Wowza! I think this might be my favorite cycloidal drive I have seen so far! Great job!! I really like that you used the 3mm rods and bearings as well as supporting the output shaft radially and axially! Great job!!
Hi Thanks for the kind words! I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Yea i just happened to have a bunch of dowel pins sitting around and decided to use them when i was designing the drive. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Agree, it had almost zero backlash and cycloidal drives are much easier to manufacture by CNC milling. Btw, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Very nice. My latest cycloid iteration uses dowel pins too. But instead of pushing then into the plastic, I use sleeve bushings (3mm ID 5mm OD 3mm length) so they can spin freely. That makes the spacing a bit wider, so I can't get the high you got.
Hi if you are interested in trying out this cycloidal drive, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Very nice results. Good idea to put non-3d printed friction surfaces for more durability and quieter, smoother output. I myself like to use PTFE or HDPE tubes rollers in my 3d printed slew bearing and people cannot believe they are 3d-printed because of how smooth they are.
Thanks! Actually if you look closer into the cycloidal disk drawings. It is actually designed for "Contact-rolling" just like a spur gear. So there is no actual sliding in theory which should help with the wear and tear. You can read more on my website if you are interested!.
@@gedr7664 PTFE tube 4mm OD from 3d printer boden tube, or 6mm OD from the thicker 2.85mm filament boden tubes, readily available, some very cheap. Alternatively, even cheaper are HDPE tubes, from refrigerator water line or water filter water line, usually are 6mm. I can usually make one 6 inch diameter slew bearing that can easily handle 10KGF dynamic load for a robotic application for about $2. Nylon or Delrin are also options but are higher friction than HDPE, though PTFE is the lowest friction. But beware that your printer must be tuned accurately otherwise be ready for lot of hand machining and smoothing time. I spent days tuning and reprinting, then once I got the dimensions exactly, I can print bearings with minimal machining and almost as smooth as metal bearings. If you want to go even stronger you can go for solid core PTFE/HDPE rods or metal rods but it will cost lot more as you have to get from special industrial stores. Also make sure for cutting the roller, you make a tool so they are perfect cuts without imperfect edges, otherwise the bearing won't be smooth. Everything has to be tuned if you want the super smooth bearings.
@@thewhiteowls1320 Yes I know, cycloidal are contact rollers, but there is sliding, though it seems as if they are not sliding. Just like in gears, the overall gear is not sliding, but the teeth themselves are sliding against each others. That's why many put small bearings instead of rollers, but like you did metal rollers are good alternative.
@@preddy09 i see looks like quite a fair amount of work to get the roller dimensions right. I usually add a 0.15mm offset. how about yours? What material do you print your roller in? I am kinda hoping for something similar to delrin/acetal...
How is the durability? Plastic gear over roller pins is probably ok with some grease, but the plastic over plastic output pins could wear out. Really nice design btw.
If i arent wrong, i believe you meant those bronze nuts that can be pressed into 3Dprinted PLA? yea that sound like a really good idea and a much faster process. Btw, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Hi, thanks for your great video! when does a cycloid drive use pins, and when rollers on the outside ring gear? I see you have used pins on the outside, and others use rollers. Cheers!
Thanks for the kind thoughts. The wobbling disk has a cycloidal profile which will contact and roll around the pin rather than sliding past (just like a spur gear). Hence a fixed pin work just fine. I guess others aren't aware about this.
