Absolutely amazing, as always ! I’m a software developer and I’m really excited to start creating softwares for your machine, especially for video capture.
This type is my favorite DIY reducer. I wonder how they would stack up against regular harmonic drives if made professionally; if you had a nice cast urethane flex spline with embedded kevlar weave to transfer the torque, I imagine the torque density could be quite competitive, and it should be cheaper to mass produce than notoriously finicky metal splines. Also not the most inaccessible DIY project, to try a cast urethane version.
Also; my intuition would be that given a much more flexy configuration than a metal cup closed on one end, much bigger teeth would be viable and more appropriate for this kind of setup, to get more torque out of the softer materials. Is tooth size something youve tried to optimize for yet?
@@ZeroBacklashRobotics Sure, bigger ratios means smaller teeth. But 30/31 would be the same as 60/62, no? Bigger teeth should mean bigger peak output torque, all else being equal. An open ended spline like this, made from a flexy material like urethane, should be able to accomodate quite big flex amplitudes, accomodating nice and deep teeth, and high torques. Curious how much torque youd be able to get out of teeth of this size!
@Eelco, its not the same, with 30/32 teeth you will get more torque indeed, because your wide teeth can increase with the same perimeter, but your ratio will decrease to 15:1, to get the same ratio increasing teeth there is no other way to increase the diameter size of your splines.
@@ZeroBacklashRobotics I said 30/31 but yeah you need 2 tooth difference with 2 contact points I suppose. Or do you? If the spline had two sections with a one-tooth offset between the sections, that should work I think? Like 30/32 -> 31/33
There is a similar type of gear called ferguson's paradox gear that behave almost exactly like strain wave gear. But without needing to use a flex spline. All three components are solid and fixed. It also have similar high reduction ratio as a strain wave gear. Essentially it's compose of two outer ring gears with one ring that has one more gear teeth than the other align co-axially. Then there is one smaller planet gear nested within that rotate with the input shaft. The two outer ring gears have tooth of slightly different sizes, when the planet gear came into contact with the two outer ring gears, it push the gear teeth together one side and push it away slightly on the other side. The gear ratio of this system is = number_of_teeth_in_ring_gear_1 / (number_of_teeth_in_ring_gear_1 - number_of_teeth_in_ring_gear_2) , so if you have 51 tooth in ring 1 and 52 tooth in ring 2, you get a reduction ratio of 51. Ferugson paradox gear is easier to make than strain wave gear, since there is no need for flex spline. And second advantage is that the gear tooth pressure angle can be lower compare to strain wave, which lead to greater gear efficiency.
that seems to be a variation on a cycloid drive or I am not visualizing this right ok sorry got it. not cycloidal at all but also plays on some tolerances, basically the gears do not mesh completely right.
Hi @Keen Heat, thanks for your comment. I did not know about the existence of this type of gear. It seems very interesting, mainly because having a difference of one tooth between both gears, the ratio is double than strain wave gear, but keep in mind being able to reduce the size of the main gears, you still need to add the planetary gear, which also is a lot of space. Do you know if ferguson's gear has only one planetary? Because if thats the case you can get a lot of vibrations at high frequencies, since the system is not mechanically balanced. On the other hand, I don't know how this type of reduction would behave, in terms of backlash. As the gears are solid, I suppose that the backlash could not be completely eliminated.
The original Ferguson's paradox doesn't mesh quite right (poor efficiency), but if you adjust the module/pitch of the two halves then it does. This is called compound planetary gear, and there are many examples on youtube. You can mix and match different planet/ring tooth counts to get a wide variety of output ratios, and even drive it via the carrier which eliminates the loss of the sun-planet interaction and creates a through-hole for wires in a robot arm, but results in higher RPM of the planet gears. Cycloidal drive is a closely related mechanism where the eccentric functions as the carrier, but it uses a different output method which is only possible due to the one tooth difference between ring and disc, and allows you to get relatively high reduction in a single stage. It is possible to make a split ring cycloidal with different pin counts and two discs of different lobe count glued together like a compound planet, but it has a lot of inherent problems. The big advantage of strain wave is that the flexible component allows it to be true zero-backlash. Cycloidals can get pretty close, but only if machined extremely precisely, so unless you have a lot of money it will have backlash like any other gearbox.
A drawback of the Fergusons paradox is that only a few teeth are contact at any time while the strain wave has very many distributing the stress and thus the strain wave can transmit much more torque in a smaller package. The symmetry of the strain wave is a decided advantage as well.
