Experiments with Cycloidal Drives
HTML-код
- Опубликовано: 14 мар 2021
- Check out Fantom Wallet at: store.fantomwa... and use coupon code 'BRUTON' for 10% off.
Version 2: • 3D Printed Cycloidal D...
I thought it was time I did some testing for Cycloidal Drive Reducers for robotics, in particular my Robot Dogs. Belt drives have worked ok for me up to now, but I'd like to get a higher reduction in a smaller space. I'm trying to make something that is 3D printable and also reliable enough to withstand the dog stamping on the ground.
CAD for this version: github.com/XRo...
You can support me on Patreon or buy my Merchandise:
***************************
Patreon: / xrobots
Merchandise: teespring.com/...
***************************
Affiliate links - I will get some money of you use them to sign up or buy something:
***************************
Matterhackers 3D printing supplies: www.matterhacke...?aff=7500
Music for your RUclips videos: share.epidemics...
***************************
Other socials:
***************************
Instagram: / xrobotsuk
Facebook: / xrobotsuk
Twitter: / xrobotsuk
***************************
CAD and Code for my projects: github.com/XRo...
Huge thanks to my Patrons, without whom my standard of living would drastically decline. Like, inside out-Farm Foods bag decline. Plus a very special shoutout to Lulzbot, Inc who keep me in LulzBot 3D printers and support me via Patreon.
HARDWARE/SOFTWARE
Below you can also find a lot of the typical tools, equipment and supplies used in my projects:
Filament from: www.3dfuel.com/
Lulzbot 3D Printers: bit.ly/2Sj6nil
Lincoln Electric Welder: bit.ly/2Rqhqos
CNC Router: bit.ly/2QdsNjt
Ryobi Tools: bit.ly/2RhArcD
Axminster Micro Lathe: bit.ly/2Sj6eeN
3D Printer Filament: bit.ly/2PdcdUu
Soldering Iron: bit.ly/2DrNWDR
Vectric CNC Software: bit.ly/2zxpZqv
Why not join my community, who are mostly made up of actual geniuses. There’s a Facebook group and everything: / 287089964833488
XROBOTS
Former toy designer, current RUclips maker and general robotics, electrical and mechanical engineer, I’m a fan of doing it yourself and innovation by trial and error. My channel is where I share some of my useful and not-so-useful inventions, designs and maker advice. Iron Man is my go-to cosplay, and 3D printing can solve most issues - broken bolts, missing parts, world hunger, you name it.
XRobots is the community around my content where you can get in touch, share tips and advice, and more build FAQs, schematics and designs are also available. 
Наука
Version 2 of my Cycloidal Drive is coming up in a couple of weeks!
hype is real
Genuinely really looking forward to it James, I love videos like this where you're experimenting and innovating with robotics.
Crazy thing :) you are amazing ! Have you seen this guy robot arm with cycloidal motor ruclips.net/video/4IfH93HYcpo/видео.html
Why not a "normal" planetary gear?
Great
This is becoming increasingly challenging to follow. I like it VERY much.
Strange, I am following along perfectly. Though I do have problems following along with all the jargon for code, and coding programs. This is a breath of fresh air for me, compared to the other videos.
I’d like to to point out that this is not an actual cycloidal transmission, even the animation doesn’t not show it correctly. The cycloidal disc has a specific shape to it which can’t be designed with a basic cad sketch - nonlinear. It has to be calculated and interpolated (e.g. matlab to dxf to cad). I built a full metal version for a robot link myself (video..), there is no room for shape imperfections :) The actual shape will improve output speed consistency, load distribution around the circumference. In addition: Multiple disc will help to reduce load induced eccentricity and vibration..improve the efficiency (back drivability). As a hint: every roller has contact to the cycloidal disc at any position in rotation. Might seem difficult, it can be described - for “each“ position - by means of ruler and compasses though
Aforementioned video link ruclips.net/video/pzTwEX9iBq4/видео.html
Is there a technical name for this type of 'cycloidal' variant, or is it just a bastardized version that only works if the materials are somewhat compliant or have large tolerances?
