It's unbelivable. I was trying to find or build the cycloidal drvies for my robotic arm. But, this is the best cycloidal drvie ever i have seen. Perfect. Thank you for your kind explanation and free publishing
I want to thank you for sharing this knowledge with us, as you can see I am not the only one who wanted to have this knowledge more digestible. I once had a full-time engineer trying to make one and we couldn't do it. Thank you very much.
Yes that will be quite helpful. Currently, I am going at rather slow speed so vibration has yet to be significant. I guess i will need to implement your suggestions when I am doing high speed rotation. Currently, having a single disk seems to distort the torque pin quite badly and lead to significant backlash, i am guess that a second disk would help balance out the forces.
@@thewhiteowls1320 Two cycloidal discs do not cancel all orders of vibration. A three gear system with each gear set at 120 deg from the other provides (almost) perfect balance. Our robotics team has used 3D printed, three gear cycloid drives, multiple times. They are driven with a motor with a free speed of 5676 rpm and produce no vibration. Also, since roughly 1/3 of the 'teeth' of each gear are in contact at any one time, three gears provide almost 100% tooth contact. Finally, we use as many output pins as possible to spread the load from the holes in the printed gears. For example our 48:1 cycloid drive has 24 output pins. Hope this is a help to anyone designing cycloid drives.
Only the one axial bearing for the output won't be enough, you'll need a second ball or tapered roller bearing to account for radial and angular loads. Your design won't be able to stay on axis and have a lot of friction and deflection with radial and angular loads.
Hi techer , Thanks for your guide. I have question, in section 'Design the Cycloidal Disk' step 2, 31/49=0.632, but in your guide set value to 0.622, could you please tell me why ?
@@thewhiteowls1320 Hi Techer, I cannot refer your guide to redreaw the "Cycloidal Disk" on Solidworks , I always hung on step 7, Solidworks cannot offset profile on parameter 1.5mm , maxmium value is 0.7 , however same parameter can be set on Fusion 360, but profile cannot closed, cannot extrute to a cube. Clould you please share a guide on solidworks? Thanks and Best Regards.
Thanks for kind words. It took me several attempts to get to this final design and I learnt alot. I try to document what i learn on my website. Feel free to drop by!
the secondary lines are for drawing the cycloidal profile. In actual mathematics, the cycloidal profile is created by a smaller circle rotating around a bigger circle so the secondary are formed by a small circle rotating around bigger circle.
hi man, very cool design! I'm trying building one from your stl with PLA but i'm having the problem that the cycloidal disk can't move because of lack of enough space from 3mm pins. The disk is stuck. I notice with multiple prints that maybe the holes for the pins are a little too small resulting in bad positioning of pins. Do you have any suggestions? Thankyou!
Hi thanks for liking the design. typically a printed part is slightly oversized around 0.1mm extra on each side. You have to try print some simple sample and measure if the actual size match the designed size and apply offset to your 3dprinter accordingly. My Prusa does it for my automatically so i have no issues, i am not sure about other printer. And in another video of mine, i drilled the hole through 3mm drill bits to ensure it has enough clearance for the pins. hope it helps!
@@thewhiteowls1320 I got the same problem as davidenardi536 experienced. It seems the set of 'cover & disk casing' and 'gear & disk casing' are not matched. cover and gear are unexpectedly lardge and disk casing is small. Thus, i decided to fix the size of the disk casing but, it might be able to affect the whole reducer in a negative way. I am worried but I will also try to fix it :)
@@익명-n5d I increased the size of pins holes and I increased the circle's diameter where they are generated and after several prints I managed to obtain a very good zero backlash result. I think that this could be the best 3d printable drive for NEMA 17 on youtube. I also tried very a promising harmonic drive solution (th-cam.com/video/DlZzknQdTlU/w-d-xo.html&ab_channel=FlexibleBlade) but it is more exprensive because of bearings and more fragile.
Any indication of the efficiency? A similar design in youtube with steel pins instead of bearings has an efficiency of 35% which sounds like a lot of waste. Also do you have any ideas if the wear on cycloid discs would be acceptable? Thanks
I should read your post before apply steel pins to my design instead of bearing. Using PLA as gear material is not a good idea, too soft and friction too large. My gearbox efficiency is 34.9%, which makes my motor can only operate at very low velocity. After I switched to bearing I got 86.5% efficiency, but because I use it to lift heavy load, my gear wear immediately and causing huge backlash.
mathematically, the cycloidal disk does not slide against the pin but roll over it instead, very much like spur gear. It gave me the confidence to stick to dowel pin as long as i am able to manufacture the cycloidal disk as close to the calculated profile as much as possible.
