
Slip ring idea is sweet, it looks like its already getting beat up. Would the voltage loss on magnetic induction work? Is there enough room between the plates for the magnetic fields for it without getting in the way of the Hall Effect sensor? Hmm, trying to think of some other ideas, you could modify the housing and use small conductive V-belts for low friction high surface area contacts, or like if you could get a tiny cable to have a rather seamless joint. A v-belt would be interesting, you might be able to get both polarities on one! I don't know if such a thing exists or not, or if it was made, how durable it would be.
I've been trying to use no belts at all on this project, so not sure a v-belt would do much. I am trying to think if there's a way to get one more conductive surface for actuator-to-actuator syncing using I2C or CAN bus since ESP32 has CAN bus built-in.
@@JohnLauerGplus What about magnetic induction?
@@sinpwnzorz Only problem I see there is that to get magnetic induction you need decent surface area for the coils on both PCBs. I don't have much surface area to work with given those planetary gears. I could see a coil on the top part of the actuator maybe working embedded into the top surface, but the bottom part of the actuator I can't see where that would go? You need those coils to be within 1 mm of each other for good coupling. Also, not sure if 20V could work either because induction requires PWM'ing the voltage. Doubt there's good off-the-shelf stuff to do that, but don't know really.
@@JohnLauerGplus Valid space concerns, without looking at the CAD files I can't say for certain, just brainstorming as your design has inspired me. However, traditionally, magnetic induction coils are used on the "stacked ring" configuration you're familiar with. (Two rings on one finger). It should work by doing something like small ring inside larger ring, I've looked this up a bit, and I can't find anyone else configuring it like that, probably because it sucks. However it may suck just not enough to work here... an out of the box idea perhaps? Based on the strengths of the magnetic fields in that configuration there will be more exaggerated efficiency loss than a stacked configuration, but if you overcompensated for it and used voltage regulators you should be able to accommodate it.
Based on the documentation below, one of the modules I use and it's data sheet, distance only affects amperage rather than voltage. (Up to a certain point I'm sure) That would still be an issue but easier to overcome than needing to transformer up voltage and regulate it constantly. You would likely need to print a jig, and make your own wire wraps to build custom induction rings, though I think it would be fairly doable.
The only tricky part would be making custom transmitters / receivers for the coils, which I have no idea how to do... but am now curious so I'm going to do some digging.
Example part:
www.robotshop.com/ca/en/12v-600ma-wireless-charging-module.html
Documentation:
elecfreaks.com/estore/download/EF03092-User-specified.pdf
You are amazing, really huge efforts explains alot of knowledge and experience as well as nice design capabilities
It really looks like industrial robot arms like Fanuc or ABB
You probably already know this but there's a Hall Effect sensor built into the ESP32.
Of course trying to get the geometry to work with this sensor is probably not worth the bother.
Thanks for sharing your progress with us. Your robot arm is looking great.
Yeah, I tried to use that and it was impossible.
Amazing job ! Well showed and explained. Really impressive work.
A big truck load of thumbs up (👍) for you.
Thanks!
Wow, fantastic project! Really looking forward to seeing how this evolves. Best of success to you!
Damn this really is amazing, cant wait to see the final build. Mine is coming along nicely.
How far along are you? Did you get to the electronics yet?
Finished up the three lower joints and wrote some custom software for them and my computer for controlling. Using the first kind of PCB you made.
Got it working propperly and it can move to a desired x, y, z position now from the computer.
@@dagsvikeleven1589 chilipeppr.com/arm is coming along as the control software if you want to check it out.
Love the project, well done! Looking forward to the next update :)
i ve been lookin at all 3 parts, such a great job!
Fantastic work John.
this needs more views, super awesome!
Got the web interface going well recently too, so this is heating up even more.
@@JohnLauerGplus which coding language u are using
@@vijaygharat2968 Lua. See the Github repo.
Are there course of this language
This is a really amazing design
very impressive. subbed. did you test the torque? whats the limiting factor? gears, motor or casing?