@@rdyer8764 I‘m not aware of a suiting name. It somehow appears as a cycloidal without being a cycloidal :) The cycloidal itself defines a specific shape that’s created by rolling one shape in relation to an other one and tracking a point one the moving element- e.g. circle over circle or circle over a straight line.
@@antitriangular2828 That's pretty damn cool! I think that equation gave me a mathgasm! I love seeing stuff from people way smarter than I.
@@rdyer8764 Hah. Unlikey, I just happened to have been working on an anti-backlash variant of Paul Gould's work after printing his version last week.
In fact, Paul Gould includes his inventor files and the equation is embedded in the defining profile for the disks. I ended up looking up and re-doing it myself because I don't like using equations I don't understand (also - I couldn't get the ratios I wanted to work correctly, and wanted to understand why!).
Im glad to see you working on the robot dog again
I am in my 30s with absolutely no engineering background and yet I've somehow stumbled into Gear RUclips and I can't stop watching
If you optimize your teeth design you can make it contact 100% around the cycloidal disc, that's where the real benefits are with this type of drive
correct, I see no reason why this should not work more perfect with real gear tooth design instead of these round teeth. But certainly much more efforts. I have seen this only once. All other designers do it as you do (with the half circles as teeth)
I started out in the computer industry as a field service engineer in 1970 and some of the mechanical contraptions around then were unbelievable, I found electronics much easier to follow.
Here we are over 50 years later and I'm still discovering these amazing mechanical devices.
Yes yes yes, i love seeing These cycloidal bois popping up all over RUclips. I am currently working on an actuator where i integrated a 16:1 cycloidal gearbox into the stator bore of a 8108 BLDC motor to make it super compact.
excellent project 👌😎 very curious about V 2
Yes it is , ... *** family guy characters staring in your eyes ***
ruclips.net/video/tgEOpl880KM/видео.html for V 2
I don't know if I can hear it because of having young ears, but I love the high pitched sound that the motor makes when you were initially testing it - very cool.
I can definitely hear it, and it sounds awesome. If you can hear the high-pitched whine of a CRT TV, you can hear this.
This has got to be one of the best example of a 3D printed cycloidal drive already. definitely excited for two cycloidal discs and more bearings!
Thanks, it's coming up in 2 weeks
I was a vibration analyst in a previous life and these things are my worst enemies.
Thats nothing xD
I saw CNC-machines that shook the building they stood in because they accilerated and stoped so fast^^
The building was checked regulary to withstand it but it was crazy. They once tested a machine not bolted to the floor and it moved like an arms length in a few minutes.
@@benjamin4321 veritasium has a video or two on vibrations
@@JL-pc2eh Your run of the mill Haas has to be anchored to the concrete beneath because the rapids produce enough inertia to rock the machine when it comes to a stop. It is one of their huge downsides. Higher quality machines built with larger and heavier frames don't suffer from the same rigidity issues. Even if the Haas is anchored, the vibration from the lack of dampening will translate into surface finish issues on the part being machined.
industrially they are built with multiple stages out of phase to counteract the imbalances, which he mention around 10:50
@@samrevers742 Found it. ruclips.net/video/rEoc0YoALt0/видео.html
It was Steve Mould
the shift going from the small inner gear to the 10-leaf-clover was very easy to follow. well done mate
Absolutely fantastic engineering. This is the kind of video that inspires people to learn... well everything, physics, math, statics, dynamics. It's amazing what youtube brings to the world through intelligent creators like you.
I especially liked this video, James! The way you walked through the problem, iteratively introducing each part of the solution was very informative and satisfying.
A cleverly designed 2 stage belt system can be very compact and strong. The trick is to use a very small difference in cog diameter, have the second stage coaxial with the first stage, and then spin the casing that holds the cogs. So the casing becomes the input shaft, one of the cog shafts becomes the casing, and the other (coaxial) cog shaft becomes the output shaft.
Been watching Paul for months. Fantastic upgrade
Hey, I did a version of a Cycloidal Gearbox with all printing files available and a Fusion 360 script which lets you create the reduction ratio on the fly... it is backdrivable!