For anyone wondering, this is still the best low-cost 3d printed reducer for a nema 17 - well worth the price for the design, very printable and professional.
@@GregMatoga enough that the strength of the material at spline tooth scales is the limiting factor, it's basically the only critical thing, also keep in mind threads rip out of PLA or whatever at like 5nm torque so...
Have you gathered any performance metrics yet? Max holding and dynamic torque, stiction, backlash? I'm trying to benchmark my own printed actuators against this, lol Edit.. also, the holy grail of features... Is it backdrivable?
if the flexible spline has the same number of teeth as the fixed spline, but a smaller diameter, it means the teeth on the flex spline are smaller and are not meshing right isn't it ? How can this work, just because of the extra space between teeth ? It just seems this only works because of imprecision in the manufacturing, doesn't it ?
@bschwand, what you say is correct. Due to the intrinsic design this cannot be otherwise, the teeth of the flexible ellipse have to be slightly smaller. This difference is practically negligible. The diameter is slightly reduced but the height of the teeth remains the same, so the deformation only occurs in the width of the tooth and not in the height. Given the module of these teeth 0.8, this difference is imperceptible and does not affect the operation. If we were talking about a real metal strain wave gear, we would be talking about something else since the tolerances are lower, but since this is PLA doesn't affect.
@@ZeroBacklashRobotics ok so this really works because of the inherent flex and give that the material allows, and would be impractical with, say all stainless steel parts. Got it.
What a beautiful and clever construction! I agree with what others have said already (I still say it tho :P), the ball-bearing-flex-spline combination is a such great idea.
@@ZeroBacklashRobotics hehehe.. now you got me wondering about variations of that design.. maybe the ball bearings could become the flex-spline by themselfes.. hmmmmm.. *thinking in progress* :D
@@ZeroBacklashRobotics I struggle a bit to put that thought in proper english words but let's give it a try. I had that funny idea yesternight that a row of balls could act as a flex-spline.. one ball pushing the other so they go around (like they already do).. but with enough clearence to travel a little outside/inside. So, in a way they could be seen as a flexible material.. Then I thought, what if teeth were not tooth-shaped but of circular (or speherical) shape.. which, maybe, could lead to a design like yours.. but without the flex-spline itself, if that makes any sense. Anyways, I tried to scribble something down yesterday and now I think it might been a silly idea (things either have way too much clearance or completely blocking each other). I struggle a bit describing what I'm thinking about. However I might try to scribble another variation of it tonight or tomorrow and if that looks any bit more promising, I might drop you a link here (if that's ok with you) :)
@Simon Lindauer I know what you mean, but this would not work. Because flex spline acts as one unique component so is pushed and pulled every 90 degrees. Using the balls as teeth will work only in the fase that the oval hub is pushing but not when is pulling, because the balls are not linked each other so they behave like independent components. So It will be risky to have the ball stuck because is not pulling back. I don’t know if explained well xD
Is very compact but it has some backlash in theory because the output has the same amount of teeth but has to be bigger due to the oval flex spline right? I think that’s why they are not used more often also not as much teeth can be engaged or dont line up properly compared to „normal“ HD.
Hi Adrian, after printing the flex spline I could say the first hours is pretty good, almost not noticeable. But since its done in PLA with intensive work gets deformed a bit and can have a bit of backlash. I didn’t try with other materials, nylon will improve a lot I think.
Hi @ProjectHivemind I didn’t measure at that level, but If I try to move the axis is quite stuck. Visually is difficult to see it. You need to apply a lot of torque to see something.
@ZeroBacklash Thank you so much for your effort in designing and showing the operation. I have pushed your CAD files. I have one request , can you please upload the STEP model of the full gearbox model ? Since the parts are in stl format , it is difficult to modify or edit or parts or to use it in any assembly.
I've purchased the stl files and printed with great success. Thank you. Is it possible to get the STEP files? I can convert STL into STEP and bring into CAD but the number of surfaces are insane.
Great project! I 3D printed (all parts with PETG) based on your design but found flex spline too rigid. It even split in two after forcing it... Did you print flex spline with another filament?
Does the flex spline hold up to the forces without splitting in half (along where the circular hub meets the spline? The ball bearings with oval hub is a genius ide BTW.. Out of all DIY harmonic wave gears I have seen
@@ZeroBacklashRoboticsI'm also curious about how you create smaller diameters with the same teeth, I think it will affect the module of the flexspline, right?
