You should do a final episode! It could help with maybe students who want to do robotics in the future (like me). Videos like that with the simplified robot would be much appreciated by not only me but a lot of people. Nonetheless, we thank you for keeping us updated and helping us with all the videos. Keep up the good work!
Nice work John! I've been working on a multi-axis arm with planetary gears also and since I have no idea how these joints are supposed to be designed I've gone through a lot of trial and error. Taking a look at your design is really helping me improve my design (or pretty much copy yours) so thank you for sharing this
Working on it +Mikael! The next step now that the actuators are assembled is to get the electronics going. Using ESP32's in each joint to generate stepper pulses and then will use ChiliPeppr's Cayenn protocol to upload all Gcode moves to each joint. Stay tuned.
@@JamecBond not if you want a touch sensor in each actuator for record and playback. Also, they are $3 each. That's so cheap you can sprinkle them everywhere. Also, running wires is a serious problem and this avoids that.
@03:35 You can simply create a press fit cover for the inside which will in turn hold all the nuts in place. Falling nuts into a PG system is no fun and will not be an issue if the cover is placed.
Good idea. It would add some height to the system, but probably could make the cover be 0.6mm height, so wouldn't add much. Those planetary gears do rub along the top where those nuts are. The top sun gear does push itself up towards the roof of the gearbox. So do the planetary gears, so the smoother the roof, the better. A cover could help that be even smoother than having those nuts exposed. As I've seen the grease squeezes up into the nut holes over time. Although, I do think it makes sense to have some gunk repositories where over time the shreds of plastic can bunch up and get pushed into depressed areas to help give longevity.
ive heard some bad things about mixed vaseline and plastics, it supposedly breaks the plastic down. I personally use white lithium grease. not sure if its true, but better safe than sorry. great video btw
I was worried about that. I searched online and couldn't find much data on the correct lubricant. If vaseline breaks down the plastic, then indeed something else should get used. I'll try the white lithium grease like you suggest.
@@JohnLauerGplus I definitely second the suggestion of white lithium grease. You'll commonly find it used in joysticks and things because it's plastic-safe. 👍
This is amazingly inventive in the use of materials when considering the potential accuracy of the end product, have you tried making a new arm with the arm, that'd make for a good investment.
Nice one. Had that exact gearing idea in my head a few years ago for a 'planned' robot idea that I never found time for. Nice to see it's out in the wild via the makerspace hive mind. Hoping to share another gearing design soon that will be better still for robotics (true zero backlash, high reduction, and very high load capabilities)
Ooh. Tell me more. It would be nice to have an even stronger gear reduction box. This one is decent, but the robot arm could be much stronger. Do you have any links to any designs? Or videos you've done? Or concepts?
@@JohnLauerGplus not yet but hopefully will have something to share in a few months or so. I'll shoot you a link once I'm ready to share it with the world :) probably via a Kickstarter if prototyping goes as planned.
the high torque nema 17 is not meant to be on rotation, its the medium torque in there... the high torque is on the shoulder, as it has to support all of the weight of the arm...
Beautiful, I'm quite curious about the next video where you hook up the bluetooth and use it!! Great design, nice to also see the small fans inside the housing. Nice details of your work.
I just wanna point out that metric isn't the 'english' measurement, mostly everyone in the world uses it except for US. Imperial is actually the 'english' measurement that they moved out of which is 'inches, feet, etc.'
You don't sand gears, you "lap in" gears. Apply a fine lapping compound to all the gears and run them until they get quiet. The lapping compound is basically grease and sanding media and the teeth of the gears will rapidly wear away the spots that interfere on the gear teeth. This has been the acceptable practice since modern gears were invented.
Interesting. I am working on a BLDC version of the actuator now. I'm printing the new compound planetary gear in nylon and indeed I am breaking in the gears by running them raw against each other and it is smoothing them out. However, what lapping compound would you use for plastic to make this process faster/better? I can't find any when you search google for "3d printed gear lapping compound".
@@JohnLauerGplus Loctite has a brand called "Clover." I'd say a nice 200-300 grit would do it. The Lower the grit the more abasive and faster, but finish is not as pretty... higher grits will give a smother, prettier finish, but take longer. After the gears quieten, remove and clean them thoroughly. Add real grease and they will be perfectly mated! THIS is old-school gear perfection!