I am a big fan of Cycloidal Drives! Keep it up!! 😀
I got interested in these Cycloidal drives to use for a self built CNC lathe for the C-axis. It would have to be clutched to the spindle when activating C-axis. I think you could use the same motor for normal spindle drive (high speed) and mechanically shift to C axis mode (high rigidity, low speed positional control). The reason to use a cycloidal gear is that it has almost no backlash and you can get very high reduction and rigidity (if made from something like steel, not PLA, and with tight tolerances).
The point is however, on a lathe you often want a through spindle bore. The cycloidal gear would need to have a hole in the middle and I think I have found a way how to do it. Take a look at the output shafts. They basically are excenters aswell, so you could lose the excenter in the middle. Also those output shafts (that are the input as well) don't have to be inside of the cycloidal disc itself, but imagine a disc that does the exentric wiggly thingy (not actually rotating) and the cycloidal disc is "glued" onto it. Then the outer case actually becomes the driven output.
Just some ideas. First idea works well in CAD, second one I'll have to dive in to and do some designing.
I recently experimented with cycloidal drives myself because I find the principle really fascinating. I didn't do it with such a high-torque or clean design but here's two interesting things I discovered:
If you use a correctly calculated cycloidal shape with well-tuned tolerances you can *drastically* reduce the eccentricity(wobbling movement). This not only reduces the strain on the motor and other parts, it also allows for more compact designs as - like you showed in the video - the holes for transmitting the movement from the gear to the output shaft depend heavily on the eccentricity. The drawback of this is of course that tolerances become more significant.
The second thing is probably not as relevant for your project because it's more suited for high gear ratios and hurts the ability to backdrive the mechanism, but double-staged cycloidal drives can be simpler to print and build as the output is automatically a non-wobbling rotation without the need of this transmission via holes in the cycloidal disk. So it can again be made smaller. I combined those two aspects to print a tiny double-stage gear with a ~1:1000 reduction, with an eccentricity of just 0.6mm! It's not very practical but it was more of a quick feasability experiment and a lot of fun to make, and interestingly works more smoothly than my failed first try using a single stage.
super cool stuff, you are a master of design
The best revolution in James Videos!
I see what you did, there. It's a joke.
That's going to be an unending joy, once your accellerometers get a load of 4 or more of these gearboxes.
Version 2 comes up in a couple of weeks with 2 discs in
That's really neat! I'd never even heard of a gearbox like this, I just thought it was going to be a planetary gearbox. Thank you for introducing me to a different type of drive that can be 3d printed, and looks so interesting when it's running!
I knew you would get to cycloidal drives eventually. Amazing work.
CNC mill a version of it out of metal when you have a model that works like you want and use that to make it both smaller and more solid. Epic cycloidal gear by the way :D
I came here to tell him that too. I love my 3d printer but there's only so much its capable of. Gears not really one of them for this.
@@santosvella he want to make his robot 3D printables.
@@santosvella It's good if you want to do testing. One you have something that works, then you go for the milled metal parts. If that works or not, you tweak it, rinse repeat and soon you have your working device or prototype.
@@taleg1 Yup, that's what i use my 3d printer for. Wish i had a cnc though.
@@santosvella Maybe one day soon we'll get new 3D printers that are able to do multi materials that include ceramics and metals. I know we can do that now with some really expensive gear, but imagine what we could try with 3D printers that could created items of durability.
I saw this demo of a very very custom and very advanced 3D printer being researched that more or less could do any material. It could even do 3D milling. It was insane and the cost was in the millions. But we'll get there one day.
James, You Boggle my Mind more and more with every Video you Post... Your a Modern Day Genius to the power of 10 !!!!
Really interesting video as always James
Loved this. Looking forward to v2.
I am so glad you are designing and testing these types of gears. Dare I say you are putting the ole Bruton spin on it (pun intended). Looking forward to next version!!
Pretty cool to see one of these built dogs home built.
Cycloidal drive is basically a lob sided subset of planetary gear drive method. Planetary gear set so much stronger, balanced, tighter, efficient and reversible.
Fantastic video! This is a brilliant explanation of both the function of a cycloidal drive AND your thorough design and iteration process. Really like it - looking forward to seeing v2!!