Hi, I found your harmonic reducers very interesting, I would be interested in buying the files, but first I would like to ask you if they can withstand radial loads such as the tension of a toothed belt applied to a pulley system
Great! Really nice and compact. I’m trying to figure out how to create a camera slider with a 3-axis, and this harmonic driver can be helpful in my project. Just for curiosity, I saw in another video on your channel a driver to know the exact position of the stepper motor. Is it possible to use both on a 3-axis camera slider project, or seem too much engineering?
How do change the gear size? If you have to have the same number of tooth in the static and flex gear you have to change the module. Won't that introduce some losses?
Hi Dan, the bearing is to keep centered the movable circular spline, I didn’t measure the torque, theoretical torque, motor is 39 Ncm x 30 ratio = 11,7 Nm but on top we need to add the efficiency, which is not measured. And taking into consideration is plastic PLA will be lower I would say 50% so I would guess around 5-6 Nm.
Thanks, not sure what happened with my grammar there 😅 oddly mine spins now but in the same direction as the input, it should be spinning in the opposite direction. What am I doing wrong?
very impressive to see such a compact design which seems to have a lot of torque. @ZeroBacklash: is it planed to release the stl files for this build or make a video on how you designed the drive? I would highly appreciate this :)
@@ZeroBacklashRobotics I’m not in need of these Stl files right now but if I was I would pay north of $50 for them. To show my support I will be buying them and I would hope every other subscriber would as well. I feel you are selling yourself short, charge more.
Thanks @gordon6029 much appreciated. Unfortunately not all people think like you. I know is low price… But want to keep it as affordable as possible for most of my subscribers. Thanks again for your support :)
@@ZeroBacklashRobotics Thanks! I'm building simple star tracking motor for old equatorial mount for astrophotography. This might have some use. Cool project and keep on posting!
@@ZeroBacklashRobotics I've just started learning mechanics and the math side of engineering. I've always just winged it when designing something in Fusion360 but this is the first time I actually need to do precision. Your way of using bearing balls was very cool!
Hi Edward, I don’t have a lot of experience with those two materials. But Nylon as far as I know is quite strong to abrasive task, and also its slightly flexible. For this application you don’t need that much, flex spline only deform half mm.
How did you come up with the geometry for the flex spline and the fixed splines? I am trying to design a split circular spline strain wave drive, but I can't visualize how it is possible for the flex spline to interface with both housings if they are different sizes. And if they are the same size, how are you achieving that with a different number of teeth?
Out of curiosity, is this something new you are testing or have you been using them already in the robot arm? Also, have you or do you plan on testing it for durability? The other ones I have seen on yt don't last long.
Hi Nicholas, this is just aside project (experiment) maybe I will use in future projects, its not related to the 6DOF robot, I want to test durability, backlash and electrical consumption in further videos.
Hi @BHARGAV, its noisy because its running at high frequency. It can be noiseless using the TMC StealthChop functionality, but of course at a reduced speed.
started my own (copy) design in F360 ... , questions What material you printed the flex spline with? "Slightly smaller" flex, is it just a scale down of the fixed spline? If i design the driving spline with 2 more theeths, do i have to scale it down also? What ellipse ratio do you use for the driving wave hub? Thanks
Hey I have a dummy question, I have calculated (considering nema 17 max speed at 600rpm) a max speed of around 20 rpm with this reducer without using micro steps, does anybody there mounted the gearbox to tell me if I’m right? Thanks.
Hi, you can find the link in the description, there you find the STL's, list component and explode view assembly: cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
I haven't seen a strain wave gear system that is backdrivable. some cycloidal gear systems are backdrivable because they don't have quite such a high ratio.
Hi Nadav, in this particular case is not backdrivable. Two main reasons, high reduction ratio and its made in plastic, which gives more friction in comparison with the real ones. But real ones most of them they are, you can find plenty of videos where people can back drive them.
Wow, cool design. But does it mean that both non-flex gears have different teeth size? they are both the same diameter but with different teeth count. Is it a big problem? Do you use PLA for flex part? You really inspired me to design harmonic drive now. I was more focused on multistage planetary gears before because they looked more promising, but this kind of solution you have a so damn cool and simple.
Hi @Dmitriy, happy to encourage you to start with strain wave gears. They are unique, these reductors are the core of the mechanics in robotics. Only flex spline is slightly smaller in tooth wide, height is the same, but this is not noticeable since this mechanism is done in PLA, 3D printing tolerances don't go that far.
Hi Manu, If you are still interested the design can be purchased here: cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive Best regards, ZeroBacklash
Excellent design. I love the minimalist approach. Have you considered using Acetal/Delrin balls instead of steel ball bearings? Or, for a contact line rather than a contact point (which should increase torque transmission and reduce wear), use Acetal dowel pins (McMaster offer Acetal dowel pins whose length is equal to its diameter). Cheers!
