What are Eccentrically Cycloidal Gears?
Вставка
- Опубліковано 18 бер 2022
- Designing and building a 3d Printed 15:1 eccentric cycloid gearbox / actuator to increase the torque from a stepper motor.
CAD and STLs are available at: github.com/roTechnic/Eccentri...
These are the supplies i used:
NEMA 17 stepper: amzn.to/3gtK3Qd
Arduino Mega: amzn.to/3nx3Oui
5mm steel rod: amzn.to/3HMBGeV
Thin wall bearings: amzn.to/3gtYLqG
and: amzn.to/3H7JPZY
1.75mm PLA+ filament: amzn.to/3HvzVT5 - Наука та технологія
At 0:48 I patted myself on the back for spotting the weakness when the eccentric disc was at the peak of the tooth.
At 1:12 I burst out laughing, when I realised where this was going to end up.
Of course, I had seen the final design in the thumbnail, but the progression from step to step made it blindingly clear.
Love your approach, Thanks.
Totally brilliant explanation of how this gear can be derived geometrically from a normal cycloidal gearbox. The concept has of course been known for a long time, for example from a recent Bosch patent. But the idea is probably many decades older.
That's a great design :D It's very refreshing to see some new ideas in the space of 3D printed gear actuators. Maybe you could save some space if the single tooth gear could engage internal cycloid gear.
oh, now that’s a brilliant idea. that’s what I love about doing videos - I get the best feedback ever!
one big advantage of this approach over putting it on the inside is that this doesn't require large bearings, pins, etc. this can be scaled up to very large gear ratios while still using cheap 608 bearings.
Although, I could see the potential for a large gear that has a solid disk on one side, possibly pinching the edge of the gear between the one lobed gear and bearings on the outside
just saw the video, wanted to write the same... so are there updates on this?
@@CM-mo7mv sorry - didn't see your comment when it came in! I did a follow up on this: ua-cam.com/video/xv5cd7Bg7Uk/v-deo.html
The 'pinion' seems to be pushed laterally by the main gear, due to the incline. Perhaps it both were herringbone, the forces would balance out, keeping the pinion from being pushed.
absolutely love the explanations, the pace and the material in the description box to replicate the teachings!!!
Absolutely. Discovery channel-like
Have you considered a herringbone approach - it may result in keeping axial forces to a minimum.
no, I had not.
that’s a great idea! if you just mirror the pin and disk in the vertical direction then that would create it. I’ll have to give it a go. thank you
Wondering what is a herringbone approach ?
@@BHARGAV_GAJJAR
//////
\\\\\\
@@BHARGAV_GAJJAR when the gears are mirrored to achieve chevron shaped profile like this: »
That is fascinating. Thank you for the fantastic explanation. The sequential nature of the explanation was excellent.
2:30 I think there is some sliding here beside rolling. Still an awesome design.
If you wanted to eliminate the sliding wear, you could use the 180 degree phased approach at 1:08 but use a ball race around each pinion. roller camshafts in auto engines are made like this.
So cool! I will try modelling my own one. Your channel is brilliant!
thank you very much. let me know how your modelling goes!
the fact that the small gear has only 1 tooth is amazing and i'm sure this would allow for some very high reduction gearboxes.
this is amazing, i wonder why nobody uses this design.
i love the video and explanation.
and you are very pleasant to listen to, reminds me of david attenborough.
thank you.
I think the main reason it’s not used is that it’s really really difficult to machine these forms using standard subtractive methods.
however, with 3d printing they are easy to make.
I mean, there may well be other reasons too, I haven’t done a load of testing on this yet, but I’ll cover that in a future video.
Brilliant video, great design. You've inspired me to try making my own :D
You are an excellent teacher, thank you so much for sharing these videos!
what a nice and exotic design, very fascinating to look at!
thank you very much
I like your idea! however out of curiosity, how is this different now to a standard gear ? Does the shape offer any benefits over two gears?
What a beautiful idea and interesting presentation. TY for sharing.
Thank you so much!