Yes harmonic drives are the gold standards in robotics but they need specialized equipment to manufacture. I am just a home based hobbyist so cycloidal drives is the best I can come out with. I am glad you are back with robotics!
This 3d printed drive with all the expensive hardware in it (giant thrust bearings etc) will end up costing more than an all-metal simple planetary, and close to as much as a metal professional cycloidal. Unless the expensive metal bearing components can be reduced, these things are basically fun but pointless.
That's actually one of the best, if not the best explainer on how cycloidal drives work.
Thank you, I am glad you found it easy to understand. I was concerned the last time that I might over-explain.
So much complexity and precision components. I'm glad you figured out the simpler version to this.
thanks! hope you have found it useful!
It's unbelivable. I was trying to find or build the cycloidal drvies for my robotic arm. But, this is the best cycloidal drvie ever i have seen. Perfect. Thank you for your kind explanation and free publishing
Hi thanks for your kind comments. I look forward to your creations too. feel free to subscribe so you get notified when i post new contents. thanks!
The most simple and compact design I have seen by far.
yes cycloidal drive is stunningly elegant and simple
I want to thank you for sharing this knowledge with us, as you can see I am not the only one who wanted to have this knowledge more digestible. I once had a full-time engineer trying to make one and we couldn't do it. Thank you very much.
The easiest and the best explanation of a cicloidal drive on the whole internet
you should add the second cycloidal disk 180 degree out-of-phase to eliminate vibration.
James Bruton reference?
@@godamkeras4144 SUMITOMO reference.
Yes that will be quite helpful. Currently, I am going at rather slow speed so vibration has yet to be significant. I guess i will need to implement your suggestions when I am doing high speed rotation. Currently, having a single disk seems to distort the torque pin quite badly and lead to significant backlash, i am guess that a second disk would help balance out the forces.
@@thewhiteowls1320 Two cycloidal discs do not cancel all orders of vibration. A three gear system with each gear set at 120 deg from the other provides (almost) perfect balance. Our robotics team has used 3D printed, three gear cycloid drives, multiple times. They are driven with a motor with a free speed of 5676 rpm and produce no vibration. Also, since roughly 1/3 of the 'teeth' of each gear are in contact at any one time, three gears provide almost 100% tooth contact. Finally, we use as many output pins as possible to spread the load from the holes in the printed gears. For example our 48:1 cycloid drive has 24 output pins. Hope this is a help to anyone designing cycloid drives.
Simple and Easy sounds as impropriate words in this case.
In plastic it all works, the plastic is deformed.
But when it’s in metal, the slightest inaccuracy, 0.01, will break the mechanism.
Only the one axial bearing for the output won't be enough, you'll need a second ball or tapered roller bearing to account for radial and angular loads. Your design won't be able to stay on axis and have a lot of friction and deflection with radial and angular loads.
It would be interesting to see if you could laser cut or cnc the large gear piece, that might increase performance a bit.
Yes the large gear piece is cnc milled from delrin. it does help with the accuracy compared to FDM printed ones.
Beautiful gearbox
What's is the weights of the gearbox and Max output torque with 3d printed material
Thanks for the compliments. I am sorry i have yet to test it.
Fantastic video! This looks the perfect size and gearbox ratio for my project. I cannot get on your website though. has it moved?
Stunning easy explanation, thank you!
THANK YOU FOR CONTRIBUTE!
I've been looking for something like this! Thanks for the great video!
thanks, cant wait to see what you will create! lets stay in contact!
Great job on the design and the explanation. Thank you for sharing!
Thanks glad u found it useful!
do u know where can i buy a cycloidal bearing ?
It's very simple and understandable, thank you
Hi techer , Thanks for your guide. I have question, in section 'Design the Cycloidal Disk' step 2, 31/49=0.632, but in your guide set value to 0.622, could you please tell me why ?
my mistake, thanks for pointing out ! it should be 0.632
@@thewhiteowls1320 Hi Techer, I cannot refer your guide to redreaw the "Cycloidal Disk" on Solidworks , I always hung on step 7, Solidworks cannot offset profile on parameter 1.5mm , maxmium value is 0.7 , however same parameter can be set on Fusion 360, but profile cannot closed, cannot extrute to a cube. Clould you please share a guide on solidworks?
Thanks and Best Regards.