2:16 you quoted stratus productions as that video source which is technically correct as that’s the url, but the channel name is rctestflight
what he said
Yip I was confused. Had no idea that was his channel url as the video links are always shortened ruclips.net/video/QVep9iSXVeo/видео.html
Weird, I was sure for second he was crediting a freebooted version of the video. The more you know
Great design
Thanks for sharing your experience😀👍
the quality of these videos is astonishing.
thats by far the best 3d printed one I've seen!
Awesome! Im so impressed it’s back-drivable
Since I recently got a 3d printer, I now fully realize just how much you are able to get done on one within a week. A lot of my prints have averaged at 24 hrs per print. But even with smaller prints, they take longer than I thought. You have several printers now to help you get more done, but even with 1 printer, you seemed to work fast.
Try 1mm nozzle and either .25mm or .4mm layer height, 45mm/s, 210-220C for PLA. Thinner layers allow steeper unsupported overhangs. Supports can add a lot of print time and material use, so design to minimize them. If you have something high up on a model that needs support, see if you can build one in that angles up at 45 degrees from a nearby surface. Much more efficient than a tower all the way from the build plate.
The best version I've seen so far!
Such a very elegant design you made, Bravo !
Im no engineer James but I am a mechanic and we use molybdenum disulfide grease for CV shafts, diff installs and I rebuild my old tools and use it on the sintered gears. My theory is because of the nature of that shit, moly grease should be perfect for a 3d printed gearbox if you use hard-wearing plastics because it'll soak up into the print lines and it sort of embeds the gears with a harder metal so they last longer. I don't understand this as I say I am a humble mechanic and I know near to nothing about 3d printing but I hope that this may help in future.
Love Owen.
What a really neat design!
Hi James, you can totally get to 10:1 reduction if you use a pulley force multiplier. Essentially if you replace a big pulley with a large bearing fix one end of the belt back on the body while putting an idler on the leg so that for 2 times the movement of the belt you will have 1 time the travel of the leg.
Wonderful stuff! You got me hooked with the balance gyro ;D
I absolutely loved this! Your channel inspired me to study engineering and this video really reminded me why I made that decision. Thank you!
It's time people switched to bearings pressed onto dowel pins. They're mass-produced precision parts, and all they need is a reamed hole in the plastic parts to receive them for repeatable interference fit
1:21 - those headless cricket dogs are terrifying
Wow you're on another level, amazing well done.
Fantastic product excellent staying power i am in awe!!
4kg weight: Mai wa mou shinderu
cam: Nani?!
You previously had issues with the motors in Opendog 2 heating the plastic, so you fitted those metal plates to dissipate the heat. Might be worth doing similar again.
Probably, although this is 10:1 and the belts were 5:1, so I get the same torque for half the power/heat.
This was very informative. It helped me with the exams. Thanks!
Since you're sticking the shaft through the center with a cam
To stop the vibrations you could mount a weight on the shaft that matches the weight of the gear and stick it 180⁰ out of phase. That way you don need a crank that has to go through two gears. Though getting them balanced perfectly would be difficult. (but you do have the sauce for that in the current design due to the cutout hole)
Awesome video. Excited for the next episode! Subbed
That’s pretty damn impressive sir.
The cycloidal drive gets it's name from the shape of the gear teeth, a cycloid. One of the big perks of this type of gear is that there is pure rolling contact between gears of this type, rather than the sliding that occurs with involute gears. This means that those outer ring bearings should be redundant, and the fact they are turning is a sign the tooth profile is not correct.
Glad to see you considering these though, maybe also look at 3D print+belt "strain wave gears", I don't personally like them because they work by fudging some of the tolerances and clearances more than the more industrial version but they seem to work.
Also, you can "invert" the drive and have the outer gear as the output, if you stationary fix those pins that go through the inner cycloidal gear.