What a neat idea. Nice implementation too. I agree that the backlash should be low since the gear is always engaged. I wonder about the power losses in the gear train. It is reminiscent of a worm drive, which typically has high losses for higher gear ratios.
thank you very much.
yeah, but the difference here is that worm drives work through sliding friction, but this works only with rolling friction which has much lower losses
@@roTechnic : I'm a little confused by that claim. The pin (at least when it's contacting at its largest diameter) has to slide across the surface of the large gear, yes? And wouldn't that give you sliding friction? What am I missing?
@@roTechnic I'd have to disagree about the "only rolling friction" part. Since this is a 1:N gearing, just like in a worm gear AND height of both gears are the same, the length of contact of the small gear must be much greater. In fact this is another form of a worm gear but with parallel rotation axes. Thanks for introducing me to your idea though, might be useful for me sometime.
If you start out with a small wheel rolling on the outside of a bigger wheel. Then offset the small wheel and modify the bigger wheel to fit. Now the small gear is "bigger" and the big gear is smaller. So if the small gear is "rolling" on the outside of the bigger gear, the gear ratio is smaller than originally and when the small gear is pointing out, the ratio is even higher. But in the final situation, the ratio must be uniform: The big gear MUST be moving uniformly. So... Alas, but there is friction and sliding involved in this type of gear.
Note: it's seems quite good at being back driven too, (at least what I've seen in the 11 old video here on YT)...
Thinking further, you could use multiple smaller motors, even build a planetary type configuration, to build smaller/stronger drives.
yeah, I really should have mentioned how easily backdriven it is. whoops!
I wonder if this could be used in a planetary configuration and if that would allow higher reductions…
This is a beautiful video! Loved it!
Awesome. Maybe it could be inverted to save space..? place the input shaft off-center within an output ring gear with an internal helical (or herringbone) cycloidal profile
Would be interesting to see how small the parts/curves could be printed for larger reductions and more compact mechanisms
thanks for your comments - both are great ideas - I’m definitely going to try shrinking it in the future and may see if I can get the pin internal to the disk. the only question is how to hold the top of the pin - at the moment it has a bearing on it which mounts into the housing and stops any flex
Fantastic, great job man!
If you make the teeth by alternating internal cycloids (dedenda r1) and external cycloids (addenda r2)
then the driving pin can be made a cycloid gear too (with the radii swapped for dedenda and addenda). I'd choose r1 == r2 though.
Giving the driving pin more than only just one tooth (e.g. five smaller teeth) makes more teeth intermesh simultaneously.
(=> Less twist needed to cover a tooth period. Or more than a full period covered.)
Also: With alternating internal and external cycloids at the addendum to dedendum crossover point the pressure angle is always zero.
=> No forces pressing the gears apart!
This sounds really interesting, but I'm not sure I understand- do you have a reference?
What does that even look like?
Nice video shot, thanks for sharing it :)
I’d be curious to see how this transfers to a planetary gearbox - would it work at all? Would it improve upon the already compact gear reduction? Would it hurt performance?
try it bro!
How would you transfer it? Also, I want to make a 3d Printed Cycloidal Drive Box. Not able to decide whether to print this 1 tooth gearbox or a full fledged one. This seems like a simple gear attachment that is split to be continuous. Whereas the cycloidal drive has a complete different mechanics.
Beautiful. Video and mechanism 😊
Amazing ! Great ingenuity !
Thank you very much!
This is a stroke of genius🤯
There is definitely sliding friction. If there wasn’t this would just be a normal gear.
excellent work!!
Correction. This does not roll, it slides, which means it has high friction.
This is such s great channel!! Keep it up!! 😊
Thank you! Will do!
You can also 3D print brushless stepper motors now. Maybe put in an optical sensor to measure rotation? The optical sensors for mice are dirt cheap and should be able to measure the rough PLA surface going by. Might be a really cool fully printed joint with embedded motor.
Very nice, thank you sir for this share, brilliant!
Pretty cool, the one thing to consider is a cycloïdal drive is a lot more radially compact relative to this. Perhaps there’s a better way to package it?
Great explanation
May be you can save more space by making the driven gear as internal gear.
Very cool... There was a model with a crank on thingiverse at one point and I printed it. But with no bearings, it was a terrible drive and I dismissed it outright..