I have not gotten a chance to test torque yet. Where the actuator gives is the stepper motor will skip steps. You can crank the power to about 300mA on these 35mm steppers to get more out of them. They do heat up fast at that amperage, but the fan and heatsink actually really help compensate for that if you did a short move. These are not back drivable, so you can put the motor to sleep once you're done moving and it cools nicely. Running at slow steps and full steps, not micro steps, gives really good power. It's hard for me to stop with my fingers on the actuator, if not impossible. The plastic is super strong. The gears are super strong.
Simply incredible
Brilliant work.
Very nice work. You are learning a lot and so we are. Thank you!
This is amazing work. Thank you!
Hey John, since the closing of G+ I haven’t been able to catch up with you regularly. I’ve lost a number of folks I regularly keep up with. Are you using any particular forum like you did on G+ previously? I’ve had to just use some other ones on FB.
Sadly the G+ forum has not been reproduced. There is this forum forum.makerforums.info/c/ChiliPeppr but almost nobody visits it from the G+ group.
John Lauer Yes, I signed up for maker forum, but I don’t visit much since there’s little activity. I’ll just come back here. Nice to see you doing stuff on Fusion. Did you find all those designs for the electronics online or did you design each yourself? When I designed my new controller box it would take my computer forever to calculate any move of any component within the design. Very nice design on here!
@@georgeallen1751 Oh, I found almost every component online, whether it was GrabCAD or right from the manufacturer's website or now Digikey lets you filter results on whether CAD models exist. I did have to make the DC to DC model, but i think that was it.
John Lauer Thanks for the info! I can use that!
WOW! just incredible.
Why did you move to acrylic instead of the airsoft bbs? You mentioned it runs smoother. Was is the diameter, or precision, or material? I'm making some slew bearings and I've been prototyping with airsoft.
Awesome! I’m getting a 3D printer soon, and hope to create this so that I can remove prints automatically. This would allow to to print multiple things, without any human interaction.
i think it is possible to program your g-code so that it pushes the printed product. Hope it helps.
Drone Videos search for a conveyer belt printbed thats best
The arm itself looks a little crappy but the tech inside is brilliant! I'm blown away by the flexible ring gear, slip rings and the touch dial. You should seriously consider starting a robotics company. ;)
So nice! What effective torque do you get from these actuators?
Thank you thank you. Keep making content this is great! Will we possibly get to see a collaboration video with Dr.Kerr and yourself?
Have there been any updates? Would love to see what your built for a gripper.
That's damn impressive. although for jogging each part, what do you think about a mini 4 axis joy stick attached to each part so you can just pull it around, or just have it log the positions its been manipulated into?
I tried the mini joystick in each actuator and it took up too much space.
Thank you so much for sharing this work.
That antibacklash solution is brilliant. Have you considered to use resin 3d print for gears?
I don't have a resin 3D printer, but would love to try it. Do you have one? Would you be up for trying it? Is light cured resin really strong enough for gears?
@@JohnLauerGplus no, I don't. But those resin printers have way better precision, which should help
@@cnc-ua True on resolution, but gears take a beating, so does light cured resin really hold up? The gears on these actuators are actually really good. It's the stepper motor that's the weakest part.
@@JohnLauerGplus there are different resins. Some of them are really tough and bullet proof. But I have no experience, so this is just a theory
Sounds odd, but I think this is the perfect solution for making stop motion videos. I just don't know if the arm would be able to support and DSLR and lens?
Hello John, I am wondering if you are still working on this project. I am going to make this arm to integrate as a part of a kind of kinetic sculptural project, but I have some questions regarding the motors you have recommended, because these use different voltages and I do not know if this issue is handled. I'd be pleased if you can get in touch with me.
John, I have built the robot arm and I would like to upgrade it with your actuators, which is a real cool design btw. I have watched a bunch of your videos and am impressed by your skills. My problem is I am old school I worked in custom electronics in the 80s, I either did wire wrap or I laid out a PCB using a clear sheet and black layout tape at 4x scale then sent that off to the board makers. Having recently retired I am getting back into electronics and I am not sure what I need to do to have the pc boards you designed made. I have ordered all of the parts from the BOM, I am sure I need to have 3 boards made, the round pc board, the round touch sensor board and the small pc board. I am not sure which files I need to send to get the boards made. Thank you for any assistance you can offer.