The use of the steel balls is just brilliant. Together with the bearings I imagine this gear can take quite a load and not wear out too fast. Your 3d prints look pretty shit though. I hope the design comes up for sale somewhere.
Hi Paul, Thanks for your comment. The 3D print is not my best example xD this pieces must be carefully printed. They are small and need precision. If you are still interested the design can be purchased here: cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive Best regards, ZeroBacklash
Wow!
Putting the bearing balls around the input gear is awesome and saves a lot of space!
THIS is what I was waiting for for two years now! Cheers!
I am happy, people get inspired about this stuff ;) Thanks Haensel.
This is easily the most professional and beutiful design of a 3D printed strain wave gearbox. I would like to print my own.
Impressive to say the least. I am very much looking forward to the release of the STLs. Keep up the good work!
Thank you switch!
Really nicely done! Best 3D printed one I've seen on YT!
Can't wait to see them in action
The best DIY harmonic drive I've seen. Just downloaded from the link :)
Very nice design. It's simple, compact and effective.
Congratulations
Thank you Terrestre :)
Absolutely amazing, as always !
I’m a software developer and I’m really excited to start creating softwares for your machine, especially for video capture.
Fantastic!!! Brilliant! Can't wait for the STL files :)
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
This type is my favorite DIY reducer. I wonder how they would stack up against regular harmonic drives if made professionally; if you had a nice cast urethane flex spline with embedded kevlar weave to transfer the torque, I imagine the torque density could be quite competitive, and it should be cheaper to mass produce than notoriously finicky metal splines. Also not the most inaccessible DIY project, to try a cast urethane version.
Also; my intuition would be that given a much more flexy configuration than a metal cup closed on one end, much bigger teeth would be viable and more appropriate for this kind of setup, to get more torque out of the softer materials. Is tooth size something youve tried to optimize for yet?
Hi Eelco, thanks for your input. The tooth size is as small as possible to get the higher ratio without skipping teeth.
@@ZeroBacklashRobotics Sure, bigger ratios means smaller teeth. But 30/31 would be the same as 60/62, no? Bigger teeth should mean bigger peak output torque, all else being equal. An open ended spline like this, made from a flexy material like urethane, should be able to accomodate quite big flex amplitudes, accomodating nice and deep teeth, and high torques. Curious how much torque youd be able to get out of teeth of this size!
@Eelco, its not the same, with 30/32 teeth you will get more torque indeed, because your wide teeth can increase with the same perimeter, but your ratio will decrease to 15:1, to get the same ratio increasing teeth there is no other way to increase the diameter size of your splines.
@@ZeroBacklashRobotics I said 30/31 but yeah you need 2 tooth difference with 2 contact points I suppose.
Or do you? If the spline had two sections with a one-tooth offset between the sections, that should work I think? Like 30/32 -> 31/33
What can I say? Your design is sooooooo impressive! As always!
Thank you sandoftime :)
Great looking design! It would be great to see some testing, such as the torque, backlash and reliability after many hours of running.
Hi @IV Projects, Its planned in the agenda :) in the coming videos, thanks for your comment ;)
Best 3d printed design I have seen. I hope you release the stl files, I would love to make one :)
Video sounds better at x1,25 speed... good work!
damn i would have never tought to use bearing balls.... amazing idea!
There is a similar type of gear called ferguson's paradox gear that behave almost exactly like strain wave gear. But without needing to use a flex spline. All three components are solid and fixed. It also have similar high reduction ratio as a strain wave gear. Essentially it's compose of two outer ring gears with one ring that has one more gear teeth than the other align co-axially. Then there is one smaller planet gear nested within that rotate with the input shaft. The two outer ring gears have tooth of slightly different sizes, when the planet gear came into contact with the two outer ring gears, it push the gear teeth together one side and push it away slightly on the other side. The gear ratio of this system is = number_of_teeth_in_ring_gear_1 / (number_of_teeth_in_ring_gear_1 - number_of_teeth_in_ring_gear_2) , so if you have 51 tooth in ring 1 and 52 tooth in ring 2, you get a reduction ratio of 51. Ferugson paradox gear is easier to make than strain wave gear, since there is no need for flex spline. And second advantage is that the gear tooth pressure angle can be lower compare to strain wave, which lead to greater gear efficiency.
that seems to be a variation on a cycloid drive
or I am not visualizing this right
ok sorry got it. not cycloidal at all but also plays on some tolerances, basically the gears do not mesh completely right.