Thanks to you, I now have another model in the queue 😋
I'm actually very interested in the long term wear characteristics of this and how it compares to other designs.. any plans for a stress test?
Thanks again for a great video and "avenging" this design for me😉
thank you very much.
before I choose one of these designs for my robot arm build I’m going to do a stress test of them - leave them running for a few days. I’ll publish the results in a future video.
oh, and let me know how the print goes
Wish somebody could make a video for onshape. For the single tooth how is the center found.
Thank You!
This is seriously groundbreaking
Very interesting design! Is there any noticeable backlash in the finished actuator? Do you think lubricant would help reduce losses?
I’ve got a bit of backlash at one point in the rotation. I think it’s due to my 3d printer needing a bit of calibration and the arc which it occurs in is only a couple of degrees. everywhere else has no noticeable backlash.
and, yes a bit of PTFE silicone grease would help greatly, but it looks dreadful on video 😎
WOW that is cool
great design , but what about all the friction involved in this system ?
cycloidal design permit the use of bearing that remove all friction ;)
thank you very much.
as it’s a cycloidal design, the pin rolls around the disk instead of sliding. most of the friction comes just from inaccuracies in the 3d print..I’d say that this has lower friction than my cycloidal drive of a similar size
@@roTechnic ho !! yes now i understand better !! thanks for the info , i think i will go for made one :)
So cool 👍
I love this.
Looks a bit like a lego's version of a duplo helical wormdrive 👍
I tried making an internal version but it didn't end up saving much space without drastically shrinking the gear reduction. I did make a mirrored/herringbone version though for my own joint. I'll post the link below
ua-cam.com/video/soCIkuO6VmQ/v-deo.html
Love your design. I’m hitting the same issues myself with the internal version, but I think that by reducing the contraction and pin size I’m getting somewhere. Did you find with your internal version you got an extra node on the disc?
@@roTechnic yep, one extra node. I believe that works in our favor though for extra reduction.
Awesome video but what settings did you use in your script to get that original cycloid shape?
Lovely
Very cool
thanks
Clever design. Nice explanation. Thanks!
Cool
Nice and tntriguing design! But i wonder, is it 1 to 1 equivalent to the "normal" cycloidial gear? In terms of engagement? If that design is as strong (in terms of breakage) as the normal version (for equal sized gears), that could be a viable alternative!
interesting! If by strong you mean "how much torque can it transmit?" then the normal cycloidal drive wins because of the increased surface area in contact at any given time (especially if you have multiple offset rotors). However, in my testing, I've found that the limiting factor is the motor rather than the gearbox - both these designs will handle the torque generated by a NEMA17 stepper. If you start using brushless motors then you may see a difference...
Thank you very much for sharing, fascinating! Is it true that you are using a 42 Ncm holding torque motor, so the expected efficiency at stall is 65% (4.2Nm / 6.3Nm)? I am curious about how the efficiency compares to typical spur gearing in general. Equally as curious about the cycle life/wear, since you mentioned that the interaction is rolling instead of sliding.
Hi, that's partially true, although in my experience, stepper motor holding torque is dependent on the voltage, current, driver choice etc. Instead of working out the efficiencies I've chosen to use the same setup on a number of different reduction mechanisms and see what the torque is - for my purposes, this works fine but YMMV. Comparing to a spur gear reducer is difficult as the efficiency will depend on the pressure angles and amount of backlash (and of course the quality of the prints) but I'd say this is as good if not slightly better if you are aiming for very low backlash - mainly because the form is more tolerant of the way that 3D printers produce parts. Wear on these is supposed to be negligible as long as you chose the right lube.
In a practical scenario it would be best to make multiple angle offset pins instead of a smooth curvy one because it is hard to machine the pinion pin.
Yes, you're absolutely right - if this was being made out of metal, using subtractive technology it would definitely be easier to make multiple "steps" instead of a smooth surface. With 3d printing (additive) the smooth surface means it prints easily and doesn't need supports.
ignoring cnc (the obvious answer) this could be treated like a thread with a huge pitch (22mm, not sure if my lathe could do that) and a massive round thread form
consider, if you will, an eccentrically cycloidal planetary gearbox.
hey! I am new to this space, and am documenting every cool and possibly useful mechanism i see in a notebook. just wondering, how do you find this stuff? just through reading papers about a specific mechanical system?