The bb's seem pretty annoying to deal with. Have you tried just printing spheres in place? I did that with some print in place bearings I made and they work very well.
That's a cool idea, but those need to be really smooth, so that would worry me you'll never get the smoothness and roundness that you need. Albeit you are sort of averaging them together, so maybe it "mostly" works?
I will say in subsequent designs I just went with cheap off-the-shelf metal bearings because they're worth the simplicity, smoothness, and tightness beyond a 3D printed bearing and they're not that expensive on the low end, but 10 times better than a 3D print bearing.
I think it would be interesting to look at options for bearings, but the way these actuators are all integrated is very nice and trying to get an off-the-shelf bearing in there may not be as elegant. However, I'd love to see if anyone could CAD out a potential design.
@@CinemaTablet I don't think that would be smooth or friction-free or tight enough. Ultimately you do have plastic against plastic since those airsoft BB's are plastic.
after you break in the gears, you need to wash the plastic dust out of them as it will act as an abrasive along with the Vaseline breaking down your plastic. I would invest in some rubiks cube lubricant from cubicle labs.
I don't think steel balls would work much better, but perhaps. Didn't try it, but I think steel against plastic may wear in an odd way. Plastic against plastic seems to wear nicely. Also, these bearings turn out really tight yet smooth rotation, so that's not the part of the design I'm worried about. What I don't like is the power of the stepper motor.
Very cool. For the bbs, (or however it's spelled), I wonder if printing a separate assembly - like a self-contained thrust bearing - might solve the problem.
I am now using PETG on my latest design as you can see in Part 3 ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html, but I am hearing robot guys using 3D printing like the flex on PLA better for gears. I'm not quite sure yet.
not bad but for 3d printing you make hard robot arm. With 3 motors can be done. because you dont use 4 or 5 axis. you print in 2 dimension X and Y(vertical) and Z is leveling. you cant print on X and Z plane(horizontal). EXAMPLE of ROBOT ARM is: Human arm. Entire body is the Z axis from the shoulder start the arm and use it in horizontal way.The elbow is where 3rd motor will stay. You will not have specific X and y motor axis. Motor X and Y must work toghether to make lines and curve. Z motor axis is independent.
Because wiring robots is a huge pain and I need/want a temp sensor in each actuator, fan controller, hall sensor for end stop, ws2812 LED for status, and touch sensor for local jogging and programming. We now know how to run signals over the air, rather than needing copper, so can eliminate many wires that way. ESP32's are dirt cheap and a novel way to go these days. Yes, it's all theory until it comes together, but might as well give it a go.
GREAT Stuff !.. and a lot of great info on building it. However, I noticed you didn't put ( or at least I didn't see?) you put vaseline around the airsoft BB's?? Also.. I built one a few years ago, similar to this. but I have never seen anyone use vaseline? I'm curious how long it last? as I used a tube of Lithium White grease that works really really slick and well.. still lasts after a few years!! But cost a lot more than a jar of vaseline . LOL.. so if it works as good. .I'll try vaseline next time !!
Latest design now has a hall effect endstop. I'm getting about 500 Hz PWM signal to stepper driver before these 35mm steppers freak out. It's looking like 1lb could be achieved, but I'll be doing a payload measurement in Part 4 video in the future. Here's Part 3 now. ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
I think either one is fine. I used PETG in the latest design, but I'm going back to PLA. I still find ABS kind of hard to work with. The strength of the plastic isn't the issue, it's the strength of the stepper motor that's the limiting factor on the actuators.
Cool, what's your agriculture robotics project? I'd love to see more of an open source community sprout (pun intended) around this design so we all benefit.
Backlash is minimal given how tight these gears are fitted. Those airsoft BB bearings don't move laterally at all, so they're pretty decent. You'd be surprised how little backlash.
@0212 Nice job but why don't you adjust the 3D model to create a counter-sink for the screws. I don't see an issue here since these were printed in-house.
Yes, I agree. That base should be raised by about 0.5mm or even 1mm to sink the M3 screws further. The downside is the sun gear height is exactly the shaft gear height to get full contact, so you'd end up with the sun gear 0.5 or 1mm higher, but that probably is an ok tradeoff.
@@JohnLauerGplus Nope, no trade off at all. shit the base of the PG system by making it a little thicker to compensate for the inner nut cover. that way everything inside stays the same length.