May I have your email? I want to send my document to you, If possible could you please guide me and indicate my wrong settings.
@@yilongchen5133 should be contact (at) ewhiteowls.com
The analytical search for the path of least resistance is intelligence.
Great job, i will try to do the same!
Great! looking forward to your creations!
Nice. How many attempts?
Thanks for kind words. It took me several attempts to get to this final design and I learnt alot. I try to document what i learn on my website. Feel free to drop by!
In Design the Cycloidal Disk,what is the length of the secondary lines .which drawn in diagram
the secondary lines are for drawing the cycloidal profile. In actual mathematics, the cycloidal profile is created by a smaller circle rotating around a bigger circle so the secondary are formed by a small circle rotating around bigger circle.
Excellent, thank you!
Great explanation, thank you!
Nice. How do you design the convolution or gear for the pins to mesh all perfectly?
hi man, very cool design! I'm trying building one from your stl with PLA but i'm having the problem that the cycloidal disk can't move because of lack of enough space from 3mm pins. The disk is stuck. I notice with multiple prints that maybe the holes for the pins are a little too small resulting in bad positioning of pins. Do you have any suggestions? Thankyou!
Hi thanks for liking the design. typically a printed part is slightly oversized around 0.1mm extra on each side. You have to try print some simple sample and measure if the actual size match the designed size and apply offset to your 3dprinter accordingly. My Prusa does it for my automatically so i have no issues, i am not sure about other printer. And in another video of mine, i drilled the hole through 3mm drill bits to ensure it has enough clearance for the pins. hope it helps!
@@thewhiteowls1320 thankyou!. I will try! 😀
@@thewhiteowls1320 I got the same problem as davidenardi536 experienced. It seems the set of 'cover & disk casing' and 'gear & disk casing' are not matched. cover and gear are unexpectedly lardge and disk casing is small. Thus, i decided to fix the size of the disk casing but, it might be able to affect the whole reducer in a negative way. I am worried but I will also try to fix it :)
I noticed that the two files, STL files and STEP files, are not same in the same zip file.
@@익명-n5d I increased the size of pins holes and I increased the circle's diameter where they are generated and after several prints I managed to obtain a very good zero backlash result. I think that this could be the best 3d printable drive for NEMA 17 on youtube. I also tried very a promising harmonic drive solution (th-cam.com/video/DlZzknQdTlU/w-d-xo.html&ab_channel=FlexibleBlade) but it is more exprensive because of bearings and more fragile.
Awesome, thank you! 🙏
Any indication of the efficiency? A similar design in youtube with steel pins instead of bearings has an efficiency of 35% which sounds like a lot of waste. Also do you have any ideas if the wear on cycloid discs would be acceptable? Thanks
I should read your post before apply steel pins to my design instead of bearing. Using PLA as gear material is not a good idea, too soft and friction too large. My gearbox efficiency is 34.9%, which makes my motor can only operate at very low velocity. After I switched to bearing I got 86.5% efficiency, but because I use it to lift heavy load, my gear wear immediately and causing huge backlash.
mathematically, the cycloidal disk does not slide against the pin but roll over it instead, very much like spur gear. It gave me the confidence to stick to dowel pin as long as i am able to manufacture the cycloidal disk as close to the calculated profile as much as possible.
Is this system reversible?
Good.
helpful.
can you post the stl files of how to connect with the other parts of the base of the robot please?
:)
seems to me you should have another thrust bearing below the cycloidal disk as well if you are going to have the one on top.
Not necessary.
Is this read by AI text to voice?
Definitely
That cycloidal drive is going to wiggle a lot.
Harmonic drives no longer in use? Sorry, I got no robotic knowledge update since 20 years ago.
Yes harmonic drives are the gold standards in robotics but they need specialized equipment to manufacture. I am just a home based hobbyist so cycloidal drives is the best I can come out with. I am glad you are back with robotics!
@@thewhiteowls1320 3D printer should be your best friend. No more relying on CNC machining.
@@mtay65 haha i kinda like cnc machining tho
I tried hard for this Cycloidal driver but I couldn't succeed :(
This 3d printed drive with all the expensive hardware in it (giant thrust bearings etc) will end up costing more than an all-metal simple planetary, and close to as much as a metal professional cycloidal.
Unless the expensive metal bearing components can be reduced, these things are basically fun but pointless.
Bad audio after minute two. Gets better again later.
so easy a rank amateur like myself can do it
:)
ohh shit not robot voice.