0:08 Подскажите, пожалуйста, возможна ли тут передача вращения в обратном направлении? То есть в данном случае внутренний эксцентриковый вал (шестерня) крутит наружную шестерню. А если приложить вращательный момент на наружную шестерню, то будет ли она вращать внутреннюю?
thats so dang cute!
Just curious: Is this effectively the same thing as a worm gear (but with very shallow screw pitch so it can be backdriven)?
not really. in a work drive there is sliding friction so it wears. in this, there is ideally only rolling friction which is much lower
Using two conical bearings behind each plate, the pin gear can be moved to the inner of the big gear, while being rigid, compact, and protected (bearings will be covered). If the tooth angle can be increased, the big gear will shrink even more. Angle is the main ratio driver?
You could combine the above with 90 degree conical gears to bring the motor parallel or a centered worm gear to have the motor inside the arm. Alternatively, belt or gear drive the motor to integrate it vertically in the arm. Apart from adding ratio reduction of ie. /2, this way even a second synched motor can be added, to drive both sides of the pin.
That's a really good point about the angle being the main ratio driver - I had not considered that before. Thank you
Congratulations for your wonderful solution: Backdrivable, compact, strong and with not any sliding friction.
Why do you use such large and expensive bearings?
thank you.
I’m using the large bearings because I want to eventually have this in a robot arm, and the larger the diameter of the bearing the more it and the plastic will resist twisting forces.
What if you moved the smaller gear into the larger gear (that is have the “teeth” on the inside of a circle, not outside.
Hahaha! I did exactly that in the next video! Let me know what you think of it...
WOW
great. make an ivolute one. inside out one
Wouldn't you be able to move the pin to the inside of the disc?
This is a great video! I am trying to make an EC gear using your tutorial with Python scrip and fusion 360 but I am trying increase the diameter of the gear. What is the best way to increase the diameter of the EC gear?
Thank you for the help!
Hiya, the best way to increase he size of the gear is to change the number of pins or the pin circle radius. increasing either of these will increase the size of the gear, but only the number of pins will change the reduction ratio of the gear.
@@roTechnic I have been trying to recreate your base that uses the EC gearing for my robotic arm but, I am have a bit of trouble. Do you mind sharing the specs of the larger EC gear used in the base (pin_radius, pin_circle_radius, number_of_pins, contraction)?
Thank you for the help!
Nice video, I want to build the Hand Crank gear model. What metal hardware are you using? I'm guessing bearing spec are: Deep Groove Ball Bearings 5mm Bore 16mm OD 5mm Thick; Is this correct ?
I have just put my make into operation and am thrilled: Almost no play, quiet run and driveable backwards with ingeniously simple construction. The origin of this concept could be a Russian patent?
Glad to hear you've made one! yes I found the concept at www.ec-gearing.com/ and worked out how to design a similar system myself
Hi could you put the pinion into the big gear, to make it more compact and put the teeth of the big gear into the so created gear ring?
Yeah, a couple of other folks have suggested that and I think it's a great idea.
The only thing to work out is how to hold the top of the pinion steady - currently it has a bearing on the top which is held by the housing - any ideas on how to do this if it's enclosed by the larger gear?
@@roTechnic you can have a bridge tower come up from the base to support the top, the housing would have to have room for it (or a hole in it's center) to give clearance, but it should be straighforward to do.
Yep, you're right - why didn't I think of that 😅
For example, planetary gears work fine with single sided bearings.
@@falkgerbig7787 yeah, they do, but this has a fairly large moment at the top which needs a bearing to counteract it
What is the lowest gear ratio you can create with these cycloidal gears? Can a single tooth gear turn with another single tooth gear that is mirrored? Would it be possible to create a 2:1 ratio for these gears? Also how do you generate the shape at 0:39? Sorry for all the questions lol
you could probably go down to about 4:1 reduction, the shapes produced are not really suitable for low reduction ratios as the smaller the reduction, the more precise the shape has to be, and especially if you are 3d printing it, you'd be hard pushed to get the precision you need. The shape is generated using a plugin I wrote for fusion 360 and I covered how it works and how to use it in a previous video "Designing a cycloidal drive in Python and Fusion 360".