I probably need to make a video of making the PCBs. I have also recently discovered that jlcpcb.com will even solder your basic parts to your board for you for a few extra dollars, so now I'm working on tweaking the design of the PCB to put all components on one side and switch to the standard parts in their library so that you get like a 70% assembled/soldered board in the mail. Since making that video could take several weeks until it's out, you could just watch a video on using jlcpcb to get through the ordering of the custom boards.
John Lauer thank you, I will check it out, my fingers aren’t as agile as they use to be 30 years ago so a pre assembled board with the small parts sounds good.
Very nice project, good job
Welcome. I have a question about the universal robot arm, how is the mechanism of moving the arm by hand and then storing the movement, explain this please
Thanks for the video, are we getting new video with new arms
Hello John, congratulations for the amazing job done. I'm looking forward to see new update. I had some problem of melting gear with pla, i'll try with ABS. I'm also working for add a magnetic encoder inside the actuator, however I think the static mechanic torque is pretty good.
How is gone with the project? For the development of other actuator what do you think about using Mechaduino or makerbase closed loop servo ? King regards
I solved the melting plastic with the active cooling with the fan, heatsink, and code running on the ESP32.
On the closed loop, I do think that would make sense, so the notion is to look at an encoder somewhere in the mechanism to know that steps are getting missed. Not sure best design. Would love to see if any others are able to get as far as I did and run some experiments.
Hi @@JohnLauerGplus , do you have some example code. I try to search it in the repo but didn't found
@@Alby32mx The code is in the lua folder in the repo. You also need to upload the firmware to your ESP32 from the firmware folder in the repo.
Hi would be good to see the arm complete and moving
Very well made
Really great Stuff, keep it up.
Nice work!
awesome work!!! I'll try to do mine, thanks for share!
Nice, really professional!
PS. There is an electrical conductive filament that we could use to print the slip ring plus any other internal wiring!
www.blackmagic3d.com/Conductive-p/grphn-pla.htm I made some Nike sandals that I could plug my rife machine into instead of having to hold and step on pads. I just walk around with them plugged in! LOL
pss. Look at ClicBot it uses a wireless feature to supply power to each add on motor. I think it might help your design to mimic?
Cool project... have you considered using AS5600?
I sort of did, then tossed the idea out becuase there was nowhere good to place it on the actuator. After seeing the video, any thoughts where I could place it? I do think I still may run into issues in the future where you need to detect missed steps and the robot should know it's got too heavy of a load. An absolute encoder could be the trigger for that and solve the homing problem. Have you used the AS5600? With 12-bit DAC, is it really precise, or noisy and indeterminate such that you end up with about half of that resolution or worse for knowing your absolute position? I have one in my electronics stash but never really played with it much.
@@JohnLauerGplus I got my AS5600 from ebay a few days back... Still haven't opened the package... Will let you know how it goes when i get a chance to play with it... Which motor are you using for this project?
@@nakulsolanki The 35mm stepper motor is in the BOM on the Github repo, but really I have 3 different motors I've been using. The expensive one is here www.moonsindustries.com/p/35mm-permanent-magnet-stepper-motors/35pm048s8-08001-000004611120001819, but I needed some beater/test ones too and they seem to work just as well. This one is really strong and almost as good as the Moon's but you have to grind off the gear with a Dremel www.aliexpress.com/item/32975464158.html
and then there's this super cheap one that works fairly well but isn't as strong and doesn't need any grinding www.aliexpress.com/item/2054451318.html
Well done John! 😀
Keep up the work
You can use 2 metal springing balls to receive power on motors. This is more effective (or two rolls)
I'm not sure what you mean? You mean in place of a slip ring? Any examples out there?