@@pahafico2 i am not sure if i understood your questions. Mainly the virtue of a cycloidal profile is that is was create from a epicycloid which involves a small circle rolling on the surface of a bigger ones. Hence, the cycloidal profile created also allow it to contact and roll around the pin rather than sliding. Perhaps manufacturing limits meant that the real profile are slightly off from the cycloidal profile and may have slide slightly on the rods but it should be minor and decent bearings are rather expensive, kinda hard for me to buy 49 pieces of decent bearing per drive.
thanks for the kind words. I have just bought a weighing machine, hopefully i can use it for some torque testing soon. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Precision, stepper motors at full stepping modes do 200 steps/revolution In a 48:1 reduction ratio that's 9600 steps per revolution WITHOUT microstepping, so this makes it far, far more detailed
Thats a very good question, it made me think alot. I agree that stepper motors arent the best choice for robotics because of its weight. Ideally, i should have used BLDC. However, stepper motor have a much higher holding torque and it is also much easier to control. For stepper motor, i only had to count the number of steps to take while BLDC requires additional sensors. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/ And great question! Hope i had answered them well.
@@sus7382 Yeap thats one advantage too! you seems very experienced with stepper motors. Currently my 3dprinted drives have some backlash so it is not quite able to do 9600 steps per revolution that well. Or at least I have yet to track.
@@thewhiteowls1320 heh, thanks, I plan on building my own 6 axis arm soon and as such I've learnt a lot about steppers, and as you mentioned before they are quite heavy, but only in some areas. You can offset weight for robotic applications and use belts to drive the power of the motor to a different location without sacrificing strength and accuracy, but by gaining some mechanical complexity.
@@sus7382 Very good points on keeping motor and drives near axis of rotation to minimize moment while transmitting the torque forward to where it is needed. I totally agree with u on mechanical complexity, it is really hard to design mechanism that is effective, light-weight, easy to manufacture and assembly. I had completed a design on my own 6dof robotic arm. you can check it out on my instagram @ewhiteowls
Hi, there is nothing special. Just some simply denatured ethanol that I bought from hardware shops. The main purpose of the ethanol is to keep the drill bits cool as PLS warp easily and could not hold the threads well if the drill bits were too hot.
HI Thanks for your kind words. I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/ I look forward to what this community will do with my design. hope to hear from you soon.
Thanks for your kinds word, if you are interested, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
I believe not, since the gear stays relatively stationary to the axle because of the reduction. Another way to see it: The input shaft would just spin freely in the cycloidal gear, not forcing it to walk along the outer surface.
Hey Aniket, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/ I really look forward to what you will do with my design! keep in the loop
Sadly i have not yet done anything to test. So far just with my hands i dont feel any back lash. Cant say the same when the drives are loaded. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
@@thewhiteowls1320 I write some comments under your blog , I have questions on your tourital , about "Step 5 : Offset the spline" , the orignal spline and the offseted spline are totally different.
@@wei1224hf yea... the offset is offset towards the center so it naturally have a different curve. perhaps you can send me a screenshot of your attempt?
HI Sam, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
@@allip673 Yes my bad. didnt check before i published. I have wrote an article on this drive and updated the items list. you can find out more here www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Excellent demo- exxplain of harmonic drive assembly with motor
Kami
Cycloidal Drive, not Harmonic Drive
Wowza! I think this might be my favorite cycloidal drive I have seen so far! Great job!! I really like that you used the 3mm rods and bearings as well as supporting the output shaft radially and axially! Great job!!
Hi Thanks for the kind words! I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
There is definitely an output, its the white circle that surrounds the stepper motor input shaft.
nice use of dowel pins, I currently use em to make crossed roller slew bearing
Yea i just happened to have a bunch of dowel pins sitting around and decided to use them when i was designing the drive. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Wave reducers have undeniable advantages.
Agree, it had almost zero backlash and cycloidal drives are much easier to manufacture by CNC milling. Btw, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Very nice. My latest cycloid iteration uses dowel pins too. But instead of pushing then into the plastic, I use sleeve bushings (3mm ID 5mm OD 3mm length) so they can spin freely. That makes the spacing a bit wider, so I can't get the high you got.
i see... thats a very good technique to minimize space. Typical bearings are at least 8mm in diameter.