Hi there! You cannot imagine how this kind of things makes me happy. Keep going that way! I just bought 3d printer and started with Drone( arduino) I would like see more of this project. I would like know which CPU you are using and how you program it. Best regards!
That makes more sense. I've always had really good luck Amazon filament. For me, around 200 maybe 210 if it's been out, it goes down nice. Have you ran a PID tune?
i didnt get the reason y u used planetary gear train at each joint, is it to have sufficient torque( to take up the further assembly loads) and speed reduction ?
You are correct in your assumption. It creates the torque you need. The extra benefit is you get better accuracy too because the steppers are 48 steps per rotation, but now you control 48*38.6 positions per rotation.
For the high detailed planetary gears you say you used a 0.3mm layer height, why not go with something like a 0.1mm? It would take longer but it seems like it'd be much less sanding to do, right?
No, I used a 0.3mm nozzle with 0.06mm layer height. Creality Ender 3 magic numbers are 0.06 for full steps so I use that and they end up coming out with near perfection. I barely have to sand/file anymore. Use the latest anti-backlash forgiving design for the small actuators. I haven't made a video of those yet. www.thingiverse.com/thing:3566678
I use Amazon pla which seems less soft compared to hatchbox pla, but find I have to extrude it at 225 for pla to get perfect extrusion on the 0.3mm nozzle at that layer height.
Yeah, you can use any motor you want, but you'd end up redesigning the 3d parts a bit for fit. If you do try out a NEMA 14 and rework any of the CAD, could you please share it because even I would like to see some stronger motors on these actuators so we have choice. The stepper motor driver in this design is 2A capable so the electronics are ready to go for bigger motors.
No encoders. You can very readily rely on stepper motor steps to execute steps accurately without needing encoders, just like on a CNC machine. However, there will be joysticks to train the robot and record the step position to generate Gcode.
@@JohnLauerGplus Then you going to need end stops to have a known position to start at (like CNC machines). Do you have those? If you get skips (too much torque) then steppers can skip and you will get off on your assumed position.
@@MakunaRGBIC I got the endstop hall sensor working in Part 3. Super cool. I am still considering adding an encoder though, but can't find one that would fit the actuator design/format. Any ideas? I would want to encode the reduced output, not the stepper motor input.
hi, can you show me the algorithm flowchart? I follow your project slow enough but I still cant understand the whole process of the robot and how the robot works?
Take a look at this thread in the chat group. 1.2lbs for the small actuator. forum.makerforums.info/t/1-2-lbs-of-lifting-power-out-of-small-actuator/78653/4
I think this could achieve about 1lb, but haven't tested that far yet. Still working on electronics to see how powerful the steppers can be on these actuators.
So far the arm is like $95. No, doing this in metal would be very hard given the intricate details. Maybe 3D metal printing could work, but the weakest point isn't the plastic, rather it's the stepper motors. So the material doesn't matter as much as you think. Going stronger would require more gear ratios. Harmonic drive would be amazing if we could all figure out how to print a harmonic drive.
That's going to take a while to get all of that going. You will be able to send Gcode to the robot arm from ChiliPeppr from a new workspace, so you'll be able to visually see the robot arm move in the 3D viewer. A typical Gcode command would look like "G1 F200 X2000 Y3000 Z500 A234 B111 C222 D300" where G1 is a move that supports feeedrate which is the F value. XYZ are the base, shoulder, and elbow so going from bottom of robot up. ABC are the 3 wrists from base towards gripper. D is the gripper.
Something is wrong with Thingiverse right now. I just logged into my account there and all of my makes are gone. I have to imagine this is just temporary.
That's bad, i really like the arm, i want to make some modifications for other application i am working on. Is there any other way i could get the models.@@JohnLauerGplus
There will be 2 power wires going to each joint. They're not wired up yet. I may end up with a sync signal wire as well unless I'm happy with the Bluetooth or WiFi command sending being close enough on timing.
BTW, check out my new approach of doing a slip ring in the actuator for the 20v power so this turns out to be a zero wire robot. I'm not perfectly happy with it yet because I get some blips on the power, but it may pan out if I can get the design tweaked. ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
Watch the part 1 video to see how they were constructed. Also, I did not design the CAD on the planetary gears, rather Jeff Kerr (LoboCNC) on Thingiverse did. I'm not sure which CAD program he used for the helical gears, but they came out awesome.