Hi @Keen Heat, thanks for your comment. I did not know about the existence of this type of gear. It seems very interesting, mainly because having a difference of one tooth between both gears, the ratio is double than strain wave gear, but keep in mind being able to reduce the size of the main gears, you still need to add the planetary gear, which also is a lot of space. Do you know if ferguson's gear has only one planetary? Because if thats the case you can get a lot of vibrations at high frequencies, since the system is not mechanically balanced. On the other hand, I don't know how this type of reduction would behave, in terms of backlash. As the gears are solid, I suppose that the backlash could not be completely eliminated.
You can have several planetaries
There are animations on youtube
The original Ferguson's paradox doesn't mesh quite right (poor efficiency), but if you adjust the module/pitch of the two halves then it does. This is called compound planetary gear, and there are many examples on youtube. You can mix and match different planet/ring tooth counts to get a wide variety of output ratios, and even drive it via the carrier which eliminates the loss of the sun-planet interaction and creates a through-hole for wires in a robot arm, but results in higher RPM of the planet gears.
Cycloidal drive is a closely related mechanism where the eccentric functions as the carrier, but it uses a different output method which is only possible due to the one tooth difference between ring and disc, and allows you to get relatively high reduction in a single stage. It is possible to make a split ring cycloidal with different pin counts and two discs of different lobe count glued together like a compound planet, but it has a lot of inherent problems.
The big advantage of strain wave is that the flexible component allows it to be true zero-backlash. Cycloidals can get pretty close, but only if machined extremely precisely, so unless you have a lot of money it will have backlash like any other gearbox.
A drawback of the Fergusons paradox is that only a few teeth are contact at any time while the strain wave has very many distributing the stress and thus the strain wave can transmit much more torque in a smaller package. The symmetry of the strain wave is a decided advantage as well.
Super compact / friction free design. I have seen a lot of similar Strain Wave Gear, but never that good ! Great job, need stls !
For anyone wondering, this is still the best low-cost 3d printed reducer for a nema 17 - well worth the price for the design, very printable and professional.
I am going to use 3 of these as my z axis to raise and lower my bed. Just gotta figure out the mountjng solution and belt solution
So how many Nm you're getting?
@@GregMatoga enough that the strength of the material at spline tooth scales is the limiting factor, it's basically the only critical thing, also keep in mind threads rip out of PLA or whatever at like 5nm torque so...
Can you publish the STLs on Thingiverse or Printables? This is a great compact drive!
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
@@ZeroBacklashRobotics booooo
Great design. Top notch
Thanks Dylan!
Have you gathered any performance metrics yet? Max holding and dynamic torque, stiction, backlash? I'm trying to benchmark my own printed actuators against this, lol
Edit.. also, the holy grail of features... Is it backdrivable?
if the flexible spline has the same number of teeth as the fixed spline, but a smaller diameter, it means the teeth on the flex spline are smaller and are not meshing right isn't it ? How can this work, just because of the extra space between teeth ?
It just seems this only works because of imprecision in the manufacturing, doesn't it ?
@bschwand, what you say is correct. Due to the intrinsic design this cannot be otherwise, the teeth of the flexible ellipse have to be slightly smaller. This difference is practically negligible. The diameter is slightly reduced but the height of the teeth remains the same, so the deformation only occurs in the width of the tooth and not in the height. Given the module of these teeth 0.8, this difference is imperceptible and does not affect the operation. If we were talking about a real metal strain wave gear, we would be talking about something else since the tolerances are lower, but since this is PLA doesn't affect.
@@ZeroBacklashRobotics ok so this really works because of the inherent flex and give that the material allows, and would be impractical with, say all stainless steel parts. Got it.
Brilliant design!!
I printed your design and it works pretty well. Thanks🥰
Thanks for sharing naoa :)
A really clean and compact drive! Love it!
What a beautiful and clever construction! I agree with what others have said already (I still say it tho :P), the ball-bearing-flex-spline combination is a such great idea.
Hi Simon, thank you, I loved the idea from first time it came into my mind :)
@@ZeroBacklashRobotics hehehe.. now you got me wondering about variations of that design.. maybe the ball bearings could become the flex-spline by themselfes.. hmmmmm.. *thinking in progress* :D
Hey Simon, don’t get your point of view 🧐 what exactly you mean?