Hi if you are interested in trying out this cycloidal drive, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Very nice results. Good idea to put non-3d printed friction surfaces for more durability and quieter, smoother output. I myself like to use PTFE or HDPE tubes rollers in my 3d printed slew bearing and people cannot believe they are 3d-printed because of how smooth they are.
can I ask where you purchase them?
Thanks! Actually if you look closer into the cycloidal disk drawings. It is actually designed for "Contact-rolling" just like a spur gear. So there is no actual sliding in theory which should help with the wear and tear. You can read more on my website if you are interested!.
@@gedr7664 PTFE tube 4mm OD from 3d printer boden tube, or 6mm OD from the thicker 2.85mm filament boden tubes, readily available, some very cheap.
Alternatively, even cheaper are HDPE tubes, from refrigerator water line or water filter water line, usually are 6mm. I can usually make one 6 inch diameter slew bearing that can easily handle 10KGF dynamic load for a robotic application for about $2.
Nylon or Delrin are also options but are higher friction than HDPE, though PTFE is the lowest friction.
But beware that your printer must be tuned accurately otherwise be ready for lot of hand machining and smoothing time. I spent days tuning and reprinting, then once I got the dimensions exactly, I can print bearings with minimal machining and almost as smooth as metal bearings.
If you want to go even stronger you can go for solid core PTFE/HDPE rods or metal rods but it will cost lot more as you have to get from special industrial stores.
Also make sure for cutting the roller, you make a tool so they are perfect cuts without imperfect edges, otherwise the bearing won't be smooth. Everything has to be tuned if you want the super smooth bearings.
@@thewhiteowls1320 Yes I know, cycloidal are contact rollers, but there is sliding, though it seems as if they are not sliding. Just like in gears, the overall gear is not sliding, but the teeth themselves are sliding against each others. That's why many put small bearings instead of rollers, but like you did metal rollers are good alternative.
@@preddy09 i see looks like quite a fair amount of work to get the roller dimensions right. I usually add a 0.15mm offset. how about yours? What material do you print your roller in? I am kinda hoping for something similar to delrin/acetal...
How is the durability? Plastic gear over roller pins is probably ok with some grease, but the plastic over plastic output pins could wear out. Really nice design btw.
Great work! Is this easily backdrivable?
both the design and tutorial and clear and concise, great job! can you back drive it with your hand? have you tested the torque output?
Have you looked into heat set inserts instead of cutting threads into the plastic?
If i arent wrong, i believe you meant those bronze nuts that can be pressed into 3Dprinted PLA? yea that sound like a really good idea and a much faster process. Btw, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Hi, thanks for your great video! when does a cycloid drive use pins, and when rollers on the outside ring gear? I see you have used pins on the outside, and others use rollers. Cheers!
Thanks for the kind thoughts. The wobbling disk has a cycloidal profile which will contact and roll around the pin rather than sliding past (just like a spur gear). Hence a fixed pin work just fine. I guess others aren't aware about this.
@@pahafico2 i am not sure if i understood your questions. Mainly the virtue of a cycloidal profile is that is was create from a epicycloid which involves a small circle rolling on the surface of a bigger ones. Hence, the cycloidal profile created also allow it to contact and roll around the pin rather than sliding. Perhaps manufacturing limits meant that the real profile are slightly off from the cycloidal profile and may have slide slightly on the rods but it should be minor and decent bearings are rather expensive, kinda hard for me to buy 49 pieces of decent bearing per drive.
Greta design:) have you made any torque testing?
thanks for the kind words. I have just bought a weighing machine, hopefully i can use it for some torque testing soon. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Very cool. I'm a little confused as to why you're using a stepper motor for this application though, can you elaborate?
Precision, stepper motors at full stepping modes do 200 steps/revolution
In a 48:1 reduction ratio that's 9600 steps per revolution WITHOUT microstepping, so this makes it far, far more detailed
Thats a very good question, it made me think alot. I agree that stepper motors arent the best choice for robotics because of its weight. Ideally, i should have used BLDC. However, stepper motor have a much higher holding torque and it is also much easier to control. For stepper motor, i only had to count the number of steps to take while BLDC requires additional sensors. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/ And great question! Hope i had answered them well.