Really? I thought it was the opposite, that involute gear teeth have theoretically perfect rolling contact, and cycloidals always have some amount of sliding. How are you generating your cycloidal curves? I've been using these equations, but I'm not sure how to analyze the exact amount of sliding that happens and/or whether there's an ideal pin radius to minimize it.
x = R * cos(t) + e * cos(N * t)
y = R * sin(t) + e * sin(N * t)
Then inset by pin radius
R = ring radius
N = number of pins
e = eccentricity
t = 0...2pi
@@dekutree64 It's been a while but I wrote an openscad script to generate cycloidal gears. Essentially you are tracking a point on the circumference of a circle as it rolls on another circle.
But it's the other way around, cycloidal gears have pure rolling contact and a variable pressure angle (goes from 90 to 0 if memory serves) which is what makes them bad for power transmission, but good for low friction and wear, so they are sometimes used in mechanical watches.
Involute gears have a constant pressure angle, which makes them good for power transmission but there is sliding contact. en.wikipedia.org/wiki/Involute_gear
There is rolling contact, but only for the instant where the line of action crosses the line connecting the two gear axles.
involute.pro/fileman/Uploads/sliding%20velocity.pdf
this also describes it in more detail
@@H34... But does generating the curve by rolling a circle mean that you have rolling contact against stationary pins as in a cycloidal drive? Seems like the pins would need to roll like the circle that you generated the curve from. i.e. the perimeter length of each disc tooth is equal to the pin circumference, so in a theoretically perfect machine where the disc maintains constant contact with all pins at all times, the pins would make one revolution for each turn of the input shaft/advance of one disc tooth. In a real machine the disc is only in contact with ~1/3 of the pins at any given time, but they should still be rolling at the same angular velocity as the input shaft during that time... unless I'm failing to visualize it properly (which is possible)
@@dekutree64 yes, you have rolling contact against a stationary pin. Imagine a fixed pin, and a pipe (loosely) sleeved over it, you can roll the pipe around the pin without sliding them against each othef and it sort of orbits and rotates, which is exactly what cycloidal drives do.
@@H34... Hmm, I guess the problem with my mental simulation was that when rolling the cylinder to create the curve, the disc is stationary, whereas in the working machine the movement of the disc is equivalent to rolling the pins in the opposite direction.
So then are the roller pins in some professionally made cycloidals just to reduce cost by allowing for lower precision machining of the discs, and to compensate for other sources of imperfection such as deformation under load, thermal expansion, wear over time, etc.?
Thanks for the help! I'll have to compare the curve generated by those equations I posted to one generated by cylinder rolling and see if there's any difference.
I love this design!
Super cool! I've been trying to make my own cyclodial drive for a tiny jumping robot for a while now.
Nice job James! I love watching your robot videos.
This seems very promising. Good job
You can use M3 of M4 bolts of 12.9 steel quality instead of M6 overkill bolts. Smaller bearings would also allow you to make the device much smaller and lighter for a better impulse response. Looking forward to see it in action!
All the bolts are M4
It also has the bonus of sounding cool as hell, come on, I saw this video and thought you were going to be making a spacecraft haha
This is so cool! I've been thinking that cycloidal drives would be perfect for OpenDog for a while, but I always figured either there was some missing feature or you'd get around to them eventually. I'm really excited to see V2!
Make the cam shaft out of metal. Instead of a straight pole off center, make a centered pole with a kink. This will also make it simpler to secure on the other end. The craftsmanship is good enough as is for the rest to hold up to higher torque.
that's really cool. like 3d printed version of a mazda rotary engine in reverse, and rotary engines stack together with offset crankshaft too.
For a similar result, but with everything running perfectly "round" have a look at "Split-Ring Epicyclic Gear" (or "differential epicyclic gearset", "compound-planet epicyclic", lots of names for the same device). It is also back-drivable (but there is more friction).
You might also consider the "Strain wave gearing" that is also compact, has massive reduction-ration and no backlash (and is supposed to be back-drivable, even if this doesn't seem obvious) ; it is 3D printable but requires one "gear" to be printed with TPU or equivalent, because is relies on deformation to transmit the motion.
Way of explanation is high level sir
Hey man great content really well explained all throughout. Also loved the modelling you did for the wallet ;) take care
Epic levels of under extrusion on those prints dude
You have a cnc router, why not use that and cut all the parts in aluminium
Even might be smaller than 3D printed
I really want to make it accessible. If I can make it 3Dprinted, then I can always upgrade later.