@@ZeroBacklashRobotics I struggle a bit to put that thought in proper english words but let's give it a try. I had that funny idea yesternight that a row of balls could act as a flex-spline.. one ball pushing the other so they go around (like they already do).. but with enough clearence to travel a little outside/inside. So, in a way they could be seen as a flexible material.. Then I thought, what if teeth were not tooth-shaped but of circular (or speherical) shape.. which, maybe, could lead to a design like yours.. but without the flex-spline itself, if that makes any sense.
Anyways, I tried to scribble something down yesterday and now I think it might been a silly idea (things either have way too much clearance or completely blocking each other). I struggle a bit describing what I'm thinking about. However I might try to scribble another variation of it tonight or tomorrow and if that looks any bit more promising, I might drop you a link here (if that's ok with you) :)
@Simon Lindauer I know what you mean, but this would not work. Because flex spline acts as one unique component so is pushed and pulled every 90 degrees. Using the balls as teeth will work only in the fase that the oval hub is pushing but not when is pulling, because the balls are not linked each other so they behave like independent components. So It will be risky to have the ball stuck because is not pulling back. I don’t know if explained well xD
Getting into more detail with each vedio
Yees, thats the idea ;)
@@ZeroBacklashRobotics thanku for inspiring and sharing ..... Very detailed work
Is very compact but it has some backlash in theory because the output has the same amount of teeth but has to be bigger due to the oval flex spline right? I think that’s why they are not used more often also not as much teeth can be engaged or dont line up properly compared to „normal“ HD.
Very nice! It would be interesting to see load current on the motor with and without the HD.
That would be a good check yes. To see how much current takes to move all the stuff.
Best design👍🏻 good job
I am most interested in how much the backlash is. Have you checked?
Hi Adrian, after printing the flex spline I could say the first hours is pretty good, almost not noticeable. But since its done in PLA with intensive work gets deformed a bit and can have a bit of backlash. I didn’t try with other materials, nylon will improve a lot I think.
@@ZeroBacklashRobotics Is there a safe operating load, where the deformation doesn't happen (or is much slower)?
Hi @ProjectHivemind I didn’t measure at that level, but If I try to move the axis is quite stuck. Visually is difficult to see it. You need to apply a lot of torque to see something.
Looks like ideal solution for Voron V2 Z-Drives
Exactly what im thinking. Or trident with belted z, or lead screw could work to. Have you done it to your voron z axis?
@ZeroBacklash Thank you so much for your effort in designing and showing the operation. I have pushed your CAD files. I have one request , can you please upload the STEP model of the full gearbox model ? Since the parts are in stl format , it is difficult to modify or edit or parts or to use it in any assembly.
I've purchased the stl files and printed with great success. Thank you. Is it possible to get the STEP files? I can convert STL into STEP and bring into CAD but the number of surfaces are insane.
I think it goes without saying, to really appreciate this video you have to enable CC, and turn the volume down completely.
Excellent design 👍
Great project! I 3D printed (all parts with PETG) based on your design but found flex spline too rigid. It even split in two after forcing it... Did you print flex spline with another filament?
He used PLA, have a look at the link in the description for all printing settings
Really cool job !!!
Does the flex spline hold up to the forces without splitting in half (along where the circular hub meets the spline? The ball bearings with oval hub is a genius ide BTW.. Out of all DIY harmonic wave gears I have seen
Hi mvadu, thanks for your comment. The flex spline is holding yes, but I didnt push to the limit yet, but first tests went quite well.
Beautiful.
I know you dont release files, but could you give us the modulus and scale factors for the teeth profiles ?
Hi Insane, modulus is 0.8 what do you mean with scale factor?
@@ZeroBacklashRobotics The flexspline and second stage ring need to be scaled a little bit so the teeth mesh right ?
@@ZeroBacklashRoboticsI'm also curious about how you create smaller diameters with the same teeth, I think it will affect the module of the flexspline, right?
Hi, I found your harmonic reducers very interesting, I would be interested in buying the files, but first I would like to ask you if they can withstand radial loads such as the tension of a toothed belt applied to a pulley system
When you say the flex spline is smaller than the outside 60... how much smaller? how did you find the right size?
Great! Really nice and compact.
I’m trying to figure out how to create a camera slider with a 3-axis, and this harmonic driver can be helpful in my project.
Just for curiosity, I saw in another video on your channel a driver to know the exact position of the stepper motor. Is it possible to use both on a 3-axis camera slider project, or seem too much engineering?
How do change the gear size? If you have to have the same number of tooth in the static and flex gear you have to change the module. Won't that introduce some losses?
gorgeous. why the first bearing? what is the output torque?