@@sus7382 Yeap thats one advantage too! you seems very experienced with stepper motors. Currently my 3dprinted drives have some backlash so it is not quite able to do 9600 steps per revolution that well. Or at least I have yet to track.
@@thewhiteowls1320 heh, thanks, I plan on building my own 6 axis arm soon and as such I've learnt a lot about steppers, and as you mentioned before they are quite heavy, but only in some areas.
You can offset weight for robotic applications and use belts to drive the power of the motor to a different location without sacrificing strength and accuracy, but by gaining some mechanical complexity.
@@sus7382 Very good points on keeping motor and drives near axis of rotation to minimize moment while transmitting the torque forward to where it is needed. I totally agree with u on mechanical complexity, it is really hard to design mechanism that is effective, light-weight, easy to manufacture and assembly. I had completed a design on my own 6dof robotic arm. you can check it out on my instagram @ewhiteowls
Can u please provide details of the lubricant alcohol u r using while drilling into 3D Printed parts.
Hi, there is nothing special. Just some simply denatured ethanol that I bought from hardware shops. The main purpose of the ethanol is to keep the drill bits cool as PLS warp easily and could not hold the threads well if the drill bits were too hot.
i appreciate ur word on this u should at least accept donations for providing to the community
Thanks for your appreciation. Its okay we can get by without donations
HI Thanks for your kind words. I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/ I look forward to what this community will do with my design. hope to hear from you soon.
Nice
Thanks!
Thanks for your kinds word, if you are interested, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Just wondering, instead of make a input shaft with a offset, can't you just offset the hole in the middle of the cycloidal gear?
I believe not, since the gear stays relatively stationary to the axle because of the reduction. Another way to see it: The input shaft would just spin freely in the cycloidal gear, not forcing it to walk along the outer surface.
Hey, do you have the files available only in stl format? I might try to machine some of the parts like you, step files would be so much better!!
Yea you should be able to download them at my website. I have machined them from HDPE and delrin and you can see the assembly video in my channel
Hey Aniket, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/ I really look forward to what you will do with my design! keep in the loop
How much backlash does it have?
Sadly i have not yet done anything to test. So far just with my hands i dont feel any back lash. Cant say the same when the drives are loaded. BTW, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
cool !
thanks!
@@thewhiteowls1320 I write some comments under your blog , I have questions on your tourital , about "Step 5 : Offset the spline" , the orignal spline and the offseted spline are totally different.
@@wei1224hf yea... the offset is offset towards the center so it naturally have a different curve. perhaps you can send me a screenshot of your attempt?
@@thewhiteowls1320 Trying to send my SolidWorks design screenshots , but don't know your email
@@thewhiteowls1320 how to draw from the offset position? Seems Solidworks don't have the function to offset towards the center
what dowel length do you use?
Hey thanks for the questions! It have 3mm diameter and 8mm in length
@@thewhiteowls1320 You should update the parts list on your website, and maybe provide a direct link from the video.
HI Sam, I have shared the design process and step, stl and fusion360 files on my website at the bottom of my article www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
@@allip673 Yes my bad. didnt check before i published. I have wrote an article on this drive and updated the items list. you can find out more here www.ewhiteowls.com/2022/02/the-ultimate-guide-to-design-cycloidal-drives-the-beating-heart-of-robotic-arms/
Use alcohol to keep parts cool, what an idea! My ass has been havenly saved by you my dear!
hopes that help. it is great when the alcohol evaporates off quickly, leaving you behind some dry dust to vacuum away easily.
Взял винт руками со стола, сунул в пассатижи, потом воткнул в корпус. А сразу руками вставить нельзя было?
Hanzhen harmonic drive gear , machine gear reducer, over 30 years experience , robot gear