@@3dsafety247 PLease post some video if/when you get your build. I would love to see how others are getting their own version assembled. Did you get to see Part 2 and Part 3 yet?
Your audio quality is fantastic, it feels like you're talking to me in person with headphones on!
You realise I got a better mic for £30?
You should do a final episode! It could help with maybe students who want to do robotics in the future (like me). Videos like that with the simplified robot would be much appreciated by not only me but a lot of people. Nonetheless, we thank you for keeping us updated and helping us with all the videos. Keep up the good work!
Yeah, but a final video on this would take it to like Part 30. So this will take years.
Nice work John! I've been working on a multi-axis arm with planetary gears also and since I have no idea how these joints are supposed to be designed I've gone through a lot of trial and error. Taking a look at your design is really helping me improve my design (or pretty much copy yours) so thank you for sharing this
Guess credit for the design is Jeff, but your contribution helps immensely. Thanks again!
@@jaredgray7872 Check out the latest flex spline design of the actuator I'm now using. It's even better. ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
Your robot arm inspired me to use water tube as the casing for my own robot arm.
Nice idea with the “mismatched” planetary gears. Now, you could add a potentiometer to register the absolute angle.
👍👍👍
Very nice work. Looking forward to seeing how it works out. Super tricky project, so well done for tackling it.
Fun project. Excited to see it running.
*eyes glued to the screen!*
I love technical builds like this :D
You should check out latest part of the build ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
Please show how it works, it would be nice.
Working on it +Mikael! The next step now that the actuators are assembled is to get the electronics going. Using ESP32's in each joint to generate stepper pulses and then will use ChiliPeppr's Cayenn protocol to upload all Gcode moves to each joint. Stay tuned.
@@JohnLauerGplus thanks for the answers keep up the good work. Very interesting content. 👍
Wow..... this is really good job. I hope your effort easier access to robot technology for anyone around the globe. #globalpride
@@mohdyuzairin8719 Thank you.
@@JamecBond not if you want a touch sensor in each actuator for record and playback. Also, they are $3 each. That's so cheap you can sprinkle them everywhere. Also, running wires is a serious problem and this avoids that.
@03:35 You can simply create a press fit cover for the inside which will in turn hold all the nuts in place. Falling nuts into a PG system is no fun and will not be an issue if the cover is placed.
Good idea. It would add some height to the system, but probably could make the cover be 0.6mm height, so wouldn't add much. Those planetary gears do rub along the top where those nuts are. The top sun gear does push itself up towards the roof of the gearbox. So do the planetary gears, so the smoother the roof, the better. A cover could help that be even smoother than having those nuts exposed. As I've seen the grease squeezes up into the nut holes over time. Although, I do think it makes sense to have some gunk repositories where over time the shreds of plastic can bunch up and get pushed into depressed areas to help give longevity.
Great work, can't wait to see it move!
ive heard some bad things about mixed vaseline and plastics, it supposedly breaks the plastic down. I personally use white lithium grease. not sure if its true, but better safe than sorry. great video btw
I was worried about that. I searched online and couldn't find much data on the correct lubricant. If vaseline breaks down the plastic, then indeed something else should get used. I'll try the white lithium grease like you suggest.
@@JohnLauerGplus I definitely second the suggestion of white lithium grease. You'll commonly find it used in joysticks and things because it's plastic-safe. 👍
I thought the same.
I believe it depends on the type of plastic, some plastics react to it and some don't. Would depend on what type of filament he used for the print.
Can't go wrong with silicone grease. Used in distillation lab equipment because of how little it reacts with other chemicals.
Verry good design of bearings
Thanks for doing this. I am printing an actuator as we speak.
Check out the new anti-backlash 50% better torque small stepper motor actuator at www.thingiverse.com/thing:3566678
This is amazingly inventive in the use of materials when considering the potential accuracy of the end product, have you tried making a new arm with the arm, that'd make for a good investment.
Nice one. Had that exact gearing idea in my head a few years ago for a 'planned' robot idea that I never found time for. Nice to see it's out in the wild via the makerspace hive mind. Hoping to share another gearing design soon that will be better still for robotics (true zero backlash, high reduction, and very high load capabilities)
Ooh. Tell me more. It would be nice to have an even stronger gear reduction box. This one is decent, but the robot arm could be much stronger. Do you have any links to any designs? Or videos you've done? Or concepts?