Hi Dan, the bearing is to keep centered the movable circular spline, I didn’t measure the torque, theoretical torque, motor is 39 Ncm x 30 ratio = 11,7 Nm but on top we need to add the efficiency, which is not measured. And taking into consideration is plastic PLA will be lower I would say 50% so I would guess around 5-6 Nm.
Do you know how much angular play this assembly has?
Is it degrees, minutes or seconds?
If the flex spline doesn't even move then how do it spin the output??
Hi @ryanmckenna2047, because the dynamic circular spline (end axis) has 2 teeth more that the flex spline. That produces the moment.
Thanks, not sure what happened with my grammar there 😅 oddly mine spins now but in the same direction as the input, it should be spinning in the opposite direction. What am I doing wrong?
I bought the model but the belt don’t fit the exncloser. Can you please help me?
Great work! Would love to build one for myself.
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
How long would the flex spline printed from pla last?
very impressive to see such a compact design which seems to have a lot of torque.
@ZeroBacklash: is it planed to release the stl files for this build or make a video on how you designed the drive? I would highly appreciate this :)
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
@@ZeroBacklashRobotics I’m not in need of these Stl files right now but if I was I would pay north of $50 for them. To show my support I will be buying them and I would hope every other subscriber would as well. I feel you are selling yourself short, charge more.
Thanks @gordon6029 much appreciated. Unfortunately not all people think like you. I know is low price… But want to keep it as affordable as possible for most of my subscribers. Thanks again for your support :)
What does "half" mean in half movement circular spline?
Is the flexible spline printed with same PLA, or a flexible filament?
This is really neat! Did you print the spline, or make it from something else?
Hi @wow all is 3D printed, 0.4 mm nozzle in PLA.
What a nice design. Will you release the stls or sell them in some way?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
@@ZeroBacklashRobotics thank alot for sharing this files! For some reason I can't load the website. Is there something wrong with the server?
Mmm I’ve just tried and in my phone is working..
@@ZeroBacklashRobotics yeah you are ri8it works for me as well
Great design, would u mind sharing the stls/step?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
Hey i was wondering what kind of motor driver do you use?
I could think a lot of use cases for this. Will you release plans sometime?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
@@ZeroBacklashRobotics Thanks! I'm building simple star tracking motor for old equatorial mount for astrophotography. This might have some use. Cool project and keep on posting!
Thanks to you @ex1tium to support ZeroBacklash channel :)
@@ZeroBacklashRobotics I've just started learning mechanics and the math side of engineering. I've always just winged it when designing something in Fusion360 but this is the first time I actually need to do precision. Your way of using bearing balls was very cool!
Happy you like it 😎
I was going to purchase your model on Cults3D but their login seems to be down. Do you have the files up on anywhere else?
Hi @jaytee0108, did it work finally?
@@ZeroBacklashRobotics Yes, I eventually was able to login and download it. Thank.
Great! Thanks ;)
Hello, why is it better to choose nylon instead of tpu for flex spline? Is it because of the consideration of longer service life？
Hi Edward, I don’t have a lot of experience with those two materials. But Nylon as far as I know is quite strong to abrasive task, and also its slightly flexible. For this application you don’t need that much, flex spline only deform half mm.
That is a very interesting design. Is the flex spline pla or something else? Have you released the cad files for this?
Hi, chris, its pure PLA. You can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
How did you come up with the geometry for the flex spline and the fixed splines? I am trying to design a split circular spline strain wave drive, but I can't visualize how it is possible for the flex spline to interface with both housings if they are different sizes. And if they are the same size, how are you achieving that with a different number of teeth?
Will I get everything if I buy it from 3d cult or only the 3d printer file
Great design! 👍 is it backdrivable?
No Mishin, it would break PLA before it moves.
Out of curiosity, is this something new you are testing or have you been using them already in the robot arm?
Also, have you or do you plan on testing it for durability? The other ones I have seen on yt don't last long.
Hi Nicholas, this is just aside project (experiment) maybe I will use in future projects, its not related to the 6DOF robot, I want to test durability, backlash and electrical consumption in further videos.
what filament is the flex spine made of?
Hi, what is the steel balls diameter? Thank you in advance! best
Hi Veselin 5mm diameter ;)
Hey! very cool what you do! Is the belt also printed or bought? can you share stl files?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
The belt is also printed in PLA.
no backlash is good but is it noisy ?
Hi @BHARGAV, its noisy because its running at high frequency. It can be noiseless using the TMC StealthChop functionality, but of course at a reduced speed.