@@JohnLauerGplus I try create pic to you later. Ok?
drive.google.com/open?id=1bSY4Je9cvTTw9S1i4J6flVdxAK5qOpen sorry i don`t have cad on my work - but something like on pic (kid pic but i think it more informative)
@@РодинаМояРоссия-ч3л I see what you mean. The balls in your diagram are very similar to the brushes in the 3D printed design, so I'm not sure they provide much benefit. I've considered springs behind the brushes, but the plastic acts as a spring with the shape of the brushes. Plus, it's easier to to just print, than have to buy parts and assemble.
John Lauer yes like brushes. Well - as you wish but this is good way to solve problem frictions. ;)
can you give a cost estimate for this design. i am a computer engineering student building a mobile robot arm with robotic vision. Being from Pakistan cost are a huge factor in choice of design and components.
Great work, But if I can ask you about the dynamic model of your robot arm, and if I can have your approval in order to study it.
Yeah, this is all open source so go for it.
Oh, the dynamic model is that you train it on each axis to the position you want using the touch dial. Then record the point or all 6 points from each actuator. Then play it back. The play back just moves each actuator over the same length of time, so the longest actuator moves becomes the long pole in the tent. The shorter moves just slow down so they all end up at the end of the move at the same time.
Nice work 👍🏽
Thanks
Is the flex spline really durable, please? Does it break after one hour? I saw a comment on your thingiverse. Thanks in advance for your reply :)
No, it's durable. Plastic has a flexibility to it like other materials do. This doesn't bend it beyond any point of real stress.
Is there any BOM for the Robot? Eg. for screews and nuts?
I would actually have used the AS5047P 14bit high res positioning sensor with one ofe trinamoc drivers for accurate positioning and stop/ homing
I talk about the attempt to use the Trinamic driver in the video. It failed. The cost is too high at $5 per chip with zero benefits and only negative downsides. The chips blow like crazy, so you have to add 8 protection diodes, which don't really fit into the actuator and expand the amount of components you have to worry about. The sensorless homing does not work. The DRV8825's are actually fairly silent at 4 microsteps, which is what I'm running them at, so the silence of the Trinamic's isn't super compelling for the pain and cost.
I am considering how to do absolute positioning, which would eliminate the need for the hall sensor, but there's no good solution given the format of the actuator. You want to measure the position on the reduced planetary gear output, not on the stepper motor shaft, so none of these magnetic chips do the job as far as I can find. Also, that chip you recommend is $12 so that's really expensive for this robot arm.
This is very cool.
How do you sink all the stepper motors together
I was so excited by your videos that I forgot to subscribe
Perfect job!!!
Thanks!
Did you manage to calculate inverse kinematics?
great job
Well thank you.
Curious... what state you in? Near Vegas?
oh Man U had the upgrade the pcd man oh
THIS IS AMAZING !
Really nice., is the motor stepper or DC motor?
Stepper
Hello! Good job! I think.. this is the 3D printer - new technique....arms with printing heat. Try!
For the Copper Tape Split Ring, did you attach the sensors and wires directly to the copper tape?
No, all of the sensors are done over wireless. Just the power is done over the split ring. I don't like it though as I get little blips of power loss, so will be abandoning it.
@@JohnLauerGplus Do you have an update of the changes that you made? Or how can you reduce the power loss? Would love to get your input on how to transmit power without wires!
@@cassandratom8986 I have not made any updates yet as I'm moving to a brushless motor for the actuator which has a slew of challenges. On the power, I'm not sending that wirelessly, I'm just running wires. It's the signals that were wireless, not the power. However, with moving to wired power, it may also be worth just running wires for the signals too.
Would it be asked too much to make this all metric instead of a units mix and match?
I'd like to get it all metric too. I used Jeff Kerr's design and he used mostly imperial, but any new stuff I'm doing is metric as I think that should be the gold standard these days given we're a global ecosystem with so many parts out of China that we just have to settle on metric.
Any resources for biped robot using Arduino and small servo motors, for newcomers like me?
Thanks for video!))) Good job!))
super cool
like it
Amazing !
can you carve wood with it?