@@JohnLauerGplus not yet but hopefully will have something to share in a few months or so. I'll shoot you a link once I'm ready to share it with the world :) probably via a Kickstarter if prototyping goes as planned.
@@JohnLauerGplus in the meantime check out Jeff's new design ua-cam.com/video/sdRGrTHq4hA/v-deo.html
the high torque nema 17 is not meant to be on rotation, its the medium torque in there... the high torque is on the shoulder, as it has to support all of the weight of the arm...
Yes, i think i messed up putting the larger motor in the base. I caught that too. Good eye.
Beautiful, I'm quite curious about the next video where you hook up the bluetooth and use it!! Great design, nice to also see the small fans inside the housing. Nice details of your work.
Nicely done, keeping an eye on this one
Now all we gotta do is get a ROS package going for this!
Totally agree
Totally agree
Sir this is the very nice project and I want to implement it too but I'm unable to find the STL of all actuators ... Could you please help me please 🥺
Adjust the current , that screaming is due to too high set voltage regulator on the stepper driver. The stepper motors shall be dead silent
I just wanna point out that metric isn't the 'english' measurement, mostly everyone in the world uses it except for US. Imperial is actually the 'english' measurement that they moved out of which is 'inches, feet, etc.'
Yeah that's what he said, the model is in english "imperial" however the nema holes are in metric
You don't sand gears, you "lap in" gears. Apply a fine lapping compound to all the gears and run them until they get quiet. The lapping compound is basically grease and sanding media and the teeth of the gears will rapidly wear away the spots that interfere on the gear teeth. This has been the acceptable practice since modern gears were invented.
Interesting. I am working on a BLDC version of the actuator now. I'm printing the new compound planetary gear in nylon and indeed I am breaking in the gears by running them raw against each other and it is smoothing them out. However, what lapping compound would you use for plastic to make this process faster/better? I can't find any when you search google for "3d printed gear lapping compound".
@@JohnLauerGplus Loctite has a brand called "Clover." I'd say a nice 200-300 grit would do it. The Lower the grit the more abasive and faster, but finish is not as pretty... higher grits will give a smother, prettier finish, but take longer.
After the gears quieten, remove and clean them thoroughly. Add real grease and they will be perfectly mated!
THIS is old-school gear perfection!
Why not use servo motors and forget the gears?
That wooden table is so nice
That's so funny. Although the video is all about the robot arm, I do think the table behind it helps glorify the whole thing.
I'm glad I've found your channel!
The bb's seem pretty annoying to deal with. Have you tried just printing spheres in place? I did that with some print in place bearings I made and they work very well.
That's a cool idea, but those need to be really smooth, so that would worry me you'll never get the smoothness and roundness that you need. Albeit you are sort of averaging them together, so maybe it "mostly" works?
I will say in subsequent designs I just went with cheap off-the-shelf metal bearings because they're worth the simplicity, smoothness, and tightness beyond a 3D printed bearing and they're not that expensive on the low end, but 10 times better than a 3D print bearing.
Cool project, looking forward to next video!
maybe you could fine adjust the extrusion multiplier so that you don't end up with a lot of sending.
Looks interesting! What about just plain sliding bearings?
I think it would be interesting to look at options for bearings, but the way these actuators are all integrated is very nice and trying to get an off-the-shelf bearing in there may not be as elegant. However, I'd love to see if anyone could CAD out a potential design.
Well, I just ment to use plastics sliding against each other. Without any real bearings.
@@CinemaTablet I don't think that would be smooth or friction-free or tight enough. Ultimately you do have plastic against plastic since those airsoft BB's are plastic.
after you break in the gears, you need to wash the plastic dust out of them as it will act as an abrasive along with the Vaseline breaking down your plastic. I would invest in some rubiks cube lubricant from cubicle labs.
silicone based would be best
It looks very nice, but I would probably have used steel balls for the bearing. Why did you choose to go with plastic pellets?
I don't think steel balls would work much better, but perhaps. Didn't try it, but I think steel against plastic may wear in an odd way. Plastic against plastic seems to wear nicely. Also, these bearings turn out really tight yet smooth rotation, so that's not the part of the design I'm worried about. What I don't like is the power of the stepper motor.
Very cool. For the bbs, (or however it's spelled), I wonder if printing a separate assembly - like a self-contained thrust bearing - might solve the problem.
Perfect! I realy appreciate your work. Keep uploading!