Will I get the item if I bought it from cult 3d of just the printing files
Hi, cani ask what torque can it handle? :)
started my own (copy) design in F360 ... , questions
What material you printed the flex spline with?
"Slightly smaller" flex, is it just a scale down of the fixed spline?
If i design the driving spline with 2 more theeths, do i have to scale it down also?
What ellipse ratio do you use for the driving wave hub?
Thanks
can I know bout extra preparations when I purchase that stl file? like bearings, bolts, etc
Are you planning to release on Thingiverse?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
Hey I have a dummy question, I have calculated (considering nema 17 max speed at 600rpm) a max speed of around 20 rpm with this reducer without using micro steps, does anybody there mounted the gearbox to tell me if I’m right? Thanks.
Nice! Will you share files and components list?
Hi, you can find the link in the description, there you find the STL's, list component and explode view assembly:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
Great project sir. Can you share the step files. ???
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
This is really awesome! Would you release the STL at some point? Would be happy to pay for it... 🙂
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
@@ZeroBacklashRobotics wohoooo!!! Thank you very much! Of course I bought it 🥰
Thank you Andi ;)
@@ZeroBacklashRobotics no no… thank YOU for publishing. You put a lot of effort in this. 👍🥰
Is it back drivable?
No
I haven't seen a strain wave gear system that is backdrivable. some cycloidal gear systems are backdrivable because they don't have quite such a high ratio.
Hi Nadav, in this particular case is not backdrivable. Two main reasons, high reduction ratio and its made in plastic, which gives more friction in comparison with the real ones. But real ones most of them they are, you can find plenty of videos where people can back drive them.
does this work the other way ?
could be a great second gear for battery cars :)
Hi Maly, if you mean if it is backdrivable is not possible. The reduction is quite high to do that.
Where is the STL file I can download ?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
Is this zero backlash?
Not measured yet. But first impression is pretty close ;)
Bearing sizes ????
Have you measured the backlash on this? :D
Hi, its good, only small flexion of the PLA is noticeable, but I am wondering after some use, PLA will deform. Nylon should be better choice.
@@ZeroBacklashRobotics I think the ball bearings are rubbing on themselves. I think grease may be good :) idk.
@wow hahah yes but I didn’t have for the video xD I agree with you it would increase the performance.
Wow, cool design. But does it mean that both non-flex gears have different teeth size? they are both the same diameter but with different teeth count. Is it a big problem? Do you use PLA for flex part?
You really inspired me to design harmonic drive now. I was more focused on multistage planetary gears before because they looked more promising, but this kind of solution you have a so damn cool and simple.
Hi @Dmitriy, happy to encourage you to start with strain wave gears. They are unique, these reductors are the core of the mechanics in robotics. Only flex spline is slightly smaller in tooth wide, height is the same, but this is not noticeable since this mechanism is done in PLA, 3D printing tolerances don't go that far.
`hello,
nice and compact hd! Can you share your designs? I plan to use it and modify it to make 3d printer extruder.
Thanks
Hi Manu,
If you are still interested the design can be purchased here:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
Best regards, ZeroBacklash
Excellent design. I love the minimalist approach. Have you considered using Acetal/Delrin balls instead of steel ball bearings? Or, for a contact line rather than a contact point (which should increase torque transmission and reduce wear), use Acetal dowel pins (McMaster offer Acetal dowel pins whose length is equal to its diameter). Cheers!
Where can I find the STL files?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
What is it used for?
Hi Anabolic, whatever that requires precision and torque. Robotics mainly...
Will you make the CAD or stl files available? This would be amazing for a project I’m working on
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
nice! are the STL available?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
@@ZeroBacklashRobotics thanks!!
AWESOME
That music and that computer voice. Immediately thought of a scam... sadly there's no (official) visible dislike count anymore.
STL?
Hi, you can find the link in the description:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
Hi very nice
try that with magnetic gear system, not mechanical
off-axis rigid circle gear planetary gear
The use of the steel balls is just brilliant. Together with the bearings I imagine this gear can take quite a load and not wear out too fast.
Your 3d prints look pretty shit though. I hope the design comes up for sale somewhere.
Hi Paul,
Thanks for your comment. The 3D print is not my best example xD this pieces must be carefully printed. They are small and need precision.
If you are still interested the design can be purchased here:
cults3d.com/en/3d-model/tool/3d-printed-strain-wave-gear-harmonic-drive
Best regards, ZeroBacklash
wtf is that intro
How long till that belt has worn down form friction on the toothed gear with 2 extra teeth?