@@jamesbruton this is patent worthy
@@atharvbhalerao3062 No Cycloidal Drive have been in use for a weary long time. Nothing novel here.
@@jamesbruton I'm really thankful for that!
If you use prime numbers of teeth for the gears you can get the teeth to wear out more evenly.
Glad I came here looking for potentially useful comments
The sound for the cycloidal sun gear is a bit noisy, you could try to change to using Alloy 910 nylon filament with a petg core to reduce the noise and wear. With your bearing design I doubt there would be any wear with nylon.
Mounting the bearings on shoulder screws instead of bolt threads works nicely.
Properly shaped gear teeth would probably help too.
Consider, bearings without sealing with less friction probably better suited for this project. Great job!
great video and technique, but your 3D printed parts look like they have bad layer adhesion.
I think if you tweaked your printed to get perfect layer adhesion your parts would be much stronger.
Maybe use PETG and increase your heat to 220c
It's PLA at 230' and it's fine.
this is what I call a great youtuber
I can not express my happiness and relieve when I looked right and saw "Cycloidal Drive V2"
There's also a better testing video that was published a couple of weeks later.
I look forward to your next video
Thanks, the next one on this is scheduled for Tuesday 30th
Cycloidal drives done properly should have virtually zero backlash. It is their defining characteristic.
I look forward to seeing version 2. I certainly think once you sort out some of the bugs and create a good design it will work well for lots of your projects.
I thought that cycloidal gear reductions couldn't be back-driven, but I was reaching that conclusion based on OskarPuzzle's "11 million to one gearing". It's nice to be able to update my mental models.
Amazing work as always!
This is very nice!! Great job!!
i subed and liked just because this video made my day and was super wholesome
Thanks, V2 is coming up in a couple of weeks.
Me too!
This is an absolutely brilliant job. Well done 👍
Very interesting video, well edited. Thanks for that
Think this could use some M10 smooth rod:
- Ditch the exterior bearings and run the rotor directly up against the rod. Rotot on steel + silicone lube should be fairly low friction
- Same for the bearings inside the rotor: just put small smooth rods through each hole
- Finally, even the centre axle could be replaced. Then add a 3D printed cam and drill a fixing bolt through the rod
= smaller, fewer parts by ditching all the bearings and easy enough to build if you have a rotary tool with a cutting disc
You could save a lot of space by replacing the ball bearings on the ring and the output disc with plain bearings (aka bushings) running on smooth steel shafts. They can be 3d printed from nylon or graphite filled pla. Obviously this arrangement will have a shorter service life and more friction but the cost, size and weight savings may outweigh that, especially since well designed plastic plain bearings actually last a really long time.
This japanese invention used in Hajabusa blanes. Usually has two inner teeth in 180° ankle from each other to reduce resonance, and tear when there is allways atleast 2 inner teeth touching outer teeths in opposite sides. The durablity is better.
This would make a good bed shaker alarm. I'd buy one.
This looks like a nightmare child, but its a great proof of concept.
Would it be possible to use one of the holes in the motor for the center cam axis? It does extend the offset a little, but it is metal instead of plastic
While the cycloidal drive option is a really nice one, I'm personally not convinced that you couldn't reach the same goal with a planetary gear set. Regardless though.. I love the solution :)
To balance the assembly, you can put a counterweight on the motor (without needing to add a second cycloidal disc).
Have you considered making this reduction unit "turned 90 degrees" with flat discs with cycloidal grooves & ball bearings between?
You could mill an aluminium part to replace the one that broke. Some things aren't meant to be 3d printed.
You can push the output pin on brushless motors through the body, to convert them to work inboard rather than outboard.
@James Bruton
Hi fellow engineer. I just came across your channel and really enjoy what you are doing. Subscribed!
A cool mechanism! Could you just use a planetary gear out of a battery drill and use the belt drive as well for strength
2 cams, 2 counter rotating discs, no vibration, stonks!
What type of filament do I need to 3d print those belts in the beginning?