I always found PLA was prone to cracking, I now exclusively use PETG ... just a thought
I am now using PETG on my latest design as you can see in Part 3 ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html, but I am hearing robot guys using 3D printing like the flex on PLA better for gears. I'm not quite sure yet.
not bad but for 3d printing you make hard robot arm. With 3 motors can be done. because you dont use 4 or 5 axis. you print in 2 dimension X and Y(vertical) and Z is leveling. you cant print on X and Z plane(horizontal).
EXAMPLE of ROBOT ARM is: Human arm. Entire body is the Z axis from the shoulder start the arm and use it in horizontal way.The elbow is where 3rd motor will stay. You will not have specific X and y motor axis. Motor X and Y must work toghether to make lines and curve. Z motor axis is independent.
Awsome voiceover
Why did you not put the drivers and boards in the bottom and wire it through? Great job btw.
Because wiring robots is a huge pain and I need/want a temp sensor in each actuator, fan controller, hall sensor for end stop, ws2812 LED for status, and touch sensor for local jogging and programming. We now know how to run signals over the air, rather than needing copper, so can eliminate many wires that way. ESP32's are dirt cheap and a novel way to go these days. Yes, it's all theory until it comes together, but might as well give it a go.
@@JohnLauerGplus thanks man
@@williamalkhoury3417 Part 3 shows you it all coming together--thus answering your original question. ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
In my experience, Zyltech's PLA is amazing. even Uncle Jessy suggests it.
GREAT Stuff !.. and a lot of great info on building it. However, I noticed you didn't put ( or at least I didn't see?) you put vaseline around the airsoft BB's?? Also.. I built one a few years ago, similar to this. but I have never seen anyone use vaseline? I'm curious how long it last? as I used a tube of Lithium White grease that works really really slick and well.. still lasts after a few years!! But cost a lot more than a jar of vaseline . LOL.. so if it works as good. .I'll try vaseline next time !!
Very nice! Encoders? Limit switches? Payload max? Movement speed?
Latest design now has a hall effect endstop. I'm getting about 500 Hz PWM signal to stepper driver before these 35mm steppers freak out. It's looking like 1lb could be achieved, but I'll be doing a payload measurement in Part 4 video in the future. Here's Part 3 now. ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
Amazing videos 🎉 are you using PLA?
Yes, that was all PLA because I only had an Ender 3 at the time so that was my goto.
Would you recomend to use pla or abs, because normaly abs should be stronger.
I think either one is fine. I used PETG in the latest design, but I'm going back to PLA. I still find ABS kind of hard to work with. The strength of the plastic isn't the issue, it's the strength of the stepper motor that's the limiting factor on the actuators.
Ok, I will try with PLA. Trank u for your answer.
Nice prints. Thank you!
Absolutely awesome, can't wait to see more! (Especially the control interface!)
Oh damn, I wish I have some free time to try to make your arm, nice job.
This is actually genius! And controlling it via BT makes it modular too!
Great video for me. I'm working on a robotics project for precision agriculture and received pretty good ideas. Thank you.
Cool, what's your agriculture robotics project? I'd love to see more of an open source community sprout (pun intended) around this design so we all benefit.
@@JohnLauerGplus I'm looking at a "something" to select and pick up fruits. Initially.
If possible, I would like to share in private. May be through Linkedin. Thank you.
This is amazing!!. So clean and robust. What is the load it can carry? Any tests carried on that?
It can carry 1 pound.
@@JohnLauerGplus what would be the possible changes for 4 pound payload.? Motors and higher reduction ratios I guess?
@@sharath_naik yes, I think that would be the trick but you may have to move to metal to do that.
Absolutely awesome work, what abouth backlash? Thanks for sharing
Backlash is minimal given how tight these gears are fitted. Those airsoft BB bearings don't move laterally at all, so they're pretty decent. You'd be surprised how little backlash.
And now the latest design has an anti-backlash flex spline in the actuator to help solve backlash. ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
@0212 Nice job but why don't you adjust the 3D model to create a counter-sink for the screws. I don't see an issue here since these were printed in-house.
Yes, I agree. That base should be raised by about 0.5mm or even 1mm to sink the M3 screws further. The downside is the sun gear height is exactly the shaft gear height to get full contact, so you'd end up with the sun gear 0.5 or 1mm higher, but that probably is an ok tradeoff.