@@roTechnic Alright, thanks so much for your answer! With that in mind, do cycloidal gears only produce gear ratios of integers n:1 or are there gears that could mesh that could have a ratio of say 9:4?
@@inoojo4780 I'm sure it's possible to design a pinion with more than 1 tooth which would give you varying ratios like you've described. But I'm afraid I haven't spent any time on the maths for it :(
@@roTechnic Got it, thanks again!
can you place the pin within the toothweel? should make for a significantly smaller design.
this is a fairly low gear ratio (15:1) and we've seen a lot of people do backdrivable cyclodial gears up to 20:1 or so, how high a gear ratio do you need to go before it's no longer backdrivable?
This drive has less friction than the cycloidal drives I've made in the past, but I've got no real way to measure the losses due to friction - and it's these losses which will hinder backdrivability. Have you got any ideas on how to measure them?
@@roTechnic You can measure force (ie. torque) in and force out i guess. Any difference from the calculated one (deriving from the ratio) should be losses due to friction or other causes. You can also find papers on friction in ie. helical gears (long reads) on the subject online.
What if the drive gear was on the inside?
great idea, I’m going to try that for the next video!
If you don't mind, could you post to github the python code for what is shown at 0:45 in the above video? I ask so that I can change the number of valleys to 12 from 15 (and all affected geometry), in a circumference of 42mm. Please. I've been beating my head against the wall since you posted this, and made dozens of not-quite-exact-enough iterations through Kentucky windage, to no avail. I don't mean to bother, and have therefore done all I know to do first. Thank you for your excellent work and kind attention. I will hold my breath. :D
Hiya, I'm using the same code as in my cycloidal drive video - it's posted here: github.com/roTechnic/CycloidalDesign
You should just need to adjust the pin diameter to something like 15mm and the PCD to your gear diameter.
Let me know if you have any trouble changing the parameters - I'm always happy to help!
@@roTechnic Thanks! So I'm doing something wrong. From the calculations on screen of your cycloidal drive video ua-cam.com/video/y9vLVXjz2c0/v-deo.html "Designing a cycloidal drive in Python and Fusion 360" at 2:09, I plugged the following values into the fusion 360.py code from the github link:
pin_radius = 7.5 (bc PCD 15mm), pin_circle_radius = 21 (bc PCD 42), number_of_pins = 12, and left contraction (which I think I need to eliminate altogether to get what you have above at 0:45, but I won't be surprised to be incorrect) at .2
Idk how to upload an image of the result, but suffice to say it's nothing like 0:45 above. So I then tried reversing the convex to concave by applying the changes you made on your video ua-cam.com/video/xv5cd7Bg7Uk/v-deo.html "an easier robot arm actuator..." starting at 2:06, but still get the result that you did (how odd) of convex and not what I'm trying to get at 2:02 of same video ( ua-cam.com/video/xv5cd7Bg7Uk/v-deo.html ) or BETTER, at 0:45 in the video above these comments, "What are Eccentrically Cycloidal Gears?".
Help please?
@@Upekrithen I think your problem is that you are trying to fit too many pins which are too large into too small a pin circle radius. If I take your numbers and increase the pin circle radius to 40 then it looks about right. If this doesn't help, feel free to email me on the email in the about section of my channel, you can attach pictures that way.
One other quick hint would be to prototype your numbers using "pyplot cycloid.py" which i have put in the same repo. this will show you more detail about what everything would look like, and then once you have something which looks good you can take the numbers over into the fusion add in
@@roTechnic Thank you very much. I realized that (as you said) the measurements are in centimeters, so I changed pin_radius = .125, pin_circle_radius = 1.5 number_of_pins = 11 and contraction = 0.2, and that very nearly gave me what I need for the disc. I appreciate your time and assistance.
Would efficiencies be better if it was actually a stack of bearings?
I don’t think so. in this mechanism, the pin is rolling over the disk, so there should be minimal friction. if it was sliding - like a work drive does - then bearings would definitely help.
however, I’m always up for being wrong about these things 😂 so if you disagree, please let me know
@@roTechnic I hear this sliding vs rolling friction in cyclodial gear discussions a lot, but then James Burton mentioned that going from sleeves to bearings for his pins made a big difference (although he admitted it wasn't quite a proper cyclodial gear at the time). It will be interesting to see how it holds up over time (power it and see how long it can last)
there will be friction induced by the non canonical shapes due to manufactiring tolerances, 3D printing in this case.
you showed the amount you could lift with that stepper before it started skipping steps, but without knowing how strong the stepper is by itself, that's not that useful a number
how much weight could the stepper lift if the arm was directly attached to it? How close does this come to the theoretical perfect gearing (i.e. 15:1 gear allowing it to lift 15x as much)
the stepper is rated at 42Ncm, but the torque produced is very dependant on speed (and current, driver choice etc). That's why I'm not really comparing it to a bare stepper, I'm really just looking for a total package which will give me the torque I need for an robot arm project I'm working on.
If the driver provides the maximum current for the stepper, it would deliver its maximum torque at stall. Thus one could calculate the efficiency, but I don't think that this really matters for a hobby application.
big brain
thank you!
Красиво
thank you
and now I'm trying to figure out how to cut this on manual machines
Yeah, you and me both! I can see how it could be done with a 4th axis on a CNC mill, but maybe there's a way with a helical milling attachment on a manual mill. Let me know if you come up with anything!!!
Isn't this a worm gear and wheel, just with odd parameters?
the main drawback of worm gears is that they transmit torque using sliding friction which makes backdriving more difficult as the reduction ratio increases and also has power losses as the torque increases.
EC gears on the other hand teamsmit torque using rolling friction (well, they do within the tolerances of the printer) which is markedly more efficient and allows for back driving even at high reduction ratios
@@roTechnic Is there a variant of this like a ballscrew?
@@daliasprints9798 Never seen one, and I can't visualise how it might work as ballscrews have to slide to move. However, it should be possible to make a low friction, high reduction rack and pinion with this technique to give linear motion - that could be an interesting project...
I have a question .... why aren't you making MORE VIDEOS???
This is more like a "worm gear" than an actual worm gear.
That has no torque at all.
It uses 90% of the energy if not more
I garuntee the motor holds 10lbs if not more
So a helical gear lol
I disagree with characterizing this as any sort of cycloidal drive. A cycloidal drive works by pushing a bearing into an angled surface, producing torque by rolling downhill in a circle. This video uses the geometries of cycloidal drive gears, but it rotates a circle against a gear. It is basically a 1-tooth pinion gear. It does not „roll downhill“ anywhere.
In a cycloidal gear, the force pushing the cycloidal gear parts apart at the sliding surface is converted with high efficiency to torque. In this gear arrangement, similar separating forces are present, but because of no rolling on the sliding surface, they are lost to friction.
I suspect that if you where to make a custom involute gear with 1 tooth as a helical pinion gear, it would work better than this.
Hi, I'm seeing this a year later, hope you still reply... If you extend the big wheel with blades to be driven by water flow this design could be used to generate electricity. Wouldn't that be a more efficient than current designs?
Isn't that "just" a worm gear?
No, worm gears work by sliding friction between the worm and the wheel. This works through static contact and should have minimal friction and therefore wear
i commend mechanical engineers, this makes no sense to me. which is why i went with electrical engineering 😭
It's called a worm gear, not a "single tooth" gear
it's not a worm gear, worm gears work through sliding friction which makes the whole system very inefficient. this works through rolling friction which means that the transmission is a lot more efficient than worm drives
@@roTechnic The distinction between a worm and a helical gear is that at least one tooth persists for a full rotation around the helix. If this occurs, it is a 'worm'; if not, it is a 'helical gear'. A worm may have as few as one tooth.
en.m.wikipedia.org/wiki/Gear
Words are sound and smoke.
May be you can save more space by making the driven gear as internal gear.