@@JohnLauerGplus Nope, no trade off at all. shit the base of the PG system by making it a little thicker to compensate for the inner nut cover. that way everything inside stays the same length.
Hi there! You cannot imagine how this kind of things makes me happy. Keep going that way! I just bought 3d printer and started with Drone( arduino)
I would like see more of this project. I would like know which CPU you are using and how you program it.
Best regards!
Then you should watch part 2 and part 3!
John Lauer thanks John! Subscribed!
3:21
Nice
🤣 i thought the same and ... 3:33
That's a lot of work........Good job 👍
Hello
I have printed gearboxes
But I have a problem that the gears are large and cannot be installed in the gearbox
Print the gears at 99% or even 98% then.
@@JohnLauerGplus Thank you
What is the output torque with the planetary gearbox included?
She Looks Nice!
You should do an ASMR 3D printing! Voice is very soothing. All love from Pakistan!!! 😃
Wonder if you can make same with one stepper and air pushing from set stepper to get same movement less power parts?
Nicely done!
That makes more sense.
I've always had really good luck Amazon filament. For me, around 200 maybe 210 if it's been out, it goes down nice.
Have you ran a PID tune?
i didnt get the reason y u used planetary gear train at each joint, is it to have sufficient torque( to take up the further assembly loads) and speed reduction ?
You are correct in your assumption. It creates the torque you need. The extra benefit is you get better accuracy too because the steppers are 48 steps per rotation, but now you control 48*38.6 positions per rotation.
@@JohnLauerGplus thanks for your timely reply
Wow, beautifully engineered. Great design and walk though as well. 👍
Thanks!
For the high detailed planetary gears you say you used a 0.3mm layer height, why not go with something like a 0.1mm? It would take longer but it seems like it'd be much less sanding to do, right?
No, I used a 0.3mm nozzle with 0.06mm layer height. Creality Ender 3 magic numbers are 0.06 for full steps so I use that and they end up coming out with near perfection. I barely have to sand/file anymore. Use the latest anti-backlash forgiving design for the small actuators. I haven't made a video of those yet. www.thingiverse.com/thing:3566678
@@JohnLauerGplus What slicing program do you use?
@@JohnLauerGplus thats impressive, fine prints below .1 give me some nasty whiskers, even with retraction set to 5mm and 200°
@@Hotdog80085 latest cura release
I use Amazon pla which seems less soft compared to hatchbox pla, but find I have to extrude it at 225 for pla to get perfect extrusion on the 0.3mm nozzle at that layer height.
cannot open the steppers web links, is there a new BOM list?
How much is the payload Wright?
Is it possible to use a NEMA 14 Bipolar 200 Steps Rev 35×28mm 10V 0.5 A Phase stepper motor for the small acuators?
Yeah, you can use any motor you want, but you'd end up redesigning the 3d parts a bit for fit. If you do try out a NEMA 14 and rework any of the CAD, could you please share it because even I would like to see some stronger motors on these actuators so we have choice. The stepper motor driver in this design is 2A capable so the electronics are ready to go for bigger motors.
I thought I spotted a encoder in one scene, but you didn't talk about if you were using them or not.
No encoders. You can very readily rely on stepper motor steps to execute steps accurately without needing encoders, just like on a CNC machine. However, there will be joysticks to train the robot and record the step position to generate Gcode.
@@JohnLauerGplus Then you going to need end stops to have a known position to start at (like CNC machines). Do you have those?
If you get skips (too much torque) then steppers can skip and you will get off on your assumed position.
@@MakunaRGBIC Yes, looking at using magnet hall sensors as endstops.
@@MakunaRGBIC I got the endstop hall sensor working in Part 3. Super cool. I am still considering adding an encoder though, but can't find one that would fit the actuator design/format. Any ideas? I would want to encode the reduced output, not the stepper motor input.
So impressive. Thanks for sharing!
Excellent video. Which 3D did you use and/or recommend?
hi, can you show me the algorithm flowchart? I follow your project slow enough but I still cant understand the whole process of the robot and how the robot works?
Beautiful Project. Congratulations.
Did you test, how much the robot arm can uplift (approximately) ?
Take a look at this thread in the chat group. 1.2lbs for the small actuator. forum.makerforums.info/t/1-2-lbs-of-lifting-power-out-of-small-actuator/78653/4
This is so bad ass!
Great job with the build, looks really neat.
Thanks for sharing and as always Keep Building👍
What is the maximum payload for it
You should try casting metal
i don't think casting would be precise enough
may i ask you 3d CAD blueprint file?
What payload are you trying to achieve?
I think this could achieve about 1lb, but haven't tested that far yet. Still working on electronics to see how powerful the steppers can be on these actuators.
How long does it take for you to print all the parts and what printer did you used to print those????
10 days and Creality Ender 3
What is the cost for this arm until this moment? Do you have in plan to do a “metal” version?
So far the arm is like $95. No, doing this in metal would be very hard given the intricate details. Maybe 3D metal printing could work, but the weakest point isn't the plastic, rather it's the stepper motors. So the material doesn't matter as much as you think. Going stronger would require more gear ratios. Harmonic drive would be amazing if we could all figure out how to print a harmonic drive.
its look amazing ... nice work ... you have public models ?
They're in the description.
Nice one mate
How much was your 3D printer? And would you advise buying a 2nd hand ?
hello,
It's an impressive design. can you also do a video on inverse kinematics...
That's going to take a while to get all of that going. You will be able to send Gcode to the robot arm from ChiliPeppr from a new workspace, so you'll be able to visually see the robot arm move in the 3D viewer. A typical Gcode command would look like "G1 F200 X2000 Y3000 Z500 A234 B111 C222 D300" where G1 is a move that supports feeedrate which is the F value. XYZ are the base, shoulder, and elbow so going from bottom of robot up. ABC are the 3 wrists from base towards gripper. D is the gripper.
Hello, where can i get the 3d models of the your arm. I tried the link in the description, it is throwing error.
Something is wrong with Thingiverse right now. I just logged into my account there and all of my makes are gone. I have to imagine this is just temporary.
That's bad, i really like the arm, i want to make some modifications for other application i am working on. Is there any other way i could get the models.@@JohnLauerGplus
@@Balajireddypanta Looks like Jeff Kerr still has it on Printables. www.printables.com/model/132260-we-r24-six-axis-robot-arm
It's a really cool project, but the shaky camera drives me crazy!!
Very impressive, and looking forward to your updates. Probably me being an idiot, but how is power/signal fed through each joint?
There will be 2 power wires going to each joint. They're not wired up yet. I may end up with a sync signal wire as well unless I'm happy with the Bluetooth or WiFi command sending being close enough on timing.
BTW, check out my new approach of doing a slip ring in the actuator for the 20v power so this turns out to be a zero wire robot. I'm not perfectly happy with it yet because I get some blips on the power, but it may pan out if I can get the design tweaked. ua-cam.com/video/4o3d7_WZ_DQ/v-deo.html
Why are so many people using stepper motors for their robot arms instead of brushless DC?
Because they have holding power and can step.
Can you share the GCode for the #D printed parts?
They're all on Thingiverse in the description.
amazing work, you should print a tripode for your camera ahaha.
what if you connect the ARMs direct to the NEMA17 Axis ? Why do you need all these gears? isnt Nema strong enough and you need reductions?
Stepper motors are not strong enough.
How much weight it can lift...?
Great job!! Has you finish it?
Can someone explain me which is the final ratio used for the small joint and for the medium joint? Thank you :)
Best to check out all of the Thingiverse links. Each item has the ratio in the description.
How did you contruct the planetary gears? I mean besides cad the maths?
Watch the part 1 video to see how they were constructed. Also, I did not design the CAD on the planetary gears, rather Jeff Kerr (LoboCNC) on Thingiverse did. I'm not sure which CAD program he used for the helical gears, but they came out awesome.
What type of 3d printer you used?
How many DOF or axis is this arm?
can you maybe link all the electronics here i relly wanna print that robot arm. relly nice projekt thats insane
how much weight it can handle it ? Could it carry 0,5 kg without any failure issue ?
Yes, it should be able to handle just about 0.5kg, but it is still untested.
@@JohnLauerGplus What is the Gear Ratio for large actuator ? I wanna use "Step Motor Nema 23 2.7 Nm" motor for large actuator
@@3dsafety247 38.4:1 is the ratio. The Nema 23 should work quite well.
@@JohnLauerGplus Thank you for all informations. You inspired many people with this project (incluiding me). I will share when I finished assembly.
@@3dsafety247 PLease post some video if/when you get your build. I would love to see how others are getting their own version assembled. Did you get to see Part 2 and Part 3 yet?